Volume 37 Number 3 September 2003 Published by The Raptor Research FoundatioFi, Inc, THE RAPTOR RESEARCH FOUNDATION, INC. (Founded 1966) OFFICERS PRESIDENT: Brian A. Millsap SECRETARY: Judith Henckel VICE-PRESIDENT: David M. Bird TREASURER:' Jim Fitzpatrick BOARD NORTH AMERICAN DIRECTOR #1: Jeff Smith NORTH AMERICAN DIRECTOR #2: Laurie J. Goodrich NORTH AMERICAN DIRECTOR #3: Ted Swem INTERNATIONAL DIRECTOR#!: Beatriz Arroyo INTERNATIONAL DIRECTOR #2: Ricardo Rodriguez-Estrella OF DIRECTORS INTERNATIONAL DIRECTOR #3: Steve Redpath DIRECTOR AT LARGE #I : Jemima ParryJones DIRECTOR AT LARGE #2: Petra Bohael Wood DIRECTOR AT LARGE #3: Michael W, Collopy DIRECTOR AT LARGE #4: Carol McIntyre DIRECTOR AT LARGE #5: Robert N. Rosenfield DIRECTOR AT LARGE #6: Daniel E. Varland EDITORIAL STAFF EDITOR: James C. Bednarz, Department of Biological Sciences, P.O. Box 599, Arkansas State University, State University, AR 72467 U.S.A. ASSISTANT EDITOR: Rebecca S. Maul ASSOCIATE EDITORS James R. Betthoff Juan Jose Negro Clint W. Boaj, Marco Restani Fabrtzio Sergio Ian G. Warkentin Joan L. Morrison BOOK REVIEW EDITOR: Jeffrey S. Marks, Montana Cooperative Research Unit, University of Montana, 'Missoula, MT 59812 U.S.A. SPANISH EDITOR: Cesar Marquez Reyes, Instituto Humboldt, Colombia, AA. 094766, Bogota 8, Colombia EDITORIAL ASSISTANTS: Jaimie Varner, Joan Clark 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, short communications, letters to the editor, 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% X 11 in.) or standard international, white, bond paper, with 25 mm (1 in.) mar- gins. The cover page should contain a title, 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 title, 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 ligure 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 (7th ed., 1998) 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 1999). 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. 36(4), and are available from the editor. Submit manuscripts to J. Bednarz at the address listed above. COVER: Red-shouldered Hawk {Buteo lineatus). Painting by James Coe. Visit www.Jamescoe.com to view more of his artwork. Contents Behavior and Prey of Nesting Red-shouldered Hawks in Southwestern Ohio. Cheryl R. Dykstra, Jeffrey L. Hays, Melinda M. Simon, and F. Bernard Daniel 1 77 Red-shouldered Hawk {Buteo lineatus) Abundance and Habitat in a Reclaimed Mine Landscape. Melissa J. Balcerzak and Petra Bohall Wood 188 Efficacy of Male Goshawk Food-delivery Calls in Broadcast Surveys on Vancouver Island. Erica L. McClaren, Patricia L. Kennedy, and Phillip L. Chapman 198 Reproductive Success, Environmental Contaminants, and Trophic Status of Nesting Bald Eagles in Eastern Newfoundland, Canada. Laura Dominguez, William Montevecchi, Neil M. Burgess, Joe Brazil, and Keith A. Hobson 209 Habitat and Nest-site Use by Red-tailed Hawks in Northwestern Wyoming. Roger N. Smith, Stanley H. Andeiwn, Steven L. Cain, and Jeffrey R. Dunk 219 Nest and Roost Habitat Characteristics of the Grey-faced Buzzard in Northeastern China. Wen-Hong Deng, Gao Wei, and Zheng Guang-Mei 228 Food Habits and Foraging Ecology of American Kestrels in the Semiarid Forests of Central Argentina. Jose Heman Sarasola, Miguel Angel Santillan, and Maximiliano Adrian Galmes 236 Short Communications Vertebrate Prey in the Diet of Flammulated Owls in Northern Utah. M. David Oleyar, Carl D, Marti, and Markus Mika 244 Nesting Habitat of Cooper’s Hawks in Northern Great Plains Woodlands. Melvin P. Nenneman, Todd A. Grant, Marriah L. Sondreal, and Robert K Murphy 246 Dynamics and Temporal Variation in Age Structure at a Communal Roost of Egyptian Vultures (Neophron i^cnopterus) in Northeastern Spain. Antoni Margalida and Jennifer Boudet 252 Diet of Three Sympatrig Owls in Steppe Habitats of Eastern Kazakhstan. Joan Navarro, Jose Antonio Sanchez-Zapata, Martina Carrete, Francisco Botella, Andrei Gavrilov, Sergei Sklyarenko, Jose Antonio Donazar, Olga Ceballos, and Fernando Hiraldo 256 Letters An Unusual Case of Adoption of a Golden Eagle (Aquila chrysaetos) Chick in the Mala Fatra Mountains, Northwestern Slovakia. Martin Koriian, Milol Majda, Metod Macek, and J^ Kornan 259 SiBUCIDE AND CANNIBALISM IN THE BOOTED EAGLE {HjERAAETUS PENNATUS) IN THE TIETAR VALLEY, Central Spain. Ignacio S. Garcia Dios 261 Book Review. Edited by Jeffrey S. Marks 262 The Raptor Research Foundation, Inc. gratefully acknowledges funds and logistical support provided by Arkansas State University to assist in the publication of the journal. THE JOURNAL OF RAPTOR RESEARCH A QUARTERLY PUBLICATION OF THE RAPTOR RESEARCH FOUNDATION, INC. VOL. 37 SEPTEMBER 2003 No. 3 J Raptor Res. 37(3) rUV-lS? © 2003 The Raptor Research Foundation, Inc. BEHAVIOR AND PREY OF NESTING RED-SHOULDERED HAWKS IN SOUTHWESTERN OHIO Cheryl R. Dykstra^ U.S. Environmental Protection Agency, National Exposure Research Laboratory, Cincinnati, OH 45268 U.S.A. Jeffrey L. Hays RAPTOR, Inc., 1586 Covered Bridge Rd., Cincinnati, OH 45231 U.S.A. Melinda M. Simon 9016 Winthrop, Cincinnati, OH 45249 U.S.A. F. Bernard Daniel U.S. Environmental Protection Agency, National Exposure Research Laboratory, Cincinnati, OH 45268 U.S.A. Abstract. — ^We used direct observations to quantify prey types, prey delivery rate, and adult and nestling behavior at nests of Red-shouldered Hawks {Buteo lineatus) in suburban southwestern Ohio. Twenty-one nests were observed for 256 hr in 1997-2001. Small mammals made up the largest percentage of the identified prey (31.5%), followed by reptiles (22.7%), invertebrates (18.8%), amphibians (17.7%), birds (6.9%), and fish (2.5%). Season-long prey delivery rate averaged 3.4 ± 0.6 prey items delivered per 4- hr observation period, orll6±19g biomass delivered per 4-hr observation period. Weekly prey delivery rate showed no correlation with the age of the nestlings {P > 0.05). Adult attendance at the nest and time adults spent brooding nestlings both were negatively correlated with nestling age (P< 0.05). Time adults spent feeding nestlings was negatively correlated with nestling age {R3 = 0.92, P = 0.002), while time nestlings spent feeding themselves was positively correlated with nestling age {R3 = 0.92, P = 0.003) . These data may serve as a baseline for assessing prey delivery rates and behavior of populations of Red-shouldered Hawks throughout the lower Midwest. Key Words: Red-shouldered Hawk, Buteo lineatus; behavior, diet, prey, raptor. CONDUCTA Y PRESAS DEL GAVILAN DE HOMBROS ROJOS DURANTE SU ANIDACION EN EL DEL SUROESTE DE OHIO Resumen. — Utilizamos observaciones directas para cuantificar las presas, su tasa de entrega y la conducta de adultos y polluelos en nidos de gavilanes de hombros rojos {Buteo lineatus) en las localidades subur- banas del suroeste de Ohio. Se observaron veintiun nidos durante 256 horas entre 1997-2001. El por- centaje mas grande de presas identificadas estuvo representado por mamiferos pequenos (31.5%), se- guido por reptiles (22.7%), invertebrados (18.8%), anfibios (17.7%), otros pajaros (6.9%) y peces (2.5%). El promedio de la tasa de entrega de presas a lo largo de la temporada fue 3.4 ± 0.6 items de presa entregados en un periodo de 4 horas de observacion, o 116 ± 19gde biomasa entregada, en un periodo de 4 horas de observacion. La tasa semanal de entrega de presa no mostro correlacion con la edad del polluelo (P > 0.05). La asistencia de los adultos en el nido y el tiempo empleado por los adultos empollando, estuvieron correlacionados negativamente con la edad del polluelo (P < 0.05). El ' Present address: 7715 Mitchell Park Dr., Cleves, OH 45002 U.S.A.; e-mail address: cheryldykstra@juno.com 177 178 Dykstra et al. VoL. 37, No. 3 tiempo que los adultos emplearon alimentando al polluelo tuvo una correlacion negativa con la edad del polluelo {R^ = 0.92, P = 0.002), mientras que el tiempo que gastaron los polluelos alimentandose a si mismos tuvo una correlacion positiva con su edad = 0.92, P = 0.003). Estos datos pueden servir como una base para valorar las tasas de entrega de presa y la conducta de poblaciones de gavilanes de hombros rojos a traves del Medio Oeste bajo. [Traduccion de Cesar Marquez] The diet of Red-shouldered Hawks {Buteo linea- tus) has been documented in several locations us- ing either observational data or prey remains in pellets or nests (Craighead and Craighead 1956, Portnoy and Dodge 1979, Bednarz and Dinsmore 1985, Howell and Chapman 1998), but only a few researchers have attempted to measure prey deliv- ery rate by adult Red-shouldered Hawks to nest- lings (Snyder and Wiley 1976, Portnoy and Dodge 1979, Penak 1982), and only one study has quan- tified and reported other behaviors of adults and nestlings during the breeding season (Portnoy and Dodge 1979). Quantitative information on food delivery rates and adult and nestling behaviors is of particular interest because productivity may be related to prey delivery rate. Prey delivery rates are usually indicative of prey availability in the environment (Newton 1979, Collopy 1984) and food supply is generally the most signihcant factor determining breeding rates for raptors (Newton 1979). Adult and nestling behavior also may be influenced by prey availability, in that time spent foraging is a function of prey availability, and nestling growth and development may be related to food provi- sioning rate (Dykstra 1995). Because of its restricted distribution due to hab- itat loss within the state of Ohio, the Red-shoul- dered Hawk is currently a “Species of Special In- terest” in the state (Ohio Department of Natural Resources pers. comm.) and is similarly classified in other midwestern and eastern states (Titus et al. 1989, Castrale 1991). Prey delivery rates and other behavioral data for the Red-shouldered Hawk may be useful in evaluating the viability and conserva- tion status of this species, particularly if they can be used as indicators of the adequacy of prey avail- ability within an environment. Our objectives were to quantify the prey delivery rate and other behaviors of adult and nestling Red- shouldered Hawks in southwestern Ohio from hatching through fledging. This information may be useful for comparison to other populations in the lower Midwest and other parts of the breeding range. Study Area and Methods Study Area. The study area in southwestern Ohio is a hilly, unglaciated area in the Interior Plateau ecoregion (Omernik 1987). The hills are dissected by many small streams located in ravines and by two large rivers, the Great Miami River and the Little Miami River. Native for- ests are dominated by second-growth oak-hickory (Quer- cus spp., Carya spp.) and beech-maple {Fagus grandifoha, Acer saccharum) associations, with lowland, riparian forests characterized by sycamores {Platanus occidentalis) and beech. Elevation ranges from ca. 140-270 m. The study area consisted of Hamilton County, Cler- mont County, and southwestern Warren County, Ohio, however, the nests studied were actually located in a wide band of suburban development surrounding the city of Cincinnati, Ohio. Suburban areas varied from densely- populated (residential lots ca. 20 X 35 m) to sparsely- populated ( >2. 5-ha residential lots, as well as undevel- oped private land) . Most residences and other buildings were surrounded by lawns and other nonnative vegeta- tion, but residences tended to be located on level ground, with steep slopes and riparian areas left in native vegetation. Areas of public land within the study area contained no residences, but were heavily used for sports and other recreation. Nests and Nest Selection. Nests in this study were se- lected from those known from a larger study of suburban Red-shouldered Hawk productivity and nest site selection (Dykstra et al. 2000a). We selected nests based on their visibility from a suitable blind location and their accessi- bility. Nests were selected for study after nestlings had hatched. We studied six nests in 1997, six nests in 1998, five nests in 1999, four nests in 2001 (Table 1). We ob- served for two or three breeding seasons at some nests (or at alternate nests within the same territory): two ter- ritories were each monitored in two years and two were each monitored in three years. Thus, we studied a total of 15 independent territories (Table 1); of these, five were located within or very near (225 m away, N = \) the floodplain of the Little Miami River, and 10 were m upland locations, primarily near small streams. Observed nests were built primarily in sycamores {N — 12; 67%), but also in red oak (Quercus rubra; N = 2), white ash {Fraxinus americana; N = 1), sugar maple {Acer saccharum; N= 1), and common locust, {Robinia pseudoacacia; N = 1), and one nest was built on the roof of a three-story apartment building in a busy complex of such buildings (Hays 2000, Dykstra et al. 2001a). Mean height of the nests was 15.2 ± 0.8 m above the ground {N = 17 mea- sured nests). Nest Observations. At most sites, nests were observed from blinds made of camouflage-colored canvas and na- tive plant materials, erected 31-85 m (x = 49 m) from the nests {N = 12). At other nests, we used temporary September 2003 Red-shouldered Hawk Behavior 179 Table 1. Observation schedules and reproduction at Red-shouldered Hawk nests studied 1997 to 2001. Nest Name W.AR Number of 4-hr Observation Blocks Age Observed ( wk) Brood Size^ Number of Young Fledged Birch Rail Run 1999 4 2-6 2 2 Cones 1998 2 4 3 0 Fields-ertel 1997 4 2-6 2 2 Gaines 2001 2 6 2 2 Lake Bern 1999 4 2-5 4 3 Mapleknoll 2001 2 4 2 2 Maryknoll 1997 2 2-3 2 0 Nisbet 1997 4 2-6 3 3 Nisbet 1998 4 3-6 3 3 Nisbet 1999 4 1-5 3 3 North Wesselman 2001 3 2-6 4 4 Pineview 1998 3 1-6 3 3 Pineview 1999 3 1-4 3 3 Roof 1998 3 2-4 4 4 Sheed 1997 2 5-6 4 4 Sheed 1998 4 2-6 3 3 Sheed 1999 4 1-5 3 3 Tanager Hills 1997 2 1-2 4 0 Twin Willow 2001 1 5 2 2 Woodsong 1997 3 1-4 3 0 Woodsong 1998 4 1-5 lb 1 ^ Brood size during most observation periods. Brood size during wk 1 may have been smaller if not all chicks had hatched, or, possibly, larger. Brood size during wk 1 was 2 or 3 young. blinds created from patio furniture {N = 3) or sat inside a vehicle parked on a driveway (N = 4). At two nests where birds were very accustomed to humans, no blinds were used. Where possible, we placed blinds upslope from the nests, improving the view of nest contents. We conducted observations during the brood-rearing peri- ods, between 1 May and 3 June 1997, 27 April and 1 June 1998, 27 April and 5 June 1999, and 4 May and 10 June 2001. Each observation period lasted 4 hr; “morning” observations were conducted between 0700 and 1300 H {N = 35), “afternoon” observations were conducted be- tween 1130 and 1800 H {N = 28), and a single obser- vation was conducted 1000—1400. We observed most nests {N = 14) 3-4 times, approximately once per wk from the time of hatching until the first nestling fledged. Because three nests failed partway through the nestling period, we were unable to complete studies on them, and we added four nests partway through the season; these nests were observed 1-2 times each (iV= 7; Table 1). We chose to observe fewer times at many nests rather than make intensive studies of a few nests in order to minimize the influence that individual birds/pairs might have on the behavior data. We monitored and recorded behaviors of nestlings and adults, with the time of occurrence and duration (nearest min) . The specific behaviors quantified were: adult atten- dance at the nest, adult prey delivery to nest, adult feed- ing nestlings, adult brooding nestlings, and nestlings feeding themselves. We defined adult attendance at the nest as the total amount of time at least one adult was present in the nest itself. Adult feeding nestlings includ- ed all the time an adult fed one or more nestlings and all the time an adult fed itself and nestlings. Brooding behavior was defined as an adult covering 50% or more of the total number of nestlings. Nestlings feeding them- selves included all time that 50% or more of the nestlings in the brood were feeding themselves. We recorded prey deliveries by adults to the nestlings. We attempted to identify the taxonomic class and the size of each prey item. Prey were categorized into one of eight classes: amphibian, reptile, fish, bird, mammal, in- sect, crustacean, or oligochaete. Further identification to species or species-group was recorded if possible. Prey were also categorized to one of four size-classes: 0—15 cm, 16—30 cm, 31—46 cm, >46 cm; smaller size categories were recorded when possible. Hatch date was estimated by back-calculating from the approximate age at banding, based on secondary length (Penak 1982), or by observation of adult behavior. The date of hatch of the oldest nestling was used when as- signing nestling ages for data analyses. We defined Day 1 as hatch date and Week 1 as 1—7 d post-hatch. Data Summary and Analyses. Total minutes spent in brooding, feeding, and adult presence at the nest were tallied, and reported as a percent of total time observed (240 min/session) . Adult and nestling behaviors were re- 180 Dykstra et al. VoL. 37, No. 3 ported as a function of approximate nestling age in weeks. At five nests, we observed twice within a single week of nestling age; behavioral data at these sites were averaged to provide a single mean value for each week. Multiple years of data from territories observed in more than one year were considered to be independent for purposes of data analyses. We observed a total of 256 hr, 64 4-hr sessions. We tallied the prey delivery rate at each nest each week (prey deliveries/nest/ 4-hr observation period), and pooled study nests by nestling age for a study-wide weekly prey delivery rate. We also calculated a season-long mean prey delivery rate for each nest. Records of prey type and size were combined for all nests and all years. In addition, we included data on prey deliveries recorded during shorter observation periods (1-3.5 hr) totaling 44 hr, at eight nests 1998-2001. (Oth- er behavioral data from these shorter observations were not included in the above analyses). Biomasses of prey items were estimated based on published data on Red- shouldered Hawk diet (Craighead and Craighead 1956, Howell and Chapman 1998) and other measures of prey Item masses (Hiles and Jones 1941, Mohr 1947, Steenhof 1983, Jayne and Bennett 1990, Sibley 2000), as well as field measurements on prey items captured within the study area (C. Dykstra and J. Hays, unpubl. data). When the species of a prey item was unknown, but the genus or genus and size were known (e.g., Microtus vole, medi- um Rana frog) , we used the mass of an appropriate spe- cies known to be present in the study area. For calcula- tion of biomass-delivery rates, unidentified prey of the 0- 15 cm size {N = 20) were assigned the mean mass for all identified prey size 0-15 cm (28 g), unidentified prey of the size 16-30 cm {N = 2) were assigned the mean mass for identified prey of size 16-30 cm (52 g), and uniden- tified prey of unknown size class {N = 24) were assigned the overall mean mass for all known prey items (36 g). Distributions of prey-size classes for identified and un- identified prey did not differ (Pearson Chi-square, P > 0 05). Nestling and adult behaviors, including prey delivery rates, were reported as a function of nestling age (wk). Weekly values for each behavior at each nest were log- transformed to account for non-normality, then averaged to produce a study-wide weekly mean value. Study-wide weekly values were compared to nestling age using linear regression. Season-long mean prey delivery rates, calculated in both numbers of prey items and estimated biomass, were log-transformed and compared to brood size using linear regression. Taxonomic classes of prey delivered at ripar- ian zone nests and upland nests were compared using a Chi-square test for independent distribution; we com- bined prey items in classes Crustacea, Insecta, and Oli- gochaeta into the category “invertebrates,” because the numbers of items in some of these classes were very small. When distribution of prey types differed, we used Bonferroni Z-test to examine which prey types were de- livered more or less often than expected at the two hab- itats: flood plains or upland (Neu et al. 1974, Byers et al. 1984). Results are presented as mean ± SE. All statistical tests were conducted using SYSTAT (Wilkinson 1988) and were considered significant at the P < 0.05 level. Results Reproductive Rate. Red-shouldered Hawks at observed nests had reproductive rates of 2.2 ± 0.3 young per nest where eggs hatched {N =21; Table 1) and 2.8 ± 0.2 young/ successful nest {N = 17), which was similar to the rate we measured for the entire study area (2.7 ± 0.2 young per successful nest, 1997-99, Dykstra et al. 2000a). Prey Deliveries. There was no difference in the number of prey items or estimated biomass deliv- ered in morning observations and in afternoon ob- servations (number of items; t — 0.73, df = 61, P = 0.47; estimated biomass: t — 0.10, df = 61, P = 0.92), so data from both observation periods were combined for analyses below. Sample sizes were in- sufficient to test for differences between years, so data from all observation years were combined. Weekly prey delivery rate showed no correlation to the age of the nestlings (number of prey items, = 0.27, P = 0.29, N = 6 wk, and estimated biomass, P2 = 0.23, P = 0.34, N = 6 wk; Fig. 1). Season- long prey delivery rate averaged 3.4 ± 0.6 prey items delivered per 4-hr observation period, or 116 ± 19 g biomass delivered per 4-hr observation pe- riod {N = 21 nests). Season-long prey delivery rate was weakly correlated to brood size when delivery rate was measured as number of prey items (P^ = 0.14, one-tailed P = 0.048, N = 21; Fig. 2); how- ever, the relationship was dependent on one da- tum (which included an unusual day when 21 items, mostly earthworms, were delivered in 4 hr) and was not significant when that datum was re- moved (P2 - 0.07, one-tailed P = 0.14, N = 20). Season-long prey delivery rate was not correlated to brood size when delivery rate was measured as estimated biomass (P^ = 0.07, one-tailed P — 0.13, 21 ). Eighty-two percent of all prey items delivered were identified to taxonomic class (203 of 249 de- liveries). Small mammals (Class Mammalia) made up the largest percentage of the identified prey, in terms of numbers of items and biomass (Table 2) Thirty-two percent (65 of 203 deliveries) of iden- tified prey items were further classified to species or genus (Table 2). The types of prey items delivered to riparian nests differed significantly from those delivered to upland nests (x^ = 22.04, df = 5, P = 0.001). Spe- cifically, riparian nests received fewer invertebrates September 2003 Red-shouldered Hawk Behavior 181 Age of Nestlings (wk) Figure 1. Prey delivery rate to nests by adult Red-shouldered Hawks as a function of nestling age. (A) Prey delivery rate measured as the number of items delivered per 4-hr observation period. (B) Prey delivery rate measured as the estimated biomass of prey delivered per 4-hr observation period. Mean ± SE. Figure 2. Prey delivery rate by adult Red-shouldered Hawks as a function of brood size. Each point represents the mean season-long prey delivery rate for a single nest. Prey delivery rate was weakly correlated with brood size (R^ = 0.14, one-tailed P = 0.048, but see text for details). 182 Dykstra et al. VoL. 37, No. 3 Table 2. Prey items delivered to Red-shouldered Hawk nests in southwestern Ohio and identified to taxonomic class or species, 1997 to 2001. Prey Types Number Identified to Species or Genus Number Identified TO CI.A.SS (%)" Biomass of Prey Identified to Class in g (%)" Class Mammalia 64 (31.5) 4084 (55.9) Vole {Microtus spp.) 12 Mice {Peromyscus spp.) 4 Chipmunk {Tamias stnatus) 5 Eastern gray squirrel {Sciurus carolinensis) 4 Eastern mole {Scalopus aquaticus) 1 Eastern cottontail {Sylvilagus floridanus) 1 Class Reptilia'^ 46 (22.7) 549 (7.5) Garter snake ( Thamnophis sirtalis) 4 Class Amphibia 36 (17.7) 1731 (23.7) Bullfrog {Rana catesbeiana) 4 Other frogs {Rana spp.) 29 Class Aves 14 (6.9) 506 (6.9) Class Osteichthyes 5 (2.5) 253 (3.4) Yellow perch {Perea jiavescens) 1 Class Oligochaeta'^ 32 (15.8) 160 (2.2) Class Crustacea^ 4 (2.0) 28 (0.4) Class Insecta 2 (1.0) 2 (<0.1) Total 65 203 7313 ^ Percent of 203 total prey items that were identified to class. Percent of 7313 total g of prey items that were identified to class. Prey consisted entirely of snakes. ^ Prey consisted entirely of unidentified earthworms. *= Prey consisted entirely of unidentified crayfish. than expected, based on the overall distribution of prey types. Adult Behavior. Adult attendance at the nest was negatively related to nestling age = 0.96, P — 0.001, N — 6 wk; Fig. 3A). Similarly, time adults spent brooding nestlings declined significantly as nestlings aged — 0.94, P — 0.001, A = 6 wk; Fig. 3B). Time adults spent feeding nestlings also was negatively correlated with nestling age — 0.92, P = 0.002, A - 6 wk; Fig. 4). Nestling Behavior. Time nestlings spent feeding themselves was positively correlated with nestling age — 0.92, P = 0.003, A = 6 wk; Fig. 4), but time nestlings spent at all feeding behaviors (being fed and feeding themselves) did not vary as nes- tlings grew {R^ = 0.07, P = 0.61, A = 6 wk). Discussion Prey Types. Diet of eastern Red-shouldered Hawks {B. 1. lineatus) has been documented in sev- eral locations using either observational data or prey remains in pellets or nests (Craighead and Craighead 1956, Portnoy and Dodge 1979, Bed- narz and Dinsmore 1985, others). For Red-shoul- dered Hawks, observational data provide a more accurate and quantitative description of diet than prey remains in pellets, due to the prevalence in the diet of highly-digestible amphibians, which fre- quently are under-represented in pellets or nest re- mains (Portnoy and Dodge 1979, Bednarz and Dinsmore 1985, Welch 1987). Diet of Red-shoul- dered Hawks varies widely from location to loca- tion, and from year to year at the same location (Bednarz and Dinsmore 1985), underscoring the ability of this species to adapt to available condi- tions. In published studies, mammals comprised 17-72% of nesting Red-shouldered Hawks’ diet (by number of prey items), amphibians 12-46%, rep- tiles 3-24%, birds 0-8%, fish 0-3%, and inverte- September 2003 Red-shouldered Hawk Behavior 183 Figure 3. (A) Adult attendance at nests of Red-shouldered Hawks as a function of nestling age. Adult attendance was negatively related with nestling age {R^ = 0.96, P = 0.001). (B) Brooding behavior of adult Red-shouldered Hawks as a function of nestling age. Time adults spent brooding nestlings declined significantly as nestlings aged {R^ = 0.94, P = 0.001). Mean ± SE. Adult Feeding -♦"Nestling Feeding Figure 4. Feeding behavior of adult and nestling Red-shouldered Hawks as a function of nestling age. Time adults spent feeding nestlings was negatively related with nestling age {R^ = 0.92, P = 0.002) and time nestlings spent feeding themselves was positively related with nestling age {R^ = 0.92, P = 0.003). Mean ± SE. 184 Dykstra et al. VoL. 37, No. 3 brates 0-26% (observational studies only: Portnoy and Dodge 1979, Penak 1982, Bednarz and Dins- more 1985, Parker 1986, Welch 1987, Howell and Chapman 1998, this study). The abundance of aquatic prey in the Red-shouldered Hawk diet can be as high as 60% of the diet (Howell and Chap- man 1998), a factor that is likely related to the pref- erence of this species for nesting near wetlands, ponds or streams (Woodrey 1986, Bosakowski et al. 1992, Moorman and Chapman 1996, Dykstra et al. 2000a, 2001b). The hawks we observed in south- western Ohio delivered more mammalian prey and fewer aquatic prey items than those observed in bottomland forests further south in Missouri and Georgia (Parker 1986, Howell and Chapman 1998). This is likely due to the primarily upland nature of our study site; even birds nesting in the floodplain of the Little Miami River had access to only a narrow corridor (150-300 m) of river and bottomland forest, surrounded by slopes and up- land habitats. The only difference in diet between upland sites and floodplain sites, fewer earthworms delivered at floodplain sites, was probably due to the paucity of suburban development within the floodplain itself. We observed that earthworms were delivered in rapid succession (10 worms in 61 mm at one nest, 10 worms in 70 min at another nest) at nest sites in lawns or adjacent to lawns, following a rain. Prey species observed and identified were gen- erally similar to those reported in other studies of Red-shouldered Hawks, consisting primarily of voles and mice, garter and other snakes, and Rana frogs, with a few exceptions. On one occasion, an adult delivered a small nest containing two uniden- tified passerine nestlings, which were removed in- dividually from the nest materials and eaten by a nestling; delivery of nests containing nestlings has been reported for Swallow- tailed Kites {Elanoides forficatus) Coulson 2001), but not for Red-shoul- dered Hawks, to our knowledge. Unusual prey spe- cies found as prey or prey remains in nests (visited in a related study) included Eastern Screech-Owl ( Otus asio) and southern flying squirrel ( Glaucomys volans). We also observed Red-shouldered Hawks feeding on an unidentified road kill within the study area, and on the carcass of a deer {Odocoileus virginianus) lying beside a road in the Hocking Hills region of south-central Ohio, about 180 km east of Cincinnati. Prey Delivery Rate, It was unclear whether Red- shouldered Hawk food delivery rates increased with increasing brood size in this study. Although the number of prey deliveries per 4-hr was weakly correlated with brood size, the estimated biomass of those deliveries was not. These data should be interpreted with caution, because our study design was such that each nest was observed for a limited amount of time (3—4 4-hr periods at most nests; Table 1), so the mean (season-long) prey delivery rate at each nest had a fair amount of variance. The mean 4-hr prey delivery rate for all Red- shouldered Hawk nests, 3.4 prey items per 4-hr or 116 g biomass per 4-hr, was somewhat higher than that measured in 220 hr of observation at eight nests in southwestern Quebec, 2.5 items per 4-hr or 91.1 g per 4-hr for broods of 1-4 nestlings (Pe- nak 1982). However, hawks in southwestern Que- bec were nesting at the northern limit of their range and had low nesting success (only 13% of the pairs observed fledged all the young that hatched; Penak 1982), so it is possible that these feeding rates were suboptimal. The 4-hr prey delivery rates we documented also may be used to estimate prey delivery rates/d (14- 15 hr daylight). Because we detected no difference between morning and afternoon delivery rates, the mean 4-hr rate is likely representative of all hours we observed, 0700-1800 H. However, prey delivery rates were likely lower in very early morning and in the evening (Portnoy and Dodge 1979), so we estimated that daily prey delivery rates were ap- proximately three times our measured 4-hr rates, or 10.2 items per d and 348 g biomass/d. This was slightly higher than rates measured at two Califor- nia nests each containing two nestlings, 19.6 g/hr or about 294 g/d (Snyder and Wiley 1976). As a caveat, we note that we were unable to identify or even classify to size a small number (10%) of the prey deliveries we observed, so our estimation of prey biomass was not exact and may be biased in some way. Prey delivery rate to nestlings may be an indi- cator of the adequacy of the prey base (Newton 1979), an important component of habitat quality for raptors. Individual raptors or populations of raptors living in marginal habitat with low prey availability may have lower prey delivery rates to their nestlings, which can result in nestling mor- tality and, hence, lower adult productivity (Dykstra et al. 1998). Lower prey delivery rates also may re- sult from aberrant parental behavior due to con- tamination of the adults by toxins, such as organ- ochlorines (Fox et al. 1978, McArthur et al. 1983, September 2003 Red-shouldered Hawk Behavior 185 Kubiak et al. 1989) . Thus, lower prey delivery rates may be an early indicator of subtle problems in a population or its environment. However, it is important to have baseline infor- mation on prey delivery rates from a productive population in order to make any assessments of other populations. In the absence of other data, nesting behaviors alone cannot indicate whether a population is viable, but we have investigated sev- eral other aspects of the ecology of Red-shouldered Hawks in southwestern Ohio. Despite their some- what atypical suburban habitat, the population does not appear to be compromised in any way that we have measured. Compared to more rural populations in south-central Ohio and elsewhere, the suburban birds reproduce at relatively high, consistent rates (Dykstra et al. 2000a, 2000b) , nest at a high nest density (Dykstra et al. 2000a) , choose suitable nest sites similar to those in rural areas (Dykstra et al. 2000a), and inhabit home ranges that are typical in size for Red-shouldered Hawks, although they are less forested than those mea- sured elsewhere (Howell and Chapman 1997, Dykstra et al. 2001a). Unfortunately, we have only limited data on some aspects of population dynam- ics such as post-fledging survival and long-term population stability. Preliminary analysis of post- fledging survival using banding data indicated that about 50% of the fledglings survived the first year (C. Dykstra, J. Hays, and M. Simon, unpubl. data), which was similar to the survival rate determined by Henny (1972) using banding data. However, in the long term, the population may be decreasing; anecdotal data suggest that the suburban birds of southwestern Ohio may have lost nesting habitat in the past 20-30 yr, as urbanization has proceeded (Dykstra et al. 2000a). Thus, with the exception of long-term population stability, which is unknown, most evidence suggests that the Red-shouldered Hawks in the southwestern Ohio study area prob- ably comprise a productive population inhabiting a suitable environment at this time and, hence, data from this population may serve as a baseline for evaluating prey delivery rates and nesting be- haviors of other populations throughout the Mid- west and elsewhere. Other Behaviors. The nestling period of Red- shouldered Hawks may be divided into three dis- tinct behavioral stages (Newton 1979), as evi- denced by our behavior data (Figs. 3 and 4) . In the early nestling stage, wk 1 and 2 after hatch, adults were present in the nest ca. 70% of the day, and much of that time was spent brooding, particularly in wk 1. Nestlings were inactive, and unable to feed themselves; feeding was done by the adults and comprised about 10% of the day. In the middle nestling stage, wk 3 and 4 after hatch, behaviors were transitional. Adult attendance declined rap- idly, as did brooding, which had nearly ceased by wk 4. There was no change in the overall amount of time spent feeding, but the time adults fed nest- lings declined to 4-8% of the day, and nestlings fed themselves for 4% of the day. In the late nest- ling stage, wk 5 and 6 after hatch, brooding had ceased and adults were present in the nest primar- ily to deliver food. Nestlings were active and fed themselves most of the time, on average about 11% of the day. Behaviors measured in our study were quite sim- ilar to those documented in central Massachusetts using a time-lapse movie camera in 1974. In that study, adult attendance rates were >80% in wk I, then declined rapidly to <5% by the end of wk 3 (Portnoy and Dodge 1979); this was consistent with our study except that southwestern Ohio birds still attended the nest 42% of the day during wk 3 and 21% during wk 4. Time adults spent feeding nestlings and nest- lings spent feeding themselves in the Massachusetts study also were similar to those recorded in this study. In wk 1 and 2, adults fed chicks for ca. 7% of the day (estimated from graphical data, Portnoy and Dodge 1979), compared to 10% of the day in southwestern Ohio. In wk 3 and 4, Massachusetts nestlings fed themselves for about 6% of the day while adults fed them for only 4% of the day (Port- noy and Dodge 1979), compared to 4% for self- feeding and 6% for adult-feeding in southwestern Ohio. In wk 5 and 6, nestlings at both locations fed themselves for about 1 1 % of the day (Portnoy and Dodge 1979, this study). It is not known whether the slightly higher rates of adult attendance and adult-feeding in wk 3 and 4 recorded in south- western Ohio were statistically significantly differ- ent from those in Massachusetts, or, if so, whether they represent real behavioral differences or are artifacts caused by differences in study designs (e.g., camera vs. direct observation, number of hours and times of day recorded) . Summary. Aberrant food delivery rates or other nesting behaviors may be an early indicator of sub- tle problems in a raptor population or its environ- ment. Measurements of prey delivery rates and nesting behaviors, as well as measurements of pro- 186 Dykstra et al. VoL. 37, No. 3 ductivity, can be used to compare data from mar- ginal populations to benchmark data from a healthy population. We measured prey delivery rates and other nesting behaviors in an apparently productive population of suburban Red-sboul- dered Hawks inhabiting a suitable environment in southwestern Ohio. These data may serve as a base- line for evaluating behaviors of other populations throughout the Midwest and elsewhere, particular- ly where the Red-shouldered Hawk is classified as a species of special concern because of habitat loss. Acknowledgments We are grateful to David Dashiells, Karen Keller, and Kelly Krebs for assistance with observations, and to Polly Dornette for help with analysis of behavioral data. Special thanks to the many landowners who gave us access to private property. This research was supported in part by RAPTOR Inc., Martin and Julie Wilz of Hamilton County, Brad and Marsha Lindner of Hamilton County, and by an appointment to the Postgraduate Research Participa- tion Program at the National Exposure Research Labo- ratory administered by the Oak Ridge Institute for Sci- ence and Education through an interagency agreement between the U.S. Department of Energy and the U.S. Environmental Protection Agency. We thank Peter Bloom and two anonymous reviewers for valuable comments that improved this paper. Literature Cited Bednarz, J.C. and J.J. Dinsmore. 1985. Flexible dietary response and feeding ecology of the Red-shouldered Hawk, Buteo lineatus, in Iowa. Can. Field-Nat. 99:262- 264. Bosakowski, T., D.G. Smith, and R. Speiser. 1992. Status, nesting density, and macrohabitat selection of Red- shouldered Hawks in northern New Jersey. Wilson Bull 104:434-446. Byers, C.R., R.K. Steinhorst, and P.R. Krausman. 1984. Clarification of a technique for analysis of utilization- availability data. /. WildL Manage. 48:1050-1053. Castrate, J.S. 1991. Eastern woodland buteos. Pages SO- SO in B.G. Pendleton, D.L. Krahe, M.N. LeFranc, Jr., K. Titus, J.C. Bednarz, D.E. Andersen, and B.A. Mill- sap [Eds.], Proceedings of the Midwest Raptor Man- agement Symposium and Workshop. Natl. Wildl. Fed., Washington, DC U.S.A. CoLLOPY, M.W. 1984. Parental care and feeding ecology of Golden Eagle nestlings. Auk 101:753-760. CouLSON,J.O. 2001. Swallow-tailed Kites carry nests con- taining passerine nestlings to their own nests. Wilson Bull 113:340-342. Craighead, JJ- AND F.C. Craighead, Jr. 1956. Hawks, owls, and wildlife. Stackpole Co., Harrisburg, PA and Wildlife Management Institute, Washington, DC U.S.A. Dykstra, C.R. 1995. Effects of contaminants, food avail- ability, and weather on the reproductive rate of Lake Superior Bald Eagles {Haliaeetus leucocephalus) . Ph.D dissertation, University of Wisconsin-Madison, Madi- son, WI U.S.A. , M.W. Meyer, D.K. Warnke, W.H. Karasov, D.E. Andersen, W.W. Bowerman, IV, and J.P. Giesy. 1998. Low reproductive rates of Lake Superior Bald Eagles: low food delivery rates or environmental contami- nants? J. Great Lakes Res. 24:32-44. , J.L. Hays, F.B. Daniel, and M.M. Simon. 2000a. Nest site selection and productivity of suburban Red- shouldered Hawks in southern Ohio. Condor 102A01- 408. , , , and . 2000b. Breeding bi- ology of suburban Red-shouldered Hawks in southern Ohio. Ohio Birds Nat. Hist. 1:137-144. , , , and . 2001a. Home range and habitat use of suburban Red-shouldered Hawks in southwestern Ohio. Wibon Bull. 113:308-316. , F.B. Daniel, J.L. Hays, and M.M. Simon. 2001b Correlation of Red-shouldered Hawk abundance and macrohabitat characteristics in southern Ohio. Condor 103:652-656. Fox, G.A., A.P. Gilman, D.R. Peakall, and F.W. Anderka 1978. Behavioral abnormalities of nesting Lake On- tario herring gulls./. Wildl Manage. 42:477-483. Hays, J.L. 2000. Red-shouldered Hawks nesting on hu- man-made structures in southwest Ohio. Pages 469- 471 in R.D. Chancellor and B.-U. Meyburg [Eds.], Raptors at risk: Proceedings of the V World Confer- ence on Birds of Prey and Owls. WWGBP, Berlin, Ger- many and Hancock House Publishers, Surrey, BC Canada. Henny, C.J. 1972. Analysis of population dynamics of se- lected avian species with special reference to changes during the modern pesticide era. U.S. Bureau of Sport Fisheries and Wildlife (Wildl. Res. Rep. No. 1) Washington, DC U.S.A. Hiles, R. and FW. Jones. 1941. Age, growth, and pro- duction of the yellow perch, Perea flavescens (Mitchill) of Saginaw Bay. Trans. Am. Fish. Soc. 70:102-122. Howell, D.L. and B.R. Chapman. 1997. Home range and habitat use of Red-shouldered Hawks in Georgia. Wil- son Bull 109:131-44. AND . 1998. Prey brought to Red-shoul- dered Hawk nests in the Georgia Piedmont. / Raptor Res. 32:257-260. Jayne, B.C. and A.F Bennett. 1990. Scaling of speed and endurance in garter snakes: a comparison of cross- sectional and longitudinal allometries. J. Zool. Lond. 220:257-277. Kubiak, T.J., HJ. Harris, L.M. Smith, T.R. Schwartz, D.L. Stalling, J.A. Trick, L. Sileo, D.E. Docherty, AND T.C. Erdman. 1989. Microcontaminants and re- productive impairment of the Forster’s Tern on Green Bay, Lake Michigan — 1983. Arch. Environ. Con- tam. Tox. 18:706-727. September 2003 Red-shouldered Hawk Behavior 187 McArthur, GA. Fox, D.B. Peakall, and BJ.R. Philogene. 1983. Ecological significance of behavior- al and hormonal abnormalities in breeding ring doves fed an organochlorine chemical mixture. Arch. Envi- ron. Contain. Tox. 12:343-353. Mohr, C.O. 1947. Table of equivalent populations of North American small mammals. Am, Midi. Nat. 37: 223-249. Moorman, C.E. and B.R. Chapman. 1996. Nest-site selec- tion of Red-shouldered and Red-tailed Hawks in a managed forest. Wilson Bull. 108:357-358. Neu, C.W., C.R. Byers, and J.M. Peek. 1974. A technique for the analysis of utilization-availability data. J. Wildl, Manage. 38:541-545. Newton, I. 1979. Population ecology of raptors. Buteo Books, Vermillion, SD U.S.A. Omernik, J.M. 1987. Ecoregions of the conterminous United States. Ann. Assoc. Am. Geographers 11 Parker, M.A. 1986. The foraging behavior and habitat use of breeding Red-shouldered Hawks in southeast- ern Missouri. M.S. thesis. University of Missouri-Co- lumbia, Columbia, MO U.S.A. Penak, B.L. 1982. Aspects of the nutritional ecology of the Red-shouldered Hawk {Buteo lineatus lineatus) in southwestern Quebec. M.S. thesis, McGill University, Ste.-Anne-de-Bellevue, Quebec, Canada. Portnoy, J.W. and W.E. Dodge. 1979. Red-shouldered Hawk nesting ecology and behavior. Wilson Bull. 91. 104-117. Sibley, D.A. 2000. The Sibly guide to birds. National Au- dubon Society and Alfred A. Knopf, Inc., New York Snyder, N.F.R. and J.W. Wiley. 1976. Sexual size dimor- phism in hawks and owls of North America. Ornithol Monogr. No. 20. American Ornithologists’ Union and Allen Press, Inc., Lawrence, KS U.S.A. Steenhof, K. 1983. Prey weights for computing percent biomass in raptor diets. Raptor Res. 17:15—27. Titus, K., M.R. Fuller, D.F. Stauffer, and J.R. Sauer 1989. Pages 53-64 in B.G. Pendleton, M.N. LeFranc, Jr., and M.B. Moss [Eds.], Buteos. Proceedings of the Northeastern Raptor Management Symposium and Workshop. Natl. Wildl. Fed., Washington, DC U.S.A. Welch, RJ. 1987. Food habits of the Red-shouldered Hawk in Wisconsin. Passenger Pigeon 49:81—92. Wilkinson, L. 1988. Systat; the system for statistics. Systat, Inc., Evanston, IL U.S.A. WOODREY, M.S. 1986. Characteristics of Red-shouldered Hawk nests in southeast Ohio. Wilson Bull. 98:466— 469. Received 7 October 2002; accepted 24 April 2003 J Raptor Res. 37(3): 188-1 97 © 2003 The Raptor Research Foundation, Inc. RED-SHOULDERED HAWK {BUTEO LINEATUS) ABUNDANCE AND HABITAT IN A RECLAIMED MINE LANDSCAPE Melissa J. Balcerzak and Petra Bohall Wood^ West Virginia Cooperative Fish and Wildlife Research Unit, U.S. Geological Survey, and Division of Forestry, West Virginia University, P.O. Box 6125, Morgantown, WV 26506-6125 U.S. A. Abstract. — Fragmentation of the landscape by large-scale mining may affect Red-shouldered Hawk {Buteo lineatus) populations by reducing the amount of forested habitat available in a landscape and by creating fragmented forest patches surrounded by reclaimed mine lands. We examined habitat char- acteristics and relative abundance of Red-shouldered Hawks in reclaimed mine landscapes within four treatments; early-successional grassland habitat, mid-successional shrub/pole habitat, late-successional fragmented forest habitat, and late-successional intact forest habitat. We quantified microhabitat char- acteristics within an 11.3-m-radius plot centered on 156 vegetation plots throughout the four treatments. We surveyed 48 stations on and adjacent to three mines for Red-shouldered Hawks using standardized broadcast call techniques during February 2000^anuary 2001 and measured landscape characteristics within 1000-m buffer zones centered on each station from digitized aerial photographs. Mean abun- dance of Red-shouldered Hawks was significantly higher in the intact forest {x = 0.07 detections/point, SE = 0.03) than the grassland {x = 0.01, SE = 0.01) treatment, but did not differ from the fragmented forest {x = 0.03, SE = 0.01) or shrub/pole {x = 0.03, SE = 0.01) treatments. Most microhabitat characteristics in both fragmented and intact forest differed from shrub/pole and grasslands. Amount of wetland was the most important characteristic determining presence of Red-shouldered Hawks in a forest-dominated landscape. More wetlands in the landscape may provide abundant amphibians and reptiles, which are important in the diet of Red-shouldered Hawks. Key Words: Buteo lineatus; Red-shouldered Hawk, habitat use, landscape, micro habitat] mining. ABUNDANCIA Y HABITAT DEL GAVILAN DE HOMBROS ROJOS {BUTEO LINEATUS) EN UN PAISAJE DE MINERIA RESTAURADO Resumen. — La fragmentacion del paisaje por la mineria a gran escala puede afectar a las poblaciones del gavilan de hombros rojos {Buteo lineatus) mediante la reduccion de la cantidad de habitat de bosque disponible en un paisaje, creando parches de bosque fragmentado rodeados por tierras de mineria restauradas. Examinamos las caracteristicas del habitat y la abundancia relativa del halcon de hombros rojos en paisajes de mineria restaurados, con cuatro tratamientos: habitat de pradera de sucesion tem- prana, habitat de arbustos de sucesion media, habitat de bosque fragmentado de sucesion tardio, y habitat de bosque intacto de sucesion tardio. Cuantificamos las caracteristicas micro habitat en una parcela de 11.3 m de radio, en 156 cuadrantes de vegetacion con los cuatro tratamientos. Inspeccio- namos 48 estaciones para el gavilan de hombros rojos dentro y en sitios adyacentes a tres minas, utili- zando tecnicas de emision de vocalizaciones estandardizadas durante febrero del 2000-enero 2001. Se midieron las caracteristicas del paisaje dentro de zonas de amortiguacion de 1000 m centradas en cada estacion determinada a partir de fotografias aereas digitalizadas. La abundancia media del gavilan de hombros rojos fue apredablemente mas alta en el bosque intacto {x = 0.07 deteccion/ punto, SE = 0.03) que en el tratamiento de la pradera {x = 0,01, SE = 0.01), pero no difirio de los tratamientos de bosque fragmentados {x = 0.03, SE = 0.01) ni del de arbustos {x = 0.03, SE = 0,01). La mayoria de las caracteristicas de los micro habitat en bosques fragmentados c intactos difirieron de los de arbustos y praderas. La existencia de humedales fue la caracteristica mas importante para determinar la presencia de gavilanes de hombros rojos en un paisaje dominado por bosque. El aumento de humedales en el paisaje podria proporcionar mayor abundancia de anfibios y reptiles, los cuales son muy importantes en la dieta del gavilan de hombros rojos. [Traduccion de Cesar Marquez] ^ Corresponding author’s e-mail address: pbwood@wvu.edu 188 September 2003 Red-shouldered Hawks and Reclaimed Mines 189 The eastern subspecies of the Red-shouldered Hawk {Buteo lineatus lineatus) is considered primar- ily a forest species that breeds in large tracts of contiguous, mature forest (Hall 1983, Buckelew and Hall 1994, Crocoll 1994), although some pop- ulations thrive in suburban landscapes (Dykstra et al. 2001). Although quantitative data are lacking, trends suggest a long-term decline in the numbers of Red-shouldered Hawks in several northern states, including New Jersey, Illinois, Iowa, New York, Michigan, and Wisconsin (Peterson and Cro- coll 1992, Crocoll 1994). A number of states attri- buted the decline primarily to reduction of closed canopy forest due to logging, agriculture, urban development, and drainage of wooded wetlands; and secondarily to competition with Red-tailed Hawks {Buteo jamaicensis) (Peterson and Crocoll 1992). Red-shouldered Hawk populations in West Virginia appear stable (Sauer et al. 2001) with con- firmed breeding in 34 out of 455 Breeding Bird Survey blocks and probable or possible breeding in 123 blocks (Buckelew and Hall 1994). Nest-site selection studies of Red-shouldered Hawks have concluded that they select large, ma- ture stands of deciduous trees that have a well de- veloped canopy layer and reduced understory e.g., (Portnoy and Dodge 1979, Titus and Mosher 1981, Bednarz and Dinsmore 1982, Morris and Lemon 1983, Bloom et al. 1993). Nest sites also have been associated with floodplain forest near areas of open water (Bent 1937, Titus and Mosher 1981, Bednarz and Dinsmore 1981, 1982, Moorman and Chapman 1996). Other studies have shown Red- shouldered Hawks to nest in upland, hardwood forests rather than floodplain forest, but still in close proximity to water (Crocoll and Parker 1989, Belleman 1998). Bednarz and Dinsmore (1982) found that Red-shouldered Hawks in Iowa used up- land forest habitat immediately adjacent to flood- plains, and they suggested that this adjacent up- land forest may be used when floodplain forest is limited in area. No previous studies have examined the effect of mountaintop mining on Red-shouldered Hawks, although the loss and fragmentation of forest and creation of edge by the mountaintop mining pro- cess may have negative effects on this species. Clearing of forests and creation of fragmented for- est patches surrounded by reclaimed grassland may decrease the suitability of mountaintop mined areas and lead to lower abundance of forest raptor populations that occupy and breed in large blocks of intact forest. In a study on 12 reclaimed surface mines in northern West Virginia and southern Pennsylvania, Mindell (1978) noted that although American Kestrels {Falco sparverius) and Red-tailed Hawks were relatively common, other raptor spe- cies such as Red-shouldered Hawks were much less common. On reclaimed surface mines in northern West Virginia, Forren (1981) found no Red-shoul- dered Hawks using artificial perches and made no mention of their presence on the surface mine at any time. In these studies, reclaimed mines were approximately 0.7-40 ha in size, while mountain- top mines typically are much larger and encompass hundreds to thousands of hectares. Creation of edge also may increase competition between Red- shouldered Hawks and Red-tailed Hawks for nest- ing areas, because forest fragmentation and large open areas favor the latter species (Bednarz and Dinsmore 1982, Preston and Beane 1993, Moor- man and Chapman 1996). To assess the impact of these relatively large-scale mountaintop mines on a forest raptor, we quantified relative abundance of Red-shouldered Hawks and examined landscape characteristics important to this species in south- ern West Virginia. Study Area We quantified Red-shouldered Hawk relative abun- dance and habitat characteristics at three mountaintop mines in the southern West Virginia counties of Boone, Logan, Kanawha, and Fayette. Reclaimed areas at each mine encompassed 2431 ha, 2180 ha, and 1819 ha. Study areas were divided into four treatments: intact forest, fragmented forest, young reclaimed mine (grassland), and older reclaimed mine (shrub/pole). The intact for- est treatment included mature forest on ridgetops, mid- slopes, and along first and second order streams and was located adjacent to the mines and within the same river drainage basin, but was undisturbed by mining activities. The fragmented forest treatment included mature forest on midslopes and along first and second order streams that were almost entirely surrounded by reclaimed mine land. Grassland areas, estimated to be 5-19 yr of age, consisted primarily of grasses and scattered shrubs Shrub/ pole areas were made up of shrub and pole-sized vegetation and ranged from 13-27 yr of age. Lack of suc- cession on some older grassland sites resulted in an over- lap in age for the latter two treatments. Wetlands were scattered throughout the grassland and shrub/ pole treat- ments, the majority were “fill” ponds located at the bot- tom of reclaimed hillsides to collect sediment due to ero- sion. The fragmented forest treatment also contained fill ponds that were located at the bottom of reclaimed grass- land hillsides adjoining the forest patch. The intact and fragmented forest areas were com- prised of mature Appalachian hardwood species typical of mixed mesophytic forest (Strausbaugh and Core 190 Balcerzak and Wood VoL. 37, No. 3 1977). Common species included red and white oak ( Quercus rubra and Q. alba) , pignut and shagbark hickory {Cary a glabra and C. ovata), yellow poplar {Liriodendron tulipifera), and American beech {Fagus grandifolia) . Pri- mary vegetation on grassland areas was tall fescue {Festuca arundinacea) , sericea {Lespedeza cuneata), and autumn ol- ive {Elaeagnus umbellatd); but also included multiflora rose (Rosa multiflora), legumes such as birdsfoot trefoil {Lotus corniculatus) and purple vetch {Vida americana) , and occasionally patches of planted wheat {Triticum aes- tivum L.). There were two distinct shrub/pole treat- ments. The first type consisted mostly of black locust {Ro- binia pseudoacada) , European black alder {Alnus glutinosa), and Rubus species. The second type included primarily coniferous species such as red pine {Pinus resi- nosa) , Scotch pine {P. sylvestris) , and pitch pine {P. rigida) . Methods We quantified Red-shouldered Hawk presence and rel- ative abundance at 48 sampling points located in 14 stream drainages with 1-6 points within each drainage. Each of the 14 stream drainages were Si 000 m apart. Within a drainage, points were spaced approximately 350-1320 m (x = 733 m) apart. Because there were no instances of a Red-shouldered Hawk being detected mul- tiple times among the points <800 m apart, we believe that lack of independence between sample points did not influence our data. Sample points were divided equally among the four treatment types with 12 points in each treatment. Because one mine lacked the shrub/pole hab- itat, the number of sample points at each of the three mines (19, 18, and 12) varied slightly. All 48 points were sampled monthly (February 2000-January 2001) over a 4— 6-d period with at least three treatments sampled on a given day to minimize temporal variability between treat- ments. A set of points representing at least three treat- ment types for each mine was assigned to be surveyed on a given day. The original order that those points were sampled was randomly established during the first survey. In subsequent surveys, the order in which points were sampled was systematically varied throughout the sam- pling day. Broadcasting conspecific vocalizations has been shown to be an effective way to survey targeted raptor species (Rosenfield et al. 1988, Mosher et al. 1990, Kennedy and Stahlecker 1993). During winter months, we conducted broadcast surveys from one-half hr after sunrise until 1600 H because raptors can be active throughout the day during cooler weather. During summer months, we con- ducted surveys from one-half hr after sunrise until 1300 H, because shifts in raptor activity in the afternoon may reduce the detectability of certain raptor species such as Red-tailed Hawks and Acdpiters (Bunn et al. 1995). Two observers trained in identification of raptors by sight and sound were present at every survey. During the survey period, both observers simultaneously watched and lis- tened for raptors. Data recorded included latency (time from start of survey until first raptor detection), raptor species detected, age and sex (if possible), behavior dur- ing detection (perch and call, flyby and call, silent perch, silent flyby, vocal only), time each individual bird was seen, estimated distance bird was from observer, and treatment in which a bird was first detected. Surveys were not conducted in inclement weather (moderate to heavy rain, fog, or wind). Broadcast surveys lasted 10 min and were based on the Fuller and Mosher (1987) protocol. We used a TOA Transistor® megaphone speaker^ (Frederick Goertz Ltd., Victoria, BC) with an attached CD player for broadcast- ing calls. One observer held the speaker 1.5 m above the ground and rotated 120° between each broadcast. Broad- cast call volume was adjusted between 100-110 db at 1 m from the speaker. Using the same broadcast volume, Mosher et al. (1990) estimated that calls could be heard ca. 0.75 km from the speaker. Before and after leaf-out, mean decibel readings at 50 m from the speaker were similar between the four treatments (Balcerzak 2001) . Vo- calizations alternated between Great-horned Owl {Bubo virginianus) and Red-shouldered Hawk calls during broadcasting. Both calls were recorded from the Peterson Field Guide to Bird Songs of Eastern and Central North America CD (Peterson Field Guide Series 1990). Mosher and Fuller (1996) found that Great-horned Owl vocali- zations were able to elicit responses from a variety of rap- tor species on a single survey. McLeod and Andersen (1998) compared conspecific and Great-horned Owl calls in attracting Red-shouldered Hawks and found that Red- shouldered Hawks responded to Great-horned Owl vo- calizations, although response rates were higher for con- specific vocalizations earlier in the season and in the day. Therefore, we used both calls within a survey period to maximize detection of a number of raptor species (Bal- cerzak 2001) as well as specifically target Red-shouldered Hawks. We randomly selected one call to start the first survey each day. We then alternated the starting call throughout the entire daily session. For landscape-level habitat analysis, each survey point was geographically referenced using a global positioning system (GPS) receiver. Using a Geographic Information System (GIS), we determined landcover from U.S. Geo- logical Survey (USGS) Digital Orthophoto Quadrangle (DOQ) images within 1000-m radius buffer zones placed around each broadcast station. The lOOO-m radius (3.14 km^, 314 ha) zone was based on the maximum home range size for Red-shouldered Hawks (Bednarz and Dins- more 1981, Crocoll 1994). Although there was some over- lap between the 1000-m buffer zones around broadcast points within a stream drainage, we completed landscape analyses for each individual point because we felt it would better reflect the habitat available to a bird detected at a given point. We classified landcover within the buffer zones into nine types: early-successional grassland habi- tat, mid-successional shrub/pole habitat (coniferous and deciduous), late-successional forest habitat (coniferous and deciduous), development, roads/bare ground, open water, and wetland. The ESRI® ArcView extension Patch Analyst 2.2 (ESRI, Redlands, CA) was used to analyze habitat composition of the 1000 m buffer zone and to calculate estimated values for 19 landscape variables (Ta- ble 1). We summarized vegetation data collected at 156 points spaced ^250 m apart within the 14 stream drainages into ^ Use of tradenames does not imply endorsement by the Federal Government. September 2003 Red-shouldered Hawks and Reclaimed Mines 191 Table 1. Abbreviations and descriptions of 19 variables used in landscape analyses of habitat use of Red-shouldered Hawks in southern West Virginia during February 2000-January 2001. Abbreviation Description DEVELOPMENT EARLY-SUCC. GRASS LATE-SUCC. CONIFER LATE-SUCC. DECIDUOUS MID-SUCC. CONIFER MID-SUCC. DECIDUOUS ROAD WATER WETLAND FOREST No. FOREST PATCHES FOREST PATCH SIZE CORE AREA CORE PATCH SIZE EDGE DENSITY DIVERSITY INDEX EVENNESS INDEX DISTANCE TO WATER DISTANCE TO TREATMENT Amount of human development (in ha). Includes residential and commer- cial buildings. Amount of early successional habitat (in ha) . Consists of mostly grasslands and scattered shrubs. Amount of late successional habitat (in ha). Consists of mature, coniferous forest, mostly Eastern hemlock ( Tsuga canadensis) . Amount of late successional habitat (in ha). Consists of mature, deciduous forest, mostly northern red oak (Quercus rubra), white oak (Q. alba), yellow poplar {Liriodendron tulipifera) , and American beech {Fagus grandifolia) . Amount of mid-successional habitat (in ha) . Consists of coniferous shrub/ pole, mostly red pine {Pinus resinosa) , Scotch pine {R sylvestris) , and pitch pine {P. rigida). Amount of mid-successional habitat (in ha). Consists of deciduous shrub/ pole, mostly black locust (Robinia pseudoacacia) , European black alder {Al- nus glutinosa) , and Rubus species. Amount of primary/ secondary roads and bare ground (in ha) . Includes all paved and most gravel/ dirt roads and any area without vegetation. Amount of open water (in ha) . Includes all streams, rivers, lakes, and ponds. Amount of emergent palustrine wetland (in ha). Amount of total mature forest (in ha). LATE-SUCC. CONIFER and LATE- SUCC. DECIDUOUS combined. Number of total FOREST patches. Mean patch size of FOREST. Total core area of FOREST in the landscape (lOO-m buffer used). Mean patch size of total core area of FOREST in the landscape. Edge density. Measure of amount of edge relative to the landscape area. Shannon’s Diversity Index. Measure of relative patch diversity in the land- scape. Shannon’s Evenness Index. Measure of patch distribution and abundance in the landscape. Distance to nearest WATER or WETLAND (in m) . Distance to next habitat/ treatment type (in m). 23 variables for use in microhabitat analyses (Table 2). Four vegetation subplots of II.3-m radius (0.04 ha) were sampled per point with one subplot centered on the point and three subplots ca. 35 m away from the center spaced 120° apart (0°, 120°, and 240°). We recorded tree species and tree diameter at breast height (DBH) >8 cm within each 11.3-m-radius subplot. Within a 5-m radius centered on each subplot, we counted shrubs, saplings, and poles taller than 0.5 m and with DBH <8 cm. We measured six categories of ground cover and six catego- ries of canopy cover using a “hit or miss” method with an ocular sighting tube at five points along four 11.3-m transects that radiated north, south, east, and west from the center of the vegetation subplot (James and Shugart 1970). Canopy cover height classes were visually estimat- ed. For each subplot, canopy height and slope were mea- sured with a clinometer, aspect was measured with a com- pass, and elevation was derived with a GPS receiver. We used analysis of variance (ANOVA) to compare mean abundance of Red-shouldered Hawks among treat- ments. Mean abundance was calculated as the mean number of detections per point per month across all treatments. In the ANOVA model, mean abundance was the dependent variable, while treatment, season, and the interaction between treatment and season were indepen- dent variables. We defined three seasons: winter (Decem- ber-March), summer (April^uly), and migration (Au- gust-November). We also used ANOVA to compare microhabitat characteristics across treatments. The ANO- VA models included the microhabitat variables as the de- pendent variable; independent variables were treatment, mine, and the interaction between treatment and mine. When differences between treatments were detected by ANOVA, we used the Waller-Duncan K-ratio t test to de- termine where those differences occurred. We used Student’s Mests to compare mean values of landscape variables between points with and without Red- shouldered Hawk detections. We also used stepwise lo- gistic regression to identify landscape characteristics im- portant to presence of Red-shouldered Hawks. Similar to 192 Balcerzak and Wood VoL. 37, No. 3 Table 2. Abbreviations and descriptions of 23 microhabitat variables compared across treatments in southern West Virginia during February 2000-January 2001. Abbreviation Description ASPECT SLOPE ELEVATION DISTANCE TO EDGE CANOPY HEIGHT SNAG Ground cover: BARE GROUND GREEN WATER LITTER MOSS WOODY DEBRIS Tree class categories: SAPLING POLE MID-SIZED TREE LARGE TREE Canopy cover: CANOPY 0-3 m CANOPY 3-6 m CANOPY 6-12 m CANOPY 12-24 m CANOPY >24 m VERTICAL STRUCTURE Aspect (degrees) taken at the center of plot. Slope (degrees) taken at the center of plot. Elevation (m) taken at center of plot. Distance to nearest microhabitat edge (m) . Includes roads, openings, and streams, or any other significant edge in forest microhabitat. Average canopy height (m). Number of snags/ dead trees. Percent of ground area covered by no vegetation. Percent of ground area covered by vegetation. Percent of ground area covered by water. Percent of ground area covered by leaf litter. Percent of ground area covered by moss. Percent of ground area covered by woody debris. Number of trees <2.5 cm DBH. Number of trees >2.5-8 cm DBH. Number of trees >8-38 cm DBH. Number of trees >38 DBH. Mean canopy cover for <0.5-3 m. Mean canopy cover for >3-6 m. Mean canopy cover for >6-12 m. Mean canopy cover for >12-18 m. Mean canopy cover for >24 m. Sum of all canopy “hits” with ocular sighting tube (min = 0, max = 120). Mitchell et al. (2001), we did not eliminate any variables a priori from the analyses because little is known about Red-shouldered Hawk habitat use in a reclaimed mine landscape. The Hosmer and Lemeshow (H/L) Good- ness-of-fit test was used to determine the ht of the logistic regression model. The entry and stay levels (a levels) for the logistic regression models were set at 0.3 and 0.1, respectively. We initially developed logistic regression models using all 48 points in all four treatments to de- termine if broad-scale features enabled Red-shouldered Hawks to persist in a reclaimed mine landscape. We then developed models for just the two forest treatments to more specifically examine features of fragmented and in- tact forest. All statistical analyses were completed using the SAS® GLM, LOGISTIC, and TTEST procedures (SAS® Insti- tute 1991). We set the significance level for all statistical tests at a = 0.1. Results From February 2000-January 2001, we recorded 30 detections of Red-shouldered Hawks on broad- cast surveys, with 19 of the 48 points having at least one detection. Of the four detections in shrub/ pole, three were aural and one was visual. In frag- mented forest, eight detections were aural, two were visual, and four were both. In intact forest, five detections were aural, two were visual, and four were both. Mean abundance of Red-shouldered Hawks was significantly different among treatments {F — 2.57, df = ?>, P — 0.053). The multiple comparison test indicated that mean abundance was significantly higher in the intact forest {x = 0.07 detections/ point, SE — 0.03) than the grassland {x — 0.01, SE = 0.01) treatment but did not differ from the frag- mented forest {x — 0.03, SE = 0.01) or shrub/pole {x = 0.03, SE = 0.01) treatments. As expected, most microhabitat characteristics in both fragmented and intact forest differed from shrub/pole and grasslands (Table 3). Comparing fragmented and intact forest, the primary differ- ences in microhabitat characteristics were greater Table 3. Means and standard errors (SE) of microhabitat characteristics in grassland {N = 44), shrub/pole {N = 33), fragmented forest {N = 36), and intact forest {N = 49). Within a row, means with different letters differ at T' < 0.05 (Waller-Duncan ^-ratio test). September 2003 JlED-SHOUI.DERED HAWKS AND ReCIAIMED MINES 193 a. GO 1 — ' 1 — < rH ^ O O CO o o GO o o o o o rH O O O O O err rH rH O rH GO o o rr o o o o o o o o o o o o o o o o o o o o o o o o o o CM o o o o o o o o o d d o o O o o o o o o o o O o O o o o V V V V V V V V V V V V V V V CD ID CD GO 00 OV CD CO 00 CO 00 no 00 CD GO 00 o> CD CD CD 00 r-H CO t-H o OD o iq vq CM rH CO cq CM CM OO d 'cH d vri d U OD no CO VO no C30 CM CM f— 1 CO CO CD CD OD CM r~ H 2 ; H 2 u s I r-1 o Ph rt § a: C/D to C/D tol C/D tol C/D to C/D GO CD OD OD Xr no C5D 1 \ o ]> cO CD o q GST iO CD CM no q 1—H CO no X^ q rH CO q d CM 1-H 1-H d d d rH d i-H d d CM O Xr. no vn CM CO GST CM CM <1 pp <1 < GO lO rH 00 t> o t> ^ cq cn lio rH GO ai CM cd GO CD CO CM ^ GO CM OD 00 GO cn ^ CM ' — ^ 00 JDr OD O CM J> GO O c6 < u pa <; CD lO 00 CO Cr rH 00 rH CD d o 00 00 ^ rH Xr CD ^ GO O O i> lO rH t-H Hf J> d CM J> CD pa < VO CO CM O CD CM O rH CD GO I — ( O rH o o d u o o d <1 u pa pa < d GO CO <1 -q < CDI CM pa pa <1 CM no x^ CO no CO X^ CD q q OD 00 GO q q CD CD ■cX d CM no X-^ CM CM X>^ no GO 00 no tX no CD no GO CD GO CM CM no CO CD Xr GO CM GO q OO no 00 q Xq GO rX GO rH d d d vd no od CM CM CM VO CD OO < < < < C < <1 t-H o CD X^ x^ O CO CD CM o CM Xq q rH CM OD q q CM h4 no T— H OD d od VD GO o CM 'cX no CO CD O'J C5D hCh GO CD CM 00 CM CM o CD OD X>; q rH q q Xq q q q q rH d d d CM od d CM CM CM OD CM pa < pa O pa pa pa pa PQ U O pa CM CD no CD CO o 00 GO o o 00 Xr CM q rH q q CO q \ — 1 q q x^ q CO CM no d d rH d no OD CM d d CM 00 x^ x~ GO CM CM t-H 00 o o 00 CM CO CM o r q rH q CO o q Xr; q o 1 1 1 rH CM d rH d d X^ od d d 1 1 1 o T ^ CM pa pa pa u pa pa pa u U pa GO GO tX CO o m CO o X' Xr rH rH q O x^ rH q q 1 1 1 Xr^ CM d 00 rH d x^ od d d 1 1 1 00 i> X-. Xr ^ to HP M <1 [/D to Q CANOPY 12-24 m — — 2.84 C 0.86 56.01 B 2.68 63.34 A 2.07 280.81 <0.0001 CANOPY 18-24 m — — 0.11 C 0.08 41.39 B 2.97 51.28 A 3.06 180.95 <0.0001 CANOPY >24 m — — 0.00 B 0.00 16.15 A 2.48 18.06 A 2.14 36.62 <0.0001 VERTICAL STRUCTURE — — 13.98 B 1.47 59.63 A 1.29 60.09 A 1.39 339.75 <0.0001 194 Balcerzak and Wood VoL. 37, No. 3 Table 4. Results of Student’s Hests showing mean and standard error (SE) of the 19 landscape variables within 1000-m buffer zones for presence/absence of Red-shouldered Hawks at 48 broadcast survey points in southern West Virginia during February 2000-January 2001. All Treatments^ Forest Treatments*’ Presence Absence Presence Absence Variable Mean SE Mean SE Mean SE Mean SE DEVELOPMENT 0.82 0.40 1.94 0.67 0.38 0.26 2.55 1.27 EARLY-SUCC. GRASS‘S 93.50 19.41 102.36 15.83 108.23 27.25 52.18 17.01 LATE-SUCC. CONIFER 19.41 4.06 16.52 3.40 23.65 5.40 17.86 4.64 LATE-SUCC. DECIDUOUS'’ 152.70 17.92 160.53 17.38 161.68 27.55 227.90 20.87 MID-SUCC. CONIFER" 3.24 2.21 2.02 1.26 0.00 0.00 0.06 0.03 MID-SUCC. DECIDUOUS 31.44 10.24 16.64 5.38 7.42 4.30 4.74 2.67 ROAD 9.10 1.51 10.47 2.10 8.68 2.27 5.99 1.99 WATER 1.79 0.45 1.54 0.25 1.70 0.44 1.01 0.37 WETLAND" 0.58 0.15 0.54 0.12 0.82 0.23 0.28 0.12 FOREST" 172.11 19.58 177.05 17.75 185.34 29.42 245.76 19.71 No. FOREST PATCHES 5.84 0.89 5.69 0.84 5.73 1.17 5.77 1.55 FOREST PATCH SIZE 61.95 20.65 80.87 20.34 83.73 34.47 140.06 39.09 CORE AREA 91.50 20.10 103.93 18.59 111.27 32.71 170.53 29.43 CORE PATCH SIZE 61.59 20.04 75.58 19.23 88.53 32.46 136.82 35.33 EDGE DENSITY 220.51 13.77 213.00 15.08 212.13 20.61 187.86 30.65 DIVERSITY INDEX 2.99 0.11 2.82 0.13 2.93 0.16 2.51 0.24 EVENNESS INDEX 0.73 0.02 0.72 0.01 0.70 0.02 0.70 0.02 DISTANCE TO WATER 378.02 102.78 407.75 76.97 437,37 173.82 600.85 149.12 DISTANCE TO TREATMENT 434.79 183.13 581.51 161.14 612.05 298.01 1034.15 318.51 ® Includes all four treatments: grassland, shrub/pole, fragmented forest, and intact forest, N = 48. Includes fragmented forest and intact forest treatments only, N = 24. Indicates significant difference between means for tbe forest treatment at a = 0.1. amounts of water, litter, and low canopy (CANOPY 0-3 m, CANOPY 3-6 m) cover in fragmented for- est. Intact forest generally had greater amounts of the higher canopy covers (CANOPY 12-18 m, CANOPY 18-24 m), although supercanopy cover (CANOPY 24 m) was similar between intact and fragmented forest. At the landscape level, no variables were re- tained as predictors of Red-shouldered Hawk pres- ence when all treatments were included in the lo- gistic regression analyses. For the fragmented and intact forest points, the amount of wetland (WET- LAND) in the landscape was the only predictor of presence (H/L Goodness-of-fit = 0.498, ^ 3.580, P = 0.059) . Additionally, wetlands occurred at 14 of the 19 points where Red-shouldered Hawks occurred, and wetland habitat was more abundant (t = —2.21, df = 22, P = 0.038) at forested points with Red-shouldered Hawk presence (Table 4). Student’s t-tests indicated that early-successional re- claimed grassland habitat (EARLY-SUCC. GRASS) was more abundant, while late-successional forest (LATE-SUCC. DECIDUOUS) , mid-successional co- niferous forest (MID-SUCC. CONIFER), and total mature forest (FOREST) were less abundant at for- ested points where Red-shouldered Hawks were present (Table 4). Discussion Red-shouldered Hawks had significantly higher abundance in the intact forest treatment than the grassland treatment, but not compared to the frag- mented forest or shrub/pole treatments. Obser- vations of birds outside the treatment at the survey point were dropped from analyses, therefore, 18 of 23 detections (nine intact forest, four fragmented forest, four shrub/pole, one grassland) were used in statistical comparisons among treatments. When all 23 detections of Red-shouldered Hawks were classified into the habitat treatment where the bird occurred: 10 were in intact forest, eight in frag- mented forest, four in shmb/pole, and one in grasslands. Many studies have shown these birds to nest primarily in contiguous, mature forest habitat September 2003 Red-shouldered Hawks and Reclaimed Mines 195 (Bednarz and Dinsmore 1981, Morris and Lemon 1983, Belleman 1998). During the study, only one Red-shouldered Hawk responded to broadcast sur- veys in fragmented forest during the breeding sea- son (Balcerzak 2001). Overall, more Red-shoul- dered Hawks responded in intact forest during the breeding season (four detections), including one instance where two birds were observed together during a survey and may have been a nesting pair. Additionally, two-three Red-shouldered Hawks (possibly juveniles) were observed at an intact for- est site in August, so breeding may have occurred in the area. Thus, it is possible that Red-shouldered Hawks may be using fragmented patches of forest for hunting or roosting, but using intact forest for nesting. Although analyses of microhabitat char- acteristics suggested few major differences between fragmented and intact forests, the amount of water as ground cover was greater in fragmented forests which may provide better foraging opportunities. Red-shouldered Hawks were detected in the shrub/pole and grassland treatment only during migration and winter periods (Balcerzak 2001). Some studies have reported greater use of more open areas and woodland edges by Red-shoul- dered Hawks during the winter months as com- pared to summer months (Bohall and Collopy 1984, Grocoll 1994). The only significant predictor of Red-shouldered Hawk presence at the landscape level was the amount of WETLAND within the 1000-m buffer zone for points within the two forest treatments. WETLAND was not significant for the overall lo- gistic regression models that included all 48 points and all treatments. Many of the ponds and wet- lands in the study area occurred within the early- successional grassland found on the reclaimed mine sites. Throughout the year of surveys, only one juvenile Red-shouldered Hawk was detected flying over the open grassland of the mine, so any significance WETLAND may have had on Red- shouldered Hawk presence for the overall land- scape regression models was mitigated by the low occurrence of Red-shouldered Hawks in the grass- lands. The significance of the variable WETLAND for points within fragmented and intact forest, where a majority of Red-shouldered Hawk obser- vations occurred, suggests that Red-shouldered Hawks are frequenting landscapes that contain a higher amount of wetlands, but that also are asso- ciated with mature forest. Amount of open water (WATER) and DISTANCE TO WATER/wetland were not significant in any of the landscape re- gression models, suggesting that these two vari- ables were not as important as the amount of pa- lustrine emergent wetland habitat present in the landscape. Wetlands have been associated with macrohabi- tat of Red-shouldered Hawks in a number of stud- ies. Both Bednarz and Dinsmore (1981) and Bo- sakowski et al. (1992) found wetlands to be significant in the landscape surrounding Red- shouldered Hawk nests, while Howell and Chap- man (1997) determined that a majority of Red- shouldered Hawk home ranges were within bottomland forest containing perennial streams, seasonal pools, and beaver ponds. Dykstra et al. (2001) found Red-shouldered Hawk response to broadcast surveys correlated with the number of small ponds within stream corridors. Red-shoul- dered Hawk association with wetlands may be re- lated to diversity and abundance of amphibian prey. A majority of prey items delivered to Red- shouldered Hawk nests were prey associated with aquatic or moist habitats (Bednarz and Dinsmore 1981, Howell and Chapman 1997). Snakes, toads, lizards, and frogs are common prey for Red-shoul- dered Hawks in the eastern U.S. (Peterson and Crocoll 1992, Crocoll 1994). Many of the ponds and wetlands associated with the reclaimed grass- land and fragmented forest in our study did con- tain populations of frogs, in particular bullfrogs {Rana catesbeiana) . Although most variables associated with the re- claimed mine present on the study sites (EARLY- SUCC. GRASS, MID-SUCC. DECIDUOUS DIS- TANCE TO TREATMENT) were not significant at a landscape scale in regression analyses, there was significantly more EARLY-SUCC. GRASS, and sig- nificantly less MID-SUCC. CONIFER, LATE-SUCC. DECIDUOUS, and FOREST at forest sites in the 1000-m buffer zone where Red-shouldered Hawks were present. This suggests that Red-shouldered Hawks can tolerate a measure of fragmentation (forest edge) within the landscape, which some studies have indicated may be important for for- aging (Bednarz and Dinsmore 1982, Moorman and Chapman 1996). Although forested points where Red-shouldered Hawks were present con- tained more grassland and less forest overall, gen- erally, multiple detections of Red-shouldered Hawks occurred at more intact forest points and few Red-shouldered Hawks were ever observed within reclaimed mine habitat, and then only dur- 196 Balcerzak and Wood VoL. 37, No. 3 ing the non-breeding season (Balcerzak 2001). The reclaimed portions of the three mines in our study were 1800-2500 ha in size, so such a large expanse of open land with little to no cover may have deterred Red-shouldered Hawks from using this habitat, regardless of the high amount of wet- lands present. Generally, the grasslands had very few perches, especially adjacent to wetlands, which could explain low use of this area by hunting Red- shouldered Hawks. Bloom et al. (1993) found that Red-shouldered Hawks in California only used non-wooded areas when perches were present. Al- though wetlands were a significant factor deter- mining Red-shouldered Hawk presence, this was only in combination with some mature forest pre- sent, indicating that the large size of the grasslands and lack of perches are likely the main reasons Red-shouldered Hawks were rarely observed using this habitat and associated wetlands. In summary, our study, similar to others, found that the amount of wetland within a landscape con- taining mature forest is an important habitat char- acteristic determining presence of Red-shouldered Hawks. Many of the wetlands found along drain- ages of forested valleys within the landscape origi- nated with fill ponds created to control soil erosion from reclaimed mine sites. Thus, it may be that Red-shouldered Hawks have taken advantage of the increased amount of wetlands created by past mining activities, but also need an undetermined amount of mature forest present. Interestingly, amount of grasslands and various measures of for- est fragmentation were not important in predicting presence, even though Red-shouldered Hawks were rarely observed in grassland and shrub/pole habitat (Balcerzak 2001). Differences in seasonal use and limited sample size may have influenced this lack of pattern. The large size of the grasslands and lack of available perches for hunting likely contributed to low Red-shouldered Hawk use of re- claimed areas. More wetlands in the landscape may provide for a large population of amphibians and reptiles, important in the diet of Red-shouldered Hawks. Use of reclaimed lands by Red-shouldered Hawks might be enhanced and accelerated by planting hardwood trees near ponds and wetlands. ACKNO WEED GMENTS This research was conducted as part of an Environ- mental Impact Assessment examining the effects of mountaintop mining on terrestrial vertebrates. Funding was provided by the West Virginia State Legislature and the Environmental Protection Agency. We also thank Arch Coal, Cannelton, and Amherst Corporation for al- lowing access to their properties. Ark Land Company provided field housing. The West Virginia Cooperative Fish and Wildlife Research Unit, BRD/USGS at West Vir- ginia University provided logistical support and field equipment. Special thanks go to the many people who helped with surveys, in particular Dorothy Tinkler and John D. Anderson, and to Mike Strager for providing help and guidance with CIS analysis. Thanks also to Rog- er J. Anderson, Cheryl Dykstra, John W. Edwards, and two anonymous reviewers for helpful comments on this manuscript. This is West Virginia Agricultural and For- estry Experiment Station scientific article no. 2847. Literature Cited Balcerzak, MJ. 2001. Raptor abundance and diversity and Red-shouldered Hawk {Buteo lineatus) habitat characteristics on reclaimed mountaintop mines in southern West Virginia. M.S. thesis, West Virginia Uni- versity, Morgantown, WV U.S.A. Bednarz, J.C. andJ.J. Dinsmore. 1981. Status, habitatuse, and management of Red-shouldered Hawks in Iowa. /. Wildl Manage. 45:236-241. AND JJ- Dinsmore. 1982. Nest-sites and habitat of Red-shouldered and Red-tailed Hawks in Iowa. Wilson Bull. 94:31-45. Belleman, B.A. 1998. Red-shouldered Hawk breeding ecology and habitat use in central Minnesota. M.S. thesis. University of Minnesota, St. Paul, MN U.S.A. Bent, A.C. 1937. Life histories of North American birds of prey. Part I. U.S. Natl. Mus. Bull. 167. Bloom, P.H., M.D. McCrary, and M.J. Gibson. 1993. Red-shouldered Hawks home-range and habitat use in southern California./. Wildl. Manage. 57:258-265. Bohall, P.G. and M.W. Collopy. 1984. Seasonal abun- dance, habitat use, and perch sites of four raptor spe- cies in north-central Florida. J. Field Ornithol. 55:181- 189. Bosakowski, T., D.G. Smith, and R. Speiser. 1992. Status, nesting density, and macrohabitat selection of Red- shouldered Hawks in northern New Jersey. Wilson Bull. 104:434-446. Buckelew, A.R. and G.A. Hall. 1994. Red-shouldered Hawk Buteo lineatus. WV breeding bird atlas, Univer- sity of Pittsburgh Press, Pittsburgh, PA U.S.A. Bunn, A.G., W. Klein, and K.L. Bildstein. 1995. Time- of-day effects on the numbers and behavior of non- breeding raptors seen on roadside surveys in eastern Pennsylvania./. Field Ornithol. 66:544-552. Crogoll, S.T. 1994. Red-shouldered Hawk {Buteo linea- tus). In A. Poole and F. Gill [Eds.], The birds of North America, No. 107, The Academy of Natural Sciences Philadelphia, PA; The American Ornithologists’ Union, Washington D.C. U.S.A. and J.W. Parker. 1989. The breeding biology of Broad-winged and Red-shouldered Hawks in western New York. / Raptor Res. 23:125-139. Dvkstra, C.R., F.B. Daniel, J.L. Hays, and M.M. Simon. September 2003 Red-shouldered Hawks and Reclaimed Mines 197 2001. Correlation of Red-shouldered Hawk abun- dance and macrohabitat characteristics in southern Ohio. Condor 103:652-656. Forren, J.D. 1981. Artificial perch use by raptors on re- claimed surface mines in West Virginia. M.S. thesis, West Virginia University, Morgantown, WV U.S.A. Fuller, M.R. andJ.A. Mosher. 1987. Raptor survey tech- niques. Pages 45-46 in B.A. Giron Pendleton, B.A. Millsap, K.W. Cline, and D.M. Bird [Eds.], Raptor management techniques manual. Natl. Wildl. Fed., Washington, DC U.S.A. Hall, G.A. 1983. West Virginia birds: distribution and ecology. Spec. Publ. Carnegie Mus. Nat. Hist. No. 7, Pittsburgh, PA U.S.A. Howell, D.L. and B.R. Chapman. 1997. Home range and habitat use of Red-shouldered Hawks in Georgia. Wil- son Bull. 109:131-144. James, F.C. and H.H. Shugart. 1970. A quantitative meth- od of habitat description. Audubon Field Notes 24:727- 736. Kennedy, PL. and D.W. Stahlecker. 1993. Responsive- ness of nesting Northern Goshawks to taped broad- casts of three conspecific calls. J. Wildl. Manage. 57: 249-257. McLeod and D.E. Andersen. 1998. Red-shouldered Hawk broadcast surveys: factors affecting detection of responses and population trends./. Wildl. Manage. 62: 1385-1397. Mindell, D.P. 1978. Habitat use by Red-tailed Hawks in surface mined areas. M.S. thesis, West Virginia Uni- versity, Morgantown, WV U.S.A. Mitchell, M.S., R.A. Lancia, and J.A. Gerwin. 2001. Us- ing landscape-level data to predict the distribution of birds on a managed forest: effects of scale. Ecol. Appl. 11:1692-1708. Moorman, C.E. and B.R. Chapman. 1996. Nest-site selec- tion of Red-shouldered and Red-tailed Hawks in a managed forest. Wilson Bull. 108:357-368. Morris, M.M.J. and R.E. Lemon. 1983. Characteristics of vegetation and topography near Red-shouldered Hawk nests in southwestern Quebec, f. Wildl. Manage 47:138-145. Mosher, J.A. and M.R. Fuller. 1996. Surveying wood- land hawks with broadcasts of Great-horned Owl vo- calizations. Wildl. Soc. Bull. 24:531-536. , M.R. Fuller, and M. Kopeny, 1990. Surveying woodland raptors by broadcast of conspecific vocali- zations./ Field Ornithol. 61:453-461. Peterson Field Guide Series. 1990. Peterson field guide to bird songs of eastern and central North America. Cornell Laboratory of Ornithology/Interactive Au- dio, Houghton Mifflin Co., Boston, MA U.S.A. Peterson, J.M.C. and S.T. Crocoll. 1992. Red-shoul- dered Hawk, Buteo lineatus. Pages 333-351 in K.J. Schneider and D.M. Pence [Eds.], Migratory non- game birds of management concern in the Northeast. U.S. Dept. Inter., Fish and Wildl. Serv., Newton Cor- ner, MA U.S.A. Portnoy, J.W. and W.E. Dodge. 1979. Red-shouldered Hawk nesting ecology and behavior. Wilson Bull. 91. 104-117. Preston, C.R. and R.D. Beane. 1993. Red-tailed Hawk (Buteo jamaicensis) . In A. Poole and F. Gill [Eds.], The birds of North America, No. 52, The Academy of Nat- ural Sciences, Philadelphia, PA; The American Orni- thologists’ Union, Washington, DC U.S.A. Rosenfield, R.N., J. Bielefeldt, and R.K. Anderson. 1988. Effectiveness of broadcast calls for detecting breeding Cooper’s Hawks. Wildl. Soc. Bull. 16:210- 212 . SAS Institute, Inc. 1991. SAS/STAT Guide for personal computers, 6th Ed. SAS Institute, Inc. Cary, NC U.S.A. Sauer, J.R., J.E. Hines, and J. Fallon. 2001. The North American breeding bird survey, results and analysis 1966-2000. Version 2001.2, USGS Patuxent Wildlife Research Center, Laurel, MD U.S.A. Strausbaugh, P.d. and E.L. Core. 1977. Flora of West Virginia. Seneca Books Inc., Morgantown, WV U.S.A. Titus, K. and J.A. Mosher. 1981. Nest-site habitat select- ed by woodland hawks in the central Appalachians. Auk 98:270-281. Received 29 May 2002; accepted 21 May 2003 J. Raptor Res. 37(3):198-208 © 2003 The Raptor Research Foundation, Inc. EFFICACY OF MALE GOSHAWK FOOD-DELIVERY CALLS IN BROADCAST SURVEYS ON VANCOUVER ISLAND Erica L. McClaren^ Department of Fishery and Wildlife Biology, Colorado State University, Fort Collins, CO 80523 US. A. Patricia L. Kennedy^ Department of Fishery and Wildlife Biology, Colorado State University, Fort Collins, CO 80523 US. A. Phillip L. Chapman Department of Statistics, Colorado State University, Fort Collins, CO 80523 US. A. Abstract. — ^We conducted broadcast experiments at occupied Northern Goshawk {Accipiter gentilis laingi) nest sites on Vancouver Island, British Columbia, during the 1998 (V = 8) and 1999 [N = 11) breeding seasons to examine the potential of the untested male food-delivery call to improve detection rates. We compared the male food-delivery call to standard vocalizations used to locate Northern Gos- hawk nests during each breeding phase. An adult alarm and male food-delivery call were broadcast during the courtship (1999 only), nestling, and fledgling-dependency breeding phases, whereas a ju- venile-begging call was broadcast only during the latter phase, when young were sufficiently developed to respond to calls. Northern Goshawks were detected at 52% {N = 88) of all broadcast trials. The male food-delivery call did not improve detection rates throughout the breeding season. Detection rates were lowest (40%) during courtship and highest (75%) during the fledgling-dependency phase. The distance we detected Northern Goshawks from nests with male food-delivery and alarm calls increased between courtship and nestling phases to the fledgling-dependency phase when the majority of detections shifted from adults to fledglings. Breeding phase did not influence the probability of detecting goshawks with male food-delivery and alarm calls. Broadcasting the juvenile-begging call within the fledgling-depen- dency phase increased the probability of detecting Northern Goshawks relative to the other two call types. The alarm and juvenile-begging calls remain the most effective for detecting Northern Goshawks on Vancouver Island during the nestling and fledgling-dependency periods, respectively. Dense coastal vegetation and rugged terrain may have interfered with our ability to project broadcast calls and to detect Northern Goshawks. The efficacy of broadcast surveys in Pacific Northwest forests during the nestling phase may be improved by spacing broadcast stations and transects at 200-m intervals, rather than the current standard of 300-m intervals, when detections occur close to nests. Broadcast stations and transects could be spaced 400-m apart during the fledgling-dependency phase when fledglings are detected farther from nest sites. Key Words: Northern Goshawk, Accipiter gentilis laingi; alarm call, broadcast surveys', juvenile-begging call, male food-delivery call, mixed models. EFICACIA DE LAS VOCALIZACIONES DE ENTREGA DE ALIMENTO DEL MACHO DE AZOR SEPTENTRIONAL EN INVESTIGACIONES HECHAS EN LA ISLA DE VANCOUVER Resumen. — Llevamos a cabo experimentos con la emision de vocalizaciones en los sitios de anidacion ocupados por el azor septentrional {Accipiter gentilis laingi) en la Isla de Vancouver, British Columbia, durante las temporadas de crianza de 1998 (V = 8) y 1999 {N = 11) con el proposito de examinar el potencial de las vocalizaciones del macho a la entrega de alimento y con el fin de mejorar las tasas de deteccion de las mismas. Comparamos la vocalizacion de entrega de alimento del macho para su estan- darizacion y para localizar nidos del azor septentrional durante cada fase de cria. Una vocalizacion de alarma del adulto y otra de entrega del alimento del macho se transmitieron durante el cortejo (solo ^ Present address: Ministry of Water, Land, and Air Protection, 2080A Labieux Rd., Nanaimo, BC V9T 6J9 Canada. ^ Corresponding author’s present address: Eastern Oregon Agricultural Research Center, Union Experiment Station, PO. Box E, 372 So. 10th, Union, OR 97883 U.S.A.; e-mail address: Pat.Kennedy@orst.edu 198 September 2003 Broadcast Surveys for Goshawks 199 en 1999), la anidacion, y las fases de apareamiento y dependencia de volanton, mientras que una vocalizacion de un juvenil pidiendo aliraento se transmitio solo durante la ultima fase, cuando los juveniles estuvieron lo suficientemente desarrollados para responder a las vocalizaciones. Los azores septentrionales se detectaron en 52% (N = 88) de todos los ensayos de emision. La vocalizacion de entrega del alimento del macho, no mejoro las tasas de deteccion a traves de la temporada de cria. Las tasas de deteccion fueron bajas (40%) durante el cortejo y altas (75%) durante la fase de dependencia del volanton. La distancia detectada desde los nidos con las vocalizaciones de entrega de alimento y de alarma, aumentaron entre el cortejo y la fase de anidacion hasta la fase de dependencia del volanton, cuando la mayoria de las detecciones cambiaron de adultos a volantones. La fase de cria no influyo en la probabilidad de discernir las llamadas de entrega de alimento y de alarma. La emision de la vocali- zacion de juveniles pidiendo alimento dentro de la fase de dependencia de volanton aumento la pro- babilidad de detectar a los parientes del azor septentrional de otros dos tipos de vocalizaciones. Las vocalizaciones de alarma y las de juveniles pidiendo alimento son las mas efectivas para detectar azores septentrionales en la Isla de Vancouver durante los periodos de cria y de dependencia del volanton, respectivamente. La densa vegetacion costera y el terreno escabroso pudieron haber interferido en nuestra habilidad de proyectar la emision de las vocalizaciones y para detectar los azores. La eficacia en las emisiones en los bosques del noroeste del Pacifico durante la fase de cria pueden mejorarse espa- ciando las estaciones de transmision y los transectos en intervalos de 200 m, en lugar de los intervales estandar actuales de 300 m, cuando las detecciones se dan cerca de los nidos. Las estaciones de emision y los transectos se podrian espaciar 400 m, durante la fase de dependencia de volanton cuando estos se detectan mas lejos de los sitios de anidacion. [Traduccion de Cesar Marquez] Techniques used to sample avian populations have come under recent scrutiny (Nichols et al. 2000, Rosenstock et al. 2002, Thompson 2002). In particular, scientists are concerned with sampling methods that generate abundance estimates that assume equal (and often 100%) detection rates across all species, age groups, habitat types, and time periods (Anderson 2001). Without incorpo- rating detection probability functions into abun- dance estimates to adjust for these factors, the re- sults produced are suspect, at best. Unbiased and accurate abundance estimates for avian popula- tions are essential to monitor changes in popula- tion abundance, elucidate avian-habitat relation- ships, and detect population responses to environmental change (Rosenstock et al. 2002, Thompson 2002). Abundance estimates for songbirds derived from point-count methodology have been the target of most sampling criticism. However, abundance es- timates derived from broadcast surveys (also re- ferred to as acoustic lure, call playback, or call re- sponse surveys) may be equally problematic because they are based on similar assumptions. Broadcast surveys are used to detect several elusive bird taxa including waterbirds (Legare et al. 1999, Erwin et al. 2002), frogmouths (Smith and Jones 1997), owls (Bosakowski and Smith 1998, Reid et al. 1999, Hardy and Morrison 2000), and hawks (Mosher and Fuller 1996, Bosakowski and Smith 1998, McLeod and Andersen 1998). Broadcast surveys of conspecific calls are the most widely used technique to detect breeding Northern Goshawks {Accipiter gentilis; hereafter re- ferred to as goshawk) throughout North America (New Mexico/Arizona: Kennedy and Stahlecker 1993, Arizona: Joy et al. 1994, Washington: Watson et al. 1999, British Columbia: McClaren 2001, Min- nesota: Roberson 2001). Through broadcast survey experiments at known, occupied nests in Arizona and New Mexico, Kennedy and Stahlecker (1993) demonstrated broadcast surveys were 89% effective at detecting breeding goshawks and their young throughout the breeding season. They showed the alarm call was most effective during the nestling phase and the juvenile-begging call was most effec- tive during the fledgling-dependency phase. Using similar experimental techniques in Washington, Watson et al. (1999) elicited 56% detection rates from breeding adults and their young. This sug- gests the effectiveness of broadcast surveys to de- tect breeding goshawks varies and may be influ- enced by habitat type, with detection rates being lower in the dense, coastal forests of the Pacific Northwest. Variable, and perhaps, low, goshawk detection rates from broadcast surveys limit our ability to dis- cern population trends and the influence of forest harvesting on breeding-habitat suitability. As a re- sult, goshawk rates of population change and hab- itat associations remain unclear and are controver- 200 McClaren et al. VoL. 37, No. 3 sial (Crocker-Bedford 1998, Kennedy 1998, Smallwood 1998). In an effort to increase goshawk detection rates through broadcast surveys, we pro- vide the first rigorous test of the male goshawk food-delivery call. This vocalization has been pho- netically described as kek. . .kek. . .kek (Penteriani 2001), guck (Schnell 1958), or chuuck (Squires and Reynolds 1997) and will be defined in this paper as the male food-delivery call. Male goshawks use this vocalization frequently throughout the year to facilitate pair contact and prey deliveries (Penteria- ni 2001). For this reason, and from our field ex- perience where adult females and fledglings seemed very responsive to males giving food-deliv- ery calls (P. Kennedy, unpubl. data), we postulated this call would enhance our ability to detect breed- ing goshawks using broadcast surveys. Our objectives were to modify Kennedy and Stahlecker’s (1993) broadcast experiment to: (1) test the effectiveness of broadcasting a male food- delivery call, against an adult alarm and juvenile- begging call, for detecting goshawks at occupied nest sites during the breeding season; and (2) pro- vide the first estimate of detection rates of the gos- hawk subspecies A. g. laingi in the dense coastal forests of Vancouver Island, British Columbia, Ca- nada. This subspecies was federally listed in 2000 as Threatened in Canada (Cooper and Chytyk 2000) and Red-listed provincially in 1993 (Ministry of Environment, Lands and Parks 2000) . Methods Study Area. Forty goshawk nest areas were located on Vancouver Island, British Columbia between 1994-99 through goshawk inventory efforts and reports by forest company personnel and the public. We conducted broad- cast experiment trials at 19 occupied nests within nest areas that were distributed widely throughout Vancouver Island. Nest sites were situated in the coastal western hemlock (CWH; Tsuga heterophylla) biogeoclimatic zone, the most productive temperate rainforest region in Ca- nada (Pojar et al. 1991). The dominant tree species were western hemlock and Douglas-fir {Pseudotsuga menziesii), although western red cedar ( Thuja plicata) , amabilis hr {Abies amabilis), and red alder {Alnus rubra) were also abundant. Vancouver Island has rugged mountains dis- sected by many creek drainages. Elevations of nest sites ranged from 150-700 m. Mean daily temperatures range from 4.1°C in winter (October-April) to 14.3°C in sum- mer (May-September) . Mean monthly precipitation ranges from 40 mm in July to 234 mm in December, with a mean annual total of 1409 mm. Most precipitation falls as rain (Environment Canada 1998). Broadcast Trials. We conducted broadcast trials at 8 and 1 1 goshawk nests, respectively, June to mid-August 1998, and April and mid-August 1999. These sample sizes reflect all known occupied (determined during court- ship) and active (determined during the nestling and fledgling-dependency phases) goshawk nests on Vancou- ver Island during these 2 yr. We considered nest areas occupied if females, radio-tagged by other investigations in 1997 and 1998, were present near nest sites or if un- tagged females were observed or heard vocalizing near nest sites. We considered nests active if nestlings or fledg- lings were observed. Our dehnitions of active and occu- pied are based on McClaren et al. (2002). In both years, we conducted trials during the nestling (June) and fledg- ling-dependency (early July to early August) stages of gos- hawk breeding phenology. In 1999 only, we also con- ducted trials during the courtship period (March to mid-April). We did not conduct broadcast trials during incubation because previous studies demonstrated fe- male raptors were less likely to respond to broadcasts dur- ing this period (Fuller and Mosher 1981, Rosenheld et al. 1988), and broadcasts may disturb incubating females and cause egg loss. Also, we only conducted broadcast trials during the initial 25 d of the fledgling-dependency period when fledglings remain within 200-300 m of nests (Kenward et al. 1993, Kennedy et al. 1994). We added active nest sites to the experiment, as they were located, and deleted nest sites when they failed {N = 2) because goshawks are less likely to remain near nests after nest failure (Kennedy and Stahlecker 1993). To prevent pseu- doreplication and habituation to broadcast calls, broad- cast trials in 1999 only occurred at nest sites that held different breeding females than in 1998. We were less concerned about habituation of breeding males to broad- cast calls because we expected most detections from the alarm call would be from females (Kennedy and Stablec- ker 1993) and because the male food-delivery and juve- nile-begging calls target adult females and fledglings. When we confirmed nests were occupied by females or were active, we established transects following the exper- imental design of Kennedy and Stahlecker (1993). We spaced eight broadcast stations at 100-m intervals along 700 m transects, using a hip chain to measure distances. Transects were offset perpendicular from nest trees by 100 m to simulate a more realistic survey situation where the probability of transects intersecting nest trees is low (Kennedy and Stahlecker 1993). Thus, the first (farthest from nest tree) and last (closest to nest tree) broadcast stations were 707 m and 100 m from the nest tree, re- spectively. We oriented transects perpendicular to slopes to minimize topographic interference with sound projec- tion. Most broadcast stations were entirely within the for- est to minimize variation in detection abilities and sound projection among habitat types. We used a commercially available adult goshawk alarm call (Western Bird Songs, Peterson Field Guides, Hough- ton Mifflin Company, Boston, MA) because we wanted our results to be comparable to other broadcast survey experiments and we wanted to simulate methodology used to locate breeding goshawks throughout North America. Trade name products are mentioned through- out the document to provide complete descriptions of methods. The authors’ institutions neither endorse these products nor intend to discriminate against products not mentioned. In the absence of commercially-available re- cordings, we used a juvenile-begging call recorded by September 2003 Broadcast Surveys for Goshawks 201 A.C. Stewart (Ministry of Sustainable Resource Manage- ment, Victoria, BC) from Vancouver Island, and a male food-delivery call recorded by M. Robinson from a cap- tive male goshawk in Waterford, Wisconsin. Broadcast calls were professionally recorded onto compact discs (CDs), background noise was removed, and recording levels were standardized so that broadcasting could occur at 100-110 dB (digital sound level meter model 33-2055: InterTan Canada Ltd., Barrie, Ontario, C-weighting [C- weighting is used to measure low frequency sounds that are >85 dB] 1 m from the audio source (Fuller and Mosher 1987) without distortion. We assumed that dif- ferences among call types were greater than regional var- iation within call types. We used a portable CD player attached via a coaxial cable to a TOA® transistor megaphone (model: SPA-603, TOA Corporation, Kobe, Japan) to broadcast calls. Calls were played from 1 -m above ground for 6 calling bouts of 10-12 sec separated by 30 sec of silence. We randomly determined the initial direction of the megaphone, and then rotated 120° to the right or left so that a full 360° was covered twice. We modified Kennedy and Stahlec- ker’s (1993) methods by offsetting a second set of calls 60° from the first to increase the area covered by broad- casts. We also altered their design by following each broadcasting period with 5 min of looking and listening, thus providing 9 min of observation at each broadcast station. This modification was recommended for Vancou- ver Island by the Resource Inventory Committee (1997) and is consistent with field observations of goshawk de- tections following shorter broadcast sessions (E. Mc- Claren, unpubl. data). During broadcast trials, E. Mc- Claren recorded all goshawk detections in both years to avoid observer bias. She purposefully and systematically looked and listened in all directions to eliminate detec- tion bias associated with prior knowledge of nest loca- tions. Broadcast trials began at station 1 and were ter- minated as soon as goshawks were detected. We avoided visiting nests after trials to prevent goshawks from asso- ciating us with broadcast calls. We broadcast adult alarm and male food-delivery calls during the courtship, nestling, and fledgling-dependency periods, whereas we broadcast the juvenile-begging call only during the fledgling-dependency period, when young were sufficiently developed to respond to this call. There were no silent walk-in controls for this experiment because broadcasting conspecific calls has been demon- strated to increase goshawk detection rates (Kennedy and Stahlecker 1993, Watson et al. 1999). Instead, we com- pared the male food-delivery call to the standard vocali- zations used to locate goshawk nests during each breed- ing phase (Arizona: Joy et al. 1994, Oregon: United States Forest Service 1994, British Columbia: Resource Inven- tory Committee 1997, Minnesota: Kennedy and Ander- sen 1999, Alaska: Titus et al. 1999) to see if it enhanced detectability. The alarm call was compared with the male food-delivery call during the courtship and nestling phas- es. In the fledgling-dependency phase we compared the male food-delivery call with the juvenile-begging call. We randomized broadcast trials at nests within each breeding stage (courtship, nestling, and fledgling-depen- dency) and within groups of nests that were geographi- cally close, enabling sampling >1 transect/ day. Once broadcast trials were initiated with one call type at a nest site, they were continued every 2 d until all call types were broadcast for that breeding phase. This design pre- vented differential detection rates from advances in chick development, and minimized temporally correlated de- tections. Trials were conducted 0800-2000 H; we did not design this experiment to test the influence of time of day on goshawk detection rates. We terminated trials m heavy rain or winds exceeding 20 km/hr. Trials inter- rupted by weather {N = 2) were repeated within 1-2 d At each broadcast station we recorded date, weather parameters (wind, cloud cover, cloud ceiling, precipita- tion, temperature), start/end time, detection type (au- ditory only, visual only, auditory and visual), detection latency (time in sec from start of broadcast session to detection), as well as gender (male, female, unknown) and age (adult, juvenile, unknown) of detected gos- hawks. We considered goshawk detections between sta- tions (N = 8 ) to be associated with the previous broadcast station. Latencies were calculated from the start of that station’s broadcast session until the time of detection. Statistical Analyses. We evaluated the success of broad- cast surveys, relative to broadcast call type and breeding phase, in 3 ways: (1) detection rates; (2) the distance of detections from occupied nests, as this influences the likelihood of locating nest sites; and (3) the probability of detecting a goshawk, with each call type, during each breeding phase. Detection latencies were also analyzed in relation to breeding phase and broadcast call type to determine the optimal amount of time a surveyor should spend at each broadcast station. Goshawk detection rates were calculated as the num- ber of goshawk detections per number of broadcast trials. We used a chi-square analysis to test for differences in detection rates among broadcast calls and breeding phas- es. However, this analysis treats broadcast trials conduct- ed at the same nest sites with the same call types in dif- ferent breeding phases independently, as other studies have done (Kennedy and Stahlecker 1993, Watson et al. 1999). We reanalyzed the data using mixed models, which incorporates the influence of these repeated mea- sures on experimental results. Failure to include repeat- ed measures in models in the following analyses could cause P-values to be inaccurate. We analyzed the influence of broadcast call type and breeding phase on detection distance (from occupied nests) and detection latency with mixed linear regression models (Littell et al. 1996). Because the experimental design was unbalanced (juvenile-begging calls were only broadcast during one breeding phase in both years, and the courtship phase was tested in only 1 yr) , analyses were performed on three data subsets (Table 1). Preliminary analyses provided no evidence that the fixed effects of year, year X broadcast call, and year X breeding phase influenced detection distance (Table lb [year: g = 0.01, P = 0.94; year X broadcast call: ij lo = 0.10, P = 0.76; year X breeding phase: G 3 = 0.60, P = 0.49]; Table Ic [year: G ^ = 1.03, P = 0.33; year X broadcast call. -^ 2,15 “ 0.31, P = 0.74]). Therefore, we pooled the 1998 and 1999 data except when the courtship phase was in- cluded. Fixed effects in mixed linear regression models included breeding phase, broadcast call type, and their interaction. Random effects were nest site and its inter- 202 McClaren et al. VoL. 37, No. 3 Table 1. Three data matrices used in mixed linear and logistic regression models to accommodate an unbalanced experimental design. Gall Type/Year Breeding Phase a) Alarm and male food-delivery calls. 1999 courtship nestling fledgling-dependency b) Alarm and male food-delivery calls. 1998, 1999 nestling fledgling-dependency c) Alarm, male food-delivery, and juvenile-begging calls. 1998, 1999 fledgling-dependency actions with broadcast call type and breeding phase. The term nest site X breeding phase accommodated the in- fluence of repeated transects at the same nest sites (using different call types) within the same breeding phase on experimental results. Likewise, the term nest site X call type incorporated the repeated measures associated with broadcasting alarm and male food-delivery calls during three breeding phases. We square-root transformed de- tection distances and latencies to decrease the hetero- geneity of variances associated with large values. We as- sessed normality assumptions of mixed linear regression models with studentized residual versus predicted plots and concluded the models in the above analyses were appropriate for the data sets. We compared least square mean detection distances and latencies of broadcast calls within and between breeding phases using pairwise t- tests. We did not perform multiple comparison adjust- ments because comparisons were chosen a priori and sample sizes were small. Therefore, we controlled com- parison-wise error rates in our analyses. The influence of broadcast call type and breeding phase on the probability of detecting a goshawk (0 = no detection, 1 = detection) was analyzed using mixed lo- gistic regression models (McCulloch and Searle 2001). We analyzed the three data subsets (Table 1) combining years for analyses that did not include the courtship phase because all year and year interactions were nonsig- nificant in the previous analyses. As with mixed linear regression models, the fixed effects included broadcast call type, breeding phase, and their interaction. Nest site, the random effect in these models, was used to incor- porate variability in detection rates caused by differences in detection probabilities among individual goshawks. Designating nest site as a random effect accommodated problems associated with repeated measures on the same nest sites (sampling the same nest sites using alarm and male food-delivery calls, during three breeding phases). We assumed random effects in mixed logistic models were normally distributed. All analyses were performed using SAS Version 7.0 (SAS Institute 1989). Results Detection Rates, Goshawks were detected on 52% of broadcast trials {N = 88). In courtship, de- tection rates were 40% for both the male food-de- livery and alarm calls (Table 2). In the nestling phase, detection rates were 60% with the alarm call and 40% with the male food-delivery call, but this difference was not statistically significant (x^ — 1.20, P = 0.27; Table 2). Detection rates were high- est (75%) in the fledgling-dependency phase when the juvenile-begging call was broadcast. However, detection rates were not significantly different among the three call types within this phase (x^ = 3.56, P = 0.17; Table 2). Goshawks were detected at least once at all but one nest site, after broadcast trials were completed. For all breeding phases and broadcast calls com- bined, 83% of goshawk detections were only audi- tory, 2% were only visual, and 15% were auditory and visual. During the courtship phase (pooled over broadcast call type), all detections were audi- tory. In the nestling phase, 53% of detections were only auditory and 40% were a combined auditory and visual detection. Detections during the fledg- ling-dependency phase were primarily auditory Table 2. Goshawk detection rates (detections/total trials) during broadcast experiments on Vancouver Island, British Golumbia, 1998-99. Broadcast Gall Breeding Phase Gourtship Nestling Fledgling-Dependen(W Gall Totals Male food-delivery 2/5 6/15 8/16 16/36 Alarm 2/5 9/15 7/16 18/36 Juvenile-begging NS^ NS 12/16 12/16 Phase totals 4/10 15/30 27/48 46/88 Not sampled. September 2003 Broadcast Surveys for Goshawks 203 Time (sec) at Broadcast Station Figure 1. Cumulative probability of detecting goshawks as a function of time spent at broadcast stations during broadcast experiment trials on Vancouver Island, British Columbia, 1998-99. Our study is compared to other stud- ies that spent 3 min at broadcast stations (Kennedy and Stahlecker 1993, Watson et al. 1999). only (96%). Similarly, detections for each broad- cast call, pooled over breeding phase, were mainly auditory only [male food-delivery call (81%), alarm call (72%), juvenile-begging call (100%)]. Only 39% of detections during broadcast trials were of adult goshawks. During courtship, only adults were available for detection. In the nestling phase, 80% of detections were of adults whereas in the fledgling-dependency phase, 93% of detections were of fledglings. Pooled over breeding phase, the male food-delivery call primarily generated fledgling detections (62%), whereas the alarm call primarily generated adult detections (67%) . All de- tections from the juvenile-begging call were of fledglings. Detection Latency. We recorded 63% of detec- tions at broadcast stations within 3 min of initiating broadcast calls and 90% of detections within 5 min of initiating broadcast calls (Fig. 1). In other words, 63% of detections occurred after we broadcast three sets of calls and 90% of detections occurred after we broadcast six sets of calls plus a 1 min listening period. Mean detection latencies did not significantly differ between call types within the nestling and fledgling-dependency phases (all pair- wise comparison P-values > 0.05) . Detection Distance from Occupied Nests. All de- tections during the courtship and nestling phases with the male food-delivery call were 141 m from nests (Figs. 2a, 2b). However, in the fledgling-de- pendency phase we detected goshawks with the male food-delivery call throughout transect dis- tances and as far as 707 m from nests (Fig. 2c). From courtship through fledgling-dependency. 6 1 ■ II ■ Ual? »or<- iXlVPrv 100 141 224 316 412 510 608 707 Distance from Nest (m) (a) Courtship >. u c cr £ c o ■ AUrtr ■ UaM Fcui cMOvor/ 100 141 224 316 412 510 608 707 Distance from Neat (m) (b) Nestling 6 S' 5 c * & ^ LL 3 I : ■ AJem ■ FooO- ijHnwry r. Ju-rfnrttM-lMiggt'tg dLiilii ]L 100 141 224 316 412 510 608 707 DIatance from Nest (m) (c) Fledgling-dependency Figure 2. Frequency of goshawk detections as a function of distance from occupied nests, Vancouver Island, Brit- ish Columbia, 1998-99. Graphs are presented for the (a) courtship, (b) nestling, and (c) fledgling-dependency phases. goshawks were detected with alarm calls at pro- gressively farther distances from nests. However, greater than 70% of alarm call detections were within 316 m of nests during all breeding phases. With the juvenile-begging call, goshawks were de- tected 141-707 m from nests but most frequently they were detected at 316 m from nests (Fig. 2c). Breeding phase influenced the distance we de- tected goshawks from occupied nests (Fig. 3) . The most dramatic pattern we observed was with the male food-delivery call. The mean distance we de- tected goshawks with the male food-delivery call increased from the courtship (t = 3.07, P = 0.01) and nestling (t = 3.64, P = 0.003) phases to the fledgling-dependency phase (Fig. 3). Mean detec- tion distances were similar (p 2 ,iv ~ 0.79, P = 0.47; 204 McCiaren et al. VoL. 37, No. 3 600 -I 500 - 400 300 200 - 100 - - - Male Food-deli\ery — ■ — Alarm A Ju\«nile-begglng 0 J Courtship Nestling Fledgling- de pendency Breeding Phase Figure 3. Mean ± SE goshawk detection distance (m) from occupied nests during broadcast experiment trials on Vancouver Island, British Columbia, 1998-99. Fig. 3) for the three calls in the fledgling-depen- dency phase. Detection Probability. The probability of detect- ing goshawks did not significantly depend on which call type was broadcast within the courtship and nestling phases {t = —0.28, P = 0,78, N ~ 56), nor did detection probabilities change for a given call type between breeding phases {t = 0.73, P = 0.47, N — 56). However, in the fledgling-depen- dency phase, broadcasting the juvenile-begging call increased the probability of detecting goshawks rel- ative to alarm and male food-delivery calls {t = 1.97, P = 0.07, N= 48). Discussion Efficacy of the Male Food-delivery Call. The probability of detecting goshawks or their young was not increased by broadcasting the male food- delivery call during the courtship, nestling, or fledgling-dependency breeding phases, relative to the standard alarm (courtship, nestling) and juve- nile begging (fledgling-dependency) calls. Al- though not statistically significant, detection rates were higher when alarm and juvenile-begging calls were broadcast during the nestling and fledgling- dependency phases, respectively, relative to broad- casts of the male food-delivery call. Non-statistical differences in our detection rates may reflect small sample sizes as a result of relatively low breeding densities of goshawks on Vancouver Island. How- ever, the 20% and 25% difference in detection rates we observed between the male food-delivery call and alarm and juvenile-begging calls during the nestling and fledgling-dependency phases, re- spectively, may reflect biologically meaningful dif- ferences. Thus, we recommend the continued use of standard calls until a more effective call type is identified. The male food-delivery call may not be as effec- tive as alarm and juvenile-begging calls in broad- cast surveys because it is naturally a call of low pitch and volume that is given by the male when he is delivering food to the nest (Schnell 1958, Squires and Reynolds 1997, Penteriani 2001). Con- sequently, broadcasting this call at 100-110 dB may be unrealistic and may alter the call’s identity. The male’s physical presence in the nest stand, as well as his food-delivery call, may stimulate the re- sponse. Because this call is used for pair contact, goshawks may utilize individual variation in this call to recognize their mates compared to the alarm and juvenile-begging calls which have more generalized usage. Therefore, when we broadcast a recording from Wisconsin on Vancouver Island, females may have been less responsive to our re- cording. However, Roberson (2001) later tested the same recording of the male food-delivery call in Minnesota and reported lower detection rates than our study. This suggests that our results are not an artifact of dialect. Geographic Variation in Detection Rates. Over- all, detection rates for alarm and juvenile-begging calls in this study were lower than those reported by Kennedy and Stahlecker (1993). Kimmel and Yahner (1990) and Watson et al. (1999) also re- ported lower detection rates than Kennedy and Stahlecker (1993). For example, Kennedy and Stahlecker (1993) reported detection rates of 93% with the alarm call during the nestling phase, com- pared to 60% (this study), 37% (Watson et al. 1999), and 48% (Kimmel and Yahner 1990). Ken- nedy and Stahlecker (1993) also reported higher detection rates during the fledgling-dependency phase with the juvenile-begging call (85%) than re- corded in this experiment (75%) and by Watson et al. (1999; 74%). Regional variation in goshawk detection rates reinforces that local detection probability functions should be incorporated when broadcast data are used to monitor changes in population abundance, elucidate goshawk-habitat relationships, and detect population responses to environmental change. Lower detection rates in the Pacific Northwest compared to the southwestern United States sug- September 2003 Broadcast Surveys for Goshayvtcs 205 gest that transmission of broadcast calls in coastal forests may be hindered by vegetation and topog- raphy. These factors may also reduce an observer’s ability to detect goshawks, visually and aurally, in coastal forests. Other studies have also expressed, but have not documented experimentally, con- cerns regarding lower goshawk detection rates with broadcast surveys in coastal forests (southeast Alas- ka: Iverson et al. 1996; western Oregon: DeStefano and McCloskey 1997; western Washington: Bosa- kowski and Vaughn 1996). A large body of litera- ture from songbird broadcast experiments docu- ments the scattering of sound by reflective surfaces such as foliage and tree trunks (Fotheringham and Ratcliffe 1995, Brown and Handford 2000). Tree density within goshawk nest areas on Vancouver Is- land (Ethier 1999) is higher than in New Mexico (Siders and Kennedy 1996) which may degrade broadcast calls. Many songbirds use sound degra- dation to gauge the distance of an intruder from their territory (Fotheringham and Ratcliffe 1995, Holland et al. 1998). Similarly, goshawks may gauge the distance of broadcast calls and if calls appear far away, they may be less responsive. Survey Design and the Probability of Detecting a Goshawk. It is important to streamline broadcast surveys so that they occur when they are most ef- fective. Depending on the objective of broadcast surveys, efficacy will be measured by number of detections, number of occupied nests located or both. Breeding phase, call type, distance between broadcast stations and transect lines, and the amount of time spent at each broadcast station will influence the success of broadcast surveys and the amount of time, effort and money expended. Breeding phase and call type. Detection rates with alarm and male food-delivery calls were similar be- tween the nestling and the fledgling-dependency breeding phases. Our results were similar to con- clusions made by Kimmel and Yahner (1990) who broadcast goshawk alarm calls during the nestling and fledgling phases. Conversely, breeding phase influenced the probability of detecting goshawks throughout the breeding season in experiments conducted by Kennedy and Stahlecker (1993). However, they compared differences in combined detection rates from alarm and wail calls during the nestling phase to rates from alarm, wail, and juvenile-begging calls during the fledgling-depen- dency phase. Thus, it is difficult to ascertain wheth- er Kennedy and Stahlecker (1993) obtained signif- icant results because the effectiveness of individual call types differed between breeding phases or be- cause overall detection rates significantly increased between the nestling and fledgling phases. Total detection rates in our study were also greater m the fledgling-dependency phase compared to the nestling and courtship phases. Breeding phase influenced the distance gos- hawks were detected from occupied nests with the mean distance of detection for alarm and male food-delivery calls increasing between the nestling and fledgling-dependency phases. These trends are consistent with other broadcast experiments on goshawks (Kimmel and Yahner 1990, Kennedy and Stahlecker 1993, Watson et al. 1999). Increased mean detection distances with alarm and male food-delivery calls between breeding phases re- flects the shift from adults comprising most detec- tions during the courtship and nestling phases, to primarily fledglings during the fledgling-depen- dency phase. In general, adults are secretive and reveal their presence when intruders or male gos- hawks are perceived as being close to nests, where- as fledglings often approach observers on transects because male food-delivery (this study), wail (Ken- nedy and Stahlecker 1993), and juvenile-begging calls (Kennedy and Stahlecker 1993, Watson et al. 1999, this study), probably suggest the likelihood of obtaining food. Although detection rates were not significantly different among breeding phases in our study, de- tection rates were 75% in the fledgling-dependen- cy phase, compared with 60% in the nestling phase. However, it is more difficult to locate oc- cupied nests during the fledgling-dependency phase because detections occur farther from nests. To maximize the probability of locating occupied nests, broadcast surveys should be conducted a minimum of twice throughout the breeding season (once during each of the nestling and fledgling- dependency phases). Nest areas should be sur- veyed a minimum of two consecutive nesting sea- sons because goshawk nest areas are not always occupied annually (Kennedy and Stahlecker 1993). Distance between broadcast stations and transect lines. The distances goshawks are detected from active nests influences the optimal spacing of broadcast stations and transect lines. Given that detection rates generally decrease when observers are farther from nests (Kennedy and Stahlecker 1993, Watson et al. 1999, this study), observers are less likely to detect goshawks as the spacing between broadcast stations and transects is increased. Kennedy and 206 McClaren et al. VoL. 37, No. 3 Stahlecker (1993) recommended broadcast sta- tions be located 300-m apart on parallel transects separated from each other by 260 m, and stations on adjacent transects should be offset by 130 m to maximize coverage, because they assumed that gos- hawk detections were maximum within 100-200 m of occupied nests. Results from this study and Wat- son et al. (1999) suggest that broadcast surveys in dense, coastal forests could be improved during the nestling phase if broadcast stations and tran- sects are separated by 200 m, with parallel transects being offset from one another by 100 m. Because fledglings are detected at greater distances from nests during the fledgling-dependency phase, broadcast stations and transects conducted during this time could be separated by 400 m to maximize survey efficiency. Staggering adjacent transects by half the distance between broadcast stations maxi- mizes the area covered by calls (Joy et al. 1994) . Time spent at broadcast stations. The time we spent at broadcast stations also appeared to influence our likelihood of detecting goshawks. In our ex- periment, 37% {N — 14) of detections occurred beyond the 3 min/ station practiced by other re- searchers (Kennedy and Stahlecker 1993, Watson et al. 1999). Our results suggest the optimal amount of time/station is 5 min in dense coastal forests. Sampling for 5 min/broadcast station or six calls plus a 1 min listening period, increases the probability of detecting goshawks at nearby occu- pied nests, while enabling broadcast surveys to be conducted more efficiently than when observers spend 9 min/station. However, 9 min/station is recommended if surveyors wish to maximize de- tection probabilities without time constraints. Acknowledgments Forest Renewal British Columbia and the Habitat Con- servation Trust Fund provided financial assistance for this research. TimberWest, Weyerhaeuser, Canadian For- est Products, and Western Forest Products reported nest sites, provided funding and enabled this project to run smoothly. D. Doyle secured funding and provided invalu- able logistical support. We are grateful to M. Bigg, C. Darimont, D. Klinka, D. Koshowski, S. McDonald, T. Van Enter, K. Wickert, and several others who contributed countless hours of field assistance and reported goshawk nest site locations. M. Baker, B. Van Horne, and K. Wil- son offered important study design input and feedback on earlier drafts of this manuscript. T. Bosakowski, G. Hayward, D. Stahlecker, J. Squires, and K. Titus also re- viewed an earlier draft of this manuscript and their com- ments improved the clarity of the manuscript. Literature Cited Anderson, D.R. 2001. The need to get the basics right in wildlife field studies. Wildl. Soc. Bull. 29:1294-1297 Bosakowski, T. and D.G. Smith. 1998. Response of a for- est raptor community to broadcasts of heterospecific and conspecific calls during the breeding season. Can Field-Nat. 112:198-203. and M.E. Vaughn. 1996. Developing a practical method for surveying Northern Goshawks in man- aged forests of the western Washington Cascades. West. J. Appl. For. 11:1 09-1 13. Brown, TJ. and P. Handford. 2000. Sound design for vocalizations: quality in the woods, consistency in the fields. Condor 102:81-92. Cooper, J.M. and P. Chytyk. 2000. Draft status report on the Queen Charlotte Goshawk Accipiter gentilis laingi in Canada. Committee on the Status of Endangered Wildlife in Canada, Ottawa, ON Canada. Crocker-Bedford, D.C. 1998. The value of demographic and habitat studies in determining the status of Northern Goshawks {Accipiter gentilis atricapillus) with special reference to Crocker-Bedford (1990) and Ken- nedy (1997)./. Raptor Res. 32:329-335. DeStefano, S. and j. McCloskey. 1997. Does vegetation structure limit the distribution of Northern Goshawks in the Oregon coast ranges? J. Raptor Res. 31:34—39. Environment Canada. 1998. Canadian climate normals 1961-1990: Campbell River airport, British Columbia Retrieved September 13, 2001 from the World Wide Web: http:/ /www.cmc.ec.gc.ca/climate/normals/ BCC002.HTM. Erwin, R.M., CJ. Conway, and S.W. Hadden. 2002. Spe- cies occurrence of marsh birds at Cape Cod National Seashore, Massachusetts. Northeast. Nat. 9:1-12. Ethier, T.J. 1999. Breeding ecology and habitat of North- ern Goshawks {Accipiter gentilis laingi) on Vancouver Island: a hierarchical approach. M.S. thesis, University of Victoria, Victoria, BC Canada. Fotheringham, J.R. and L. Ratcliffe. 1995. Song deg- radation and estimation of acoustic distance in Black- capped Chickadees {Parus atricapillus). Can. J. Zool, 73:858-868. Fuller, M.A. and J.A. Mosher. 1981. Methods of detect- ing and counting raptors: a review. Stud. Avian Biol. 6. 231-246. AND . 1987. Raptor survey techniques. Pag- es 37-65 in B.A. Giron Pendleton, B.A. Millsap, K.W. Cline, and D.M. Bird [Eds.], Raptor management techniques manual. Nat. Wildl. Eed. Sci. Tech. Ser No. 10, Washington, DC U.S.A. Hardy, P.C. and M.L. Morrison. 2000. Factors affecting the detection of Elf Owls and Western Screech-Owls. Wildl. Soc. Bull. 28:333-343. Holland, J., T. Dabelsteen, and S.B. Pedersen. 1998. Degradation of wren Troglodytes troglodytes song: imph- September 2003 Broadcast Surveys for Goshawks 207 cations for information transfer and ranging./. Acoust. Soc. Am. 103:2154-2166. Iverson, G.C., G.D. Hayward, K. Titus, E. DeGayner, R.E. I.OWEEI., D.C. Crocker-Bedford, re. Schempf, AND J. Lindell. 1996. Conservation asse.ssment for the Northern Goshawk in southeast Alaska. Gen. Tech. Rep. PNW-GTR-387, USDA, USES, Pacific Northwest Research Station, Portland, OR U.S.A. Joy, S.M., R.T. Reynoi.ds, and D.G. Leslie. 1994. North- ern Goshawk broadcast surveys: hawk response vari- ables and survey costs. Stud. Avian Biol. 16:24-30. Kennedy, P.L. 1998. Evaluating Northern Goshawk {Ac- cipiter gentilis atricapillus) population trends: a reply to Smallwood and Crocker-Bedford. J. Raptor Res. 32: 336-341. and D.E. Andersen. 1999. Research and monitor- ing plan for Northern Goshawks (Accipiter gentilis atri- capillus) in the Western Great Lakes Region. Re- trieved March 29, 2000 from the World Wide Web: http:// www.fw.umn.edu/CO-OP/download/ default.html. AND D.W. Stahiecker. 1993. Responsiveness of nesting Northern Goshawks to taped broadcasts of 3 conspecific calls./. Wildl. Manage. 57:249-257. , J.M. Ward, G.A. Rinker, and J.A. Gessaman. 1994. Post-fledgling areas in Northern Goshawk home ranges./. Avian Biol. 16:75-82. Kenward, R.E., V. Marcstrom, and M. Karlbom. 1993. Post-nesting behaviour in goshawks, Accipiter gentilis: I. The causes of dispersal. Anim. Behav. 46:365-370. Kimmel, J.T. AND R.H. Yahner. 1990. Response of North- ern Goshawks to taped conspecific and Great Horned Owl calls./. Raptor Res. 24:107-112. Legare, M.L., W.R. Eddleman, P.A. Buckiey, and K. Col- leen. 1999. The effectiveness of tape playback in es- timating Black Rail density./. Wildl. Manage, 63:116— 125. Littell, R.C., G.A. Milliken, W.W. Stroup, and R.D. WOLEINGER. 1996. SAS system for mixed models. SAS Institute, Cary, NC U.S.A. McClaren, E.L. 2001. Factors influencing Northern Gos- hawk detectability and reproduction on Vancouver Is- land, British Columbia. M.S. thesis, Colorado State University, Fort Collins, CO U.S.A. MgClaren, E.L., P.L. Kennedy, and S.R. Dewey. 2002. Do some Northern Goshawk nest areas fledge more young than others? Condor 104:343-352. McCulloch, C.E. and S.R. Searle. 2001. Generalized, linear, and mixed models. Wiley Series in Probability and Statistics. John Wiley and Sons, Inc., New York, NY U.S.A, McLeod, M. and D.E. Andersen. 1998. Red-shouldered Hawk broadcast surveys: factors affecting detection of responses and population trends./. Wildl. Manage. 62: 1385-1397. Ministry of Environment, Lands and Parks. 2000. 2000 Red and Blue lists for amphibians, reptiles, birds and mammals. Victoria, BC Canada. Mosher, J.A. and M.R. Fuller. 1996. Surveying wood- land hawks with broadcasts of Great Horned Owl vo- calizations. Wildl. Soc. Bull. 24:531-536. Nichols, J.D., J.E. Hines, J.R. Sauer, F.W. Fallon, J.E Fallon, and P.J. Heglund. 2000. A double-observer approach for estimating detection probability and abundance from point counts. Auk 117:393-408. Penteriani, V. 2001. The annual and diel cycles of gos- hawk vocalizations at nest sites. /. Raptor Res. 35:24— 30. Pojar, J.E., K. Klinka, and D.A. Demarche 1991. Coastal western hemlock biogeoclimatic zone. Pages 95-112 in D.V. Meidinger and J.E. Pojar [Eds.], Ecosystems of British Columbia. Spec. Rep. Ser. No. 6. British Co- lumbia Ministry of Forests Research Branch, Victoria, BC Canada. Reid, J.A., R.B. Horn, and E.D. Forsman. 1999. Detec- tion rates of Spotted Owls based on acoustic-lure and live-lure surveys. Wildl. Soc. Bull. 27:986-990. Resource Inventory Committee. 1997. Standard inven- tory methodologies for components of British Colum- bia’s biodiversity: raptors. Version 1.1. Ministry of En- vironment, Lands and Parks, Wildlife Branch, Victoria, BC Canada. Roberson, A.M. 2001. Evaluating and developing survey techniques using broadcast conspecific calls for Northern Goshawks in Minnesota. M.S. thesis. Uni- versity of Minnesota, St. Paul, MN U.S.A. Rosenfield, R.N., J. Bielefeldt, and R.K. Anderson 1988. Effectiveness of broadcast calls for detecting breeding Cooper’s Hawks. Wildl. Soc. Bull. 16:210- 212 . Rosenstock, S.S., D.R. Anderson, K.M. Giesen, T. Leu- kering, and M.F. Carter. 2002. Landbird counting techniques: current practices and an alternative. Auk 119:46-53. SAS Institute. 1989. SAS/STAT user’s guide. Version 6.12. Fourth Ed. SAS Institute, Cary, NC U.S.A. ScHNELL, J.H. 1958. Nesting behavior and food habits of goshawks in the Sierra Nevada of California. Condor 60:377-403. Siders, M.S. and P.L. Kennedy. 1996. Forest structural characteristics of Accipiter nesting habitat: is there an allometric relationship? Condor 98:123-132. Smallwood, K.S. 1998. On the evidence needed for list- ing Northern Goshawks {Accipiter gentilis) under the Endangered Species Act: a reply to Kennedy. /. Raptor Res. 32:323-329. Smith, G.C. and D.N. Jones. 1997. Vocalisations of the marbled frogmouth I: descriptions and analysis of sex differences. Emu 97:290-295. Squires, J.R. and R.T. Reynolds. 1997. Northern Gos- hawk {Accipiter gentilis) . Pages 1-32 in A.. Poole and F. Gill [Eds.], The Birds of North America, No. 298. The Academy of Natural Sciences, Philadelphia, PAU.S.A., 208 McClaren et al. VoL. 37, No. 3 and The American Ornithologists’ Union, Washing- ton, DC U.S.A. Thompson, W.L. 2002. Towards reliable bird surveys: ac- counting for individuals present but not detected. Auk 119:18-25. Titus, K., C. Flatten, and R. Lowell. 1999. Goshawk ecology and habitat relationships on the Tongass Na- tional Forest. 1998 annual report. Study SE-4-2. Alaska Department of Fish and Game, Division of Wildlife Conservation, Juneau, AK U.S.A. United States Forest Service. 1994. Goshawk monitor- ing, management, and research in the Pacific North- west region: 1994 status report. USDA, USFS, Port- land, OR U.S.A. Watson, J.W., D.W. Hays, and D.J. Pierce. 1999. Efficacy of Northern Goshawk broadcast surveys in Washing- ton State./. Wildl. Manage. 63:98-106. Received 15 July 2002; accepted 21 April 2003 J Raptor Res. 37(3) :209-218 © 2003 The Raptor Research Foundation, Inc. REPRODUCTIVE SUCCESS, ENVIRONMENTAL CONTAMINANTS, AND TROPHIC STATUS OF NESTING BALD EAGLES IN EASTERN NEWFOUNDLAND, CANADA Laura Dominguez and William A. Montevecchi Biopsychology Programme, Memorial University of Newfoundland, St. John’s, Newfoundland AIB 3X9 Canada Neil M. Burgess^ Canadian Wildlife Service, Environment Canada, 6 Bruce Street, Mount Pearl, Newfoundland AIN 4T3 Canada Joe Brazil Inland Fish & Wildlife Division, Department of Tourism, Culture and Recreation, P.O. Box 2007, Corner Brook, Newfoundland A2H 7S1 Canada Keith A. Hobson Canadian Wildlife Service, Environment Canada, 115 Perimeter Rd., Saskatoon, Saskatchewan S7N 0X4 Canada Abstract. — Eastern Newfoundland contains a large breeding population of Bald Eagles (Haliaeetus leu- cocephalus) . Our objectives were to determine the reproductive success, contaminant levels, and trophic status of Bald Eagles in eastern Newfoundland, and to compare contaminant levels of eagles in indus- trialized and nonindustrialized coastal bays (Placentia and Bonavista bays, respectively). Bald Eagle breeding density in eastern Newfoundland was 0.09 occupied nests/km shoreline, and reproductive success suggested a healthy population (1.1 young/occupied nest). Geometric mean concentrations of PCBs and DDE in nestling plasma were significantly higher in Placentia Bay (PCBs = 30.0 ng/g [wet wt]; DDE = 9.0 ng/g) than in Bonavista Bay (PCBs = 10.0 ng/g; DDE = 2.0 ng/g). Within Placentia Bay, concentrations of PCBs and DDE in nestlings were negatively related to the nest distance from the former U.S. naval base at Argentia. Geometric mean concentration of mercury in whole blood of eagle nestlings was 87 ng/g (wet wt). Analysis of prey remains collected at nests and of stable-carbon '‘^C) and nitrogen (’'’N/’'^N) isotope ratios in nestling blood indicated some differences between nests in food sources, but not in relative trophic level. Prey remains collected at 30 Bald Eagle nests were composed of 64% birds, 29% fish, and 7% mammals. Contaminant burdens in Bald Eagles in eastern Newfoundland were lower than those in populations from more industrialized areas of North America and were below thresholds for reproductive impairment. These results provide an essential baseline for future monitoring of the eagle population as industrial development increases in Placentia Bay. Key Words: Bald Eagle, Haliaeetus leucocephalus; contaminants', diet, stable isotopes', Newfoundland', pro- ductivity. EXITO REPRODUCTOR, CONTAMINANTES AMBIENTALES, Y ESTATUS TROFICO DE AGUIIAS CALVAS EN NIDIFICANTES EN EL ESTE DE TERRANOVA (NEWFOUNDLAND), CANADA Resumen. — El este de Terranova alberga una gran poblacion re.sidente grande de Aguilas Calvas {Ha- liaeetus leucocephalus) . Los objetivos de la investigacion fueron; determinar el exito reproductor, los ni- veles de contaminacion, y el estatus trofico de las Aguilas Calvas en el este Terranova y comparar los niveles de contaminacion de las aguilas de bahias costeras industrializadas y no industrializadas (Bahias de Placentia y de Bonavista, respectivamente). La densidad del Aguila Calva en reproduccion en Terra- nova oriental fue de 0.09 nidos ocupados/km de costa y el exito reproductor resulto ser el de una poblacion sana (1.1 juvenil/nido ocupado). La media geometrica de las concentraciones de PCBs y DDE en plasma de polluelos fue significativamenta mas alta en la Bahia de Placentia (PCB = 30.0 ng/ g [peso fresco]; DDE = 9.0 ng/g) que en la Bahia de Bonavista (PCB = 10.0 ng/g; DDE = 2.0 ng/g). ^ Corresponding author’s e-mail address: neil.burgess@ec.gc.ca 209 210 Dominguez et al. VoL. 37, No. 3 En la Bahia de Placentia, se observo una relacion inversa entre las concentraciones de PCBs y DDE en polluelos y las distancias desde los nidos a la antigua base naval de los EE.UU. en Argentia. La media geometrica de la concentracion de mercurio en la sangre de polluelos fue de 87 ng/g (peso fresco). El analisis de restos de presas recolectadas en los nidos y las proporciones de los isotopos de carbono (i^C/^^C) y nitrogeno en sangre de polluelos, indicaron algunas diferencias entre las fuentes de alimento de los nidos, pero no en el nivel trofico relativo. Los restos de presas recogidos en 30 nidos de Aguila Calva estuvieron compuestos por aves (64%), peces (29%) y mamiferos (7%). Los niveles de contaminacion en las Aguilas Calvas del este de Terranova fueron mas bajos que en las de poblaciones de areas mas industrializadas de Norteamerica y se encontraron por debajo del umbral a partir del cual afectan a la reproduccion. Estos resultados proporcionan una base esencial para el futuro seguimiento de la poblacion de aguilas, a medida que aumente el desarrollo industrial en la Bahia de Placentia. [Traduccion de Cesar Marquezs, Laura Dominguez y Rafael Mateo] Bald Eagle {Haliaeetus leucocephalus) populations were affected seriously throughout North America from the 1950s-1970s by the widespread use of or- ganochlorine pesticides (OCs) such as DDT (Grier 1982, Colborn 1991). Although many breeding populations began recovering in the 1980s-1990s (Grier 1982, Wiemeyer et al. 1993, Bowerman et al. 1995), contaminants still appear to be a limiting factor in the reproductive success of several North American populations (Anthony et al. 1993, 1999, Welch 1994, Bowerman et al. 1995). One of the densest breeding concentrations of Bald Eagles in eastern North America is in Placen- tia Bay, Newfoundland (Dominguez 1999). Placen- tia Bay is also home to several industrial sources of chronic pollution. OC, polychlorinated biphenyl (PCB), and metal pollution was found at the for- mer U.S. naval base in Argentia, which closed in 1994 (JWEL 1996). Eluoride and phosphorus pol- lution was caused by the phosphorus plant in Long Harbour (Osbourne 1978), which closed in 1989. Other industrial developments in the bay include an oil refinery at Come-By-Chance, a crude oil storage facility at Whiffen Head, a shipyard at Marystown, and associated heavy shipping traffic. Chronic oil pollution in Placentia Bay has been documented by the high incidence of oiled sea- birds regularly washing up along its shores (Mon- tevecchi and Tuck 1987, Lock et al. 1994, Wiese and Ryan 2003). Contaminants transported atmo- spherically or through ocean currents are other possible sources of chronic pollution (Eurness 1993, Anthony et al. 1999). A recent study con- firmed that PCBs and metals from the former U.S. naval base at Argentia are bioaccumulating in the local marine food web (JWEL 1996). However, there has been no monitoring of contaminant ac- cumulation in top predators, such as Bald Eagles, that are sensitive to the presence of toxic chemicals that biomagnify up food chains (Howells et al. 1990). This study was undertaken to assess the repro- ductive success, contaminant exposure, and tro- phic status of nesting Bald Eagles prior to further industrial development in Placentia Bay. For com- parison, we measured the same parameters in a breeding population of Bald Eagles 80 km to the north in Bonavista Bay, which has had no major industrial development. We hypothesized that concentrations of PCBs and DDE would be higher in eagle nestlings in Pla- centia Bay than in Bonavista Bay, and higher in eagle nestlings located closer to the former naval base than in those further away along Placentia Bay. Contaminant concentrations were measured in Bald Eagle nestlings because they are good bioindicators for assessing local contamination over a relatively short period (Welch 1994), and they have been used widely in contaminant moni- toring studies across Canada and the United States (Anthony et al. 1993, Bowerman et al. 1995, Elliott and Norstrom 1998, Kumar et al. 2002). The tro- phic status of nesting Bald Eagles was compared by studying the composition of prey remains at nests and by analyzing stable-carbon and nitrogen iso- tope ratios in nestling blood samples (Hobson and Montevecchi 1991, Hobson et al. 1994). Methods Placentia Bay, located on the southeastern coast of Newfoundland (Fig. 1), is one of the richest fishing areas in Newfoundland waters, and has important seabird col- onies, wintering and staging grounds for millions of sea- birds and seaducks, breeding and wintering grounds for Bald Eagles, and feeding areas for several species of ma- rine mammals (LEML and OL 1984, Montevecchi and Tuck 1987, Cairns et al. 1989). Bonavista Bay, located on the northeast coast of Newfoundland (Fig. 1), is also rich in fish, seabird, seaduck, eagle, and marine mammal pop- ulations (Deichmann and Bradshaw 1984). The smdy September 2003 Contaminants in Bald Eagles 211 area in Bonavista Bay was located within the boundaries of Terra Nova National Park (TNNP). Reproductive Surveys. We studied Bald Eagle breeding density, nest success, and productivity in Placentia Bay in 1996 and 1997 and in Bonavista Bay in 1997. Definitions of these terms and others used to describe reproductive parameters follow those of Postupalsky (1974) and Kozie and Anderson (1991). We conducted a minimum of two surveys per breeding season along the coastline by air- craft or boat, extending 50-200 m inland from the shore- line, depending on topography. Boat surveys were con- ducted by cruising at 20-50 m from shore, in a 6-m fiberglass boat at a speed of 10-20 km/hr. Boat surveys m late April— early May assessed breeding population size and nest occupancy. Total shoreline surveyed was mea- sured with Mapinfo (Mapinfo Corp., Troy, NY) to cal- culate breeding density (number of occupied nests per kilometer of shoreline). Aerial surveys were conducted by helicopter (Bell 206) in Placentia Bay and by fixed- wing aircraft (Twin Otter) in Bonavista Bay, with cruising speeds of 120-140 km/h and flying altitudes of 50-100 m above tree line. Aerial survey schedules and method- ology followed standard methods (Postupalsky 1974, Grier 1982, Fraser et al. 1983). Aerial surveys in early June assessed clutch/brood size and chick development, and nest vi.sits in late June-early July assessed nest success and productivity. Nest visits involved climbing into the nests to collect blood samples from the chicks for con- taminant analyses. Contaminant Analyses. Blood samples from one nes- tling per nest (Welch 1994) were collected in late June and early July in 1996 and 1997 in Placentia Bay and m 1997 in Bonavista Bay. Approximately 12-15 ml of blood was drawn from the left brachial vein directly into hepa- rinized vacuum tubes when eaglets were 6-8 wk old. From each sample, 4-6 ml of whole blood were frozen in nitric-acid-rinsed cryovials for mercury analysis. The remaining blood was centrifuged and 3-6 ml of plasma were frozen in acetone-hexane-rinsed glass vials for OC analyses. Samples were stored at — 20°C until analyzed. Further details on blood collection and processing are provided by Dominguez (1999). Plasma and whole blood samples were sent to the Ca- nadian Wildlife Service National Wildlife Research Cen- ter (NWRC) in Hull, Quebec, for OC and mercury anal- yses. Measurement of 21 OC pesticides and metabolites and 42 PCB congeners in plasma was conducted by gas- chromatography mass spectrometry (Hewlett Packard model 5890 gas chromatograph, coupled to a Hewlett Packard model 5971 Ni®^ electron capture detector, Palo Alto, CA) . Methods follow those of Norstrom and Won (1985). Recoveries of internal standards ranged from 74— 104%. Detection limits were 0.1 ng/g wet weight (wetwt) for both OCs and PCBs. Plasma lipids were measured by 212 Dominguez et au. VoL. 37, No. 3 Table 1. Reproductive success of Bald Eagle breeding pairs in Placentia and Bonavista Bays, Newfoundland, 1996-97. Location Year No. Occu- pied Nests No. Success- ful Nests Nest Success (%) No. Young/ No. Young/ Occupied Nest Successful Nest Placentia Bay 1996 31 25 81 1.3 1.6 Placentia Bay 1997 27 18 67 0.9 1.4 Bonavista Bay 1997 12 10 83 0.9 1.1 sulpho-phospho-vanillin reaction (Frings et al. 1972). To- tal mercury in blood was analyzed by cold-vapor atomic- absorption spectrophotometry (Perkin-Elmer 3030-AAS equipped with Varian VGA-76 hydride generator, Shel- ton, CT; Scheuhammer and Bond 1991, Neugebauer et al 2000). National Research Council of Canada DOLT-2 and DORM-2 were used as standard reference materials to ensure the accuracy of the mercury results. Two blanks were included in each set of digestions. All blood samples were analyzed in duplicate. Recoveries of reference ma- terials were within the certified range. The detection lim- it for mercury was 40 ng/g (wet wt). Field procedural blanks were included in all chemical analyses. All contam- inant concentrations are presented in ng/g (wet wt). Trophic Analyses, Prey remains and a sample of fine nest material (to check for Fish scales and other small remains) were collected from each nest bowl and the nest-site area, and were classified as fish, bird, or mam- mal. Composition of prey remains is expressed as percent occurrence of minimum number of individuals for each of the three prey classes (Todd et al. 1982, Knight et al. 1990). Regurgitated pellets below nests could not be re- trieved because most nests were located on cliffs over- hanging the sea. Stable nitrogen isotope ratios (^■''N/^'^N) were used to compare the dietary trophic level of Bald Eagles between and within study areas (Hobson et al. 1994), and to investigate relationships between contami- nant concentrations and trophic level (Braune et al. 2002). Stable carbon isotope ratios (^‘^C/^^C) were used to investigate food sources (i.e., freshwater vs. marine, benthic vs. pelagic marine food chains; Hobson et al. 1994). Dried blood samples were loaded into tin cups and combusted at 1850°C in a Robo-Prep elemental an- alyzer (Europa Etd., Crewe, England) interfaced with a Europa 20:20 continuous-flow isotope-ratio mass spec- trometer (Hobson et al. 1999). Two standards (egg al- bnmin) were measured in sequence for every five un- knowns. Analytical error is estimated to be ±0.3%o for ''’N and ±0.1 %o for ^'^C analyses. Stable isotope ratios in blood samples are reported in delta notation as parts per thousand according to the following: 8V (%o) = [{RsumpiJ f^standard) “ D X 1000 whcrc X Is >"’N or ^'^C and R IS the corresponding ratio '■''^N/^'*N or '■'^C/^^C. Rsiandard for ^'’N and ^'^C are atmospheric N^ (AIR) and the PDB standard, respectively. Statistical Analyses. Contaminant data were loguptrans- formed to achieve a normal distribution, and geometric means were calculated. Contaminant data were not lipid- normalized for statistical analyses because there were no correlations between plasma lipid concentrations and log-transformed concentrations of contaminants (e.g., r = 0.07, 0.2, and 0.2 for DDE, PCB, and mercnry, respec- tively). When samples had contaminant concentrations below the analytical detection limit, and 50% or more of the samples had values above detection limits, a value of one-half the detection limit was assigned (Anthony et al 1999). One-way analysis of variance (ANOVA) was used to test differences in the mean contaminant concentrations be- tween years in Placentia Bay, and between Placentia and Bonavista Bays. Within Placentia Bay, relationships be- tween contaminant concentrations in nestling tissues and nest distance (km) to the former U.S. naval base were assessed using linear regression. Differences in composi- tion of prey remains between years in Placentia Bay and between stndy areas were tested with a chi-square test of independence and Fisher’s exact test, respectively. Dif- ferences in the means of 8'‘’N and S^-^C values between study areas were tested with a one-way ANOVA. Linear regression was used to assess relationships between con- taminant concentrations and isotopic ratios. Multiple re- gression analysis was used to assess the relative strength of associations between contaminant concentrations, nest distance to the former naval facility and trophic differ- ences reflected by stable isotope ratios. Randomization tests (Adams and Anthony 1996) were conducted (10 000 runs) to prodnce reliable P values when residuals from ANOVAs did not show a normal distribntion (test results indicated as Prand)- ^ value of a = 0.05 was used in all tests. Statistical tests were conducted with Minitab (Mi- nitab Inc., State College, PA) and SYSTAT (SPSS Inc , Chicago, IL). Results Reproductive Success. We identified 55 and 13 breeding territories in Placentia and Bonavista Bays, respectively; some of them containing more than one nest. Breeding density was 0.10 and 0.06 occupied nests/km of shoreline for Placentia and Bonavista Bays, respectively (overall mean = 0.09 occupied nests/km). Nest success and productivity (Table 1) were not significantly different between study areas (x^ = 0.46, P — 0.5 and 7 / 41 = 2.5, P = 0.1, respectively). Overall, mean nest success in eastern Newfoundland was 76% and mean produc- tivity was 1.1 young/occupied nest. Organochlorines. Plasma concentrations of OCs were low in all samples in both Placentia and Bon- September 2003 Contaminants in Bald Eagi.es avista Bays (Table 2). In Placentia Bay, there were no significant differences in mean concentrations between years (e.g., 20 = 0.58, Prand = 0.1 and ^ 1,20 ^ 1-9, Prand = 0.1 for DDE and PCBs, respec- tively). Therefore, results were pooled for 1996 and 1997 and mean values were used for nests sam- pled in both years. Geometric mean concentra- tions of plasma DDE and PCBs were significantly higher in Placentia Bay than in Bonavista Bay (Pi_ 2 i = 4.0, Prand = 0.03 and P^ 21 = 6.3, P^and ~ 0.003, respectively; Table 2). There were no significant differences in mean concentrations of other OCs (Table 2). Dieldrin; j&p'-DDD; pp'-\yUI-, pentachlo- robenzene; 1234- and 1245-tetrachlorobenzene; oc- tachlorostyrene; a-, 3- and y-hexachlorocyclohexane; Pfln5^chlordane; and tris(4-chlorophenyl) methanol were detected in less than half of the plasma sam- ples. In Placentia Bay, DDE and PCB concentra- tions were negatively related to nest distance from the former U.S. naval site at Argentia (P^ = 0.4, P — 0.07 and = 0.3, P = 0.02, respectively; Eig. 2). There were no significant relationships be- tween concentrations of other OCs and distance from the naval site. Mercury. No significant differences were found in geometric mean mercury concentrations in whole blood of eagle nestlings between years in Placentia Bay (Pi ^o 0.001, P^^nd ^ 0.3), or be- tween Placentia and Bonavista Bays (Pi 21 = 1-7, P = 0.2). The overall geometric mean concentration for blood mercury in nestlings in eastern New- foundland was 87 ng/g {N — 23, range: 50-250 ng/g wet wt). There was no association between mercury concentrations and nest distance from the naval base in Placentia Bay (Fig. 2). Trophic Status. Most birds found in prey re- mains were Black-legged Kittiwakes {Rissa tridacty- la) , and Herring Gulls {Larus argentatus) , or murres {Uria spp.) in both Placentia and Bonavista Bays. Fish species included yellow-tail flounder {Li- manda ferruginea) , wolffish {Anarhichas spp.), scul- pin {Myoxocephalus spp.), lumpfish {Cyclopterus lum- pus), Atlantic cod {Gadus morhua), redfish {Sebastes spp.), Atlantic herring (Clupea harengus), and American lobster {Homarus americanus) . The only mammal remains identified were snowshoe hare {Lepus americanus). The minimum percent occur- rence of avian, fish, and mammalian prey in the remains at 22 Placentia Bay eagle nests and eight Bonavista Bay nests averaged 64, 29, and 7%, re- spectively. There were no significant differences in the frequency of occurrence of bird and fish re- C o PQ d cS • ^ OJ u s o cc be S 'he 'Tj CQ o 6 "Oh V, ■qh dj TfC ^ d Q d CO Z £ CO q op »o ' — ' cO d (M q CO o C20 1 — ^ I i-C I ' — I lO o irj GO 05 co to qj 00 ^ 02 « ^ « Z P o ^ ^ “ oi ^ d ^ CO — s ^ CO .5fH Id ^ I-H ^ CO CO I ’d I CO 1-H ® td 06 d e- I ao OJ 1-H I d CO in d O) d o PQ ® Abbreviations are as follows: XPCB = sum of 42 PCB congeners, toonachlor = trans-nonachlor, HCB = hexachlorobenzene, HE = heptachlor epoxide. A, B = means sharing the same letter are significantly different, P^and < 0.05. ■^ND = below detection limit (0.1 ng/g, wet wt). 214 Dominguez et al. VoL. 37, No. 3 Distance from Argentia (km) Figure 2. Regressions of contaminant concentrations (ng/g, wet wt) in plasma or whole blood of Bald Eagle nestlings in Placentia Bay and nest distance (km) from the former U.S. naval base at Argentia, Newfoundland. mains either between years in Placentia Bay {N — 80, = 0.1, P — 0.2) or between Placentia and Bonavista Bays (N = 102, ^ 0.9, P = 0.3). Stable isotope ratios were determined in 13 blood samples from Placentia Bay and six samples from Bonavista Bay collected in 1997. Geometric mean values (±SD) in nestling blood were not significantly different between Placentia and Bonavista Bays (16.0 ± 0A%o and 15.7 ± 0.2%o, respectively; G, 17 - 1-69, P= 0.2). 8 values were not significantly related to plasma DDE concentra- tions = 0.004, 7*) 15 = 0.44, P — 0.5), plasma PCB concentrations — 0.002, G,i5 ~ 0.03, P = 0.8), blood mercury concentrations (R^ = 0.002, 73,15 ~ 0.02, P — 0.9), or nest distance from the naval site in Placentia Bay (R^ = 0.03, i^.ii ~ 0.2, P = 0.6). Geometric mean 8^^^G values (±SD) in nestling blood were greater in Bonavista Bay than in Placentia Bay ( — 17.8 ± 0.2%o and —18.3 ± 0.5 %o, respectively; E) = 5.0, P — 0.05). 8^®C val- ues were negatively associated with plasma DDE {R^ = 0.36, = 9.3, P = 0.008) and PCB con- centrations (i?^ = 0.36, 73 15 = 8.7, P = 0.009) in nestlings from both bays. There was no relation- ship between 8^®C values and blood mercury con- centrations (7?2 — —0.09, 73,15 ~ 2.1, P = 0.2). Within Placentia Bay, 8^^C values were positively re- lated to nest distance from the naval base (R^ — 0.48; 73,9 = 8.3; P = 0.02). Multiple regression results indicated that plasma DDE concentrations in Placentia Bay eagles were associated with nest distance from the naval base but not with 8^3C values (R^ = 0.66, 73,8 ^ 10.5, P = 0.01; 73 8 = 0.4, P = 0.5; respectively). Similar analysis showed that plasma PCB concentrations were associated more with nest distance from the naval base than with 8^"^C values, although neither association was statistically significant (7?^ = 0.46, T'l.s ~ 3-4, 7* = 0.1; 73,8 = 0.001, P — 0.9; respec- tively) for Placentia Bay nestlings. Discussion Concentrations of OCs in plasma of Bald Eagle nestlings were low in both study areas in eastern Newfoundland. Observed plasma DDE and PCB concentrations were below thresholds associated with depressed Bald Eagle productivity (i.e., <0.7 young/occupied nest), namely 107 ng/g PCBs and 26 ng/g DDE (Bowerman et al. 1995). Levels of DDE, PCBs, and other OCs were lower than those found in plasma of nestling eagles in more indus- trialized regions such as the Columbia River Estu- ary, Great Lakes, and Maine (geometric means: 80-200 ng/g PCB, 20-100 ng/g DDE; Anthony et al. 1993, Bowerman 1993, Matz 1998, Donaldson et al. 1999) but were similar to those found in Brit- ish Columbia and interior areas of the Midwest and Ontario (5-50 ng/g PCB, 3-24 ng/g DDE; Bower- man 1993, Elliott 1995, Donaldson et al. 1999). We September 2003 Contaminants in Bald Eagles 215 conclude that OC concentrations in eagle nestlings in eastern Newfoundland are below levels associ- ated with population-level impacts. Mean DDE and PCB concentrations were signif- icantly higher in Placentia Bay than in Bonavista Bay, with an increasing trend in Placentia Bay in nests closer to the Argentia naval base. Concentra- tions of other OCs were similar in both study areas. Our findings support the hypotheses that the Ar- gentia naval base is a source of PCB contamination to the local marine environment, and that part of the contamination detected in the eagles had a lo- cal origin at the base (although we acknowledge the smaller sample size from the reference site in our comparative analyses). The U.S. Navy con- structed the Naval Air Station, Dockyard, and as- sociated fuel, ordnance, communication, and sup- port facilities at Argentia in the early 1940s (Argentia Remediation Group 1995). These facili- ties were gradually closed or abandoned by the American military between 1968 and 1994. Soil, groundwater, marine sediments, and biota at Ar- gentia were contaminated with petroleum hydro- carbons, OCs, PCBs, or metals, which leaked from fuel storage tanks, an electrical power plant, elec- trical transformers, several landhlls, and equip- ment storage areas (Argentia Remediation Group 1995, JWEL 1996). Our results suggest that PCB contamination of marine hsh observed around Ar- gentia (JWEL 1996) is moving up the food web into Bald Eagles (although not reaching toxic con- centrations in eagle nestlings). PCB contamination has been frequently ob- served around military installations in North America, Europe, and Asia (de March et al. 1998, Gregor et al. 2003, Kuzyk et al. 2003). At these con- taminated sites, PCBs are often found in soil adja- cent to buildings with electrical equipment, in and around dumps, and migrating from these sources down drainage channels into the ocean (Gregor et al. 2003) . In a contaminant study in the Aleutian Islands of Alaska, Anthony et al. (1999) reported higher levels of PCBs in Bald Eagle eggs from nests that were located on islands that had a former mil- itary base than in eggs from islands that did not. Blood mercury levels in Newfoundland Bald Ea- gle nestlings were similar to those in Elorida and coastal Maine nestlings (geometric means: 82-130 ng/g; Welch 1994, Wood et al. 1995), but lower than those in Oregon and Washington nestlings (230-1200 ng/g; Wiemeyer et al. 1989, Anthony et al. 1993). No threshold value for adverse effects is available in the literature for blood mercury con- centrations in Bald Eagles. The blood mercury concentrations in Newfoundland eagles were less than those (ca. 1000 ng/g in 5-wk-old chicks) that led to sublethal effects on appetite, body condi- tion, behavioral activity, anemia, and organ histol- ogy in dosed Great Egrets (Ardea alba) (Bouton et al. 1999, Spalding et al. 2000a, 2000b). Dietary habits and trophic level can influence contaminant concentrations in eagles. Bald Eagles that feed mainly on seabirds, especially gulls, are known to bioaccumulate higher concentrations of contaminants than eagles feeding mainly on hsh or terrestrial herbivores (Kozie and Anderson 1991, Welch 1994, Anthony et al. 1999). Our anal- ysis of bird, hsh, and mammal remains at eagle nests indicated that the trophic status of eagles was similar in Placentia and Bonavista Bays. Prey re- mains collected at Newfoundland nests were simi- lar to other coastal populations of Bald Eagles, where seabirds predominate in the diet (Welch 1994, Todd et al. 1982, Knight et al. 1990). Because analysis of Bald Eagle prey remains overestimates the dietary importance of birds, medium-sized mammals, and bony hsh, while underestimating consumption of small hsh and mammals (Mers- mann et al. 1992), it is important to limit compar- isons to similar studies of prey remains. Traditionally, combined analyses of prey re- mains, regurgitated pellets, and direct observation of feeding events are required to assess dietary composition accurately for Bald Eagles (Mersmann et al. 1992). However, stable isotopic analyses of bird tissues is another method available to quanti- tatively assess their trophic status (Hobson et al. 1994). Analyses of 8^^N values in nestling blood samples showed no dietary trophic differences among nestlings, and no relationships between contaminants and 8^^N values. Although 8^’’N val- ues of food web components may differ among lo- cations, we feel that the observed differences in contaminant levels are not due to differences in dietary trophic level among nestlings. This conclu- sion is supported by the prey remains analysis. In contrast, 8^^C values were greater in Bonavista Bay, and in nests further away from the Argentia naval base in Placentia Bay. Greater 8^"^C ratios in birds in marine ecosystems have been related to diets with more inshore or benthic prey, while low- er 8^^C ratios have been related to more offshore and pelagic-based diets (Hobson et al. 1994). Nests in Bonavista Bay are located in a long sound with 216 Dominguez et al. VoL. 37, No. 3 an estuary at the head of the bay. In Placentia Bay, nests further away from the naval base are also lo- cated near an estuary at the head of the bay, while nests closer to the base are located on islands near- er the mouth of the bay. This difference in geo- graphical distribution and proximity to freshwater inputs could explain the differences in values between study sites and within Placentia Bay. The associations found between values and PCB and DDE levels in nestling plasma indicate a pos- sible dietary cause for differences in contaminant concentrations. However, the multiple regression analysis suggests that proximity to the Argentia na- val base was more important than values in influencing DDE and PCB concentrations in young eagles. Given our small sample sizes, the snap-shot nature of our dietary study, the opportunistic for- aging strategies of Bald Eagles (Stalmaster 1987), and the changing availability of different prey through the breeding season (e.g., spawning of dif- ferent species of hsh), there remains much to in- vestigate about Bald Eagle foraging ecology, stable isotope patterns, and their association with con- taminant levels in Newfoundland. Breeding density of Bald Eagles in eastern New- foundland (0.090 occupied nests/km shoreline) was similar to coastal British Columbia (0.086 nests/km; Hodges and King 1984), greater than coastal New Brunswick (0.014 nests/km; Stocek and Pearce 1981) and Nova Scotia (0.047 nests/ km; MacDonald and Austin-Smith 1989), but less than the maximum densities seen along the south- east coast of Alaska (0.28 nests/km; Hodges 1982). High breeding densities and reproductive output of Bald Eagles in marine ecosystems are often as- sociated with high quality and availability of nest- ing habitat and with food abundance (Hansen 1987). We cannot draw a definitive conclusion on the reproductive status of the Bald Eagle population in eastern Newfoundland from our study because a minimum of 3 yr of data collection is recommend- ed to reliably determine mean productivity (Elliott 1995). However, our data from the two years in Placentia Bay and the one year in Bonavista Bay indicate that reproductive success was high and are suggestive of a healthy breeding population, i.e., >1 young/occupied nest (Wiemeyer et al. 1993, Bowerman et al. 1995). The breeding densities and reproductive success of Bald Eagles in eastern New- foundland can presumably be attributed to high availability and quality of food, perches, and nest- ing habitat (Chandler et al. 1995), and to relatively low contaminant levels in the environment. We found no evidence that current contaminant levels are impairing reproduction in Bald Eagles in Placentia and Bonavista Bays. Nevertheless, Bald Eagles would be an excellent bioindicator of any ecological impacts of future industrial develop- ment in Placentia Bay (e.g., proposed nickel smelt- er at Argentia), and of any ecological benehts of on-going remediation of existing contaminated sites at Argentia. This study provides the necessary baseline information on Bald Eagle productivity and contamination for any future monitoring pro- gram. Acknowi.edgments Memorial University of Newfoundland (MUN), the Newfoundland and Labrador Inland Fish & Wildlife Di- vision, Terra Nova National Park, the Canadian Wildlife Service (CWS), the Mountain Equipment Co-op, and the Hibernia Management and Development Corporation provided funding and/ or in-kind support for this study Public Works and Government Services Canada (PWGSC) funded the chemical analyses of samples. Hen- ry Won and Ewa Neugebauer at the CWS National Wild- life Research Centre performed the chemical analyses, and Garth Parry carried out the stable isotope analyses in the Department of Soil Science, University of Saskatch- ewan. We thank Scott Gilliland, Geoff Goodyear, Hugh Broders, Christine Pound, Darrock Whitaker, Dave Slade, and Ed Lodder for assisting in the field, the fishermen in Placentia Bay for their information on Bald Eagle nest locations, and Glen Troke from PWGSC for his support of the project. Richard Elliot, Keith Warnke, Rick Spaul- ding, and Clint Boal provided helpful comments which improved this manuscript. Literature Ch ed Adams, D.C. and C.D. Anthony. 1996. Using randomi- zation techniques to analyze behavioral data. Anim Behav. 51:733-38. Anthony, R.G., M.G. Garret, and G.A. Schuler. 1993. Environmental contaminants in Bald Eagles in the Columbia River Estuary. /. Wildl. Manage. 57:10-19. , A.K. Miles, J.A. Estes, and P.B. Isaacs. 1999. Pro- dirctivity, diets, and environmental contaminants in nesting Bald Eagles from the Aleutian Archipelago Environ. Toxicol. Chem. 18:2054—2062. Argentia Remediation Group. 1995. Phase Ill/phase IV environmental site assessment, US Naval Facility Ar- gentia, Newfoundland: Report 2, site characterization. Report No. 20000-413. Argentia Remediation Group, Dartmouth, NS Canada. Bouton, S.N., PC. Frederick, M.G. Spalding, and H McGill. 1999. Effects of chronic, low concentrations of dietary methylmercury on the behavior of juvenile great egrets. Environ. Toxicol. Chem. 18:1934-1939. Bowerman, W.W. 1993. Regulation of Bald Eagle {Ha- September 2003 CONTAMINANl’S IN BALD EAGI.ES 217 liaeMus leucocephalus) productivity in the Great Lakes Basin: an ecological and toxicological approach. Ph.D. dissertation, Michigan State University, East Lansing, MI U.S.A. , J.R Giesy, D.A. Best, and V.J. Kramer. 1995. A review of factors affecting productivity of Bald Eagles in the Great Lakes region: implications for recovery. Environ. Health Perspect. 103(Suppl. 4):51-59. Braune, B.M., G.M. Donai dson, and K.A. Hobson. 2002. Contaminant residues in seabird eggs from the Ca- nadian Arctic: II spatial trends and evidence from sta- ble isotopes for intercolony differences. Environ. Po- llut. 117:133-145. Cairns, D.K., W.A. Montevecchi, and W. Threlfall. 1989. Researcher’s guide to Newfoundland seabird colonies, 2nd Ed. Memorial Univ. Nfld. Occas. Pap. Biol. 14. Chandi.er, S.K., J.D. Fraser, D.A. Buehler, and J.K.D. Seegar. 1995. Perch trees and shoreline development as predictors of Bald Eagle distribution on Chesa- peake Bay. /. Wildl. Manage. 59:325-332. CoLBORN, T. 1991. Epidemiology of Great Lakes Bald Ea- gles. /. Toxicol. Environ. Health 33:395-453. Deichmann, K.H. and D.B. Bradshaw. 1984. Terra Nova National Park: resource description and evaluation. Parks Canada, Glovertown, NF Canada. DE March, B.G.E., C.A. de Wit, and D.C.G. Muir. 1998. Persistent organic pollutants. Pages 183—371 in S.J. Wilson, J.L. Murray, and H.P. Huntington [Eds.], AIVIAP assessment report: Arctic pollution issues. Arc- tic Monitoring and Assessment Programme, Oslo, Norway. Dominguez, L. 1999. Reproductive success and environ- mental contaminants among Bald Eagles in Placentia and Bonavista Bays, Newfoundland. M.S. thesis, Me- morial University of Newfoundland, St.John’s, NF Ca- nada. Donald.son, G.M. , J.L. Shutt, and P. Hunter. 1999. Or- ganochlorine contamination in Bald Eagle eggs and nestlings from the Canadian Great Lakes. Arch. Envi- ron. Coniam. Toxicol. 36:70-80. Elliott, J.E. 1995. Environmental contaminants in Bald Eagles on the coast of British Columbia: exposure and biological effects. Ph.D. dissertation, University of British Columbia, Vancouver, BC Canada. and RJ. Norstrom. 1998. Chlorinated hydrocar- bon contaminants and productivity of Bald Eagle pop- ulations on the Pacific Coast of Canada. Environ. To- xicol. Chem. 17:1142-1153. Fraser, J.D. , I.D. Frenzel, J.E. Mathisen, F. Martin, and M.E. Sliough. 1983. Scheduling Bald Eagle reproduc- tive surveys. Wildl. Soc. Bull. 11:13-16. Frings, C.S., T.W. Fendley, R.T. Dunn, and C.A. Quenn. 1972. Improved determination of total serum lipids by the sulpho-phospho-vanillin reaction. Clin. Chem. 18:673-674. Furness, R.W. 1993. Birds as monitors of pollutants. Pag- es 86-143 in R.W. Furness and JJ-D. Greenwood [Eds.], Birds as monitors of environmental change Chapman & Hall, New York, NY U.S.A. Gregor, D., J. Stow, D. Kennedy, K. Reimer, and C. Oli- SON. 2003. Local sources of contaminants in the Ca- nadian Arctic. Pages 158-183 in T. Bidleman, R. Mac- donald, and J. Stow [Eds.], Canadian Arctic Contaminants Assessment Report II: sources, occur- rence, trends and pathways in the physical environ- ment. Indian & Northern Affairs Canada, Ottawa ON Canada. Grier, J.W. 1982. Ban of DDT and subsequent recovery of reproduction in Bald Eagles. Science 218:1232- 1235. Hansen, A.J. 1987. Regulation of Bald Eagle reproductive rates in southeastern Alaska. Ecology 68:1387-1392. Hobson, K.A. and W.A. Montevecchi. 1991. Stable iso- topic determinations of trophic relationships of Great Auks. Oerofogia 87:527-531. , J.F. Piatt, and J. Pitogchelli. 1994. Using stable isotopes to determine seabird trophic relationships J Anim. Ecol. 63:786—798. , M. Drever, and G. Kaiser. 1999. Norway rats as predators of burrow-nesting seabirds: insights from stable isotope analysis./. Wildl. Manage. 63:14-25. Hodges, J.I., Jr. 1982. Bald Eagle nesting studies in Sey- mour Canal, southeast Alaska. Cowior 84:125-127. and J.G. King. 1984. Bald Eagle breeding popu- lation survey of coastal British Columbia, f. Wildl Manage. 48:993—998. Howells, D., D. Calamari, J. Gray, and P.G. Wells. 1990. An analytical approach to assessment of long- term effects of low levels of contaminants in the ma- rine environment. Mar. Pollut. Bull. 21:371-375. JWEL. 1996. Marine sampling off landfill 1, former Unit- ed States naval facility, Argentia, Newfoundland JWEL Report No. 84336. Jacques Whitford Environ- ment Ltd., St.John’s, NF Canada. Knight, R.L., PJ. Randolph, G.T. Allen, L.S. Young, AND RJ. WiGEN. 1990. Diets of nesting Bald Eagles, Haliaeetm leucocephalus, in western Washington. Can Field-Nat. 104:545-551. Kozie, K.D. and R.K. Anderson. 1991. Productivity, diet, and environmental contaminants in Bald Eagles nest- ing near the Wisconsin shoreline of Lake Superior Arch. Environ. Coniam. Toxicol. 20:41—48. Kumar, K.S., K. Kannan, J.R Giesy, and S. Masunaga 2002. Distribution and elimination of polychlorinated dibenzo-p-dioxins, dibenzofurans, biphenyls, and p,p'- DDE in tissues of Bald Eagles from the Upper Pen- insula of Michigan. Environ. Sci. Technol. 36:2789- 2796. Kuzyk, Z.A., N.M. Burgess, J.R Stow, and G.A. Fox. 2003. Biological effects of marine PCB contamination on black guillemot nestlings at Saglek, Labrador: liver biomarkers. Ecotoxicology 12:183-197. Ledrew Environmental Management Ltd. and Oceans 218 Dominguez et al. VoL. 37, No. 3 Ltd. 1984. Southeastern Newfoundland marine re- sources: a literature review. Newfoundland and Lab- rador Dept, of Development, St. John’s NF Canada. Lock, A.R., R.G.B. Brown, and S.H. Gerriets. 1994. Gaz- etteer of marine birds in Atlantic Canada: an atlas of seabird vulnerability to oil pollution. Canadian Wild- life Service, Dartmouth, NS Canada. MacDonald, RR.N. and PJ- Austin-Smith. 1989. Bald Eagle, Haliaeetus leucocephalus, nest distribution on Cape Breton Island, Nova Scotia. Can. Field-Nat. 103: 293-296. Matz, A.C. 1998. Organochlorine contaminants and Bald Eagles Haliaeetus leucocephalus in Maine: investigations at three ecological scales. Ph.D. dissertation, Univer- sity of Maine, Orono, ME U.S.A. Mersmann, T.J., D.A. Buehler, J.D. Fraser, and J.K.D. Seegar. 1992- Assessing bias in studies of Bald Eagle food habits, y. Wildl. Manage. 56:73-78. Montevecchi, W.A. and L.M. Tuck. 1987. Newfoundland birds: exploitation, study, conservation. Publ. Nuttall Ornithol. Club 21, Cambridge, MA U.S.A. Neugebauer, E.A., G.I,. Sans Cartier, and B.J. Wake- ford. 2000. Methods for the determination of metals in wildlife tissues using various atomic absorption spectrophotometry techniques. Can. Wildl. Serv. Tech. Rep. 337. Norstrom, RJ. AND H.T. Won. 1985. Long term preser- vation of egg and tissue homogenates for the deter- mination of organochlorine compounds: freezing ver- sus freeze drying./. A.s.soc. Off. Anal. Chem. 68:129-135. Osborne, J.M. 1978. Invertebrate diversity Long Harbour, Placentia Bay, Nfld.: eight years after a pollution crisis. Surveillance Report EPS 5-AR-78-2, Environment Pro- tection Service, Environment Canada, Atlantic Re- gion, St. John’s, NF Canada. Postupalsky, S. 1974. Raptor reproductive success: some problems with methods, criteria, and terminology'. Raptor Res. Rep. 2:21—31. SCHEUHAMMER, A.M. AND D. Bond. 1991. Factors affect- ing the determination of total mercury in biological samples by continuous flow cold vapour atomic ab- sorption spectrophotometry. Biolog. Trace Element Res. 31:119-129. Spalding, M.G., PC. Frederick, H.C. McGill, S.N. Bou- ton, AND L.R. McDowell. 2000a. Methylmercury ac- cumulation in tissues and its effects on growth and appetite in captive great egrets./. Wildl. Dis. 36:411- 422. , , , , L.J. Richey, I.M. Schu- macher, C.G.M. Biackmore, and J. Harrison. 2000b. Histologic, neurologic, and immunologic effects of methylmercury in captive great egrets./. Wildl. Dis. 36: 423-435. Stalmaster, M.V. 1987. The Bald Eagle. Universe Books, New York, NY U.S.A. Stocek, R.F. and P.A. Pearce. 1981. Status and breeding success of New Brunswick Bald Eagles. Can. Field-Nat. 95:428-433. Todd, C.S., L.S. Young, R.B. Owen, and E.J. Gramlich. 1982. Eood habits of Bald Eagles in Maine. /. Wildl. Alanage. 46:636-645. Welch, L.J. 1994. Contaminant burdens and reproduc- tive rates of Bald Eagles breeding in Maine. M.S. the- sis, Univeusity of Maine, Orono, ME U.S.A. Wiemeyer, S.N., R.W. Frenzel, R.G. Anthony, B.R. McCleliand, and R.L. Knight. 1989. Environmental contaminants in blood of western Bald Eagles./. Rap- tor Res. 23:140—146. , C.M. Bunck, and CJ. Stafford. 1993. Environ- mental contaminants in Bald Eagle eggs (1980-84) and further interpretations of relationships to productivity and shell thickness. Arch. Environ. Contam. Toxicol. 24: 213-227. Wiese, F.K. and P.C. Ryan. 2003. The extent of chronic marine oil pollution in southeastern Newfoundland waters assessed through beached-bird surveys 1984- 1999. Mar. Pollut. Bull. 46: (in pre.ss). Wood, P.B., J.H. White, A. Sfeffer, J.M. Wood, C.E. Face- mire, and H.F. Percival. 1995. Mercury concentrations in tissues of Florida Bald Eagles. /. Wildl. Manage. 60: 178-185. Received 26 September 2002; accepted 30 May 2003 Associate Editor: Clint W. Boal J. Raptor Res. 37(3) :219-227 © 2003 The Raptor Research Foundation, Inc, HABITAT AND NEST-SITE USE BY RED-TAILED HAWKS IN NORTHWESTERN WYOMING Roger N. Smith ^ Grand Teton National Park, Division of Science and Resource Management, P.O. Drawer 1 70, Moose, WY 83012 U.S.A. Stanley H. Anderson Wyoming Cooperative Wildlife Research Unit, P.O. Box 3166, University of Wyoming, Laramie, WY 82071 U.S.A. Steven L. Cain Grand Teton National Park, Division of Science and Resource Management, P. O. Drawer 1 70, Moose, WY 83012 U.S.A. Jeffrey R. Dunk^ Grand Teton National Park, Division of Science and Resource Management, P. O. Drawer 1 70, Moose, WY 83012 U.S.A. Abstract. — We described Red-tailed Hawk {Buteo jamaicensis) habitat in the valley portion of Grand Teton National Park during 1992-94 at three spatial scales: (1) nest tree, (2) microhabitat centered around nest trees, and (3) within the home range. Nests were found in six tree species, 22 (36%) were in coniferous and 38 (64%) were in deciduous trees. Nest-tree height, nest height, nest-tree diameter- at-breast-height (DBH), and percentage height of nest in the tree means were 22 m, 16 m, 50 cm, and 72.7%, respectively. Nests were located predominantly on northwestern aspects and on moderately-steep slopes. Microhabitat selection was determined by comparing nest sites with randomly-located, non-nest sites. We measured nest-site inicrohabitat characteristics in 0.04-ha circles centered on nest trees and two paired non-nest sites at 60 independent locations. Compared to non-nest sites, nest-sites had a higher number of trees, taller trees, larger DBH and basal area of trees, and greater canopy cover. Among nest sites, 50 of 60 nest trees were the tallest trees in the plot, and 48 of 60 had the largest diameter. Discriminant function analysis indicated that canopy cover, basal area of trees, and distance to forest edge were the most important variables in distinguishing nest sites from non-nest sites. We calculated cover of six vegetation classes within 100% minimum convex polygon home ranges of seven breeding Red-tailed Hawk pairs, and within a 10 640 ha core study area. The proportion of habitat cover types in home ranges varied considerably among individual pairs. Among all home ranges, 75% were non- fo rested open grassland or shrub and 25% were forested. Habitat characteristics important to Red-tailed Hawks on our study area were mature forests of mixed conifer and deciduous trees on moderately-steep hillsides adjacent to expansive openings consisting of either sagebrush (Artemm« spp.) -grassland flats or remnant agricultural sites. Landscape-scale factors other than vegetation cover-type proportions may be more important in determining home range size and location within our study area. Key Words: Red-tailed Hawk; Buteo jamaicensis; nesting habitat:, habitat use, home range. Grand Teton Na- tional Park; Wyoming. HABITAT Y USO DEL SITIO DEL NIDO DEL GAVILAN DE COLA ROJA EN EL NOROESTE DE WYOMING Resumen. — Describimos el habitat del gavilan de Cola Roja {Buteo jamaicensis) en el valle del Grand Teton National Park durante 1992-94, en tres escalas espaciales: (1) arbol con nido, (2) microhabitat ^ Present address: The Raptor Fund, 1680 Quarterhorse Drive, Jackson, WY 83011 U.S.A.; e-mail address: rogersmith@wyoming.com ^ Present address: Department of Natural Resources Planning and Interpretation, Humboldt State University, Areata, CA 95521 U.S.A. 219 220 Smith et al. VoL. 37, No. 3 centrado alrededor de arboles con nidos y (3) dentro del rango ocupacion. Se encontraron nidos en seis especies de arboles, 22 (36%) fueron en coniferas y 38 (64%) fueron en arboles deciduos. La altura del arbol con nido, la altura del nido, el diametro a la altura del pecho (DAP) del arbol con nido y el porcent^e de la altura del nido de las raedias de los arboles fueron 22 m, 16 m, 50 cm, y 72.7%, respectivamente. Los nidos se localizaron predominantemente en vertientes del noroeste y en pendien- tes moderadamente empinadas. La seleccion de micro habitats se determino comparando los sitios que tenian nidos, con sitios sin nidos localizados al azar. Se midieron caracteristicas de los micro habitats, de los sitios con nido, en circulos de 0.04 ha con centre en los arboles con nido y dos sitios pareados sin nidos en 60 lugares independientes. Al comparar los sitios sin nidos, con los sitios con nido, estos ultimos tuvieron un numero mas alto de arboles, arboles mas altos, DAP y area basal de arboles mas grandes, y mayor cobertura del dosel. Entre los sitios con nidos, 50 de 60 arboles con nido fueron los arboles mas altos en el cuadrante y 48 de 60 tuvieron el diametro mas grande. El analisis de la funcion de discriminacion indico que la cobertura del dosel, el area basal de los arboles y la distancia al borde del bosque fueron las variables mas importantes para distinguir los sitios con nido de los sitios sin nido. Se calculo la cobertura de seis clases de vegetacion dentro del 100% del poligono convexo minimo del rango de ocupacion de 7 parejas del gavilan de cola roja en reproduccion dentro de un area de estudio de 10 640 ha. La proporcion de los tipos de cobertura de habitat en los rangos de ocupacion, variaron considerablemente entre parejas. Entre todos los rangos de ocupacion, el 75% fueron praderas abiertas sin arboles o arbustos y 25% fueron praderas sin arboles. Las caracteristicas de los habitat importantes para los gavilanes de cola roja en el area del estudio, fueron bosques maduros con coniferas y arboles deciduos en laderas moderadamente empinadas, adyacentes a regiones ampliamente pobladas con Ar- temisia spp., pastizales o remanentes de regiones agricolas. Otros factores de la escala del paisaje ademas de las proporciones del tipo de cobertura de la vegetacion podrian ser mas importantes para determinar el tamaho del rango de ocupacion y la ubicacion dentro del area del estudio. [Tradducion de Cesar Marquez] Red-tailed Hawks {Buteo jamaicensis) are widely distributed in North America and breed and winter m a wide variety of habitats (Bent 1937, Palmer 1988, Preston and Beane 1993). Previous studies have quantified Red-tailed Hawk habitat use at multiple scales, including perch sites (Preston 1980), nest-tree sites (Mader 1978, Titus and Moser 1981, Bednarz and Dinsmore 1982, Speiser and Bo- sakowski 1988), on a landscape level (Stout et al. 1998), and within nesting (Janes 1984), and over wintering (Preston 1990) territories. Little is known, however, about Red-tailed Hawk habitat use in the intermountain region of the western United States. During 1990 and 1991 we moni- tored raptor and Common Raven (Corvus corax) productivity along the floodplain forests of the Snake River and its tributaries in Grand Teton Na- tional Park (GTNP). During those two years, we recorded zero fledglings from 12 Red-tailed Hawk nests. We also documented a high density of breed- ing Common Ravens, whose breeding populations had increased dramatically since the 1940s (Dunk et al. 1994). In addition, during 1990-91, we peri- odically found dead hawk chicks below nests. We thus began a broad-based ecological study of Red- tailed Hawks in northwestern Wyoming designed to investigate the causes of low productivity. In par- ticular we evaluated the effects of parasites on Red- tailed Hawk reproduction (Smith et al. 1998) as well as evaluating breeding habitat use patterns of both Common Ravens (Dunk et al. 1997) and Red- tailed Hawks. Herein, we describe our findings of the associations of Red-tailed Hawks to vegetation at three spatial scales: (1) the nest tree, (2) micro- habitat within 0.04 ha vegetation plots centered around nest trees, and (3) within the home range. Study Area The study area was located in northwestern Wyoming within the valley portion of GTNP (43°40'N, 110°43'W), a high mountain valley that is surrounded by the Teton Mountain Range to the west, the Gros Ventre Mountains to the east, the Yellowstone Plateau to the north, and the town of Jackson to the south. We conducted home range analyses within a 10 640 ha core study area that included Moose, Blacktail Butte, and east to the foothills of Shad- ow Mountain. Vegetation, climate, and landscape fea- tures within the study area have been described by Craig- head and Craighead (1956), Dunk et al. (1997), and Smith et al. (1998). Elevation of nests within the study area ranged from 1934—2164 m {N = 74). Meihods The study was conducted from March 1992 to August 1994. Nests were located in April and May by systematic annual searches of the core study area and surrounding areas. Terminology used in this study follow recommen- dations by Postupalsky (1974) and Steenhof and Kochert September 2003 Red-tailed Hawk Habitat Use 221 Table 1. Structural habitat variables measured in 0.04 ha vegetation plots centered on nest trees and at randomly- located non-nest plots. Habitat Variable Description Nest tree spp Nest height Nest DBH Percent nest height Number trees Tree height Canopy cover^ Ground cover"* Shrub cover"* Distance to edge Foliage ht 0-0.3 m’* Foliage ht 0.3-3 m*" Basal area Slope Aspect Species of nest tree Height of nest tree (m), measured with a clinometer Diameter at breast height of nest tree (cm) Height of nest in tree divided by nest-tree height X 100 Total number of trees in plot, measured with a clinometer Height of trees in plot, including nest trees, measured with a clinometer Percent canopy closure in the plot Percent ground cover in the plot Percent shrub cover in the plot Distance from plot center to nearest forest opening Understory foliage density height 1 (0-0.3 m) Understory foliage density heights 2-4 combined (0.3-3 m) Basal area of all trees in plot Slope of plot, measured with clinometer (degrees) Direction of slope, measured with compass ^ Eight transects 45° apart were walked, stopping every 2 m. At each stop we sighted directly above and below through an ocular sighting tube with cross hairs at one end and recorded presence or absence of vegetation. Modified understory foliage density estimated at four levels with a drop cloth (Noon 1981) placed at the edge of the plot in each of four cardinal directions. Density heights 2-4 did not differ significantly; thus, we used the mean of layer 2-4 combined. (1982). An occupied nest refers to a nest where; (1) eggs were laid, or young were raised; (2) we saw fresh green boughs on the nest rim or in the nest with adult hawks present; (3) two adults observed in the nest, either perched or actively maintaining the nest; (4) an adult in incubation posture in the nest. All habitat data were col- lected from occupied nests and the territories they were within. Nest Tree and Microhabitat Characteristics. We mea- sured five nest-tree variables and 12 structural-habitat var- iables (Table 1) at nest sites at least two weeks after fledg- ing and before drying of understory vegetation. We defined the nest site as the 11.3-m radius (0.04 ha) circle centered on the nest tree (Noon 1981). For each nest site, we located two non-nest sites in a random compass direction and distance 25-200 m from the nest and mea- sured the same 12 structural habitat variables (Table 1). We maintained that a non-nest site located within 200 m from the nest was within a Red-tailed Hawk territory and thus was available to a nesting pair. We calculated means for each pair of non-nest site variables and compared them to nest site variables to investigate nest-site affini- ties. Non-nest sites were rejected if they did not contain at least one tree with structure suitable to support a nest. Sampling techniques followed those presented by James and Shugart (1970) and Dunk et al. (1997). Home Range Delineation. Within the 10 640 ha core study area, we delineated Red-tailed Hawk home ranges by direct observation of unmarked male birds at seven breeding territories. Males and females of each breeding pair were distinguished by differences in molting pat- terns, plumage characteristics or size. We used a spotting scope and binoculars to observe movement patterns and continuously recorded time and activity data. Observa- tion periods ranged from 2-5 hr, and were equally dis- tributed between 0500 and 2000 H MST. Observations began in early April and continued through the fledging period or until the nesting attempt failed. We plotted the birds’ position every 10 min and recorded locations on 1:24 000 topographic maps assisted by aerial photo- graphs. The 10-min time interval was chosen to minimize dependency between successive locations (White and Gar- rott 1990). We used the ArcView extension ANIMAL MOVEMENT (Hooge and Eichenlaub 1997), to calculate 100% minimum convex polygon (MCP) home range es- timates. We determined that home ranges defined by 100% MCPs best suited our needs because our primary goal was to define the outer-most boundary of use and describe general vegetation patterns therein. We realized that several potential biases could have af- fected our home range delineations. First, by relying on direct observations of unmarked individual birds, terri- tories could be: (1) underestimated in size as a result of birds sometimes being out of view (Village 1982, Baekken et al. 1987), (2) erroneously described due to failure to identify correctly the focal individual (Altmann 1974), or (3) to record accurately the location on a map of a bird that was greater than 300-400 lateral m from the observ- er. Our study area was conducive to direct observations, however, because it contained relatively open habitat and little topographic relief. We further mitigated potential biasing effects by careful selection of two or more obser- vation points within each home range, and found that on average, birds were out of view during only 7.2% (SE = 1.39, range = 0-66 min) of each observation session Further, males in each territory were easily identifiable from a combination of individual morphological charac- teristics and because of rigorous territorial behavior, 222 Smith et al. VoL. 37, No. 3 0 ^ ^ ^ ^ 0 10 20 30 40 Hours Observation Figure 1. Mean cumulative percentage of home range size determined by hours of observation on seven Red- tailed Hawk territories in Grand Teton National Park, 1993-94. Horizontal bars indicate ± one SD of error. thus, erroneous identification of individuals was probably rare. Territory size could also be underestimated by fo- cusing observations on only one sex of the territorial pair. For example, during courtship, incubation, and early brood-rearing, female Red-tailed Hawk’s movements were largely restricted to the nest site, while the male foraged for both sexes and defended their territory (Bent 1937, Newton 1979, Preston and Beane 1993). De- lineation of home ranges is dependent upon obtaining a sample of locations large enough to accurately represent the area used (e.g., Bekoff and Mech 1984). Therefore, we addressed this by ensuring that for each territory ob- served, an a.symptote in home range size plotted against observation time was obtained (Fig. 1). Home Range Characteristics. Within each home range and the core study area, we used a ground-truthed digital vegetation cover-type database (Despain 1990) to cate- gorize habitat into six classes: (1) coniferous forest, (2) deciduous forest, (3) mixture of deciduous and conifer- ous forest, (4) mixture of timber and shrub, (5) sage/ shrub, and (6) grass/ meadow. A digital elevation model was used to categorize aspect into eight classes and slope in four classes: 1-10%, 11-20%, 21-30%, and 31-40%. Statistical Procedures. Univariate comparison of habi- tat variables at nest and non-nest sites were completed using two-tailed paired Student’s Uests. Comparisons be- tween deciduous and coniferous nest tree characteristics were conducted using two-sample Hest. We used a step- wise discriminant-function analysis (DFA) with a forward- variable selection and a tolerance value of 0.15 to arrive at the best subset of habitat-related predictor variables most useful in separating nest plots from paired non-nest sites. We used a jackknife classification procedure and variables with canonical structure coefficients greater than 0.30 from the first stepwise discriminant procedure in a final discriminant analysis. To assess the effectiveness of the DFA, we calculated Cohen’s Kappa statistics for the classification matrix. This statistic indicates whether the model classified observations into their respective groups significantly better than by chance (Titus et al. 1984). We tested the randomness of the aspect of sites using Wat- son-Williams’s (1956) multisampling testing of angles and Rayleigh’s z-test (Zar 1984) . Due to the small number of observed home ranges, we opted not to test hypothe- ses regarding these data, and present descriptive mea- surements of habitat use, slope, and aspect at the land- scape scale. Descriptive statistics are presented as mean ± SE. Statistical analyses were conducted using Systat software (Wilkinson 1997). Results Nest-tree Characteristics. We located 22, 29, and 23 occupied nests in 1992, 1993, and 1994, respec- tively. Of these 74 occupied nests, 12 were reoc- cupied twice during the three years of study, and two nests were occupied all three years. Habitat data from nests that were occupied in more than one year were included in our analyses only once. Nests were located in a variety of habitats, includ- ing riparian bottoms, timbered hillsides consisting of mixed conifer and deciduous trees, and in trees along irrigation ditches located in remnant agri- cultural areas. We found Red-tailed Hawk nests in six tree species; 22 (36.6%) were in coniferous and 38 (63.3%) were in deciduous trees (Table 2). One nest was in a snag. Two other tree species found in our study area, but not used by nesting Red- tailed Hawks were limber pine {Pinus flexilis) and subalpine fir {Abies lasiocarpa) . Mean nest-tree height and mean nest height were higher in co- niferous than deciduous trees {t — 6.67, P< 0.001, X = 26.98 m vs. x — 18.15 m, and = 6.05, P < 0.001, X — 20.95 m vs. X — 13.62 m), respectively. We found no difference in nest-tree DBH or per- cent height of the nest in the tree between conif- erous and deciduous nest trees. Douglas Fir {Pseu- dotsuga menziesii) nest trees were taller than the other five nest-tree species (ifi 54 = 13.2, P< 0.001, X — 30.8 m, SE — 1.32; Tukey HSD multiple com- parisons). Mean nest diameter and depth were 81.3 cm (SE = 0.43, range = 75-88 cm) and 39.0 cm (SE = 0.91, range = 29-61 cm) respectively {N = 50 nests). There was no difference in nest size (diameter) among the three years (P] 47 = 0.78, P = 0.46). Microhabitat Characteristics. Eifty of 60 (83.3%) nest trees were the tallest trees in the nest-site plot. Nest trees had the largest DBH in 48 of 60 (80%) plots. Mean aspect of slope at nest plots and non- nest plots differed from a uniform distribution September 2003 Red-tailed Hawk Habitat Use 223 Table 2. Mean characteristics of Red-tailed Hawk nests {N = 60) in northwest Wyoming, 1992—94 (standard errors in parentheses). Tree Species Percent Number of Total Trees Nest Trees Tree Diameter ( cm) Tree Height ( m) Nest Height ( m) Percent Tree Height* ; Mean SE Mean SE Mean SE Mean SE Quaking aspen {Populus tremuloides) 22 37.5 33.4 1.8 17.6 0.84 12.7 0.67 72.2 (0.01) Narrowleaf cottonwood {Populus angustifolia) 16 25.8 63.5 6.3 19.0 1.5 15.0 1.3 78.1 0.02 Douglas-fir {Pseudotsuga menziesii) 10 16.1 70.6 4.8 30.8 1.3 24.8 1.3 80.5 0.02 Lodgepole pine {Pinus contorta) 6 9.6 38.2 5.1 24.3 1.5 18.2 1.3 75.1 0.03 Blue spruce (Picea pungens) 4 6.4 57.6 4.9 23.3 1.9 17.7 1.6 76.0 0.02 Engelman spruce {Picea engelmannii) 2 3.2 46.5 3.5 23.3 0.4 16 1.0 69.0 5.1 All species 60 50.2 2.8 21.4 0.81 16.3 0.73 76.1 0.01 ^Nest height divided by tree height X 100. (Rayleigh’s z = 27.3 and 43.3, respectively; P < for five variables (Table 3) . We found no difference 0.001), and the two sample means did not differ between distances to forest edge at nest sites versus from each other (T^ijys = 3,89, P > 0.25; Watson- random sites from our univariate Kests. However, Williams test for two samples; Zar 1984), The mean discriminant function analysis included distance to slope aspect for the nest and non -nest plots was edge : along with canopy cover and basal area in a 347°. Nests were typically situated on 10-20% model that best explained the differences between slopes. Univariate paired Mests analyses indicated nest and non-nest plots (Wilk’s Lamdba - 0.82, P differences between nest plots and non-nest plots < 0.0001). Using the reduced model derived by Table 3. Mean microhahitat characteristics in 0.04 ha plots centered around Red-tailed Hawk nest sites (N = 60) and available sites (N = 118) in Grand Teton National Park, 1992-94. Nest-Site Non-Nest Site Statistic Variables Mean SE Mean SE t P Percent Canopy cover® 51.7 2.4 34.4 2.4 6.8 <0.01 Basal (mV0.04 ha) 0.92 0.09 0.53 0.05 5.3 <0.01 Diam. breast ht® 29.8 1.9 24.9 1.0 2.9 0.01 Tree height® 14.1 0.67 12.5 0.47 2.7 0.01 Number of trees 13.3 1.2 11.1 0.91 2.1 0.03 Ground cover® 72.2 2.6 74.2 2.1 0.50 0.61 Shrub cover® 20.4 2.3 21.5 2.1 0.71 0.47 Foliage ht 0-0.3 m 68.8 3.6 70.5 3.1 0.58 0.56 Foliage ht 0.3-1 m 39.6 3.1 36.4 2.7 1.0 0.32 Foliage ht 1-2 m 26.9 2.9 24.9 2.4 0.67 0.50 Foliage ht 2-3 m 26.5 2.9 24.1 2.3 0.86 0.39 Distance to edge 21.7 3.2 22.9 3.4 0.24 0.81 Aspect^ NW (347°) — NW (347°) — — — Percentage slope 22.6 3.2 19.1 5.7 0.23 0.81 Mean of plot means. '’Cardinal Direction, (mean azimuth, Watson-Williams 1956). 224 Smith et al. VoL. 37, No. 3 Table 4. Two-group discriminant function analysis of Red-tailed Hawk nest-site characteristics compared to non-nest (available) sites in northwest Wyoming for years 1992-94 combined. Group Means (SE) Structure Variable Nest-Plot Non-Nest Plot Coeeficients Canopy cover 51.7 (2.4) 34.4 (2.4) 0.83 Basal area 0.92 (0.09) 0.53 (0.05) 0.41 Distance to forest edge 21.7 (3.2) 22.9 (3.4) 0.44 Sample size Wilk’s lambda 60 0.82 118 Approx. F-statistic 12.5^ Cohen’s kappa Percent correct classification 43/60 = 71.6 4.9 82/118 = 69.5 Overall % correct classification 70.2 P < 0.0001 the stepwise DFA (Table 4) , jackknife classification procedures successfully predicted 71 % of nest sites and 69% of non-nest sites. The overall correct clas- sification was 70.2%, which differed only slightly from a random classification rate (Titus et al. 1984). Home Range Characteristics. On average it took 25 hr (range = 11.6-44.3, SE — 4.7, N = 7) of observation to delineate 90% of the home range, and 32 hr (range = 13.1-49.2, SE = 5.5, N = 7) for 100% delineation (Fig. 1). Total observation time by nest ranged from 24.5-73.5 hr {x = 47.4, SD = 17.1). Mean MCP home range size was 241 ha (SD = 119 ha, range = 181-480 ha; Table 5). This mean was similar to the mean of 225 ha (SD = 1.0 ha, range = 77-385 ha) obtained for eight nesting pairs of Red-tailed Hawks on the same study area in 1947 (Craighead and Craighead 1956). Among the seven home ranges, the per- centage of non-forested cover ranged from 48- 95% (x = 76%). Also, we observed considerable variation in the percentage cover of shrubland and grass meadow vegetation types (5-83% and 4-63%, respectively), while the total percentage cover of each were relatively equal (Table 5). The core study area was characterized by 76% open vegeta- tion, primarily sagebrush {Artemisia tridentata), bit- terbrush {Purshia tridentata), grasslands, and rem- nant agriculture, and 24% forested vegetation that included both coniferous and deciduous trees (Ta- ble 5). Relative to the core study area, four of seven home ranges contained considerably more grass/ meadow cover type, and five of seven contained considerably less sage/shrub cover type. Similarly, two of seven home ranges contained slopes greater than 11% (range = 0—26%). The core area con- tained more area of south and southeast aspects than the combined home ranges. Discussion We observed Red-tailed Hawks nesting only in trees. In another Wyoming study area, 40% {N — 10) of Red-tailed Hawks nested on cliffs (MacLaren et al. 1988). Most, (83%) of the nests in our study area were situated high in forest canopy, in the top one-fourth of the tree, and in the tallest tree in the immediate area. These findings corroborate other studies that suggest Red-tailed Hawks select nest- sites that afford unobstructed access and a high viewpoint of the surrounding area (e.g., Orians and Kuhlman 1956, Bednarz and Dinsmore 1982, Preston and Beane 1993, Stout et al. 1998). Dis- tance to forest edge and slope may also be impor- tant factors. For example, Speiser and Bosakowski (1988) reported Red-tailed Hawk nests significantly closer to forest openings than random plots in both young and mature hardwood forests, report- ing a mean distance to forest edge of 155 m. Con- versely, Titus and Mosher (1981) detected no dif- ference in distance to edge between Red-tailed Hawk nest plots and random plots in western Maryland, and they reported a mean distance of 235 m to the nearest forest edge. Like Titus and Mosher (1981), we detected no difference between distances to forest edge at nest sites versus random sites. However, on average, nests in our study area were within only 15 m of large patches of sage- brush, grasslands or remnant cultivated fields, sug- gesting an attraction of open areas. Several studies have reported nest sites situated September 2003 Red-tailed Hawk Habitat Use 225 Table 5. Minimum Convex Polygon (MCP) home range sizes (100%) and vegetation cover for seven breeding pairs of Red-tailed Hawks in northwestern Wyoming, 1992-93. Vegetative Cover in Hectares (%) Nest Territory Days Obs. Hours Obs. MCP (ha) Non- Forested^^ Forested CONIE. Forest Decid. Forest Mixed Forest Forest/ Shrub Sage/ Shrub Grass/ Meadow Green 92 7 26.4 180.9 146 34.9 9,3 1.3 17 7.3 51.6 94.4 (80.7) (19.3) (5.1) (0.7) (9.4) (4.0) (28.5) (52.2) Hunter 92 7 24.5 162.6 78.1 84.6 6.4 33.8 1.5 42.9 37.7 40.4 (48) (52) (3.9) (20.8) (0.9) (26.4) (23.2) (24.8) Lobo 92 10 31.5 480.1 373.3 106.9 32.0 51.7 0 23.2 355.3 18 (77.7) (22.3) (6.7) (10.8) (4.8) (74) (3.7) Moulton 93 15 73.5 318.1 299.5 18.6 0 18.6 0 0 263.8 35.7 (94.5) (5.5) (5.9) (82.9) (11.3) Blacktail 93 10 41.6 145.0 99.8 45.1 34.4 1.3 3.4 6.0 7.9 91.9 (68.9) (31.1) (23.7) (0.9) (2.3) (4.1) (5.5) (63.4) S. Green 93 12 57.2 203.8 177 26.7 8.4 0 14.5 3.8 40.6 136.4 (86.9) (13.1) (4.1) (7.1) (1.9) (19.9) (67) Wilson 93 14 60 196.7 154 42.6 15.4 9.6 10.7 6.9 47.6 106.4 (78.3) (21.7) (7.8) (4.9) (5.4) (3.5) (24.2) (54.1) Mean 10.7 47.4 241 192.3 51.3 15.1 16.6 6.7 12.9 117.5 74.7 (76.4) (23.5) (7.3) (6.3) (3.6) (6.4) (36.9) (39.5) SE 3.1 17.1 119.3 42.0 12.2 5.0 7.4 2.7 5.7 52.8 16.5 (6.1) (5.6) (2.9) (2.8) (1.4) (3.4) (11.1) (9.7) Totals 75 322.1 — 1327.7 359.4 105.9 116.3 47.1 90.1 823.1 523.2 (78.7) (21.3) Study area — — — 7750.8 2805.3 1203.1 664.6 211.9 725.7 4970 2651.6 (72.8) (26.4) (11.3) (6.2) (2.0) (6.8) (46.7) (26.9) “* Non-forested areas (ha) within the study area also included: barren ground = 25, burned = 15, water = 88, and urban = 86 ha on sloped hillsides (Titus and Mosher 1981, Bed- narz and Dinsmore 1982, Speiser and Bosakowski 1988). Titus and Mosher’s (1981) data implied that Red-tailed Hawks select for nest sites on steep hill- sides, and Speiser and Bosakowski (1988) suggest- ed that Red-tailed Hawks may gain better soaring capability and an easier, downward departure flight from the nest on steeper hillsides relative to nests on flat terrain. Although we found no difference in slope between nest and non-nest sites, nests were typically situated on moderately-sloped (x — 22%) hillsides, which also facilitated nest access and views of the surrounding terrain. We also regularly observed Red-tailed Hawks actively foraging while perched at the nest and tending nestlings (i.e., searching and stooping from nests during prey capture attempts), suggesting that foraging effi- ciency may be important in nest site selection. We believe that, in addition to tree height, both prox- imity to edge and slope were important for Red- tailed Hawks in our study area for the reasons dis- cussed above, but that our technique for locating random plots (within 200 m of nest sites) may have biased our comparisons for these variables. Indeed, it is likely that the site-specific combination of tree height, slope, and distance to edge, as well as nest- tree branch structure, determines nest access. Nest and non-nest sites in our study area were found predominately on north aspects and similar slopes (Table 3), however, this similarity may be a function of the relatively close proximity of a non- nest plot (within 200 m) from a nest tree. Two eco- logical factors may be relevant here. First, in our study area’s temperate and relatively arid environ- ment, moisture regimes associated with slope effect (Knight 1994) favored more well-developed forest- ed areas on north aspects, simply providing the stands with more abundant large trees that were selected by hawks. Second, nests on north aspects may provide some protection from our study area’s prevailing southwest winds and associated rain and snow. Similarly, several other studies have noted the importance of aspect in reducing possible neg- ative effects of prevailing weather on nestling rap- 226 Smith et al. VoL. 37, No. 3 tors (Newton 1979, Titus and Mosher 1981, Bed- narz and Dinsmore 1982, Speiser and Bosakowski 1988, Preston and Beane 1993, Stout et al. 1998). We observed large variation in the proportion of habitat types used among the seven home ranges. The absence of strong vegetation preferences, oth- er than open grass/meadow cover type, at the landscape scale may indicate that: the number of home ranges sampled were inadequate, or upon arrival onto territories, these birds are first select- ing habitat from the scale of the nest tree and the microhabitat site second. If the hawks are selecting habitat at a larger scale, it appears to be smaller than that of the entire home range. Bingham and Noon (1997) found “core areas” of ca. 200 ha to be used much greater than other areas within the home ranges of Northern Spotted Owls (Strix oc- adentalis). Perhaps Red-tailed Hawks also select habitat at some intermediate scale between area immediately surrounding the nest and the home range boundary. For Red-tailed Hawks, it appears that the availability of adequate nesting trees and their proximity to foraging areas are important characteristics that influence their selection of breeding areas. Selection of open habitat with trees or forested components is a consistent finding in other studies of Red-tailed Hawk habitat (see Pres- ton and Beane 1993), although Bednarz and Dins- more (1982) investigated Red-tailed Hawk habitat features within 3.14 km^ “home ranges” in Iowa and found home ranges typically contained equal proportions of woodland, pasture and croplands. But consistent with other studies, Bednarz and Dinsmore argue that open areas, particularly pas- ture land, provided the key foraging habitats for nesting Red-tailed Hawks in Iowa. The importance of specific nest-site and home range variable comparisons among multiple studies IS diminished in the absence of similar comparisons of larger landscape patterns. Other components, in- cluding vegetative structure and physiognamy, food availability (Luttich et al. 1970, Marquiss and New- ton 1981), and the presence of territorial conspe- cifics and other competitors (Rothfels and Lein 1983), each undoubtedly influence nest-site and home range selection. For future evaluations of Red-tailed Hawk habitat, it would seem prudent to focus more effort at spatial scales between: (1) the microhabitat scale and the home range, and their influence on Red-tailed Hawk populations, espe- cially where human-caused landscape changes are likely to occur (Craighead and Mindell 1981), and (2) landscape features surrounding the home range. Nevertheless, when considering this and oth- er Red-tailed Hawk habitat studies, common pat- terns emerge that underscore the importance of nest access, and open areas associated with forest/ non-forest ecotones. Acknowledgments We conducted this study with funds provided by Grand Teton National Park and the University of Wyoming. We are especially grateful to Tom McFetters for his field as- sistance and companionship during the 1993 and 1994 field seasons. We also thank E. Anderson, M. Baptiste, L. Garlman, M. Creel, A. Harvey, G. Horne, G. Montopoli, M. Reid, G. Smith, R. Whalen, and Y. Wolfers, for their assistance during this study. We thank Derek Craighead and Joel Berger for editorial comments on earlier drafts, and C. Preston, H. Garner, and an anonymous referee for making very helpful comments on the manuscript. Literature Cited Altmann, J. 1974. Observational study of behavior: sam- pling methods. Behavior 49:227—265. Baekken, B.T., L.O. Nvbo, and G.A. Sonerud. 1987. Home-range size of Northern Hawk Owls: depen- dence on calculation method, number of tracking days, and number of plotted perchings. Pages 145- 148 in R.W. Nero, RJ. Clark, R.J. Knapton, and R.H. Hamre [Eds.], Biology and conservation of northern forest owls. United States Forest Service General Technical Report RM-142. Bent, A.C. 1937. Life histories of North American birds of prey. Part. 1. U.S. Nat. Mus. Bull. No. 167. Bekoff, M. and L.D. Mech. 1984. Simulation analyses of space use: home range estimates, variability, and sam- ple size. Behav. Res. Meth. Instr. Comp. 15:32-37. Bednarz, J.C. andJJ. Dinsmore. 1982. Nest-sites and hab- itat of Red-shouldered and Red-tailed Hawks in Iowa. Wilson Bull. 94:31-45. Bingham, B.B. and B.R. Noon. 1997. Mitigation of hab- itat “take”: application to habitat conservation plan- ning. Conserv. Biol. 11:127-139. Craighead, JJ- AND F.C. Craighead. 1956. Hawks, Owls, and Wildlife. Stackpole Co., Harrisburg, PA and Wild- life Management Inst., Washington, DC U.S.A. Craighead, F.C. and D.P. Mindell. 1981. Nesting raptors in western Wyoming, 1947 and 1975./. Wildl. Manage 45:865-872. Despain, D.C. 1990. Cover types of Grand Teton National Park. Electronic Arcinfo grid file. Grand Teton Na- tional Park, Moose, WY U.S.A. Dunk, J.R., S.L. Cain, M.E. Reid, and R.N. Smith. 1994 A high breeding density of common ravens in north- western Wyoming. Northwest. Nat. 75:70—73. , R.N. Smith, and S.L. Cain. 1997. Nest-site selec- tion and reproductive success in Common Ravens. Auk 114:116-120. September 2003 Red-tailed Hawk Habitat Use 227 Hooge, P.N. and B. Eichentaur. 1997. Animal move- ment extension to Arcview, Ver. 1.1. Alaska Biological Science Center, U.S. Geological Survey, Anchorage, AK U.S.A. James, F.C. and H.H. Shugart. 1970. A quantitative meth- od of habitat description. Audubon Field Notes 24::7 21- 736. Janes, S.W. 1984. Influences of territory composition and interspecific competition of Red-tailed Hawk repro- ductive success. Ecology 65:862—868. Knight, D.H. 1994. Mountains and plains: the ecology of Wyoming landscapes. Yale Univ. Press, New Haven, CT U.S.A. Luttich, S., D.H. Rusch, E.C. Meslow, and L.B. Keith, 1970, Ecology of Red-tailed Hawk predation in Alber- ta. Ecology 51:190-203. MacLaren, P.A., S.H. Anderson, and D.E. Runde. 1988. Food habits and nest characteristics of breeding rap- tors in southwestern Wyoming. Great Basin Nat. 48: 548-553. Mader, W.J. 1978. A comparative nesting study of Red- tailed Hawks and Harris’ Hawks in southern Arizona. Auk 95:327-337. Marquiss, M. and I. Newton. 1981. A radio-tracking study of the ranging behavior and dispersion of Eu- ropean Sparrowhawks. /. Anim. Ecol. 51:111-133. Newton, I. 1979. Population ecology of raptors. Buteo Books, Vermillion, SD U.S.A. Noon, B.R. 1981 . Techniques for sampling avian habitats. In D.E. Capen [Ed.], The use of multivariate statistics in studies of wildlife habitat. USDA, USES, General Technical Report RM-87. Rocky Mountain Forest and Range Experimental Station, Fort Collins, CO U,S.A. Orians, G. and F. Kuhlman. 1956. Red-tailed Hawk and horned owl populations in Wisconsin. Condor 58:571- 385. Palmer, R.S. 1988. Handbook of North American birds, Diurnal Raptors Part 2. Yale Univ. Press, New Haven, CT U.S.A. POSTUPALSKY, S. 1974. Raptor reproductive success: some problems with methods, criteria, and terminology. Pages 21-31 mF.N. Hamerstrom, Jr., B.E. Harrell, and R.R. Olendorff [Eds.], Management of raptors. Rap- tor Research Report No. 2. Raptor Research Foun- dation, Inc., Vermillion, SD U.S.A. Preston, C.R. 1980. Differential perch site selection by color morphs of the Red-tailed Hawk {Buteo jamaicen- sis). Auk 97:782-789. . 1990. Distribution of raptor foraging in relation to prey biomass and habitat structure. Condor 92: 107- 112 . AND R.D. Beane. 1993. Red-tailed Hawk {Buteo ja- maicensis). In N. Poole and F. Gill [Eds.], The birds of North America, No. 52. The Academy of Natural Sci- ences Philadelphia, PA and The American Ornithol- ogists’ Union, Washington, DC U.S.A. Rothfels, M. and M.R. Lein. 1983. Territoriality in sym- patric populations of Red-tailed and Swainson’s hawks. Can. J. Zool. 61:60-64. Smith, R.S., S.L. Cain, S.H. Anderson, J.R. Dunk, and E.S WiLLLAMS. 1998. Blackfly-induced mortality of nes t ling Red-tailed Hawks. Auk 115:368-375. Speiser, R. and T. Bosakowski. 1988. Nest site preferenc- es of Red-tailed Hawks in the highlands of southeast- ern New York and northern New Jersey, y. Field Ormt- hol. 59:361-368. Steenhof, K. and M.N. Kochert. 1982. An evaluation of methods used to estimate raptor nesting success. J Wildl. Manage. 46:885-893. Stout, W.E., R.K. Anderson, and J.M. Papp. 1998. Urban, suburban and rural Red-tailed Hawk nesting habitat and populations in southeast Wisconsin./. Raptor Res 32:221-228. Titus, K. and J.A. Mosher. 1981. Nest-site habitat select- ed by woodland hawks in the central Appalachians Auk 98:270-281. , , and B.K. Williams. 1984. Chance-cor- rected classification for use in discriminant analysis: ecological applications. Am. Midi. Nat. 111:1-7. Village, A. 1982. The home range and density of kestrels in relation to vole abundance. /. Anim. Ecol. 51:413— 428. Watson, G.S. and E.J. Williams. 1956. On the construc- tion of significance tests on the circle and the sphere. Biometrica 38:330—336. White, G.C. and R.A. Garrott. 1990. Analysis of wildlife radio-tracking data. Academic Press, San Diego, CA U.S.A. Wilkinson, L. 1997. SYSTAT 8.0: the system for statistics. SYSTAT, Inc., Evanston, IL U.S.A. Zar, J.H. 1984. Biostatistical analysis. Prentice-Hall, En- glewood Cliffs, NJ U.S.A. Received 7 October 2002; accepted 3 June 2003 J Raptor Res. 37(3) :228-235 © 2003 The Raptor Research Foundation, Inc. NEST AND ROOST HABITAT CHARACTERISTICS OF THE GREY-FACED BUZZARD IN NORTHEASTERN CHINA Wen-Hong Deng^ Ministry of Education Key Laboratory for Biodiversity Science and Ecological Engineering, College of Life Sciences, Beijing Normal University, Beijing, 100875, China Gao Wei School of Life Sciences, Northeast Normal University, Changchun, 130024, China Zheng Guang-Mei Ministry of Education Key Laboratory for Biodiversity Science and Ecological Engineering, College of Life Sciences, Beijing Normal University, Beijing, 100875, China Abstract. — The habitat of the Grey-faced Buzzard {Butastur indicus) has diminished substantially be- cause of forest management for timber production and farmland reclamation in recent years. An un- derstanding of the characteristics of nest and roost sites of this bird is important for its conservation. We studied Grey-faced Buzzards during their breeding season in Zuojia Nature Reserve, Jilin Province, China, from March 1996-August 1998. This species selected both nesting and roosting sites in mixed- deciduous forests that contained Korean larch {Larix olgens) more frequently than available in the study area. Most nests were located in Korean larches and in Chinese pines {Pinus tabulaeformis; 70%), whereas nests in broadleaf trees were relatively infrequent (30%). Eight nests (75%) were located on the upper third of a slope, three nests (19%) were located on the middle third, and one nest (6%) on the lower third. Buzzards used 11 tree species for roosting; however, (58%) of all roosts were located in three tree species (Korean larch, Chinese pine, and River birch \_Betula nigra]). Higher canopy closure and taller trees best separated roost sites from random plots with a discriminant analysis. Mean roost height was 9.6 ± 0.5 m. Roost trees averaged 11.2 ± 0.6 m in height with mean DBH (diameter at breast height) of 16.9 ± 0.3 cm. Buzzards selected nest and roost sites in forests characterized by tall (>12 m) conifers, hardwood understory, and high canopy closure (>70%). Key Words: Butastur indicus; habitat use, Grey faced Buzzard; nest site, nest tree, roost site. CARACTERISTICAS DEL HABITAT DE LOS DORMIDEROS Y DE ANIDACION DEL BUITRE DE CARA GRIS DEL NORESTE DE CHINA Resumen. — El habitat del buitre de cara gris {Butastur indicus) ha disminuido substancialmente en los ultimos ahos, a causa del manejo del bosque para la produccion de madera y la recuperacion de tierras para agricultura. Una comprension de las caracteristicas de los sitios de nido y de los dormideros de esta ave, es importante para su conservacion. Estudiamos el buitre de cara gris durante su temporada de cria en la Reserva Natural de Zuojia, Provincia de Jilin, China, desde marzo del 1996 hasta agosto del 1998. Esta especie escogio sitios de anidacion y dormideros en los bosques deciduos mixtos que contenian una mayor frecuencia de Larix olgens en el area del estudio. l.a mayoria de los nidos se localizaron en Larix olgens y Pinus tabulaeformis; 70%, mientras que los nidos que estaban en arboles de hojas anchas fueron relativamente poco frecuentes (30%). Ocho nidos (75%) fueron localizados en el tercio superior de una pendiente, tres nidos (19%) fueron localizados en el tercio medio, y un nido (6%) en el tercio inferior. Los buitres utilizaron 11 especies de arboles como dormideros; sin embargo el 58% fueron localizados en tres especies de arboles Larix olgens, Pinus tabulaeformis y Betula nigra. La cobertura del dosel y los arboles mas altos fueron separados de los dormideros mediante la utilizacion del analisis discriminante. La media de la altura de los dormideros fue de 9.6 ± 0.5 m. Los arboles que sirvieron como dormideros tuvieron una altura promedio de 11.2 ± 0.6 m, con media DAP (diametro a la altura del pecho) de 16.9 ± 0.3 cm. Los buitres escogieron nidos y dormideros en bosques carac- ^ E-mail address: dengwh@bnu.edu.cn 228 September 2003 Grey-faced Buzzard Eagle Habitat 229 terizados por altas coniferas (>12 m), arboles del sub-dosel, de madera dura con copas cerradas y alias (>70% alto de dosel). [Traduccion de Cesar Marquez] The large number of published reports that de- scribe habitat selection and habitat characteristics among birds attests to the enormous variation ob- served and to the biological importance of this top- ic. In addition to its importance, an understanding of the response by birds to environmental habitat change is necessary before conservation strategies can be developed and implemented (Schmutz 1989). A theoretical framework for habitat selec- tion has been provided by Fretwell and Lucas (1970), and functional and theoretical aspects of habitat selection have been summarized by Cody (1985). Within these frameworks, nest and roost characteristics are very important factors related to avian habitat selection. Among subtropical birds, raptors are one of the least-studied groups and relatively little is known about their nest and roost characteristics. Grey- faced Buzzard (Butastur indicus) is a summer mi- grant in northeast China (Cheng 1987, Anony- mous 1988). It seems that most of the Grey-faced Buzzards that breed in northeastern China migrate to Okinawa, Taiwan, the Philippines, Indonesia, Malaysia or nearby regions to winter (Chang 1980, Cheng 1987, Ehimekensibu et al. 1989, Deng et al. 1997). This buzzard has been listed as a threatened species in National Second Class Protected Species in China Data Book of Endangered Animals (Zheng and Wang 1998). The habitats of this spe- cies have been substantially reduced because of forest management for timber production and farmland reclamation in recent years (Zheng and Wang 1998). An understanding of the characteris- tics of nest and roost sites of this bird is particularly important for its conservation. However, very little is known about nesting ecology of Grey-faced Buz- zards either in China or in other areas. In this pa- per we describe the nest and roost characteristics of this raptor in northeastern China. Our null hy- potheses were: (1) no difference exists between nesting sites and randomly-placed plots within the study area, (2) no difference exists between the immediate nest site and general habitat within nest stands, and (3) no difference exists between roost- ing sites and randomly-placed plots within the study area. Study Area The study area, ca. 84 km^ in size, was located in Zuojia Natural Protection Area and included the Tumengling Mountains and Zhujia Mountains ranging from the east- ern ChangBai Mountains to the western plain (126°!'- 127°2'N, 44°6'-45°5'E). Elevation at the site ranged from 200-500 mask The climate is east monsoon, character- ized by hot, dry summers and cold, snowy winters. The vegetation within the study area was quite diverse, al- though the existing forest is secondary. The most com- mon trees present in the study area were Mongolian oak {Quercus mongolica), dahur birch {Betula daxmrica), Man- churian linden {Tilia mandschurica) , Japanese elm {Ulmus japonica) , Scotch pine {Pinus sylvestris) , Korean larch {Fi- nns koraiensis) and Masson pine {Finns massoniana) (Deng et al. 1997). In the study area, hawthorn raspberry {Rubus crataegifolius) , dahurian rose {Rosa dahurica), Ko- rean rose {Rosa doreana), willowleaf spiraea {Spiraea sali- cifolia), ural falsespiraea {Sorbaria sorbifolia), prickly rose {Rosa acicularis), amur barberry {Berberis amurensis), amur honeysuckle {Lonicera maacki), manchur honeysuckle {Lonicera ruprechtiana) , and Sakhalin honeysuckle {Loni- cera maximowiczii) dominated the shrub layer. The study area consists of ca. 30% open habitat and 70% forest habitat. Methods Survey Methods. We surveyed the study area at least four times each year, 20 March-20 August, 1996—98, us- ing conventions suggested by Newton and Marquiss (1982) and Steenhof (1987) to describe occupancy and activities at nesting sites. We attempted to locate all buz- zard nests within the study area. We used behavioral cues and systematic searches of potential nest substrates to lo- cate nests. We determined the sex of buzzards by body size and plumage characteristics (Deng et al. 1997). We established six transects in the study area (x = 3 km, range = 2-5 km) . Each transect was 50 m wide and par- allel to the forest edge. Each transect was divided into 100 m segments from one end to the other. Sites were surveyed within a 4-hr period beginning 30 min after sun- rise by walking along each side, with 5-min stops at each 100 m interval. We walked along one side and came back along another one. A nesting area was considered occu- pied if a territorial pair or evidence of a territorial pair (such as observations of an incubating bird, nest con- struction, or nest maintenance) was observed, otherwise the area was classified as unoccupied. We located roosts (the perch location where a bird spends the night) by observing buzzards at roost sites. Only roost sites where buzzards remained stationary upon initial detection were used in analyses. We recorded eagle nest and roost lo- cations with GPS receivers and plotted these on geologic survey maps to the nearest 10 m using Universal Trans- verse Mercator (UTM) coordinates. Most of our habitat-sampling protocol was adopted from Seamans and Gutierrez (1995). We measured site characteristics in sample plots centered on buzzard nest 230 Deng et al. VoL. 37, No. 3 trees and roost trees and at random locations in forest habitat. Detailed vegetations were collected in 12 nest stands and 12 random plots. Nesting habitat data were collected at four sample points 25 m from each nest in each cardinal direction. Also, four sample points were located in each cardinal direction and at a random dis- tance between 100 m and 1000 m from the nest tree to represent available habitat. This sampling was designed to address nest-site level habitat selection occurring with- in a hypothetical Grey-faced Buzzard home range. In each nest stand, four sample points were measured with- in both nesting and available habitat to increase the num- ber of points available for use in the assessment of model stability (Mueller-Dombois and Ellenberg 1974). We cat- egorized the forest type (conifer, if the proportion of co- nifers was >70%; mixed conifer/broadleaf, if the pro- portion of conifer and broadleaf was near equal; broadleaf forest, if the proportion of broadleaf was >70%) and slope position (lower, middle, upper third) at each site, and measured 12 habitat characteristics. At each plot center, we estimated slope aspect with a com- pass, slope angle (%) with a clinometer, and relative can- opy closure (%) with a spherical densiometer. We mea- sured tree DBH (diameter at breast height in cm) with calipers, and tree height with a clinometer. We used a varible radius-plot method (Mueller-Dombois and Ellen- berg 1974) to estimate basal area (m^/ha) of conifers and broad-leaf trees. In addition, we recorded nest and roost tree species, percent height (roost height/tree height) relative to the height of the tree, and distance from nest trees to roost trees. We used the variance of tree height and variance of tree diameters of all trees in each sample plot as an index of forest structural hetero- geneity. Data Analysis. We pooled data among years after find- ing no difference using a series of Kruskall Wallis tests (Zar 1984), with sequential Bonferroni adjustments (Rice 1990). We compared forest types and slope position of eagle sites and random sites using chi-square analysis. We estimated the mean slope aspect of eagle nests and roosts using circular statistics (Batschelet 1981), and compared aspect of eagle nests and roosts vdth random sites using a Watson-Williams test (Zar 1984). We assessed univariate normality of the variables using skewness, kurtosis, and probability plots. We assessed the equality of variance of variables between groups using an T-max test. We used logarithmic and square-root transformations to normal- ize variables and equalize variances for variables that de- viated from normal distribution. For analyses, we only used those variables which approximated a normal dis- tribution and had comparable variances between groups, either before or after transformation. We tested the null hypotheses of no difference in variable means between eagle nest and roost plots and random plots using mul- tivariate analysis of variance (MANOVA, Stevens 1996). For the MANOVA, we used Wilk’s Lambda to compare linear combinations of variables between used and ran- dom sites. If the MANOVA was significant, we tested in- dividual variables using a series of t tests with sequential Bonferroni adjustments. We used discriminant analysis (DA, Stevens 1996) to model data, to estimate which characteristics contributed the most to differences be- tween groups. We used cross validation (Capen et al. 1986) to evaluate the stability of the DA model. We took detailed measurements of all nest trees (age, height, DBH, canopy and nest height) using the same techniques used for measuring trees in sample plots. We estimated nest tree age by extracting a core sample with an increment borer and counting the rings. We used cir- cular statistics to estimate mean orientation of the nest relative to the tree trunk. We used a chi-square analysis to test for differences in tree species between nest and random tree distributions. We used paired-sample t tests (Zar 1984) to compare nest tree height and DBH to ran- dom trees within nest stands. Results Roost-site Characteristics. Of 86 roost sites we measured, we used 58 (one each for 30 males and 28 females) as independent samples for analysis. The distribution of forest types at Grey-faced Buz- zard roosts differed from random sites throughout the study area (x^ = 9.06, df = 2, P < 0.05), with most roosts in the mixed conifer/ river birch forest type. Position of roosts sites on the slope differed from a random distribution (x^ = 9.06, df = 2, P < 0.05). Forty-seven roosts (81%) were located on the middle third, eight (14%) roosts were on the upper third, and three roosts were (5%) on the lower third of the slopes. Mean aspect at roost sites differed from aspect at random sites (P = 8.94, df = 1, 112, P < 0.05; mean aspect — 6.8°, mean vec- tor length = 0.60, angular deviation = 68.5°). Buz- zards used 11 tree species for roosting, however over half (58%) of all roosts were located in three tree species (Korean larch, Chinese pine, and river birch). Mean roost height was 9.6 ± 0.5 m in the roost tree. Roost trees averaged 11.2 ± 0.6 m tall in height with mean DBH of 16.9 ± 0.3 cm. Roost plots differed from random plots (MA- NOVA; Wilk’s Lambda = 0.51, F - 16.8, df = 10, 109, P < 0.05). The ^-tests indicated most variables differed between roost and random plots (Table 1 ) . Higher canopy closure and taller trees best sep- arated roosts from random plots in the DA (Table 2) . The pooled DA correctly classified 78.6% of the roost and random plots (Table 3). Nest-site Characteristics. We found 12 Grey- faced Buzzard nests and used all nest sites as in- dependent samples for analysis (Fig. 1). The dis- tribution of forest types at nests differed from random sites (x^ = 15.62, df = 2, P < 0.05), with most nests in the mixed-conifer/river birch forest type. Position of nest sites on the slope differed from an independent distribution (x^ “ 11.26, df = 2, P< 0.05). Eight nests (75%) were located on September 2003 Grey-faced Buzzard Eagle Habitat 231 Table 1. Habitat characteristics at Grey-faced Buzzard roost sites (A^ = 58) and random plots (N = 58) in the Zuojia and Tumengling mountains, northeastern China, 1996-98. Variable Mean ± Roost SD Random^ P P Tree height (m) 13.8 1.5 9.1 -h 1.2 5.63 <0.01 Tree DBH (cm) 28.7 11.3 21.2 8.6 2.84 0.01 Tree basal area (m^ha“b 10.6 -h 4.9 9.3 + 5.8 0.98 0.46" Canopy closure (%) 81.9 H- 18.5 62.4 H- 13.1 8.56 <0.01 Tree height variance 1.9 -1- 1.6 0.8 -t- 0.5 4.29 <0.01 Tree DBH variance 3.9 -h 1.6 2.1 + 0.9 2.91 0.01 Distance from water (m) 348.6 ± 57.9 401.2 H- 345.5 0.42 0.61 Tree crown volume (m^) 12.2 + 5.5 7.9 + 3.8 3.21 <0.01 Slope angle (%) 22.5 11.4 13.6 H- 6.8 8.27 <0.01 Random sites were located throughout the study area by using CIS. Degrees of freedom = 98. ^ No significant difference. the upper third of the slope, three nests (19%) were located on the middle third, and one nest (6%) on the lower third. Mean slope aspect at nest sites was northerly (mean aspect = 295°, mean vec- tor length — 0.37, angular deviation = 56.5°), and differ significantly from random sites (F = 15.13, df = 1, 22, P< 0.05). Nest plots differed from random plots (MANO- VA; Wilk’s Lambda = 0.68, F - 7.96, df = 8, 23, P < 0.05) . The Rests indicated that four of the eight variables differed between nest and random plots (Table 3) . Larger DBH, taller trees, greater canopy closure, and greater basal area of mature trees best separated nest sites from random sites in the DA (Table 3). Nest-tree Characteristics. Grey-faced Buzzard nests were located in five tree species. Fifty percent {N — 6) of nests were located in Korean larches, 25% (N = 3) were in Chinese pines, and 8% each were in river birch, Mongolian oak, Manchurian linden {Tilia mandschurica) , respectively. Mean as- pect deviation of the nests in the trees was north- westerly (mean aspect — 342.5°, mean vector length = 0.51, angular deviation — 56.5°), but did not differ from a random distribution (z = 1.2, T* > 0.05). The distribution of random tree species differed from nest tree species (x^ — 8.9, df — 2, P < 0.05). Nest trees were larger, denser, and taller than trees randomly located within the nest stand (Table 4). Table 2- Mean habitat characteristics at Grey-faced Buzzard nest {N = 12) and random plots {N = 12) in the Zuojia and Tumengling mountains, northeastern China, 1996-98. Variable Mean ± Nest SD Random^ P P Tree height (m) 15.1 -h 2.2 10.6 + 1.7 8.24 <0.01 Tree DBH (cm) 31.9 ± 9.2 25.2 -h 6.7 3.13 0.01 Tree basal area (m^ha^^) 17.6 5.1 9.6 -h 3.7 4.94 <0.01 Canopy closure (%) 84.5 ± 15.6 62.7 22.0 8.76 <0.01 Tree height variance 1.5 -1- 1.1 1.3 -1- 0.6 0.87 0.42" Tree DBH variance 3.2 ± 1.8 3.1 0.8 0.69 0.49" Distance from water (m) 229.5 -h 112.6 314.2 258.6 0.28 0.87" Tree crown volume (m^) 13.8 6.9 8.1 -h 3.8 3.47 <0.01 Slope angle (%) 25.5 8.5 21.3 -H 7.1 1.22 0.06" ^ Random sites located in each cardinal direction from nest at a random distance between 100 and 1000 m. Degrees of freedom = 94. No significant difference. 232 Deng et al. VoL. 37, No. 3 Table 3. Discriminant analysis results of habitat characteristics at Grey-faced Buzzard roost and nest plots in the Zuojia and Tumengling mountains, northeastern China, 1996-98. Variable Roost Plots {N = 58) Nest Plots (N = 12) Mean Rank Mean Structure Coefficient^ Pooled Data Structure Coefficient"^ Mfan Rank Mean Strucfure Coefficient"* Pooled Data Structure Coefficient"* Tree height (m) 1.9 0.57 0.64 2.5 0.54 0.62 Tree DBH (cm) 6.3 0.21 0.33 6.6 0.29 0.28 Tree BA (m^ha"b 8.7 0.16 0.13 9.2 0.11 0.09 Canopy closure (%) 1.2 0.76 0.81 1.0 0.80 0.83 Tree height variance 2.7 0.53 0.58 3.6 0.35 0.59 Tree DBH variance 4.9 0.29 0.45 5.7 0.39 0.36 Distance from water (m) 9.9 0.04 0.03 9.9 0.07 0.01 Tree crown volume (m^^) 3.1 0.42 0.53 3.9 0.55 0.44 Slope angle (%) 1.3 0.68 0.77 1.8 0.68 0.72 “ Structure coefficient is correlation between a single variable and discriminant function. Discussion Most Grey-faced Buzzard nest and roost sites were found on the upper third of north-facing slopes. This corresponded to the distribution of mature mixed conifer/ river birch forests on the study area. In addition, most nest and roost sites had an understory of Mongolian oak, which con- figure 1. The study area and locations of the nest sites (solid squares indicate nest location) of the Grey-faced Buzzard in Zuojia Nature Reserve, northeastern China. tributed to the forest structure. Raptors have shown some selection for slopes, but slope orien- tation patterns were not always consistent (Klopfer 1965, Delannoy and Cruz 1988, McIntyre and Ad- ams 1999, Nijman et al. 2000). In selecting nest sites, Grey-faced Buzzards avoid slopes with south- ern aspects. We suggest that the birds are selecting exposures to insulate the nest against hot weather conditions during incubation and to place their nests close to hunting habitat. Data from accipiter studies (Shuster 1980, Moore and Henny 1983, Speiser and Bosakowski 1987) also show an obvious avoidance of southern slope aspects for nesting. The majority of nests (75%) were built in coni- fers rather than in deciduous hardwood trees. Grey-faced Buzzards preferred to build their nests in dense canopy closure of Korean larches and Chinese pines (Table 4). Broad-leaf trees were rarely used as nest trees despite the preference for broad-leaf trees in nest stands. We propose that broad-leaf trees are partly avoided because they sel- dom have a larger triple and quadruple primary crotches. All nest trees were generally greater in DBH and height when compared to those random trees (Table 4). Large raptors require large tree- forks to place the nest in (Newton 1979, Mader 1982). Grey-faced Buzzard nests were always situ- ated below or in the bottom quarter of the tree canopy, which had an open branch structure. Rap- tors nest in the lower quarter of the canopy with open branch structure probably to allow the buz- zards access to the nest both above and within the canopy; this positioning may also hide the nest September 2003 Grey-faced Buzzard Eagit Habitat 233 Table 4. Characteristics of Grey-faced Buzzard nest (N = 12) and random trees {N = 12) in the Zuojia and Tu- mengling mountains, northeastern China, 1996-98. Variable Mean ± Nest Trees SD Random Trees^ A P-Vaiue Age (yr)'^ 56.2 ± 5.1 54.8 ± 3.4 0.16 0.91 Tree height (m) 18.8 ± 3.1 14.3 ± 2.6 4.23 <0.01 Tree DBH (cm) 32.5 ± 9.7 24.9 ± 9.2 3.44 <0.01 Tree crown volume (m^) 3.6 ± 1.1 2.4 ± 0.7 3.16 <0.01 Nest height (m) 12.5 ± 3.4 — — — Random trees were located a random distance (10-100 m) in a random direction from the nest tree, t values from matched pair test, significant at a = 0.05. Estimated by extracting at core sample with an increment borer and counting the rings. from potential predators (Selas 1996, Malan and Robinson 2001), Cerasoli and Penteriani (1996) suggested that tree-nesting raptors might select trees for their size and structural features, such as a tall and open canopy, that allow unobstructed access to nests. In this study, however, Grey-faced Buzzard seem to prefer dense to open canopy for nesting. In our study, Grey-faced Buzzards selected nest and roost sites primarily in mixed conifer/river birch forests in the largest and tallest trees, with relatively high canopy closure. Forests composed of larger trees with high variation in tree heights may provide an accessible prey base for buzzards and provide protection from potential disturbance from human activities. This kind of vegetation structure allows more small animals to occur and also prevents humans from entering. In addition, Grey-faced Buzzards are heat intolerant and may require mature, multi-storied forests to thermoreg- ulate effectively (Feng 1991, Deng et al. 1997). The middle partition of north-facing slopes, forested with multi-storied mixed conifer habitat, may pro- vide suitable cooler microclimates for buzzards. However, habitats such as hardwood forests seem to lack the complete vertical structure of most typ- ical nest and roost sites. Trees frequently used for roosting were often those with dense foliage or high canopy closure. In contrast, available trees rarely or never used for roosting appeared to provide little cover. In addi- tion to concealing birds from potential human dis- turbance, the dense cover of most roost sites prob- ably provided a favorable microclimate. We found that individuals often used the same roost site on successive days, although our presence may have disturbed the buzzards. Although we climbed up each nest tree to measure the characteristics of nes- tlings in every five days during nestling period, the buzzards did not change their roost sites. Belthoff and Ritchison (1990) suggested that Eastern Screech-Owls (Otus asio) did not use the same roost site on successive days, and suggested reuse of sites could attract potential predators. However, repeat- ed use of roost sites by Grey-faced Buzzards is very common, probably because they have few preda- tors in the study area (Deng et al. 1997). Also, suit- able roosts with dense foliage may be extremely limited in the study area. The forest tracts inhabited by buzzards were large in extent as exemplified by the comparatively long distances (>15 km) to human habitation (Deng et al. 1997). Nest sites were also more often found in dense forests, where little disturbance of any kind occurred and that may have had a greater density and diversity of prey species than edge ar- eas (Kojima 1982, Ricketts and Ritchison 2000). However, we found most nest sites closer to logging roads (or discernable trails) than random sites: six nests (50%) were very close (<30 m) and four (33%) were within 60 m. In our study, forests roads often represented the break in deep contiguous forests. In dense contiguous forests, logging roads may aid the buzzards by providing open flyways. On several different occasions, we observed buz- zards flying, perching, and plucking prey along logging roads in the study area. Speiser and Bosa- kowski (1987) found Northern Goshawks (Acciptter gentilis) often nested near logging roads in north- ern New Jersey and southern New York. Raptors often nest near logging roads or in an exposed po- sition that allows easy access to and from the nest to deliver sticks and prey. Moore and Henny (1984) pointed out the im- 234 Deng et al. VoL. 37, No. 3 portance of past experience (success or failure) in nest site selection by raptors, but at least for first-' time nesters, the role of early experience and im- printing to the natal habitat may be of primary im- portance (Newton 1979). For a variety of nonpasserines, studies of marked individuals have shown that nesting sites are more likely to be re- occupied in years following successful nesting at- tempts and abandoned after nesting failures (e.g., Newton and Marquiss 1982, Marks 1986, Thors- trom and Quixchan 2000) . We did not quantify for- aging habitat of the buzzards in our study area. Available information indicated that Grey-faced Buzzards commonly forage in open areas (e.g., pas- tures, marshes, paddy fields) where they find most of their prey (Kojima 1982, Ehimekensibu et al. 1989). Their foods mainly include frogs, reptiles, rodents, and some birds (Cheng 1987, Ching et al. 1989, Severinghaus 1991). According to our obser- vations, Grey-faced Buzzards often hunt from perches, typically at a top of dead tree; once de- tected the buzzards then dive down to capture prey. Because our study did not involve marked Grey-faced Buzzards among years, whether the same individual birds reoccupied nests is open to question. Future studies should examine Grey- faced Buzzard Eagles in more detail and should address the nature and extent of individual varia- tion in habitat use. Additional factors such as prey, density, prey accessibility, and competition with other raptors need to be addressed in future stud- ies of the nestling ecology of Grey-faced Buzzards. Acknowled gments We thank Liu Yang, Yan-hui Li, Ren-kai Song, Lu-ye Tang, Ya-mei Yu for help in the field. Wei-dong Luo and Bao-xue Zhang were very patient in helping with data collection. Earlier drafts of this manuscript were im- proved by comments from Zheng-wang Zhang and Yan- yun Zhang. The staff of Zuojia Natural Protection Area provided facilities during the study period. We especially thank the editors and the anonymous referees of the Jour- nal of Raptor Research for their excellent and helpful re- marks on the manuscript. Literature Cited Anonymous. 1988. A study of the northward migration of Grey-faced Buzzard Eagles passing PaKua Moun- tain in the spring. Ann. Taipei Wild, Bird Soc. 12:12- 18. Batschelet, E. 1981. Circular statistics in biology. Aca- demic Press, San Francisco, CA U.S.A. Belthoff, J.R. and G. Ritchison. 1990. Roosting behav- ior of postfledging Eastern Screech-Owls. Auk 107: 567-579. Capen, D.E., J.W. Fenwick, D.B. Inkley, and A.C. Boyn- ton. 1986. Multivariate models of songbird habitat in New England forests, Pages 171-176 in]. Verner, M.L Morrison, and C.J. Ralph [Eds.], Wildlife 2000: mod- eling habitat relationships of terrestrial vertebrates. Univ. of Wisconsin Press, Madison, WI U.S.A. Cerasoli, M. and V. Penteriani. 1996. Nest-site and ae- rial point selection by Common Buzzards {Buteo buteo) in Central Italy./. Raptor Res. 30:130-135. Chang, W.F. 1980. Travellers who do not need visas, the Grey-faced Buzzard Eagle. Environment 1:58-64. Cheng, Tso-hin. 1987. A synopsis of the avifauna of Chi- na. Science Press, Beijing, China. Ching, H.C., S.T. Yu, and C.H. Chiang. 1989. A survey of Grey-faced Buzzard Eagle hunting in ManChou Area. KenTing National Park, Taipei, China. Cody, M.L. 1985. Habitat selection in birds. Academic Press, New York, NY U.S.A. Delannoy, C.A. AND A. Cruz. 1988. Breeding biology of the Puerto Rican Sharp-shinned Hawk {Acdpiter stria- tus Venator). Auk 105:649-662. Deng, W.H., X. Wang, and W. Gao. 1997. Breeding hab- itat selection of the Grey-faced Buzzard Eagle./. North- east Normal Univ. 9:57-61. Ehimekensibu, Kochikensibu, Oitakensibu, and Miyaza- kikensibu. 1989. The migration route of Grey-faced Buzzard Eagle Butastur indicus from Shikoku to Kyu- shu. Strix 6:125-131. Feng, H.L., G.Q. Xiang, and K.Q. Hang. 1991. Obser- vation on breeding Grey-faced Buzzard Eagle. Chi- nese Avian Studies, No. 3., Chinese Science Press, Bei- jing, China. Fretwell, S.D. and H.L. Lucas. 1970. On territorial be- havior and other factors influencing habitat distribu- tion in birds. I. Theoretical development. Acta Biotheor 9:16-36. Kojima, Y 1982. Territory and territorial behavior of the Grey-faced Buzzard Eagle Butastur indicus. Tori 30: 117-147. Klopfer, P. 1965, Behavioral aspects of habitat selection, a preliminary report on foliage preferences of birds. Wilson Bull. 77:376-381. Mader, W.J. 1982. Ecology and breeding habitats of the Savanna Hawk in the Llanos of Venezuela. Condor 84. 261-271. Malan, G. and E.R. Robinson. 2001. Nest-site selection by Black Sparrowhawks Acdpiter melanoleucus: impli- cations for managing exotic pulpwood and sawlog for- ests in South Africa. Environ. Manage. 28:195-205. Marks, J.S. 1986. Nest-site characteristics and reproduc- tive success of Long-eared Owls in southwestern Ida- ho. Wilson Bull. 98:547-560. McIntyre, C.L. and L.G. Adams. 1999. Reproductive characteristic of migratory Golden Eagles in Denali National Park, Alaska. Condor 101:115-123. Moore, K.R. and C.J. Henny. 1983. Nest site character- September 2003 Grey-faced Buzzard Eagle Habitat 235 istics of three coexisting Accipiter hawks in northeast Oregon. Raptor Res. 17:65-76. AND J. Henny. 1984. Age specific productivity and nest site characteristics of Cooper’s Hawks. Northwest Sci. 58:290-299. Mueller-Dombois, D. and H. Ellenberg. 1974. Aims and methods of vegetation ecology. John Wiley and Sons, New York, NY U.S.A. Newton, I. 1979. Population ecology of raptors. Buteo Books, Vermillion, SD U.S.A. and M. Marquiss. 1982. Fidelity to breeding area and mate in sparrowhawks Accipiter nisus. J. Anim. Ecol. 5:327-341. Nijman, V., S. Van Balen, and R. Sozer. 2000. Breeding biology of Javan Hawk-Eagle Spizaetus bartelsi in west Java, Indonesia. Emu 100:125-132. Rice, W.R. 1990. A consensus combined p-value test and the family-wide significance of component tests. Bio- metrics 46:303-308. Ricketts, M.S. and G. Ritchison. 2000. Nesting success of yellow-breasted Chats: effect of nest site and terri- tory vegetation strucure. Wilson Bull. 112:510-516. SCHMUTZ, J.K. 1989. Hawk occupancy of disturbed grass- lands in relation to models of habitat selection. Condor 91:362-371. Seamans, M.E. and R.J. Gutierrez. 1995. Breeding hab- itat of the Mexican Spotted Owl in Tularosa Moun- tains, New Mexico. Coiirfor 97:944-952. Selas, V. 1996. Selection and reuse of nest stands by spar- rowhawk {Accipiter nisus) in relation to natural and manipulated variation in tree density./. Avian Biol. 56-62. Severinghaus, L.L. 1991. The status and conservation of Grey-faced Buzzard-Eagles and Brown Shrikes migrat- ing through Taiwan. Pages 203-223 in T. Salathe, [Ed.]. ICBP Tech. Publ. 12. Taiwan, China. Shuster, W.C. 1980. Northern Goshawk nesting require- ments in the Colorado Rockies. West. Birds 11:89-96 Speiser, R. and T. Bosakowski. 1987. Nest-site selection by Northern Goshawks in northern New Jersey and southeastern New York. Condor 89:387-394. Steenhof, K. 1987. Assessing raptor reproductive success and productivity. Pages 157-170 in B.A. Giron Pen- dleton, B.A. Millsap, K.W. Cline, and D.M. Bird [Eds.], Raptor management techniques manual. Sci- entific and Tech. Series No. 10, Nad. Wildl. Fed., Washington, DC U.S.A. Stevens, J. 1996. Applied multivariate statistics for the social sciences, 3rd Ed, Lawrence Erlbaum Publishers, Mahwah, NJ U.S.A. Thorstrom, R. and a. QuixchAn. 2000. Breeding biol- ogy and nest site characteristics of the Bicolored Hawk in Guatemala. Wilson Bull. 112:195-202. Zar, J.H. 1984. Biostatistical analysis. Prentice-Hall, En- glewood Gliffs, NJ U.S.A. Zheng, G.M. and Q.S. Wang. 1998. China red data book of endangered animals. Chinese Science Press, Bei- jing, China. Received 2 October 2002; accepted 18 May 2003 J. Raptor Res. 37(3) :236— 243 © 2003 The Raptor Research Foundation, Inc. FOOD HABITS AND FORAGING ECOLOGY OF AMERICAN KESTRELS IN THE SEMIARID EORESTS OE CENTRAL ARGENTINA Jose HernAn Sarasota^ Miguel Angel SantillAn and Maximiliano AdriAn Galmes Centro para el Estudio y Conservacion de las Aves Rapaces en la Argentina, Avda. Uruguay 151, FCEyN — Universidad Nacional de La Pampa, 6300 Santa Rosa, La Pampa, Argentina Abstract. — The annual diet of the American Kestrel {Falco sparverius) was studied by analyzing 705 pellets and prey remains collected in a semiarid forest area of central Argentina. The diet composition fluctuated seasonally but was composed primarily of invertebrates, followed by small mammals, birds, and reptiles. Based on biomass, vertebrate prey was the major component of the diet in winter and spring. We found significant numerical differences between prey consumed by adult and nestling kes- trels. Percentages of small mammal prey were higher in the diet of adults than in the nestlings, but rodents consumed by nestlings were larger in size than consumed by adults. Nestlings also ate more birds and reptiles than adults. Based in biomass contribution of vertebrate and invertebrate prey to the diet during non-breeding and breeding seasons, American Kestrels in semiarid forest of Central Argen- tina were more carnivorous than insectivorous predators. Differences between adult and nestling diets emphasized the importance of collecting samples from perches and nests when kestrel food habits are being studied during the breeding season. Keywords: American Kestrel, Falco sparverius; /ood habits', breeding season-, semiarid forest. Central Argentina. HABITOS DE ALIMENTACION Y ECOLOGIA DE FORRAJEO DE CERNICALOS AMERICANOS EN LOS BOSQUES SEMIARIDOS DEL CENTRO DE ARGENTINA Resumen. — Estudiamos la composicion anual de la dieta del Cernicalo Americano {Falco sparverius) en un area de bosque semiarido del centro de Argentina a traves del analisis de 705 egagropilas y restos presas. La composicion de la dieta fluctuo estacionalmente pero fue compuesta en primer orden por invertebrados, seguidos en importancia por mamiferos, aves y reptiles. En terminos de biomasa, los vertebrados fueron predominantes en la dieta en invierno y primavera. Encontramos diferencias signi- ficativas entre las presas consumidas por los cernicalos adultos y sus polios. Los porcentajes de ocurren- cia de pequehos mamiferos fueron mayores en la dieta de los adultos, pero los roedores consumidos por los polios fueron mas grandes. Ademas los polios consumieron un mayor numero de aves y de reptiles que los adultos. Tomando como base el aporte en biomasa de vertebrados e invertebrados durante las estaciones reproductivas y no-reproductivas, el Cernicalo Americano se presenta como un predador carnivoro en lugar de insectivoro para los bosques semiaridos del Centro de Argentina. Las diferencias observadas entre las dietas de los adultos y los polios demuestran la importancia de conducir muestreos tanto en perchas como en nidos cuando se estudia el habito alimenticio de este cernicalo durante la estacion reproductiva. [Traduccion de los autores] The American Kestrel {Falco sparverius) is widely distributed in North, Central, and South America inhabiting many different habitats and landscapes (Brown and Amadon 1968). In Argentina, the kes- trel occupies subtropical forests in the north through shrubland steppes in Patagonia (Narosky ^ Present address: Department of Applied Biology, Esta- cion Biologica de Dohana (CSIC), Avda. Maria Luisa s/n, Pabellon del Peru, 41013 Sevilla, Spain; e-mail ad- dress: sarasola@ebd.csic.es and Yzurieta 1989, De la Pena and Rumholl 1998). The ecology and behavior of kestrels have been studied extensively (e.g., Roest 1957, Balgooyen 1976, Koplin et al. 1980, Rudolph 1982, Smallwood 1987, Varland et al. 1993). The food habits of this species have been documented in various habitats, but most of these studies have been carried out in Central and North America (see Heintzelman 1964 for a review, Jenkins 1970, Balgooyen 1976, Cruz 1976, Collopy and Koplin 1983). In the southern- most extreme of its range, the diet of kestrels has 236 September 2003 Food Habits of American Kestrels 237 been examined only in Chile (Greer and Bullock 1966, Yahez et al. 1980, Simonetti et al. 1982) and there is one anecdotal record on the stomach con- tents from northeast Argentina (Beltzer 1990). All of these studies have quantified the diets, but no attempt has been made to assess the diet based on prey biomass. In addition, research has focused on the diet either during the winter or the breeding season, and no systematic study of the annual var- iation diet within a particular habitat has been completed. Similarly, studies in diet of kestrels dur- ing the breeding season have employed several dif- ferent approaches (Lamore 1963, Balgooyen 1976), but none include a comparative analysis be- tween prey consumed by adults and nestlings. Here, we report the food habits of the American Kestrel in the semiarid forest of central Argentina and analyze the numerical and biomass contribu- tions of vertebrate and invertebrate prey to the diet. We examine seasonal variations in the diet and show the differences in prey consumed by adults and nestlings during the breeding season. Study Area This study was conducted in Parque Luro Reserve (36°55'S, 64°16'W), located in La Pampa province of cen- tral Argentina. The reserve (7604 ha) consists mainly of xerophytic open forests of calden {Prosopis caldenia) which represents the characteristic landscape of the Es- pinal biome in the semiarid pampas of Argentina (Ca- brera 1976), Forest areas in the reserve differ structurally, m part because of soil features, but mostly because of the effects of human disturbance. Stipa spp. are tbe domi- nant herbaceous species, and Condalia microphylla, Lycium chilense, L. gillesianum, and Schinus fasciculatus are the common shrub species when a middle stratum is present. Fieldwork was conducted in the tourist area of the re- serve (450 ha) where forest fragmentation has been greater due to the clearing of native vegetation for the development of tourist and recreation facilities. Broad, open areas of natural Stipa grassland are common in this part of the reserve. Habitats surrounding the reserve consist of agricultural areas planted with crops, and pe- rennial and annual pastures. Historically, these semiarid forests have been characterized as having hot summers and cold winters with low humidity and low annual rain- fall, typically concentrated in spring and summer. How- ever, unusually excessive rains have occurred during the last 11 years and the mean annual precipitation has in- creased to 791 ± 336 mm {N = 11). Methods We collected American Kestrel pellets monthly from March 2000 to February 2001. Pellets and prey remains were collected beneath kestrel perches throughout the year and from bve nest boxes occupied by pairs of kes- trels during the breeding season. Pellets from adult kes- trels were collected in perches located in three of the five territories where nest boxes were occupied. Pellets and prey remains found in nest boxes were assumed to represent the diet of nestlings and collected during the post-hatching period until the time when young kestrels left. Only fresh and compact pellets were collected both from perches and nest boxes. Pellets collected were hydrated and broken apart by hand and remains of prey items were separated for iden- tification. Mammals and lizards were identified on the basis of skulls and dentaries using reference collections and keys (Pearson [1995] for mammals and Cei [1986] for lizards) to the species and subfamily levels, respec- tively. Arthropods were identified using mandibles, heads, elytra, and any other parts that allowed identifi- cation to the subfamily level. In order to estimate the minimum number of individual prey in each sample, skulls were counted for mammal prey; skulls, legs, and feet were used for birds; skulls and tails were used for lizards; and whole heads, feet, elytra and mandibles were used for insects. When only hairs, bones, feathers, or scales were found in a pellet, one individual prey was counted in the sample and classified as unidentified. To compute prey biomass, we obtained the mean body mass of adult small mammal species from Tiranti (1992) and Tiranti (unpubl. data) provided data for the body mass of juvenile tuco tucos (Ctenomys spp.). The mean body mass of the bird species was obtained from Flora (1933). Unidentified passerine body mass was calculated as the arithmetic mean of the known species that we re- corded in the diet (Flora 1933, Salvador 1988, 1990). Liz- ard and arthropod biomass were obtained from speci- mens collected in the study area in three pit-fall traps (Corn 1994). Mean weights were calculated from six liz- ards {Teius spp. and Cnemidophorus spp.), 56 orthopter- ans, seven mantids (Mantidae), 46 ants (Formicidae), 18 homopterans, 37 coleopterans, 21 spiders (Araneae) and 15 scorpions (Bothruridae) . For unidentified mammals, reptiles, and birds, we assigned them the average mass of prey contained in each taxa for the same sample. We used chi-square analysis (Zar 1996) to test differ- ences in diet between seasons and between the diets of adults and nestling kestrels during the breeding season. For this analysis, prey taxa were grouped into four major classes (vertebrates, Orthopterans, Coleopterans, and others) and their frequencies were pooled by season (21 March-21 June = Autumn, 21 June-21 September = Winter, 21 September-21 December = Spring, and 21 December-21 March = Summer). Correlation analysis was used to examine the relationship among percentage of Coleopterans and Orthopterans in the diet. These per- centages were subjected to an angular transformation prior to analysis (Zar 1996). Levin’s index of food niche breadth (Marti 1988) was calculated for each sample as follows: B = l/S]Li where p^ is the proportion of prey in different categories To compare diet breadth between samples with different numbers of prey categories, we calculated the standard- ized food niche breadth following Colwell and Futuyma (1971). Bg|-j^ (Bobs Bmin)/ (Bjnax B^vin) j whcCC is the minimum niche breadth possible {N = 1), B„bs = number of prey types observed, and B^^^ ~ A! This index ranged from 0—1. In order to evaluate differences in size of prey consumed by adult and young American Kestrels, 238 Sarasola et al. VoL. 37, No. 3 we calculated the mean prey size by summing the prod- ucts of the number of individual prey items and their mass and then dividing by the total number of prey in the sample. Our statistical test comparing this parameter was based on the assumption that prey sizes consumed were normally distributed. Comparative studies have demonstrated that pellet analysis is subject to a number of biases for specific raptor species (e.g., Mersmann et al. 1992). Although no similar studies have been conducted to assess these biases in the analysis of the diet of the American Kestrel, our results must be considered potentially biased in representing some prey types in the diet. Results During the study period, we collected 284 pellets and 421 prey remains. A total of 3127 prey items were identified from three vertebrate and six in- vertebrate classes. Arthropods comprised about 93% of the total prey items followed by rodents (6%), birds (1%), and reptiles (<1%). In terms of biomass, rodents accounted for 47% of the diet followed by insects (42%), birds (8%), and reptiles (3%). Seasonal Diet Composition. We found signifi- cant differences in the diet of American Kestrels between seasons (x^ = 256, df = 9, P < 0.01). Or- thopterans, mainly grasshoppers, were the most common prey item in all the seasons, but their oc- currence in the diet varied seasonally (Table 1). The percentage of Orthopterans in diet was nega- tively, but non-significantly, correlated with the per- centage of Coleopterans (r = —0.90, N — 4, P = 0.09; Fig. 1), which were only important in winter and spring when the percentage of Orthopterans decreased in the diet. The greatest food niche breadth and mean prey size were recorded during the breeding season due to the frequent occur- rence of mammals, birds, and reptiles in the diet. The biomass of vertebrate prey was higher than for invertebrates in winter (64.4%) and spring (68.6%; Fig. 2), but the biomass of Orthopteran prey was higher (70.9%) in summer than all other prey groups combined. Diets of Adult and Nesding Kestrels. On a nu- merical basis, we found significant differences be- tween the diets of nestling and adult kestrels dur- ing the breeding season (x^ = 85.2, df = 3, P < 0.01) (Table 2). The standardized food niche breadth of nestling was greater than that of adults, but the mean prey size was approximately the same for both age groups (Table 2). Although similar percentages of invertebrates were found in both diets, nestlings consumed more birds and reptiles than their parents, while adults ate comparatively more rodents than nestlings. In addition, a parti- tioned analysis of the standardized food niche breadth resulted in similar values for adults and nestlings for invertebrate prey (adult breadth = 0.20 vs. nestling breadth = 0.21), but a greater in- dex for nestlings when we considered only verte- brate prey (adult breadth = 0.22 vs. nestling breadth — 0.39). Considering the mean size of ver- tebrate prey, prey consumed by nestlings were also larger than prey consumed by adults (36.1 g and 26.8 g, respectively). Discussion We found American Kestrels in semiarid forests of central Argentina to be generalized predators of invertebrate and small vertebrate animals. Insects occurred in the diet throughout the year with de- creasing percentages of one insect prey offset by the increased ingestion of another insect group, making the total percentage of arthropods almost constant over the seasons. Based on our observa- tions of insects and small vertebrates in the study area, we suggest that these fluctuations in inverte- brate prey consumed were due to seasonal changes in their availability rather than due to prey selec- tion by the kestrels. From 2225 orthopterans and coleopterans collected seasonally throughout the study period, only 6% of them were recorded dur- ing winter while 47% of the total were collected in summer. The kestrels preyed most heavily on large prey such as vertebrates, only in winter when the prey were breeding. As far as rodents, adults preyed most heavily on tuco tucos (juvenile body mass = 80 g) which they delivered to their young, and on vesper mice (Calomys spp.; body mass — 16 g), which they ate themselves. Even though all of the tuco tucos consumed were juveniles, they were considerably heavier when compared to the kes- trel’s body mass and with the body masses of avail- able alternative mammalian prey. Similar prey sizes have been reported by Lamore (1963) and Balgoo- yen (1976), both of whom analyzed the diets of kestrels during the breeding season. Selection of different-sized rodent prey during the breeding season may be due to the different energy require- ments of adult and nestling kestrels. Balgooyen (1976) found that a brood of four nestling kestrels needed almost twice the prey mass required by their parents between hatching and post-fledging (3973 g for four nestlings vs. 2142 g for the pair of September 2003 Food Habits of American Kestrels 239 Table 1 . Seasonal composition of the diet of American Kestrels (Falco sparverius) in semiarid forest in Parque Luro Reserve, central Argentina. N = Number of Prey and % = Percent Frequency. Mass (g) Autumn Winter Spring Summer N % N % N % N % Vertebrates Rodents (subtotal) 32 4.7 75 7.8 67 6.3 6 1.4 Calomys spp. 16 10 1.5 14 1.5 21 2.0 1 0.2 Oligoryzomys flavescens 22 2 0.3 8 0.8 4 0.4 — — Akodon molinae 38 — — 6 0.6 1 0.1 — — Akodon azarae 22 3 0.4 5 0.5 6 0.6 — — Graomys griseoflavus 61 — — — — 1 0.1 — — Reithrodon auritus 74 — — — — — — 1 0.2 Eligmodontia typus 17 — — 1 0.1 1 0.1 — — Ctenomys spp. 80 — — — — 5 0.5 — — Rodents unidentified 17 2.5 41 4.3 28 2.6 4 1.0 Reptiles (subtotal) 1 0.1 — — 11 1.0 — — Polychrotidae 32 — — — — 1 0.1 — — Teidae 29 — — ■ — — 10 0.9 — — Reptiles unidentified 1 0.1 — — — — — — Birds (subtotal) 1 0.1 1 0.1 27 2.5 — — Passeriformes 28 — — 1 0.1 23 2.2 — — Columbina picui 47 — — — — 3 0.3 — — Bird unidentified 1 0.1 — — 1 0.1 — — Invertebrates Orthoptera (subtotal) 516 75.8 438 45.5 617 57.8 338 81.1 Acrididae 2.5 491 72.1 374 38.9 342 32.1 303 72.7 Tettigoniidae 2 14 2.1 8 0.8 25 2.3 11 2.6 Gryllidae 1.2 11 1.6 56 5.8 250 23.4 24 5.8 Mantodea Mantidae 1.4 33 4.8 8 0.8 11 1.0 26 6.2 Hymenoptera Formicidae 0.1 55 8.1 301 31.3 157 14.7 10 2.4 Homoptera Cicadidae 1.3 2 0.3 — — 70 6.6 8 1.9 Coleoptera (subtotal) 37 5.4 133 13.8 89 8.3 27 6.5 Scarabaeidae 1.1 26 3.8 44 4.6 64 6.0 19 4.6 Carabidae 1.3 11 1.6 89 9.3 24 2.2 6 1.4 Curculionidae 1 — — — — — — 1 0.2 Cerambycidae 1 — — — — — — 1 0.2 Tenebrionidae 1 — — — — 1 0.1 — — Arachnids Scorpionida 8 4 0.6 6 0.6 5 0.5 2 0.5 Araneae 1 — — — — 13 1.2 — — Total number of prey items 681 962 1067 417 Food niche breadth 1.79 3.48 5.00 1.82 Standardized food niche breadth 0.07 0.19 0.20 0.07 Mean prey size (g) 3.47 3.71 4.43 2.74 240 Sarasola et al. VoL. 37, No. 3 Figure 1. Relationship between percentage of Orthop- terans versus percentage of Coleopterans in the diet of American Kestrels (each data point represents a season) . The regression line is drawn only to emphasize the trend observed. adults). High-energy demands of the nestlings might also explain why adults ate few birds, but fed a large number of birds to their nestlings. Birds such as House Sparrows {Passer domesticus) have been shown to have higher fat and gross energy content than voles {Microtus pennsylvanicus) , and grasshoppers {Melanoplus femur rubrmn) (Bird et al. 1982). The kestrels we studied appeared to be rather opportunistic in their feeding behavior when we consider the occurrence of tuco tucos in their diet. This fossorial, medium-sized rodent spends most of its life underground living in long and complex burrow systems feeding on roots and grasses (Red- ford and Eisenberg 1992). Kestrels probably prey on the young tuco tucos when they disperse from parental burrows. Quantitative analyses of prey have shown the in- sectivorous nature of kestrels. However, an inverse relationship occurs when invertebrate and verte- brate biomasses in the diet are considered throughout the seasons. Jaksic et al. (1981) and Jaksic and Delibes (1987) have classified kestrels as carnivorous/insectivorous (in that order) and within the insect feeding guild, while Jaksic et al. (1993) classified kestrels as an omnivorous preda- tor on the basis of dietary data collected during the breeding and nonbreeding (wintering) sam- pling periods. In all these studies, kestrels were tro- phically closer to Burrowing Owls {Speotyto cunicu- laria), which Silva et al. (1995) have suggested belong to a carnivorous instead of an insectivorous guild, based on the biomass dominance of verte- brate prey in the breeding and nonbreeding sea- sons (see Bellocq 1988). Vertebrate prey reflected the same prevalence in the kestrel diet in our study. When we combined our data into nonbreed- ing (autumm and winter) and breeding (spring Figure 2. Percentage of biomass by season of the four major groups (Vertebrates, Orthopterans, Coleopterans, and other) of prey in the diet of American Kestrels in Parque Luro Reserve, Argentina. September 2003 Food Habits of American Kestrels 241 Table 2. Composition of the diets of adult vs. nestling American Kestrels during breeding in Parque Luro Reserve Data from breeding season 2000 is pooled by age class {N = Number of Prey). Adults Nestlings N Percent Frequency Percent Biomass N Perceni Frequency Percent Biomass Vertebrates Rodents (subtotal) 56 10.1 62.2 11 2.2 26.5 Calomys spp. 21 3.8 14.1 — — — Oligoryzomys flavescens 4 0.7 3.7 — — — Akodon molinae 1 0.2 1.6 — — — Akodon azarae 6 1.1 5.5 — — — Graomys griseoflavus 1 0.2 2.6 — — — Eligmodontia typus 1 0.2 0.7 — — — Ctenomys spp. — — — 5 1.0 17.1 Unidentified rodents 22 4.0 34.1 6 1.2 9.5 Reptiles (subtotal) 1 0.2 1.2 10 2.0 12.5 Polychrotidae — — — 1 0.2 1.4 Teidae 1 0.2 1.2 9 1.8 11.1 Birds (subtotal) 3 0.5 4.7 24 4.7 30.3 Passeriformes 1 0.2 1.2 22 4.3 26.3 Columbina picui 1 0.2 2.0 2 0.4 4.0 Unidentified birds 1 0.2 1.6 — — — Invertebrates Orthoptera (subtotal) 254 45.7 25.8 363 71.0 25.0 Acrididae 234 42.1 24.5 108 21.1 11.5 Tettigoniidae 11 2.0 0.9 14 2.7 1.2 Gryllidae 9 1.6 0.5 241 47.2 12.3 Mantodea Mantidae 5 0.9 0.3 6 1.2 0.4 Hyraenoptera Formicidae 153 27.5 0.6 4 0.8 — Homoptera Cicadidae 1 0.2 0.1 69 13.5 3.8 Coleoptera (subtotal) 75 13.5 3.6 14 2.7 0.8 Scarabaeidae 61 11.0 2.8 3 0.6 0.1 Carabidae 13 2.3 0.7 11 2.2 0.6 Tenebrionidae 1 0.2 0.0 — — — Arachnids Scorpionida 4 0.7 1.3 1 0.2 0.3 Araneae Total number of prey items Food niche breadth Standardized food niche breadth Mean prey size (g) 4 0.7 556 3.44 0.13 4.30 0.2 9 1.8 511 3.38 0.17 4.59 0.4 242 Sarasola et al. VoL. 37, No. 3 and summer) seasons, the percentage of vertebrate biomass was higher than invertebrates during the nonbreeding season (54%), but also during the breeding season (55%). Therefore, we agree with Jaksic et al. (1981) and Jaksic and Delibes (1987) and consider the kestrel to be a carnivorous/insec- tivorous predator. Finally, demonstration of significant differences between the diets of adult and nestling kestrels in- dicated that food habit studies of this raptor should take into account age-related differences in the diet. If the diet during the breeding season is to be described, samples of pellets and prey re- mains should be collected both at the nest site and from perches used by the adult kestrels. Acknowled gments We thank Ramon A. Sosa, Juan Jose Maceda and Agus- dn Lanusse for field assistance and Laura Bragagnolo for help with the pellet analysis. We are grateful to Sergio Tiranti for assisting us with rodent identification. Javier Seoane, Carlos Rodriguez, Marc Bechard, Karen Wiebe, and two anonymous reviewers provided criticisms and helpful suggestions to an earlier version of this manu- script. The Direccion de Recursos Naturales, Gobierno de La Pampa made this work possible by giving permis- sion to carry out fieldwork in Parque Luro Reserve. This study was supported by a grant (PI 122) from the Depar- tamento de Recursos Naturales, Facultad de Ciencias Ex- actas y Naturales, Universidad Nacional de La Pampa. The senior author had a fellowship for graduate students from the Universidad Nacional de La Pampa. Literature Cited Balgooyen, T.G. 1976. Behavior and ecology of the American Kestrel {Falco sparveriush.) in the Sierra Ne- vada of California. Univ. Calif. Publ. Zook 10.S. Bellocq, M.I. 1988. Dieta de Athene cunicularia (Aves, Strigidae) y sus variaciones estacionales en ecosiste- mas agrarios de la Pampa, Argentina. Physis Secc. C 46: 17-22. Beltzer, A.H. 1990. Biologia alimentaria del halconcito comun Falco sparverius en el valle aluvial del rio Pa- rana Medio, Argentina. Hornero 13:133-136. Bird, D.M., S.K. Ho, and D. Pare. 1982. Nutritive values of three common prey items of the American Kestrel. Comp. Biochem. Physiol. 73:513-515. Brown, L.H. and D. Amadon. 1968. Eagles, hawks and falcons of the world. Vols. I and II. MacGraw-Hill, New York, NYU.S.A. Cabrera, A.L. 1976. Regiones fitogeograficas argentinas. Enciclopedia Argentina de Agricultura y Jardineria. Editorial Acme, Buenos Aires, Argentina. Cei, J.M. 1986. Reptiles del centro, centro-oeste y sur de la Argentina: Herpetofauna de las zonas aridas y se- miaridas. Museo Regionale di Scienze Natural!, Mo- nografie IV, Torino, Italy. COLLOPY, M.W. AND J.R. KoPLiN. 1983. Diet, capture suc- cess, and mode of hunting by female American Kes- trels in winter. Condor 85:369-371. Colweel, R.K. and D.J. Futuyma. 1971. On the measure- ment of niche breadth and overlap. Ecology 52:267- 276. Corn, P.S. 1994. Straight-line drift fences and pitfall traps. Pages 109-117 in W.R. Heyer, M.A. Donnelly, R.W. McDiarmid, L.C. Hayek and M.S. Foster [Eds.], Measuring and Monitoring Biological Diversity. Stan- dard Methods for Amphibians. Smithsonian Institu- tion Press, Washington, DC U.S.A. and London, U.K. Cruz, A. 1976. Food and foraging ecology of the Amer- ican Kestrel in Jamaica. Cowdor 78:409-412. De La Pena, M.R. and M. Rumboll. 1998. Birds of south- ern South America and Antarctica. Harper Collins, London, U.K. Fiora, a. 1933. El peso de las aves. Hornero 5:174-188. Greer, J.K. and D.S. Bullock. 1966. Notes on stomach contents and weights of some Chilean birds of prey Auk 83:308-309. Heintzelman, D.S. 1964. Spring and summer sparrow- hawk food habits. Wilson Bull. 76:323-330. Jaksic, F.M., H.W. Greene, and J.L. Yanez. 1981. The guild structure of a community of predatory verte- brates in central Chile. Oecologia 4:9:21-28. AND M. Delibes. 1987. A comparative analysis of food-niche relationships and trophic guild structure in two assamblages of vertebrate predators differing in species richness: causes, correlations, and conse- quences. Oecologia 71:461-472. , P. Feinsinger, AND J.E. JiMENEZ. 1993. A long- term study on the dynamics of guild structure among predatory vertebrates at a semiarid Neotropical site. Oikos 67:87-96. Jenkins, R.E. 1970. Food habits of wintering sparrow- hawks in Costa Rica. Wilson Bull. 82:97-98. Koplin, J.R., M.W. CoLLOPY, A.R. Bammann, and H. Le- VENSON. 1980. Energetics of two wintering raptors. Auk 97:795-806. Lamore, D.H. 1963. Prey of a sparrowhawk family when raising young. Wilson Bull. 75:461. Marti, C.D. 1988. A long-term study of food-niche dy- namics in the Common Barn-Owl: comparations with- in and between populations. Can. J. Zool. 66:1803- 1812. Mersmann, T.J., D.A. Buehler, J.D. Fraser, and J.K.D. Seegar. 1992. Assessing bias in studies of Bald Eagle food habits. J. Wildl Manage. 56:73-78. Narosky, T. and D. Yzurieea. 1989. Birds of Argentina and Uruguay, a field guide. Asociacion Ornitologica del Plata, Buenos Aires, Argentina. Pearson, O.P. 1995. Annotated keys for identifying small mammals living or near Nahuel Huapi National Park or Lanin National Park, southern Argentina. Masto- zoologia Neotrop. 2:99-148. Redford, K.H. AND J.F. Eisenberg. 1992. Mammals of the September 2003 Food Habits oe American Kestrels 243 neotropics. The southern cone. Vol. 2. Univ. Chicago Press, Chicago, IL U.S.A. Roest, A.I. 1957. Notes on the American Sparrowhawk. Auk 74:1-19. Rudolph, S.G. 1982. Foraging strategies of American Kestrels during breeding. Ecology 63:1268-1276. Salvador, S.A. 1988. Datos de peso de aves Argentinas. Hornero 13:78-83. . 1990. Datos de pesos de aves Argentinas 2. Hor- nero 13:169—171. Silva, S.L, 1. Lazo, E. Silva-aranguiz, P. Meserve, and J.R. Gutierrez. 1995. Numerical and functional re- sponse of Burrowing Owls to long-term mammal fluc- tuations in Chile./. Raptor Res. 29:250-255. SiMONETTi, J.A., H. Nunez, and J.L. Yanez. 1982. Falco sparverius L.: rapaz generalista en Chile Central. Bol. Mus. Nac. Hist. Nat. CAzZe 39:119-124. Smallwood, J.A. 1987. Sexual segregation by habitat m American Kestrels wintering in southcentral Florida, vegetative structure and responses to differential prey availability. Con99.99% in our Utah studies), this owl takes invertebrate prey, there is little doubt that it is highly adapted for an insectivorous life- style and its ecological role as a predator centers on its insectivorous feeding behavior. Resumen. — Mientras estudiamos la ecologia reproductiva de Otus flammeolus en el norte de Utah entre 1994-2001, se encontro la presencia de 13 vertebrados presa. Restos de aves, roedores y murcielagos fueron encontrados en nidos ocupados por el buho, adicionalmente se capto, con una camara de video, la entrega de un mamifero pequeno en el nido. Los vertebrados han sido raramente reportados como presas de estos buhos, los cuales abar- caron 0.01% de la dieta del buho. [Traduccion de Cesar Marquez] Acknowledgments We would like to thank D. Arsenault, J. Belthoff, G Hayward, and D. Holt for reviews that helped to greatly improve this manuscript. Literature Cited Goggans, R. 1986. Habitat use by Flammulated Owls in northeastern Oregon. M.S. thesis, Oregon State Univ , Corvallis, OR U.S.A. Holt, D.W. 1996. On winter records and vertebrate prey in Flammulated Owls. J. Raptor Res. 30:46-48. , J.A. Hoy, and RL. Wright. 1987. Occurrence and first nest record of Flammulated Owls in Mon- tana. /. Raptor Res. 21:120-121. Linkhart, B.D. and R.T. Reynolds. 1994. Peromyscus car- cass in the nest of a Flammulated Owl. J. Raptor Res 28:43-44. Marti, C.D. 1997. Flammulated Owls breeding in decid- uous forests. Pages 262-266 in J.R. Duncan, D.H Johnson, and T.H. Nicholls [Eds.], Biology and con- servation of owls of the Northern Hemisphere: pro- ceedings of the second international owl symposium. USDA Gen. Tech. Rep. NC-190, St. Paul, MN U.S.A. McCallum, D.A. 1994. Flammulated Owl. In A. Poole and F. Gill [Eds.], The birds of North America, No. 93, Academy of Natural Sciences, Philadelphia, PA and American Ornithologists’ Union, Washington, DC U.S.A. . 1996. Diurnal sight records of Flammulated Owls and possible vertebrate prey in winter: the case for caution. J! Raptor Res. 30:49-51. Oleyar, M.D. 2000. Flammulated Owl {Otus flammeolus) breeding ecology in aspen forests of northern Utah: including responses to ski area development. M.S. thesis, Boise State Univ., Boise, ID U.S.A. Reynolds, R.T. and B.D. Linkhart. 1987. The nesting 246 Short Communications VoL. 37, No. 3 biology of Flammulated Owls in Colorado. Pages 239- 248 in R.W. Nero, R.J. Clark, R.J. Knapton, and R.H. Hamre [Eds.], Biology and conservation of northern forest owls. USDA For. Serv. Gen. Tech. Rep. RM-142, Ft. Collins, CO U.S.A. . 1992. Flammulated Owls in ponderosa pine: ev- idence of preference for old growth. Old-growth for- ests in the Southwest and Rocky Mountain Regions: proceedings of a workshop. USDA For. Serv. Gen. Tech. Rep. RM-213, Ft. Collins, CO U.S.A. Verner, J. 1994. Current management situation: Flam- mulated Owls. Pages 10-13 in G.D. Hayward and J. Verner [Eds.], Elammulated, Boreal, and Great Gray Owls in the United States: a technical conservation assessment. USDA For. Serv. Gen. Tech. Rep. RM-253, Ft. Collins, CO U.S.A. Received 27 August 2002; accepted 1 May 2003 Associate Editor: James R. Belthoff J Raptor Res. 37(3);246-252 © 2003 The Raptor Research Foundation, Inc. Nesting Habitat of Cooper’s Hawks in Northern Great Plains Woodlands Melvin P. Nenneman^ and Todd A. Grant U.S. Fish and Wildlife Service, J. Clark Salyer National Wildlife Refuge, 681 Salyer Road, Upham, ND 58789 U.S.A. Marriah L. Sondreal^ and Robert K. Murphy U.S. Fish and Wildlife Service, Des Lacs National Wildlife Refuge Complex, 8315 Highway 8, Kenmare, ND 58746 U.S.A. Key Words: Cooper’s Hawk; Accipiter cooperii; Great Plains; North Dakota; nesting habitat. Cooper’s Hawk {Accipiter cooperii) nesting habitat has been described quantitatively in the eastern (Titus and Mosher 1981, Bosakowski et al. 1992), midwestern (Wig- gers and Kritz 1991, Garner 1999, Trexel et al. 1999), southwestern (Fischer 1986, Kennedy 1988, Boal and Mannan 1998), and western (Reynolds et al. 1982, Moore and Henny 1983, Asay 1987) regions of North America, but not for the northern Great Plains, where the hawk has been considered a species of special concern due mainly to its perceived scarcity (Nenneman et al. 2002) . Woodland habitat within the northern Great Plains was scarce prior to European settlement in the early 1900s, limited primarily to gallery forests along major rivers (Coues 1897, Stewart 1975). Since then, tree coverage has increased. Eor example, woodland area in north- western and north-central North Dakota more than dou- bled during 1938-91 (T. Grant and R. Murphy unpubl. data). Cooper’s Hawk nesting habitat in the region may differ from that in other regions due to isolation of avail- ^ Present address: Valentine National Wildlife Refuge, HC 37 Box 37, Valentine, NE 69201 U.S.A.; e-mail ad- dress: Melvin_Nenneman@fws.gov ^ Present address: N8386 County Highway D, Belleville, WI 53508 U.S.A. able woodland habitats, lower tree species diversity, and differences in structural complexity of woodlands. Re- cent studies have shown Cooper’s Hawks to be markedly plastic in their nesting habits (Boal and Mannan 1998, Bielefeldt et al. 1998, Trexel et al. 1999). However, spe- cies-habitat relationships inferred from data collected in one region may not adequately predict Cooper’s Hawk use of woodland habitat in another region (Kennedy 1988, Trexel et al. 1999). Our objectives were to: (1) describe quantitatively hab- itat characteristics of Cooper’s Hawk nest-sites across the Souris River basin of North Dakota, (2) compare nest site habitat used to woodland habitat available, and (3) com- pare Cooper’s Hawk nesting habitat in North Dakota to that in other regions. Study Area and Methods Our study area was the Souris River basin in north- central North Dakota (48°40'N, 101°25'W; Ward, Ren- ville, and McHenry counties; Fig. 1). The area is char- acterized by level to rolling plains found within the Drift Plain physiographic region (Bluemle 1991). Climate is subhumid continental, with mean monthly temperatures ranging from — 15°C in January to 20°C in July. Mean annual precipitation is about 42 cm, most of which falls as rain between April and September (U.S. Fish Wildl. Serv. unpubl. data). The Souris River flows along a 110-km “loop” from Canada south into north-central North Dakota, then back north again (Bluemle 1991; Fig. 1). In 1994—95, we September 2003 Short Communications 247 I — I National Wildilife Refuge (NWR) intensively Searched Areas 0 20 40 so Kilomctars Figure 1. Cooper’s Hawk study area in the Souris River basin of North Dakota, including two intensive search areas where nesting habitat use and availability were compared: (A) the southern one-half of Des Lacs National Wildlife Refuge (3951 ha) and (B) the far southern one-fifth of J. Clark Salyer National Wildlife Refuge (5263 ha). located Cooper’s Hawk nests in woodlands along the western one-half of this loop, the adjoining Des Lacs Riv- er, and intermittent tributaries (“coulees”) of these riv- ers. Des Lacs and Upper Souris National Wildlife Refuges (NWR) were included, as well as privately-owned wood- lands in Ward and southwestern McHenry counties. Des Lacs NWR has about 740 ha of woodland (10% of the refuge), mostly on north- and east-facing slopes of cou- lees. The Souris River has a narrow (<0.2 km wide) band of woodlands along its meandering bottoms, and wooded coulees in adjacent uplands. In 1996, we located Cooper’s Hawk nests along the eastern one-half of the Souris River loop, focusing our search for nests on the far southern part of J. Clark Salyer NWR (about 4000 ha of woodland, 17% of the refuge; Fig. 1) . Here, the Souris River bisects an upland sandhills complex derived from the shores of glacial Lake Souris (Bluemle 1991). We also located nests in the Souris River floodplain woodlands adjacent to this sandhills complex. Woodlands in our study area represented the naturally- occurring deciduous woodlands available in the northern Great Plains region: quaking aspen (Populus tremuloides)- bur oak (Quercus macrocarpa) woodland in the sandhills, green ash {Fraxintis pennsylvanica)-Amcricaai elm ( Ulmus americana) on north and east slopes of coulees, and green ash-American elm woodland in the river floodplain. The only other woodland habitat available within our study area was planted tree shelterbelts around farmsteads. Woodlands were searched regardless of their perceived suitability for nesting (Trexel et al. 1999). We searched woodland habitat on foot during late April-June and found nests mainly by scanning individual trees (Nen- neman et al. 2002). We completely searched all woodland habitat within two intensive search areas arbitrarily se- lected to determine Cooper’s Hawk nesting density (Des Lacs NWR, 3951 ha and J. Clark Salyer NWR, 5263 ha; Nenneman et al. 2002). We restricted comparisons of nest-site habitat use and availability to these two areas. Additional nests were located by searching parts of other woodlands scattered throughout the study area. In Ward County, this included areas around nest sites originally identified during a 1986-87 breeding bird atlas (G. Ber- key and R. Martin unpubl. data) . We quantified woodland habitat with techniques de- scribed by James and Shugart (1970) and Noon (1981). A 0.04-ha circular plot, centered on the nest tree, was the basis for nest habitat measurements (James and Shugart 1970, Trexel et al. 1999; Table 1). We measured habitat only once at each nest area (he., the area within 0.4 km of the first nest found [Rosenfield et al. 1995]), to avoid interdepen- dency among sampling points (Bosakowski et al. 1992). To assess nest habitat use by Cooper’s Hawks, we com- pared habitat characteristics at nest sites with those at ran- dom woodland plots. We confined this analysis to nest sites located within our two intensive search areas. Woodlands at Des Lacs NWR were roughly linear (mostly <0.1 km wide), so we chose three to six plots centered at 200-m intervals from a random starting point (firom 0-40 m from 248 Short Communications VoT. 37, No. 3 Table 1. Mean habitat characteristics of 48 Cooper’s Hawk nest sites across north-central North Dakota, 1994-96. Characteristic Description X (SE) Nest tree Nest height Nest tree height % nest height DBH Nest canopy cover Nest site Height (m) from ground to top of nest (meter tape) Height (m) of nest tree canopy (Abney level or rangefinder) (Nest height/tree height) X 100 Diameter of nest tree (cm) at breast height % canopy cover within a 1-m horizontal radius of the nesU Canopy height Stand DBH Tree density Basal area Shrub density % canopy cover % ground cover Slope Distance to water Distance to edge Stand age Mean height (m) of canopy of three trees in study plot Mean DBH (cm) of canopy trees in the study plot Number of canopy trees >7.5 cm DBH per ha. Cross-sectional area (m^/ha) of canopy trees Number of shrub stems (<7.5 cm DBH and >1.5 m tall) /ha % of area over plot occluded by canopy tree foliage*’ % of ground in plot covered by ground layer foliage’’ Slope (degrees) of plot (Abney level) Distance (m) to nearest permanent water source (pacing or measured from aerial photographs) Distance (m) to nearest opening in woodland canopy (pac- ing or measured from aerial photographs) Determined from core samples on the two largest trees^ Habitat (%) within 1 km of nest site Woodland Cropland Grassland Wetland Open Water Urban/other Percent of landscape covered by trees Percent of landscape covered by annually-tilled cropland Percent of landscape covered by native mixed-grass prairie (grazed or idle) , planted grassland, or perennial forage crop Percent of landscape covered by riparian wet meadow or pot- holes Percent of landscape covered by rivers, lakes, or impound- ments Percent of landscape covered by cities, roads, etc. 9.2 (0.3) 13.2 (0.4) 69.7 (1.6) 24.7 (1.0) 77.4 (2.9) 13.3 (0.4) 16.0 (0.3) 1156.3 (64.0) 25.9 (1.5) 1364.1 (152.8) 63.9 (2.2) 72.1 (2.7) 3.0 (0.8) 494.5 (72.0) 23.9 (2.8) 63.5 (5.1) 24.6 (2.2) 13.5 (2.5) 49.9 (2.4) 4.5 (1.3) 5.7 (1.5) 2.0 (0.9) Ocular estimate. ’’ Forty ocular tube readings (James and Shugart 1970). Measured at 14 nest sites. the woodland edge), on transects that bisected (linearly) the woodlands. We chose the nearest tree on a random compass bearing from the plot center to compare to nest trees. In sandhill woodlands at J. Clark Salyer NWR, we paced a random distance (up to 40 m) and direction in the field from random x-y coordinates identified on aerial photographs (1992, 1:7920), then selected the nearest tree along a random compass bearing as the center of the ran- dom plot (modified from Titus and Mosher [1981]). We determined percentage area of woodland, grassland, crop- land, wetland, open water on rivers, and urban/ other land use within 1 km (314 ha) of each nest and random plot using aerial photographs (1:21120, 1:7920) and an elec- tronic planimeter or a dot grid. We used standard statistical software (Hintze 1995) to conduct our data analyses. We grouped variables into subsets that characterized similar features, and conduct- ed a multivariate analysis of variance (MAN OVA) on each of these subsets. Variables characterizing the nest or cen- ter tree (tree height and DBH) comprised one subset, and variables characterizing the nest or plot stand (can- opy height, stand DBH, tree density, and percent canopy cover) comprised the second subset (Tables 2 and 3) Using these subsets and the remaining variables, we made seven (six at J. Clark Salyer NWR) simultaneous comparisons between nest sites and random plots for each intensive search area. To maintain an experiment- wise alpha of 0.10, we applied a Bonferroni correction (Sokal and Rohlf 1995); thus an alpha of 0.0143 was used for these simultaneous comparisons. When differences were detected, we compared individual variables with 2- tailed Mests, or if data were non-normal, we used non- parametric rank tests (Mann-Whitney U-test) or unequal variance tests (Aspin-Welch [Hintze 1995]). Variables that did not logically fit into subsets (e.g., slope) were compared using 2-tailed t-tests. We used Chi-square good- ness-ol-fit tests to compare species composition of nest September 2003 Short Communications 249 Table 2. Mean (SE) values of habitat characteristics at Cooper’s Hawk nest sites {N = 11) and random plots {N = 32) at Des Lacs National Wildlife Refuge, North Dakota, 1994-95. Habitat Characteristic Nest Sites Random Plots pa pb Nest tree or center tree Height (ra) 12.4 (0.8) 8.6 (0.6) 0.002 <0.001* DBH (cm) 22.1 (1.6) 17.3 (1.7) 0.004 Nest site or random plot Canopy height 12.4 (0.6) 11.4 (0.6) 0.247 Stand DBH 15.0 (0.4) 15.4 (0.6) 0.090 0.821 Tree density (no. /ha) 1054.5 (99.9) 755.5 (49.8) 0.017 % canopy cover 64.5 (4.1) 58.3 (3.1) 0.233 Shrub density (stems/ha) 1120.5 (224.9) 2010.2 (226.1) 0.009 Slope (degrees) 7.0 (2.3) 12.6 (0.9) 0.043 Distance to water (m) 422.9 (167.9) 302.4 (73.8) 0.358 Distance to edge (m) 22.2 (4.7) 22.0 (3.1) 0.979 Stand age (years) 59.9 (4.7) 72.6 (3.2) 0.041 ^P-value is for group MAN OVA or Rest (rows that share a common P-value comprise a MANOVA group); experiment-wise alpha = 0 10 . P-value is for Aspin-Welch Unequal-Variance Rest, except * denotes Mann-Whitney Cftest. trees used to that available based on center trees of 0.04- ha random plots. Results We measured 48 Cooper’s Hawk nests found during 1994—96. Cooper’s Hawks tended to nest in woodland sites with dense canopies and high tree densities (Table 1). Can- opy cover directly over individual nests was also high, with the exception of a nest in an aspen snag. Although shrubs were common at most nest sites (Table 1), shrub stem den- sities were highly variable, ranging from almost none in grazed woodlands to 4125 shrub stems/ha at one sandhill woodland site. Most nests (78%) were on slopes <5°, even though steeper wooded slopes (5-25°) were typical of cou- lee and some sandhill woodlands. Cooper’s Hawks nested consistently close to woodland edge (Table 1). Distance to water, however, was more variable, ranging from 34—2000 m. Mean stand age (determined at three floodplain and 11 coulee nest sites) was 64 yr (Table 1 ) , but nests occurred m stands as young as 35 yr. Cooper’s Hawks typically placed their nests within the lower part of the tree canopy, at slightly more than two- Table 3. Mean (SE) values of habitat characteristics at Cooper’s Hawk nest sites (N =18) and random plots {N = 40) in sandhill woodlands at J. Clark Salyer National Wildlife Refuge, North Dakota, 1996. Habitat Characteristic Nest Sites Random Plots pa pb Nest tree or center tree Height (m) 12.7 (0.4) 9.8 (0.4) <0.001 <0.001 DBH (cm) 23.8 (1.4) 17.5 (1.1) <0.001 Nest site or random plot Canopy height 12.7 (0.4) 10.7 (0.4) <0.001 Stand DBH 15.9 (0.6) 15.5 (0.7) 0.012 0.387* Tree density (no. /ha) 1277.8 (112.0) 1042.5 (60.2) 0.076 % canopy cover 54.2 (3.0) 50.6 (2.2) 0.351 Shrub density (stems/ha) 2265.3 (230.0) 2510.3 (183.9) 0.410 Slope (degrees) 1.0 (0.7) 5.1 (0.9) <0.001* Distance to water (m) 677.7 (89.0) 665.3 (61.3) 0.909 Distance to edge (m) 25.2 (6.3) 16.3 (2.2) 0.073 P-value is for group MANOVA or Rest (rows that share a common P-value comprise a MANOVA group); experiment-wise alpha = 0 10. P-value is for Aspin-Welch Unequal-Variance Rest, except * denotes Mann-Whitney U-test. 250 Short Communications VoL. 37, No. 3 thirds of tree height (Table 1). Heights of nest trees and surrounding trees were similar (^4 = 0.23, P = 0.822), but nest tree DBH was larger than stand DBH (^4 = 9.78, P < 0.001). Nests were primarily in green ash (46%) or quaking aspen (31%); one (2%) was in an aspen snag, and remaining nests (21%) were in five other tree species. Grassland, woodland, and cropland comprised most land cover within 1 km of Cooper’s Hawk nests (Table 1) . Wetlands, open water, and urban lands generally com- prised a small portion of land cover around nests (means <6%), but were major components of land cover (31- 44%) surrounding a few nests. At Des Lacs NWR, nest trees were taller and larger than those available (Table 2). Nest sites had lower shrub den- sity than random plots. Cooper’s Hawks also appeared to select sites with less slope than that available. Two-thirds of the nests at Des Lacs NWR were on slopes <10°, com- pared to only one-third of the random plots. Nest tree species were used in proportion to their availability (x^.s = 1.66, P > 0.50). Other variables were similar between nest sites and random plots at Des Lacs NWR. At J. Clark Salyer NWR, nest trees again were taller and larger than expected based on random sampling (Table 3) . Canopy height was greater, and slope was less at nest sites than at random plots. Tree species were used for nesting in proportion to their availability (x^3 = 1.94, P > 0.50). Discussion Cooper’s Hawks breeding in North Dakota selected nest sites with less canopy cover and with shorter nest trees of lesser DBH compared to most other regions (Ta- ble 4). These differences may be due, in part, to struc- turally-inherent contrasts between woodlands in North Dakota when compared to other regions. Most wood- lands in the northern Great Plains are relatively young (<100 yr), and the cool, subhumid continental climate (Stewart 1975:3) may slow tree growth. Although nest trees in this study were smaller than in other regions, Cooper’s Hawks still placed their nests at about two-thirds of the nest tree height. Cooper’s Hawks nested in areas with a surprisingly small amount of woodland cover (as little as 1.0% within 1 km of nest), although most nested in areas with ^10% woodland cover. Cooper’s Hawks have also nested in an area in northwestern North Dakota where woodland may comprise only 2% of the landscape (Peterson and Mur- phy 1992). We observed two parallel patterns of Cooper’s Hawk nest-site use on our two intensive search areas; typically, .sites with the greatest tree density and with little or no slope were used. Other authors (Reynolds et al. 1982, Bo- sakowski et al. 1992) have noted that most Cooper’s Hawks nest in relatively flat terrain. Steeper slopes in our study area tend to have shorter trees due to drier and more poorly-developed soils. Trees on slopes also have relatively more exposed canopies than those on level ground. U O OJ GT) u u u X u V I S-i OJ (X o o u V •kJ u o Ph ai 3 w w u Pi L O C/!5 PM E O H H p: ^ sag 52 g? O) CO nj CO 00 CT> ^ CM p ^ p ^ in ca 00 oj -o ^ S n ■M i O rd n .22 g p H S pq H T3 P OJ u 00 00 CTj 00 P cd P P O CO n 00 OC <05 01 OJ X OJ u Ml o o 'o V £ c p cd "cd o PQ M .IP O oOGO^«ooiTMk,t^mcNaii>GH T—i CN IO 00 t-H rH rH H 01 1-H 0 cd 0 01 00 1 1 1 1 1 1 1-H H Q CM CM CM lO 1 1 1 1 1 1 m c/l W t-H Q H Z CM Pi u 0 oi 1 01 0 CD 1 0 0 1 00 i-H CO u 5 GO 00 d 1 id d 1 d CO I-H Pi u CO 00 j> 00 i> CO CO 00 J> w Pm u U -< O] 22 1 0 ^ CO cq Oi GO q q 1 % s id 1 d kH cri d 1-H G id 1 I CQ "s CM GO CM GO GO GO 1-H 1 — H > 00 GO i-n 0 CM 00 10 CM CM 04 1-H CM lO GO kk ^H kH 10 1-H 1-H 1> 0 0 10 CD 1> CM CO kH 0 CO CN pq fa f 1 d G GO GO CD G d J> 1-H Q CM CM GO GO CO lO 1-H CO H ri Z w H C/} H E 1> n GO iq 00 1 1 q 1 1 U W 0 oi id G G d 1 1 1 1 d 1 1 Pi u u dO GO CO CO CO CO GM irj O CO oi id 01 CO to r-H ^ c4 c4 N CM (M CM CM CM t kM kH CM kH CM kH 01 oieocdiricdcMiiO'dtdidGo cd cd O cd Q kP kP M o !U CTl V Z Ji M T3 P cd U p o he o M O w p o he OJ M o o u O oj he QJ j-O O g c/D C/D a a Percent nest height = (nest heigh t/nest tree height) X 100 Data unavailable September 2003 Short Communications 251 Distance to woodland edge was similar among nests and between nest sites and random plots on both inten- sive search areas, but was far less than reported elsewhere in North America (Table 4). Cooper’s Hawk habitat in our study was relatively edge-dominated, with floodplain and coulee woodlands consisting of narrow, linear belts of timber and sandhill woodlands composed of timber interspersed within a grassland matrix. Despite these con- straints of woodland configurations, North Dakota Coo- per’s Hawks placed nests near the center of woodlands, relatively “far” from edges. Distances from nests to water were extremely variable. Reynolds et al. (1982) and Boal and Mannan (1998) suggested that a nearby water source is important to nesting Cooper’s Hawks in arid western and southwestern North America. However, our data and results from other studies (Titus and Mosher 1981, Bo- sakowski et al. 1992, Trexel et al. 1999) indicate that nest- ing near a permanent water source is not required in more temperate areas. Our data, combined with those from other studies, in- dicate that nesting Cooper’s Hawks exploit a wide array of woodland habitats across their range, including sub- urban and urban areas (Rosenfield et al. 1995, Boal and Mannan 1998), pine plantations (Wiggers and Kritz 1991, Rosenfield et al. 2000), and deciduous and coniferous forest (Titus and Mosher 1981, Reynolds et al. 1982, Bo- sakowski et al. 1992, Trexel et al. 1999). In our study area, Cooper’s Hawks used natural woodlands, planted shel- terbelts (R. Murphy unpubl. data), and even tall shrubs (Sondreal and Murphy 1998). In the northern Great Plains, such flexibility apparently has allowed Cooper’s Hawks to exploit nesting habitat only recently available, and to become perhaps the most abundant breeding rap- tor in the region’s sparsely distributed woodlands (Nen- neman et al. 2002). Resumen. — El gavilan de Cooper {Acdpiter cooperit) ha sido poco estudiado como una especie que se reproduce en las Grandes Planicies, debido a esto, los responsables del ma- nejo de los recursos en la region deben referirse a descrip- ciones cuantitativas de otras partes de Norteamerica para identificar los habitats probables de anidacion. Durante 1994—96 se midieron las caracteristicas del habitat en 48 sitios de anidacion del gavilan de Cooper, en una variedad de tipos de bosque en el centro-norte de Dakota del Norte. Los nidos se construyeron principalmente en arboles de Fra- xinus pennsylvanica, 46% y de Populus tremulmdes, 31%, en sitios de alta densidad de los mismos (0 = 1156 arboles/ ah) y con una pendiente pequefia (0 = 3.0°) . El gavilan de Cooper utilizo los sitios con menos cobertura de dosel (0 = 64%) y que estuvieran mas cerca a los bordes de bosque (0 = 24 m) que lo previamente reportado. Comparado con los habitats disponibles del bosque, el gavilan de Cooper anido en bosques con menos pendiente y ubico los nidos entre los arboles mas grandes y altos. Los gavilanes de Coo- per en periodo de anidacion utilizaron los bosques con los atributos estructurales que fueran relativos a otras regiones. Al igual que con los hallazgos recientes en otros habitats, los datos apoyan la idea que el gavilan de Cooper explota una amplia serie de habitats de bosque para su anidacion. [Traduccion de Cesar Marquez] ACKNO WEED GMENTS We thank those who assisted in nest searching and data collection, especially M. Seykora, D. Huesmann, G. Ber- key, A. Buckmeier, S. Heitzner, and Y Heitzner. G. Berkey and R. Martin kindly provided many nest area records for Ward County. R. Rosenfield advised on methods. J Bielefeldt, E. Madden, N. Niemuth, and M. Restani gra- ciously reviewed early drafts of the manuscript, and com- ments from C. Boal, H. Garner, S. DeStefano, and an anonymous reviewer improved the final draft. G. Sar- geant (USGS-Northern Prairie Wildlife Research Center) provided statistical advice. M. Sondreal and M. Nenne- man were undergraduate interns under auspices of the College of Natural Resources, University of Wisconsin- Stevens Point during the 1994 and 1995 field seasons, respectively. Financial support was provided by the U S Fish and Wildlife Services Refuges Division and Chal- lenge Cost Share Program, North Dakota Game and Fish Department’s nongame program, and the North Dakota Falconers’ Association. Literature Cited Asay, C.E. 1987. Habitat and productivity of Cooper’s Hawks nesting in California. Calif. Fish Game 73:80-87. Bielefeldt, J., R.N. Ro.senfield, W.E. Stout, and S.M. Vos. 1998. The Cooper’s Hawk in Wisconsin: a review of its breeding biology and status. Passenger Pigeon 60: 111 - 121 . Bluemle, J.P. 1991. The face of North Dakota. North Da- kota Geol. Surv. Bismarck, ND U.S.A. Boal, C.W. and R.W. Mannan. 1998. Nest-site selection by Cooper’s Hawks in an urban environment./. Wildl Manage. 62:864—871. Bosakowski, T, D.G. Smith, and R. Speiser. 1992. Nest sites and habitat selected by Cooper’s Hawks, Acdpiter cooperii, in northern New Jersey and southeastern New York. Can. Field-Nat. 106:474-479. CouES, E. (Ed.). 1897. New light on the early history of the great Northwest: the manuscript journals of Al- exander Henry and of David Thompson 1799-1814. Vol. 1. The Red River of the North. F.P. Harper, New York, NY U.S.A. Fischer, D.L. 1986. Daily activity patterns and habitat use of coexisting Accipiter hawks in Utah. Ph.D. disserta- tion, Brigham Young Univ., Provo, UT U.S.A. Garner, H.D. 1999. Distribution and habitat use of Sharp-shinned and Cooper’s hawks in Arkansas. J. Raptor Res. 33:329-332. Hintze, J.L. 1995. Number Cruncher Statistical Systems (NCSS 6.0) . Kaysville, UT U.S.A. James, F.C. and H.H. Shugart, Jr. 1970. A quantitative method of habitat description. Audubon Field Notes 24' 727-735. Kennedy, P.L. 1988. Habitat characteristics of Cooper’s 252 Short Communications VoL. 37, No. 3 Hawks and Northern Goshawks nesting in New Mex- ico. Pages 218-227 in R.L. Glinski, B.G. Pendleton, M.B. Moss, M.N. LeFranc, Jr., B.A. Millsap, and S.W. Hoffman [Eds.], Proc. Southwest Raptor Manage. Symp. and Workshop, Instit. for Wildl. Res., Natl. Wildl. Fed., Natl. Wildl. Fed. Sci. and Tech. Ser. No. 11. Washington, DC U.S.A. Moore K.R. and C.J. Henny. 1983. Nest site characteris- tics of three coexisting Accipiter hawks in northeast- ern Oregon. Raptor Res. 17:65-76. Murphy, R.K. 1993. History, nesting biology, and preda- tion ecology of raptors in the Missouri Coteau of northwestern North Dakota. Ph.D. dissertation, Mon- tana State Univ., Bozeman, MT U.S.A. Nenneman, M.P., R.K. Murphy, and T.A. Grant. 2002. Cooper’s Hawks, Accipiter cooperii, nest successfully at high densities in deciduous woodlands of the north- ern Great Plains. Can. Field-Nat. 116:573-577. Noon, B.R. 1981. Techniques for sampling avian habitats. Pages 42-52 in D.E. Capen [Ed.], The use of multi- variate statistics in studies of wildlife habitat, U.S. For. Ser. Gen. Tech. Rep. RM-87. Rocky Mountain Forest and Range Exp. Station. Fort Collins, CO U.S.A. Peterson, D.J. AND R.K. Murphy. 1992. Prey delivered to two Cooper’s Hawk {Accipiter cooperii) nests in north- ern mixed grass prairie. Can. Field-Nat. 106:385-386. Reynolds, R.T., E.C. Meslow, and H.M. Wight. 1982. Nesting habitat of coexisting Accipiter in Oregon. J. Wildl. Manage. 46:124-138. Roseneield, R.N., J. Bielefeldt, J.L. Affeldt, and D.J. Beckmann. 1995. Nesting density, nest area reoccu- pancy, and monitoring implications for Cooper’s Hawks in Wisconsin./. Raptor Res. 29:1-4. , , S.A. Sonsthagen, and T.L. Booms. 2000. Comparable reproductive success at conifer planta- tion and non-plantation sites for Cooper’s Hawks in Wisconsin. Wilson Bull. 112:417-421. Sokal, R.R. and FJ. Rohlf. 1995. Biometry: the princi- ples and practice of statistics in biological research W.H. Freeman and Co., New York, NYU.S.A. SoNDREAL, M.L. and R.K. MuRPHY. 1998. Low Cooper’s Hawk, Accipter cooperii, nest in dense shrubs. Can. Field- Nat. 112:339-340. Stewart, R.E. 1975. Breeding birds of North Dakota. Tri- College Center for Environmental Studies. Fargo, ND U.S.A. Titus, K. andJ.A. Mosher. 1981. Nest-site habitat select- ed by woodland hawks in the central Appalachians. Auk 98:270-281. Trexel, D.R., R.N. Rosenfiei.d, J. Bielefeldt, and E.A Jacobs. 1999. Comparative nest site habitats in Sharp- shinned and Cooper’s hawks in Wisconsin. Wilson Bull 111:7-14. WiGGERS, E.P. and KJ. Kritz. 1991. Comparison of nest- ing habitat of coexisting Sharp-shinned and Cooper’s hawks in Missouri. Wilson Bull. 103:568-577. Received 1 February 2002; accepted 5 June 2003 Associate Editor: Clint W. Boal J RaptcrrRes. 37(3):252-256 © 2003 The Raptor Research Foundation, Inc. Dynamics and Temporal Variation in Age Structure at a Communal Roost of Egyptian Vultures {Neophron percnopterus) in Northeastern Spain Antoni Margalida^ and Jennifer Boudet GEPT. Ap. 43. E-25520 El Pont de Suert (Eleida), Spain Key Words: Egyptian Vulture, Neophron percnopterus; Age structure, Communal roost, Spain. The Egyptian Vulture {Neophron percnopterus) is a small vulture that occasionally gathers in communal roosts (Ce- ballos and Donazar 1990, Donazar et al. 1996). The ex- istence of such roosts seems to be linked to abundant and regular food sources and also to the species’ gregar- ious behavior under such conditions (Donazar 1993, ^ E-mail address: margalida@gauss.entorno.es Donazar et al. 1996). In small cathartid vultures, Raben- old (1983) has suggested that this behavior would be an adaptation in order to increase an individual’s foraging efficiency on dispersed and unpredictable food supplies. The Spanish population of Egyptian Vultures repre- sents 80% of the total in the European Union, and has undergone a significant decline in recent years (Del Mo- ral and Marti 2002). Communal roosts are thus impor- tant in terms of conservation, particularly bearing in mind that a large fraction of the preadult population concentrates in them (Donazar et al. 1996). The under- lying causes leading to the formation and disappearance September 2003 Short Communications 253 Table 1. Monthly variations in the number {x ± SD) of Egyptian Vultures recorded at the roost between March and September 1998. Month Total Individuals (Range) Adults (Range) Immature (Range) March {N = 4) 0 0 0 April (N = 4) 1 ± 1.7 (0-4) 0.75 ± 1.3 (0-3) 0.25 ± 0.4 (0-1) May {N = 6) 8 ± 5.6 (0-16) 2 ± 1.5 (0-5) 6 ± 4.3 (0-11) June (N = 3) 17.3 ± 2.4 (14-19) 11.7 ± 2.1 (9-14) 5.7 ± 0.9 (5-7) July {N= 5) 5.4 ± 3.5 (0-10) 3 ± 2 (0-6) 2.4 ± 1.6 (0-4) August (N = 4) 13.2 ± 5.5 (6-21) 7 ± 3.7 (2-11) 6.3 ± 2.4 (4-10) September (N = 2) 2.5 ± 2.5 (0-5) 1 ± 1 (0-2) 1.5 ± 1.5 (0-3) of such roosts, the possibility that local variations exist, and the movement of birds between roosts (Donazar et al. 1996), are reasons that justify the need for further information on the structure and dynamics of such ag- gregations in this species. In this paper we describe the results obtained during a study on the dynamics and temporal variation in age structure at an Egyptian Vulture roost in Catalonia (north-eastern Spain) carried out for two consecutive sea- sons. We also discuss the factors leading to its formation and disappearance. Study Area and Methods The study was carried out in Catalonia, northeastern Spain. The breeding population of Egyptian Vulture in this region typically includes 34-40 pairs and its status has remained stable since the middle 1980s (D. Garcia and A. Margalida unpubl. data) . Even though the terri- tories of breeding pairs have been monitored regularly, the first communal roost of this species in the area was not found until 1997 (Margalida 1997). Seven Egyptian Vulture pairs have their nests within a 10-km radius of the roost. Food availability is high and is a consequence of extensive livestock rearing of sheep, goats, and cattle (Margalida et al. 1997). The area where the roost is found has a mean annual rainfall of around 800 mm, with mean temperatures of 23°C (mean high) and 6°C (mean low). Dominant vegetation is of Mediterranean type with scattered forests consisting of white oak ( Quer- cus humilis) , holm oak ( Quercus ilex) , and Scotch pine {Pi- nus sylvestris) . In 1997, when the roost was formed, it was monitored weekly (Margalida 1997) and this was continued during 1998—99. Monitoring consisted of one visit per week for the period between March and September, with a total of 50 visits (28 in 1998 and 22 in 1999). June 1998, and April and August 1999 received only three visits each, which were spaced out every 10 d, so as to cover the whole month in a homogeneous way. In September, at the time of the post-breeding migration, observations were carried out during the first 10 days. In 2000, visits to the roost were carried out more or less sporadically until its disappearance was confirmed at the end of June. Visits usually started in the early evening (1800 H, solar time), as birds began to arrive and settle in, and were carried out until dusk (1900-2000 H), when birds stopped flying in the proximity of the roost. The number and age of the birds present was recorded in every visit Age was established according to plumage following Ce- ballos and Donazar (1990), Telia (1991), and Donazar et al. (1996). Only two age categories were identified, be- cause no young-of-the-year were found. These were* adults (individuals showing complete adult plumage, >4- yr birds) and immature (individuals showing some im- perfection or with dark transitional plumage, 2-yr and 3- yr birds; Donazar 1993). Observations were carried out using 20-60 X telescopes at a distance of over 500 m in order to minimize disturbance to the birds. Results In the study area, arrival of the first breeding individ- uals took place on 5 March 1998 and on 26 February 1999. In 2000, the first individual was observed on 3 March. In 1998, the first birds (three adults and one im- mature) started using the roost on 28 April (previous observation with no birds seen was on 19 April), and m 1999 (11 adults plus one immature) on 26 April (during observations on 19 April no birds were seen). The last birds were seen at the roost on 1 September (two adults and three immature) in 1998 and on 11 August in 1999 (one adult and one immature) . Location of the sleeping vultures changed between the years. While in 1997 they chose to perch on Scotch pines, a birch {Betula pendula), and a white oak (Margalida 1997), in 1998 and 1999 they chose a small forest of oak and, later, an isolated black poplar (Populus nigra). The latter was used in all the observations between May and September. In 2000, all the observations were made on the same poplar that had been used in 1999. Distance between the roost and the nearest predictable food source (a rubbish dump where carcasses were available regularly) was of some 2 km. Altogether, no differences were noticed in the number of birds using the roost in the two study years. In 1998, the maximum number of birds sleeping simultaneously at the roost was 21 on 7 August (Table 1). In 1999, the maximum number of birds was also 21, on 5 June (Table 2). However, both in 1998 and in 1999, the mean largest 254 Short Communications VoL. 37, No. 3 Table 2. Monthly variations in the number {x ± SD) of Egyptian Vultures recorded at the roost between March and September 1999. Month Total Individuals (Range) Adults (Range) Immature (Range) March {N = 2) 0 0 0 April (N = 3) 9.3 ± 6.8 (0-16) 5.5 ± 6.0 (0-11) 0.5 ± 0.8 (0-5) May {N - 5) 7.4 ± 7.3 (0-21) 4.2 ± 3.7 (0-11) 3.2 ± 3.6 (0-10) June {N = 4) 17.3 ± 2.6 (14-21) 11.3 ± 1.1 (10-13) 6 ± 3 (3-11) July {N = 4) 11.5 ± 5.7 (2-17) 4.8 ± 2.3 (1-7) 6.8 ± 3.5 (1-10) August {N = 3) 0.7 ± 0.9 (0-2) 0.3 ± 0.5 (0-1) 0.3 ± 0.5 (0-1) September {N = 1) 0 0 0 number of birds using the roost occurred in June, with 17.3 ± 2.4 birds {N = 3) and 17.3 ± 2.6 birds {N= 4), respectively (Tables 1 and 2). In terms of age classes, the largest number of adults were seen in June of both years. In contrast, birds in non-breeding plumage mostly con- centrated in roosts in August in 1998 and in July in 1999 (Tables 1 and 2) . With respect to birds of breeding age, differences between the years were found in May, with higher numbers in 1999, and in August, with less birds in 1999 (Tables 1 and 2). Differences in the number of non-breeding individuals were found in August, with few- er birds in 1999 (Tables 1 and 2). Ninety-five (50.3%) of the 189 birds seen in 1998 (N = 28 visits) were adult and 94 (49.7%) were immature (Table 1). In 1999, 108 (59.3%) of the 182 birds seen {N = 22 visits) were adult and 74 (40.7%) were immature (Table 2). No significant differences were found between the years with respect to the number of adult birds pres- ent vs. immature (x^i = 2.73, P = 0.099). Monthly vari- ation (using only the highest single count made in each month) in both years neither was statistically significant (1998: = 5.93, P = 0,31; 1999: x\ = 2.56, P = 0.63). In both 1998 and 1999 there was a very similar pattern m the percentages of adults and immatures present at the roost (Fig. 1) and significant differences were only Figure 1. Monthly variation in the proportion of birds of breeding age (white bars) vs. birds of non-breeding age (black bars) during the period April-September of 1998 and 1999. found in May, when the reverse was found (25% adults vs. 75% immature in 1998 and 56.7% adult vs. 43.4% immature in 1999; x^i = 7.58, P< 0.001). In 2000 there were no systematic observations, but the presence of birds in the area of the roost was positively confirmed. The last observation of birds using the roost was on 7 June 2000, when 1 1 individuals were present. Later checks in July and August confirmed the absence of the species in the roost. Discussion The first Egyptian Vultures were present at the roost approximately one and a half months after they arrived at the breeding grounds (beginning of March) . This co- incides with the species’ incubation period, because lay- ing usually takes place during the first half of April (Don- azar 1993, pers. observ.). The birds seen in the first period were mostly adults; immatures joined the roost progressively at a later stage. The fact that the largest number of adults present were found in June (during the chick-rearing period) might correspond to their need to optimize foraging efficiency on food sources that are predictable in both space and time (Ceballos and Donazar 1988), On the other hand, adult birds that had failed in their breeding attempts and that had subse- quently joined the roost might also cause this increase in numbers. The larger proportion of immature birds from May onward might be due to the late arrival of those non- breeding migrants, as has been suggested by Ceballos and Donazar (1990). In this species, roosts may play a strong socialization role, as has been suggested in Black Vultures {Coragyps atratus; Parker et al. 1995). Associations are favored be- cause they facilitate advantages in food finding and ad- vantages in accruing coalition members in contests over food (Rabenold 1987, Parker et al. 1995). Social organi- zation could facilitate the evolutionary stability among communally-roosting birds, but more data on this topic are needed. Contrary to the dynamics observed in other Spanish roosts, where there is a progressive increase in birds from September 2003 Short Communications 255 June-August (Ceballos and Donazar 1990, Donazar et al. 1996), in the roost that we studied, the largest number of birds was found in June of both years. The decrease in the number of birds present in July and August, much higher in 1999, might be linked to a progressive decrease in food availability in the area and by the increase in the availability in food resources in neighboring zones (e.g., Ebro valley) . Low food availablity would have caused the displacement of birds to richer areas. In fact, the small numbers of birds observed in August-September 1999 was due to the fact that the food source they had been exploiting was progressively reduced (between March and May rubbish was dumped at the landfill site once a week, while between June and September no rubbish was dumped). Also, no food was available at all in July 2000 (pers. observ.). This lack of food, which would partly ex- plain the observed differences in temporal pattern with respect to other roosts, may also explain the absence of observations of young of the year. These juveniles, having left their nests between the end of July and the beginning of August, did not visit the roost. Our results also differ from those obtained in island population; where, due to the species’ sedentary habits, the largest concentrations take place in the period between October and February (de Pablo 2000, Donazar et al. 2002) . There seems to be a link between roosts that would be due to a socializing component in this behaviour rather than to food availability (Donazar et al. 1996). Even though use of roosts is dynamic and birds do visit other roosts in the same year, we believe that the disappearance of the roost was caused by the progressive decrease and total disappearance of the available food source. The for- mation of this roost coincided with the appearance of a food source that was available regularly in space and time (Margalida 1997), a factor that has also been recorded in other Spanish roosts (Ceballos and Donazar 1990, Donazar et al. 1996). The roost under study disappeared when no more food was available at the rubbish dump the birds had been exploiting. The roost was located on trees that were close to the food source and in an area with little human disturbance. These factors might have determined the selection of the specific roost site (Ce- ballos and Donazar 1990). However, contrary to other roosts located on isolated trees (Ceballos and Donazar 1990, Donazar et al. 1996) this selection did not occur until one year after the roost was formed. Our observa- tions suggest that the formation and exact location of the roost might undergo a spatial selection process that would end in the choice of the most appropriate loca- tion. Conservation Problems. The Egyptian Vulture roost we studied was formed in 1997 and disappeared in June 2000 just as the food source ceased to exist. This factor accounts for the disappearance of other roosts in Spain (Telia 1991), which shows the species’ strong dependen- cy on these roost sites and the importance that maintain- ing them may have in terms of conservation. The species has undergone a 30% decline in the last 20 years in Spain (J. Donazar pers. comm.), so urgent measures are needed to reverse this trend. Egyptian Vul- ture survival might be closely linked to the maintenance and protection of roost sites, which are typically in quiet areas and close to predictable food sources (Ceballos and Donazar 1990, Telia 1991, Margalida 1997). A large part of the population decline might be due to the appear- ance of a nonnatural mortality factor such as poison. Pro- tection of these areas is thus key to the conservation of the species. Also an important part of the non-breeding population concentrates at these roosts and exploits the available food. It is essential to protect the areas where the roosts are located and to maintain the food sources to which they are associated, as our results suggest that the roost will remain as long as there is an abundant and regular food supply. For this reason, solutions must be found that allow for the coexistence of the necessary san- itary regulations of waste dumps and the conservation needs of scavenging raptors (see Telia 2001). Resumen. — Describimos la dinamica y variacion temporal en la estructura de edades de un dorraidero comunal de Alimoches en Cataluna. La formacion y desaparicion del dormidero coincidio con la aparicion y desaparicion de la fuente de alimento. El dormidero se formo a menos de 2 km de la fuente predecible de alimento de la cual se abastecian y cambio en varias ocasiones su ubicacion Los primeros individuos ocuparon el dormidero a finales de abril y las maximas concentraciones se produjeron en junio. Debido a la regresion sufrida por la especie duran- te los ultimos 20 anos y la importancia que tienen estos lugares para su conservacion, es necesario proteger estos emplazamientos y compatibilizar la normativa sanitaria con el mantenimiento de los puntos de alimentacion de los que dependen. [Traduccion de autores] Acknowledgments We wi.sh to thank J. Feixa, P. Pelayo, and R. Pelayo for their interest and help. Comments of J.C. Bednarz, J.A. Donazar, and L. Gangoso improved the manuscript. C. Carboneras translated the text into English. Literature cited Ceballos, O. and J.A. Donazar. 1988. Actividad, uso del espacio y cuidado parental en una pareja de Alimo- ches {Neophron percnopterus) durante los periodos de dependencia de los polios. Ecologia 2:275-291. AND J.A. DonAzar. 1990. Roost-tree characteris- tics, food habits and seasonal abundance of roosting Egyptian Vultures in northern Spain./. Raptor Res. 24 19-25. de Pablo, F. 2000. Estatus del Alimoche Neophron perc- nopterus en Menorca, Islas Baleares. An. Ornitol. Balears 15:2-9. Del Moral, J.C. and R. Marti (Eds.). 2002. El Alimoche 256 Short Communications VoL. 37, No. 3 Comun en Espana y Portugal (I Censo Coordinado) . Ano 2000. Monografia n° 8. SEO/BirdLife, Madrid, Spain. Donazar, J.A. 1993. Los buitres ibericos. Biologia y con- servacion. J.M. Reyero, Madrid, Spain. , O. Ceballos, and J.L. Tella. 1996. Communal roosts of Egyptian Vultures {Neophron percnopterus): dy- namics and implications for the species conservation. Pages 189—201 in J. Muntaner and J. Mayol [Eds.], Biologia y Conservacion de las Rapaces Mediterra- neas. Monografias No. 4 de la SEO, Madrid, Spain. , C.J, Paiacios, L. Gangoso, O. Ceballos, M.J. GonzAlez, and F. Hir al do. 2002. Conservation status and limiting factors in the endangered population of Egyptian Vulture {Neophron percnopterus) in the Canary Islands. Biol. Conserv. 107:89-98. Margalida, a. 1997. Aparicion de un dormidero comu- nal de Alimoche {Neophron percnopterus) en Cataluha (NE Espafia). Bull GCA 14:73-77. , D. Garcia, and R. Heredia. 1997. Estimacion de la disponibilidad trofica para el Quebrantahuesos {Gypaetus barbatus) en Cataluha (NE Espana) e imph- caciones sobre su conservacion. Dohana Acta Vertebr 24:235-243. Parker, P.G., T.A. Waite, and M.D. Decker. 1995. Kin- ship and association in communally roosting black vultures. Anim. Behav. 49:395-401. Rabenold, P.P. 1983. The communal roost in Black and Turkey Vultures: an Information Center? Pages 303- 321 in S.R. Wilbur and J.A. Jackson [Eds.], Vulture biology and management. Univ. of California Press, Berkeley, CA U.S.A. . 1987. Recruitment to food in black vultures: ev- idence for following from communal roosts. Anim. Be- hav. 35:1775-1785. Telia, J.L. 1991. Dormideros de alimoches en el Valle Medio del Ebro. Actas I Congreso Internacional sobre Aves Carroheras: 69-74. AEDENAT-CODA, Madrid, Spain. . 2001. Action is needed now, or BSE crisis could wipe out endangered birds of prey. Nature 410:408. Received 15 April 2002; accepted 25 May 2003 /. Raptor Res. 37(3):256-258 © 2003 The Raptor Research Foundation, Inc. Diet of Three Sympatric Owls in Steppe Habitats of Eastern Kazakhstan Joan Navarro, Jose Antonio Sanchez-Zapata,^ Martina Carrete, and Francisco Botei.la Area de Ecologia, Departamento de Biologia Aplicada, Universidad Miguel Hernandez, Ctra Beniel km 3.2, Orihuela, Alicante, Spain Andrei Gavrilov and Sergei Sklyarenko Institute of Zoology, Akademgorodok, 480060 Almaty, Kazakhstan Jose Antonio Donazar, Olga Ceballos, and Fernando Hiraldo Department of Applied Biology, Estacion Biologica Donana, CSIC, Pabellon del Peru, Avda. Maria Luisa s/n, 41013 Sevilla, Spain Key Words: Eurasian Eagle-Owl, Bubo bubo; Little Owl, Athene noctua; Long-eared Owl, Asio otus; Asia, diet, steppe habitats. We studied the diet and food-niche overlap of three sympatric owls, the Eurasian Eagle-Owl {Bubo bubo), the Long-eared Owl {Asio otus) and the Little Owl {Athene noctua), during the breeding season in semiarid steppe habitats of eastern Kazakhstan and compared their food habits with other localities in the western Paleartic. These ^ Corresponding author’s e-mail address: toni@umh.es remote steppe habitats have been little studied and their breeding raptor communities resemble those of Mediter- ranean Europe. Study Area and Methods The study was conducted between 12 and 28 June 1999 in eastern Kazakhstan in central Asia. The climate is con- tinental, with very cold winters (when temperatures re- main under 0°C for months) , and warm summers. Such conditions can also be considered as arid or semiarid with annual rainfall <300 mm. Extreme temperatures and rainfall limit the growth of tree species (Walter 1981) and the landscape is dominated by steppe and semidesert September 2003 Short Communications 257 Table 1. Number and percent occurrence in the diet of three owl species during the breeding season in eastern Kazakhstan. N = number of prey. Prey Items Eurasian Eagle-Owl Long-eared Owl Little Owl N Percent N Percent N Percent Birds 54 17.3 2 2.5 8 4.1 Sturnus roseus 25 8.0 2 2.5 0 0 Falco naumanni 4 1.3 0 0 0 0 Unidentified 25 8.0 0 0 8 4.1 Reptiles 19 6.1 1 1.2 3 1.5 Lacertidae 2 0.6 0 0 1 0.5 Colubridae 9 2.9 1 1.2 0 0 Unidentified 8 2.6 0 0 2 1.0 Mammals 195 62-5 74 91.3 109 55.3 Microtidae 82 26.3 39 48.1 39 19.7 Muridae 12 3.9 27 33.4 8 4.1 Hemiechinus auritus 45 14.4 1 1.2 1 0.5 Sciuridae 39 12.5 6 7.4 8 4.1 Soricidae 1 0.3 0 0 0 0 Unidentified 16 5.1 1 1.2 53 26.9 Invertebrates 44 14.1 4 5.0 77 39.1 Coleoptera 13 4.2 2 2.5 53 26.9 Orthoptera 19 6.1 2 2.5 11 5.6 Unidentified 12 3.8 0 0 13 6.6 Total 312 81 197 plains and hills with herbaceous vegetation and small bushes, whereas trees are confined to foothills of large mountains (Tian Shan in the south, Altai and Alatul in the east) or around human settlements and river valleys. Pellets were collected in two different habitat types in the study area (Sanchez-Zapata et al. 2003): (1) Seminat- ural grasslands. Here the landscape was dominated by extensive livestock use and widespread croplands, but vil- lages are small with <500 inhabitants. There are many abandoned fields and degraded steppe areas. (2) Dry steppes. This includes large areas of natural dry steppes with little or no human presence. Trees are lacking and the vegetation is dominated by grasses and forbs {Arte- misia spp., Limonium spp., Salsola spp.. Ephedra spp., Hal- oxylon spp.). There are also sparse rocky outcrops. Fresh pellets were collected under perches around nests of the three species. Overall, we analyzed 108 pel- lets of the Eurasian Eagle-Owl collected in nine localities, 40 pellets of Long-eared Owls collected in two localities, and 107 pellets of Little Owls collected in four localities. All prey items were classified into 15 different categories (at the level of species, family, or order) included in four broad classes; birds, reptiles, mammals (Corbet 1978), and invertebrates. Diversity was quantified by means of the Levin’s Index {D = 1/X p^) whereas food niche over- lap was calculated using the Mac Ar thur-Levins ’ s Index modified by Pianka (0,y = S pkipk/^ (S 2 pkf) (Marti 1987, Krebs 1989). Results The diet of the eagle-owl included mostly mammals and birds (Table 1). The main prey species were voles (Microtidae) and medium-sized mammals such as hedge- hogs {Hemiechinus auritus) and hamsters {Citellus Sa- ms sp., Alactaga sp., Pygerethmus sp.) and Rosaceus Star- lings {Sturnus roseus). Long-eared Owls preyed almost exclusively upon small rodents such as mice (Muridae) and voles, whereas Little Owls fed mainly on small ro- dents and invertebrates, mostly beetles (Coleoptera) . The diversity of the diet of eagle-owls {D = 7.63) and Little Owls {D — 5,10) was higher than that of Long-eared Owls {D = 2.85) . Food niche overlap was higher between eagle and Long-eared Owl (O = 0.73), than between Little Owls and the two other species (O = 0.58 and 0.47, re- spectively) . The diet of the eagle-owl was more diverse in steppe habitat than that found in Mediterranean countries, where the species feeds mainly on rabbits ( Oryctolagus cu- niculus) , one of the staple food sources for top predators of Mediterranean ecosystems in Western Europe (Delibes and Hiraldo 1981). The absence of rabbits and the di- urnal habits of similar medium-sized rodents in steppe habitats of Central Asia may be responsible for this dif- ference. As a consequence, small rodents, hedgehogs, 258 Short Communications VoL. 37, No. 3 and birds (including raptors) play an important role in the diet, similar to that found in other semiarid Mediter- ranean habitats when rabbits are scarce (Jaksic and Marti 1984, Serrano 2000, Martinez and Zuberogoitia 2001) and other semiarid or lower-latitude Nearctic biomes (Donazar et al. 1989, Marchesi et al. 2002). Long-eared Owls had a similar diet to that reported for other populations in the Western Palearctic, with voles and mice as the main prey species, with the exception of a few localities in Mediterranean and suburban habitats where birds may also play an important role in the diet (Garcia and Cervera 2001). The diet of the Little Owl was also similar to that de- scribed in the Western Palearctic (Mikkola 1983), al- though invertebrates were more important in Mediter- ranean countries (Mahez 1983) and rodents were more important in temperate-climate areas including cold de- serts (Herrera and Hiraldo 1976). Our results point out that despite regional differences in prey communities, owls did not show large intercontinental differences in food-niche metrics, as suggested in previous broader- scale studies (Marti et al. 1993). ReSUMEN. — Se estudia la dieta de tres rapaces nocturnas; el mochuelo {Athene noctua), buho chico (Asio otus) y buho real {Bubo bubo) en habitats esteparios de Kazastan. Los restos de presas se obtuvieron del analisis de egagro- pilas recogidas en las proximidades de los nidos durante el periodo reproductor. Los invertebrados y microma- miferos fueron la presa principal de Athene noctua, mien- tras que Asio otus predo casi exclusivamente sobre micro- mamiferos y Bubo bubo presento una dieta muy diversificada, aunque basada en aves y mamiferos de me- diano tamano. La dieta de las tres especies fue mas sim- ilar a la observada en latitudes templadas y desiertos frios que a la descrita para ecosistemas mediterraneos semia- ridos. [Traduccion de los autores] Acknowledgments We are very grateful to Ana Emilia Lopez and Alicia Montesinos for their help with the pellet analyses. The Institute of Zoology of Almaty (Kazakhstan) provided valuable help in organizing the field work and a refer- ence skull collection. Literature cited Corbet, G.B. 1978. The mammals of the palearctic re- gion. A taxonomic review. British Museum. Cornell Univ. Press, London, U.K. Delibes, M. and F. Hiraldo. 1981. The rabbit as a prey in the Iberian Mediterranean ecosystems in western Europe. In K. Myers and C.D. Mclnnes [Eds.], Pro- ceedings of the World Lagomoph Conference, Guelph, Canada. Donazar, J.A., F. Hiraldo, M. Delibes, and R.R. Estre- lla. 1989. Comparative food habits of the Eagle Owl {Bubo bubo) and the Great-horned Owl {Bubo virginia- nus) in six Nearctic biomes. Ornis. Scand. 30:398-306. Garcia, A. and F. Cervera. 2001. Notas sobre lavariacion estacional y geografica de la dieta del buho chico. Ar- deola 48:54—80. Herrera, C.M. and F. Hiraldo. 1976. Food-niche and trophic relationships among European owls. Ornis Scand. 7:29—41. Jaksic, EM. and C.D. Marti, 1984. Comparative food habits of Bubo owls in Mediterranean-type ecosys- tems. Condor 86:288-296. Krebs, C.J. 1989. Ecological methodology. Harper & Row Publ., New York, NY U.S.A. MAnez, M. 1983. Espectro alimenticio del Mochelo co- mun {Athene noctua) en Espana. Alytes L275-289. Marchesi, L., F. Sergio, and P. Pedrini. 2002. Costs and benefits of breeding in human-altered landscapes for the Eagle-Owl Bubo bubo. Ibis 144:164. 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. National Wildlife Federation, Wash- ington, DC U.S.A. , E. Korpimaki, and F.M. Jaksic. 1993. Trophic structure of raptor communities: a three-continent comparison and synthesis. Pages 47-137 m D.M. Pow- er [Ed.], Current Ornithology, Vol. 10. Plenum Pub- lishing Corporation, New York, NY U.S.A. Martinez, J. A. and I. Zuberogoitia. 2001. The response of the Eagle Owl {Bubo bubo) to an outbreak of the rabbit haemorrhagic desease.y. Ornithol. 142:204-211. Mikkoia, H. 1983. Owls of Europe. T. &: A.D. Poyser, Berkhamsted, U.K. Sanchez-Zapata, J.A., M. Carrete, A. Gavrilov, S. Sklya- renko, O. Gebali.os, J.A. Donazar, and F. Hiraldo 2003. Land use changes and raptor conservation in steppe habitats of Eastern Kazakhstan. Biol. Conserv 111:71-77. Serrano, D. 2000. Relationship between raptors and rab- bits in the diet of eagle owls in southern Europe; com- petition removal or food stress. J. Raptor Res. 34:305- 310. Walter, H. 1981. Los sistemas ecologicos de lbs conti- nentes. Omega, Barcelona, Spain. Letters J. Raptor Res. 37(3):259-260 © 2003 The Raptor Research Foundation, Inc. An Unusual Case of Adoption of a Golden Eagle (Aquila chrysaetos) Chick in the Mala Fatra Mountains, Northwestern Slovakia Golden Eagle {Aquila chrysaetos) productivity was intensively studied in Slovakia in the period 1993-2000. Of nests containing at least two chicks that were not disturbed by man {N — 15), all exhibited brood reduction (Kropil and Majda 1996, Pages 486-494 in B.-U. Meyburg and R.D. Chancellor [Eds.], Eagle studies. WWGBP, Berlin, Germany; J. Korhan unpubl. data), which is high in comparison to other regions (McGrady 1997, Birds of the western palearctic, Update 1:99-114). Brood reduction was observed from hides or discovered when the nests were checked after the egg hatched and one chick was found killed. In addition. Golden Eagle productivity in Slovakia has been impacted by illegal nest thefts for decades (Kornan, M. et al. 1998, Bird Crime in Europe 4:11-13), which caused a 15-29% reduction in nesting productivity during the study period {N = 67). Therefore, nest wardens or guards were closely monitoring some sites (Korhan, J. et al. 1995, Buteo 7:163—168; Korhan, J. et al. 1996, Buteo 8:137-142). In order to reduce losses to brood reduction, the Slovak State Nature Conservation authorities manipulated two-chick broods to mitigate siblicide (cainism) and increase annual eagle productivity. The highest risk of cainism occurs when the chick IS <6 d old; however, it may occur until the chick is ca. 2 wk old. In two-chick broods, the older or younger chick was taken into captivity and either hand-reared or reared by a foster parent (eagle or other raptor). Then after the period of 5-6 wk, the nestling was returned to the original nest to be reared by its natural parents or fostered into nests of other pairs (Hrtan and Mihok 1985, Milvus 2:71-76; Kulstrunk and Trpak 1987, Pages 235-238 inj. Sitko and P. Trpak [Eds.], Dravci 1985. Sbornikz ornitologicke konference P0erov 14-16. 11. 1985, Statni ustav pamatkove pece a ochrany pfirody v Praze ve spolupraci s Okresnim vlastivednym muzeem J.A. Komenskeho, Moravsk}4n orni- tologickym sdruzenim v Pferove ve Statnim zemedelskem nakladatelstvi v Praze. Praha, CSSR; Simak and Mihok 1989, pages 176-179 inJ. Lukasek, I. Otahal, and P. Trpak [Eds.], Zachranne chovy a odchovy Novyjicin 1987. Statni ustav pamatkove pece a ochrany pfirody v Praze ve spolupraci s Okresnim vlastivedn)4n muzeem v Novem Jicine. Novy Jicin, CSSR; Simak and Mihok 1996, Pages 463-468 in B.-U. Meyburg and R.D. Chancellor [Eds.], Eagle studies WWGBP, Berlin, Germany). This letter describes an unusual case of chick adoption by foster parents that had lost their offspring and nest due to severe weather. In fact, this is the first repeated chick adoption to the unoccupied nest ca. 690 m away from the original nest, which was destroyed. In addition, we note that the attempt to foster the chick to the original recon- structed nest was not successful. Finally, in the case we describe, there was an unusually long, 7-d period between the time of death of the original chick and time of adoption by the foster eagle pair. As a result of heavy rain on the night of 26 May 2000, a Golden Eagle nest containing one 3-wk-old chick was dislodged and fell from the supporting tree. The nest was situated in a fir tree {Abies alba) near the village of Zazriva, in the Mala Eatra Mountains, Slovakia. Two days later, a ranger of the Mala Fatra National Park checked the area around the nest tree, but was unable to find the chick. At 2000 H on the following day (29 May 2000) two rangers spent about 1 hr rebuilding the nest at the original location in the tree, using mainly material from the nest. Thick branches were used to improve stability, the nest cup was lined with green branches and prey remains (e.g., feathers, bones) were spread on the outer edge of the nest. On 30 May 2000, a 3-wk-old Golden Eagle chick was removed from a nest containing a brood of two, and placed in the newly-reconstructed nest at 0530 H. A hide was constructed about 400 m away from the nest. Adult eagles were not seen near the new nest for the remainder of the day, and the chick was left on the nest overnight. On the morning of 31 May 2000 an adult female eagle was seen flying from the slope opposite the nest, but did not visit it Later that day, we climbed the nesting tree to feed the chick. At 2030 H we took the nestling eagle from the nest because no eagles had visited the nest. Despite the lack of success, we decided to make another attempt at fostering this chick to this pair. Before 1700 H, on 1 June 2000, the same chick was placed in an alternate nest in this eagle home range that was built on a cliff about 690 m from the nest that had collapsed (Fig. 1). The position of this nest, which had been successful in 1999, was quite open, and the chick in the nest could be seen from far away (probably >1 km). We restored and tidied up the nest cup. At 1440 H on 2 June 2000 an adult female Golden Eagle visited this nest, and immediately the nestling stood and begged for food. The female remained at the nest about 5 min. At 1600 H the female returned 259 260 Letters VoL. 37, No. 3 Figure 1. Locations of the original nest that failed (Nest 1) and the alternate nest (Nest 2) where a 3-wk-old chick was adopted by the resident pair of Golden Eagles near Zazriva, Slovakia. carrying rabbit-sized prey, and fed the chick for ca. 30 min and then flew away. Afterwards, the adult eagles attended the chick in a normal manner. The nestling fledged during the last week of July. The first manipulations of Golden Eagle broods in Slovakia were carried out in 1978 (Simak and Mihok 1996). The younger chicks were removed and fostered to other eagle pairs that had clutches destroyed, had abandoned unhatched eggs, or had suffered nesding mortality of a single-chick brood. Fostering was successful 74.3% {N = 35) of the time (Korhan, J. 1995, Buteo 7:163-168; J. Korhan unpubl. data; Korhan, J. et al. 1995, Korhan, J. et al. 1996) in the period 1993-2000. The fostering failures were caused by siblicide, natural predation, or unknown causes. One attempt to foster a chick to an empty nest was not successful, the chick was killed by the male and consequently eaten by both adult birds (Hrtan and Mihok 1985). According to the authors, the pair probably did not lay the clutch. The failure of this adoption was probably related to the nonbreeding condition of the pair. Comparable adoption of an eagle chick to an unoccupied, alternate nest, as described in this letter, has not been described elsewhere as far as we could determine through our literature review. An adoption of a young Golden Eagle into a nest where the chick died and the nest was vacant for at least 2 d and possibly 5 d was conducted in the Snake River Birds of Prey Area, Idaho, U.S.A., in 1976 (M. Kochert pers. comm.). In this case, the foster chick was adopted and fledged successfully. We are grateful to Mike J. McGrady, Michael N. Kochert, and the anonymous referee for reviewing the manuscript and making useful suggestions. — Martin Kornan, Center for Ecological Studies, 013 62 Veike Rovne 14, Slovakia; e- mail address: mkornan@stonline.sk and Milos Majda, Metod Macek, and Jw Kornan, State Nature Conservancy of the Slovak Republic, Lazovna 10, 974 01 Banska Bystrica. Received 9 May 2002; accepted 13 April 2003 September 2003 Letters 261 J. Raptor Res. 37(3) :261 © 2003 The Raptor Research Foundation, Inc. SiBLICIDE AND CANNIBALISM IN THE BoOTED EaGLE {HiEEAAETUS PENNATUS) IN THE Tietar Vallly, Central Spain Facultative siblicide is assumed to occur in broods of the Booted Eagle {Hieraaetus pennatus; Simmons 1988, Ibis 130:339-357), although only minimal sibling aggression has been described in this species (Brown and Amadon 1968, Eagles, hawks, and falcons of the world. Country Life Books, London, U.K.; Steyn and Grobler 1981, Ostrich 52A08- 118; Steyn 1982, Birds of prey of Southern Africa. David Philip, Johannesburg, South Africa). Furthermore, there are no records of cannibalism for this species. Here I report two instances of both siblicide and subsequent canni- balism recorded in the Tietar River valley (central Spain, 4°42'W, 40°40'N) in 2000, during observations made from hides. Detailed information of the study area can be found in Garcia Dios and Vinuela (2000, Ardeola 47:183-190) In a nest with two chicks, where observations started on 29 May 2000, the chicks were of similar size and about 5 d old. During my next visit (5 June), one chick was much larger than the other. On 7 June, the smaller chick had an obvious bare skin patch on its back, as if the down had heen plucked out. On 8 June, the larger chick harassed and pecked the head and back of its sibling constantly; the adult birds were at the nest for short periods, but did not stop the sibling aggression. On the early morning of 10 June, the small chick begged less frequently and less loudly than during previous visits and the large chick pecked the smaller chick’s head and back. A few hours later, presumably soon after the small chick died, the adult female started to feed the surviving chick with the remains of Its sibling, providing 257 individual morsels. The second record concerned a nest that contained two chicks about 2- and 4-d-old on 14 June 2000. Neither sibling aggression nor wounds were observed on that date. On 16 June, the female fed both chicks, but the small chick pecked at the head and back of its sibling repeatedly in the morning. On 18 June, the small chick initiated sibling aggressive behavior; the large chick pecked at its sibling less often but used more force. The adult female stopped the agnostic behavior by brooding both chicks. On 19 June, adults were not seen at the nest when 1 began observations. The small chick was more active than its larger sibling, it begged and moved around the nest, but no aggressive behavior was observed. The adult female brought one small prey, which was eaten by the large chick. Soon after, the small chick attacked but the large chick replied by pecking more often and also more forcefully than its sibling. This time, the female did not stop the sibling aggression. On 19 June, I climbed up the nest and found no prey remains. I made an examination of the chick’s bodies and no wounds were observed, although the small chick was weak and emaciated. On 20 and 21 June, I monitored the nest continuously from dawn to dusk and observed no prey deliveries to the nestlings. In the morning on 22 June, I observed the adult female pecking the head of the small chick, which seemed to be dead. During 4 hr the female brooded the two chicks. After that, she picked up the dead chick and started to feed the surviving chick, which was begging. The surviving chick ate 206 food morsels and the female 14. In this population of Booted Eagles, unlike in other areas (e.g., southern Africa; Steyn 1982), brood-reduction seems to be frequent; sibling aggression has been regularly observed in almost all of the nests monitored {N = 21) and at least two additional suspected cases of siblicide occurred during 2001. I am grateful to Juan Gomendio who allowed me to do this study. This work could not have been carried out without the help of Manuel Garcia Tornero and Javier Munoz Familiar. Javier Vinuela and Juan Jose Negro critically reviewed previous drafts. This paper is dedicated to warden Marcos, from the town of Guisando, as a model of what exploiting and preserving nature should be. — Ignacio S. Garcia Dios, Instituto de Investigacion en Recursos Cine- geticos (C.S.I.C.-U.C.L.M.), Libertad 7“, 13004 Ciudad Real, Spain; Present address: C. /Plaza del padre Felipe Fer- n^dez n° 1-2° Piso, 05416 EL ARENAL (Avila); e-mail address: isgdios@eresmas.com Received 13 November 2001; accepted 13 April 2003 Associate Editor: Juan Jose Negro BOOK REVIEW J Raptor Res. 37(3):262-263 © 2003 The Raptor Research Foundation, Inc. Cool North Wind: Morley Nelson’s Life with Birds of Prey. By Stephen Stuebner. 2002. Caxton Press, Caldwell, ID. xx + 431 pp., 17 color photo- graphs, numerous black-and-white photographs. ISBN 0-87004-426-5. Cloth, $24.95, — ^Admirers of Morley Nelson will enjoy the stories told in this book, as will all other raptor enthusiasts. This ac- count describes the life of a man of action, a man with a fierce passion for birds of prey. It traces his introduction to the family farm in North Dakota, where Morley first became acquainted with wild birds of prey, and his beginnings in falconry with a Red-tailed Hawk {Buteo jamaicensis) taken from a nest on the farm. From the first encounter with that hawk, there was no turning back for Morley. His life has revolved around birds of prey for more than seven decades, and the world is a better place for it. This book makes a case for inclusion of Mor- ley Nelson among the world’s all-time raptor “pro- tectors,” and in the process of ranking Morley Nel- son’s accomplishments in comparison with other notable conservationists, may actually do Morley a disservice. Morley was born in 1917 and raised on a farm in North Dakota. Many of the experiences of living on a farm shaped his life, including lessons about killing animals for food, dealing with all sorts of animals, and overcoming hardships with resolve and determination. A defining moment in young Morley’ s life occurred while he rode his horse “Slim” across a pasture that held a pond. The horse inadvertently caused four Green-winged Teal {Anas crecca) to flush from the pond, and Morley was amazed to see a tremendously exciting phe- nomenon of nature: a Peregrine Falcon (Falco per- egrinus) diving out of the blue, striking one of the teal with an explosion of energy and feathers, grab- bing the teal out of the air, after inverting in flight, and flying off with it. Later, Morley had opportunity to tell his father and grandfather about what he had seen, and he expressed a fervent desire to obtain a hawk and try to train it. With his father’s blessing, Morley went out and located the nest of a pair of Red-tailed Hawks, and with a fair amount of effort, climbed the tree and pulled an eyas to take home. Morley used his own sensibility and ingenuity, and support from his father and grandfather, to figure out how to “man” the hawk, train it to hunt with him, and to take game on the family farm. Morley became a self-taught falconer and began at age twelve a lifelong pursuit that made him famous, led to his personal connections with governors and people in entertainment and high places, and enabled him to do what he could to educate the public and protect wild raptors and their habitats. Cool North Wind describes the progress of Mor- ley’s life as a student at North Dakota State Uni- versity, where he obtained a bachelor of science degree and specialized in soil science and hydrol- ogy. Morley answered an advertisement for a job with the Soil Erosion Service in New Mexico and moved there, where he was pleased to meet Luna Leopold, then a professor at the University of New Mexico. Contact with the son of famed ecologist Aldo Leopold made a lifelong impression on Mor- ley. The attack on Pearl Harbor pulled Morley, like many other Americans of his age, into World War II, and he became an officer in the famed Tenth Mountain Division. This elite Army division re- quired very strenuous physical training in moun- tainous survival-type settings. Morley proved to be a perfect fit for this assignment, and he perfected mountaineering skills that later were very useful in climbing to Prairie Falcon {Falco mexicanus) and Golden Eagle {Aquila chrysaetos) nests. Morley eventually entered combat with his division and performed superbly, even when wounded in com- bat. Yet, he maintained a bit of a nonconformist streak, going absent without leave during a period of convalescence from a war wound in order to take a jeep ride to check out raptor nests in Yu- goslavia. Cool North Wind continues as it traces Morley’s life after the war, his falling in love with his first wife, Betsy Ann, and the arrival of their three chil- dren. Morley resumed what was to become a very successful career with the Soil Conservation Ser- vice. But raptors and falconry continued to play a 262 September 2003 Book Review 263 dominating role in his life. His children grew up with Prairie Falcons and Golden Eagles in the house, as well as Ferruginous Hawks {Buteo regalis) and many other raptors. Rehabilitation of injured raptors became a part of the “normal” life in the Nelson household. Over the years, various unique challenges and opportunities presented themselves, as Morley be- came regionally and even nationally known for his work with birds of prey and his passion for them. Morley became a consultant in the preparation and filming of a number of feature films, including those made by the Disney Corporation. These films featured birds of prey that were trained by Morley and allowed for hunting and other spectacular ae- rial scenes that depicted the lives, natural history, and magnificence of various species of raptors. Morley and his sons eventually formed their own production company and have produced their own high-quality hlms featuring birds of prey. These fabulous film productions have educated the pub- lic on natural history of birds of prey, as well as promoting admiration, protection, and conserva- tion of the birds and their habitats. Morley was able to translate his intimate knowl- edge and appreciation of the unique Snake River nesting habitat as one of the richest breeding areas for birds of prey in North America into establish- ment of a National Conservation Area for these birds, which were at risk of losing their critical hab- itat to the growing agriculture industry. He was also hired by Idaho Power to help redesign power poles and structures to minimize dangers of electrocu- tion to Golden Eagles and other large raptors. Morley used a mock power pole setup at his resi- dence to experiment with his own live eagles and thus learn design techniques that would minimize risk of electrocution. All the while, Morley used his voice and clout to call for an end to shooting of raptors and for an increase in facilities to care for sick and injured raptors. Cool North Wind tells some stories of alter- cations between Morley and citizens who made the mistake of threatening to shoot wild raptors, in which case harsh words were exchanged and fights were narrowly averted. This book contains many, many good stories. But it has some flaws. Very little is said about the actual practice of falconry by Morley Nelson. Falconry is the taking of live quarry with trained birds of prey, and very little is said about Morley’s hunting with trained falcons, which detracts from the book’s po- tential appeal to falconers. There are inaccuracies in the book, such as a story of Morley’s training in the Aleutian Islands prior to WWII combat, where Morley encountered a Peale’s Peregrine Falcon. The Peale’s Peregrine is incorrectly described as an “arctic” subspecies of the Peregrine Falcon. The principal flaw of this biography of such an impressive man is the attempt to rank his accom- plishments with those of other raptor “protection- ists.” By making such an attempt, further scrutiny is invited. For instance, although it is true that Morley Nelson consulted with the power industry in Idaho to develop transmission facilities that would be safer for raptors, what is not emphasized is that by so doing, the power industry has been able to ward off federal regulation that could have imposed more stringent safety standards for rap- tors. Even the successful preservation of the Snake River Birds of Prey National Conservation Area tells only part of the story. Compromises were made with local land users, including cattlemen and the Idaho National Guard, in order to obtain political consent to preserve the Snake River Birds of Prey Area from agricultural development. As a result, those “conservation” lands have never been managed primarily for the benefit of the resident birds of prey, despite the name of the BLM-man- aged area. Consequently, a tremendous negative change has occurred in the conservation area over the past few decades, and habitat is now altered considerably. Shrubsteppe habitats are highly de- graded, invasive cheatgrass is reducing lagomorph populations, fire regimes have become catastroph- ic, Golden Eagle and Prairie Falcon numbers are in long-term decline, and no one has found a so- lution for protecting the Snake River Birds of Prey National Conservation Area from its own manage- ment. Morley Nelson is now well aware of the de- clines of eagles and falcons, and he surely appre- ciates that saving the lands from cultivation was not the permanent solution he might have hoped for. As a biography of a fascinating life, this book deserves reading by all raptor enthusiasts. It pro- vides a history of a man who truly is an American original. Morley Nelson has won respect from es- sentially everyone he has encountered during his long life, and his fierce passion for birds of prey has been contagious and inspiring to countless ad- mirers from all walks of life. — Stan Moore, 4220 Langner Avenue, Santa Rosa, CA 95407 U.S.A. The Wheeler Guides Raptors of North America Brian K. Wheeler « « # » # # # # » # « « # # # « # # « « * 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 These books are the best and most complete guides to North American hawks, eagles, and other raptors ever published. Abundantly illustrated with hundreds of full-color, high-quality photographs, they are essential for anyone seeking to identify these notoriously hard-to- identify birds. The range maps are the most accurate and largest ever produced for North American raptors. Raptors of Eastern North America Raptors of Western North America "Brian Wheeler's background as a raptor artist and photogra- pher combined with his exten- sive field expe- rience provide the foundation for an excellent book. The detailed plum- age descrip- tions of each species and sub-species. "A highly signif- icant contribu- tion to the field. Wheeler has accomplished what no other guide has done for any group of birds. The quality of photography, scope and depth on plum- ages, races, and variations, are unequaled. The range maps are stellar — the best ever done for any guide. It is the ultimate raptor reference and will be for decades to come." — Clay Sutton 559 color illustrations. 37 maps. Cloth $45.00 ISBN 0-691-11598-2 Not available from Princeton in the Commonwealth (except Canada) and the European Union 600-1- newly published raptor images plus extensive range maps and migration data will appeal to the general birding community and ensure that this guide becomes the reference book of raptor enthusiasts." — Ned Harris 622 color Illustrations. 56 maps. Cloth $49.50 ISBN 0-691-1 1599-0 Not available from Princeton in the Commonwealth (except Canada) and the European Union Also Available A Photographic Guide to North American Raptors Brian K, Wheeler and William 5 . Clark Paper $24.95 ISBN 0-691-1 1 644-X 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 PRINCETON University Press 800-777-4726 • READ EXCERPTS ONLINE WWW.BIRDS.PRINCETON.EDU B UMTEDSTATES IPOmLSERVKET^ Statement of Ownership, Management, and Circulation (Required by 39 USC 3685) 2. Publication Number 3. Filing Dale Raptor Research 0 6 9 2 - 1 0 1 6 4 Issue Frequency Quarterly 5. Number of Issues Published Annually 4 6. Annual Subscription Pries $33.00 7 Complete Mailing Address of Known Office of Publication (Not printer) (Street, city, county, state, and Zlp*4) Raptor Research 810 E. 10th Street p o Box aea Contact Person Karen Ridgwav Telephone (785)843-1234 ext. 136 5 Complete Mailing Address of Headquarters or General Business Office of Publisher (Not printer) Raptor Research 810 E- 10th Street pn Rr%y:^fiR 6 Full names and Complete Mailing Addresses of Publisher, Editor, and Managing Editor (Do not leave blank) Publisher (Name and complete mailing address) Jim Fitzpatrick Treasurer Camenter Nature Center Editor (Name and complete mailing address) James C. Bednarz, Department of Biological Sciences. P.O. Box 599, ArkariRas State tiniversih/ Managing Editor (Afe/ne a/rd compiB^ mafiing addnss) 10 Owner (Do ixil leave blank. If Ihe publication is owned by a corporation, give the name and address of the corporation immediately followed by the names and addresses of all stockholders owning or bolding 1 peroonl or more oflbe total amount of stock. If not owned by 9 corporation, give the names and addresses of the individual ovmers. if owned by a partnership or other unmoorporated firm, give Ita name and address as well as those of each IndiNddual owner. If the publlcaiton is published by a nonprofit organization, give Its name and address.) Pull Name Complete Mailing Address Raptor Research 810 E. 10th Street P.O. Box 368 Lawrence, KS 66044 1 1 Known Bondholders, Mortgages, end Other Security Holders Owning or Holding 1 Percent or UoreofTotalAmounlof Bonds, Mortgages, or Other Secunties, If none, check box ► Lii None Full Name Complete Mailing Address 12 Tax Status (For compleUon by nonprofit organizations authorized to mall at special rates) (Check one) The purpose, function, and nonprofit status of this organization and the exempt etatus for federal Income tax purpoees: 0 Has Not Changed During Preceding 12 Months 1 I Has Changed During Preceding 12 Months (Publisher must submit eiqilanation of change with this statement) 13. PuMication Title Raptor Research 14. Issue Date for Circulation Data Below July-03 15, Extent and Nature of Circulation Average No. Copies Each Issue During Preceding 12 Months Actual No. Copies of Single Issue Published Nearest to Filinq Date a. Total Number of Copies (Net press run) 1375 1250 b. Paid and/or Requested Circulation (1) Paid/Requested Outside- County Mail Subsi^ptions stated on Form 3541. (inctuda acfi«rf/ser's proof and BKchanae copies) 806 761 (2) Paid ln-(Dour>ty Subscriptions (IncJuda adverf/spr's proof and exchange copies) 0 0 (3) Sales Through Dealers and Carriers, Street Venders, Counter Sales, and Other Non-USPS Paid Distribution 0 0 (4) Other Classes Mailed Through the USPS 391 331 c. Total Paid and/or Requested Circulation (SurT; of 15b(1) , (2f,(3), and (4))^ 1197 1092 d. Free Distribution by Mall (SamptBS.complim enlary, and other hee) (1) Outside-County as Stated on Form 3541 2 2 (2) In-County as Stated on Form 3541 0 0 0> Other Classes Mailed Through the USPS 0 0 a. Free Distribution Outside the Marl (Canters or oF/rer means) 0 0 F. Total Free Dislrlbutiai (Sum of 1 Sd and I5e} ^ 2 2 g. Total Distribution fSu/n of 75ca/)d 15F) ^ 1199 1094 h. Copies not Distributed 177 156 1. Total (Sum of 15g. and h.) ^ 1375 1250 Percent Paid and/or Requested Circulation (15c divided bv 15o times 100) 99.8 99.6 ,16. Publication of Statement of Ownership ' [Zl Publication required. Will be printed in the_ Jssue of this publlcatiMi. i I Publication not required. 17. Signature and Title of Editor, Publisher, Business Manager, or Owner o/i>^Jos. 1 certify that all information furnished on this ferm Is true and complete. I understand that anyone who furnishes False or misleading Infomiatton an this form or who oiT^ts material or Information requested on the form may be subject to criminsl senclions (including fines and hnprtaonmenl) and/or dvil sanctions (including civil penalties) Instructions to Publishers 1. Complete and file ore cc^jy of this Form with your postmaster annually on or before October 1. Keep a copy of the completed form for your records. 2. In Cases where the stockholder or security holder Is a trustee. Include in Items 10 and 11 the name of the person or corporation For whom Ihe trustee Is acting. Also Indude the names and addresses of individuals who are stockholders who own or hold 1 percent or more of the total amwniof bonds, mortgages, or other securities of the publishing corporation, In item 11. if none, check the box, Use blank sheets if more space is required. 3. Be sure to furnish ah circulation Information called for In item 15, Free circulabon must be shown in Items 15d, e.and F, 4. Item 15h., Coptee not Distributed, must Include (1) newsstand copies originally stated on Form 3541, and returned to the publisher, (2) estimated returns From news agents, and (3), copies for office use, leftovers, spdled, and all other copies not distributed. 5. If the publication had Periodicals authorization as a general or requester publication, this Statement of Ownership, Management, and Circulation must be published; It must be printed in any issue In October or, If the publication is not published during Oct^er, the First issue printed after October. 6. In item 16, indicate the date of the Issue in which this Statement of Ownership will be published. 7. Item 17 must be signed. Fa,i7ufB fo file orpvtiish a sfafemert ofowners/iip may iead to suspansion of sacond-ctass airthorizsaion. PS Form 3526, September 1C96 (See Instructors or> Reverse) Computerized Facsimile PS Form 3526, September 19S8(Reverse) Computerized Facsimile orders:8oo-722-246o Phone:434-263-4842 fax:434-263-4842 Buteo Books is the largest retailer of Ornithology books in North America^ with over 2,000 in~-print titles in stock. A FEW BOOKS ON BIRDS OF PREY AND FALCONRY Handbook of the Birds of the World. The first seven volumes of this projected sixteen- volume work are available for $ 1 85 each. Volume 2 includes diurnal raptors, Volume 3 includes owls. Volume 8, the first to cover Passerines, due Summer 2003. A Photographic Guide to North American Raptors. Wheeler & Clark. $24.95 (New Eastern and Western editions by Wheeler are due September 2003.) Ecology and Conservation of Owls. Ian Newton. $95 Birds of Prey: Health and Disease. Third Edition. Cooper. $96.95 North American Falconry and Hunting Hawks. Beebe & Webster. $55 Buteo Books has a wide selection of used and out-of-print titles available. Our stock changes daily, so call to check availability of that scarce title you're seeking. In 1992 the American Ornithologists' Union, in partnership with the Academy of N atural Sciences of Philadelphia, undertook the publication of species accounts for each of the more than 700 species which breed in the United States and Canada. These illustrated reviews provide comprehensive summaries of the current knowledge of each species, with range maps and an extensive list of references. Buteo Books is pleased to offer these individual species accounts for $7.50 each. All 716 profiles are listed in taxonomic order on our website (conforming with the Seventh Edition of the A.O.U. Check-list of North American Birds). Singles may be ordered by contacting Buteo Books. Shipping and handling is $4 for the first profile and $1 for each additional profile to a maximum of $10 per order. RARE AND OUT-OF-PRINT ORNITHOLOGY BIRDS OF NORTH AMERICA SERIES Visit our website for more information: www.buteobooks.com allen@buteobooks.com 3130 Laurel Road; Shipman, VA 22971; USA A Telemetry Receiver Designed with The Researcher in Mind What you've been waiting fori ftnjlly, a highly scniituc 999 channel synlhc^i.'ed telemetry receiver that wreiglrs leis than 13 ounces, is completely user programmable and offers variable scan rales over all frequencies. For each animal being tracked, the large LCD display provides not only the frequency (to IOOM 2 ) and channel number, but also a ; character alphanumeric comment field and a digital signal strength meter. Stop carrying receivers that are tin* si/e of a lunch box or cost over SlSOO. The features and performance of the new R-1000 pocket sized telemetry receiver will impress you. and the price will convince you. Other Features Include: • factory tuned to any 4MHz wide segment m the 148-174MH/ Band • Very high sensitivity of -l^SdBm to l^iOdBm • Illuminated display and keypad for use in low light or darkness • liser selectable scan rates from 1-30 seconds in 1 second steps • Rechargeable batteries operate the receiver for 12 hours and can b<* replaced with standard AA Alkaline batteries in the field. Both 12vdc and llOvac chargers arc included. • 6.r (15.5cm) high. 2.6’ (6.6cm) wide. 1.5" (3.8cm) deep. • 3 year warranty • 1 day delivery $ 695.00 Please specify desired 4MH/ wide segment in the 148 174MM/ band 2004 ANNUAL MEETING The Raptor Research Foundation, Inc. 2004 annual meeting will be held on 9-13 November 2004 in Bakersfield, California. For information about the meeting see the following website: http://www. calhawkingclub.org/fleld_meet/34th_annual/ or contact Rick Holderman (parabuteol@cox.net). 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 Rap Urr Research (ISSN 0892-1016) is published quarterly and available to individuals for $33.00 per year and to libraries and institutions for $50.00 per year from The Raptor Research Foundation, Inc., 14.377 117th Street South, Hastings, Minnesota 55033, U.S.A. (Add $3 for destinations outside of the continental United States.) Periodicals postage paid at Hastings, Minnesota, and additional mailing offices. POSTMASTER: Send address changes to The Journal of Raptor Research, OSNA, RO. Box 1897, Lawrence, KS 66044-8897, U.S.A. Printed by Allen Press, Inc., Lawrence, Kansas, U.S.A. Copyright 2003 by The Raptor Research Foundation, Inc. Printed in U.S.A. © This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). Raptor Research Foundation, Inc., Awards Lifetime Achievement Awards The Tom Cade Award recognizes an individual who has made significant advances in the area of captive prop- agation and reintroduction of raptors. Nomination packets can be submitted at any time. Contact: Brian Walton, Predatory Bird Research Group, Long Marine Laboratory, University of California, Santa Cruz, CA 95064 U.S.A.; tel. 408-459-2466; e-mail: walton@cats.ucsc.edu. The Fran and Frederick Hamerstrom Award recognizes an individual who has contributed significantly to the understanding of raptor ecology and natural history. Nomination packets can be submitted at any time. Con- tact: Dr. Clint Boal, Texas Cooperative Fish and Wildlife Research Unit, BRD/USGS, Texas Tech University, 15th Street & Boston, Ag Science Bldg., Room 218, Lubbock TX 79409-2120 U.S.A.; tel. (806) 742-2851; e-mail: cboal@ttacs.ttu.edu. Student Recognition and Travel Assistance Awards The James R. Koplin Travel Award is given to a student who is the senior author and presenter of a paper or poster to be presented at the RRF annual meeting for which travel funds are requested. Contact: Dr. Patricia A. Hall, 5937 E. Abbey Rd. Flagstaff, AZ 86004 U.S.A.; tel. 520-526-6222; e-mail: pah@spruce.for.nau.edu. Application Deadline: due date for meeting abstract. The William C. Andersen Memorial Award is given to the students who are senior authors and presenters of the best student oral and poster presentation at the annual RRF meeting. Contact: Laurie Goodrich, Hawk Mountain Sanctuary, 1700 Hawk Mountain Road, Kempton, PA 19529 U.S.A.; tel. 610-756-6961; email: goodrich@hawkmountain.org. Application Deadline: due date for meeting abstract; no special application is needed. Grants For each of the following grants, complete applications must be submitted to the contact person indicated by 15 February. Recipients will be notified by 15 April. The Dean Amadou Grant for $200-400 is designed to assist persons working in the area of distribution and sys- tematics (taxonomy) of raptors. Contact: Dr. Carole Griffiths, 251 Martling Ave., Tarrytown, NY 10591 U.S.A.; tel. 914-631-2911; e-mail: cgriff@liu.edu. The Stephen R. TuUy 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. Kim Titus, Alaska Department of Fish and Game, Division of Wildlife Conservation, P.O. Box 240020, Douglas, AK 99824 U.S.A.; e-mail: kimt@lishgame.state.ak.us. The Leslie Brown Memorial Grant for up to $1,000 to support research and/or dissemination of information on birds of prey, especially to proposals concerning African raptors. Contact: Dr. Jeffrey L. Lincer, 9251 Golondrina Dr., La Mesa, CA 91941 U.S.A.; e-mail: jefflincer@tns.net.