The Journal of Raptor Research Volume 36 Number 4 December 2002 Published by The Raptor Research Foundation* Inc THE RAPTOR RESEARCH FOUNDATION, INC. (Founded 1966) OFFICERS PRESIDENT: Brian A. Miu sap SECRETARY: Judith Henckel VICE-PRESIDENT: Keith L. Bildstein TREASURER: Jim Fitzpatrick BOARD OF DIRECTORS NORTH AMERICAN DIRECTOR #1: Philip Detrich NORTH AMERICAN DIRECTOR #2; Iaurie J. Goodrich NORTH AMERICAN DIRECTOR #3: Jeff P. Smith INTERNATIONAL DIRECTOR #1: Eduardo Inigo-Elias INTERNATIONAL DIRECTOR #2: Ricardo Rodriquez-Estrella INTERNATIONAL DIRECTOR #3: Beatriz Arroyo DIRECTOR AT LARGE #1: Jemima ParryJones DIRECTOR AT LARGE #2: Petra Bohall 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; Ed Henckel EDITORIAL STAFF EDITOR: James C. Bednarz, Department of Biological Sciences, P.O. Box 599, Arkansas State University, State University, AR 72467 U.S.A. ASSOCIATE EDITORS James R. Beithoff Clint W. Boai. Joan L. Morrison Juan Jose Negro Marco Resi ani Ian G. Warkentin Troy I. Wellicome BOOK REVIEW EDITOR: Jeffreys. 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: Rebecca S. Maul, Kristina Baker, 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 figure legends on a separate page. Each illustration should be centered on a single page and be no smaller than final size and no larger than twice final size. The name of the author (s) and figure number, assigned consecutively using Arabic numerals, should be pencilled on the back of each figure. Names for birds should follow the A.O.U. Checklist of North American Birds (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: Bald Eagles {Haliaeetus leucocephalus) . Painting by John Schmitt. Contents Foraging Ecology of Nesting Bald Eagles in Arizona, w. Grainger Hunt, Ronald e. Jackman, Daniel E. Driscoll, and Edward W. Bianchi 245 Vernal Migration of Bald Eagles from a Southern Colorado Wintering Area. Alan R. Harmata 256 Does Northern Goshawk Breeding Occupancy Vary with Nest-stand Character- istics ON THE Olympic Peninsula, Washington? Sean p. Finn, Daniel e. Variand, and John M. Marzluff 265 Subordinate Males Sire Offspring in Madagascar Fish-Eagle {Hauaeetus vocife- ROIDES ) PoLYANDROUS BREEDING GROUPS. Ruth E. Tingay, Melanie Culver, Eric M. Hallerman, James D. Fraser, and Richard T. Watson 280 Nesting and Perching Habitat Use of the Madagascar Fish-Eagle. James Berkeiman, James D. Fraser, and Richard T Watson 287 Use of Vegetative Structure by Powerful Owls in Outer Urban Melbourne, Victoria, Australia — Implications for Management. Rayiene Cooke, Robert Waiiis, and John White 294 Nest-site Selection of the Crowned Hawk-Eagle in the Forests of Kwazulu- NATAL, South Africa, and TaI, Ivory Coast. Gerard Malan and Susanne Shultz 300 Short Communications Juvenile Dispersal of Madagascar Fish-Eagles Tracked by Satellite Telemetry. Simon Rafanomezantsoa, Richard T. Watson, and Russell Thorstrom 309 Prey of the Peregrine Falcon (Falco Peregiunus cassini) in Southern Argentina and Chile. David H. Ellis, Beth Ann Sabo, James K. Tackier, and Brian A. Millsap 315 An Elevated Net Assembly to Capture Nesting Raptors. Eugene A. Jacobs and Glenn A. Proudfoot 320 Florida Bald Eagle {Hauaeetus leucocephalus) Egg Characteristics. M. Alan Jenkins, Steve K Sherrod, David A. Wiedenfeld, and Donald H. Wolfe, Jr. 324 Osprey Ecology in the Mangroves of Southeastern Brazil. Robson Silva e Silva and Fabio Olmos 328 Diet of Breeding Tropical Screech-Owls (Otus chouba) in Southeastern Brazil. Jose Carlos Mottajunior 332 Letters Comments of the First Nesting Record of the Nest of a Slaty-backed Forest Falcon {Micrastur MIRANDOUEJ ) IN THE ECUADORIAN AMAZON. Russell Thorstrom 335 Micrastur or Acupiter, That is the Question. Tjitte de Vries and Cristian Melo 337 Book Reviews. Edited by Jeffery S. Marks 338 Information For Contributors 340 Index to Volume 36 344 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. 36 December 2002 No. 4 /. Raptor Res. 36(4):245-255 © 2002 The Raptor Research Foundation, Inc. FORAGING ECOLOGY OF NESTING BALD EAGLES IN ARIZONA W. Grainger Hunt/ Ronald E. Jackman, and Daniel E. Driscoll Predatory Bird Research Group, Long Marine Laboratory, University of California, Santa Cruz, CA 93060 U.S.A. Edward W. Bianchi Entrix, 7919 Folsom Blvd., Suite 100, Sacramento, CA 93826 U.S.A. Abstract. — ^We studied foraging ecology of nesting Bald Eagles (Haliaeetus leucocephalus) in Arizona during 1987-89, with emphasis on the influence of dams and river flow regulation. We examined diet, foraging modes, habitat selection, fish abundance, and factors associated with fish availability. Based on biomass, prey remains yielded 76% fish, 14% mammals, and 10% birds. On rivers, eagles primarily caught live fish as they spawned or foraged in shallow water, whereas, on reservoirs, most fish were obtained as carrion or as they floated moribund on the surface. Fish communities differed among river reaches and reservoirs, and ecological and life-history characteristics influenced vulnerability and sea- sonal differences in exploitation, Water temperature, a principal factor determining fish community structure among eagle territories, was also associated with temporal differences in fish availability, as was flow and turbidity. Few prey sources remained constant throughout the reproductive cycle, and prey and habitat diversity buffered temporal changes in prey availability. We conclude that dams benefit breeding eagles to the extent that they create water temperature discontinuities and additional aquatic habitats, some that support large populations of fish. However, environments modified by dams are not necessarily better for Bald Eagles than those on free-flowing sections of rivers; our data show that Bald Eagle reproduction in the two settings is nearly identical. Key Words: Bald Eagle, Haliaeetus leucocephalus; dams; habitat selection; home range, piscivory; radiotelem- etry; rivers. ECOLOGIA DEL FORRAJEO DE AGUILAS CALVAS NIDIFICANDO EN ARIZONA Resumen. — Estudiamos la ecologia de forrajeo de aguilas calvas {Haliaeetus leucocephalus) nidificando en Arizona durante 1987-89, con enfasis en la influencia de los embalses y la regulacion del flujo de los rios. Examinamos la dieta, modos de forrajeo, seleccion de habitat, abundancia de peces, y factores asociados con la disponibilidad de peces. Tomando como base la biomasa, los restos de presas arrojaron 76% peces, 14% mamiferos, y 10% aves. En los rios, las aguilas ante todo capturaron peces vivos cuando estos desovaban o forrajeaban en aguas someras, mientras que, en los reservorios, la mayoria del pescado fue obtenido como carrona o cuando flotaban moribundos sobre la superficie. Las comunidades de peces difirieron entre los limites de los rios y las caracteristicas ecologicas y de su historia de vida las cuales influyeron en la vulnerabilidad y las diferencias estacionales en la explotacion. La temperatura del agua, un factor primordial determinante de la estructura de la comunidad ictica entre los territorios de las aguilas, fue asociada ademas con las diferencias temporales en disponibilidad de peces, tal como lo fue con el flujo y la turbidez. Pocos recursos de presas permanecieron constantes a traves de todo el ciclo reproductivo, y las presas y diversidad de habitats amortiguaron los cambios temporales en la disponibilidad de presas. Concluimos que los embalses benefician las aguilas que estan reproduciendose en el sentido en que crean discontinuidades en la temperatura y habitats acuaticos adicionales, algunos ^ Present address: The Peregrine Fund, 5568 West Flying Hawk Lane, Boise, ID 83709 U.S.A.; e-mail address: regniarg@aol.com 245 246 Hunt et al. VoL. 36, No. 4 de los cuales soportan grandes poblaciones de peces. Sin embargo, los ambientes modificados por las represas no son necesariaraente mejores para las aguilas pescadoras que aquellos que estan en secciones de libre flujo en los rios; nuestros datos muestran que la reproduccion de las aguilas pescadores en los dos escenarios son cercanamente identicos. [Traduccion de Cesar Marquez] Population persistence in raptors and other ter- ritorial birds depends on an aggregate of breeding locations that contribute to above-replacement-rate reproduction (Hunt and Law 2000). Conserving high-quality sites requires knowledge of key com- ponents, some being physiographic, others de- pending on the ecology, behavior, and life-history characteristics of associated biota. Of particular in- terest are factors relating to food acquisition. Re- productive success requires that breeding pairs have sustained access to prey within efficient com- muting distance (Royama 1970). For species with prolonged breeding cycles (ca. 5 mo for North American eagles), during which numerous phe- nological events transpire, food continuity may in- volve switching from one prey type to another over the course of the nesting season (Jamieson et al. 1982, Edwards 1988). Prey remains collected from nests of inland- breeding Bald Eagles {Haliaeetus leucocephalus) of- ten show dietary diversity. This has been observed not only for populations as a whole, but for indi- vidual nests (Todd et al. 1982, Jackman et al. 1999). Thus, it is tempting to hypothesize that prey variety and, by inference, environmental variation within the foraging range are important components of Bald Eagle territories in some regions (Grubb 1995). From January 1987-June 1989, we investigated the effects of dams and flow regulation on the nest- ing population of Bald Eagles in central Arizona (Driscoll et al. 1999). We obtained information on the spatial and temporal aspects of foraging. We recorded shifts of eagle prey use, foraging behav- ior, and ranging patterns that followed temporal variation in factors influencing prey availability. We found that water temperature and clarity, the struc- ture of various riverine and lacustrine habitats, and life-histories and behavior of prey species were in- terrelated with respect to food availability. Study Area General Description. Our study centered on Bald Ea- gle breeding territories along the Salt and Verde rivers in central Arizona (Fig. 1), a generally open, desert land- scape of the Upper and Lower Sonoran Life-Zones (Lowe 1964, Brown 1998). Eagles nested at elevations ranging from ,329-1719 mask Riparian environments in these re- gions are composed of Sonoran Riparian Deciduous For- est and Woodlands Biome, Sonoran Riparian Scrubland Biome, and the Sonoran Interior Strands Biome. Up- lands in the Lower Sonoran Life Zone are all within Son- oran Desertscrub Biome (Brown 1998). Upper Sonoran Life Zone (Brown 1998) vegetative composition near Bald Eagle breeding territories includes Great Basin Co- nifer Woodland, Interior Chaparral, and Semidesert Grassland Biomes. Mean annual precipitation ranges from 39 cm at higher elevations to 25 cm in the low desert where temperatures may reach 50°C. The central Arizona landscape has been greatly altered by human activity. Cattle grazing, particularly after railway development in the late 1800s, resulted in dramatic ero- sion (Hastings 1959, Hastings and Turner 1965, Hayden 1965). This and woodcutting reduced riparian forests to a scattering of isolated groves and trees. Soil loss drained near-surface aquifers, creating drier soil conditions, and livers became muddy torrents following rains. With the increased need for flood control, water storage, and ir- rigation, hve impoundments were constructed on the Salt River and two on the Verde River during the early 1900s. Riverine environments downstream of the reser- voirs were changed by flow regulation and sediment fil- tration, and upstream by migrations of fish populations such as common carp {Cyprinus carpio) and catfish (Ic- taluridae) out of the reservoirs. Fisheries. Three native species of fish, appropriate for eagle exploitation, remain in substantial numbers within the study area: desert sucker {Catostomus insignis), Sonora sucker (C. clarki), and roundtail chub {Gila robusta) (Minckley 1973). Introduced species of potential impor- tance to eagles in rivers and/ or reservoirs include chan- nel catfish {Ictalurus punctaLus), bullhead {Ameiurus ne- bulosus and A. natalis), flathead catfish {Pylodictis olivaris) , common carp, black crappie {Pomoxis nigromaculatus) , yellow bass (Morone mississippiensis) , largemouth bass {Mi- cropterus salmoides), smallniouth bass {M. dolomieui), blue- gill {Lepomis macwchirus) , green sunfish {L. cyanellus), and walleye (Stizostedion vitreum). Fish distribution in central Arizona, as elsewhere, is strongly influenced by water temperature (Vannote et al 1980, this study). Trout (Salmonidae) inhabit cool head- waters. Suckers, smallmouth hass, and then channel cat- fish increase in abundance as water warms downstream With increasing water temperature, carp and catfish be- come the primary species in size categories suitable for Bald Eagle foraging. When a river enters a reservoir, per- ciforms (hass, perch, and crappie) predominate, al- though carp and catfish cemtribute importantly to overall fish biomass in the reservoir. Water temperature and vol- ume released from the reservoir influence the fish com- munity below the dam. Cool releases from the hypolim- nion of deep, stratified reservoirs favor sucker populations. If the reservoir is shallow or unstable and December 2002 Bald Eagle Foraging Ecology 247 Figure 1. The Salt and Verde river systems of central Arizona. The six names in italics identify breeding territories where Bald Eagle foraging was studied with radio telemetry. fails to maintain a cool hypolimnion, or if releases issue from the epilimnion, warm releases favor carp and catfish in the river reach below the dam. During spring, spawn- ing runs of carp and catfish from downstream reservoirs may augment riverine fish populations. Study Sites. We studied Bald Eagle foraging at six breeding territories (defined here to include the forag- ing range) chosen to compare regulated with unregulat- ed (i.e., free-flowing) environments (Fig. 1). All six ter- ritories have been occupied by eagles since the 1970s (Driscoll et al. 1999). Two breeding territories (Ladders and East Verde) were on free-flowing rivers far from res- ervoirs. Two other pairs (Bardett and Blue Point) occu- pied settings in which all flows were regulated by dam releases, i.e., reservoirs releasing cold water and fed by other dams upstream. The remaining two breeding ter- ritories (Horseshoe and Pinal) included both reservoir systems and the free-flowing rivers that fed them. Methods Telemetry. We used radio-controlled bow nets, power snares, and noosed fish (Jackman et al. 1993, 1994) to capture territorial eagles for radio-tagging. We attached 65-g transmitters with a backpack configuration of teflon ribbons adjoined with cotton string over the carina to permit eventual loss of the radio (Hunt et al. 1992, McClelland et al. 1994). Mercury (activity) switches changed the pulse rate when tilted from near-vertical to horizontal (Kenward 2001). We refined signal interpre- tation on the basis of frequent visual verification. From the early brooding period to fledging, tracking teams collected data in 8-d sessions separated by 6-d pe- riods of no data collection. The objective was to obtain uninterrupted, minute-by-minute records (time lines) of movements and activities of radio-tagged eagles. To over- come the bias associated with observer location, trackers spread out to strategic viewing sites and maintained com- 248 Hunt et al. VoL. 36, No. 4 munication by hand-held radios. Tracking teams of three to five members later conferred to eliminate duplicate data points. To reference eagle locations, habitat vari- ables, and prey, we marked river centerlines and reservoir shorelines depicted on USGS topographic maps to show 1-km and 0.1-km intervals. If we could not identify the eagle’s location to the level of even a 1-km segment, we marked the location within larger zones positioned be- tween familiar landmarks. We measured variables at prey-strike points as soon as possible after a foraging event but without disturbing the eagle. Recorded information included attack method and aquatic habitat type, as follows: “pools” are depres- sions in the streambed, with hydrologic control in the downstream end and low current velocities relative to prevailing streamflow; “runs” are moderately deep, usu- ally narrow channels with relatively fast current, but little or no white water; “riffles” are characterized by shallow, fast-moving water flowing down gradients and over sub- strates usually no larger than small boulders; “pocket wa- ter” usually contains larger boulders, with fast water in- terspersed across the width of the stream among frequent pockets of quiet water; “cascades” are steep gradient white water with less than 10% quiet water (Hunt et al. 1992). Microhabitat features measured at strike points in- cluded water depth, water temperature, and turbidity (Secchi disk). We collected evidence (e.g., scales) of prey species identity, and estimated prey size and whether it was obtained alive or as carrion. Prey Delivery, Observers recorded prey deliveries to the nests from points permitting clear views of nest bowls and at distances of 125-400 m, typically beginning when young were 2-3 wk of age. Prey items were assigned to general taxonomic categories (e.g., class), then to more specific categories (e.g., family, species), where possible. For fish, diagnostic features included fin and scale char- acteristics, body and mouth shapes, jaw configurations, barbel presence, caudal peduncle thickness, and mark- ings. Observers noted their confidence in each identifi- cation; items identified with low confidence were as- signed to a higher taxonomic level. We sometimes confirmed identification with body parts (e.g., scales) col- lected at eagle foraging sites. We did not distinguish be- tween desert and Sonora suckers. We estimated prey size by comparing the item with the length of the eagle’s bill or with objects of known size in the nest. Analysis of Tracking Data. Time line tracking data con- sisted of 22 742 records of the movements and activities ol nine radio-tagged adults in six breeding territories, for a mean of 2327 records per tagged eagle. We recorded the number of minutes an eagle remained at a location and the frequency of visits to each location. The first mea.sure (time) offered a relatively poor estimate of area use compared to relocation frequency. Consider an ex- ample in which an eagle loafed for 146 min at a location 300 m downstream of the nest (in sight of the nest), then flew 2 km upstream where it perched for 5 min at each of three locations, some 200 m apart. At the last of these, the eagle caught a fish, after which it returned to the nest area where it spent 62 rain. Clearly, a time-based assessment awards small significance to the foraging area where the eagle spent only 2% of its time. By contrast, the relocation-based appraisal recognized each of the 0.1- km segments (or larger zones) where the eagle perched. Weighting each location equally as a measure of use was supported by our data: in 80% of observed foraging events, the eagle perched within sight of the foraging location just before the event, and in 70% of cases, the eagle appeared to have seen the prey before leaving the perch. Each perching visit to a O.Tkm segment received a score of one point. If the eagle left a location and en- tered another O.l-km segment along the river, but then returned to the original location, the latter received an- other point. Segments visited repeatedly thereby received the most points. The large number of nest visits over- shadowed other relocation scores; however, eliminating the nest area from the analysis was inappropriate because the nest vicinity was often an important foraging area. Therefore, we used the prey delivery data to estimate the percentage of foraging events occurring in the nest vicin- ity. This became the relocation percentage for the nest area, and percentages for all other areas within the home range were adjusted accordingly so the total was 100%. Collection and Analysis of Prey Remains. Collection. We collected prey remains within and below nests at five of the six breeding territories studied (one nest was on an inaccessible pinnacle) and at 16 other breeding sites in Arizona. Collections occurred during the middle to late part of the brood-rearing period and again after the young had fledged. We attributed variation in the amount of remains present to removal by the adults (ob- served) and to the activities of woodrats (Neotoma spp.) and other scavengers. We collected ca. 2 L of fine nest lining from each nest for scale analysis to detect soft- boned fish (e.g., trout, chub) (Jackman et al. 1999). Analysis. We collected 5—10 individuals of differing sizes of each fish species expected to occur in the eagles’ diet. We weighed and measured each fish, then parboiled it to remove all flesh. We weighed, dried, and labeled all bones for reference. We developed regression equations to relate bone length to total body length for each species (Hunt et al. 1992, Jackman et al. 1999). Because unattached bones were often from the same fish, we used the following procedure to avoid duplication: first, we determined the 95% confidence intervals from each equation to deter- mine the probable range of total length represented by the bones. For a given collection and species, we calcu- lated fish total length from each bone and then sorted all like bones. Bones with the most entries and relatively low confidence intervals were examined first. We grouped pairs of these bones of the same size (<1.0 mm difference or <5 mm for broken bones), e.g., left and right opercula. We marked each pair and the remaining odd entries as individual prey items. We next matched the opcrcula(s) from one fish with other bones whose confidence intervals for total body lengths overlapped with those computed from the opercula. Because differ- ent parts of the same fish had specific proportional re- lationships (e.g,, ±2.0 mm for sucker opercula and clav- icles), we eliminated those entries that were eclipsed by the confidence intervals, a procedure that left the fewest numbers of unmatched parts and, thus, the fewest pos- sible number of individual fish represented. Results un- December 2002 Baed Eagi.e Foraging Ecology 249 derestimated total fish numbers to the extent that matched parts may have been from different fish. We calculated total mass for the selected (nondupli- cate) fish prey items, using length-to-mass equations from this study and from Carlander (1969, 1977) and Becker (1983). We subtracted the mass of bones and scales (from regression equations) plus 5% of total mass (esti- mated unavailable biomass) to calculate the edible bio- mass for each prey item. We identified nonfish remains from museum reference collections and then used stan- dard body mass for each species less 10% for inedible parts. Fish Sampling. Objectives of fish sampling within eagle territories were to identify: (1) relative abundance of prey fish, (2) seasonal changes in their distribution, with em- phasis on availability to eagles (e.g., fish moving into shal- low water), (3) spawning and its effect on fish availability, and (4) effects of water management on prey fish avail- ability. Fish abundance, activity, and distribution. We conducted both roving and fixed-point visual surveys. In roving sur- veys, one or two biologists (depending upon flow) walked along the river bank, noting abundance and activity of fish, aquatic habitat, depth, location (within standard 0.1- km segment), and water temperature. We also observed fish activity and behavior from fixed points. Prior to sur- veying, we compiled information on fish communities in the various river reaches, tributaries, and reservoirs, in- cluding Arizona Game and Fish Department (AGFD) and U.S. Fish and Wildlife Service reports and field data from D. Henrickson (AGFD), M. Jakle (U.S. Bureau of Reclamation), and C. Zeibell (Arizona Cooperative Fish- eries Unit) . To verify our observational data and to determine go- nadal development, we sampled fish in representative habitats with gill nets and throw nets, and occasionally by snorkeling surveys. We removed all collected fish from the system, many of which were used in the prey refer- ence series. We conducted an electrofishing survey at the East Verde territory just after the eagle nesting season. Aquatic habitat surveys. We surveyed aquatic habitat dis- tribution in four of the six eagle territories. We mapped sections of rivers and tributaries within eagle home rang- es into basic habitat units, including pools, runs, riffles, pocket water, and cascades, as defined by Hunt et al. (1992). We differentiated between two types of riffles: channel-riffles, which become runs during moderate flow increase, and bar-riffles which remain as riffles under a variety of flows. Bar-riffles are characterized by the pres- ence of a gravel/ cobble bar oriented diagonally or per- pendicularly to flow. As flow increases, water depth and velocity increase only partially in bar-riffles, whereas the amount of shallow water increases overall due to spread- ing of water across the gravel/ cobble bar. We mapped reservoirs according to distributions of shallow water ar- eas. We obtained river flow and reservoir water surface elevation data from agencies maintaining gaging stations. Carrion and Waterbird Surveys. We conducted peri- odic surveys on rivers and reservoirs to assess temporal availability of carrion. We sampled representative stretch- es of reservoir shoreline where we expected carrion to accumulate, e.g., coves, bends, and especially where riv- ers entered reservoirs. We slowly followed shorelines by boat, identifying and measuring all carrion fish, birds, and mammals encountered. We noted factors contribut- ing to death, e.g., trauma, evidence of spawning, fishing paraphernalia. On rivers, we selected one to three 100- m areas in each territory where carrion was likely to ac- cumulate, and with particular attention to channel bends. We surveyed for waterbirds from one or two points offering wide views per territory, or by making counts while traveling along water bodies (e.g., during carrion surveys), noting the species and numbers pres- ent. Results Diet. We identified 19 species of fish, 26 birds, 16 mammals, and three reptiles from (1) the re- mains of 2601 prey individuals collected from nests, under perches, and after foraging events at 23 breeding territories, and (2) observations of 713 prey items delivered to nests (Table 1). Mean biomass percentages for each class in remains from all sites were 75.5% fish, 14.3% mammals, and 10.2% birds. Four groups accounted for nearly all fish biomass: catfish (mainly channel catfish), suck- er (desert and Sonora suckers), carp, and perci- forms (mainly largemouth bass, black crappie, and yellow bass). Seven taxa exceeded 15% of fish bio- mass at one or more of 23 territories sampled in central Arizona: sucker at 12 territories, carp at 12, channel catfish at 10, largemouth bass at six, flat- head catfish at three, crappie at two, and yellow bass at two. Comparisons of prey remains with prey deliver- ies over similar time frames consistently showed that biomass estimates from remains overrepre- sented mammals and birds over fish, and catfish over suckers and perciforms (Hunt et al. 1992). In three territories where items and deliveries were within comparable time frames, 7 of 56 fish (10.3%) identified in remains were suckers, where- as 124 of 342 (33.2%) fish deliveries were suckers (X^ = 13.1, df — 1, 7* = 0.0003). Remains versus delivery ratios for catfish in these samples were 24: 56 (42.8%) and 56:342 (16.4%) (x^ = 47.6, df = 1, P < 0.0001). An experiment involving a blind sample of 45 fish fed to a captive eagle supported our field data in that soft-boned fishes tended to be underrepresented, e.g., 100% of carp appeared in the remains, 80% of catfish, 60% of the some- what softer-boned suckers, and only 8% of trout (T. Gatz and M. Jakle, unpubl. data). As expected, suckers were the most common prey for pairs nesting on cool, free-flowing reaches nearest the headwaters or at sites offering access to regulated river reaches downstream of hypolim- 250 Hunt et al. VoL. 36, No. 4 Table 1. Prey biomass estimates from prey remains and observed prey deliveries (in italics) for fl Bald Eagle territories on the Salt and Verde rivers where sample sizes exceeded 40 items. Letters in parentheses refer to dam releases from the cool hypolimnion (C) or the warm epilimnion (W). Percent Biomass Breeding Area Setting N Suckers Carp Catfish Perci- forms Mam- mals Birds Other Ladders Free-flowing river 79 8 48 20 1 18 5 0 deliveries 130 45 33 17 0 3 0 2 East Verde 95 5 47 27 2 12 5 2 deliveries 103 14 49 17 8 6 0 6 Redmond 156 5 18 55 1 12 4 5 Final Free-flowing river 107 5 19 47 13 8 8 0 deliveries and reservoir 46 0 10 55 27 0 3 5 Horseshoe 95 1 8 36 31 4 19 1 deliveries 48 0 34 11 40 2 0 13 Blue Point Regulated river (C) 85 10 7 21 22 18 22 0 deliveries and reservoir 152 28 2 9 41 8 8 4 Bartlett 47 55 11 18 7 9 0 0 deliveries 234 66 2 9 15 6 0 2 Orme Regulated river 56 45 4 3 2 34 12 0 Et. McDowell 62 66 5 5 1 18 5 0 Chff Regulated river (W) and reservoir 45 0 46 18 27 5 4 0 “76” Creek 59 5 38 2 1 41 13 0 netic dam releases (Table 1). Perciforms were tak- en mainly in the reservoirs. Eagles obtained carp primarily in warm, free-flowing reaches upstream of reservoirs and in a river fed by epilimnetic re- leases (Table 1). Catfish (channel and flathead) were widely utilized, the highest numbers taken from free-flowing river reaches and in a reservoir (Alamo) with only seasonal inflow (Haywood and Ohmart 1986). Estimates of mammal biomass from remains ex- ceeded 25% at six breeding territories. Most fre- quently identified were black-tailed jackrabbit {Le- pus californicus) and cottontail rabbit (Sylvilagus audubonii). We recorded only 32 mammals among 713 prey items observed delivered to nests (4.5% of items) despite a more substantial representation m prey remains from those nests (18.3%). We at- tribute this disparity to the greater use of mammals early in the breeding season and to biases associ- ated with bone persistence. Waterbirds were more important to Bald Eagles in early winter than during the nesting season, par- ticularly at territories containing reservoirs. In win- ter, the percentages of birds observed taken (as compared with fish) were 50% in December, 56% in January, and 13% in February, as compared with 5% in March, 1 % in April, and 0% in May. The most commonly recorded birds taken among 30 identi- fied were American Coots {Fulica americana, N— 15) and Eared Grebes {Podiceps nigricollis, N = 8). Conditions of Prey Acquisition. Eagles took some fish species only in riverine conditions, oth- ers only from reservoirs, and some from both. Of the seven important fish taxa recorded during prey delivery observation, numerical ratios of their ori- gin in rivers versus reservoirs at four territories containing both environments were as follows (riv- er : reservoir) : yellow bass (0:31), crappie spp. (0; 40), largemouth bass (2:22), flathead catfish (3:5), channel catfish (4:17), sucker spp. (105:0), and carp (3:30). In reservoirs, eagles obtained most fish as carrion (or moribund), i.e., 66% of 125 fish of known status (excluding piracies) were obtained from reservoirs as carrion, whereas 12% of 201 fish from rivers were carrion. Suckers. At least 83% {N = 114) of suckers were alive when taken, 5% were pirated, 3% were car- rion, and 9% were of unknown status. Bald Eagles December 2002 Bald Eagle Foraging Ecology 251 caught them mainly in riffles while they spawned or foraged. Of 64 depth measurements at strike points for live suckers, 80% were in water <30 cm in depth; mean depth at strike points of these 51 shallow water captures was 16.4 cm (SD ± 7.3). Carp. Eagles obtained carp from both rivers and reservoirs. In rivers, eagles caught them in the shal- lows of runs and riffles, 17 of 20 strikes in water less than 36 cm deep. We were unable to deter- mine under what conditions carp were captured in reservoirs. Catfish. We estimate from prey collections that channel catfish contributed almost three times the biomass as flathead catfish (301 182 g versus 92 304 g, respectively). Excluding piracies, we observed nesting Bald Eagles taking channel catfish on 51 occasions: 75% on reservoirs and 25% on rivers. On reservoirs, 81% of 26 catfish of known status were obtained as carrion, whereas on rivers, 27% of 11 were carrion. Although the sample of con- ditions at strike points for live channel catfish in rivers was small, it differed from those noted for other species. Eagles captured five in pocket water, three in runs, and none from riffles. The mean of six depth measurements was 58 cm (SD ± 28.1). We occasionally observed catfish swimming near the surface in riverine pools (Van Daele and Van Daele 1982), and “blooms” of carrion channel cat- fish (ca. 20 cm long) appeared in late spring at two reservoirs (Horseshoe and Roosevelt) . Perciforms. At four breeding territories contain- ing reservoirs (Bartlett, Saguaro, Horseshoe, and Roosevelt), we recorded delivery of 61 largemouth bass, 51 black crappie, 41 yellow bass, and 14 oth- ers (mainly sunfish). Eagles obtained these perci- forms mainly as carrion from the reservoir or as they lay moribund at the surface. Of 76 dead or moribund perciforms found in carrion surveys on these same reservoirs, 29 were yellow bass, 15 large- mouth bass, 13 black crappie, eight bluegills, four smallmouth bass, four green sunfish, and three walleye. The yellow bass and black crappie were apparent victims of spawning stress, whereas many of the largemouth fatalities were angler-related. Birds and Mammals. Coots were attacked when they foraged in the grassy shallows of reservoirs. Eagles caught grebes and waterfowl either by stooping repeatedly at groups in open water or by approaching low (<1 m) over the surface, snatch- ing the prey in passing. We observed no attempts at live mammals. Habitat Use. We observed eagles foraging in rif- fles disproportionate to their occurrence along ihe river. At Ladders, 54% of 58 prey captures were in riffles, the latter composing only 5% of riverine habitat within the 22-km foraging range of the ea- gles. At East Verde, 46% of 61 observed foraging events were in riffles, compared to 15% availability within 19 river km. Along 2-3 kilometers of river at Bartlett, 73% of 119 observed foraging attempts were in riffles compared with 8% availability. At Blue Point, 32% of 28 attempts were in riffles com- pared with 4% availability. The frequency of riffle use in part reflected the high proportion of sucker captures in those territories. As noted, eagles cap- tured suckers mainly in riffles, carp most often in runs, and channel catfish in pocket water and runs. In 162 measurements of turbidity at strike points for live fish, 136 (84%) were in water that was “clear to the bottom.” When rivers became turbid during prolonged periods of snowmelt, eagles tended to forage elsewhere, such as in clear trib- utaries. At the four breeding areas containing both riv- erine and reservoir environments, eagles mostly foraged from reservoirs (>50% of locations). At two territories where eagles nested on the river 3.6 km and 7.0 km from reservoirs, 51% and 61% of total relocations were on the reservoirs, respective- ly. For a radio-tagged pair whose nest was situated where a river entered a reservoir, 85% and 86% of relocations were on the reservoir. At a territory where the nest was about 2 km from both river and reservoir, 59% of relocations were on the reservoir. Among a biomass total of 113.5 kg of delivered prey items recorded at these four territories, 28.4%, 65.9%, 93.9%, and 48.4% were obtained from the reservoir. Of 641 forage attempts record- ed at these territories, 386 (60%) were on reser- voirs. Foraging Range. Free-flowing river. We compared eagle foraging ranges at two territories (Ladders and East Verde) situated on the free-flowing Verde River far upstream of the dams and reservoirs (Fig. 1). The nests of both pairs were on cliffs overlook- ing the river: the two nests at Ladders were directly over bar-riffles, but the East Verde nest was about 1 km from the nearest bar-riffle. At both territo- ries, the null hypothesis of random selection by the eagles of 1-km segments containing bar-riffles was rejected (Chi-square tests, P < 0.005). At East Verde, 72% of mainstem relocation points were within seven 1-km segments, in the aggregate con- taining 100% of the bar-riffles within the 19 km 252 Hunt et ajl. VoL. 36, No. 4 *34 37« ^36 137 153 1' UO U’ (0 14 |„ U J2 10 « S 8 o c 6 d) g 4 0. Fast Verde Male 494 relocations 124 125 ‘>7 Ti ' - Bar Rime - I V'erde River ^ 'tritaiHries ' ''Q" _ tribSy Vei 4 e River Figure 2. Foraging range of the radio-tagged breeding male Bald Eagle at the East Verde territory. Relocation percentages in the nest vicinity were adjusted according to the proportion of observed prey deliveries from those 1- km segments (see Methods) . Open bars quantify cases in which trackers could not precisely locate the eagles; dotted lines extending laterally from open bars indicate zones of eagle occupancy for the imprecise locations (tributary relocation percentages not shown). foraging range (123-141 km), but containing only 21% of the available channel-riffle habitat (Fig. 2). At Ladders, 54% of visits within the 22 km range were within the six 1-km segments containing bar- riffles. There were clear, seasonal shifts in prey and hab- itat use (Fig. 3). For example, in March, when the Verde River was turbid, 17 of 18 prey items record- ed at the East Verde territory were mammals. In April, the East Verde male traveled up two relative- ly clear tributaries to forage on spawning suckers. By early May, he was foraging almost exclusively in the mainstem Verde River, taking carp and catfish. His use of river sections downstream of the nest peaked during the mid-point of the brood cycle, then shifted dramatically to the area upstream of the nest containing a large bar-riffle. Regulated river and reservoir. The home ranges of the Bartlett and Blue Point pairs (29 and 26 river- km, respectively) both contained a deep-release (cool) regulated river section below a reservoir fed by a regulated reach. At both breeding territories, 90 (A C 80 70 8 60 O 3050 m) mountain ranges border the valley on east and west, merging at the northern end. The Rio Grande and Conejos rivers flow through the SLV and numerous natural warm springs and wetlands that seldom freeze have made the SLV attractive to waterfowl and Bald Eagles, probably for millennia. Water develop- ments and agriculture in the 20th century have probably improved the attractiveness. Between 7 January-18 March 1980 and 1981, 15 adult Bald Eagles were captured and radio-tagged in the SLV All were captured by a modified “Lockhart” method (Miner 1975) with and without live Bald Eagle and Gold- en Eagle {Aquila chrysaetos) lure birds. Trap sites were chosen on the basis of frequency and duration of the presence of two adult Bald Eagles of distinctly dissimilar size, presumably mated, within 1.6 km. In 1981, I esti- mated the SIV winter population at 170 Bald Eagles (Harmata 1984). Gender of radio-tagged eagles was assigned by methods presented by Garcelon et al. (1985). Three were con- firmed by behavior during copulation. All Bald Eagles were radio-tagged with two-stage radio transmitters mounted proximo-ventrally on the tail. Transmitter fre- quencies were between 148.500 and 148.950 mHz. Unit life expectancy was ^5 mo. Transmitter, antenna, and mounting tab weighed 50-57 g. Telemetry receiving equipment included fixed channel and programmable receivers. Mated status of radio-tagged Bald Eagles was deter- mined by frequency and duration of time spent in the presence of another adult eagle of distinctly dissimilar size, observed copulation {N = 3) in the SLV, or associ- ation with a nest site on the breeding grounds {N = 2). Eagles were considered unmated if diurnal movements in the SLV were clearly independent of other eagles and they were not observed to roost away from communal roosts with just one other eagle, as mated birds often did Climatological data associated with days that Bald Ea- gles left the SLV were analyzed by stepwise discriminate analysis (Dixon 1981) to investigate meteorological con- ditions associated with initiation of vernal migration. Cli- matological data recorded near the geographical center of the SLV (Alamosa, Colorado) were obtained from Na- tional Oceanographic and Atmospheric Administration, Asheville, North Carolina Monthly Summary Sheets. Data for days that eagles left the SLV were compared to data for days randomly selected between 1 January-15 April 1980 and 1981 that they did not. Variables selected for comparisons among days were maximum, mean, and range of temperature, percent of clear sky and mean wind speed. Migrating eagles were followed primarily from a single 4X4 vehicle with two human trackers and a dog. An omnidirectional antenna and two element “H” yagi re- ceiving antenna were mounted on the roof of the chase vehicle. The yagi was attached to a 360° traversing mount, allowing for directional tracking while the vehicle was moving. Manpower and logistic limitations plus variability in departure dates, routes, and travel speeds of Bald Ea- gles prevented ground tracking of more than one eagle at a time. One tracker drove while the other operated the receiving equipment. Both shared navigational du- ties. Due to often high chase speeds (up to 150 km/hr) and off-road “adventures,” visual contact with migrating eagles could not be maintained, so migration behavior often could not be recorded continuously, accurately, or safely. Route, direction, and speed of the chase vehicle, therefore, often was selected primarily to maintain max- imum audio signal strength. When contact with a mi- grating eagle was lost, an aerial search was implemented using local air services, A two or three element yagi an- 258 Harmata VoL. 36, No. 4 FEMALES O O O • • MALES O o -40 -35 -30 -25 -20 -15 -10 -5 0 5 10 15 DAYS FROM VERNAL EQUINOX Figure 1. Days from vernal equinox that radio-tagged Bald Eagles departed {N= 12) their San Luis Valley, Colorado, winter ranges to initiate vernal migration, 1978 {N = 1), 1980 (N = 4), 1981 {N = 7). Open circles indicate eagles determined to be unmated, closed circles indicate eagles that were mated (see text), and square indicates eagle of undetermined mated status. tenna was taped to the wing strut (high wing) or step (low wing) of light aircraft and the area surveyed in tran- sect style with intermittent, “lazy” circles at high altitude. Frequencies were scanned to search for the target eagle and other potential migrating birds. When contact was reestablished and eagle’s status discerned as stationary, vehicle tracking resumed. Flight altitudes were estimated based on eagle’s position relative to the search aircraft and trees, geography, and structures such as buildings and radio towers while ground tracking. Eagles were monitored in their summering grounds with Beaver air- craft, snow machines, and snow shoes. Results and Discussion Initiation of Migration. Over half (58%) of ra- dio-tagged Bald Eagles with known departure dates (N = 12), initiated vernal migration within one week after the vernal equinox (Fig. 1). In 1980, radio-tagged eagles {N — 4) all left the wintering grounds within a 15-d span in March (Table 1). Range of known departure dates {N — 8) spanned 50 d in 1981, with the first confirmed departure of Table 1. Residency and mated status of radio-tagged adult Bald Eagles with known departure dates from winter ranges in the San Luis Valley (SLV), Colorado. Eagle Designation Sex Captured Date Departed SLV Mated Status'* Minimum Residency in SLV (davs) 378 F 21 Jan. 1978 28 Feb. 1978 U 39 180 F 18 Jan. 1980 21 Mar. 1980 M 63 280 M 23 Jan. 1980 24 Mar. 1980'^ M 61 380 M 23 Jan. 1980 21 Mar. 1980 M 58 480 M 27 Jan. 1980 8 Mar. 1980 U 41 181 F 7 Jan. 1981 11 Feb. 1981 u 36 281 M 9 Jan. 1981 20 Feb. 1981 u 43 381 F 9 Jan. 1981 27 Mar. 1981 M 78 481 M 11 Jan. 1981 27 Mar. 1981 M 76 581 F 15 Jan. 1981 25 Mar. 1981 U 70 881 M 15 Mar. 1981 21 Mar. 1981 ?*' 7 981 F 18 Mar. 1981 1 Apr. 1981 M 15 M = mated, U = unmated (see text). First departed on 22 March 1980, but eagle encountered a winter storm and returned to winter range. Undetermined. December 2002 Bald Eagle Vernal Migration 259 Table 2. Mean climatological values (±99% confidence interval) which discriminated {P ^ 0.01) between days radio-tagged Bald Eagles initiated northward migration from their San Luis Valley, Colorado wintering area (De- part) and days they did not (Remain). Tempera- TURE Percent Wind Range Sky Speed (°C) Clear (km/hr) Depart {N = 11) 24 (5) 50 (26) 22 (5) Remain {N = 21)=* 19 (4) 56 (20) 16 (5) Randomly selected. a radio-tagged eagle occurring in mid-February and the last in early April (Table 1). Number of days from the equinox that eagles departed was not different between genders (Mann-Whitney U = 16.50, P — 0.81). However, mated eagles departed the SLV closer to the equinox and later than un- mated eagles (Mann-Whitney U = 3.00, P = 0.03; Fig. 1). Three climatological variables discriminated be- tween days that radio-tagged Bald Eagles left the SLV on migration and those they did not (Table 2). Eagles initiated migration on days with much larger range of temperatures, more clouds, and with higher winds than other days over the winter- spring period. Migration tended to be initiated about 5 hr after sunrise regardless of the immedi- ate meteorological condition. Photoperiodism is considered to be the ultimate stimulus for the onset of migration in birds (King and Earner 1963). All radio-tagged Bald Eagles de- termined to be mated left the SLV within 12 days of the vernal equinox, indicating breeding adults may be sensitive to equal periods of light and dark. Unmated, nonbreeding, or immature eagles may be equally sensitive but may not be driven by pres- sure to procreate. In fact, first observations of sub- adult eagles in northern ranges both years were not until at least two weeks subsequent to the ar- rival of the first radio-tagged adults. Migrational movement also paralleled the northward move- ment of the 2°C isotherm (Lincoln 1979) both temporally and spatially, hinting that thermal cues may also be involved in the initiation of migration. Once the urge is kindled, proximate affectors of Bald Eagle migration appear to be coincident with incoming low pressure systems, associated wide range of temperatures, cyclonic air flows, and southerly winds; similar to conditions noted by Bagg et al. (1950) for other birds. Although Bald Eagles are apparently sensitive to local conditions when migration is initiated, they appear to be cognizant of little beyond their im- mediate environs. Eagle 280 departed the SLV on 22 March 1980 and traveled 145 km north before being stopped by severe snow squalls. He spent the night on a mountain pass and as the storm per- sisted to the north the next day, he returned to his winter range in the valley. He spent the remainder of that day and part of the next in close association with a female (distinctly larger eagle) and was ob- served to copulate during this period. He initiated migration again on 24 March, leaving the SLV for the season. Migration Routes, Destination, and Navigation. In 1980, a relatively narrow migration corridor through Colorado, Wyoming, and Montana was used by three radio-tagged eagles (Fig. 2). Eagle 380 became sedentary in north-central Saskatche- wan after 15 d of migration, six of which were spent sitting out bad weather. After four days of almost continual solitary soaring, he was found perched close to a larger eagle where he remained for several hours. The next day the pair began con- struction of a new nest on Chachukew Lake, north of Pelican Narrows, Saskatchewan. They eventually fledged one young in August 1980. In 1981, eagle 281 left the SLV on 20 February, and was followed for two days over 302 km through the mountains of western Colorado (Fig. 3) . Direc- tion was primarily northwest and 281 was the only eagle tracked that crossed the Continental Divide. Subsequent tracking of 281 was interrupted be- cause a radio of the same frequency and pulse rate as 281 was tracked for a day before I discovered it was a transmitter on a collar of a bighorn sheep ( Ovis canadensis) . Eagle 981 left the SLV on 1 April and was fol- lowed to about 65 km north of Casper, Wyoming, where she was lost in a snow storm because roads became impassable. Eagle 481 was detected ap- proaching Fort Peck Reservoir in east-central Mon- tana on 5 April. After leaving Fort Peck Reservoir, 481 flew due north about 35 km then gradually shifted to a northwest course. This route took him directly to the Missouri Coteau, a ridge line run- ning longitudinally for over 150 km in southwest- ern Saskatchewan. He continued north along the Coteau for about 100 km. Approaching the Cana- dian Shield, 481 made a sudden change in course 260 Harmata VoL. 36, No. 4 Figure 2. Partial migration routes of radio-tagged adult Bald Eagles from winter ranges in the San Luis Valley, Colorado in 1978 and 1980. Numbers indicate eagle des- ignation and year (Table 1). Summering area of eagle 380 is indicated by 0 (confirmed nest location). from north to northeast, paralleling the direction of many elongated lakes and rivers in the Canadian Shield country. Eagle 481 was subsequently fol- lowed to Reindeer Lake, Saskatchewan, arriving on 10 April. On 22 April, 481 was found by air asso- ciated with a nest site on Ramuchawie Lake in west- ern Manitoba near the Saskatchewan border (Fig. 3). Aerial surveys covering ca. 210000 km^ of north- ern Saskatchewan were conducted to detect signals of other SLV Bald Eagles in mid- to late April in both 1980 and 1981. During aerial surveys, eagle 181 was located on the Pagato River, south of Rein- deer Lake, Saskatchewan, and eagle 581 was de- tected briefly in the area of Trade Lake, south of the Churchill River (Fig. 3). The signal was weak and intermittent, which indicated either a radio Figure 3. Partial migration routes of radio-tagged adult Bald Eagles from winter ranges in the San Luis Valley (SLV), Colorado in 1981. Numbers indicate eagle desig- nation (Table 1). Summering areas are indicated by 0 (nest locations). Respective 1981 and 1982 recovery lo- cations of Bald Eagles banded in the SLV (January 1977) are indicated by (x). Tbe October 1981 recovery location of a Bald Eagle radio-tagged in the SLV in March 1981 is indicated by 781. malfunction, a dead eagle, or a bird in incubation posture. Lack of contact with other SLV eagles may have been a function of transmitter failure (one was known to fail in the SLV) , premature shedding of the transmitter (one transmitter was found still attached to broken tail feathers below a perch tree in the SLV) , mortality, incomplete survey coverage, or summering areas were located outside of north- ern Saskatchewan. However, the fact that 31% of potentially-detectable eagles (N= 13) were located within an area roughly the size of Yellowstone Na- tional Park (ca. 9300 km^) suggests that most adult Bald Eagles wintering in the SLV originated from this area of Canada. December 2002 Bald Eagle Vernal Migration 261 Mean distance (as the eagle flies) from the SLV wintering area to breeding or summer areas of Bald Eagles was 2019 km {N - 4, SD = 50). All eagles located on their summer ranges were within 102 km of each other in northeastern Saskatche- wan or northwestern Manitoba (Figs. 2 and 3). Du- ration of travel between winter and summer range for SLV Bald Eagles with known departure and ar- rival dates {N = 2) was 15 d. Although SLV Bald Eagles traveled about Vs the distance to their sum- mer ranges compared to one Bald Eagle tracked by satellite from Arizona to the Northwest Terri- tories (3032 km, Grubb et al. 1994), they complet- ed the journey in only about 40% of the time. However, the Arizona eagle was a subadult (3rd yr) and presumably was not driven by the impetus to nest, as were adult SLV eagles tracked. Geography, celestial cues, and weather may all play a role during migration of Bald Eagles. Prom- inent physiographic features such as deep canyons, rivers, and north-south oriented topography seemed to assist in visual navigation during flight (Griffin 1943). These features could have been im- printed in the memory of eagles during their ini- tial migrations and experience dictated direction during subsequent flights (kinesis theory; Mat- thews 1963). Imprinting of migration routes from wintering areas, which compliment survival during the first year, would be more adaptive than im- printing during first southward migration. Eagles did not migrate on days of total overcast, a phenomenon also noted by Gerrard and Gerrard (1982). The altitude of the cloud layer may have an effect, but overcast layers over 90 m were not experienced during tracking. Sun compass orien- tation (Kramer 1952, 1957) with time compensa- tion, commonly referred to as sun-azimuth orien- tation (Welty 1982) may, therefore, be important for orientation of adult Bald Eagles during migra- tion. True navigation (selection of a compass direc- tion toward a known goal in unfamiliar territory; Able 1980) may also be a component of Bald Eagle migration. Migrating eagles appear to avoid strong winds during migration because strong winds ap- parently influence direction of flight. Eagles did not move, except locally, during days when winds in excess of 35 km/hr occurred prior to 0900 H. East winds of 60—80 km/hr began about two hours after initiation of eagle 380’s flight on one migra- tion day and his flight path deviated well west of direct line to the eventual goal. Flight direction for the remainder of migration clearly compensated for the one day blown off course (Fig. 2). Unless tbe eagle had been exposed to the area on previ- ous migrations, true navigation is indicated. How- ever, it is not unreasonable to assume that an adult eagle over five years old, may indeed be familiar with a great portion of western North America as a result of vagaries of previous migrations. Regard- less, redundancy in navigational systems has been illustrated for homing pigeons (Columba livia) (Able 1980) and in all probability, several backup navigational systems are available to Bald Eagles, especially experienced adults. Migration Behavior. Mated adult Bald Eagles ap- parently migrated alone in the spring. Eagle 380 was seen roosting solitarily at all but one stopover location (six other eagles on the Yellowstone River near Hysham, Montana). Eagle 481 did not mi- grate with a mate, but often moved northward with other adult eagles. He roosted with other eagles three times, but two sites obviously were not com- munal or traditional because of lack of similarity to typical Bald Eagle roost sites (Kiester and An- thony 1983). Paired roosting was probably a result of facilitatory behavior influenced by poor weather and lack of daylight remaining. Subsequent obser- vations of migration flight confirmed he moved alone. Once in the boreal forest of the Canadian Shield country, local eagles appeared to “meet” him in flight and escort him through their terri- tories, but no overt agonistic encounters were ob- served. Occasional associations with other eagles appeared incidental for all radio-tagged eagles and were of short duration. Solitary migration behavior in established pairs would reduce the possibility of both members being lost in a local catastrophe. Daily migration flights consistently began be- tween 1015-1045 H MST and ended between 1715-1745 H. Mean daily movement was 180 km, but ranged from 144-435 km {N = 5) in 1980 and 33-248 km (A = 5) in 1981. Speed of migration flights recorded averaged 50 km/hr (A = 7, 22- 144 km/hr in 1980; A = 9, 20-105 km/hr in 1981) . Altitude of flights recorded ranged from 30-4572 m AGL, but most often was between 1500-3050 m. Total distances, speed, and daily duration of mi- gratory flights indicate that under optimal weather conditions. Bald Eagles can reach their breeding grounds within 6 d after leaving the SLV. Penny- cuick (1975) indicated a 2000 km migration for a bird the size of a Bald Eagle would be near maxi- mum attainable without eating, assuming a 25% 262 Harmata VoL. 36, No. 4 mass loss. During this study, no radio-tagged eagles were observed feeding during migration and mean migration distance was 2020 km. Captive eagles commonly fast more than two weeks with no ap- parent deleterious effects (Brown and Amadon 1968, pers. observ.) and wild raptors can lose up to 30% of body mass without problems (Newton 1979). These compensatory capacities undoubted- ly allowed adult Bald Eagles to reach their breed- ing grounds with sufficient energy reserve for breeding. All radio-tagged eagles arrived in their summer range at a time when lakes and most stretches of rivers were still frozen. The only areas of open wa- ter were rapids or narrows on rivers or between lakes. Radio-tagged eagles spent most of their time there, presumably foraging for fish. Other eagles observed during aerial surveys were associated with ubiquitous holes in lake ice and viscera piles of fish left by native commercial fishing operations. For- aging eagles were also seen on or near caribou {Rangifer tarandus) and moose {Alces alces) carcass- es killed by natives or wolves ( Canis lupus) . A few eagles were seen with snowshoe hare {Lepus amer- tcanus) remains. Regional Relationships. McClelland et al. (1994) noted that Bald Eagles radio-tagged in Glacier Na- tional Park, Montana, in autumn wintered west of the Continental Divide and summered in the MacKenzie River watershed of northern Alberta, northwest Saskatchewan, and Northwest Territo- ries. The summer range of a Bald Eagle that win- tered in Arizona (west of the Divide) also was in the MacKenzie River watershed (Grubb et al. 1994). McClelland et al. (1994) suggested that win- tering areas may be related to the watershed of origin and Bald Eagles be managed by application of a “Migration Flyway Concept.” Adult Bald Ea- gles radio-tagged and banded in the SLV wintering area (east of the Divide) were tracked to breeding areas in Saskatchewan and Manitoba, all in the Churchill River watershed. Jenkins et al. (1982) fol- lowed two adult Bald Eagles radio-tagged in Wyo- ming during winter. One trapped on the west side of the Continental Divide was followed to the MacKenzie River watershed, while one trapped on the east side of the Divide was followed to the Churchill River watershed, similar to those from the SLV. These data suggest a “Churchill-East Slope” Migration Flyway exists, distinct from the “Mackenzie-Intermountain” Flyway proposed by McClelland et al. (1994). Stopover areas used during vernal migration of SLV Bald Eagles were generally widely distributed. A tree of adequate size, secure from human dis- turbance in any type habitat, was all that seemed necessary for roosting. In late March 1982, Swen- son (1983) counted 232 Bald Eagles along the Yel- lowstone River between the mouth of the Bighorn River and Miles City, Montana. A site where eagle 380 roosted is in the middle of this stretch and within an area where the river seldom freezes, con- tained the most highly braided portion of channel, most heavily wooded islands, and highest Canada goose {Branta canadensis) populations of three sec- tions of river studied by Hinz (1974). Use of this section of the Yellowstone River by adults may be dictated primarily by tradition and availability of water, because migrating adults were not known to feed during this study. Areas in eastern Montana may be equally im- portant to as many or more migrating eagles as the more highly-publicized areas, where ephemeral concentrations of eagles occur in western Mon- tana. Leighton et al. (1979) estimated a population of 14 000 Bald Eagles in Saskatchewan. Some ea- gles from north-central Saskatchewan were cap- tured in autumn at Hauser Lake (Restani et al. 2000) and Glacier National Park (McClelland et al. 1982), while others passed through eastern Mon- tana during migration (Gerrard et al. 1978, Har- mata et al. 1985). Both Hauser Lake and Glacier National Park Bald Eagle concentrations are now defunct due to a collapsed, exotic food base (ko- kanee salmon, Oncorhynchus nerko), but eastern Montana habitats still support large numbers of na- tive prey (lagomorphs, ungulates, waterfowl). How- ever, lack of a concentrated food base, diffusion of roost sites, solitary habits of migrating eagles, plus dispersion of departure dates from winter (this study) and summer ranges (Harmata et al. 1985), prohibit any accurate estimate of numbers of Bald Eagles passing through eastern Montana. Relatively low numbers of eagles present at any particular time at some stopover areas in eastern Montana may belie the true importance of these areas to migrating eagles. Turnover of individuals appeared to be daily, over months. Therefore, western prai- rie states may provide important migratory habitat for a large proportion of the continental popula- tion of Bald Eagles over long periods. Acknowledgments J. Stoddart initially suggested Bald Eagle research in the SLV. Dale Stahlecker served as co-investigator during December 2002 Bai.d EACiLE Vernai, Migration 263 preliminary marking studies. P. Harmata (then age 5) helped track the first migrating Bald Eagle in 1978. G. Montopoli was instrumental for successful completion of the first year’s migration tracking. M. Lockhart provided details of eagle capture techniques and field assistance. S. Werner and E. Spettigue participated in winter and migration tracking. L. Stevenson of Pelican Narrows, Sas- katchewan, donated flight expertise and air time. Finan- cial assistance was provided by R. Koteen; Fred Jense, Dept, of Veteran’s Affairs; J. Lincer and W. Clark of the National Wildlife Federation; R. Plunkett of the National Audubon Society; D. Flath of Montana Fish, Wildlife, and Parks; 1.. Jahn of the Wildlife Management Institute and American Petroleum Institute; T. Ingram of Eagle Valley Environmentalists; S. Rainey and C. Merrit of American Wilderness Alliance; M. Malone, Montana State Univer- sity (MSU) Research Creativity Program; R. Eng, MSU Agricultural Experiment Station; and R. Moore, MSU Bi- ology Department. T. Crubb, M. Ratine, M. Restani, and D. Stahlecker made helpful comments on earlier drafts of the manuscript. “Sarge” was the dog, a great compan- ion and a loyal friend. Literature Cited Able, K.P. 1980. Mechanisms of orientation, navigation and homing. Pages 284-373 in S.A. Gauthreaux, Jr. [Ed.], Animal migration, orientation and navigation. Academic Press, New York, NY U.S.A. Bagg, A.M., W.W.EI. Gunn, D.S. Miller, J.T. Nichols, W. Smith, and F.P. Wolforth. 1950. Barometric pres- sure — patterns and spring bird migration. Wilson Bull. 62:5-19. Brodeur, S., R. Deecarie, D.M. Bird, and M. Fuller. 1996. Complete migration cycle of Golden Eagles breeding in northern Quebec. Contfor 98:29.3-299. Brown, L.H. and D. Amadon. 1968. Eagles, hawks, and falcons of the world. Vol. 2. McGraw-Hill Book Co., New York, NY U.S.A. Craig, J.R. 1981. Bald and Golden Eagle winter popula- tion surveys. Wildlife Research Report: Part I. Federal Aid Job Final Report. Colorado Division of Wildlife, Denver, CO U.S.A. Dixon, W.J. 1981. BMDP. Statistical Software. Univ. of California Press, Berkeley, CA U.S.A. Garcei.on, D.K., M.S. Martell, P.T. Redig, and L.C. Buoen. 1985. Morphometric, karyotypic, and laparo- scopic techniques for determining sex in Bald Eagles. J. Wildl. Manage. 49:595-599. Gerrard, J.M., D.W.A. Whitfield, P. Gerrard, P.N. Ger- RARD, AND WJ. Maher. 1978. Migratory movements and plumage of subadult Saskatchewan Bald Eagles. Can. Field-Nat. 92:37.5-382. and P.N. Gerrard. 1982. The spring migration of Bald Eagles in the vicinity of Saskatoon. Blue Jay 40: 56-60. Griffin, D.R. 1943. Homing experiments with Herring Gulls and Common Terns. Bird-banding 14:7-23. Grubb, T.G., W.W. Bowerman, and P.W. Howey. 1994. Tracking local and seasonal movements of wintering Bald Eagles Haliaeelus leucocephalus from Arizona and Michigan with satellite telemetry. Pages 347-358 m B.-U. Meyburg and R.D. Chancellor [Eds.], Raptor conservation today. World Working Group on Birds of Prey. Pica Press, London, U.K. Harmata, A.R. 1984. Bald Eagles of the San laris Valley, Colorado; their winter ecology and spring migration. Ph.D. dissertation, Montana State University, Boze- man, MT U.S.A. , J.E. Toepfer, and J.M. Gerrard. 1985. Fall mi- gration of Bald Eagles produced in northern Sas- katchewan. Blue Jay 43:56-62. AND D.W. Stahlecker. 1993. Fidelity of Bald Eagles to wintering grounds in southern Colorado and northern New Mexico. /. Field Ornithol. 64:129- 134. Hinz, T.C. 1974. Seasonal activity, numbers and distri- bution of Canada Geese {Branta canadensis) in the lower Yellowstone Valley, Montana. M.S. thesis, Mon- tana State University, Bozeman, MT U.S.A. Jenkins, M.A., T.P. McEneaney, I.. Hanebury, and J.R Squires. 1982. Bald Eagle (Ffaliaeetus leucocephalus) es- sential habitat on and near Bureau of Land Manage- ment lands in Wyoming. Draft final report — FY 1981 & 1982. Wintering Bald Eagles. U.S. Dept, of Interior, Fish & Wildl. Service, Denver, CO U.S.A. Kiesier, G.P., Jr. and R.G. Anthony. 1983. Characteris- tics of Bald Eagle communal roosts in the Klamath Basin, Oregon and California. J. Wildl. Manage. 47 1072-1079. King, J.R. and D.S. Earner. 1963. The relationship of fat deposition to Zugunruhe. Condor 65:200—223. Kramer, G. 1952. Experiments on bird orientation. Ibis 94:265-285. . 1957. Experiments on bird orientation and their interpretation. Ibis 99:196-227. Leighton, F.A., J.M. Gerrard, P. Gerrard, D.W.A. Whit- field, and W.J. Maher. 1979. An aerial census of Bald Eagles in Saskatchewan./. Wildl. Manage. 43:61-69. Linc:oln, F.C. 1979. Migration of birds. In S.R. Peterson and P.A. Anastasi [E.DS.], Fish & Wildlife Service Cir- cular 1 6. U.S. Dept, of Interior, Washington, DC U.S.A. Matthews, G.V.T. 1963. The orientation of pigeons as affected by the learning of landmarks and by the dis- tance of displacement. Anim. Behav. 11:310—317. McCleeland, B.R., L.S. Young, D.S. Shea, P.T. Mc- Cleliand, H.L. Ai.len, and E.B. Spettigue. 1982. The Bald Eagle concentration in Glacier National Park, Montana: origin, growth and variation in numbers. Living Bird 19:133—15.5. , L.S. Young, P.T. McClelland, J.G. Crenshaw, H.I.. Allen, and D.S. Shea. 1994. Migration ecology of Bald Eagles from autumn concentrations in Glacier National Park, Montana. Wildl. Monogr. 125. Meyburg, B-U., D.H. Ellis, C. Meyburg, J.M, Mendel- son, and W. Scheller. 2001 . Satellite tracking of two 264 Harmata VoL. 36, No. 4 Lesser Spotted Eagles, Aquila pomarina, migrating from Namibia. Ostrich 72:35-40. Miner, N.R. 1975. Montana Golden Eagle removal and translocation project. Pages 155-162 in Proc. of the Second Great Plains Wildlife Damage Control Work- shop. U.S. Fish & Wildlife Service, Denver, CO U.S.A. Newton, I. 1979. Population ecology of raptors. Buteo Books, Vermillion, SD U.S.A. Pfnnycuick, C.J. 1975. Mechanics of flight. Pages 1-75 in D.S. Earner and J.R. Bang [Eds.], Avian biology. Vol. V. Academic Press, New York, NY U.S.A. Restani, M., A.R. Harmata, and E.M. Madden. 2000. Nu- merical and functional responses of migrant Bald Ea- gles exploiting a seasonally concentrated food source. Wilson Bull. 102:561-568. Ryder, R.A. 1965. A checklist of the birds of the Rio Grande drainage of southern Colorado. Colorado State Univ. Press, Ft. Collins, CO U.S.A. Swenson, J.E. 1983. Is the northern interior Bald Eagle population in North America increasing? Pages 23- 34 in D.M. Bird, Biology and management of Bald Eagles and Ospreys. Harpell Press, Ste. Anne de Belle- vue, Quebec, Canada. Welty, J.C. 1982- The life of birds, 3rd Ed. W.B. Saunders Co., Philadelphia, PA U.S.A. Received 21 December 2001; accepted 6 July 2002 J. Raptor Res. 36(4):265-279 © 2002 The Raptor Research Foundation, Inc. DOES NORTHERN GOSHAWK BREEDING OCCUPANCY VARY WITH NEST-STAND CHARACTERISTICS ON THE OLYMPIC PENINSULA, WASHINGTON? Sean P. Finn^ Biology Department, Boise State University, 1910 University Drive, Boise, ID 83725 US. A. Daniel E. Varland^ Rayonier, 3033 Ingram Street, Hoquiam, WA 98550 US. A. John M. Marzluff College of Forest Resources, University of Washington, Box 352100, Seattle, WA 98195 US. A. Abstract. — To determine stand-level habitat relationships of Northern Goshawks (Accipiter gentilis) on Washington’s Olympic Peninsula, we surveyed all known historically-occupied sites {N = 30) for occu- pancy. We measured 45 forest-stand attributes at these sites and found, using stepwise logistic regression, that goshawks were most likely to occupy historical nest sites with high overstory depth (maximum overstory height-minimum overstory height) and low shrub cover. Forest managers can manage for high overstory depth (>25 m) and low shrub cover (<20%) by conducting a single, moderate-level thinning (leaving 345-445 trees/ha) in young, even-aged 30-35-yr-old stands. Overstory canopy and shrub cover conditions should improve over a 5—10 yr period following thinning. Values for some habitat features (i.e., percent shrub cover, percent canopy closure, and total snags/ha) in our study were near or within the range of values reported for Spotted Owls {Strix occidentalis) in young forests on the Olympic Peninsula. Thus, forest management recommendations described herein may also benefit Spotted Owls. Key Words: Northern Goshawk, Accipiter gentilis; logistic regression-, overstory depth, shrub cover, Washington-, wildlife habitat relationships, silviculture, thinning, forestry. VARIA LA OCUPACION REPRODUCTIVA DEL AZOR CON LAS CARACTERISTICAS DEL SITIO— NIDO EN LA PENINSULA OLYMPIC, WASHINGTON? Resumen. — Para determinar las interrelaciones del habitat a nivel del sitio-nido para el Azor {Accipiter gentilis) en la peninsula Olympic de Washington, estudiamos todos los sitios ocupados conocidos his- toricamente {N = 30) . Medimos 45 atributos de los sitios en bosques y encontramos, usando regresion logistica paso a paso, que estos azores probablemente ocuparon historicamente sitios nido con cubierta densa (maxima altura de la cubierta-minima altura de la cubierta) y baja cobertura arbustiva. Los administradores de bosques puedan manejar cubiertas densamente altas (^25 m) y baja cobertura arbustiva (<20%) llevando un simple, y moderado nivel de entresaca (dejando 345—445 arboles/ha) en plataformas jovenes, o incluso de edades entre 30-35 ahos. La cubierta del dosel y las condiciones de la cobertura arbustiva deben mejorar en un periodo de 5-10 ahos despues de la entresaca. Los valores para algunas caracteristicas de habitat (v.gr. Porcentaje de cobertura arbustiva, porcentaje de cerra- miento del dosel, y total de tocones/ha) en nuestro estudio estuvieron cerca o dentro del rango de los valores reportados para Strix occidentalis en bosques jovenes de la peninsula Olympic. De esta manera las recomendaciones para el manejo de los bosques que se dan aqui, pueden beneficiar ademas a los buhos. [Traduccion de Cesar Marquez] Of critical importance to the success of an or- ganism is its selection and use of resources. Selec- ^ Current address: USGS, Forest and Rangeland Ecosys- tem Science Center, Snake River Field Station, 970 Lusk Street, Boise, ID 83706 U.S.A. ^ Corresponding author’s e-mail address: daniel.varland® rayonier.com tion among available resources may be especially important in large mobile organisms that rapidly move through extensive areas and sample available resources at a relatively coarse grain (Stern 1998). Large mobile organisms living in structurally-com- plex habitats may be particularly responsive to changing conditions because the various compo- 265 266 Finn et al. VoL. 36, No. 4 nents of their habitat may singly or interactively affect their preferred breeding sites, thermal en- vironment, prey abundance and distribution, vul- nerability to predators, or their competitive status (Hilden 1965, Patton 1997). The Northern Gos- hawk {Accipiter gentilis; hereafter known as gos- hawk) is an excellent example of such an organ- ism. Goshawks inhabit boreal and temperate forests within the Holarctic region (Squires and Reynolds 1997). Because they are highly mobile, long-lived, and can take a broad assortment of prey (Squires and Reynolds 1997, Watson et al. 1998), they are able to select among many different avail- able habitats for breeding, roosting, foraging, and other activities. Much research has focused on goshawk habitat use and requirements (Block et al. 1994, Squires and Reynolds 1997), primarily in response to con- cerns over habitat alteration (DeStefano 1998) and potential population declines (Crocker-Bedford 1998). Goshawks are described as forest generalists at large spatial scales, but are a species with nar- rower habitat requirements at nest sites (Squires and Reynolds 1997). At the nest-stand scale, re- search has shown that goshawks select stands with large-diameter trees and high canopy closure, re- gardless of forest type or region (DeStefano 1998). To evaluate relationships between extant habitat and goshawk site-occupancy, we measured 45 forest characteristics in nest stands at 30 historical sites (Table 1); 29 sites were on the Olympic Peninsula and one was just south of this location (Fig. 1). Hereafter, because of the proximity of this site to the peninsula, all sites are referred to as Olympic Peninsula sites. Our objectives were to: (1) estimate current oc- cupancy and breeding rates at all historically oc- cupied goshawk nest sites on the Olympic Penin- sula, (2) describe the relationship between goshawk nest-stand occupancy and nest habitat at- tributes (see Finn et al. 2002 for descriptions at larger spatial scales), and (3) offer management recommendations based on our findings. We hy- pothesized that the 30 historical nest sites we iden- tified for study would still be occupied during our study if forest conditions at these sites had not been degraded since they were first discovered. We reasoned that habitat degradation at nest sites would result in sites being unoccupied and that sites we found to be occupied would more closely resemble forest conditions at historical sites when they were used by goshawks. SruDY Area The peninsula is composed of a central core of rugged mountains surrounded by more level, forested lowlands. Elevation ranges from 0-2420 m, although all known gos- hawk nests were restricted to elevations ranging from ca. 150-810 m. Mixed coniferous forest is the dominant veg- etation over most of the peninsula although tree species, age, and composition vary along a west-east moisture gra- dient and from natural and anthropogenic disturbances (Franklin and Dyrness 1988, Agee 1993). Western slopes are dominated by Sitka spruce {Picea sitchensis) , western hemlock ( Tsuga heterophylla) , and western redcedar ( Thu- ja plicata) whereas the central and eastern portions con- tain pure or mixed stands of western hemlock and Doug- las-fir (Pseudotsuga menziesii) , along with western redcedar and Pacific silver fir {Abies amabilis). Riparian and re- cently-disturbed areas usually contain stands of red alder {Alnus rubra), which may also grow in the understory or in tree gaps on older upland sites. Understory and shrub- layer densities vary widely and contain western hemlock, red alder, Pacific rhododendron {Rhododendron macro- phyllum), sword fern {Polystichum munitum), and salal ( Gaullheria shallon) . Vegetation on the Olympic Peninsula is influenced greatly by the management strategies of the four princi- pal landowners, resulting in a mixture of forest stands of varied serai stages. The Olympic National Park (ONP, 365 000 ha, Holthausen et al. 1995) does not engage in commercial timber harvest. Under the Northwest Forest Plan, the ONP is classified as Congressionally Withdrawn (USDA and USDI 1994). The oiympic National Forest (ONF, 254 000 ha) is managed under the Northwest For- est Plan for multiple uses (USDA and USDI 1994) in which forest management now occurs at low levels in lim- ited areas. Forest management on lands managed by the Washington Department of Natural Resources (DNR, 164 000 ha) is guided to a significant extent by a Habitat Conservation Plan (Washington State Department of Natural Resources 1997). However, the focus on these lands and on private forest lands (347 000 ha) is on com- mercial timber production and forest management. For- est cover conditions on the ca. 1.2 million ha of the Olympic Peninsula may be summarized by the percent of total area of each ownership class in nesting, roosting or foraging habitat for the Spotted Owl {Strix occidentalis) as defined by Holthausen et al. 1995: ONP — 46%, ONF = 38%, DNR = 20%, and private/other non-federal = 7%. MRtJIODS Occupancy at historical nest sites is an important mea- sure of habitat suitability because goshawks usually exhib- it high site fidelity (Crocker-Bedford 1990, Woodbridge and Detrich 1994, Squires and Reynolds 1997). We mea- sured stand attributes at historical nest sites and avoided measures at random locations to eliminate the inherent bias of most use-availability studies that statistically test what is already known; that animals are nonrandomly dis- tributed in the environment (Cherry 1998, Johnson 1999). Occupancy Surveys. We defined 30 goshawk location records as historical nest sites after reviewing all sight re- cords in state and federal databases. All historical nest December 2002 Goshawk Nest-stand Habitat 267 Table 1. Northern Goshawk survey effort at 30 historical nest sites (170 or 314 ha) on the Olympic Peninsula, Washington, 1996-98. In 1996, a 170 ha area was surveyed around each site and in 1997-98, a 314 ha area was surveyed. Site Name/ Number® 1996 No. OF Visits 1997 No. OF Visits 1998 No. of Visits Sta- Court- TUS SHIP Nest- ling Fledg- ling To- tal^ Court- ship Nest- ling Fledg- ling To- tal^ Court- ship Nest- ling Fledg- ling To- tal’’ Calawah/ Sitkum/ 1 2 O 2 2 1 5 2 2 1 5 3 1 4 Raney Creek/29 O 1 2 1 4 1 1 2 1 3 2 6 Dungeoness/16 1 1 1 1 1 3 1 1 2 Burnt Mountain/2 O 3 7 1 11 3 2 1 6 1 1 2 The Hole/30 O 1 6 1 8 3 3 1 7 1 1 2 Donkey Creek/26 O 1 6 1 8 3 2 2 7 1 2 1 4 Snow Creek/ 18 U 2 2 4 3 1 4 1 2 1 4 Morganroth Flat/20 u 1 2 1 4 1 3 1 5 2 2 4 Swede Road/4 u 1 3 1 5 3 1 4 3 1 4 Bear Creek/3 u 2 1 1 4 3 1 4 1 2 1 4 N Fork Solduc/7 o 1 1 2 Mount Zion/ 14 o 1 1 2 Wolf Creek/ 11 o 2 2 4 West Twin River/ 1 u 3 1 4 Dosewallips/ 24 u 3 1 4 Cook Reload/28 u 3 1 4 Wildcat Mountain/5 u 3 1 4 Bear Mountain/ 13 u 2 2 4 Boulder Creek/9 u 3 1 4 Iverson/ 6 u 2 2 4 Bowman Creek/ 19 o 2 2 4 Antelope Creek/ 15 o 2 2 4 Lillian River/ 17 o 1 1 2 Snahapish River/ 25 u 1 2 1 4 Big Canyon/23 u 2 2 4 Palo Alto/ 10 u 2 2 4 Bingham Creek/27 u 2 2 4 Caraco Creek/8 u 2 2 4 Dry Creek/21 u 1 1 2 4 Minnie Peterson/22 u 2 2 4 “ See Fig. 1 for location by site number. ’’ Minimum of four visits required to meet protocol. Where <4 visits shown, occupancy was determined before protocol was met In 1996, occupancy determination was made early in nesting season by Watson (Watson et al. 1998). sites: (1) were in the Washington Heritage Database, first located between 1976-94; (2) were occupied by at least one goshawk when reported; and (3) contained a large stick nest at the time of the goshawk sighting. Annual data on goshawk occupancy were unavailable for all of these sites, so no historical analyses were possible. We surveyed each historical nest site for goshawk occupancy using standardized aural broadcast surveys (Kennedy and Stahlecker 1993, Joy et al. 1994, Finn et al. 2002). We surveyed a minimum of a 1 70-ha circle surrounding 10 historical nest sites in 1996 and a 314-ha circle (1 km radius) surrounding 20 historical nest sites in 1997-98. The survey area was centered on the most recently used nest structure or, when no nest structure was found, on the UTM coordinates on record for that nest site. Be- cause goshawks are highly mobile and tend to be secre- tive, we considered a historical nest site to be occupied if at least one goshawk was visually detected within 1 km of a nest during >1 survey visit (Finn 2000). The protocol survey involved 4-1 1 survey vi.sits where calls were broadcast from each station once during nest- ing, with 1-2 of these survey visits during the fledgling stage (Table 1). Call stations were 300 m apart along tran- sects that were 260 m apart. Call stations on adjacent transects were offset by 130 m. If occupancy was deter- mined during a survey visit, protocol surveys were dis- continued but one additional site visit was made during the fledgling stage to count the number of young 268 Finn et al. VoL. 36, No. 4 * JO • ^9 * 1 « ^'2 ^ Olympic 23 ^ 24 8 * 30 14 20 ,28 17 Olympic National Park 26 ^ 18 3^ ^ 21 A 7 22 Nation^ 29 IS 16 Pacific Ocean Nest Survey Years ★ 1996 - 1998 • 1997 ^ 1998 * 6 Forest 2S 19 •••* ■ i ‘i 0 ^3 ■ Figure 1. Location of 30 historical goshawk nest sites (170 or 314 ha) on the Olympic Peninsula, Washington. Goshawk sites outside Olympic National Park and Olympic National Forest were located on land managed by the Washington Department of Natural Resources or owned by industrial timber companies. Numbers shown are histor- ical nest site numbers; these correspond to those identified by site name in Table 1. All historical sites were first discovered by happenstance, 1976-94. Each site was surveyed for goshawk occupancy in 1996-98, with 10 sites re- ceiving 3 yr of surveys and 20 receiving 1 yr. December 2002 Goshawk Nest-stand Habitat 269 fledged. With one exception, where another research group checked on nest status (Dungeoness site in 1996; Table 1), all known nests were checked for signs of oc- cupancy on 2-6 occasions (dependent on nest condition) during the nesting season. Because goshawk occupation of historical sites can vary over time (DeStefano et al. 1994, Keane and Morrison 1994), we surveyed 10 historical nest sites all three years. This provided an assessment of among-year site occupan- cy. Seventy percent of the sites maintained the same oc- cupancy status among any pairing of years, indicating that goshawk occupancy was consistent among years sam- pled (Finn et al. 2002; Table 1). Therefore, we classified all known historical nest sites on the peninsula as occu- pied if they were occupied Sil yr (N = 12) or not-occu- pied if they were not found occupied during any of the three survey years {N = 18). Classifying sites as “occupied” or “not-occupied” based on one year of surveys leaves room for misclassifi- cation. Sites not occupied during the year of survey may in fact have been occupied earlier or later when no sur- veys occurred. To address this problem we set a = 0.10 as the upper limit for significant differences between oc- cupied and not-occupied sites to counter the possibility that variances were higher in our not-occupied group of stands because of misclassification. In addition, the man- agement recommendations we provide focus on the at- tributes of occupied sites rather than on differences be- tween occupied and not-occupied sites. Habitat Analysis. To assess nest-stand habitat we mea- sured vegetation characteristics at 30 historical nest sites. We defined the nest stand as the homogeneous forest patch surrounding a goshawk nest and delineated stands by scribing boundaries along ecotones and topographic features surrounding the nest after examining 1:12 000 orthophotographs, 1:16 000 aerial photographs, and 1: 24 000 topographic maps. Boundaries were ground- truthed in the field. Historical nest stands averaged 51,4 ha in size (range = 9-146 ha). Areas within historical nest stands where habitat alteration occurred, after gos- hawk occupancy and before our study, were included in our measurements of nest-stand characteristics. Thus, our habitat measurements reflect stand conditions at the time of our surveys, not conditions when the historical nest site was originally determined to be occupied by gos- hawks. We measured 45 forest characteristics (Appendix 1) in 9—13 0.04-ha, systematically placed, circular plots (x = 10.5 plots/stand, SE = 0.26) in each nest stand using a modified USFS Region 6 Timber Stand Exam (USDA Forest Service 1989) and methods described by Husch et al. (1972) and Avery and Burkhart (1983). From plot center, two concentric plots were established: a variable- radius plot to sample trees >12.7 cm DBH (Diameter Breast Height, poletimber and sawtimber) and a fixed- radius plot to sample trees ^12.7 cm DBH (saplings and seedlings) . We estimated basal area, total stem density, and stem and snag density in six size classes (12.8-38.1, 38.2-63.5, 63.6-88.9, 90.0-114.3, 114.4-139.7, and >139.8 cm) from variable radius plots (sampled using a 40 basal area factor prism). We grouped snags into a single size class (^15.2 cm) because of the low number of snags in individual size classes. We also recorded species, DBH, total height, crown ratio, crown class, and level of mistletoe infection for each tree. Quadratic mean diameter at breast height (QDBH) was calculated as ((S DBH'^)/n)*^ We used a clinometer to estimate tree heights. Crown ratio, crown class, and mistletoe abundance were estimated visually for eacb tree in the variable plot and then averaged for the plot. A sample of 1-3 trees of each species on each plot was cored for age and 10-yr radial growth rate. Over- story and understory canopy characteristics (i.e., oversto- ry canopy closure, and maximum and minimum oversto- ry heights) were estimated by averaging four measurements recorded while facing the cardinal direc- tions. Overstory and understory canopy closure were e.s- timated using a moosehorn (Robinson 1947). Overstory and understory height and depth were the mean of four ocular estimates of the height of live branching in the two canopy layers. We used field data to calculate stand density index (SDI, Reineke 1933) and stem density of overstory (38.2-150 cm DBH) and understory (2.5-38.1 cm DBH) trees for each nest stand. All variables were averaged per plot, then per stand. Seedling and sapling densities were measured on a fixed-radius plot where all trees <12.7 cm in diameter were tallied and grouped by 2.5-cm diameter class. Mean values of height, crown ratio, crown class, and mistletoe infection were calculated for each diameter class. We es- timated density and height of shrub and herb layers, and coarse woody debris (CWD) characteristics on eight 1-m^ (Daubenmire 1959), nested plots. Plant association was assigned to all vegetation plots following Henderson et al. (1989). Statistical Analysis. In our study, the number of pre- dictor variables, 45, exceeded the experimental units, 30 Therefore, we first examined the relative differences be- tween occupied and not-occupied nest sites for each var- iable using box-and-whisker plots (Johnson 1999). We used this approach because simultaneous univariate tests increase the Type I error rate (Rice 1989) and because the extensive hypothesis testing inherent in multiple uni- variate tests is inappropriate for exploratory analyses such as we undertook (Cherry 1998, Johnson 1999). We eval- uated the box-and-whisker plots and identified variables with central tendencies that varied with occupancy. We selected a subset of variables that: (1) showed differences in central tendency between occupied and not-occupied sites, (2) had statistical integrity (approximate normal distribution, low multicolinearity), (3) had biological in- tegrity (accuracy of measurement, relevance to gos- hawks), and (4) forest managers could effectively man- age (i.e., overstory canopy closure can be managed, but percent slope cannot). The variables chosen were then evaluated as predictors of goshawk historical nest site oc- cupancy using stepwise logistic regression models (Hos- mer and Lemeshow 1989, PROC Logistic, SAS Inst. 1998) to explain variation in the binomial-response variable (occupied vs. not-occupied, a < 0.10). We compared a main-effect model to models that included selected in- teraction terms to assess their significance. Results We surveyed 10 historical sites all 3 yr (1996-98) and 20 sites during 1 yr (N = 50 annual site-sur- 270 Finn et al. VoL, 36, No. 4 Stem Density Decadence Figure 2. Habitat characteristics of goshawk nest stands (9—146 ha) at 30 historical nest sites on the Olympic Pen- insula, Washington. The historical sites associated with these stands were either not-occupied {N = 18, dark boxes) or occupied {N= 12, white boxes) by goshawks, 1996-98. Boxes depict the median score and 25% and 75% quartiles. Whiskers represent the 10th and 90th percentiles and black dots represent the 5th and 95th percentiles. veys; Table 1 ) . We confirmed presence of goshawks during 20 of these 50 site-surveys (40% occupancy rate). At the 20 site-surveys where we observed gos- hawks, we saw birds during >2 survey visits 75% {N = 15) of the time. During the other five site-sur- veys that revealed occupancy, we observed an adult goshawk during one visit. In all five cases, the bird’s behavior suggested it occupied the area (i.e., alarm vocalization or site tenacity during the ob- servation). We determined that 12 of the 30 his- torical nest sites were occupied (Table 1 ) . All gos- hawk responses were detected <300 m from a historical nest site location. Stand size at historical nest sites was 9-146 ha (x = 51.4, SE = 6.4). Occupied nest stands were smaller in size {x = 32.6 ha, SE = 5.5, range = 11.6-69.3) than not-occupied nest stands {x = 63.9 ha, SE = 10.9, range = 8.7—146.2). Historical nest stands (N = 30) were composed of large {x = 57.3 cm DBH, SE = 2.4; x height = 40.8 m, SE = 1.0), mature (x= 120-yr-old, SE = 12.5) Douglas-fir and western hemlock trees, usually in association with other conifers and occasionally with a few red al- ders. Compared to not-occupied nest stands, occupied nest stands tended to have deeper canopies (oc- cupied median overstory depth = 28.9 m, not-oc- cupied median = 21.6 m; Fig. 2) and higher can- opy closure (occupied median overstory canopy closure = 77.7%, not-occupied median = 71.3%; Fig. 2). Occupied goshawk nest stands had more large-diameter trees than did not-occupied nest stands (i.e., occupied overstory stem density me- dian = 191.9/ha, not-occupied median = 121.5/ ha; Fig. 2). Occupied nest stands generally con- tained more timber (i.e., occupied SDI median = 2204.8, not-occupied median = 1184.2; Fig. 2) and had less shrub cover than did not-occupied stands (occupied median == 15.6%, not-occupied median = 36.9%; Fig. 2). Overstory canopy closure, overstory canopy depth, overstory stem density, SDI, and percent December 2002 Goshawk Nest-stand Habitat 271 Canopy Characteristics Misc. E 25 3 0 W 2 S 1 K • 15.0 T *5 12.5 r ^ ; 5 10.0 1 tj n z 7.5 5.0 1 2.5 • £ 0.0 Figure 2. Continued. Unoccupied Occupied shrub cover met our variable selection criteria and were tested as predictors of goshawk nest stand oc- cupancy. Two of these, overstory canopy depth and percent shrub cover, were useful in distinguishing between occupied and not-occupied nest stands. We found that the equation logit {occupancy) — — 2.91 + 0.163 {overstory depth) — 0.063 {percent shrub cover) significantly described (overstory depth: Wald — 2.97, P = 0.043; percent shrub cover: Wald — 4.13, P = 0.039) and was an ad- equate fit (Hosmer and Lemeshow’s goodness of fit = 4.087, df = 8, P = 0.850) to the data on goshawk occupancy of historical stands (Fig. 3). This model including only main effects fit the data better than did any main effects plus interaction models appraised with log-likelihood ratio criteri- on. Discussion Our research indicates that occupancy of gos- hawk nest stands does vary with nest-stand charac- teristics. Our results agree with most other studies that report overstory canopy as an important fea- ture of goshawk habitat (Squires and Ruggiero 1996, Desimone 1997, McGrath 1997, Patla 1997). These authors reported on the significance of over- story canopy closure in the nest stand but we found stand-wide overstory depth (maximum overstory height-minimum overstory height) more valuable in predicting goshawk nest-stand occupancy. Deep, dense forest canopy {x = 28.7 m, 95% GI = 24.8- 32.6) may provide thermal cover (Newton 1979), protection from rain, or cover protection from predators (e.g.. Great Horned Owls {Bubo vir^ni- anus\, Reynolds et al. 1982, Squires and Reynolds 1997). On the Olympic Peninsula, occupied nest stands typically had about 50% the shrub cover of not- occupied nest stands (Fig. 2). The odds of goshawk occupancy decreased by 47% for each 10% in- crease in percent shrub cover (based on the odds ratio from the logistic regression analysis). Fur- thermore, productive goshawk nest stands had about half the shrub cover of occupied (10.6% vs. 19.0%; Table 2). Most other goshawk habitat studies have not re- ported shrub density (Speiser and Bosakowski 1987, Crocker-Bedford and Chaney 1988, Kennedy 272 Finn et al. VoL. 36, No. 4 Tree Girth Ground Cover Figure 2. Continued. Figure 3. The probability (p) of goshawks occupying a historical nest site on the Olympic Peninsula, Washing- ton, increases with increasing overstory depth and de- creasing percent shrub cover at the nest stand scale (9— 146 ha). 1988, Siders and Kennedy 1996, Desimone 1997, McGrath 1997, Penteriani and Faivre 1997) or have reported it as non-important in contributing to goshawk site occupancy (Hayward and Escano 1989, Squires and Ruggiero 1996, Patla 1997). DeStefano and McCloskey (1997), however, con- tend that the relative absence of goshawks from the Oregon Coast Range is due to the dense understo- ry conditions there, which, in turn, limit prey avail- ability. Goshawks rarely forage near their nests (Beier and Drennan 1997), so the lack of shrub cover we found in nest stands may be unrelated to prey availability. We did not measure shrub cover beyond the nest stand scale, however. At landscape scales (177-ha post-fledging area, 1886-ha home range), goshawk nest stand occupancy was predict- ed by a high proportion (60-75%) of late serai for- est (>70% canopy closure of conifer species with >10% of the canopy in trees >53 cm DBH) and reduced landscape heterogeneity (Finn et al. 2002 ). Our study may have bias because all nest sites were located opportunistically instead of as a result December 2002 Goshawk Nest-stand Habitat 273 Table 2. Nest stand (9-146 ha) habitat characteristics of occupied {N = 12) and productive {N = 8) historical nest sites of the Northern Goshawk on the Olympic Peninsula, Washington, 1996-98. Occupied a Productive'^ Variable‘s Mean SE 95% Cl Mean SE 95% Cl Mean DBH (cm) 58.8 3.7 50.6-67.0 58.2 5.4 45.5-71.0 Quadratic mean DBH (cm) 64.0 4.3 54.5-73.5 63.6 6.3 48.7-78.6 Maximum DBH (cm) 134.2 14.7 102.0-166.4 139.4 21.9 87.7-191.2 Minimum DBH (cm) 17.4 1.1 15.1-19.7 18.0 1.5 14.4-21.6 Mean tree height (m) 43.0 1.7 39.2-46.8 43.1 2.2 37.8-48.3 Crown ratio (index) 5.1 0.2 4.6-5.6 5.0 0.3 4.2-5.7 Crown class (index) 3.1 0.1 3.0-3.2 3.0 0.1 2.9-3.2 Mistletoe (index) 2.5 0.6 1. 2-3.8 1.8 0.6 0.5-3. 1 Radial growth (cm) 1.7 0.3 1.2-2.3 1.9 0.3 1. 1-2.6 Mean tree age (yr) 147.4 22.8 97.2-197.6 128.9 25.4 68.7-189.0 Maximum tree age (yr) 247.6 .33.0 175.1-320.1 229.2 37.0 141.7-316.7 Mean sapling DBH (cm) 5.3 0.7 3.9-6.7 5.1 1.0 2.7-7.4 Mean sapling height (m) 5.7 0.6 4.4-7.0 5.7 0.8 3.8-7.5 Overstory canopy closure (%) 78.4 2.9 72.1-84.7 79.0 4.1 69.3-88.8 Minimum overstory height (m) 18.6 0.9 16.6-20.6 19.9 1.0 17.6-22.2 Maximum overstory height (m) 47.3 2.0 42.9-51.7 47.0 2.7 40.6-53.4 Overstory depth (m) 28.7 1.8 24.8-32.6 27.1 2.3 21.7-32.4 Understory canopy closure (%) 13.7 3.7 5.6-21.8 13.9 5.1 1.8-26.0 Min. understory height (m) 4.6 0.9 2.7-6.5 5.8 1.0 3.2-S.3 Maxi, understory height (m) 16.5 1.1 14.1-18.9 16.6 1.5 13.1-20.2 Understory depth (m) 11.9 0.9 9.9-13.9 10.9 0.7 9.1-12.6 Percent shrub cover (%) 19.0 4.2 9.7-28.3 10.6 2.5 4.8-16.4 Mean shrub height (cm) 41.9 4.7 31.7-52.1 39.9 4.2 29.9-49.9 Percent herb cover (%) 36.5 3.2 29.5-43.5 5.0 4.2 25.1-44.9 Mean herb height (cm) 3.6 0.2 3.1-4.1 3.4 0.2 3.1-3.8 CWD cover (%) 11.0 1.3 8.1-13.9 12.7 1.5 9.0-16.3 CWD height (cm) 42.1 4.8 31.6-52.6 45.0 5.6 31.9-58.2 CWD length (m) 11.0 0.9 9.0-13.0 11.5 1.3 8.3-14.6 CWD DBH (cm) 40.6 3.6 32.7-48.5 41.2 5.1 29.2-53.2 Slope (%) 40.5 5.0 29.5-51.5 42.5 5.7 28.9-56.0 Aspect (degrees) 269.9 26.3 218.4-321.3 294.5 63.5 170.1-58.9 Basal area (m^/ha) 71.4 5.9 58.5-84.3 68.5 5.7 55.0-81.9 Sapling den. (No. /ha) 797.3 180.8 399.4-1195.2 831.3 259.3 218.1-1444.6 Small stem density (No. /ha) 286.6 41.9 194.4-378.8 297.2 63.8 146.3-448.1 Med. stem density (No. /ha) 151.0 22.5 101.6-200.4 146.4 23.4 91.2-201.7 Large stem density (No. /ha) 39.1 7.7 22.2-56.0 32.5 8.1 13.3-51.8 Ex.-large stem den. (No. /ha) 2.3 1.4 0.0-5.3 2.4 2.0 0.0-7.0 Understory stem den. (No./ha) 1083.9 219.8 600.0-1567.8 1128.5 320.9 369.6-1887.4 Overstory stem den. (No. /ha) 192.5 21.9 144.2-240.8 181.3 22.6 127.9-234.8 Live stem density (No. /ha) 485.4 47.4 381.0-589.8 488.0 67.9 327.4-648.6 Stand density index 2136.0 223.1 1644.9-2627.1 2107.1 317.6 1355.9-2858.3 Tree species richness (No.) 3.2 0.4 2.2-4.2 3.1 0.5 1.9-4.3 Percent hardwood (%) 1.9 1.2 0.0-4.4 1.2 0.6 0.0-2.7 Snag density (No. /ha) 35.8 6.0 22.6-49.0 40.0 6.7 24.2-55.8 Seedling (No. /ha) 2031.0 657.8 583.3-3478.7 2671.5 911.8 515.1-4827.9 ‘‘12 sites: eight productive (where ^1 young fledged) and four occupied with no productivity. Eight productive sites. ■= See Appendix 1 for descriptions of habitat variables. 274 Finn et ai.. VoL. 36, No. 4 of systematic searches of the full range of goshawk habitat (Squires and Reynolds 1997, Daw et al. 1998) . Our sample included all of the nests reliably reported on the Olympic Peninsula over an 18-yr period, 1976-94. Thus, though our sample is small, it IS likely adequate to represent goshawk habitat use by goshawks on the Olympic Peninsula. Fur- thermore, Daw et al. (1998) compared goshawk nest stand habitat in stands found opportunistically with those found by systematic searches in Oregon and found no differences in two key habitat vari- ables, large tree density and canopy cover. Their sample of opportunistically-located nests included nests found by individuals searching for goshawk nests with a preconceived notion of goshawk hab- itat preferences (i.e., searching likely habitat). Nests in our study, however, were found by individ- uals whose reasons for being in the held varied greatly (i.e., hikers, Marbled Murrelet {Brachyram- phus marmoratus) surveyors, foresters conducting timber cruises) and who, in nearly all cases, were focused on activities other than hnding goshawk nests. The Daw et al. (1998) study provides empir- ical evidence that the method we employed for identifying historic nest sites was adequate. While we provide useful information on the characteristics distinguishing between occupied versus not-occupied nest stands, we recognize that site occupancy is not necessarily indicative of qual- ity habitat (Van Horne 1983, Vickery et al. 1992). We believe our occupancy surveys are good indi- cators of habitat quality for goshawks because, in our study, nest-stand occupancy and reproduction were closely correlated (Finn 2000, Finn et al. 2002) . Young successfully fledged from eight of 1 2 occupied sites. Moreover, only one of the 10 sites we surveyed every year was consistently occupied, but never produced fledglings. Small-scale (e.g., nest tree, nest vicinity) habitat influences on occupancy of goshawk nest stands were not identihed in our study (Finn 2000). Thus, forest managers should focus on stand scale (this paper) and landscape scale (Finn et al. 2002) hab- itat management for goshawks. Management Impi.igaiions Goshawk nest stand size in our study averaged 32.6 ha in occupied historical sites and 63.9 ha in not-occupied historical sites, which is within the range of 10—100 ha reported by Squires and Reyn- olds (1997) for goshawk nest stands across North America. We recommend that managers who seek to address nest stand level habitat needs tailor stand size after the ranges reported here. Our research indicates that goshawk nest-stand habitat may be provided on the Olympic Peninsula by managing stands to create deep overstory can- opies and low shrub cover (Table 2, Figs. 2 and 3). Long et al. (1983) and Bailey (1996) report that large crowns can be created in dominant and co- dominant trees by thinning stands at 20-50 yr of age. Thinning reduces crown competition, thereby enhancing crown development and tree diameter growth. Thinning, however, allows more light to reach the forest floor which also promotes under- story growth (Hayes et al. 1997, Thysell and Carey 2000). Hayes et al. (1997) indicated that thinning to moderate densities facilitates crown develop- ment but limits development of understory be- cause the canopy closes rapidly. To accelerate the development of deep overstory canopies in young even-aged stands, we recom- mend that a single moderate-level thinning take place in stands 30-35 yr of age. On the Olympic Peninsula and elsewhere in western Washington and Oregon, moderate-level thinning would result in retention of 345—445 trees/ha where heavy thin- ning would result in retention of 148—247 trees/ha (L. Raynes pers. comm.). To promote deep overstory canopies at the onset of stand initiation, planting a mixture of shade tol- erant (i.e., western hemlock) and intolerant (i.e., Douglas-fir) tree species at 3-4 m spacing is rec- ommended (ca. 1000 trees/ha, L. Raynes pers. comm.). Spacing trees farther apart will reduce crown competition and may result in excessive can- opy depth (L. Raynes pers. comm.), therein cre- ating inadequate flight space for goshawks. A sin- gle, moderate-level thinning of the trees remaining in the stand (there will be some mortality) at 30- 35-yr-old across the range of diameter classes, as opposed to thinning a specific diameter class, would promote deeper forest canopies as the stand develops; this is because more growing space is available, particularly for the larger trees (L. Ray- nes pers. comm.). Once thinning has occurred, overstory canopy development and a concomitant reduction in shrub cover would occur over a 5-10 yr period. After this, stands would likely be suitable for gos- hawk nesting for as long as they were retained. Mean tree age of occupied nest stands in our study was 147 yr (A = 12, SD = 71.3, range - 51-275). December 2002 Goshawk Nest-stand Habitat 275 The four youngest occupied stands were 50-70-yr- old while the four oldest were 200-275-yr-old. We suggest that it is not stand age per se that is important to goshawk nesting, instead it is the hab- itat elements associated with older stands (this study: deep overstory canopy, low shrub cover, Squires and Reynolds 1997, DeStefano 1998: large trees with high canopy closure). The extent to which these features can be created in younger- aged stands will make forest management for gos- hawks more economically practicable. Other silvi- cultural prescriptions may work as well as those we suggest or may be more appropriate, depending on site conditions. Currently, most stands on the Olympic Peninsula are managed on a 40-50 yr ro- tation (L. Raynes pers. comm.), due primarily to a re-tooling of local sawmills to handle smaller-di- ameter logs. The importance of old forest attributes to the Northern Spotted Owl, which also inhabits western Washington forests and is sensitive to habitat loss, is well known (Gutierrez et al. 1995, Horton 1996, Irwin et al. 2000) . Goshawks, however, use a broad- er range of forest structural stages than do North- ern Spotted Owls (DeStefano 1998). We found gos- hawks nesting in stands as young as 51 yr, and Bosakowski et al. (1999) report on goshawks nest- ing in 40-54-yr-old managed stands in western Washington. In research on Northern Spotted Owl use of young forest habitat on the Olympic Peninsula, Buchanan et al. (1999) report values for some hab- itat features important to Northern Spotted Owls (i.e., total snags/ha, percent shrub cover, percent canopy closure, and coarse woody debris cover) that are near or within the range of values we found for these same features for goshawks (Table 2) . Thus, forest management as described herein may also benefit Northern Spotted Owls. Acknowled gments We want to thank those from private companies and public agencies that offered their assistance. The Wash- ington Department of Natural Resources, Rayonier, Port Blakely Tree Farms, Champion Pacific, and the University of Washington provided funding and logistical support. Additional support, equipment, and in-kind aid was pro- vided by Weyerhaeuser, the Washington Department of Fish and Wildlife, Olympic National Park, and Olympic National Forest. We thank R. Meier, L. Raynes, S. Katzer, N. Wilkins, T. McBride, J. Eskow, D. Runde, L. Hicks, S. Horton, P. Harrison, D. Hays, S. Desimone, E. Seaman, S Lemioux, and L. Young for their helpful input. Tech- nical and analytical advice was provided with enthusiasm by L. Bond, M. Cowing, S. Gossett, B. Lehman, J. Munger, S. Novak, K. Steenhof, and T. Zarriello. Special thanks to D. Yonkin for his outstanding help in the field, his ded- ication to the project, and his spirit. Superior assistance in the field was provided by B. Griffith, H. Tall, K. Bees- ley, B. Davies, T. Bloxton, J. Delap, J. Swingle, J. Wagen- knecht, and L. Vandernoot. Thanks also to M. Fuller, S Knick, J. Munger, K. Titus, C. Crocker-Bedford, R. Man- nan, and M. Restani who reviewed and greatly improved this manuscript. Literature Cited Agee, J.K. 1993. Fire ecology of Pacific northwest forests. Island Press, Washington, DC U.S.A. Avery, T.E. and H.E. Burkhart. 1983. Forest measure- ments. McGraw-Hill, New York, NYU.S.A. Baiity, J.D. 1996. Effects of stand density reduction on structural development in western Oregon Douglas- fir forests — a reconstruction study. Ph.D. dissertation, Oregon State University, Corvallis, OR U.S.A. Beier, P. and J.E. Drennan. 1997. Forest structure and prey abundance in foraging areas of Northern Gos- hawks. Ecol. Applic. 7:564-571. Block, W.M., M.L. Morrison, and M.H. Reiser (Eds ) 1994. The Northern Goshawk: ecology and manage- ment. Stud. Avian Biol. 16. Bosakowski, T, B. McCullough, F.J. Lapsansky, and M.E. Vaughn. 1999. Northern Goshawks nesting on a private industrial forest in western Washington./. Rap- tor Res. 33:240-244. Buchanan, J.B., J.C. Lewis, D.J. Pierce, E.D. Forsman, and B.L, Biswell. 1999. Characteristics of young for- ests used by Spotted Owls on the western Olympic Peninsula. Northivest Sci. 73:255-263. Cherry, S. 1998. Statistical tests in publications of the Wildlife Society. Wildl. Soc. Bull. 26:947-953. Crocker-Bedford, D.C. 1990. Goshawk reproduction and forest management. Wildl. Soc. Bull. 18:262—269 . 1998. The value of demographic and habitat studies in determining the status of Northern Gos- hawks (Accipiler gentilis atricapillus) with special refer- ence to Crocker-Bedford (1990) and Kennedy (1997) J. Raptor Res. 32:329-336. AND B. Chaney. 1988. Characteristics of goshawk nest stands. Pages 210—216 in R.L. Glinski, B. Giron Pendleton, M.B. Moss, M.N. LeFrank, Jr., B.A. Millsap, and S.W. Hoffman [Eds.], Southwest raptor manage- ment symposium and workshop. Natl. Wildl. Fed., Washington, DC U.S.A. Daubenmire, R.F. 1959. A canopy-coverage method of vegetational analysis. Northwest Sci. 33:43-64. Daw, S.K., S. DeStefano, and R.J. Steidl. 1998. Does sur- vey method bias the description of Northern Goshawk nest-site structure? J. Wildl. Manage. 62:1379-1384. Desimone, S.M. 1997. Occupancy rates and habitat rela- tionships of Northern Goshawks in historic nesting areas in Oregon. MS thesis, Oregon State University, Corvallis, OR U.S.A. 276 Finn et al. VoL. 36, No. 4 DeStefano, S. 1998. Determining the status of Northern Goshawks in the west: is our conceptual model cor- rect? J. Raptor Res. 32:342-348. , S.K. Daw, S.M. Desimone, and E.C. Meslow. 1994. Density and productivity of Northern Goshawks: implications for monitoring and management. Stud. Avian Biol. 16:88—91. AND J. McCloskey. 1997. Does vegetation struc- ture limit the distribution of Northern Goshawks in the Oregon coast ranges? J. Raptor Res. 31:34—39. Finn, S.P. 2000. Multi-scale habitat influences on North- ern Goshawk occupancy and reproduction on Wash- ington’s Olympic Peninsula. M.S. thesis, Boise State University, Boise, ID U.S.A. , J.M. Marzluff, and D.E. Varland. 2002. Effects of landscape and local habitat attributes on Northern Goshawk site occupancy in western Washington. For. Sci. 48:427-436. Franklin, J.F. and C.T. Dyrness. 1988. Natural vegetation of Oregon and Washington. Oregon State Univ. Press, Corvallis, OR U.S.A. Gutierrez, R.J., A.B. Franklin, and W.S. LaHaye. 1995. Spotted Owl (Strix occidentalis) . In A. Poole and F. Gill [Eds.], The birds of North America, No. 179. The Academy of Natural Sciences, Philadelphia, PA and The American Ornithologists’ Union, Washington, DC U.S.A. Hayes, J.P., S.S. Chan, W.H. Emmingham, J.C. Tappeiner, L.D. Kellogg, and J.D. Bailey, 1997. Wildlife re- sponse to thinning young forests in the Pacific North- west. /. For. 95:28-33. Hayward, G.D. and R.E. Escano. 1989. Goshawk nest-site characteristics in western Montana and northern Ida- ho. Gowrfor 91:476-479. Henderson, J.A., D.H. Peter, R.D. Lesher, and D.C. Shaw. 1989. Forested plant associations of the Olym- pic National Forest. U.S. For. Serv. Gen. Tech. Rep. RG-ECOL-TP 001-88, Portland, OR U.S.A. Hidden, O. 1965. Habitat selection in birds. Ann. Zool. Fenn. 2:53—75. Hoi.thausen, R.S., M.G. Raphael, K.S. McKelvey, E.D. Forsman, E.E. Siarkey, and D.E. Seaman. 1995. The contribution of federal and non-federal habitat to persistence of the northern Spotted Owl on the Olym- pic Peninsula, Washington: report of the reanalysis team. U.S. For. Serv. Gen. Tech. Rep. PNW-GTR-352, Portland, OR U.S.A. Horton, S.P. 1996. Spotted Owls in managed forests of western Oregon and Washington. Pages 215-231 in D.M. Bird, D.E. Varland, and J.J. Negro [Eds.], Rap- tors in human landscapes: adaptations to built and cultivated environments. Academic Press Ltd., Lon- don, U.K. Hosmer, D.W. and S. Lemeshow. 1989. Applied logistic regression. John Wiley and Sons, New York, NYU.S.A. Husch, B., C.I. Miller, and T.W. Beers. 1972. Forest mensuration, 2nd Ed. The Ronald Press Company, New York, NY U.S.A. Irwin, L.L., D.F. Rock, and G.P. Miller. 2000. Stand structures used by northern Spotted Owls in managed forests, y. Raptor Res. 34:175-186. Johnson, D.H. 1999. The insignificance of statistical sig- nificance testing./. Wildl. Manage. 63:763-772. Joy, S.M., R.T. Reynolds, and D.G. Leslie. 1994. North- ern Goshawk broadcast surveys: hawk response vari- ables and survey cost. Avian Biol. 16:24-30. Keane, JJ- AND M.L. Morrison. 1994. Northern Goshawk ecology: effects of scale and levels of biological orga- nization. Stud. Avian Biol. 16:3-11. Kennedy, P.L, 1988. Habitat characteristics of Cooper’s Hawks and Northern Goshawks nesting in New Mex- ico. Pages 218-227 in R.L, Glinski, B. Giron Pendle- ton, M.B. Moss, M.N. LeFrank, Jr., B.A. Millsap, and S.W. Hoffman [Eds.], Southwest raptor management symposium and workshop. Natl. Wildl. Fed., Washing- ton, DC U.S.A. and D.W. Stahlecker. 1993. Responsiveness of nesting Northern Goshawks to taped broadcasts of three conspecific calls. /. Wildl. Manage. 57:2249-2257. Long, J.N., J.B. McCarter, and S.B. Jack. 1983. A mod- ified density diagram for coastal Douglas-flr. West. J Appl. For. 3:88-89. McGrath, M.T. 1997. Northern Goshawk habitat analysis in managed forest landscapes. M.S. thesis, Oregon State University, Corvallis, OR U.S.A. Newton, I. 1979. Population ecology of raptors. Buteo Books, Vermillion, SD U.S.A. Oliver, C. D. and B. C. Larson. 1990. Forest stand dy- namics. McGraw-Hill, New York, NY U.S.A. Patla, S.M. 1997. Nesting ecology and habitat of the Northern Goshawk in undisturbed and timber harvest areas on the Targhee National Forest, Greater Yellow- stone Ecosystem. M.S. thesis, Idaho State University, Idaho Falls, ID U.S.A. Patton, D.R. 1997. Wildlife habitat relationships in for- ested ecosystems, Revised Ed. Timber Press, Portland, OR U.S.A. Penteriani, V. AND B. Faivre. 1997. Breeding density and nest-site selection in a goshawk Accipiter gentilis popu- lation of the Central Apennines (Abruzzo, Italy). Bird Stud. 44:136-145. Reineke, L.H. 1933. Perfecting a stand-density index for even aged forests. J. Agric. Res. 46:627-637. Reynolds, R.T., E.C. Meslow, and H.M. Wight. 1982 Nesting habitat of coexisting accipiters in Oregon./ Wildl. Manage. 46:124-138. Rice, W.R. 1989. Analyzing tables of statistical tests. Evo- lution 43:223-225. Robinson, M.W. 1947. An instrument to measure forest crown cover. For. Chron. 23:222-225. SAS Institute, Inc. 1998. SAS/STAT user’s guide. Re- lease 6.03. SAS Inst., Cary, NC U.S.A. Siders, M.S. and P.L. Kennedy. 1996. Forest structural Goshawk Nest-stand Habu a i 277 December 2002 characteristics of accipiter nesting habitat: is there an allometric relationship? Conrfor 98:123-1 32- Spf.tser, R. and T, Bosakowski. 1987. Nest site selection by Northern Goshawks in northern New Jersey and southeastern New York. Condor 89:387-394. Squires, J.R. and L.F. Ruggiero. 1996. Nest-site prefer- ence of Northern Goshawks in south-central Wyo- ming. /. Wildl. Manage. 60:170-177. AND R.T. Rit'VNOl.DS. 1997. Northern Goshawk {Ac- cipiter gentilis) . In A.. Poole and F. Gill [Ens.], The birds of North America, No. 298. The Academy of Natural Sciences, Philadelphia, PA and The American Orni- thologists’ Union, Washington, DC U.S.A. Stern, S.J. 1998. Field studies of large mobile organisms: scale, movement, and habitat utilization. Pages 289- 307 in D.L. Peterson and V.T. Parker [Eds.], Ecolog- ical scale: theory and applications. Columbia Univ. Press, New York, NY U.S.A. Thvneix, D.R. and A.B. Carey. 2000. Effects of forest management on understory and overstory vegetation: a retrospective study. U.S. For. Serv. Gen. Tech. Rep. PNW-488, Olympia, WA U.S.A. USDA Forest Service. 1989. Stand exam program. Field procedures guide. U.S. For. Serv. Gen. Tech. Rep., Pa- cific Northwest Region, Portland, OR U.S.A. USDA Forest Servk:e and USDl Bureau of Land Man- agement. 1 994. Record of decision for amcndmcnls to Forest Service and Bureau cjf Land Management planning documents within the range of the northern Spotted Owl. Standards and guidelines for manage- ment of habitat for late-successional and old-growth forest related species within the range of the northern Spotted Owl. Washington, DC U.S.A. Van Horne, B. 1983. Density as a misleading indicator of habitat quality. /. Wildl. Manage. 47:893-901. Vickery, P.D., M.T. Hunter, Jr., and J.V. Wells. 1992 Is density an indicator of breeding success? Auk 109‘ 706-710. Washincd'on Stale Department of Na'ilral Resources 1997. Final habitat conservation plan. Unpubl. Rep., Olympia, WA U.S.A. Watson, J.W., D.W. Hays, S.P. Finn, and P. Meeiian-Mar- TiN. 1998. Prey of breeding Northern Goshawks m Washington, y. Raptor Res. 32:297-305. WOODBRIDtiE, B. AND PJ Detrich. 1994. Territory occu- pancy and habitat patch size of Northern Goshawks in the southern Cascades of California. Stud. Avian Biol. 16:83-87. Received 2 December 2000; accepted 16 July 2002 Associate Editor: Marco Restani 278 Finn et al. VoL. 36, No. 4 !U •1— • C o OJ cn V a 1 r£3 C/> o O c 1) .rl o u C o c« O CO t3 D s.^ s c« d V B c > -D 4= m ;-N G O C C4H OJ T) 03 d d lA V X. 32 be d d !-H ^ ;xj "3 2 ^ 9" ^ S w u z g u g H g w Q d o X d d o be o t+H 1> SO d d X d Id so CM a. J3 u d K d V o o cd fl QJ CM i> CTi o; -15 u C/J X V A a u V •w d d d O d a a X •M^ 'd P3 o .M A in 13 > o 03 d o rO u 13 Sh d d ■ 1 — ( 73 Ch d 13 "a 3+H o M D d . 13 TO ^ d d 13 Cl 73 o bo d ii 13 d 0 • U OJ Lh • 1^ 73 1 I q q • ^ 73 Ui d 1) o d o d.< 13 g a. X bo 13 d ch +H QJ T3 c/5 CL U B ^ d al 9 c bo -j 3 'd d - £ ^ B 'd tH d « O ^ q o a o M 1) Cl 73 13 be d 1 13 'a d o X 13 u M d d X 1 d u 73 D M d V tH Cm O >•- M 0 M A 1 13 73 d d a d a '3 S d D X 3+H 73 5h O 13 73 g X d 5 ° • SP d3 S-. d o M A M QJ 73 d d A D ~Sl a d A Cl 00 A d B a o o d a. CL ijj a ,£p M ^ Cl'B X ^ " On Lm qj o . d 1 o -d 1) A O o OJ OJ V OJ CJ biD be be bo bo bo q q q q q q u Sm u u S-N OJ OJ V Ur OJ OJ OJ > > > U > > > q q q 0 q q q L( Lh u X OJ c/5 u u 3 3 iS 3 3 3 3 3 0 3 3 3 u u u 0 u u u 0 O 0 s 0 0 0 g ’Dh 13 d d d cr A 13 M d d a o d d o ll o M A D 13 d 13 be 'd 13 I3M M O M A 13 d O a d a 'd 13 •£P ’S 13 d-, d o w C/! U OJ > o a d a ■ dH Cl OJ X >- u O c/5 i-H OJ > o 13 M d kM Cl o d d 13 d^ M o M A u (U T3 d L Id 'd 13 U O .M A u OJ 73 d d d a 'a X 'C 13 d^ 1 o ■M A U 13 73 d d a d a 'x d Cl 13 X d^ M o M A !h 13 73 d L 1 d d O d 13 d Sh X A M d d d M d PI December 2002 Goshawk Nest-stand Habitat 279 T3 !U g o U s K OJ 2 OJ U 2/5 ^ d ^ -K P .2 tU in Id P M U g •d £ c/2 P d 00 CO I j> CM CM Q + X P3 Q ^ u CM -r I d ^ G CM Ih ■Sw.r- ct o Q P- n K P5 Q 1T3 00 CD I CM 00 CO CO p p PQ P a 02 rP. O CP p P 00 “*” GO cO ^P GO 4- OD '-' m P Q n a 02 -1-) C/5 ■ O • rH t/: d u t2 GX U P P + P u p + ; p ' 02 -I— > C/} o ‘in d u t3 G p u p p be P o 03 00 03 J-p <1 p c/2 p m T2 P d d o 12 B 12 P 03 12 'G d 2> o 03 03 03 00 ^ 03 p ^ o — : ^ d ci ^ p ^ ■d g 2 P P 02 O P p p d d 12 12 4< 12 d P rB" P 12 d m V m P P P P r^ G 12 p 12 P p 7! ^ p d Sh P GO P U P P rt X P S gPp UP • ' ■' u .a > .a ^ 22 ^ ■a|'a d ' d u d c/^ O §K 2 S 2 12 p g c 4 — I C/5 o u 0 u c/5 OJ V5 C/5 T3 OJ 4—1 cu 1 p J> CM 1-M ^ § s r a * -<-* — , CA P d d o u u 12 12 ^ a .2 a ^ pi V5 OJ bo G ’C QJ -Cl 4—* OJ o x! d d o 12 d p 12 r. c/5 o J8 (j d d u P -tJ C o G d O HJ cti 1/^ Cd d a - a ph -d C/2 12 Sr d ■u U d m O P O u m 12 U ‘u H 12 p in P-r O d d g P U P 12 be d ’p u be d 0 12 be d 12 be d 1 a a a a a a a a 'p c/5 'p c/5 a a a 12 Sr d u 12 a OJ 5-h re fH Sr 12 a 12 P Sr 12 a P OJ -M OJ c/5 C/3 .-H C/5 P • rH ^P o C/3 ’P Oh C/5 • rH P a c/5 • rH P Oh c/5 *P Cu c/5 'p Oh c/5 ‘p Oh c/5 ■p Oh d • rH P d _d p d c/5 • ^ P Oh u r5 P T in X P cr C/5 r5 d -H rS a re Oh S § Sr P Sr P S 4—1 re Cl P rS 3 u 0 a 3 u 0 a 3 u 0 P 12 r-i CZ) 3 u 0 _g u a 0 u P P O O p P P o p pq o p pq o p CQ o p pq o p pq o p pq o p 1) .a p 12 .a p pq o p Tally Tallv j d • rH 4—) c/5 P 3 u 0 Key a u CM 1— H A C/5 a tu 4-^ ■B' 'K f2 12 P 12 -kJ in 12 > P CZ) P U P .-B’ 'in d 12 P X P P P r-. ^ J dos S-o S' .S' &ai s a o G3 a a^ 02 12 in d 12 P d 12 P 12 -u in 12 .a p in 0^ t' ^ 2 d ^ 12 C/3 in Sr d 02 be S p Sr HI d d X p P P D X P P o d- • rH C/5 d OJ OJ c/3 ;-H O w c/5 OJ 6 p c /3 d 12 P be d • rH 'P d CZ) d 12 p be d * r^ P 12 12 C/3 P P P Id d P hH O i c/! .bn OJ — d a; T3 d u ^ a; ^ d ^ 12 u a d 12 Sr H d p 12 G IZ H P Cd v. p d o • rH 4—1 d *0 be ^ d d - a p p d p d d P P tu .SP ‘5 re 1) u OJ u d re C/l G o; X! 6 OJ cc c4_( - 1 S e a o re CJ biD ;-4 re z w Q CO ^ Q Q " W ^ a; G p /. Raptor Res. 36(4):280-286 © 2002 The Raptor Research Foundation, Inc. SUBORDINATE MALES SIRE OFESPRING IN MADAGASCAR EISH-EAGLE {HALIAEETUS VOCIFEROIDES) POLYANDROUS BREEDING GROUPS Ruth E. Tingay* School of Geography, University of Nottingham, Nottingham, NG7 2RD ILK. and the Peregrine Fund, 5668 Flying Hawk Lane, Boise, ID 83709 U.S.A. Meianie Culver, Eric M. Haiuerman, and James D. Fraser Department of Fisheries and Wildlife Sciences, Virginia Polytechnic Institute and State University, Blacksburg, VA 24061 U.S.A. Richard T. Watson The Peregrine Fund, 5668 Flying Hawk Lane, Boise, ID 83709 U.S.A. Abstract. — The island endemic Madagascar Fish-Eagle (Haliaeetus vociferoides) is one of the most en- dangered birds of prey. Certain populations in west-central Madagascar sometimes exhibit a third, and sometimes a fourth, adult involved in breeding activities at a nest. We applied DNA fingerprinting to assess relatedness among 17 individuals at four nests. In all nests with young, a subordinate rather than the dominant male sired the offspring. Within-nest relatedness comparisons showed that some dominant males had an apparent hrst-order relationship with the female. Between-nest relatedness comparisons showed that some adults had an apparent hrst-order relative at another nest in the study area. Findings that subordinate males contribute to breeding, and that adults in an area may be related, may require conservation measures such as translocation to assure the species’ survival. Keywords: Madagascar Fish-Eagle, Haliaeetus vociferoides; DNA fingerprinting, mating system; nest helper, polyandry. MACHOS subordinados engendran descendencia en grupos de reproduccion po- LIAnDRICA en AGUIIAS PESCADORAS DE MADAGASCAR {HALIAEETUS VOCIFEROIDES) Resumen. — El aguila pescadora endemica de la isla de Madagascar {Haliaeetus vociferoides) es una de las aves rapaces mas amenazadas de extincion. Algunas poblaciones en el occidente-centro de Madagascar exhiben algunas veces un tercero y a veces un cuarto adulto involucrado en las actividades reproductivas en un solo nido, Aplicamos un analisis de ADN para evaluar el parentesco entre 17 individuos de cuatro nidos. En todos los nidos con juveniles, un macho subordinado mas que el dominante engendro la prole. Las comparaciones de parentesco dentro de los nidos mostro que algunos machos dominantes tenian aparentemente una relacion de primer orden con la hembra. Las comparaciones entre nidos mostraron que algunos adultos tuvieron un pariente de primer orden en otro nido dentro del area de estudio. El hallazgo de que los machos subordinados contribuyen a la reproduccion, y que los adultos en un area pueden estar relacionados entre si, pueden hacer necesarias medidas de conservacion tales como traslados para asegurar la supcrvivencia de la especie. [Traduccion de Cesar Marquez] The island endemic Madagascar Fish-Eagle {Hal- laeetus vociferoides) is considered critically endan- '' These two authors contributed equally to this manu- script. ' Present address of corresponding author: Wildlife and Fisheries Science, University of Arizona, Tucson, AZ 85721 U.S.A.; e-mail address: culver@ag.arizona.edu gered (Collar et al. 1994) . With 63 known breeding pairs, and an estimated total breeding population of 100-120 pairs (Rabarisoa et al. 1997), it is among the most endangered birds of prey in the world (Langrand and Meyburg 1989, Watson et al. 1993, 1996). Madagascar Fish-Eagles exhibit an un- usual dispersal and breeding strategy, possibly re- stricting the species’ distribution and abundance 280 December 2002 Relaxedness of Madagascar Fish-Eagles 281 through limited dispersal or occurrence of in- breeding. Breeding was believed to be monoga- mous, but at 46% of known nests, a third, and sometimes a fourth adult is involved with the breeding activities of the primary pair (Watson et al. 1999). Based on banding studies at several nests (Watson et al. 1999), extra-pair birds were believed to be progeny (possibly only male) from previous years. Such delayed dispersal can result in forma- tion of cooperative breeding groups, a relatively rare breeding system among birds (Stacey and Ko- enig 1990, Ligon 1999), especially among raptors (Simmons 2000, and references therein). Ecologi- cal or behavioral factors may influence evolution of cooperative breeding strategies (Newton 1979, Oring 1986, Faaborg and Bednarz 1990, Stacey and Koenig 1990, Sherman 1995), and contribute to attendance of additional adults at Madagascar Fish- Eagle nests. Understanding dispersal and repro- ductive strategies is critical for developing a man- agement plan to ensure the species’ survival. DNA markers have been applied to a variety of questions regarding conservation of birds (Haig and Avise 1996) . DNA fingerprinting proved useful to assess relatedness at the nest (Westneat 1990, Wetton et al. 1992, Haig et al. 1993, 1994a, 1994b) and population (Triggs et al. 1992, Fleischer et al. 1994) levels, to infer species-level population ge- netic structure (Longmire et al. 1991), and to es- timate relatedness in captive stocks (Kirby 1990: 239). We used DNA fingerprinting to determine paternity among Madagascar Fish-Eagle adults at- tending a nest, and to examine the level of relat- edness among adults within and between nests. Methods Samples. We studied three trios and one quartet of fish-eagles at a site in west-central Madagascar (19°S, 44°30'E) on a daily basis during one breeding season from 24 June-5 October 1999. The area is tropical decid- uous dry forest containing several lakes (3.09-4.86 km^) and supports 11 fish-eagle territories (Rabarisoa et al. 1997). Eagles were marked and are referred to by num- ber. Nest sites are referred to by location and nest num- ber (Ankerika 4, Befotaka 2, Befotaka 3, and Soamalipo 2). A dominance hierarchy was observed at each nest based on aggressive interactions between adults. Aerial pursuits (chasing) and physical displacements from ei- ther the nest or from perches within 200 m of the nest tree, often accompanied by a distinctive ‘displacement’ call, were observed throughout the breeding period and were interpreted as signs of aggression (Tingay 2000). Males are referred to as either dominant (a), or subor- dinate ((3 or y). We were unable to establish the domi- nance hierarchy at nest site Befotaka 3. Nestlings were briefly removed from the nest at ca. 7 wk of age and banded. Blood (0.1-0.25 ml) was taken from the brachial vein (Tingay 2000), immediately placed in 4.5 ml of lysis buffer (100 mM, pH 8.0, 100 rriM EDTA, 10 niM NaCil, 0.5% SDS) in a polypropylene tube, labeled, and stored at ambient temperature. DNA Purification. Approximately 200 |xl of blood/ buffer solution was placed in 800 fxl lysis buffer for 10 min. Protein digestion was performed with 500 (xl of su- pernatant from the first step, 500 |xl of fresh lysis buffer, and 0.5 mg/ml proteinase K, with incubation at 37°C overnight. Extractions were performed in 1:1 phenol chloroform, and 24:1 chloroform : isoamyl alcohol. DNA was precipitated using cold 95% ethanol and 5% sample volume of 5M (0.082M final) ammonium acetate. DNA was resuspended in 25 |xl deionized water and stored at -20°C. DNA Fingerprinting. DNA samples were digested sep- arately with Hinfl, Rsal, and Haelll. Digests were loaded onto 1% TBE agarose gels (20 cm X 24 cm), and sub- jected to electrophoresis (Sambrook et al. 1989) at 32 V for 25 hr. Identity Sizing Standard (Lifecodes Corpora- tion, Stamford, CT) was placed in several lanes of the gel to provide molecular weight markers. DNA in the gel was stained using ethidium bromide, photographed using fW luminescence, and transferred (Southern 1975) onto a MagnaCharge 0.45 micron nylon membrane (Micron Separations Inc., Westborough, MA) . Jeffreys et al (1985) and Jeffreys (1987) minisatellite probe 33.15 was hybridized using the NICE hybridization solution (Life- codes Corporation, Stamford, CT) onto digested, im- mobilized DNA. Both the 33.15 probe and Identity Sizing Standard were labeled with NICE chemiluminescence Unhybridized probe and size standard were washed from the membrane using Quick-Light wash solutions (Life- codes Corporation) . The hybridized probe was illuminat- ed with Lumi-Phos 480 (I.ifecodes Corporation) and vi- sualized by exposure to Kodak XAR5 X-omat film. DNA Fingerprinting Analysis. Gels were arrayed with samples from individuals attending a nest adjacent to one another. If all hybridization bands observed for nestlings could have been inherited from the primary pair, we con- cluded that the primary pair was the parents. If, however, a hybridization band could be accounted for only by par- entage by a nest attendant, we concluded that an extra- pair mating had occurred. There was only one adult fe- male at each nest. The male that was most dominant and exhibited the greatest paternal investment (Tingay 2000) was considered the male of the primary pair. DNA band-sharing (Bruford et al. 1992) was calculated as S = 2n^y/ (n^ + riy), where = the number of bands shared by both individuals, = the total number of bands exhibited by individual x, and = the total num- ber of bands exhibited by individual y. Band-sharing was estimated for all combinations of individuals in this study The range of S for known parent-offspring combinations provided a quantitative expectation of how many bands must be shared before a hypothesis of familial related- ness was supported. Results Parentage Assessment of Nestlings and Juve- niles. DNA fingerprinting techniques were used to 282 Tingay et al. VoL. 36, No. 4 assess relatedness of 17 eagles at four nests. Two enzymes {HaAll and RscH) produced clearly inter- pretable results yielding a total of 34 bands scored, 24 of which were variable and 10 invariant (Table 1). Of the 24 variable bands, six were informative in determining one or more possible parents for the two nestlings at Soamalipo 2; three for the ju- venile at Befotaka 3; and seven for the nestling at Befotaka 2. Blood samples were available only for adults at Ankerika 4. A nest-by-nest assessment of parentage is presented below. Befotaka 2. Female 121, a male 118, and P male 8 attended the nest. Nestling 47 shared three var- iant /facIII and one variant Bsdi hybridization bands with adult female 121 , and two variant 7/fldlI and one variant hybridization bands with p male 8 , suggesting that subordinate P male 8 was the father of the nestling 47, and not a male 118. Befotaka 3. Female 6 , potential a male 48, and potential a male 150 attended this nest. Juvenile 128 shared one HadW band and one Rsa\ band with adult female 6 . Banding records show that ju- venile 128 fledged from this nest in 1998. Although band sharing showed it unlikely that either adult male at the nest in 1999 (48 and 150) was the fa- ther, it is highly probable that the adult female at the nest is the mother {S — 0.95 is the highest value in the study, female 6 has been recorded at this nest site every year since 1993, and no other female has been recorded at this nest). Soamalipo 2. Female 103, a male 5, p male 136, and 7 male 30 attended this nest. Nestling 68 shared one HaeWl band with adult female 103 and two bands with 7 male 30. Nestling 00 shared one HadW and one Rsoi band with adult female 103 and one HaeWl and three Rsai bands with 7 male 30. The apparent father of both nestlings is subordinate 7 male 30. Relatedness Estimates of All Adults Within and Between Nests. Among 136 pairwise comparisons, band-sharing among individuals ranged from 0.58- 0.95, with a mean value of 0.79. Partitioning pair- wise band-sharing into within- and between-nest components showed no difference (mean S — 0.80 within nests and 0.79 between nests). After ac- counting for eight known first-order relative pairs (parent-offspring, full-sibling), band-sharing was higher among first-order relatives (v = 0.87, range = 0.82-0.95) than overall (x = 0.79; Table 2). Us- ing these findings, relatedness among adults at- tending nests (male-male, male-female) was deter- mined (Table 2). Ankerika 4. Band-sharing values suggested a po- tential first-order relationship between female 113 and a male 31, but not between the female 113 and P male 34. Band-sharing suggested that the males were unrelated. Befotaka 2. Band-sharing values did not support a first-order relationship between the female and either male, nor between males. P male 8 had two bands not shared with any individual within the study population; trapping records indicate that p male 8 fledged from the Befotaka 3 nest in 1993. Befotaka 3. Band-sharing values indicated a po- tential first-order relationship between female 6 and male 150, but not between female 6 and male 48. Band-sharing suggested that the males were un- related. Soamalipo 2. Band-sharing values indicated a po- tential first-order relationship between female 103 and a male 5, but not between female 103 and the two subordinate males (P 136 and 7 30). Band- sharing between a male 5 and 7 male 30 indicated a potential first-order relationship. Relatedness estimates between nests. Comparing among nests, we observed high band-sharing val- ues between female 121 (Befotaka 2) and female 103 (Soamalipo 2), male 5 (Soamalipo 2) and male 48 (Befotaka 3), and between male 34 (Ankerika 4) and female 6 (Befotaka 3) , suggesting potential first-order relatedness between these pairs of adults. Discussion Subordinate males may have fathered all nest- lings in this study. At Soamalipo 2, one subordinate male appeared to have fathered both nestlings, however, because a male 5 and 7 male 30 are close relatives, and because of missing data for a male 5, we cannot exclude a male 5 as a possible father of one or both nestlings. At all nests, paternity by subordinates could have occurred by chance, as all attending males copulated with the female (Tingay 2000). Paternity by subordinates was surprising giv- en that dominant males invested more energy to the nesting attempt than subordinate males (Tin- gay 2000). This level of dominant male investment may be explained by the apparent first-order relat- edness of the female and the dominant male at three of four nests (Ankerika 4, Befotaka 3, and Soamalipo 2). Because 50% of alleles are shared with a first-order relative, and 25% with an off- spring of a first-order relative, then shared alleles are transmitted to the next generation if a first- December 2002 Reiatedness of Madagascar Flsh-Eagles 283 Table 1. DNA fingerprinting hybridization bands (Jeffreys 33.15 probe) observed for individual Madagascar Fish- Eagles. Bands are designated by enzyme used (H = Hae\\\ or R = Rsai) and molecular weight of bands in kilobase pairs. Sex and rank for individuals is indicated (F = female, aM = alpha male, pM = beta male, yM = gamma male, NSL = nestling, JUV = juvenile). Ankerika 4^ Befotaka 2'’ Befotaka 3'= Soamalipo 2^ F aM (3M F aM (3M ; NSL F aM? aM? JUV F aM PM yM NSL NSL Individual 113 31 34 121 118 8 47 6 150 48 128 103 5 136 30 68 00 Bands H 16.0 + + + + + + + + + + 4 4 4 4 4 4 4 H 10.7 -F + + 4 4 H 8.5 + + + + + + + + + 4 4 4 4 4 4 4 4 H 7.3 + + + + + + + 4 4 4 4 4 4 H 6.5 + + + + 4 H 6.0 -F -F -F 4 4 H 5.7 + + + + 4 4 H 5.6 + + 4 H 5.2 + 4 H 4.9 -F -F 4 4 H 4.7 + H 4.5 + + + + + + + 4 4 4 4 4 4 4 H 3.9 + + + + + + + + + 4 4 4 4 4 4 H 3.6 -F -F -F + + 4 4 4 H 3.2 + + + + + + + + 4 4 4 H 2.9 + + + + + + + + + 4 4 4 4 4 4 4 4 H 2.7 + -F + + + + + + 4 4 4 4 4 4 4 4 H 2.6 + + + + -F -F -F -F -F 4 4 4 4 4 4 4 4 H 2.2 + + + + + + -F + + 4 4 4 4 4 4 4 4 H 1.5 + + + + + + + + + 4 4 4 4 4 4 4 4 H 1.4 -F -F -F + -F -F + + + 4 4 4 4 4 4 4 4 H 1.0 + + + + 4 4 4 4 4 4 4 H 0.9 + + + + + + + 4 4 4 4 4 R 12.0 + + + + + + + + + 4 4 4 ? 4 4 4 4 R 5.2 -F -F -F -F ? R 5.0 + + + + + + 4 ? 4 4 4 R 4.7 + + 4 ? 4 4 4 R 4.5 + ? R 4.4 + -F -F -F -F 4 ? 4 R 4.2 + ? 4 4 R 3.3 + + + + + + + + 4 4 4 ? 4 4 4 4 R 1.5 + + + 4 4 ? 4 4 4 4 R 1.4 + + + -F -F -F -F 4 4 ? 4 4 4 R 1.2 4 + + + + + + + + 4 4 4 ? 4 4 4 4 Total No. bands per individual 21 21 22 20 20 22 26 21 21 20 19 21 14 19 21 20 23 ^Fifteen bands are variable at Ankerika 4 (H 7.3, H 6.5, H 6.0, H 5.7, H 5.6, H 4.9, H 3.6, H 1.0, H 0.9, R 5.2, R 5.0, R 4.7, R 4 4, R 1.5, R 1.4). All other bands are invariant. Of the variable bands at Befotaka 2, six are shared between the nestling and the female (H 10.7, H 4.5, H 3.6, H 1.0, R 5.0, R 1 5), six are shared between the nestling and the beta male (H 6.5, H 6.0, H 0.9, R 5.2, R 4.4, R 1.4); one is shared between the nestling, female, and beta male (H 3.2); and four are variable but are not observed in tbe nestling (H 5.7, H 5.2, H 4.7, R 4.5). All other bands are invariant. Of the variable bands at Befotaka 3, four are shared between the juvenile and the female (H 7.3, H 6.5, H 3.6, R 4.7); and ten are variable but are not observed in the juvenile (H 10.7, H 6.0, H 5.7, H 5.6, H 4.9, H 3.2, H 1.0, R 5.0, R 4.2, R 1.5). All other bands are invariant. Of the variable bands at Soamalipo 2, four are shared between nestling 68 and the female (H 7.3, H 3.9, H 3.2, R 5.0); two are shared between nestling 68 and the gamma male (R 4.7, R 1.4); and one is shared between nestling 68, the female, and the gamma male (H 4.5). Five bands are shared between nestling 00 and the female (H 3.9, H 3.6, H 3.2, R 5.0, R 4.4); four are shared between nestling 00 and the gamma male (H 0.9, R 4.7, R 4.2, R 1.4); and one is shared between nestling 00, the female, and the gamma male (H 4.5). Five are variable but are not observed in either nestling (H 10.7, H 6.0, H 5.7, H 5.2, H 4.9); and all other bands are invariant. 284 Tingay et al. VoL. 36, No. 4 be 2 G O S 2 ^ be cu g ^ S II . C/!) ^ 2; QO 2 " 2 s cn >•-. tu ,!-i ■o 1) fl d a o u V u d Cl o ’J} d a i - ^ be ^ I, C G X >- ^ c VS 0 a; i.1 a V d r“H 1 G 1-H B X CO G ’G CO ? G X X 2 II ' — J GCL be „ .a 2 ^ l=i G O w 2 ri ce a o -G G rt s vs •i— < d 'd a V be !=! Q Cfi G tj O G u X •J-i vs OJ 2 a II S a S G a X ^ jS X X a a; X G II 2 G X !i 0 TJ a; 0 2 u G ^ T3 G 2 t/2 o; G t3 2 _G ,0 > c a ■S G T3 X G OJ G C/D X) . V £ vs ? 5 -h Ph d V X u cu u c • I-^ cu d * vs H OJ 0 u > OJ 0 d d vs •w 0 vs XJ OJ c "d 1 OJ c T! c OJ OJ c/s vs G T3 O o 2 Cfl CG bo d W I VS CM rO u d u vs d be d a be G d cvr O 9 0 c/) J C/D O 2 2 ^ S o ^ GO GO GO. a GO o > 00 2 2i CO 2 o 2 »-o 2 ^ CO QO O I> QO J> d d ^ XI oc I> 00 00 d d d CO O 1— I X 00 00 i> 00 d d d d xO X GO 0 CM !>• 00 X X X d d d d d GO 0 M GO 0 i> C 3 D J> 00 00 d d d d d X 0 T~( 0 10 X X x> 00 X X 0 r- X d d d d d d d 00 in X 10 X i> O) X X X GO 00 d d d d d d d 00 d »oi>eoo mXi-Hi>GO oqi^oooo cc cc CO dddd ddddd ac cu. cc |b-i CM CD. GO 2 u cn D Q a G /C CO CJD X X 0 CO 03 in 00 tr^ j> i> i> 00 in X I> d d d d d d d d d d X X GO X 0 cvr X 03 i> CO m i> X X 00 i> d d d d d d d d d d d m 1 '' 1 '' CO X 0 i> CO CO (M X 0 X 00 QO X 03 X X i> 03 d d d d d d d d d d d d X 1 — H QO X Tf CO 03 X X 0 1 — 1 X" X X'' X X X X X l'^ X X X'- X d d d d d d d d d d d d d m CO X rH X m X 0 0 X GO X 00 X X^ (X 00 i> X X X x^ 00 d d d d d d d d d d d d d d X 03 X X 03 X X CO m 1 -H 0 rH CO 00 t> X-; 00 00 00 i> 00 QO x> 00 00 d d d d d 0 d d d d 0 d d d d T— H OC X 0 CC 00 GO 1-0 X 0 X QO X CM i—H 1-0 CM 0 QO QO X T— H T“^ 1 — H ! — H hJ X n. s S S 1 HH C/D a cn Ph a ca 2 Ph a a a (TL 2 CM d d 24 ’E d OJ -uj 2 G OJ CQ CO M 0 d CG M X d G -(-J 2 a G u 0 PQ X NSL 0 0.86 0.73 0.89 0.84 0.79 0.71 0.86 0.91 0.77 0.79 0.81 0.86 0.86 0.67 0.82 0.88 Deciembkr 2002 REIATEDNESS of MADACiASCAR FiSH-EAGI.ES 285 order relative reproduces successfully. At Soamali- po 2, the dominant male gained an additional ge- netic advantage by having two potential hrst-order relatives at the nest (the female and the y male). It would be advantageous to be a male at the same nest as a brother, because if either mated success- fully, then shared genes are transmitted to the next generation. Although a strategy of assisting repro- ductive efforts of close relatives may be advanta- geous for some Madagascar Fish-Eagles, apparently it is not the only strategy in use. At Befotaka 2, the dominant male was not the father, and nor was he a first-order relative of either the female or the sub- ordinate male. At Befotaka 3, a juvenile female did not disperse. This is the hrst observed instance of a female nest- ling from a previous year remaining at a nest (Ra- fanomezantsoa 1997). Flere, delayed dispersal was not associated with observed helping activity, yet the female juvenile was tolerated at the nest. Al- though inconclusive, our findings do not exclude the delayed dispersal hypothesis. Between-nest relatedness comparisons revealed that some adults had a potential close relative (par- ent-offspring or full-sibling) at another nest within the study area. This suggests that hrst-order rela- tives (excluding nestlings) are as likely to be found among nests as within a nest. We are currently investigating the full range of breeding strategies in the Madagascar Eish-Eagle. We intend to determine whether this species ex- hibits genetic monogamy or polyandry by extend- ing our sample size and duration of study. Studies of another cooperative polyandrous raptor species, the Galapagos Hawk {Buteo galapagoensis) has re- vealed mixed paternity at nests over two consecu- tive breeding seasons (Faaborg et al. 1995). How- ever, the dominance hierarchy we have observed among cooperative hsh-eagles has not been docu- mented among Galapagos Hawks, which may or may not influence the occurrence of genetic mo- nogamy within polyandrous groups of Madagascar Fish-Eagles. If delayed dispersal is obligatory in this species, recolonization of unoccupied habitats may have to be promoted by active conservation mea- sures, such as the translocation of individuals from other areas. Additionally, copulation by closely-re- lated pairs, as observed in this study, suggests that the effects of inbreeding may have to be consid- ered in conservation planning. For example, if hrst-order relatives are found to be producing off- spring, conservation managers may wish to target some of those specific individuals as likely candi- dates for translocation, in order to reduce the probability of further inbreeding and to create an opportunity for outbreeding with other, genetically dissimilar, individuals. Acknowledgments We conducted this study under The Peregrine Fund’s Madagascar Fish-Eagle and Wetland Conservation Pro- ject. We thank the Madagascar Direction des Eaux et For- ets, Tripartite Commission, Association Nationale pour la Gestion des Aires Protogees and United Nations Educa- tional, Scientific, and Cultural Organization for collabo- ration. This work was funded in part by grants from the Liz Claiborne and Art Ortenberg Foundation, Environ- ment Now, the John D. and Catherine T. MacArthur Foundation, Biodiversity Support Program, Hawk Moun- tain/Zeiss Optics 1999 Research Award, Jim Brett Global Conservation Fund and University of Nottingham. We are grateful to Professor David Parkin for technical assis- tance with DNA purification, to Jim Berkelman and two anonymous referees for their thoughtful comments on an earlier draft of this paper, and to Peregrine Fund Field Manager Loukman Kalavah for expertise in the field Literahjre Cited Bruford, M.W., O. Hanotte, J.F.Y. Brookfield, and T Burke. 1992. Single-locus and multilocus DNA finger- printing. Pages 225-269 ire A.R. Hoelzel [Ed.], Molec- ular genetic analysis of populations: a practical ap- proach. IRE Press, Oxford, U.K. Coliar, N.J., M.J. Crosby, and A.J. Stattersfield. 1994 Birds to watch 2: the world list of threatened birds BirdUife International, Cambridge, U.K. Faaborg, J. and J.C. Bednarz. 1990. Galapagos and Har- ris’ Hawks: divergent causes of sociality in two raptors Pages 359-383 in P.B. Stacey and K.W. Koenig [Eds ] , Co-operative breeding in birds: long term studies of ecology and behaviour. Cambridge Univ. Press, Cam- bridge, U.K. , P.G. Parker, U. Df.Uay, TJ. deVries, J.C. Bednarz, S. Maria Paz, J. Naranjo, and TA. Waite. 1995. Con- firmation of co-operative polyandry in the Galapagos Hawk (Buteo galapagoensis). Behav. Ecol. Sociohiol. 36: 83-90. Fieischer, R.C., C.L. Tarr, and T.K. Pratt. 1994. Ge- netic structure and mating system in the Palila, an endangered Hawaiian honeycreeper, as assessed by DNA fingerprinting. Mol. Ecol. 3:383-392. FIaig, S.M., J.R. Belthoff, and D.H. Ailen. 1993. Ex- amination of population structure in Red-cockaded Woodpeckers using DNA profiles. Evolution 47:185- 194. ? J-b- Ballou, and NJ. Casna. 1994a. Identifica- tion of kin structure among Guam rail founders a comparison of pedigrees and DNA profiles. Mol. Ecol 4:109-119. , J.R. Wai.ters, andJ.H. Plissner. 1994b. Genetic 286 Tingay et ai.. VoL. 36, No. 4 evidence for monogamy in the Red-cockaded Wood- pecker, a cooperative breeder. Behav. Ecol. Sociobiol. 23: 295-303. AND J.C. Avise. 1996. Avian conservation genetics. Pages 160-189 mJ.C. Avise and J.L. Hamrick [Eds.], Conservation genetics: case histories from nature. Chapman and Hall, New York, NY U.S.A. Jeffrey, A.J. 1987. Highly variable minisatellites and DNA fingerprints. Biochem. Soc. Trans. 15:309-317. , V. Wilson, and S.L. Thein. 1985. Individual-spe- cific “fingerprints” of human DNA. Nature 316:76-79. Kirby, L.T. 1990. DNA fingerprinting: an introduction. Stockton Press, New York, NY U.S.A. Langrand, O. and B.-U. Meyburg. 1989. Range, status, and biology of the Madagascar Sea-Eagle Haliaeetus vociferoides. Pages 269-278 in B.-U. Meyburg and R.D. Chancellor [Eds.], Raptors in the modern world. WWGBP, Berlin, Germany. Eicon, J.D. 1999. The evolution of avian breeding sys- tems. Oxford Univ. Press, Oxford, U.K. Longmire, J.L., R.E. Ambrose, N.C. Brown, TJ. Cade, T.L. Maechtle, S.W. Seegar, F.P. Ward, and C.M. White. 1991. Use of sex-linked minisatellite fragments to investigate genetic differentiation and migration of North American populations of the Peregrine Falcon {Falco peregrinus) . Pages 217-229 in T. Burke, G. Dolf, A.J. Jeffreys, and R. Wolff [Eds.] , DNA fingerprinting: approaches and applications. Birkhauser-Verlag, Ba- sel, Switzerland. Newton, I. 1979 Population ecology of raptors. Poyser, London, U.K. Oring, L.W. 1986. Avian polyandry. Pages 309-351 mR.J. Johnston [Ed.], Current ornithology. Vol. 3. Plenum Press, New York, NY U.S.A. Rabarisoa, R., R.T. Watson, R. Thorstrom, and J. Ber- KELMAN. 1997. Status of the Madagascar Fish-Eagle in 1995. Ostrich 68:8—12. RaFANOMEZANTSOA, S. 1997. Behaviour and natal dispers- al of fledgling Madagascar Fish-Eagles. Pages 403-41 1 in R.T. Watson [Ed.], Madagascar wetlands conserva- tion project, progress report 3, 1995-1996. The Per- egrine Fund, Boise, ID U.S.A. Sambrook, J., E.F Fritsch, and T. Maniatis. 1989. Mo- lecular cloning: a laboratory manual, 2nd Ed. Cold Spring Harbor Press, New York, NY U.S.A. Sherman, P.T. 1995, Social organisation of co-operatively polyandrous White-winged Trumpeters {Psophia leu- coptera) in Peru, Auk 112:296-309. Simmons, R.E. 2000. Harriers of the world: their behav- iour and ecology. Oxford Univ. Press, Oxford, U.K, Southern, E.M. 1975. Detection of specific sequences among DNA fragments separated by gel electropho- resis. J. Mol Biol 98:503-527. Stacey, P.B. and W.D. Koenig (Eds.). 1990. Co-operative breeding in birds: long term studies of ecology and behaviour. Cambridge Univ. Press, Cambridge, U.K. Tingay, R. 2000. Sex, lies, and dominance: paternity and behaviour of extra-pair Madagascar Fish-Eagles Hal- iaeetus vociferoides. M.S. thesis, University of Notting- ham, Nottingham, U.K. Triggs, S.J., M.J. Williams, S.J. Marshall, and G.K. Chambers. 1992. Genetic structure of Blue Duck {Hy- menolaimus malacorhynchus) populations revealed by DNA fingerprinting. Auk 109:80-89. Watson, R.T., J. Berkelman, R. Lewis, and S. Razafin- dramanana. 1993. Conservation studies on the Mad- agascar Fish-Eagle Haliaeetus vociferoides. Proc. Pan-Afr. Ornithol. Cong. 8:192-196. , S. Thomsett, D. O’Daniei., and R. L ewis. 1996 Breeding, growth, development, and management of the Madagascar Fish-Eagle {Haliaeetus vociferoides).] Raptor Res. 30:21-27. , S. Razafindramanana, R. Thorstrom, and S. Ra- fanomezantsoa. 1999. Breeding biology, extra-pair birds, productivity, siblicide, and conservation of the Madagascar Fish-Eagle. Ostrich 70:105-111. Westneat, D.F 1990. Genetic parentage in the Indigo Bunting: a study using DNA fingerprinting. Behav Ecol. Sociobiol. 27:67-76. Wetton, J.H., D.T. Parkin, .and R.E. Garter. 1992. Use of genetic markers for parentage analysis in Passer do- mesticus (house sparrows). Heredity 69:243-254. Received 26 November 2001; accepted 23 July 2002 J. Raptor Res. 36(4):287-293 © 2002 The Raptor Research Foundation, Inc. NESTING AND PERCHING HABITAT USE OE THE MADAGASCAR FISH-EAGLE James Berkelman^ and James D. Fraser Department of Fisheries and Wildlife Sciences, Virginia Polytechnical Institute and State University, Blacksburg, VA 24061-0321 U.S.A. Richard T. Watson The Peregrine Fund, 5668 West Flying Hawk Lane Boise, ID 83709 U.S.A. Abstract. — We documented Madagascar Eish-Eagle (Haliaeetus vociferoides) nest and perch use on lakes and rivers and compared parameters of used trees to unused reference trees. Nest and perch trees were broader and taller, had more unobstructed branches, and were less obstructed by adjacent trees com- pared to reference trees. Perch trees also were more often deciduous than reference trees. Nest sites had more shoreline perch trees than reference sites. Logistic regression models with tree height as the independent variable distinguished nest and perch trees from randomly selected reference trees. Models with number of perch trees along a 1.25 ha (50 m width) shoreline section distinguished nest sites from reference sites. These models suggest that the presence of trees >15 m tall within 50 m of the shoreline is a good predictor of Madagascar Eish-Eagle habitat use. Key Words; Madagascar Fish-Eagle, Haliaeetus vociferoides; habitat; Madagascar, nest tree, perch tree, shore- line. USO DE HABITAT DE ANIDACION Y PERCHA DEL AGUILA PESCADORA DE MADAGASCAR Resumen. — -Documentamos el uso de nidos y perchas para el aguila pescadora de Madagascar {Haliaeetus vociferoides) en lagos y rios y comparamos parametros de arboles usados con arboles no usados de referenda. Los nidos y arboles percha fueron mas anchos y mas altos, tenian mas ramas despejadas, y estaban menos obstruidos por arboles adyacentes en comparacion con los arboles referenda. Los arboles percha fueron ademas algunas veces mas deciduos que los arboles control. Los sitios nido disponian de mas arboles percha costeros que los sitios de referencia. Los modelos de regresion logistica con la altura de los arboles como variable independiente distinguieron los nidos y arboles percha de arboles control seleccionados aleatoreamente. Los modelos con numeros de arboles percha cerca a 1.25 ha (50 m de ancho) de la seccion de costa distinguieron los sitios nido de los sitios referencia. Estos modelos sugieren que la presencia de arboles >15 m de alto dentro de 50 m de la linea costera es un buen pronosticador del uso de habitat del aguila pescadora de Madagascar. [Traduccion de Cesar Marquez] With a population estimate of 99 breeding pairs (Rabarisoa et al. 1997), the Madagascar Fish-Eagle {Haliaeetus vociferoides) is one of the rarest birds of prey in the world (Meyburg 1986). Until recently, little was known about the species’ ecology and sta- tus. Langrand and Meyburg (1989) noted that the Madagascar Fish-Eagle used tall trees near water for nests and foraging perches, but prior to this study, there had been no detailed quantitative stud- ' Present address: Department of Wildlife Ecology, Uni- versity of Wisconsin, 218 Russell Labs, 1630 Linden Dr., Madison, W1 53706-1598; e-mail address: jberkelman® facstaff.wisc.edu ies of Madagascar Fish-Eagle nesting or perching habitat use. Nelson and Horning (1993) estimated from sat- ellite data that Madagascar’s forest cover had been reduced to 10.4% of the island by 1990. Nest-site availability is a key limiting factor for raptor pop- ulations (Newton 1979). Also perch-tree distribu- tion is a reliable predictor of Bald Eagle {Haliaeetus leucocephalus) distribution on the Chesapeake Bay (Chandler et al. 1995). Thus, we focused our study on both nest and perch trees, along with the sur- rounding habitat conditions. The objectives of this study were to determine characteristics of nest trees, nest sites, and perch trees used by Madagas- 287 288 Berkelman et al. VoL. 36, No. 4 Table 1. Sites where Madagascar Fish-Eagle nests and perches were investigated in the region of Antsalova, Mada- gascar, 1994. Site names are lakes unless otherwise indicated. Site Latitude, Longitude Number oe Eagle Pairs Masiadolo 18°41'S, 44°28'E H Besara 18°41'S, 44°16'E 1 Soahanina River 18°46'-48'S, 44°16'~19'E 3^ Antsahafa 18°48'S, 44°29'E P Masama 18°50'-5TS, 44“28'-29'E 2 Tsiandrora 18°58'S, 44°38'E 1 Andranolava 1'’ 19°0'S, 44°2TE 1 Befotaka 19°1'-2'S, 44°24'-25'E 3 Soamalipo 18°59'-19°2'S, 44°26'-27'E 3^ Ankerika 19°T-2'S, 44'^27'-44°28'E 4 Andranovorimirafy 19°3'S, 44°27'E 1 Andranolava 2'^ 19°4'S, 44°25'E 1 Antsakotsako 19°6'S, 44°33'E 1^ Ampozabe 19°9'S, 44°40'E 1 Bevoay 19°9'S, 44°25'E 1 Manambolo River 19°8'-9'S, 44°44'-49'S 2 Maromahia 19°11'-12'S, 44°37'-38'E 1 Bejijo 19°12'-14'S, 44°32 -33'E 1 No nest was found for one of the fish-eagle pairs at five of the sites. Two of the lakes in the study had the same name. car Fish-Eagles and to develop predictive models to identify fish-eagle nesting and perching habitat. SiUDY Area and Methods We conducted the study during the first half of the Madagascar Fish-Eagle breeding season from 21 May-14 August 1994. We investigated fish-eagle nesting and peiching habitat in a 3000 km^ area in the Antsalova re- gion of western Madagascar (18°40'-19°15'S, 44^15'- 44°50'E) that included the drainages of the Manambolo, Beboka, and Soahanina rivers west of the Bemaraha Pla- teau. Topography consisted of coastal plains and low roll- ing hills with elevations ranging from sea level to 126 m. Soils were shallow and sandy, and the vegetation was a patchwork of dry deciduous forest, savanna, wetlands, mangrove swamps, and rice paddies. The climate was sub- humid and tropical with a dry season from April-Octo- ber and a wet season from November-March. Mean an- nual rainfall in the region ranged from 1000-1500 mm (Donque 1972). We defined a nest tree as any tree in which we ob- served nest construction, incubation, or brood rearing. A nest site was the area within 300 m of the nest tree. A peich tree was any tree in which we observed adult fish- eagles perching. We measured nest and perch trees of every known Madagascar Fish-Eagle pair in the study area (Table 1). Our perch-tree sample {N = 29) was larger than our nest-tree sample (N = 24) because we did not find a nest for five of the fish-eagle pairs. We measured characteristics of fish-eagle nest trees and randomly selected reference trees to determine if trees used by fish-eagles differed from average large trees (>20 cm diameter at breast height [DBH]) near the same bodies of water. To compare nest and perch trees to available large trees, we randomly selected a reference tree for each nest or perch tree. We selected trees at the same distance from the water as the nest or perch tree To do this, we measured with a hip chain the distance from the nest tree to the nearest water (nest-water dis- tance), from perch tree to nearest water (perch-water dis- tance), and the distance along the shore between the nest and perch trees (nest-perch distance). We then ran- domly selected a shoreline reference point on the same body of water as the nest tree that we had measured (within 1 .5 km along the banks for nest trees on rivers) To select each nest reference tree, we went to the shoreline reference point and moved inland a distance equal to the nest-water distance and selected the nearest tree >20 cm in DBH as our nest reference tree. We used the same shoreline reference point to select a perch ref- erence tree by moving the nest-perch distance in the same direction (left or right) along the shoreline as that between the used nest and perch trees. We then moved inland the perch-water distance and selected the nearest tree >20 cm in DBH as our perch reference tree. We used 20-cm DBH as a minimum size for reference trees based on the minimum size of Bald Eagle perch trees on the Chesapeake Bay (Buehler et al. 1992). We measured DBH of nest trees to the nearest cm and used a clinometer to measure height to the nearest me- ter. We counted branches in the tree canopy that we es- timated to be >5 cm in diameter and unobstructed for 1 m above and below. We recorded arc of accessibility by standing at the base of the tree and using a compass to December 2002 Mada(;asc]AR Fish-Eagle Nests 289 measure the total arc (0°-360°) that was unobstructed by other trees for an estimated distance of 10 m from the trunk and 3 m below the tree’s crown (Buehler et al. 1992). We recorded nest-tree species and classified growth form following Keister and Anthony (1983). Our classification was based on the location of the lowest fork in the trunk, and whether the tree was dead. We classi- fied growth form as large if the lowest fork was in the lower third of the trunk, medium if the lowest fork was in the middle third of the trunk, and small if the lowest fork was in the upper third of the trunk. We recorded growth form as dead top if the top third of the crown was dead and as snag if the entire tree was dead and leafless, regardless of the location of the lowest fork in the trunk. We measured minimum distance of each nest tree to water with a hip chain and minimum distance to human disturbance, building, road, and fish-eagle nest from maps and aerial photos. Human disturbances included agricultural clearings, rice paddies, villages, tombs, and fishermen’s camps. Temporary, seasonal shelters that were not used during the fish-eagle breeding season were not considered buildings. There were no paved roads and few motor vehicles in the area, and the most traveled roads were traversed by less than one motor vehicle per day, even in the dry season. Oxcarts frequently were used to transport materials, so we recorded any oxcart track as a road. We considered trees ^6.1 m high and with ^30° ac- cessibility from the shoreline to be potential perch trees based on the smallest recorded perch tree used by Bald Eagles on the Chesapeake Bay (Buehler et al. 1992). We counted perch trees within 50 m of the water along a 250 m shoreline section centered on the nest tree or refer- ence tree (Chandler et al. 1995). We classified mean sur- rounding canopy height to 5-m intervals ranging from 0- 25 m based on visual observation. We measured the perch tree that we saw fish-eagles use most frequently for foraging for each of the 29 fish-eagle pairs in the study area. Eleven (37.9%) of the pairs were observed for at least 6 hr, at least once per week during the breeding season (May-October) in 1992, 1993, and 1994 as part of a related study (Watson et al. 1999). The remaining 18 (62.1%) pairs were observed for at least 6 hr, at least three times per breeding season from 1992- 94. We measured the same tree characteristics for perch trees that we measured for nest trees. We tested the null hypothesis of no difference between trees or sites used by breeding Madagascar Fish-Eagles and reference trees or sites for each of the numerical variables using the Wilcoxon signed-ranks test. We paired each fish-eagle nest or perch tree with the randomly se- lected reference tree on the same water body. We did not test for differences in distance to water because this was a criterion for selecting reference trees. We used the chi- square test of equal proportions to determine if fish-eagle habitat use was different from expected use for the fol- lowing categorical variables: tree species, deciduous ver- sus evergreen trees, growth form, and surrounding can- opy height. If >20% of expected values were <5, we used the likelihood ratio chi-square test statistic (Agresti 1990). We developed logistic regression models to predict the probability of fish-eagle use of trees and sites based on the measured habitat variables using stepwise analysis Our significance level for variables to both enter and exit models was P = 0.05. We used dummy variables for growth form categories in the logistic regression (Hos- mer and Lemeshow 1989). We constructed classification tables for each logistic regression model by using the es- timated logistic probabilities for each tree or site to pre- dict fish-eagle use (Hosmer and Lemeshow 1989). We considered trees or sites as correctly classified as used hy fish-eagles if the predicted probabilities were ^0.5. Resui.ts Nest-tree Characteristics. Nest construction, in- cubation, or brood rearing was observed at 21 (87.5%) of the 24 measured nest trees in 1994. The remaining three nest trees were used in 1993, but not in 1994. Nest trees were taller, had more unobstructed branches, and a greater arc of acces- sibility than reference trees (Table 2). Mean nest- tree DBH was more than twice that of reference trees. Twenty-two of 24 (91.7%) nest trees versus only 14 of 24 (58.3%) reference trees had a >270° arc of accessibility. Nest-tree species included Tamarindus indica {N = 7), Cordyla madagascariensis {N = 4), Adansonia sp. {N — 2), Colvillea racemosa {N — 2), Neobeguea mahafaliensis {N = 2), Acacia sp. {N = 1), Albizia greveana {N = 1), Alleanthus greveanus {N =1), Foe- tidia sp. {N = 1), Pandanus sp. {N = 1), and un- identified {N = 2). T indica was, the most frequent- ly recorded species of nest reference tree {N = 6). Its proportion among nest trees (29.2%) was not different from its proportion among reference trees (20.8%) (x^ - 0.44, df = I, P = 0.51). Pro- portions of nest trees and reference trees in each growth form class were similar (x^ — 4.58, df — 4, P — 0.33). Eight of the nest trees (33.3%) and three (12.5%) of the reference trees were decidu- ous (x^ = 2.95, df = I, P= 0.09). Eish-eagle nest-tree use was positively associated with tree height, producing a logistic regression model of 0 = 1/1+ exp 5.52 - X 0-38x, where 0 is the probability of fish-eagle use and x, is the height of tree i. This model correctly classi- fied 83.3% of 48 trees measured. Nest-site Characteristics. Number of shoreline perch trees was greater at nest sites than at random sites (Table 3). There was a positive relationship 290 Berkelman et al. VoL. 36, No. 4 Table 2. Characteristics of Madagascar Fish-Eagle nest trees, perch trees, and paired reference trees in the region of Antsalova, Madagascar in 1994. Variable Nest Trees (V = 24) X ± SE (Range) Paired Reference Trees (N = 24) X ± SE (Range) pa Perch Trees {N= 29) X ± SE (Range) Paired Reference Trees (N = 29) X ± SE (Range) pa DBH (cm) 87.8 ± 11.8 38.4 ± 4.2 <0.001 65.3 ± 7.2 36.9 ± 3.3 <0.001 (29-245) (22-114) (27-270) (21-120) Height (m) 18.7 ± 0.8 10.5 ± 0.9 <0.001 16.7 ± 0.8 9.8 ± 0.4 <0.001 (10.7-25.9) (5.0-23.3) (9.4-30.3) (4.9-15.8) No. of branches’^ 5.5 ± 0.7 3.2 ± 0.8 0.021 7.9 ± 1.2 1.8 ± 0.4 <0.001 (1-14) (0-19) (2-39) (0-15) Arc of accessibility 346.7 ± 5.4 260.2 ± 25.0 <0.001 336.7 ± 7.1 231.4 ± 21.4 <0.001 (265-360) (0-360) (190-360) (0-360) ■* Wilcoxon signed-ranks test significance level. Number of branches in the tree canopy >5 cm in diameter and unobstructed for 1 m above and below. Arc (0°-360°) that was unobstructed by other trees ^10 m of the trunk and ^3 m below the crown (Buehler et al. 1992). between fish-eagle nest-site use and the number of shoreline perch trees. The model was 9 = 1 1 + exp 3.49 - 2 where 0 is the probability of fish-eagle use and is the number of perch trees within a 1.25 ha (50 m wide) shoreline section centered on the point on the shoreline nearest nest tree i. Correct clas- sification of fish-eagle use for this model was 72.9% of 48 sites. Minimum distance to human distur- bance, minimum distance to nearest road, mini- mum distance to nearest building, and minimum distance to nearest fish-eagle nest did not differ between nest sites and random sites (Table 3) . The proportion of nest sites in each 5 m canopy height interval did not differ between nest sites and ran- dom sites (x^ = 4.93, df = 4, P = 0.30). Mean distance to water of nest trees was 70.8 m (SE = 12 6, range = 6.8-199.2 m). Perch-tree Characteristics. Perch trees were larg- er (DBH and height), had more unobstructed branches, and had a greater arc of accessibility than reference trees (Table 2). Twenty-six of 29 (89.7%) nest trees versus only 16 of 29 (55.2%) reference trees had a >270° arc of accessibility. Perch-tree species included Colvillea racemosa (N — 5), Ficus cocculifolia (N = 4), Neobeguea mahafal- lensis (N = 3), Tamarindus indica (N = 3), Albizia lebbeck (N — 2), Borassus madagascariensis {N —2), Cordyla madagascariensis (N = 2), Acacia sp. {N ~ 1), Adansonia sp. (N = 1), Cedrelopsis grevei {N ~ 1), Pandanus sp. (N — 1), Raphia sp. {N = 1), and unidentified {N — A). T. indica was the most fre- quently recorded perch reference tree species {N = 10). Its proportion among perch trees (10.3%) was smaller than among reference trees (48.3%) (x2 = 5.96, df = 1, P = 0.02). Perch trees and reference trees had similar growth forms (x^ = 8.04, df = 4, P = 0.09). Pro- portion of deciduous trees among perch trees (34.5%) was greater than among reference trees (10.3%) (x^ = 4.86, df = 1, P = 0.03). There was a positive association between fish-ea- gle perch-tree use and tree height, producing a lo- gistic regression model of 0 = 1 1 + exp 8.68 2 where 0 is the probability of fish-eagle use and is the height of tree i. This model correctly classi- fied 84.5% of 58 trees measured. Disclissicjn Nest-tree Use. Madagascar Fish-Eagles used nest trees that were taller and had a greater DBH, more unobstructed branches, and a greater arc of acces- sibility than reference trees. The suhstantial differ- ence between nest trees and reference trees in mean height and DBH suggests that the fish-eagle selects nest trees from among the largest trees available near water. By placing its nests in the tops December 2002 Madagascar Fish-Eagi.e Nests 291 Table 3. Characteristics of Madagascar Fish-Eagle nest sites {N = 24) and paired reference sites (N — 24) in the region of Antsalova, Madagascar in 1994. Variable Nest Sites x± SE (Range) Paired Random Sites X ± SE (Range) pa Minimum distance to human disturbance'^ (km) 0.8 ± 0.2 0.9 ± 0.1 0.742 (0-2.8) (0-2.8) Minimum distance to building (km) 1.8 ± 0.4 1.8 ± 0.3 0.814 (0.1-7.7) (0-5.6) Minimum distance to road (km) 1.7 ± 0.4 1.3 ± 0.3 0.055 (0-8.4) (0-5.4) Minimum distance to fish-eagle nest (km) 4.8 ± 0.9 4.3 ± 0.9 0.104 (1.3-20.3) (0.4-20.1) Number of perch trees'^ 30.8 ± 2.3 16.6 ± 1.9 <0.001 (10-53) (0-33) Wilcoxon signed-ranks test significance level. '’Human disturbances included agricultural clearings, rice paddies, villages, tombs, and fishermen’s camps. Number of perch trees within a 1.25 ha (50 m wide) shoreline section centered on the point on the shoreline nearest the nest tree We considered trees that we estimated to have a height ^6.1 m and >30° accessibility from the shoreline to be perch trees. of these trees, it maximizes accessibility and visibil- ity for foraging and territorial defense. These re- sults were consistent with those reported for other nesting Haliaeetus species (McEwan and Hirth 1979, Andrew and Mosher 1982, Anthony and Isaacs 1989, Shiraki 1994). Nest-site Use. Number of shoreline perch trees was the only variable that differed between nest sites and random sites. This suggests that the Mad- agascar Fish-Eagle, like the Bald Eagle (Chandler et al. 1995), may avoid areas without a sufficient number of foraging perches. Perch-tree Use. Perch trees were larger in height and DBH, and had more unobstructed branches, and had a greater arc of accessibility than refer- ence trees. Such trees probably have greater access and provide better visibility over water than other trees. This is consistent with Bald Eagle perch-tree use (Stalmaster and Newman 1979, Steenhof et al. 1980, Buehler et al. 1992). Madagascar Fish-Eagle perch trees were more often deciduous than ref- erence trees. In contrast with the nest-tree results, the fish-eagles in this study appeared to avoid T. indica for perching. T. indica is evergreen and often has a dense crown; therefore fish-eagles may use this species less often for perching than leafless trees or snags. Model Applications. The models we developed may be used to identify Madagascar Fish-Eagle nesting and perching habitat along lakes, rivers, and estuaries in western Madagascar. They do not apply to a sub-population of at least 16 fish-eagle pairs that nest on offshore islands at the north end of the species’ range (Rabarisoa et al. 1997). Al- though our sample size was limited, the 29 breed- ing sites sampled represent 29.3%, of the 99 known remaining Madagascar Fish-Eagle breeding sites (Rabarisoa et al. 1997). Bald Eagle management guidelines recommend conserving mature forest around existing and potential nest sites (Anthony et al. 1982, Wood et al. 1989). We offer guidelines that are more specific to the range of tree sizes and densities found in the tropical dry forest and sa- vanna habitats that surround the lakes where Mad- agascar Fish-Eagles occur. We recommend that areas with a >32 /ha density of trees >15 m tall should receive high priority for Madagascar Fish-Eagle conservation. Probability that a shoreline tree would be used by Madagascar Eish-Eagles for nesting or perching can be calcu- lated by inserting tree height into the correspond- ing logistic equation (Fig. 1 ) . Similarly, number of perch trees along a 1.25 ha (250 X 50 m) shoreline section can be used to estimate the probability that Madagascar Fish-Eagles will use the shoreline sec- tion for nesting (Eig. 1). These models are best used under the conditions that were present dur- ing this study (e.g., same eagle population density, same time of year) and apply to eagles nesting on lakes, rivers, and estuaries. Presence of tall trees close to shoreline is the best predictor of Madagascar Eish-Eagle nest-site use. The eagles often used the tallest trees near water both for nesting and for foraging perches. Rabarisoa et al. (1997) conducted Madagascar Fish-Eagle surveys from 1991-95, and found areas Probability of Use 292 Berkelman et al. VOE. 36, No. 4 Nest-tree Height (m) Perch-tree Height (m) 0 S 10 15 20 25 30 35 40 45 50 No. of Perch Trees Figure 1. Probability of Madagascar Fi.sb-Eagle use of iiesl trees, perch trees, and nest sites as a I’unclion of nest- tree height (A), perch-tree height (B), and number of shoreline perch trees (C), in the region ol Antsalova Mada- gascar, 1994. Probabilities were calculated by inserting dilferent values of the explanatory variable (tree height or number of perch trees) into the equation resulting from stepwise logistic regression analysis. Dec.ember 2002 Madagascar Fish-Eagle Nests 293 with dense forest adjacent to water that were un- occupied by fish-eagles. Watson et al. (1996) are developing means to augment the hsh-eagle pop- ulation and seek areas of unoccupied fish-eagle habitat where young eagles may be released. Our models may be used both to identify areas of suit- able, but unoccupied, fish-eagle habitat and high conservation priority areas of occupied habitat. The Tsimembo Forest surrounding Lakes Befo- taka, Soamalipo, and Ankerika, where the highest density of fish-eagles is found (Rabarisoa et al. 1997), should receive highest conservation priority. The human population density around the lakes was low until recent years when large numbers of fishermen began to migrate to the region (Watson and Rabarisoa 2000). Increased harvesting of tall shoreline trees by migrant fishermen will have a negative impact on the fish-eagles. People use the tallest trees available for dugout canoes and build- ing materials (Watson and Rabarisoa 2000) and may prevent regeneration of tall trees by harvest- ing large amounts of fuel wood to preserve fish by smoke drying. Deforestation probably has already substantially reduced the amount of fish-eagle hab- itat available, and as the human population contin- ues to increase, available habitat will continue to decrease unless steps are taken to conserve fish- eagle habitat. At^KNOWI TDGMENTS The Peregrine Fund provided financial and logistical support for this research. We thank C. Razafimahatratra, G. Raoelison, J. Mampiandra, and L. Kalavah for help with data collection. Thanks to J. Rajesy, R. Rabarisoa, R. I.ewis, P. Ravonjiarisoa, and M. Razafindrakoto for ad- ministrative and logistical support in Antananarivo and in the field. We thank C.A. Haas, J.J. Ney, R.G. Oderwald, D.F. Stauffer, R. Thorstrom, R. Tingay, and A.R. Harmata for comments on the manuscript. f ar ERA i t] RE Cited AciRESTi, A. 1990. Categorical data analysis. John Wiley and Sons, New York, NY U.S.A. Andrew, J.M. and J.A. Mosher. 1982. Bald Eagle nest site selection and nesting habitat in Maryland. J. Wildl. Manage. 46:383-390. Anthony, R.G. and F.B. Isaacs. 1989. Characteristics of Bald Eagle nest sites in Oregon. J. Wildl. Manage. 53: 148-159. , R.L. Knight, G.T. Ai.ien, B.R. McClelland, and J.I. Hodges. 1982. Habitat use by nesting and roosting Bald Eagles in the Pacific Northwest. Trans. N. Am. Wildl. Nat. Resour. Conf. 47:382-390. Buehler, D.A., S.K. Chandler, T.J. Meicsmann, J.D. Fras- er, AND J.K.D. Seegar. 1992. Nonbreeding Bald Eagle perch habitat on the northern Chesapeake Bay. Wil- son Bull. 104:540-545. Chandler, S.K., J.D. Fraser, D.A. Buehler, andJ.KD. Si:egar. 1995. Perch trees and shoreline development as predictors of Bald Eagle distribution on Chesa- peake Bay. y. Wildl. Manage. 59:325-332. Donque, G. 1972. The climatology of Madagascar. Pages 87-144 in R. Battistini and G. Richard-Vindard [Eds ], Biogeography and ecology of Madagascar. Dr. W.Junk B.V., The Hague, Netherlands. Hosmer, D.W., Jr. and S. Lemeshow. 1989. Applied lo- gistic regression. John Wiley and Sons, New York, NY U.S.A. Keister, G.P., Jr. and R.G. Anthony. 1983. Characteris- tics of Bald Eagle communal roosts in the Klamath Basin, Oregon and California. J. Wildl. Manasce. 47 1072-1079. Langrand, O. and B.-U. Meyburg. 1989. Range, status, and biology of the Madagascar Sea-Eagle Haliaeetus vociferoides. Pages 269-277 in B.-U. Meyburg and R.D. Chancellor [Eds.], Raptors in the modern world. WWGBP, Berlin, Germany. McEwan, L.C. and D.H. Hirth. 1979. Southern Bald Ea- gle productivity and nest site selection. J. Wildl. Man- age. 43:585-594. Meyburg, B.-U. 1986. Threatened and near-threatened birds of prey of the world. Birds Prey Bull. 3:1-12. Nelson, R. and N. Horning. 1993. AVHRR-LAC esti- mates of forest area in Madagascar, 1990. Int.J. Remote Sens. 14:1463—1475. Newton, I. 1979. Population ecology of raptors. Buteo Books, Vermillion, SD U.S.A. Rabarisoa, R., R.T. Watson, R. Thorstrom, and J. Ber- kelman. 1997. Status of the Madagascar Fish-Eagle Haliaeetus vociferoides in 1995. OVncA 68:8-12. Shiraki, S. 1994. Characteristics of White-tailed Sea-Eagle nest sites in Hokkaido, Japan. Condor 96:1003-1008 Stalmaster, M.V. and J.R. Newman. 1979. Perch-site pref- erences of wintering Bald Eagles in northwest Wash- ington./. Wildl. Manage. 43:221-224. Steenhof, K., S.S. Berunger, and L.H. Fredrickson. 1980. Habitat use by wintering Bald Eagles in South Dakota. J. Wildl. Manage. 44:798-805. Watson, R.T. and R. Rabarisoa. 2000. Sakalava fisher- men and Madagascar Fish-Eagles: enhancing tradi- tional conservation rules to control resource abuse that threatens a key breeding area for an endangered eagle. Ostrich 71:2-10. , S. Razaeindramanana, R. Thorstrom, and S. Ra- FANOMEZANTSOA. 1999. Breeding biology, extra-pair birds, productivity, siblicide, and conservation of the Madagascar Fish-Eagle. Ostrich 70:105-111. , S. Tiiomsett, D. O’Daniel, and R. Lewis. 1996 Breeding, growth, development, and management of the Madagascar Fish-Eagle {Haliaeetus vociferoides) J Raptor Rei 30:21-27. Wood, P.B., T.C. Edwards, Jr., and M.W. Collopy. 1989. Characteristics of Bald Eagle nesting habitat in Flori- da. /. Wildl. Manage. 53:441-449. Received 30 November 2001; accepted 9 July 2002 J. Raptor Res. 36(4);294-299 © 2002 The Raptor Research Foundation, Inc. USE OF VEGETATIVE STRUCTURE BY POWERFUL OWLS IN OUTER URBAN MELBOURNE, VICTORIA, AUSTRALIA— IMPLICATIONS FOR MANAGEMENT Raylene Cooke and Robert Wallis^ School of Ecology and Environment, Deakin University, Warrnambool Campus, Warrnambool, 3280 Australia John White School of Ecology and Environment, Deakin University, Melbourne Campus, Burwood, 3125 Australia Abstract. — The Powerful Owl (Ninox strenua) is Australia’s largest owl and is considered of least concern nationally. Although a number of studies have reported on the ecology of Powerful Owls inhabiting forests, few have focused on these owls living in urban areas. We report on the characteristics of different roost trees used by Powerful Owls in a continuum of habitats from urban Melbourne to the more forested outskirts. Records of weather conditions and daily temperatures were also analyzed to deter- mine whether the owls were selecting particular roost trees for specific climatic conditions. We found that roost-tree height and perch height was highly correlated, with the owls always roosting in the top one-third of the tree, regardless of the tree height. As ambient temperatnre increased perch height decreased, and vice-versa, but owls always roosted in the top one-third of the roost tree. Powerful Owls did not simply move up and down the one tree, but moved to more suitable trees according to the weather conditions. Hence, the species requires a structurally heterogeneous habitat to provide roost trees for different temperatures. Eurtherraore, successful management of this species in the future will require the protection of structurally diverse vegetation. Key Words: Powerful Owl, Ninox strenua; disturbance, management, temperature, urbanization-, vegetation struc- ture. USO DE LA ESTRUCTURA VEGETATIVA POR NINOX STRENUA EN EXTERIORES URBANOS DE MELBOURNE, VICTORIA, AUSTRALIA— IMPLICACIONES PARA EL MANEJO Resumen. — Ninox strenua es el biiho mas grande de Australia y es considerado nacionalmente de rnenor interes. Aunque un numero de estudios se han concentrado en su ecologia en bosques, pocos se han enfocado sobre los que habitan en areas urbanas. Reportamos las caracterlsticas del uso de diferentes arboles percha utilizados por Ninox strenua en un continuum de habitats desde el Melbourne urbano hasta los alrededores mas boscosos. Adicionalmente se analizaron los registros de condiciones climaticas y temperaturas diarias para determinar si los buhos estaban seleccionando arboles percha particulares debido a condiciones climaticas especificas. Encontramos que la altura de los arboles percha y la altura de la percha utilizada estaba altamente correlacionados con el uso del tercio mas alto del arbol, sin tener en cuenta la altura del arbol. Cuando la temperatura ambiente incrementaba la altura de la percha decrecia, y viceversa, pero los buhos siempre percharon en el tercio mas alto del arbol percha. Los buhos no se movieron simplemente hacia arriba y ab:qo del arbol, sino que se movieron a arboles mas adecuados de acuerdo a las condiciones climaticas. Por lo tanto, la espccie requierc un habitat estruc- turalmente heterogeneo que provea arboles perchas para diferentes temperaturas. Ademas de esto, el manejo exitoso de esta especie en el futuro requiere de la proteccion de vegetacion estructuralmente diversa. [Traduccion de Cesar Marquez] The Powerful Owl {Ninox strenua) is the largest Australian owl. The male is slightly larger than the female, growing to a length of 65 cm with a mass ^ E-mail address: rwallis@deakin.edu.au of up to 1700 g (Higgins 1999). The Powerful Owl is a nocturnal predator, with a diet consisting al- most exclusively of medium-sized, arboreal, mar- supial prey (Webster et al. 1999, Cooke et al. 2002). The Powerful Owl is classified nationally as of 294 December 2002 The Powfrfue Owe in Urban Environments 295 “least concern” (rated nationally of conservation significance, but at the lowest level, Garnett and Crowley 2000) , occurring at low densities in south- eastern continental Australia. Within the state of Victoria the species is listed as endangered (De- partment of Natural Resources and Environment, Victoria 1999) and threatened within the Greater Melbourne Area (Mansergh et al. 1989). Estimates of population numbers in the state of Victoria are less than 500 pairs across the state (Garnett and Crowley 2000). The Powerful Owl was once considered to be a specialist in ecological terms because of its appar- ently restricted habitat and dietary requirements (Fleay 1968, Seebeck 1976, Roberts 1977), indicat- ing that it is vulnerable to habitat modihcation and that it has specific conservation needs. Recent studies, however, have contested these earlier find- ings and consequently have questioned the degree to which the Powerful Owl is vulnerable to habitat modification and disturbance (Debus and Chafer 1994, Kavanagh and Bamkin 1994, Pavey et al. 1994, Cooke et al. 1997, Cooke et al. 2002). Urban and suburban areas surrounding Mel- bourne have been mostly cleared throughout the past 100 years, with only small patches of remnant vegetation remaining. Surprisingly, Powerful Owls still remain in some urban areas, with one known breeding pair located only 18 km from central Mel- bourne. Powerful Owls have also been recorded living in close proximity to other Australian cities, including Brisbane (Pavey et al. 1994, Pavey 1995) and Sydney (Rose 1993). Tittle research has been undertaken to determine the resources these owls require for long-term survival in urban environ- ments. Here, we describe roost tree characteristics and features of roosts used in urban and suburban areas by Powerful Owls. Results from this study are then used to identify management options for Pow- erful Owls in urban areas. The results of this study may also provide valuable information for the fu- ture management of other top-order raptors with similar ecological attributes in urban areas. Study Areas During this study, we examined how Powerful Owls used the structure of vegetation in a continuum of en- vironments ranging from urban Melbourne (two sites), through the urban fringe (three sites), and into more forested areas (one site). Each site was selected on the basis that it had a confirmed breeding pair of owls pres- ent for several years. The two sites located closest to Melbourne were the Yarra Valley Metropolitan Park (100 ha) and Warrandyte State Park (586 ha), which were urban parklands man- aged for public recreation and 18 km and 24 km north- east of central Melbourne, respectively. Both parks have been extensively modified in the past and now consist ot riparian areas and the occasional patch of remnant trees surrounded fry a matrix of revegetated woodlands. The next three sites along the continuum were One Tree Hill Reserve (143 ha), Smiths Gully (2.4 ha), and Steels Creek (21600 ha). One Tree Hill Reserve and Smiths Gully are both located 35 km from central Mel- bourne while Steels Creek is located 65 km from Mel- bourne. These three sites are all dry, open forests and consist primarily of different Eucalyptus spp. as upper can- opy trees with Acacia spp. dominating the middle story. These three sites are also regularly visited by people and also show signs of disturbance. The sixth site along our continuum was Toolangi State Forest (35 000 ha), which is located 80 km northeast of Melbourne. This forest is a relatively undisturbed wet sclerophyll forest dominated by mountain ash {Eucalyptwi regnans). Middle story species are less common in this area; however, the understory is dominated by various ferns and bracken. Methods A total of 1300 day visits were made to the six study sites between 1996-99. During these visits the roost tree in which the Powerful Owl was located was recorded Roost trees were those in which Powerful Owls spent time during the daylight hours. Here, we examined the different roost trees used by the Powerful Owl at each of the study sites and the char- acteristics of each tree used. These included the species of tree, tree height, and the diameter at breast height (DBH). Records of weather conditions and daily temper- atures were also analyzed to determine whether the owls are selecting particular roost trees for specific climatic conditions. Each study site was visited at least once weekly over a 4-yr period and each roost tree was examined for the presence of the Powerful Owl or evidence that an owl had used the tree recently. Evidence of usage included fresh whitewash (excreta) or regurgitated food pellets Temperature and weather conditions were noted, regur- gitated food pellets were collected and, in situations where the Powerful Owl was using the roost tree, the perch height was measured using a clinometer. Resutts The Powerful Owls used 179 individual roost trees at the six study sites. Twenty different tree species were used as roost trees. The main trees used for roosting were Eucalyptus spp. (54%), Aca- cia spp. (18%), and Leptospermum spp. (15%). Oth- er roost trees were hazel pomaderris {Pomaderns aspera), the introduced Monterey pine {Pinus ra- diata) , cherry ballart (Exocarpos cupressiformis ) , Christmas bush {Prostanthera lasianthos), the non- 296 Cooke et al. VoL. 36, No. 4 Table 1. Roost-tree characteristics at each of the six study sites. Values represent mean ± 1.96 SE. Site N Tree Height ( m) DBH (cm) Perch Height ( m) Yarra Valley Metropolitan Park 22 15.7 ± 2.2 55.0 ± 12.2 10.2 ± 1.9 Warrandyte 29 13.3 ± 2.3 40.3 ± 11.3 9.6 ± 2.0 One Tree Hill 22 16.2 ± 1.9 48.8 ± 10.0 12.2 ± 1.9 Smiths Gully 24 12,7 ± 1.8 37.1 ± 9.7 8.1 ± 1.1 Steels Creek 23 16.1 ± 2.2 38.5 ± 5.0 10.3 ± 1.9 Toolangi 59 13.0 ± 2.1 49.7 ± 9.6 11.2 ± 1.8 Pooled data 179 14.4 ± 0.9 45.6 ± 4.4 10.4 ± 0.8 indigenous sweet pittosporum (Pittosporum undu- latum), and swamp paperbark {Melaleuca ericifolia). To determine whether the dimensions of roost trees varied between sites we compared the tree height, roost height, and DBH of roost trees at each site (Table 1). Roost tree heights were not different among the six study sites (T 5 173 = 1.856, P = 0.104), with the mean height of roost trees being 14.4 m ± 0.9 m (±1.96 SE). Perch heights between the six study sites also did not differ sig- nificantly (fy 173 = 1.643, P = 0.15), with the mean perch height being 10.4 m ± 0.8 m (±1.96 SE). There was also no significant difference in the DBH of the roost trees between the six study sites (^ 5 ,i 7 s “ 1.52, P — 0.186). Overall, the mean DBH was 45.6 cm ± 4.4 cm (mean ± 1.96 SE). These results suggest that the trees used for roosting have similar physical dimensions at each site even though the tree species may differ between sites. 0.90 01 0.85 '5 z $ 0.80 o 0.75 % 0.70 Q. & 0. 0.65 n " 0.60 i 0.55 £ ® 0.50 0 . YVMP One Tree Hill Steels Creek Warrandyte Smiths Gully Toolangi Site Figure 1. Perch heights as a proportion of tree heights (mean ± 1 and 1.96 SE) at each of the six study sites. Plots with the same letters indicate homogeneous groups as revealed by the Student Newman-Keuls (P < 0.05). YVMP = Yarra Valley Metropolitan Park. Given the variety of tree species used by the owls for roosting, we decided to determine whether the roost trees were being used in a similar fashion among sites. Specifically, the relationship between perch height and tree height was examined. Over- all, perch height was positively correlated with tree height (r = 0.91, P < 0.001, N = 179). Hence, although the species of roost tree varied, the owls tended to perch toward the top of the selected roost tree. The perch height as a proportion of tree height varied significantly between sites (E 5 173 = 17.76, P < 0.001). Perch heights at the Yarra Valley Metropolitan Park, Warrandyte State Park, Smiths Gully, and Steels Creek were lower within the roost trees than those in Toolangi State Forest (Student Newman-Keuls test, P < 0.05; Fig. 1). Although the mean perch height as a proportion of the tree height varied among sites, roost tree height was a predictor of perch height within each site (Fig. 2). High values at all sites (except Smiths Gully) suggest that there was a strong and consistent relationship between perch height and tree height (Table 2). To further understand this relationship, a com- parison was made between perch height and dif- ferent temperature and weather conditions. When there was no precipitation a strong negative asso- ciation between temperature and perch height was found (Table 3), with the owls at all sites choosing lower perch heights as the temperature increased. On days where rainfall occurred this trend was less evident, with the owls at most sites showing no con- sistent association between perch height and tem- perature (Table 3). Discussion The Powerful Owls inhabiting the six study sites used 179 roost trees. Of these, 87% were from only Perch Height (m) Perch Height (m) Perch Height (m) December 2002 The Powereul Owl in Urban Environments 297 Yarra Valley Metropolitan Park Warrandyte State Park Perch height = -1.78 + 0.76 * tree height Perch height = -1.05 + 0.80 * tree height One T ree Hill Reserve Smiths Gully Perch height » -2.48 + 0.90 * tree height Perch height = 2.51 + 0.44 * tree height Tree Height (m) Steels Creek Perch height = -1.66 + 0.74 * tree height Toolangi State Forest Perch height = -0.38 + 0.89 * tree height Figure 2. The relationship between perch height and tree height at each of the six study sites. Lines indicate the regression lines with 95% confidence limits. 298 Cooke ei ai.. VoL. 36, No. 4 Table 2. Regression results of the relationship between perch height and tree height at all sites. Site B? df F P Yarra Valley Metropolitan Park 0.844 1, 20 114.71 <0.001* Warrandyte 0.859 1,27 171.34 <0.001* One Tree Hill 0.825 1, 20 100.33 <0.001* Smiths Gully 0.463 1, 22 20.84 <0.001* Steels Creek 0.718 1, 21 56.89 <0.001* Toolangi 0.980 1, 57 2792.09 <0.001* * Represents a significant relationship between perch height and tree height. three different genera, Eucalyptus (54%), Acacia (18%), and Leplospermum (15%). The other 13% of roost trees consisted of a variety of genera that were infrequently used by Powerful Owls. Overall, Powerful Owls roosted in 20 different tree species at the six study sites and in most cases the roost trees used were the most common species at the specific study site. This indicates that the Powerful Owls in the Yarra Valley corridor are probably us- ing abundant and available tree species rather than selecting less common species. Roost tree characteristics such as height, perch height, and DBH did not differ between the six sites. Roost tree height and perch height, however, were highly correlated, indicating a direct relation- ship between the height of the roost tree and the perch height. Powerful Owls observed at all six sites generally roosted in the top one-third of the roost tree, regardless of the tree height. Overall, these results suggested that Powerful Owls roosted in a number of tree species and they were most likely found in the most common tree species. It is probable that Powerful Owls are gen- eralists in terms of the tree species in which they will roost. The fact that the roost tree characteris- tics (e.g., perch height, DBH) were similar at all sites suggested that there was some degree of se- lection of individual trees that offer optimal roost characteristics. This was particularly highlighted by the relatively small number of roost trees used at each site compared with the number of trees avail- able. When temperature and weather conditions were considered in relation to roost tree usage, the re- sults suggested that as temperature increased perch height decreased, and vice-versa. On hot days. Powerful Owls were roosting lower in shadier sites and on cooler days they roost at higher levels, possibly to absorb sunlight. However, independent of the height of the roost tree the Powerful Owls still roosted in the top one-third of the roost tree. This result suggests that they require habitats with a large degree of structural variation to provide roost trees for different temperatures. The choice of roost trees used by the Powerful Owls in clear and rainy conditions was also exam- ined. The results showed that there was no signif- icant difference in the perch height used by the Powerful Owls at five of the six sites on wet days. Steels Creek was the only exception to this pattern. At most sites, Powerful Owls roosted in slightly low- er trees on rainy days. However, at Steels Creek the Powerful Owls actually roosted in taller, canopy trees on precipitation days. Thus, it would appear that the height at which Powerful Owls roost in different weather conditions was not as important as the amount of canopy cover provided by the specific roost tree. Results from this study also suggest that the structural diversity within a site is important, given Table 3. Correlation rc.sults from comparisons of perch height and temperature on days with and without precipi- tation. Site No PrECIEI J AI ION PRKCtEITAlION r-VAt.t.iE P-VAITIE r-VAt.UE P-VAl.UE Yarra Valley Metropolitan Park -0.80 <0.001* -0.03 0.875 Warrandyte -0.27 0.021* 0.06 0.714 One Tree Hill -0.87 <0.001* -0.06 0.065 Smiths Gully -0.68 <0.001* -0.39 0.006* Steels Greek -0.45 <0.001* 0.62 <0.001* Toolangi -0.59 <0.001* -0.18 0.112 * Represents a significant relationship between perch height and temperature. December 2002 The Powerful Owl in Urban Environments 299 that the Powerful Owls may use trees of different heights to regulate their temperature in relation to climatic conditions. Unfortunately environmental change accompanying urbanization often results in less structural diversity in vegetation, which can mean that Powerful Owls have less choice in suit- able thermal environments. What effect loss of structure will have on survival and reproduction is largely unknown, but it may in part explain why the Powerful Owl is rarely found in highly-urban- ized areas. This information is important for future man- agement of the Powerful Owl because it suggests that this species does not simply move to higher or lower branches in the one tree; rather, it moves to an alternative roost tree with more suitable struc- tural characteristics when it changes heights. Therefore, management of the vegetation in the urban areas must ensure that there is structural diversity in the vegetation. Currently, the focus of vegetation management for the Powerful Owl has been on maintaining old eucalypts (canopy layer) . However, this may not provide for the structural resource requirements of this species. Vegetation management for the Powerful Owls should, there- fore, be expanded to include the obviously impor- tant mid-story species such as Acacia and Leptosper- mum. Acknowled gments We thank Alan Webster for his on-site help, particularly for providing historical records, and for continual sup- port and other assistance. We also thank the Holsworth Wildlife Research Fund, the M.A. Ingram Trust, Birds Australia, and Deakin University for providing financial support. We also wish to acknowledge the constructive suggestions provided by referees of this manuscript. Literature cited Cooke, R., R. Wallis, A. Webster, and J. Wilson. 1997. Diet of a family of Powerful Owls Ninox strenua from Warrandyte, Victoria. Proc. R. Soc. Vic. 107:1—6. , R. Wallis, and A. Webster. 2002. Urbanization and the ecology of Powerful Owls Ninox strenua in out- er Melbourne, Victoria. Pages 100-106 in I. Newton, R. Kavanagh, J. Olsen, and I. Taylor [Eds.], Ecology and conservation of owls. CSIRO Publishing, Mel- bourne, Australia. Debus, S.J.S. and C.J. Chafer. 1994. The Powerful Owl Ninox strenua in New South Wales. Aust. Birds (Sup- plement) 28:20-38. Department of Natural Resources and Environment, Victoria, 1999. Powerful Owl Ninox strenua. Flora and fauna guarantee act. Department of Natural Resourc- es and Environment, Melbourne, Australia. Fleay, D. 1968. Nightwatchmen of bush and plain. Jaca- randa Press, Brisbane, Australia. Garnett, S.T and G.M. Crowity. 2000. The action plan for Australian birds. Canberra: Environment Austra- lia, Australian Government. Canberra, Australia. Higgins, PJ- 1999. Handbook of Australian, New Zealand and Antarctic birds. Vol. 4. Oxford Univ. Press, Mel- bourne, Australia. Kavanagh, R.P. and K.L. Bamkin. 1994. Distribution of nocturnal forest birds and mammals in relation to the logging mosaic in south-eastern New South Wales Biol. Conserv. 71:41-53. Mansergh, I.C., C. Beardsell, S. Bennett, R. Brereton, K. O’Conner, K. Sandiford, and M. Schulz. 1989. Report on the sites of zoological significance in the Upper Yarra Valley (western section) and Dandenong Ranges. Arthur Rylah Institute for Environmental Re- search, Tech. Rep. Ser. No. 90, Melbourne, Australia Pavey, C.R. 1995. Food of the Powerful Owl Ninox strenua in suburban Brisbane, Queensland. Emu 95:231-232. , A.K. Smyth, and J. Mathieson. 1994. The breed- ing season diet of the Powerful Owl Ninox strenua at Brisbane, Queensland. Emu 94:278-284. Roberts, G.J. 1977. Birds and conservation in Queens- land. Sunbird 8:73—82. Rose, A.B. 1993. Notes on the Powerful Owl Ninox strenua in New South Wales. Aust. Birds 26:134—136. Seebeck, J.H. 1976. The diet of the Powerful Owl Ninox strenua in western Victoria. Emu 76:167-170. Webster, A., R. Cooke, G. Jameson, and R. Wallis. 1999 Diet, roosts, and breeding of Powerful Owls Ninox strenua in a disturbed, urban environment: a case for cannibalism? Or a case for infanticide? Emu 99:80—83. Received 31 December 2001; accepted 7 August 2002 / Raptor Res. 36(4) :300-308 © 2002 The Raptor Research Foundation, Inc. NEST-SITE SELECTION OF THE CROWNED HAWK-EAGLE IN THE FORESTS OF KWAZULU-NATAL, SOUTH AFRICA, AND TAI, IVORY COAST Gerard Malan^ School of Life and Environmental Sciences, University of Durban-Westville, PB X5400f Durban 4000, South Africa SusANNE Shultz Population and Evolutionary Biology Research Group, School of Biology, Nicholson Building, University of Liverpool, Liverpool L69 3 GS, United Kingdom Abstract. — Structural characteristics of Crowned Hawk-Eagle {Stephanoaetus coronatus) nest sites were compared between forests in KwaZulu-Natal province, South Africa, and the Tai National Park, Ivory Coast, and key features of nesting trees and nest-placement sites were identified. Nest-tree heights and nest heights differed appreciably between three tree groups with the highest being indigenous Tai trees (x tree height = 52 m and x nest height = 36 m, N = 8 nests), followed by exotic eucalyptus (x = 44 and 22 m respectively, N = 10) and indigenous trees (x = 24 and 14 m respectively, N = 17) from KwaZulu-Natal. All the nest trees in Tai were eraergents, whereas 11 of 17 indigenous nest trees and only two of 10 eucalyptus were so in KwaZulu-Natal. Non-emergent eucalyptus nest trees were predom- inantly edge trees that may have provided easier access for flying eagles. Overall, nest forks were more accessible for flying eagles than random forks, although access did not differ between nests located in emergent and non-emergent trees. Crowned Hawk-Eagles transport long sticks and heavy prey items to their nests and access was probably the most critical feature of both the nest tree and placement of the nest. Wildlife managers must, therefore, ensure that lone-standing or emergent trees are cultivated and conserved, and flight paths to nests are kept open to allow Crowned Hawk-Eagles continued and easy access to their nests. Key Words: Crowned Hawk-Eagle, Stephanoaetus coronatus; nest access; emergent trees; nest-site selection; wildlife management. SELECCION DEL SITIO NIDO DEL AGUILA CORONADA EN LOS BOSQUES DE KWAZULU-NATAL, SUR AFRICA, Y TAI, COSTA DE MARFIL Resumen. — Las caracteristicas estructurales de los sitios nido de las aguilas coronadas {Stephanoaetus coronatus) fueron comparadas entre los bosques de la provincia KwaZulu-Natal, Sur Africa, y el parque nacional Tai, Costa de Marfil, y se identilicaron los rasgos claves de los arboles nido y de los sitios de ubicacidn de los nidos. Las alturas de los arboles nido y las alturas de los nidos difirieron apreciable- mente entre arbol y grupo de arboles siendo los mas altos los nativos Tai (x altura del arbol = 52 m y X altura del nido = 36 m, N = 8 nidos), seguido por eucaliptos exoticos (x = 44 y 22 m respectivamente, N ~ 10) y arboles nativos (x = 24 y 14 m respectivamente, N = 17) de KwaZulu-Natal. Los arboles nido de eucaliptos no emergen tes fueron predominantemente arbcdes de borde que podian proveer mas facil acceso a las aguilas en vuelo. En conjunto, las horquelas en nidos fueron mas accesibles para las aguilas que horquetas colocadas al azar, aunque el acceso no dilirio entre los nidos colocados en arboles emergentes y no emergentes. Las aguilas coronadas transportan grandes ramas y presas pesadas a sus nidos y probablementc el acceso fue el rasgo mas critico tanto para la seleccion del arbol nido como para la ubicacion del nido. Los manejadores de vida silvestre deben, por lo tanto, asegurar que arboles aislados o emergentes son cultivados y conservados, y que las vias de vuelo a esos nidos permaneceran abiertas para permitir a las aguilas coronadas continuo y facil acceso a sus nidos. [Traduccion de (iesar Marquez] Tree-nesting raptors select nesting trees and nest sites on the basis of certain structural features that ^ Present address: Department of Nature Conservation, Pretoria Technikon, P.B. X680, Pretoria 0001, South Af- rica; e-mail address: malang@techpta.ac.za hide the nest from potential predators, insulate the nest against adverse weather conditions, place the birds close to their hunting habitats and allow them easy access to the nest (Moore and Henny 1983, Speiser and Bosakowski 1987, Lilieholm et al. 1993, Burton et al. 1994, Selas 1996, Malan and 300 December 2002 Nest-seee Selection oe the Crowned Hawk-Eagle 301 Robinson 2001). Larger raptors often nest in an exposed position that allows easy access to and from the nest to deliver long sticks and heavy prey (Speiser and Bosakowski 1987, Burton et al. 1994). To further facilitate access, large raptors nest in tall emergent trees or large trees with open branch structures that respectively allow them access to the nest both above and within the canopy (Moore and Henny 1983, Burton et al. 1994, Malan and Rob- inson 2001). Unfortunately, these preferences bring large eagles in direct conflict with man as large trees often are selectively harvested for com- mercial and subsistence purposes (Bijleveld 1974, Watson and Rabarisoa 2000). In the predominantly arid South Africa, the de- pletion of indigenous forest habitats and the arriv- al of commercial exotic Eucalyptus, Pinus, Acacia, and Populus trees have had a mixed impact on the abundance and distribution of tree-nesting forest birds (Low and Rebelo 1996, Allan et al. 1997). Although by 1997 the area under commercial pulp- wood and sawlog plantations (15 186 km^) was four times larger than the existing natural forests (Van der Zel 1996, Anon. 1998), the very short rotation intervals of plantations (8-16 yr) do not allow the trees to attain the size necessary to support large stick nests. However, isolated, non-commercial stands of large exotic trees are used for nesting by indigenous birds, including raptors (Steyn 1977, Macdonald 1986, Malan and Robinson 2001). In indigenous forests, the processes of deforestation, forest fragmentation, and the selective removal of big trees from remnant patches alter the size, struc- ture, and availability of indigenous trees for nest- ing (Tarboton and Allan 1984, Allan and Tarboton 1985, Seydack 1995, Vermeulen 1999). In South Africa, the Crowned Hawk-Eagle {Ste- phanoaetus coronatus) has recently become of for- mal conservation concern and the status of the species has changed to near threatened due to the loss of its previously suitable, indigenous nesting habitat to short rotation, exotic plantations (Barnes 2000) . Throughout much of its range, the Crowned Hawk-Eagle breeds in tall evergreen for- ests but can also nest in deciduous forests and woodland-forest mosaics (Tarboton and Allan 1984). These birds prefer large forest trees for nesting and the nest is usually situated in a major fork, 8-30 m above ground but can be as high as 46 m (Steyn 1982, Tarboton and Allan 1984, Brown and Amadon 1989). In South Africa, Crowned Hawk-Eagles nest in exotic eucalyptus and pines, in addition to indigenous trees, consisting of most- ly White Stinkwoods ( Celtis africana) (Tarboton and Allan 1984, Boshoff 1988). The objective of this study was to compare Crowned Hawk-Eagle nesting sites in three tree groups and to use this information to provide re- source managers with silvicultural guidelines for providing nest-placement sites and stands for Crowned Hawk-Eagles. As the forests in the KwaZulu-Natal province. South Africa, are relative- ly small in size and patehy in distribution (Boshoff 1997, Midgley et al. 1997), nest site characteristics of these forests were compared with those from the extensive and continuous, indigenous forest of the Tai National Park, Ivory Coast. These comparisons clarify which structural features are most important when selecting a nest site, and how flexible Crowned Hawk-Eagles are with regard to these characteristics. Second, for KwaZulu-Natal, we ex- amine stand size to determine the minimal habitat requirements and compare topographical features with randomly-selected sites to determine the ea- gle’s selectivity for these features. Study Areas In South Africa, Crowned Hawk-Eagle nest sites were located for study during forest surveys in the KwaZulu- Natal (29°S, 31°E). Indigenous tree stands were surveyed in forests characterized by a 10-25 m high canopy, dis- tinct vegetation strata and numerous dominant tree spe- cies (Low and Rebelo 1996, Midgley et al. 1997). The mean annual rainfall in these forests range from 900- 1500 mm. Indigenous nest trees were sampled in the Or- ibi Gorge, Vernon Crookes, Krantzkloof, Harold John- son, and Dlinza Eorest nature reserves, Ithala Game Reserve, Hluhluwe-Umfolozi Park, all areas managed by the KwaZulu-Natal Wildlife, and the Umgeni Valley Na- ture Reserve and Tanglewood Natural Heritage Site. Ex- otic eucalyptus were surveyed in abandoned plantations, self-sown stands, and planted trees in large domestic gar- dens, but none were known from commercial sawlog or pulpwood plantations. In the Ivory Coast, all nest sites were located within a 50 km^ area around the Station de la Recherche en Ecol- ogie Tropicale (SRET, 7°00'N, 5°50'W) near the western edge of the park. The Tai National Park is the largest continuous lowland forest in West Africa (454 000 ha) and contains the last sizeable protected habitat for a number of Upper Guinea Forest endemics (Gartshore et al. 1995). The mean annual rainfall at the research sta- tion is 1800 mm. The forest was selectively logged in the early 1970s, but the structure is essentially indistinguish- able from a primary forest. The main forest canopy is 30-40 m in height, with emergent trees reaching to over 60 m (Guillaumet 1994). Methods The authors and conservation ofhcers located 27 Growned Hawk-Eagle nests in KwaZulu-Natal and eight 302 Malan and Shultz VoL. 36, No. 4 m Tai. Of the nests in KwaZulu-Natal, 19 were found by listening for the loud queee-queee soliciting calls of nest- lings or by spotting large nests (Steyn 1982). Although this non-systematic search method may bias the sample toward accessible and conspicuous nests (Daw et al. 1998), the nest-site data include nests from a wide selec- tion of forest and nest-tree types. Sampled nest trees were located in indigenous forests, exotic monocultures, and mixed forests, and nest tree species were grouped into either indigenous or eucalyptus stands. The following nest-tree characteristics were recorded from each nest site: the tree diameter at breast height (14 m, DBH) of all stems >22 cm in diameter; tree height (T), nest height (N), height of the first foliage, and the height of the first side branch (R, irrespective if It was dead or alive) . All height measurements were taken with a clinometer. For trees with buttresses {N = 3), the circumference of the tree, including the buttresses, was measured at breast height, a diameter calculated and di- vided by two as to provide a conservative estimate of DBH. The percent nest height was the proportional dis- tance the nest was placed from the top of the tree ((N/ T)100) and the percent first branch height was the pro- portional height the first branch was placed from the top of the tree ((R/T)100). Nests were classified as being placed within or below the foliage. Each nest was scored as either being posi- tioned in a main fork, against the main branch (i.e., pri- mary axis, mainly vertical), or on a side branch (i.e., sec- ondary axes, mainly horizontal). The number of branches supporting the nest was also counted. Lastly, the nest tree was classified as emergent if the branch and foliage structure protruded above the surrounding for- est. For each nest tree, we identified large and open forks, excluding the nest fork, which could possibly support a Crowned Hawk-Eagle nest that was 2 m wide and 2.5 m deep (Steyn 1982). Each of these forks was categorized as being positioned in a central fork or crotch (i.e., be- tween main branches), as against the main branch, or on a side branch, and numbered from tree bottom to top. A random fork was selected from the category of the fork that supported the nest (i.e., if the nest was against a main branch, the random fork was selected from similar forks). If the random fork was not available from the category that supported the nest, it was selected from a combined sample of the remaining two positions. For each cardinal direction (N, E, S, and W), accessi- bility of the nests and random forks were determined by whether the flying eagle would have a 30 m unobstructed approach on a horizontal plane. For a flying eagle, a flight path to the nest obstructed by foliage and/or branches thus qualified inaccessibility. We sampled characteristics of the three trees closest to the nest tree. The distance from the nest tree to each surrounding tree was measured. We calculated the mean area occupied by the trees by squaring the mean distance of the three trees from the nest tree. We then converted the result to the number of trees per hectare and cal- culated a tree density estimate at the nest tree (Phillips 1959). For KwaZulu-Natal nest trees, each nest was plotted on a 1:10 000 orthophoto (black and white aerial photo- graph with 50 m contour lines) to calculate stand size and shape. From this photo, the maximum length (A) and maximum width (B; perpendicular to the maximum length axis) of the nest stand was measured. All mea- surements from 1:10 000 orthophotos were rounded to 10 m to allow for a 1 mm measurement error. Stand shape (S) was defined as the maximum stand width di- vided by the maximum length, with values ranging from one (square shape) to zero (elongated shape). Nest stand shapes were grouped into square (Shape ^0.5) or elon- gated (Shape <0.5). The surface area of planted forests was calculated from orthophotos by measuring the stand lengths and widths. The surface area of indigenous forest stands could not be calculated from orthophotos because the edges of indigenous forests were not defined clearly and the surface area was, therefore, estimated as S = (A/ 2)(B/2)(tt). From the orthophotos, the nest trees were also categorized as being located on the edge of the stand (i.e., the first tree encountered on the edge of a forest stand). Lastly, distances to water, road, and nearest hu- man habitation were measured from the nest tree and one random tree selected from each nest stand. All data were subjected to the Kolmogorov-Smirnov one-sample test of normality. If the distribution of the data was found to be non-normal, the non-parametric Kruskal- Wallis test or Mann-Whitney Latest was employed The Tukey honest significant difference test for unequal sample sizes was employed to determine which groups are particularly different from each other after a signifi- cant Kruskal-Wallis test was obtained from the Analysis of Variance. The Pearson’s Chi-square test was used to test for patterns with categorical variables. All data were an- alyzed using the Statistica software package (StatSoft 1995) and probability levels were set at a = 0.05. Results Of the 35 Crowned Hawk-Eagle nests analyzed, 10 were located in eucalyptus and 17 in indigenous trees in KwaZulu-Natal, and eight in indigenous trees in Tai. Indigenous nest trees identified in KwaZulu-Natal included Ficus spp. (7), Chrysophyl- lum viridifolium (3), Syzygium cordatum (2), Cussonia spicala, Scholia brachypetala, and Celtis africana. The exotic nest trees were all eucalyptus. Nest trees identified in Tai included Lofira alata, Klainodoxa gabonesis, Ceiba pentandra, and Alstonia boonei. All tree characteristics measured differed among the three tree groups (Table 1). Tai and eucalyptus nest trees were taller than KwaZulu-Natal’s indige- nous trees. Tai nest trees were larger in diameter, and had higher first foliage, side branches and per- cent first side branch heights than trees from KwaZulu-Natal. Nest heights differed among all three tree groups, and the percent nest heights were lower in eucalyptus than Tai nests (Table 1). The number of branches that supported nests did not differ among tree groups (Kruskal-Wallis ^ 2,35 ~ 1.7, P = 0.41), and the eagles used on av- December 2002 Nest-site Sei.ection c:>f i he Crowned Hawk-Eagee 303 Table 1. Crowned Hawk-Eagle nest-tree characteristics of eucalyptus and indigenous tree groups in KwaZulu-Natal and Tai. Means ± one standard deviation and range in parenthesis. Values with the same superscripts indicate values that differed significantly from each other in the three-way comparison. Species Ciass Indigenous KwaZui.u-Natal Eucai.yptus RWAZuriJ-NAIAL Indigenous Tai N Kruskal- Wallis H DBH (cm) 93^' ± 29 126*> ± 32 234^'^ ± 66 35 21.6** (40-151) (83-206) (167-344) Tree height (m) 24ab + 5 44a + g 52'^ ± 14 35 25.7** (T5-34) (34-54) (33-67) Height of hrst 9“’ ± 6 12*’ ± 9 33^>'> ± 7 35 17.8** foliage (m) (1-19) (6-27) (26-42) Height of first side 9“ ± 6 15'’ ± 7 34’’ ± 9 35 17.8** branch (m) (1-20) (6-25) (25-47) Nest height (m) ]4ab + 3 22a<: + 7 ± 10 35 23.4** (7-21) (13-35) (2.5-49) Percent nest height 58 ± 14 52^' ± 13 71" ± 13 35 7.6* (43-93) (36-67) (44-90) Percent first side 36^' ± 21 32b 2: 12 66"’’ ± 13 35 branch height (4-62) (16-47) (44-78) ■^ihc = p <; 0.05, I'ukey tests. * P < 0.05. P < 0.001. erage 3 ± 1 branches (A ± SD, N = 35 trees) to nest on. Twenty-seven of 32 nests (84%) were placed within the foliage (as opposed to under the foliage), and placement of nests in relation to the foliage was not associated with tree groups (x^ 0.7, P = 0.70). In Tai, 63% of the nests {N = 8) were placed on the lowest side branch, more fre- quently than nests in indigenous (13%, N — \1) and eucalyptus trees in KwaZulu-Natal (10%, N = 10, = 9.3, P< 0.01). Of the 35 nests sampled, 20 (57%) were placed m a central fork, nine (26%) on a side branch, and six (17%) against the main branch; nest placement was not associated with tree groups (x^ = 6.3; P — 0.17). Whereas seven (78%) of the nine side- branch nests and four (67%) of the six main- branch nests were from the most abundant fork category, 18 (90%) of the 20 central-crotch nests Table 2. The emergence of Crowned Hawk-Eagle nest trees above the surrounding forest. Emergent Nest Trees Yes No Totai. Indigenous — KwaZulu-Natal 11 6 17 Eucalyptus — KwaZulu-Natal 2 8 10 Indigenous — Tai 8 0 8 Total (Percent) 21 (60) 14 (40) 35 were from trees where this fork type was not the most abundant (x^ = 6.3; P < 0.001). Overall, 22 (63%) of the 35 nests were not placed in the most abundant fork category and nest placement could not be associated with the most abundant fork cat- egory of the different tree groups (x'^ 1.0; P = 0.59). The number of forks per nest tree did not differ among the tree groups (1 1 ± 6 forks in indigenous KwaZulu-Natal trees, 14 ± 8 in eucalyptus, and 7 ± 3 in Tai nest trees; Kruskal-Wallis ~ ^-4? P = 0.16). Whereas the number of central forks per tree did not differ among tree groups (1 ± 1, Krus- kal-Wallis 772,35 = 3.2, P = 0.21), the number of side forks did (indigenous KwaZulu-Natal 5 ± 4 forks, eucalyptus 2 ± 2, and indigenous Tai 3 ± 3; Kruskal-Wallis 772 35 = 6.2, P< 0.05), although the classes were not significantly different from each other (Tukey tests, all P > 0.05). Eucalyptus had more forks against the main branch (11 ± 8) than indigenous trees in KwaZulu-Natal (4 ± 3) and Tai (3 ± 1; Kruskal-Wallis = 9.6, P< 0.01; Tukey tests, all P < 0.05) . All the nest trees in Tai and 11 of 17 indigenous nest trees in KwaZulu-Natal were emergents, whereas only two of 10 eucalyptus protruded above the forest canopy (x^ = 12.2, P < 0.05; Table 2). Overall, nests were more accessible to flying eagles 304 Malan and Shultz VoL. 36, No. 4 Table 3. The number of Crowned Hawk-Eagle nest trees, with flight paths sampled in four cardinal direc- tions, which allowed access to nest and random forks for 17 indigenous and 10 eucalyptus trees in KwaZulu-Natal, and eight indigenous trees in the Tai Forest. Number of Flight Paths None One Two Three Four x^ Nest fork 0 1 12 12 10 Random fork 4 12 8 4 7 18.6** ** p< 0.001. than were random forks (Table 3) . Access to nests could not be associated with their placement above (as emergents) or within the forest canopy (Table 4) . In KwaZulu-Natal, seven of the eight non-emer- gent eucalyptus were located on the edge of the nest stand, whereas, all the non-emergent indige- nous trees were located inside the nest stands (x^ = 7.3, P< 0.01). The mean distance from the nest tree to the nearest three trees did not differ among tree groups (indigenous KwaZulu-Natal 11 ± 6 m, eu- calyptus 12 ± 8 m, and indigenous Tai' 13 ± 5 m; Kruskal-Wallis 7^2,34 ~ 1-0, P = 0.62). The density of trees at nest sites also did not differ among tree groups (all trees 156 ± 175 trees/ha, indigenous KwaZulu-Natal 172 ± 183 trees/ha, eucalyptus 189 ±219 trees/ha, and indigenous Tai 85 ± 62 trees/ ha; Kruskal-Wallis 7^2,34 ^ 1-0, P = 0.62). The areas of 1 3 Crowned Hawk-Eagle nest stands in KwaZulu-Natal were not normally distributed and were predominantly less than 50 ha in size (Ta- ble 5). Nest stands were primarily small (20 m in width and 30-50 m in length) and elongated in shape (shape-index <0.5; Table 5). Sbape-index values could not be associated with eucalyptus and indigenous tree stands (x^ = 0.12, P = 0.73). Table 4. The number of flight path.s per Crowned Hawk-Eagle nest tree, sampled in four directions, which allowed access to nest forks in emergent and non-emer- gent trees. Number of Flight Paths None One Two Three Four Total Emergent 0 0 5 7 9 21 Non-emergent 0 1 5 3 5 14 2.4 Nest and random tree distances to topographical features were not different for all variables when comparing means (Table 6) and Crowned Hawk- Eagles were, therefore, non-selective regarding these topographical features. One tree selected for nesting was 20 m from an inhabited brick house and another 10 m from a used bitumen road. Discussion Nest-site Selection. Tree-nesting raptors assess variables such as vulnerability to predators, protec- tion against adverse weather conditions, the cost of nest building, structural features of the nest tree, and access to the nest when selecting a site to build their nest (Newton 1979, Moore and Henny 1983, Bosakowski and Speiser 1994, Burton et al. 1994). Whilst smaller raptors regularly conceal their nests to avoid predation, larger raptors are less secretive and customarily put their nests in exposed posi- tions (Speiser and Bosakowski 1987, Selas 1996). The large, 2-2.5 m wide and 2.5-3 m deep nests of the Crowned Hawk-Eagle (Steyn 1982) are very conspicuous. Although large raptors generally do not hide their nests as they can defend their nestling against predators (Moore and Henny 1983), Crowned Hawk-Eagles do suffer some nest predation from primates, especially if the nest tree can be accessed Table 5. Statistical distribution of KwaZulu-Natal Crowned Hawk-Eagle nest-stand variables and shape index with the Kolmogorov-Smirnov distribution test (K-S d) for normality. Dominant Cafegory (Percent) Mean ± 1 SD Range N K-S d Surface area (ha) 0-50 (77) 35.4 ± 73.7 0.05-250.15 13 0.42* Maximum width (m) 20 (69) 253 ± 363 20-1090 13 0.34 Maximum length (m) 30-50 (69) 759 ± 987 30-3520 13 0.32 Shape (0. 0-1.0) 0.2-0.3 (23) 0.39 ± 0.22 0.04-0.80 13 0.16 * P < 0.0.5. December 2002 Nest-site Selection of the Crowned Hawk-Eagle 305 Table 6. Topographical characteristics measured from Crowned Hawk-Eagle nest and random trees in KwaZulu- Natal {N = 13 nests). Variables Nest Tree Random Tree Distance to water (m) 83 ± 147 (10-550)^^ 159 ± 199 (10-520) 0.86 Distance to house (m) 859 ± 1661 (20-6250) 906 ± 1738 (20-6550) 0.24 Distance to road (m) 282 ± 277 (10-810) 285 ± 341 (30-1010) 0.11 ■* Mann-Whitney fTtest. ^ Mean ± 1 SD (range). from nearby trees (Tuer and Tuer 1974, Kalina and Butynski 1994) . This may be a reason why Crowned Hawk-Eagles often select emergents for nest sites, especially in Tai where eight diurnal monkey spe- cies are found at very high densities (McGraw 1998). In Tai, the tall nest trees with their wide bases and high, first side branches (mean = 34 m from the ground) may be extremely difficult, if not impossible, for monkeys to climb (T. Struhsaker pers. comm.). In KwaZulu-Natal, the presence of fewer primate species (three at the most; Smithers 1983) may have resulted in reduced predation risk for nestlings and might have allowed for the use of lower and non-emergent trees. Sites may also be selected for nesting because the foliage protects the nest against adverse weath- er conditions (Moore and Henny 1983, Bosakowski and Speiser 1994). Crowned Hawk-Eagle nests are usually situated within the leafy canopy, but the birds also nest in exposed positions in partially col- lapsed or dead trees (Steyn 1982, Tarboton and Allan 1984, Kalina and Butynski 1994). In this study, 32 nests were located in or below the foliage and therefore sheltered, whereas the remaining three nests, two of which produced young during the study, were located in exposed positions in dead trees. Hence, although there was a trend for the birds to nest in a sheltered position, other fac- tors, such as the availability of suitable nesting sites and the cost of building a new nest, probably in- fluenced the continued occupation of a nest in a dying or dead tree. Lastly, tree-nesting forest raptors may select trees for their size and structural features, such as a tall and open canopy, that allow unobstructed access to the nest (Speiser and Bosakowski 1987, Cerasoli and Penteriani 1996). The Crowned Hawk-Eagle may require a nest that is easily accessible as it needs to fly to the nest carrying sticks up to 1.2 m long and 8 cm thick (Steyn 1982). In addition. Crowned Hawk-Eagles are capable of killing prey 3-4 times their body mass, such as bushbuck ( Tra- gelaphus scriptus; Daneel 1979), which they dismem- ber and carry in parts to the nest. Brown (1966) noted that, in the Karen Forest in Kenya, these ap- proach flights were always above the canopy, very laborious and broken into short 91-137 m stints so as to rest between flights. Under these conditions, access to the nest would be critical in order to de- liver nesting material and prey to the nest. Nesting above the forest canopy in an emergent tree en- hances accessibility. The large indigenous nest trees were simple in structure and provided, on average, one central fork, 3-4 forks against the main branch and 3-5 forks on side branches for the eagles to place their nest. Notwithstanding, only 2-4 branches were used on which to build these large nests, indicating that big, primary and secondary forks were select- ed and not the smaller, multi-stemmed forks from within the canopy structure. Large raptors require a large tree-fork (crotch) to place the nest in and the more open branch structure may facilitate ac- cess (Newton 1979). The exotic eucalyptus differed from the indigenous nest trees in that they provid- ed, on average, 1 1 more forks. This was largely due to the single main-stem growth form of these com- mercially-cultivated trees. Trees from these stands were also largely of similar height, probably be- cause trees in these single species stands were planted simultaneously. As seven of eight non- emergent, eucalyptus nest trees were located on the edge of the nest stand, the eagles may circum- vent the scarcity of emergent eucalyptus by select- ing edge trees that have greater access to the nest. In conclusion, because nests located within and above the canopy were equally accessible, access seemed to be the most critical feature identifying a Crowned Hawk-Eagle nesting tree and site. The findings that indigenous nest trees in KwaZulu-Na- tal did not differ from eucalyptus or indigenous trees from Tai Forest in terms of tree density and 306 Mai AN AND Shui.tz VoL. 36, No. 4 central and side fork availability, may indicate that indigenous trees with their multiple main stems and open branch structure may be as accessible for flying eagles as emergent or edge trees. This study did not quantify the ‘openness’ and accessibility of indigenous nest trees in KwaZulu- Natal, particularly with regard to tree canopy di- ameter and volume, branch spacing and diameter, and branch angle as nests were often placed on near horizontal branches. Also, we did not exam- ine the inter-relationship between the largest-di- ameter side branches (required to support the large nests) and the position of the primary and secondary forks below or just inside the foliage, related to openness and improved access. We also did not take into account the slope at the nest site and aspect of the nest, as nests located on the downhill side of the tree may have been more ac- cessible to flying eagles. Given the floristic differ- ences between forests in Tai and KwaZulu-Natal, a comparison of nest sites Avith randomly selected sites, conducted at each nest stand, would have fur- ther contributed toward the understanding of what constitutes a suitable nest tree. Lastly, data on how easy arboreal primates can climb nest trees, partic- ularly from the base, would have added to our un- derstanding of how successful Crowned Hawk-Ea- gles are in eliminating the predation risk by nesting in tall, emergent trees. Recommendations. Wildlife managers need to manage forests by balancing timber harvesting with maintaining wildlife habitat and must employ multi-resource plans to do so (Lilieholm et al. 1993, Vermeulen 1999, Malan and Robinson 2001). These plans should incorporate both prac- tical and proactive management objectives to con- serve nesting habitat for large eagles in exotic and indigenous forests. Apart from silvicultural guide- lines, other considerations must include the prox- imity of nests to hunting habitat, their temporal and spatial distribution, and tbe effects of human disturbance on nesting eagles (Lilieholm et al. 1993). In this study, because we could not obtain reli- able data on the age of the nests and how many young were fledged from each nest over time, we did not compare the structural features of produc- tive between unproductive nests. Although other factors also influence productivity, e.g., experience of breeders, prey abundance, and persecution rates, the productivity analysis may have highlight- ed subtle differences between suitable and unsuit- able nest trees and nest-placement sites. Our rec- ommendations should therefore be treated with some caution, as nests included in this study might have been found unsuitable on the long term be- cause certain deficient nest features may have lim- ited successful reproduction. Nonetheless, the nest trees must have open- branch structures and large tree-forks. As nest trees are typically emergents or edge trees, these trees can be cultivated by felling surrounding trees or leaving the tall trees standing (Seydack 1995). In this study more than half of the Crowned Hawk- Eagle nests were positioned in a central fork, there- fore, the techniques of coppice-reduction or selec- tive pruning could be employed to cultivate a tree with the preferred 2-4 main branches. As Brown (1966) demonstrated, suitable nest trees can be identified a priori, and managers can therefore im- plement these techniques to ensure a continued supply of suitable nest trees. To qualify as a nest tree, commercial eucalyptus should be managed to reach a minimum height of 34 m and DBH of 83 cm (i.e., minimum size se- lected for nesting) . Indigenous trees selected must attain a minimum height of 15 m and diameter of 35 cm. The mean stand density of 156 trees/ha can be employed as a guideline for Crowned Hawk-Ea- gle nesting habitat. Although exotic stands must be managed to a minimum size so as not to encroach onto indigenous vegetation, nest-tree stand size it- self is nonessential. Furthermore, in plantations, nest stands cannot be placed at random as raptors often require nest trees to be located away from certain topographical features (Andrew and Mosh- er 1982, Malan and Robinson 1999). However, with regard to the proximity of nests to human dwell- ings, water bodies, and public roads, the Crowned Hawk-Eagles of KwaZulu-Natal were remarkably non-selective and tolerant. In fact, the city of Dur- ban, located in this province, harbors 12 active nests within its metropolitan boundaries. Inter- stand distances were not recorded in this study, but Crowned Hawk-Eagles are known to nest 1.8— 4.0 km apart and, thus, require only a relatively small patch of suitable hunting habitat (Tarboton and Allan 1984, Allan et al. 1996, Mitani et al. 2001, Shultz 2002). Given the size of some of the prey these eagles hunt, nest stands must preferentially be located near the hunting habitat so as to sbort- en flight distances to the nest. Because Crowned Hawk-Eagles use the same nest for 10 years or longer, often until the tree col- December 2002 Nest-site Seitctton of the Crowned Hawk-Eagle 307 lapses (Brown 1966, Steyn 1982), suitable trees should not be felled. If the nest stand or tree must be harvested, the felling should be done outside the birds’ breeding season (August-March) and 6- 12 months after the nestlings have fledged to allow sufficient time for the young bird to become in- dependent (Steyn 1982, Tarboton and Allan 1984). Given that a nest takes 4-5 mo to build (Brown 1966), replacement nest trees, located in the vicin- ity of the nest tree in use, should be cultivated long in advance to provide a suitable alternative nest. The Crowned Hawk-Eagle in South Africa has been classified as near threatened because of past exploitation of nest trees and the likely destruction of nesting habitat in the near future (Boshoff et al. 1983, Barnes 2000). Occupied nests should be closely monitored to assess the status of this species and to collect data on what constitutes productive nests. Large tree-nesting forest raptors will always have few suitable trees to nest in as large trees and forks are less abundant than smaller ones (Newton 1979). When nesting in large trees, eagles also compete directly with humans for this scarce re- source (Boshoff et al. 1983, Watson and Rabarisoa 2000). It is, therefore, no longer adequate simply to protect forests for birds of conservation con- cern, but specific efforts must be made to satisfy the nesting requirements of the Crowned Hawk- Eagle and other tree-nesting forest raptors; e.g., in South Africa the Bat Hawk {Macheiramphus alci- nus), Ayres’s Hawk-Eagle (Hieraaetus ayresii), and Fasciated Snake-Eagle {Circaetus fasciolatus) (Barnes 2000, Malan and Marais in press). Ulti- mately, a species-specific management plan should be developed for these and other tree-nesting for- est birds and the maintenance of nest-tree struc- tural and topographical features incorporated into management decisions. Acknowled gments We want to thank Natal Portland Cement and the KwaZulu-Natal Ornithological Trust for sponsoring this project. We also thank KwaZulu-Natal Wildlife for allow- ing us to work in their reserves and park. Bill Howells and David Johnson were particularly helpful in locating nest sites. Guy Tedder, Denise James, Martin Buchler, Gavin Lorry, and Bill Walker also informed us of nesting localities. In Tai, we thank Professor Ronald Noe for host- ing the Crowned Hawk-Eagle project, the Centre Suisse des Recherches Scientifiques, Centre de Recherche en Ecologie, the Project Autonome pour la Conservation du Parc National de Tai for logistical support, the Minstere de la Recherche and the Direction de la Protection de la Nature for permission to conduct the study. Funding (for S. Shultz) was provided by the Raptor Research Foun- dation, Inc., Leslie Brown Memorial Award, the Pere- grine Foundation, a Wildlife Conservation Society Re- search Fellowship, and the Leakey Foundation. Literature Cited At j an, D. and W.R. Tarboton. 1985. Sparrowhawks and plantations. Pages 167-177 mL.J. Bunning [Ed.], Pro- ceedings of the symposium on birds and man. Wits Bird Club, Johannesburg, South Africa. , D. Johnson, and T. Snyman. 1996. The crowned eagles of Durban and Pietermaritzburg. Afr. Birds Bird- ing 1:2-13. , J.A. Harrison, R.A. Navarro, B.W. Van Wilgen, AND M.W. Thompson. 1997. The impact of commer- cial afforestation on bird populations in Mpumalanga Province, South Africa — insights from bird atlas data. Biol. Conserv. 79:173-185. Andrew, J.M. and J.A. Mosher. 1982. Bald Eagle nest-site selection and nesting habitat in Maryland. J. Wildl Manage. 46:383-390. Anonymous. 1998. Commercial timber resources and roundwood processing in South Africa 1996/1997. Dept, of Water Affairs and Forestry, Pretoria, South Africa. Barnes, K.N. 2000. The eskom red data book of birds of South Africa, Lesotho, and Swaziland. BirdLife South Africa, Johannesburg, South Africa. Bijleveld, M. 1974. Birds of prey of Europe. Macmillan Press, London, UK. Bosakowski, T. and R. Speiser. 1994. Macrohabitat selec- tion by nesting Northern Goshawks: implications for managing eastern forests. Pages 46—49 in W.M. Block, M.L. Morrison, and M.H. Reiser [Eds.], The North- ern Goshawk: ecology and management. Studies in avian biology No. 16. Cooper Ornithological Society, Sacramento, CA U.S.A. Boshoff, A.F. 1988. The spacing and breeding periodic- ity of Crowned Eagles in the southern Cape Province. Bontebok 6:34—36. . 1997. Crowned Eagle Stephanoaetus coronatus. Pag- es 194—195 m J.A. Harrison, D.G. Allan, L.G. Under- hill, M. Herremanns, A.J. Tree, V. Parker, and C.J. Brown [Eds.], The atlas of southern African birds Vol. 1. BirdLife, Johannesburg, South Africa. , C.J. Vernon, and R.K. Brooke. 1983. Historical atlas of the diurnal raptors of the Cape Province (aves: falconiformes) . Ann. Cape Prov. Mus. Nat. Hist 14:173-297. Brown, L.H. 1966. Observations on some Kenya eagles. Ibis 108:531-572. and D. Amadon. 1989. Eagles, hawks, and falcons of the world. Wellfleet Press, New York, NY U.S.A. Burton, A.M., R.A. t\lford, and J. Young. 1994. Repro- ductive parameters of the Grey Goshawk {Accipiter no- vaehollandiae) and Brown Goshawk {Accipiter fasciatus) at Abergowrie, northern Queensland, Australia J Zool. Land. 232:347—363. 308 Malan and Shuitz VoL. 36, No. 4 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. Daneel, A.B.C. 1979. Prey size and hunting methods of the Crowned Eagle. Ostrich 50:120—121. Daw, S.K., S. DeStefano, and R.J. Steidl. 1998. Does sur- vey method bias the description of Northern Goshawk nest-site structure? J. Wildl. Manage. 62:1379-1384. Gartshore, M.E., P.D. Tayi or, and I.S. Erancis. 1995. Forest birds in the Cote d’Ivoire: a survey of Tai Na- tional Park and other forest and forestry plantations. BirdLife Int. Stud. Rep. No. 58. BirdLife Internation- al, Cambridge, U.K. Guillaumet, J.L. 1994, La flore. Pages 66-71 m E.P. Rie- zebos, A.P. Vooren, and J.L. Guillaumet [Eds.], Le Parc National de Tai, Cote d’Ivoire. Synthese des con- naissances. Fondation Tropenbos, Wageningen, Neth- erlands. Kalina, J. and T.M. Butynski. 1994. Natural deaths of two Crowned Eagles in Uganda. Gator 9:28-31. Lilieholm, R.J., W.B. Kessler, and K. Merrill. 1993. Stand density index applied in timber and goshawk habitat objectives in Douglas-Fir. Environ. Manage. 17: 773-779. Low, A.B. and A.G. Rebei.o. 1996. Vegetation of South Africa, Lesotho, and Swaziland. Dept, of Environmen- tal Affairs and Tourism, Pretoria, South Africa. Macdonaid, I.A.W. 1986. Do redbreasted sparrowhawks belong in the Karoo? Bokmakierie 38:3-4. Malan, G. and E.R. Robinson. 1999. The diet of the Black Sparrowhawk: hunting columbids in man-al- tered environments. Durban Mus. Novit. 24:43-47. and E.R. Robinson. 2001. Nest-site selection by Black Sparrowhawks Accipiter nielanoleucus: implica- tions for managing exotic pulpwood and sawlog for- ests in South Africa. Environ. Manage. 28:195-205. AND A.V.N. Marais. 2002. Guidelines for the de- sign and management of artificial raptor perches and exotic nest-tree sites on forestry estates. S. Afr. For. /.: in press. McGraw, W.S. 1998. Comparative locomotion and habi- tat use of six monkeys in the Tai Forest, Ivory Coast. Am. ]. Phys. Anthropol. 105:493-510. Midgley, J.J., R.M. Cowling, A.H.W Seydack, and (i.E van Wyk. 1997. Forest. Pages 278-299 in R.M. C'.owl- ing, D.M. Richardson, and S.M. Pierce [Eds.], Vtygc- tation of southern Africa. Cambridge Univ. Press, Cambridge, U.K. Mitani, J.C., WJ. Sanders, J.S. Lwanga, and T.K.L. Wind- eelder. 2001. Predatory behaviour of Crowned Hawk- Eagles {Stephanoaetus coronatus) in Bibale National Park, Uganda. Behav. Ecol. Sociobiol. 49:187-195. Moore, K.R. and CJ. Henny. 1983. Nest-site character- istics of three coexisting accipiter hawks in northeast- ern Oregon. Raptor Res. 17:65-76. Newton, I. 1979. Population ecology of raptors. T. & A.D Poyser, Berkhamsted, U.K. Phillips, E.A. 1959. Methods of vegetation study. Holt, Rinehart & Winston, Inc., New York, NY U.S.A. Selas, V. 1996. Selection and reuse of nest stands by spar- rowhawks {Accipiter nisus) in relation to natural and manipulated variation in tree density, f. Avian Biol. 27. 56-62. Sfydacr, A.H.W. 1995. An unconventional approach to timber yield regulation for multi-aged, multispecies forests. Fundamental consideration. For. Ecol. Manage 77:139-153. Shuitz, S, 2002. Population den.sity, breeding chronolo- gy and diet of Crowned Eagles Stephanoaetus coronatus in Tai National Park, Ivory Coast. Ibis 144:135-138. Smithers, R.H.N. 1983. The mammals of the southern African sub-region. University of Pretoria, Pretoria, South Africa. Spelser, R. and T. Bosakowski. 1987. Nest-site selection by Northern Goshawks in northern New jersey and southeastern New York. Cowiior 89:387-394. StatSoft. 1995. Statistica for Windows. StatSoft, Inc , 2300 East 14th Street, Tulsa, OK U.S.A. Steyn, D.J, 1977. Occupation and the use of the eucalyp- tus plantations in Tzaneen area by indigenous birds N A/l For /. 100:56-60. Steyn, P. 1982. Birds of prey of southern Africa. David Philip Publishers, Cape Town, South Africa. Tarboton, W.R. and D.G. Allan. 1984. The status and conservation of birds of prey in the Transvaal. Trans- vaal Mus. Monogr. No. 3. Transvaal Museum, Pretoria, South Africa. Tler, V. and J. Tuer. 1974. (browned Eagles of the Ma- topos. Honeyguide 80:32-41. Van der Zel, D.W. 1996. South African national forestry action plan. Dept, of Water Affairs and Forestry, Pre- toria, South Africa. Vermeulen, C. 1999. The multiple-use management of the indigenous evergreen high forests of the southern Cape and Tsitsikamma. Dept, of Water Affairs and E'orestry, Knysna, South Africa. Watson, R.T. and R. Rabarisoa. 2000. Sakalava fisher- man and Madagascar Fish-F.agles: enhancing tradi- tional conservation rules to control resource abuse that threatens a key breeding area for an endangered eagle. Ostrich 71:2-10. Received 5 November 2001; accepted 2 August 2002 Short Communications J. Raptor Res. 36(4):309-314 © 2002 The Raptor Research Foundation, Inc. Juvenile Dispersal of Madagascar Fish-Eagles Tracked by Satellite Telemetry Simon Rafanomezantsoa The Peregrine Fund, P.O. Box 4113, Antananarivo 101, Madagascar Richard T. Watson ^ and Russell Thorstrom The Peregrine Fund, 5668 West Flying Hawk Lane, Boise, ID 83709 U.S.A. Key Words: Madagascar Fish-Eagle, Haliaeetus vocifer- oides; dispersal] satellite telemetry. The Madagascar Fish-Eagle {Haliaeetus vociferoides) is critically endangered (Stattersfield and Capper 2000) with a small population limited to wetland habitats on Madagascar’s western seaboard (Rabarisoa et al. 1997). Observations of extra-pair adults at the nest and juveniles m the territories of breeding pairs (Watson et al. 1996, 1999) suggest unusual behaviors that may occur as a re- sult of overcrowding in limited suitable habitat, or if in- nate, may reduce the species’ ability to disperse into un- occupied habitat. In addition to the behavioral and evolutionary significance of understanding these obser- vations, the cause and consequence of this behavior may affect conservation interventions intended to prevent the species’ extinction. We report the results of a pilot study to measure the movements and habitat use of fledglings after they left parental territories to better understand post-fledging dispersal (Rafanomezantsoa 1998) and the occurrence of extra birds at the nest, and to assess the effect on dispersal of release of captive-raised birds (Wat- son et al. 1996, 1999, Rafanomezantsoa and Kalavaha 1999). The movements of raptors have been investigated for decades mainly by banding, direct observation in limited areas, or tracking using VHF radios (Meyburg and Mey- burg 1999). Recently, satellite telemetry has provided a method that makes possible the global location of birds over an extended period. Satellite telemetry was consid- ered the method of choice to study juvenile dispersal in Madagascar Fish-Eagles which may reach maturity at 3-4 yr of age. In addition to the extended study period, ju- veniles were expected to move distances greater than we could follow with conventional VHE-radio telemetry be- cause much of their range is inaccessible, especially dur- ing Madagascar’s wet season when lowlands may flood (Rafanomezantsoa 1997). ’ Corresponding author’s e-mail address: rwatson® peregrinefund.org Methods This study occurred in coastal floodplain wetlands of western Madagascar between the Manambolo River (19°15'S, 44°30'E) and Soahany River (18°40'S, 44°30’E), about 300 km west of the capital, Antananarivo. This area lies within the Western Malagasy phytogeographical re- gion (Humbert 1954), which is characterized by annual rainfall from 1000-2000 mm, monthly mean tempera- tures above 20°C, and elevations less than 800 m. The wet season begins in October or November and lasts through March. The dry season begins in May and lasts through September or October. The climax vegetation is tropical dry deciduous forest, but savanna grasslands, maintained by burning, dominate the landscape (Gml- laumet 1984). Most lakes in the region are floodplain lakes whose surface area varies considerably between wet and dry seasons (Kiener and Richard-Vindard 1972). During the 1997 breeding season (May-December), Platform Transmitter Terminals (PTTIOO Series®; Micro- wave Telemetry 2000, Inc., Columbia MD) with antennas hxed at 45*^, were mounted with a backpack style harness on two fledgling fish-eagles. One transmitter (PTT No 3482) was mounted on a female that was rescued from an aggressive sibling, raised in captivity, and released at fledging age (Watson et al. 1996, 1999). This bird was released by hacking (a falconry term for the process of release to the wild; Cade 2000) into unoccupied, suitable habitat at Lake Mangily (18°30'S, 44°34'E). The second PTT (No. 3480) was put on a male that fledged naturally from a fish-eagle pair on Lake Ankerika (19°03'S, 44°27'E), about 65 km south of the first. PTTs were pro- grammed for 8 hr on, 24 hr off for six cycles; 8 hr on, 96 hr off for 84 cycles; and 8 hr on, 240 hr off for the remaining life of the transmitter, to gain more frequent locations of movements made in the first year, followed by less frequent locations for the remaining life of the PTT, which we expected to be at least 4 yr. Satellite telemetry uses the ARGOS Data Collection and Location System (Meyburg et al. 1995). ARGOS as- signs PTT locations a grade according to calculated pre- cision. We used only locations graded as within 1000 m of actual position. ARGOS reports two locations, one be- ing a spurious mirror of the true location. We deduced the true location using either concurrent visual obser- vations, likelihood of sequential locations, or known hab- 309 310 Short Communications VoL. 36, No. 4 itat at each location. Locations were plotted in chrono- logical order on a topographical 1:100 000 map. The pro- gram Ranges IV® (Kenward 1990) was used to calculate distances moved and range areas based on minimum convex polygons (MCP) around locations. Results PTT No. 3480 on the male operated for 282 d, from 1 Decemher 1997-8 September 1998, providing 70 us- able locations. From these locations, the total MCP range of the male was 1040 km^ (Fig. 1). Between the fledging date on 13 October 1997 and 17 April 1998 (186 d), the male moved in a relatively small area around its nest site. During the wet season, from December-March, the male visited small lakes and flooded lowlands to the south and southeast of its nest site along the Manambolomaty River and Lakes Kakobo and Antsohaly, covering an area of 156.4 km^. By 24 April 1998, the male had moved north- northwest with one location near Lake Mangily (the fe- male’s release site) on the way. For 76 d at the beginning of the dry season, from 13 May 1998-28 July 1998, the male was located in the vicinity of the Soahanina River estuary, and associated ponds and mangroves, ranging over an area of 540.5 km^. The maximum distance re- corded from its nest was 51 km. By 5 August 1998 and through 8 September 1998, when the last usable trans- mission was received, the male was located again in the vicinity of the nest where his parents were raising a nest- ling. The PTT on the female operated for 348 d, from 15 October 1997-28 September 1998, providing 40 usable locations (Fig. 2). We received 21 usable locations from 15 October 1997-22 November 1997 when the PTT ceased transmitting for unknown reasons. It then re-start- ed 80 d later on 11 February 1998 and we received an- other 19 usable locations through 28 September 1998. About 95% of usable fixes were <5 km from the release site. Direct observation conhrmed that the bird moved around Lakes Mangily and Sarny until held observations ended 13 March 1998 due to inaccessibility from flood- ing. The maximum distance recorded from the release site was 8.2 km and the mean distance was 2.6 km. The female’s MCP range was 45 km^. Discussion Usable locations for the male were sufficient to deter- mine several aspects of the bird’s movement. Locations were usually received 10 or more days apart, which lim- ited the amount of detail that could be gained from them, Initial movements of the naturally fledged male were distances of 5-18 km to sites that may be visible to a soaring bird from its natal site. The male made one large movement of short duration 51 km to the north- west, where he remained for 76 d, making successive short movements, until returning to his natal site. All ar- eas visited were without territorial, breeding pairs. The male did not visit prominent wetland areas within his MCP range where human disturbance was high, such as Lakes Bemamba and Antsohaly. The released female’s range during the wet season, as water flooded the surrounding lowland areas, included Lakes Mangily and Sarny, other small lakes, and a portion of the Miharana River. Compared to the male, the female made only very short movements around the release site While it is possible that the low number of usable loca- tions influenced her measured range area, results were spread throughout a similar period to those of the male and observations by our field team confirmed her loca- tions during much of the study period. The differences in movement patterns of the male and female are interesting. They may be related to sexual dif- ferences in behavior, but there are a number of other factors that could influence dispersal. For example, nat- urally-raised and fledged compared with captive-raised and released by hacking might affect their behavior. So- cial interactions with neighboring conspecifics, the nat- urally fledged male had at least 10 territorial pairs near- by; whereas, the female had only one resident pair within 5 km, which could have caused a difference in dispersal behavior. Also, distribution of food during the wet-season floods might influence movements. Habitats visited by both eagles were lakes, rivers, man- groves, and lowlands temporarily flooded during the wet season. We assume that movements of fledglings were made in exploration of suitable foraging habitat because there was no seasonal pattern to the movements, and they were of short distance and duration relative to the species’ capacity for flight. Despite the potential of satellite telemetry to track birds for up to 4 yr, because no signal was detected from the PTTs after about 18 mo, we were unable to learn more about the occurrence of extra birds at the nest be- yond the observation that the male returned to its par- ent’s territory about 10 mo after fledging and remained there at least a month while the adults were caring for a nestling. Since this study, direct observation of banded birds, and molecular studies of their genetic relatedness, have revealed far more information (Tingay 2000, Tingay et al. 2002). Tracking additional birds to increase sample size need- ed to compare dispersal between naturally-fledged and captive-reared and released juveniles, and document hab- itat use during dispersal, has not been attempted. The practical difficulties and cost of captive rearing in Mad- agascar, and the cost of satellite telemetry, precluded fur- ther study. Few studies have been done on post-fledging dispersal in raptors (e.g., Walls and Kenward 1994) or comparison between captive-raised and naturally-fledged birds (e.g., Amar et al. 2000) probably for similar reasons. As satellite telemetry becomes more efficient and afford- able, these areas of study may benefit. Satellite telemetry provided us with a level of coverage and continuity, especially for long-distance movements, that we could not have achieved for the male using con- December 2002 Short Communications 311 44° 30’ E Figure 1 . Chronological movements and minimum convex polygon range of the male Madagascar Fish-Eagle fledged from a nest on Lake Ankerika. 312 Short Communications VoL. 36, No. 4 Figure 2. Minimum convex polygon range of the captive-reared female Madagascar Fish-Eagle released by hacking at Lake Mangily. December 2002 Short Communications 313 ventional VHF radios and tracking from the ground. We may have achieved similar results by VHF-radio tracking from an aircraft, but lack of landing fields and fuel in the study area precluded this approach. The pro- grammed on-off cycle of the transmitters provided fewer useful locations than we expected. A longer (e.g., 12-24 hr) on-cycle might have generated more usable locations m these birds that were more sedentary than migrating raptors. For example, two migrating Bald Eagles {Hal- iaeetus leucocephalus) with PTTs programmed for 8 hr on, 16 hr off, and 4 hr on, 44 hr off cycles generated 205 fixes in 136 d and 27 fixes in 119 d, respectively (Grubb et al. 1994). A migrating Wahlberg’s Eagle {Aquila wahl- bergi) with a PTT programmed for 8 hr on, 134 hr off generated 104 locations in 234 d (Meyburg et al. 1995). A 12-24 hr on-cycle would increase the chance of satel- lites passing over an eagle in optimum position (such as m flight) to receive the transmitter’s signal. Transmission of usable locations ceased for unknown reasons well be- fore their anticipated 4-yr lifespan. Unexplained periods of no data received added to our uncertainty of the birds’ movements. We believe satellite telemetry is valuable for tracking the movement of large birds of prey; however, there are several limitations. The accuracy of locations within the PTT’s operational period varied from day to day. The precision of the location is known to be affected by sev- eral satellite parameters and by the orientation, location, and movement of the transmitter. PTTs are at present larger and less streamlined than conventional VHF trans- mitters, and require antennas that protrude at a 45° an- gle rather than contour down the back and the tail of the bird. Also, PTTs are .substantially more expensive, though this cost may compare favorably with the cost of data collection from VHF radios which require personnel in the field and logistical support, Resumen. — En 1997, colocamos radio transmisores a dos polluelos de aguilas pescadoras de Madagascar (Haliaee- tus vociferoides) para estudiar sus movimientos y uso de habitat durante la dispersion post emplumamiento. El macho partio naturalmente del nido de sus padres dur- ante Octubre y nosotros liberamos la hembra criada en cautiverio el 15 de Octubre. El rango del aguila macho fue 1040 km^ entre el 1 de noviembre de 1997 y el 8 de septiembre de 1998 (70 localizaciones) y se movio una distancia maxima de 51 km desde su area natal, al norte del rio Soahanina, El comenzo moviendose largas distan- cias (>5 km) el 16 de Diciembre de 1997, permanecio en la vecindad del estuario del rio Soahanina por 76 dias, entonces retorn o a la vecindad del nido (<5 km) el 5 de Agosto de 1998. El rango de la Hembra fue 45 km^ entre el 15 de Octubre de 1997 y el 28 de Septiembre de 1998 (40 localizaciones) y se movio una distancia maxima de 8.2 km desde el sitio de liberacion. Los rios, lagos, mang- lares y zonas bajas temporalmente inundadas durante la temporada humeda fueron visitados por ambas aguilas. Sugerimos que los movimientos de los volantones se hi- cieron para explorar habitats de forrajeo adecuados, ya que no hay un patron estacional para los movimientos, y fueron relativamente de corta distancia y duracion con respecto a la capacidad de vuelo de la e.specie. [Traduccion de Cesar Marquez] ACKNOWLTDCiMEN'rS This study was conducted by The Peregrine Fund’s Madagascar Project with funding provided, in part, by Environment Now, the Liz Claiborne and Art Ortenberg Foundation, the John D, and Catherine T. MacArthur Foundation, the Walt Disney Company Foundation, the Little Family Foundation and other important contribu- tors. The project received in-kind support from the Na- tional Aeronautics and Space Administration, United States Fish and Wildlife Service, and other partners. We especially thank David H. Ellis, Jon W. Robinson, and Paul Howey for their cooperation and support during this project, and the constructive reviews of this manu- script by James Berkelman, Mark Puller, and Robert Leh- man. Literature Cited Amar, a., B.E. Arroyo, and V. Bretagnotle. 2000. Post- fledgling dependence and dispersal in hacked and wild Montagu’s Warriers Circus pygargus. Ibis 142:21 — 28. Cade, T.J. 2000. Progress in tran.slocation of diurnal rap- tors. Pages 343—372 in R.D. Chancellor and B.-U. Mey- burg [Eds.], Raptors at risk. Hancock House/ WWGBP, Berlin, Germany. Grubb T.G., W.W. Bowerman, and P.H. Howey. 1994 Tracking local and seasonal movements of wintering Bald Eagles Haliaeetus leucocephalus from Arizona and Michigan with satellite telemetry. Pages 347—358 in B.-U. Meyburg and R.D. Chancellor [Eds.], Raptor conservation today. The Pica Press/WWGBP, Berlin, Germany. Guiliaumet, J.-L. 1984. The vegetation: an extraordinary diversity. Pages 27-54 in A. Jolly, P. Oberle, and R Albignac [Eds.], Key environments: Madagascar. Per- gamon Press, Oxford, U.K. Humbert, H. 1954. Les territoires phytogeographiques de Madagascar, leur cartographic. Pages 195-204 in Les divisions ecologiques du monde. Centre Nation- ale de la Recherche Scientifique, Paris, France. Kenward, R.E. 1990. Ranges TV. Institute of Terrestrial Ecology, Wareham, England. Kiener, a. and G. Richard-Vindard. 1972. Fishes of the continental waters of Madagascar. Pages 477-499 in R. Battistini and G. Richard-Vindard [Eds.], Bioge- ography and ecology of Madagascar, Dr. W. Junk B V Publishers, The Hague, Netherlands. Meyburg, B.-U. and C. Meyburg. 1999. The study of rap- tor migration in the old world using satellite teleme- try. Pages 2992-3006 in N.J. Adams and R.H. Slotow 314 Short Communications VoL. 36, No. 4 [Eds.], Proc. 22nd Int, Ornithol. Congr. BirdLife South Africa, Johannesburg, South Africa. Meyburg, B.-U., J.M. Mendelsohn, D.H. Ellis, D.G. Smith, C. Meyburg, and A.C. Kemp. 1995. Year-round movements of a Wahlberg’s Eagle Aquila wahlbergi tracked by satellite. Ostrich 66:135-140, Rabarisoa, R., R.T. Watson, R. Thorstrom, and J. Ber- kelman. 1997. The status of the Madagascar Fish-Ea- gle Haliaeetus vociferoides in 1995. Ostrich 68:8-12. Rafanomezantsoa, S.A. 1997. Behavior and natal dis- persal of fledgling Madagascar Fish-Eagles. Pages 403-412 in R.T. Watson [Ed.], Madagascar wetlands conservation project. Progress report III, The Pere- grine Fund, Boise, ID U.S.A. . 1998. Contribution a 1’ etude des comportements et de la dispersion des jeunes pygargues de Madagas- car Haliaeetus vociferoides (Desmurs 1845) dans le com- plexe des 3 lacs d’Antsalova. Memoire de D.E.A. Univ- ersite d’Antananarivo, Madagascar. and L. Kalavaha. 1999. Transfert de jeunes Py- gargues de Madagascar. Pages 115-123 in A. Andri- anarimisa [Ed.]. Projet de conservation des zones humides de Madagascar. Rapport d’avancement V, 1997-1998. The Peregrine Fund, Antananarivo, Mad- agascar. Statterseield, AJ. and D.R. Capper. 2000. Threatened birds of the world. Lynx Edicions, Barcelona, Spain. Tingay, R.E. 2000. Sex, lies, and dominance: paternity and behaviour of extra-pair Madagascar Fish-Eagles. M.S. thesis. University of Nottingham, U.K. , M. CuLATR, E.M. Hallerman, R.T. Watson, and J.D. Fraser. 2002. Subordinate males sire offspring in Madagascar Fish-Eagle {Haliaeetus vociferoides) polyan- drous breeding groups. /. Raptor Res. 36:280-286. Wau.s, S.S. and R.E. Kenward. 1994. Movements of ra- dio-tagged Common Buzzards Buteo buteo in their first year. lUs 137:177-182. Watson, R.T, S. Thomsett, D. O’Daniel, and R. Lewis. 1996. Breeding, growth, development, and manage- ment of the Madagascar Fish-Eagle {Haliaeetus vocifer- oides). J. Raptor. Res. 30:21-27. , S. Razaeindramanana, R. Thorstrom, and S. Ra- fanomezantsoa, 1999. Breeding biology, extra-pair birds, productivity, siblicide and conservation of the Madagascar Fish-Eagle. Ostrich 70:105-111. Received 9 July 2001; accepted 28 June 2002 J. Raptor Res. 36(4) :315-319 © 2002 The Raptor Research Foundation, Inc. Prey of the Peregrine Falcon {Falco peregrinus cassini) in Southern Argentina and Chile David H. Ellis^ USGS Patuxent Wildlife Research Center, HC 1 Box 4420, Oracle, AZ 85623 U.S.A. Beth Ann Sabo Wildlife Forensics Services, P.O. Box 142613, Irving, TX 75014-2613 U.S.A. James K. Fackler 590 Davidson Road, Bow, WA 98232 U.S.A. Brian A. Millsap Fish and Wildlife Conservation Commission, 620 S. Meridian Street, Tallahassee, FL 32399 U.S.A. Keywords; Peregrine Falcon; Falco peregrinus; Argentina; Chile, Pallid Falcon; prey. The Peregrine Falcon {Falco peregrinus cassini) in Pata- gonia attracted wide interest two decades ago (Anderson and Ellis 1981, McNutt 1984) when there was a focus on determining the taxonomic position of the Pallid Falcon (also called Kleinschmidt’s falcon and Tierra del Fuego falcon; formerly named Falco kreyenborgi) . In 1981, how- ever, the pallid falcon was confirmed to be a pale color morph of the peregrine (Ellis et al. 1981, Ellis and Peres 1983), and since that time, little work has been conduct- ed on this color morph. Continent-wide research has continued and has yielded a fair understanding of the breeding distribution of the Peregrine Falcon in South America (Anderson et al. 1988, McNutt et al. 1988, Rise- brough et al. 1990). Also, two preliminary food habits studies on the peregrine have been completed in Pata- gonia (McNutt 1981, Peres and Peres 1985). Together those papers provided a list of 23 species observed as prey, and McNutt (1981) listed another eight species seen pursued (but not captured) by peregrines. The purpose of this paper is to assemble all that has been published on peregrine food habits for Patagonia and Tierra del Fuego and to add to that list from our 1980 and 1981 expeditions. Methods In November-December of 1980 and 1981, we traveled by motor vehicle searching for eyries in Chubut, Santa Cruz, and Rio Negro provinces of Argentina and in Ma- gallanes, Chile. Although we accessed 16 eyries, some were empty (prey remains sometimes scatter in the wind), so our totals included prey from only 11 eyries. We accessed eyries (normally by rope) and recovered re- cent prey (feathers, feet, and bones with tendons at- ^ E-mail address: david_h_ellis@usgs.gov or dcellis® theriver.com tached) but discarded those bones that were so bleached that they may be attributed to former occupants of the eyrie. Most of the prey were identified from whole feath- ers. No pellets were used in this anaylsis. We included some feathers from the base of the eyrie cliffs, but ex- cluded those that were likely molted by other occupants of the cliff. For example, several of our eyries were in old Black-faced (formerly buff-necked) Ibis {Tkeristicus cau- datus) nests within active ibis colonies. Although ibis feathers were frequently found near these eyries and even though we occasionally observed peregrines pursu- ing ibis, we viewed these attacks near eyries as displace- ment activities. Only once did we include an ibis as prey and this was after finding four fresh feathers within an eyrie which was neither beneath an ibis roost nor near an ibis nest. McNutt (1981) observed peregrines killing nestling ibis. No food habits study based on prey remains is without bias (Marti 1987, Bielefeldt et al. 1992). For peregrines, bias derives from the fact that many prey individuals are missed because prey are normally plucked before arrival at the eyrie and many defleshed carcasses are removed by the adults and deposited elsewhere. Also, several cast- ings sometimes represent a single prey item. Common prey are normally under represented in peregrine prey tallies, including our sample, because of the difficulty of totaling individuals. Our method was to derive a mini- mum count from feet, bills, remiges, and rectrices. For example, a sample of 300 feathers and assorted other remains from one species, and probably representing dozens of individuals, may yield a much smaller mini- mum count. Conversely, rare prey are likely to be over estimated in most studies including this one, because a single feather, bill, or foot can document prey that was accrued only once. Prey were placed in plastic bags and air dried by open- ing the bags in a windless situation on sunlit days and fumigated prior to identification at the U.S. National Mu- seum (USNM: Smithsonian Institute). At USNM, we as- sembled a synoptic series including all known and most of the likely prey species. Because USNM does not have examples for all plumages of all Patagonian birds, we could not determine species on nine individuals. 315 316 Short Communications VoL. 36, No. 4 Table 1. Avian prey of the Peregrine Falcon in southern Patagonia and Tierra del Fuego.^’^ Numbers refer to minimum number of items represented in remains. A plus ( + ) indicates that a taxa was documented, but the number of Items was not reported. Nomenclature follows Sibley and Monroe (1990). Family Scientific Name Common Name McNutt 198L^ Peres and Peres 19853 This Study 3 Rheidae Rhea pennata Lesser Rhea + * 1* Tmamidae Eudromia elegans Elegant Crested-Tinamou 2 Podicipedidae Podiceps major Great Grebe 1 Procellariidae Halobaena caerulea Blue Petrel + Pachyptila belcheri Slender-billed Prion -t- Pelecanoididae Pelecanoides magellani Magellanic Diving-Petrel -f Pelecanoides sp. 3 Ardeidae Nycticorax nycticorax Black-crowned Night-Heron 1 Threskiornithidae Theristicus caudatus Buff-necked Ibis *7*4 1 Anatidae Chloephaga picta Upland Goose Anas sp. 1 Anas flavirostris Speckled Teal 2 1 Anas platalea Red Shoveler 1 Falconidae Falco sparverius American Kestrel 2 Phasianidae Callus gallus domesticus Domestic Chicken 1* Rallidae Fulica leucoptera White-winged Coot 1 Charadriidae Vanellus chilensis Southern Lapwing 1 -t- 3 & 2* Charadrius falklandicus Two-banded Plover -f Charadrius modestus Rufous-chested Plover -h Oreopholus ruficollis Tawny-throated Dotterel 1 -f 7 & 1* Scolopacidae Fimosa haemastica Hudsonian Godwit -f Gallinago stricklandii Fuegian Snipe 1 Thmocoridae Thinocorus orbignyianus Gray-breasted Seedsnipe 2 -f 4 Thinocorus rumicovorus Least Seedsnipe 17 + 5 Laridae Sterna sp. 2 Sterna hirundinacea South American Tern + Columbidae Zenaida auriculata Eared Dove 16 Metriopelia melanoptera Black-winged Ground-dove 1 Columba livia Rock Dove -f Psittacidae Cyanoliseus patagonus Burrowing Parakeet 2 Enicognathus ferrugineus Austral Parakeet 4 1 Caprimulgidae Caprimulgus longirostris Band-winged Nightjar 1 Furnariidae Geositta cunicularia Common Miner 3 Upucerthia dumetaria Scale-throated Earthcreeper 2 Eremobius phoenicurus Band-tailed Earthcreeper 2 Cinclodes fuscus Bar-winged Cinclodes 1 Cinclodes patagonicus Dark-bellied Cinclodes 4 Feptasthenura aegithaloides Plain-mantled Tit-Spinetail 1 Tyrannidae Neoxolmis rufiventris Chocolate-vented Tyrant 1 + 2 Muscisaxicola macloviana Dark-faced Ground-Tyrant 1 Fessonia rufa Patagonian Negrito 1 Hirundinidae TachyrAneta nieyeni Chilean Swallow 2 Notiochelidon cyanoleuca Blue-and-white Swallow 2 Troglodytidae Troglodytes musculus Tropical House-Wren + Sturnidae Mimus patagonicus Patagonian Mockingbird 1 Muscicapidae Turdus falcklandii Austral Thrush 2 -1- 2 December 2002 Short Communications 317 Table 1. Continued. Family Scientific Name Common Name McNutt 198U Peres and Peres 19853 This Study3 Montacillidae Anthus correndera Correndera Pipit 1 Fringillidae Agelaius thilius Yellow-winged Blackbird 1 Sturnella militaris Pampas Meadowlark + Sturnella loyca Long-tailed Meadowlark 1 Molothrus bonariensis Shiny Cowbird 1 Phyrgilus gayi Grey-hooded Sierra-Finch 1 Phyrgilus fruticeti Mourning Sierra-Finch 3 Phrygilus unicolor Plumbeous Sierra-Finch 1 Melanodera melanodera Canary-winged Finch 1 Diuca diuca Common Diuca-Finch 3 Sicalis luteola Grassland Yellow-Finch 3 Carduelis barbata Black-chinned Siskin 1 Unknown 8 0 9 Total individuals 53+ unknown 102 Total identified individuals 45 unknown 93 Minimum no. species 13 17 42 ^ Not listed is a Kelp Gull (Larus dominicanus) observed as prey of a juvenile Pallid Falcon on 10 March 1979 (Ellis and Glinski 1980) ^ Nonavian prey include only a lizard {Liolaemus sp.) and a small rodent. ^ An asterisk (*) in these columns identifies prey that had not achieved adult size. It is not certain that all 7 ibis were nestlings when taken. We identified feathers by placing materials from one eyrie in a shallow white box and from prior experience sorted the feathers into piles tentatively assigned to a like- ly taxon. A representative feather was grasped by forceps then compared to specimens of likely donor species. Once a good match for size, color, and pattern was found, the pile of feathers was sorted to remove any that did not represent this species and morph. Then the pro- cess was commenced anew. After one of us completed an identification for feathers without unique color patterns (and most passerine primaries do not have bold color patterns), a second person evaluated the feathers and confirmed or rejected the identification. The most diffi- cult materials often required an evaluation extending an hour or more before a certain match was found. Occa- sionally, a feather had to be washed and blow dried be- fore comparisons could be made. All identified materials were bagged separately and archived. In comparing feathers, it was often necessary to fan the wing or tail on the museum skin; to do so without tearing the skin required holding the appendage in alignment with the body while deflecting the tip of the feather with forceps. To aid in this process, we prepared flat skins with tail and one wing fanned for about 50 spe- cies while in Argentina. For some other species, we mere- ly placed wings, tail, feet, beak, and feathers representing all body areas in a plastic bag. All specimens were deliv- ered to the Argentine Museum of Natural Sciences, Buenos Aires, Argentina, where the most valuable were retained. The remainder were released for export and shipped to the U.S.A. Results From this and the previous two studies (McNutt 1981, Peres and Peres 1985), we have documented a fair variety of the prey taken by the peregrine in Patagonia. McNutt (1981) identified 13 prey species and two other genera. Peres and Peres (1985) noted 17 species, of which 10 were new (i.e., not previously noted by McNutt [1981]) The list from our study (Table 1) includes 42 species of which 32 were not previously recorded. In summary, at least 55 prey species in 26 families are known to be taken by peregrines in Patagonia. To this can be added the Kelp Gull {Larus dominicanus) recorded as prey of a ju- venile Pallid Falcon seen on 10 March 1979 (Ellis and Glinski 1980). Discussion Our list does not represent the full range of prey taken by the Peregrine Falcon in South America, because first of all, our study included only the southern fifth of the distribution of this race. Second, other prey species are known to be taken by this falcon in the Falkland Islands (Cawkell and Hamilton 1961) and in more northerly re- gions of South America (Hilgert 1988). Third, the low numbers of individuals taken for common and likely prey species (e.g., only two Austral Thrushes [Turdus falcklan- dii\ and one Patagonian Mockingbird {Mimus patagoni- cus) ) suggest that much more variety will come from con- 318 Short Communications VoT. 36, No. 4 tmued sampling. The relationship between diversity (i.e., the number of species detected) and sample size can be characterized as beginning with a 1:1 relationship but with the plot soon leveling off and finally approaching an upper asymptote (the true maximum in the number of species taken) only after several hundred prey are tal- lied (Heck et al. 1975, Marti 1987). At present, the total for all three studies is less than 200 individuals. Further, the two most commonly taken species in Table 1 were represented by only ca. 23 and 16 individuals; before the upper asymptote is reached, we expect that the number of individuals of the most commonly-taken prey will ex- ceed 100. The list of species taken (Table 1) suggests that the peregrine is prone to capture some prey on the ground. The gosling was observed being taken on a gravel bar. Surely, the young rheas, perhaps the tinamou, and likely some of the other young birds in Table 1 were taken on the ground. In addition to the avian prey tallied, we also recorded one lizard {Liolaemus sp.) and a small mammal (Rodentia ca. 40 g); both of these would likely have been taken on the ground (or kleptoparasitized) . In South America there is no competing large falcon that hunts terrestrial prey (i.e., like the Prairie Falcon [A mexicanus] m North America and the Saker Falcon [A cherrug] in Europe and Asia) that may constrain the Peregrine Fal- con to an aerial foraging niche, so it was to be expected that the peregrine in Patagonia would take quarry on the ground more frequently than do some other races. It is obvious from the variety of oceanic species on the Peres list (Peres and Peres 1985; Table 1) that their study emphasized coastal areas. Their sample was also from an area where the pallid morph is relatively common (C. Peres pers. comm.). Their results, in comparison with our list for inland eyries, where the pallid morph is rare, suggest that pallid and dark peregrines hunt different prey. To document this potential difference (i.e., to com- pare the foraging niches of two sympatric color morphs) will surely be an interesting ecological study. Pallid and normal birds appear very different in the field. Pallid birds are less conspicuous and gull-like when seen be- neath gray, overcast skies. We propose that the pallid morph may have evolved when conditions were right for a population of pale peregrines to live in isolation from the population of normal peregrines further north on mainland South America. Resumen. — ^A partir de publicaciones previas, conocemos menos de 100 item presa que representan poco menos de 25 especies de aves para el halcon peregrine de la Patagonia {Falco peregrinus cassini). Este estudio, incluye presas de 11 nidos, anade 93 presas identificadas repre- sentando 42 especies de aves (32 no reportadas previa- mente), un lagarto, y un mamifero pequeno. Aunque do- cumentamos una considerable variedad de presas para el halcon peregrine en esta region, la frecuencia con la cual nuevas presas fueron encontradas en cada nido visitado, sugiere que la diversidad de aves tomadas fue mucho mas grande que la que se describe aqui. Esta alta diversidad, en parte, resulta de variedades palidas y de color normal que ocupan nichos de forrajeo un tanto diferentes. [Traduccion de Cesar Marquez] Acknowledgments In 1980, the U.S. Air Force funded much of our travel Our 1981 work was largely funded by the National Geo- graphic Society and an anonymous philanthropist. We thank Peter Simpson, manager of Estancia Chacabuco, Rio Negro, Argentina, for logistical support. Phil Angle identified the lizard. Jamie Jimenez, Clayton White, and David Whitacre reviewed and improved the manuscript Our thanks to staff at the USNM for allowing space for our synoptic series and patience during the extended time devoted to identifying prey. Literature Cited Anderson, C.M. and D.H. Ellis. 1981. Falco kreyenborgi — a current review. Raptor Res. 15:33-41. , T.L. Maechtle, and W.G. Vasina. 1988. The southern breeding limit of the Peregrine Falcon. Pag- es 251-253 mT.J. Cade,J.H. Enderson, C.G. Thelan- der, and C.M. White [Eds.], Peregrine Falcon popu- lations: their management and recovery. The Peregrine Fund, Boise, ID U.S.A. Bielefeldt, J., R.N. Rosenfield, andJ.M. Papp. 1992. Un- founded assumptions about diet of the Cooper’s Hawk. Condor 94:427-436. Cawkell, E.M. and J.E. Hamilton. 1961. The birds of the Falkland Islands. Ibis 103:1-27. Ellis, D.H. and R.L. Glinskj. 1980. Some unusual re- cords for the Peregrine and Pallid Falcons in South America. Condor 82:350-351. and C. Peres. 1983. The Pallid Falcon Falco krey- enborgi is a color phase of the Austral Peregrine Falcon {Falco peregrinus cassini). Auk 100:269-271. , C.M. Anderson, and T.B. Roundy. 1981. Falco kreyenborgi: more pieces for the puzzle. Raptor Res. 15. 42-45. Heck, K.L., Jr., G. van Belle, and D. Simberloff. 1975 Explicit calculation of the rarefaction diversity mea- surement and the determination of sufficient sample size. Ecology 56:1459-1461. Hii.gert, N. 1988. Aspects of breeding and feeding be- havior of Peregrine Falcons in Guayllabamba, Ecua- dor. Pages 749-755 in T.J. Cade, J.H. Enderson, C.G Thelander, and C.M. White [Eds.], Peregrine Falcon populations; their management and recovery. The Peregrine Fund, Boise, ID U.S.A. Marti, C.D. 1987. Raptor food habits studies. Pages 67- 80 in B.A.Giron Pendleton, B.A. Millsap, K.W. Cline, and D.M. Bird [Eds.], Raptor management tech- niques manual. Natl. Wildl. Fed. Sci. Tech. Ser. No 10 . McNutt, J.W. 1981. Seleccion de presa y comportamien- to de caza del Halcon Peregrino {Falco peregrinus) en December 2002 Short Communications 319 Magallanes y Tierra del Fuego. An. Inst. Patagonia 12: 221-228. . 1984. A Peregrine Falcon polymorph: observa- tions of the reproductive behavior of Falco kreyenborgi. Condor 86:378—382. , D.H. Ellis, C. Peres, T.B. Roundy, W.G. Vasina, AND C.M. White. 1988. Distribution and status of the Peregrine Falcon in South America. Pages 237-249 in T.J. Cade, J.H. Enderson, C.G. Thelander, and C.M. White [Eds.], Peregrine Ealcon populations: their management and recovery. Peregrine Fund, Boise, ID U.S.A. Peres, M. and C. Peres. 1985. Peregrine project, Argen- tina (activities 1982). Bull. WWGBP 2:109-110. Risebrough, R.W., A.M. Springer, S.A. Temple, C.M. White, J.L.B. Albuquerque, P.H. Bloom, R.W. Fyfe, M.N. Kirven, B.A. Luscombe, D.G. Roseneau, M. Sander, N.J. Schmitt, C.G. Thelander, W.G. Vasina, AND W. Walker, II. 1990. Observaciones del halcon peregrino, Falco peregrinus subspecies, en America del Sur. Rev. Bras. Biol. 50:563-574. Sibley, C.G. and B.L. Monroe, Jr. 1990. Distribution and taxonomy of birds of the world. Yale Univ. Press, New Haven, CT U.S.A. Received 7 January 2001; accepted 10 July 2002 J Raptor Res. 36(4) :320-323 © 2002 The Raptor Research Foundation, Inc. An Elevated Net Assembly to Capture Nesting Raptors Eugene A. Jacobs^ Linwood Springs Research Station, 1601 Brown Deer Lane, Stevens Point, WI 54481 U.S.A. Glenn A. Proudfoot^ Caesar Kleberg Wildlife Research Institute, Mail Stop 218, Texas A&M University-Kingsville, Kingsville, TX 78363-8202 U.S.A. Key Words: capture techniques-, dho-gaza; drop net, elevated net, mechanical owl", mist net. The use of a Great Horned Owl {Bubo virginianus) to induce mobbing behavior, in combination with a net sys- tem, has become the technique of choice for capturing many species of nesting raptors (Hamerstrom 1963, Bloom 1987, Bloom et al. 1992, Steenhof et al. 1994, Ja- cobs 1996, McCloskey and Dewey 1999). However, some species, and some individuals trapped previously, may be reluctant to stoop at the owl when conventional tech- niques (i.e., placing owl and nets near ground level) are followed (Rosenfield and Bielefeldt 1993). In testing a mounted and live Great Horned Owl to induce mobbing behavior in American Kestrels {Falco sparverius) , Card et al. (1989) found that the closer an owl decoy was placed to the nest the more aggressive the kestrels became. A decoy (mounted or live) placed near the nest of “trap shy” Cooper’s Hawks {Accipiter cooperii) was more effective than conventional techniques, but this method required climbing tree(s) and was found to be time consuming (Rosenfield and Bielefeldt 1993). Here we describe an elevated net assembly that, in combination with an owl decoy, proved successful for trapping five species of rap- tors. MATERIAI.S AND METHODS The system consisted of an aluminum telescoping pole, a horizontal cross bar, two vertical upright poles, a dho- gaza type net (Clark 1981), and a mechanical owl (equipped with two radio controlled servos that provided movement to the owl’s head and perch) as described by Jacobs (1996; Fig. 1). The cross bar that supported the net assembly was a 3-m section of conduit tubing 2.5 cm in diameter. We used a commercially available conduit- bending tool to form 90° angles 25 cm from each end of the tubing, resulting in a “U ’’-shaped cross bar. A section of sheet metal (45 cm X 45 cm) was then bent around the cross bar to form a triangular bracket that enveloped the cross bar. Using sheet metal screws, this sheet metal envelope was attached to a 25-cm section of aluminum ‘ E-mail address: lsrs@raptorresearch.com 2 Present address: Department of Wildlife and Fisheries Sciences, Room 210, Nagle Hall, Texas A&M University, College Station, TX 77843-2258 U.S.A. tubing 3 cm diameter at the center of the cross bar to form a “T” that resembled the goal posts of most Amer- ican football fields. The “T” formation provided a stable base for the net assembly and the diameter of the base of the “T” allowed for quick insertion of the cross bar to a telescoping pole. Two 2.2-m sections of conduit tub- ing 1,8 cm in diameter were inserted into the upturned ends of the cross bar to form the uprights that supported the net. We used four tethered leads about 12 cm in length, four metal rings (shower curtain rings) about 5 cm in diameter, and either wooden clothespins or mag- nets to attach the four corners of the net to the tops and bottoms of uprights. When clothespins were used, the free ends of the tethered leads were held by clothespins taped to the top two ends of the uprights. When magnets were used, a metal washer tied to the free end of the tethered leads provided support for the top two corners of the net. The bottom two leads were attached to the bottom of the uprights and remained stationary. Two metal rings were placed on each side of the net, one in the top corner and one at the half-way point. The rings attached the net to the uprights allowing the net to re- main open during set-up, The rings also allow the net to drop freely to the cross bar when a bird made contact Once contact is made with the net, the net slides down the uprights and creates a pocket entrapping the bird. We inserted a 10-cm section of tubing (vertical, 1.1 cm in diameter) into the bottom of the owl’s perch and at- tached a corresponding 30-cm section of tubing (0.9 cm diameter) to the “T.” This was positioned so that about 15 cm of the tubing extended above the center of the cross bar to support the owl’s perch. The 2-8 m telescoping pole allowed us to adjust the height of the net assembly to suit the nest site. Three guy lines attached to the base of the “T” bracket stabilized the net assembly when the telescoping pole was extended to a height greater than 3 m. We used a 1-m section of 1 .3 cm diameter conduit tubing, hammered flat at one end and cut to form a point, as support stakes for the telescoping pole and 0.5 m stakes (fashioned the same way) for the guy lines. Total cost of materials was ca. $250 (U.S.); telescoping pole ($180), net (.$20), cross bar ($10), uprights ($10), anchor stake for telescoping pole ($10), guy lines and support stakes (.$10), and miscella- neous material ($10). Because vegetation structure varied among nest sites, we elevated the net assembly to the maximum length of the support pole (8 m) or to the highest feasible level where tree branches blocked a greater extension. The net assembly was placed within 50 m of the nest tree, in 320 December 2002 Short Communications 321 Figure 1. Elevated dho-gaza net assembly used to trap small- and medium-sized raptors near their nests. view of the nest and with a clear flight path around the net. Great Horned Owl vocalizations and conspecific calls were utilized to lure nesting raptors to the net system (Bloom et al. 1992). A concealed observer was positioned nearby (<100 m) to operate the radio controls to the mechanical owl and record the sex of the adults when they were detected in the nest area (<50 m of nest). The observer (s) was able to sex all American Kestrels and Sharp-shinned Hawks (A. striatus) from plumage char- acteristics or relative size compared to their mates (Clark and Wheeler 1987). We reported trapping success by us- ing the number of birds trapped, divided by the number of birds “tested” (birds detected within 50 m of the nest) multiplied by 100. When trapping Ferruginous Pygmy-Owls {Glaucidium brasilianum) , we modified this system by downsizing the net assembly (cross bar and uprights) to 2 m X 1 m, replacing the dho-gaza with a shortened 2-shelf mist net, and replacing the mechanical owl with a conspecific mounted decoy. This modified setup was used as de- scribed above (using lure to induce mobbing behavior from nesting owls), or simply placed in front of the nest cavity entrance to capture the adults as they entered the cavity to feed nestlings. The elevated mist net was placed about 1.5 m from the nest cavity’s entrance. Results and Discussion During the breeding seasons (1994-2001) this tech- nique was “tested” on five species of small to medium sized raptors. Overall, we successfully captured 73% (113 of 154) of the individuals “tested.” Our trapping success 322 Short Communications VoL. 36, No. 4 Table 1. Comparison of capture rates of elevated net with a mechanical owl, normal net with a mechanical owl, and normal net with a live owl as a trapping technique for raptors. Elevated Net with Mechanical Owl (This Study) Ground-level with Mechanical Owl (Jacobs 1996) Ground-level with Live Owl Red-shouldered Hawk 65% (30 of 46) 54% (15 of 28) 75% (199 of 264)1 Sharp-shinned Hawk 81% (34 of 42) 77% (48 of 62) — American Kestrel 70% (21 of 30) — 71% (15 of 21)2 97% (115 of 118)1 Cooper’s Hawk 67% (2 of 3) 60% (3 of 5) 52% (32 of 62)1 Ferruginous Pygmy-Owl 79% (26 of 33) — — ' Data from Bloom et al. 1992. ^ Data from Steenhof et al. 1994. was generally similar or slightly lower than studies using a ground-level net set (height ^2.5 m) and live owl (Bloom et al. 1992, Steenhof et al. 1994) and slightly better than Jacobs (1996) found when using a normal net set and a mechanical owl (Table 1 ) . Bloom (1987) occasionally found it was difficult or time consuming to capture both sexes of Northern Gos- hawks (A. gentilis). Females were usually caught within 15 min, but males were often not captured. He speculated that the male is less aggressive toward the owl during the post-fledging period or is away from the nest (hunting) and is unaware of the owl’s presence. With Sharp- shinned Hawks and American Kestrels, we found of the birds that were present (assumed to have seen the owl), females responded more aggressively toward the owl mount than did males. Trapping success for female Sharp-shinned Hawks and female American Kestrels were 91% (21 of 23) and 79% (11 of 14), respectively. Males of both species occasionally showed a reluctance to “stoop” at the owl, resulting in a trapping success of 68% (13 of 19) for male Sharp-shinned Hawks and 63% (10 of 16) for male American Kestrels. The escape rate (per- centage of birds that hit the dho-gaza net, but were not captured) was 1.8% (2 of 113). Benefits of this technique include: (1) a high response rate; (2) a low escape rate; (3) the ability to readily adjust net height; (4) minimal space (especially when com- pared to a 12-m mist net) is required for setup; (5) this technique kept the net off the ground, and hence, re- quired less preparation time to reset the net after a cap- ture when compared to a ground-level dho-gaza net. Albanese and Piaskowski (1999) and Stokes et al. (2000) have described similar techniques that use elevat- ed mist nets to study birds that spend the majority of their time in woodland canopies. However, those studies used conventional mist nets in a manner designed for continuous use at one location. Our technique was de- signed to capture specific birds near their nest sites with a mobile apparatus that allows researchers to visit multi- ple locations with a minimal amount of setup time (ca. 15 min). Even though we conducted 113 captures without any visible injury, there is the potential for serious injury to species that “stoop” at higher speeds and are of greater mass than the Red-shouldered Hawk. A net that com- pletely releases from the uprights (Bloom 1987) as well as other modifications may be necessary to accommodate the force of a larger raptor hitting the net. Alternatively, because the “cut-down” mist net was successful in cap- turing Ferruginous Pygmy-Owls as they approached or exited the entrance to their nest cavity, this technique should be effective for trapping most small cavity-nesting birds without modification. Resumen. — Debido a su alta tasa de exito, el uso de una red de niebla en combinacion con un senuelo de Bubo virginianus, es una de las tecnicas populaces mas usadas para capturar rapaces durante su anidacion. Sin embar- go, el protocolo estandar para esta tecnica limita su efec- tividad a aves que tiene la habilidad de volar cerca al nivel del suelo (<3 m) para atropellar al senuelo. Con el fin de proveer una tecnica alternativa que pueda mejorar la tasa de captura en algunas situaciones, construimos y probamos un ensamblaje de redes elevado consistente de un poste telescopico de aluminio, una barra cruzada hor- izontal, dos postes verticales rectos, una red de niebla, y un buho raecanico. Desde 1994-2001, el exito en la cap- tura de cinco especies de rapaces de talla pequena a me- diana fue 73% (113/154 intentos). Debido a su adapta- bilidad, el alto exito de captura y el bajo costo, este sistema puede ser una herramienta beneficiosa para la investigadon de las aves. [Traduccion de Cesar Marquez] Acknowledgments We thank D. Haessly and R. Rosenfield for providing information on the effectiveness of our system on Amer- ican Kestrels and Cooper’s Hawks, respectively. J. Runke December 2002 Short Communications 323 provided the drawing in Figure 1. J. Bielefeldt, P. Bloom, J Marks, M. McMillian, and B. Woodbridge provided helpful suggestions for improving this manuscript. Literature Cited Albanese, G. and V.D. Piaskowski. 1999. An inexpensive elevated mist net apparatus. N. Am. Bird Bander 24: 129-134. Bloom, P.H. 1987. Capturing and handling raptors. Pag- es 99-123 in B.A. Millsap, K.W. Cline, B.G. Pendleton, and D.A. Bird [Eds.] , Raptor management techniques manual. Natl. Wildl. Fed., Washington, DC U.S.A , J.L. Henckel, E.H. Henckel, J.K. Schmutz, B. Woodbridge, J.R. Bryan, R.L. Anderson, P.J. De- TRicH, T.L. Maechtke, J.O. McKinley, M.D. McCrary, K. Titus, and P.F. Schempf. 1992. The dho-gaza with Great Horned Owl lure: an analysis of its effectiveness in capturing raptors./. Raptor Res. 26:167—178. Clark, W.S. 1981. A modified dho-gaza trap for use at a raptor banding station. J. Wildl. Manage. 45:1043- 1044. and B.K. Wheeler. 1987. A field guide to hawks in North America. Houghton Mifflin Co., Boston, MA U.S.A. Card, N.W., D.M. Bird, R. Densmore, and M. Hamel 1989. Responses of breeding American Kestrels to live and mounted Great Horned Owls. J. Raptor Res. 23 99-102. Hamerstrom, F. 1963. The use of Great Horned Owls in catching marsh hawks. Proc. XIII Int. Ornithol. Congr 13:866-869. Jacobs, E.A. 1996. A mechanical owl as a trapping lure for raptors. / Raptor Res. 30:31-32. McCloskey, J.T. and S.R. Dewey. 1999. Improving the success of a mounted Great Horned Owl lure for trap- ping Northern Goshawks./. Raptor Res. 33:168-169 Roseneield, R.N. and j. Bielefeldt. 1993. Trapping tech- niques for breeding Cooper’s Hawks; two modifica- tions. /. Raptor Res. 27:171-172. Steenhof, K., G.P. Carpenter, and J.C. Bednarz. 1994 Use of mist nets and a live Great Horned Owl to cap- ture breeding American Kestrels. J. Raptor Res. 28. 194-196. Stokes, A.E., B.B. Schultz, R.M. Degraaf, and C.R Griffin. 2000. Setting mist nets from platforms in the forest./. Field Ornithol. 71:57-65. Received 31 December 2001; accepted 13 July 2002 J Raptor Res. 36(4):324-327 © 2002 The Raptor Research Foundation, Inc. Florida Bald Eagle (Haliaeetus leucocephalus) Egg Characteristics M. Alan Jenkins, 1 Steve K. Sherrod, David A. Wiedenfeld, and Donald H. Wolfe, Jr. George Miksch Sutton Avian Research Center, RO. Box 2007, Bartlesville, OK 74005 U.S.A. Key Words: Bald Eagle, Haliaeetus leucocephalus; eggs\ eggshell characteristics', pesticides. The Bald Eagle {Haliaeetus leucocephalus) , south of the 40th parallel, was declared endangered by the U.S. Fish and Wildlife Service in 1967 owing to an observed pop- ulation decline in the 48 contiguous states, which was hrst noticed in Florida. A population of eagles in western peninsular Florida showed a sharp decline by 1957-58 (Broley 1958), and in eastern-central Florida the species declined by two-thirds between 1951-61 (Howell 1963). Those declines were later attributed to poisoning or the weakening effects on eggshells caused by chlorinated hy- drocarbon pesticide residues, notably DDE (Nisbet 1989). Eggs with shells thinned by DDE broke before hatching, destroying the egg and resulting in lowered productivity. From 1984—91 biologists with the George Miksch Sutton Avian Research Center collected entire clutches of Bald Eagle eggs from Florida nests as part of a project to restore nesting populations of this species to the southeastern U.S. (Sherrod et al. 1989). The eggs were incubated and hatchlings were reared in Oklahoma, and the young were released by hacking in eight south- eastern states. This paper describes the physical charac- teristics of 395 Bald Eagle eggs and the analysis of pesti- cide residues in 15 unhatched eggs collected during that project. We report the results of testing for correlations between egg characteristics, location of collection, year, sex of the eagle hatched, and information on organo- chlorine pesticides and eggshell thickness. Study Area and Methods Eggs were collected from nine counties in north-cen- tral Florida (Fig. 1). At collection, each egg was marked with a unique number, measured for length and breadth with a caliper to the nearest 0.01 mm, and weighed to the nearest 0.1 g (balance) or 0.01 g (electronic scale). After incubation, the shells of hatched eggs were air- dned and their thickness measured. Unhatched eggs were frozen for as long as a year, a method of preserva- tion which does not change the level of pesticide con- centrations even if thawing and microbial degradation occur (Stickel et al. 1984). Unhatched eggs were opened, the contents removed, the shells were air-dried, and the shell thickness measured. Thickness measurements were made on fragments of shell from three evenly spaced places near the equator. Fragments were clamped in a hemostat and viewed with a binocular microscope at lOOX. One microscope eye- 1 E-mail address: alanjenkins@ou.edu piece was equipped with a micrometer calibrated with a stage micrometer. Each shell sample was viewed with the microscope on its edge, the number of calibrations span- ning the fragment recorded, and the thickness calculated using a constant conversion factor obtained from the cal- ibration (Enderson et al. 1982). The three thickness mea- surements of each egg were averaged. Shell membranes of hatched eggs usually shrank, be- coming detached from the shell and so they could not be measured as part of the shell’s thickness. Shell mem- branes from unhatched eggs sometimes remained at- tached after drying and, where possible, shells were mea- sured with attached membranes, then the membranes removed and shell thickness re-measured. This gave shell thicknesses with and without membranes for 37 eggs. This allowed us to correct for loss of shell thickness caused when shell membranes become detached. De- tachment of the shell membranes after drying slightly reduces the measured total shell thickness because the mammillary tips from the shell come away attached to the outer shell membrane during detachment (Terepka 1963). Calculated egg volumes presented here include the shell volume, a method preferable to trying to determine inside volume (Stickel et al. 1973). The calculated vol- ume formula was determined to be within —8 to +7 per- cent of measured volumes for this species (Stickel et al 1973). Shell thicknesses may decrease during embryo de- velopment as the embryo uses calcium from the shell for bone formation (Romanoff and Romanoff 1949). How- ever, most of the decrease apparently occurs in the mam- millae core (Bond et al. 1988) and does not contribute significantly to shell thinning (Bunck et al. 1985). Cal- culated egg volume (V) was estimated using the formula V = 0.508LB^ (Stickel et al. 1973), where L = length of egg (cm) and B = breadth of egg (cm) . Calculated fresh egg mass (M) was estimated as M = 0.56227LB^. The coefficient of 0.56227 yielded the best estimate of hatch- ing date, assuming a period from the beginning of in- cubation to pip of 33.5 d. Unhatched eggs were considered fertile if they showed any sign of embryonic development upon dissection Hatch order within clutches was recorded based on pip- ping date and time. Hatch order was assumed to be iden- tical with laying order because female eagles lay eggs a few days apart and begin incubation with the first eggs, making hatching within a clutch asynchronous (Gerrard and Bortolotti 1988). However, time of pipping was only accurate within ca. 12 hr because eggs were checked twice daily for pipping. Sex of fledgling-age eagles was determined from mea- surements of bill depth and toepad length following the methods of Walborn (1991). Egg length, breadth, volume, length/breadth ratio, cal- culated fresh mass, and shell thickness were checked for 324 December 2002 Short Communications 325 Key to Counties A = Alachua O = Osceola H = Highlands Po = Polk La = Lake Pu = Putnam Le = Levy S - Seminole M = Marion V = Volusia Figure 1. Florida Bald Eagle study area and counties of egg collections. correlations with year and the latitude of the Florida county where collected to determine any latitude trend. Analysis of variance was used to compare shell thickness, volume, and hatch order by sex; volume and shell thick- ness (for all and for both sexes) by hatch order; and volume and shell thickness with fertility. Statistical tests were made using Systat statistical package, version 7.0. Organochlorine and polychlorinated biphenyl pesti- cide residues of the contents of 15 unhatched eggs were analyzed by Hazleton Labs using Food and Drug Admin- istration (1973) methods. The reported residues are based on calculated fresh egg masses. Results No significant correlations were found between egg length, breadth, volume, length/breadth ratio, calculat- ed fresh mass, and shell thickness and the year of collec- tion or the latitude of the Florida county where collected (Table 1). There was a significant increase of eggshell thickness from 1984-91 (T - 0.073, df = 392, P < 0.05) However, the increase was not signihcant if data for 1984 were excluded. The 1984 sample size {N = 18) is biased by the inclusion of an unusually thin-shelled two-egg clutch with very high pesticide residues. One of these two eggs was found broken in the nest; the other failed to develop. The eggshells we collected were 4.5% thinner than the pre-1947 Florida sample {N = 211) measured by Anderson and Hickey (1972). Residue analysis results for 23 chlorinated hydrocar- bons, percent moisture and percent lipids for 15 un- Table 1 . Summary of Florida Bald Eagle egg characteristics. Length (L) (mm) Breadth (B) (mm) Calculated Volume (mm^) L/B Ratio Calcuiated Fresh Mass ( g) Shell Thickness Without Membranes (mm) N 392 392 392 392 392 395 Minimum 64.38 48.34 82.30 1.166 90.92 0.312 Maximum 79.50 59.88 140.20 1.517 154.92 0.553 Mean 71.04 55.37 111.08 1.284 122.71 0.453 Standard Deviation 2.732 1.690 9.233 0.051 10.20 0.035 326 Short Communications VoL. 36, No. 4 Table 2. Summary of chlorinated hydrocarbon pesti- cide residues in 15 unhatched Florida Bald Eagle eggs, 1984-91. Data are corrected to calculated fresh egg mass- es Mean (ppm) Maxi- mum (ppm) Mini- mum (ppm) Stan- dard Devia- tion DDE 2.47 10.10 0.66 2.35 DDD 0.05 0.58 0.0 0.14 DDT 0.24 3.89 0.0 0.94 HCB ND* Alpha-BHC ND Gamma-BHC (lindane) ND Beta-BHC ND Heptachlor ND Aldrin ND Octachlorostyrene ND Heptachlor epoxide 0.01 0.07 0.00 0.02 Oxychlordane 0.06 0.26 0.00 0.06 Gamma-chlordane 0.04 0.15 0.00 0.04 Alpha-chlordane ND Transnonachlor 0.29 1.06 0.00 0.25 Mirex ND Dieldrin 0.10 0.66 0.00 0.15 Endrin ND Methoxychlor ND Toxaphene ND PCB 1260 5.13 15.47 0.00 5.50 PCB 1248 0.02 0.24 0.00 0.06 PCB 1254 5.82 28.15 0.00 8.03 Percent moisture 80.78 83.60 76.00 2.04 Percent lipids 5.15 12.00 2.20 2.19 ND = None detected above detection limit of 0.1 ppm or lower. hatched eggs are presented in Table 2 as benchmark data for possible future information on this population. Discussion Bald Eagle body size is known to increase with increas- ing latitude (Stalmaster 1987), but we found no correla- tion of egg size characteristics with the small latitude span we sampled. We speculated that eggs laid later in a clutch might be smaller than the first, but our data failed to detect such a trend. Likewise, sex of the eagles hatched did not correlate with size of the egg or hatch order. We expected an increase in eggshell thicknesses and lessening of chlorinated hydrocarbon residues during the years of our study because most uses of DDT in the U.S.A. have been banned; however, our thickness data did not show a change over time during the period of our study. The residues for our eggs, except one egg with high DDE residues, are below the threshold levels indicated by some authors as sufficient to affect raptor productivity (Fyfe et al. 1988, Nisbet 1989, Risebrough 1989). Mean productivity of Florida Bald Eagles during the years of egg collection was 1.10 young/occupied territory (Nes- bitt et al. 1998), and above the minimum of 1.0 young/ occupied territory that Wiemeyer et al. (1993) consid- ered as representing a healthy population. Based on these data, we believe that by 1984 the population of nest- ing Bald Eagles in the areas of Florida from which we collected eggs were reproducing at a rate that met the criteria suggested by Wiemeyer et al. (1993). This pop- ulation was relatively free of pesticide contamination and that eggshell thinning was no longer a problem. Resumen. — Las medidas y calculos de las medidas de 395 huevos de aguilas calvas {Haliaeetus leucocephalus) colec- tados en Florida desde 1985-91 fueron evaluadas por me- dio de relacion estadistica para el ano de coleccion, la- titud del condado donde fueron colectados, sexo del polluelo y orden de salida del huevo. No se encontraron correlaciones significantes. Se hicieron analisis de resi- duos de pesticidas organoclorados y bifenil policlorina- dos en 15 huevos sin empollar. Con excepcion de una nidada, los residues organoclorados fueron bajos. [Traduccion de Cesar Marquez] Acknowledgments Cooperation for this project was given by the Florida Game and Fresh Water Fish Commission, in particular Don Woods and Steve Nesbitt; The University of Florida, especially Michael Collopy, Petra Bohall Wood, and Rob- ert Rosen; and the U.S. Fish and Wildlife Service’s En- dangered Species Program and Regions 2 and 4. Grateful thanks are due to our numerous funders and the Sutton Board of Directors. Many Sutton employees worked tire- lessly to make the restoration project successful, Alan Be- ske, Gwyn McKee, and Sheryl Tatora in particular. James Enderson loaned eggshell measuring equipment. The manuscript was greatly improved by comments from Stanley N. Wiemeyer, James C. Bednarz, Sally Jenkins, and two anonymous reviewers. Literature Cited Anderson, D.W. and J.J. Hickey. 1972. Eggshell changes in certain North American birds. Pages 514-540 in K.H. Voous [Ed.], Proc. XV Int. Ornithol. Congr. Leiden, Germany. Bond, G.M., R.G. Board, and V.D. Scott. 1988. A com- parative study of changes in the fine structure of avian eggshells during incubation. Zool.J Linn. Soc. 92:105- 113. Broley, C.L. 1958. The plight of the American Bald Ea- gle. Audubon 60:162-163, 171. Bunck, M.M., J.W. Spann, O.H. Patee, and W.J. Fleming 1985. Changes in eggshell thickness during incuba- tion: implications of evaluating the impact of organ- ochlorine contaminants on productivity. Bull. Environ Contam. Toxicol. 35:173-182. Enderson, J.H., G.R. Craig, W.A. Burnham, and D.D. December 2002 Short Communications 32V Berger. 1982. Eggshell thinning and organochlorine residues in Rocky Mountain peregrines, Falco peregti- nus, and their prey. Can. Field-Nat. 96:255-264. Food and Drug Administration. 1973. Pesticide analyt- ical manual. Vol. 1, Washington, DC U.S.A. Fyfe, R.W., R.W. Risebrough, J.G. Monk, W.M. Jarman, D.W. Anderson, L.F. Kief, J.L. Linger, I.C.T. Nisbet, W. Walker, II, and BJ. Walton. 1988. DDE, produc- tivity, and eggshell thickness relationships in the ge- nus Falco. Pages 319—335 in T.J. Cade, J.H. Enderson, C.G. Thelander, and C.M. White [Eds.], Peregrine Falcon populations: their management and recovery. The Peregrine Fund, Boise, ID U.S.A. Gerrard, J.M, and G.R. Bortolotti. 1988. The Bald Ea- gle. Smithsonian Institution, Washington, DC U.S.A. Howell, J.C. 1963. The 1961 census of some Bald Eagle nests in east-central Florida. Auk 79:716-718. Nesbitt, S.A., M.A. Jenkins, S.K. Sherrod, D.A. Wood, A. Beske, J.H. White, P.A. Schulz, and S.T. Schwik- ert. 1998. Recent status of Florida’s Bald Eagle pop- ulation and its role in eagle reestablishment efforts in the southeastern United States. Proc. Annu. Conf. Southeast. A.ssoc. Fish Wildl. Agencies 52:377-383. Nisbet, I.C.T. 1989. Organochlorines, reproductive im- pairment, and declines in Bald Eagle (Haliaeetus leu- cocephalus) populations: mechanisms and dose-re- sponse relationships. Pages 483-489 in B.-U. Meyburg and R.D. Chancellor [Eds.], Raptors in the modern world. WWGBP, Berlin, Germany. Risebrough, R.W. 1989. Toxic chemicals and birds of prey: discussion at Eilat in 1987. Pages 515-525 in B.-U. Meyburg and R.D. Chancellor [Eds.], Raptors in the modern world. WWGBP, Berlin, Germany. Romanoff, A.L. and AJ. Romanoff. 1949. The avian egg. John Wiley and Sons, Inc., New York, NY U.S.A. Sherrod, S.K., M.A. Jenkins, G. McKee, D.H. Wolfe, Jr , and S. Tatom. 1989. Restoring nesting Bald Eagle Haliaeetus leucocephalus populations to the southeast- ern United States. Pages 353-357 in B.-U. Meyburg and R.D. Chancellor [Eds.], Raptors in the modern world. WWGBP, Berlin, Germany. Stalmaster, M.V. 1987. The Bald Eagle. Universe Books, New York, NY, U.S.A. Stickel, L.F., S.N. Wiemeyer, and LJ. Blus. 1973. Pesti- cide residues in eggs of wild birds: adjustment for loss of moisture and lipid. Bull. Environ. Contam. Toxicol 9 193-196. Stickel, W.H., L.F. Stickel, R.A. Dyrland, and D.H. Hughes. 1984. Comparison of methods of preserving tissues for pesticide analysis. Environ. Monit. Assess. 4- 113-118. Terepka, A.R. 1963. Organic-inorganic interrelationships in avian egg shell. Exp. Cell Res. 30:183-192. Walborn, E.B. 1991. Use of morphometric measure- ments in determining sex of southern Bald Eagles. M.S. thesis, Oklahoma State University, Stillwater, OK U.S.A. Wiemeyer, S.N., C.M. Bunck, and C.J. Stafford. 1993. Environmental contaminants in Bald Eagle eggs (1980-1984) and further interpretations of relation- ships to productivity and shell thickness. Arch. Environ. Contam. Toxicol. 24:213—227. Received 7 January 2002; accepted 25 July 2002 J Raptor Res. 36(4):328-331 © 2002 The Raptor Research Foundation, Inc. Osprey Ecology in the Mangroves of Southeastern Brazil Robson Silva e Silva^ Rua Sdo Jose 48/31, 11040-200, Santos, SP, Brazil Fabio Olmos Largo do Paissandu 100/4C, 01034-010, Sdo Paulo, SP, Brazil Key Words: Osprey; Pandion haliaetus; Brazil; diet; man- grove, wintering ecology. The Osprey {Pandion haliaetus) is a widely distributed raptor found in every continent but Antarctica. The Eur- asian nominate and North American subspecies {P. h. car- ohnensis) make extensive migrations, the latter being a trans-equatorial migrant wintering in United States (Flor- ida), Mexico, Central America and South America from Colombia to Argentina and Chile (Poole 1989, del Hoyo et al. 1994, Martell et al. 2001). In Brazil, Ospreys have been recorded in almost every state and month, but most records were made between September and April (Sick 1997). The Amazon basin, where reports are most common, seems to be the main wintering area for Ospreys in the country, but there are also clusters of records in coastal localities (Sick 1997, Olmos and Silva e Silva 2001). Band recoveries made in Brazil show the birds come from the Mid-Atlantic and northeastern USA (Poole and Agler 1987). Despite being a well-known species in its breeding grounds (del Hoyo et al. 1994), there is little information on the wintering ecology of Ospreys in South America (Saggese et al. 1996), although detailed studies have been made in Africa (Prevost 1982, Boshoff and Palmer 1983). Ospreys spend most of the year (six or more months) in their wintering grounds, so information on their ecology during this period is critical for a clearer understanding of their life-histories and the pressures laced by the birds outside the breeding season. Here, we provide an account of the ecology of Ospreys using a mangrove eco.system in southeastern Brazil, the first such study in South America. SruDY Area AND Methods The study was conducted in the mangrove ecosystem that covers the estuarine area between Sao Vicente Island and the mainland, in the coast of Sao Paulo state, south- eastern Brazil (ca. 23°53'S, 46°23'W). This area, belong- ing to the Santos and Cubatao counties, is part of the major mangrove ecosystem located in the region known as Baixada Santista, covering 120 km^, and located in one of the most populated and developed areas in Brazil (Lamparelli 1999). The local climate is hot and humid, with annual rainfall ranging from 2000 to over 2500 mm. ^ E-mail address: rsilvaesilva@uol.com.br Winter is the driest season, the lowest rainfall occurring in July-August, the highest values occurring between Sep- tember-March (Olmos and Silva e Silva 2001). For a gen- eral description of the area’s geography and environ- ment, see CETESB (1991) and Olmos and Silva e Silva ( 2001 ). The main feature in the study area is a channel (Pia- gaguera channel) bisecting it and connecting Santos Bay to the estuarine area inland. This broad (ca. 1 km wide) channel is mostly man-made and regularly dredged to allow for the passage of cargo ships serving the local steel and fertilizer plants. Because of dredging and silting, there are large mudflats along the channel. Ten naviga- tion buoys and six concrete towers dot the entire length of the channel and are used as perches by feeding Os- preys. Prey remains accumulate on top of these buoys and towers, making collecting the collection of dietary data simple. Three rivers (Quilombo, Casqueiro, and Cu- batao) empty into the channel. Diet and behavioral data were collected as part of a broader study of the mangrove avifauna conducted from March 199T-July 2000 during 231 field-days (Olmos and Silva e Silva 2001). Data on Osprey abundance through- out the year and habitat use were gathered between Au- gust 1995-November 1996, when we made 36 standard- ized bird censuses by boat along a 19.25 km transect covering all habitats present (Olmos and Silva e Silva 2001 ). Habitat classification was adapted from the mapping made by Sao Paulo state’s environmental agency (CE- TESB 1991) and represents a gradient of mangrove tree cover. Habitat categories sampled were: mangrove forest with mostly touching canopies and complete cover, de- graded mangrove with many gaps between trees making a patchwork of clearings, mangrove degraded by pollu- tion, groups of mangrove trees, and large areas of her- baceous growth dotted by dwarfed mangroves, a mosaic of mudflats (with scattered mangrove trees), and ex- posed mudflats without any tree cover. For a detailed de- scription of habitats see Olmos and Silva e Silva (2001) Feeding Ospreys were observed opportunistically with the help of spotting scopes and binoculars with the aim of identifying the captured prey. Prey remains accumu- lated on buoys and towers were recovered during regular checks and identified to the lowest possible taxonomic level by comparison to reference material. Results Ospreys were first recorded along the Piagaguera chan- nel in 1986 (Olmos 1989). Interestingly, the species was not found there earlier in the 20th century (Luederwaldt 328 December 2002 Short Communications 329 Figure L Monthly numbers of Ospreys censused in the Santos-Cubatao mangroves, southeastern Brazil, between August 1995-November 1996. 1919). It is almost impossible to tell adult birds from young ones under field conditions (Boshoff and Palmer 1983), but we never observed newly-fledged Ospreys, with their characteristic speckled plumage. Ospreys are recorded in Santos-Cubatao throughorxt the year. Census data show a peak during the austral sum- mer and a minimum of three birds during the winter (Fig. 1). Most (47%) records made during the censuses were along the Cascalho River, where channel margins are dominated by mangroves degraded by pollution (Ol- mos and Silva e Silva 2001). Table 1. Prey items of wintering Ospreys in the man- groves of Santos and Cubatao, Brazil. Unidentified fish were excluded. Species N Percent Mullets Mugil sp. 69 76.7 Rhomboid mojarra Diapterus rhombeus 16 17.8 Snooks Centropomus sp. 2 2.2 Common trahita Hoplias malabaricus 1 1.1 Brazilian mojarra Eugerres brasilianus 1 1.1 Brazil geophagus Geophagus brasiliensis 1 1,9 Total 90 100 From year-round census data. Ospreys had higher lin- ear densities along rivers bordered by mangrove forest (x = 0.71 birds/km) and mangroves degraded by pollution (0.70 birds/km) compared to mudflats (0.2 birds/km) and degraded mangrove (0.01 birds/km, Fij 440 = 14 09, P < 0.0001). Densities in mangrove forest and mangrove degraded by pollution were not different, as well as dif- ferences in densities between mudflats and degraded mangroves (Tukey HSD test, both P < 0.001). Most fish seem to be captured alive, but in November 1996, we did record an Osprey taking a dead fish floating on the surface. The most important prey were mullets (Mugil spp.) (76.7%) and rhomboid mojarra {Diapterus rhomheus) (17.8%; Table 1). Only two freshwater fish, common trahita {Hoplias malabaricus) and Brazil geopha- gus {Geophagus brasiliensis) were recorded. One mullet {Mugil platanus) left whole on a tower was 41. .5 cm TL and weighed 590 g, while a Diapterus rhombeus found m the same circumstances was 30 cm and 490 g. Based on prey remains and observations, fish smaller than 20 cm TL seem to be rare in the diet of Ospreys in southeastern Brazil. We rarely observed Ospreys hovering or flying low over the water before capturing a fish. Once, an Osprey caught a dying fish on the surface while “hovering.” In most instances, the birds would soar over the water and plunge dive rapidly after locating a fish. It was common for the birds to abort several dives (up to nine in a row) before actually touching the water. Ten out of 14 plunge dives by different birds were suc- cessful (71%), with nine fish carried away and one falling back to the water. After taking a fish, the birds would fly 330 Short Communications VoL. 36, No. 4 to one of the available towers, buoys, trees, or when the prey was too heavy, even a mudflat. All hsh were eaten head first, discarding opercula and gills. Mullet viscera were also discarded. Usually only the tail would remain but, sometimes, we would find the posterior half of a large mullet. After feeding, the bird would fly dragging its feet in the water for over TO m. Sometimes up to three birds were seen fishing in the same small area. When one was successful the others would chase it amid much call- ing, trying to steal its fish. A curious behavior was observed on 22 August 1997, at the Pia^aguera channel. One Osprey, while holding a tree branch, was seen flying and calling after another in- dividual. The second bird would perch, also calling. The first bird flew out of sight but came back quickly, always calling, without the branch. Both birds then began to soar, always calling, making several aborted dives. After one Osprey finally captured an unidentified fish, it flew away being followed by the other bird. An Osprey flying while holding a tree branch was also seen on 18 July 2000; this bird was chasing a Crested Caracara (Polyborus plancus) , but we soon lost sight of them. Contrary to the behavior reported in some wintering areas (Boshoff and Palmer 1983), Ospreys were very vo- cal in Santos-Cubatao throughout the year, especially when two birds came close to each other. It was common to see two Ospreys soaring together with Black Vultures ( Coragyps atratus) while calling. They would also call when we approached the perch of a feeding bird, which caused It to fly with its prey. Yellow-headed Caracaras {Milvago chimachima) benefit from Ospreys by scavenging fish remains from the feed- ing perches, sometimes waiting beside a feeding Osprey. Crested Caracaras are more aggressive and actively try to steal fish from the Ospreys; sometimes up to four cara- caras may join in a chase after an Osprey carrying a fish, but we never saw them being successful. Interestingly, Great Egrets {Casmerodius alhus) would expel Ospreys Irom their feeding perches, while Kelp Gulls {Larus dom- imcanus), known to steal Ospreys’ food in South Al'rica (Boshoff and Palmer 1983), ignore them in Brazil. Ospreys did not seem to be actively persecuted by local people and the many fishermen using the mangroves. Nevertheless, on 26 December 1996, a female was found with a wounded wing, perhaps the result of being shot. This bird is now in the ornithological collection of the Museu de Zoologia da Universidade de Sao Paulo (MZUSP 74346). Discussion The lack of records of newly-fledged Ospreys with their characteristic speckled plumage suggests only birds over 1 yr-old migrate to the study area. The year-round pres- ence of Osprey in Brazil raises the possibility the birds might breed in South America, but no evidence has been found so far (Sick 1997). Nevertheless, nesting in a win- tering area has been documented in South Africa (Dean and Tarboton 1983). The presence of Ospreys in Brazil during April-August show non-breeders, juveniles, stay in their wintering range. Although Sick (1997) attested, without details, that Os- preys sometimes take mammals and birds when winter- ing, the only documented food items of Ospreys in Brazil were fish, making the diet seem less diverse compared to breeding areas (Wiley and Lohrer 1973). Mullets are de- tritus-eating fish abundant in Brazilian estuaries, where they gather in shoals year-round (Menezes 1983). Their abundance, size, and the habit of sunning close to the surface make them ideal prey for a plunge diver like the Osprey. In fact, the only former report of a prey taken by an Osprey in Brazil refers to a mullet (Mugil incilis) (Martuscelli 1992), also taken in a mangrove area in southern Sao Paulo state. Mullet was also the main prey taken by wintering Ospreys in mangroves and estuaries in Senegal (Prevost 1982) and South Africa (Boshoff and Palmer 1983). The availability of mullet may probably ex- plain why Ospreys seem to be the most common in coast- al areas, especially estuaries and mangroves (Haver- schmidt and Mees 1994). Fish were also the only prey recorded as taken by Ospreys in freshwater habitats in Peru (Willard 1985) and Argentina (Saggese et al. 1996), the detritus-eating Prochilodus sp. being reported by both studies. We observed a high capture success (71%) for a limited number of foraging attempts on fish, but not out- side the 40-70% range reported by studies elsewhere (Poole 1989, del Hoyo et al. 1994). There is no direct or intensive persecution of Ospreys in southeast Brazil. Nevertheless, their reliance on detri- tus-eating fish might be problematical. The Santos-Cu- batao estuary receives the discharges of one of the major industrial areas in Brazil, well-known in the 1980s for its high pollution levels (Gutberlet 1996). Despite improve- ments, the sediments still hold high levels of heavy metals and organochlorines such as Hexachlorobenzene (HCB) and Polycyclic aromatic hydrocarbons (PAHs). For ex- ample, sediments from the Piagaguera channel hold 109.200 to 733.700 [xg/kg of benzopyrene, some of the largest concentrations in the world (CETESB 2001). These contaminants may be ingested by fish like mullet, feeding on detritus and benthic algae, and accumulated and transferred to piscivores like Ospreys. Thankfully, most contaminants seem to be trapped in the sediment and mullet samples generally have small concentrations (CETESB 2001), although this situation may change if dredging makes the compounds available again in the water column. RF..SUMEN. — Las aguilas pescadoras {Pandion haliaetus) son registradas todo el ano en los manglares de Santos-Cu- batao al suroriente del Brasil, con un pico de abundancia entre diciembre y marzo. Las aves usan todos los habitat del manglar pero las densidades lineares mas altas ocur- ren a lo largo de rios bordeados por bosques de manglar {x = 0.71 aves/km) y manglares degradados por polu- December 2002 Short Communications 331 cion (x = 0.70 aves/km). Las ^uilas pescadoras fueron registradas comiendo unicanaente pescado y mostraron una alta tasa de exito (71% de los intentos). Los salmo- netes {Mugil spp.) fueron el item de comida mas comun (77%), seguido por las mojarras romboides {Diapterus rhombeus) (18%). A pesar de la historia de alta polucion industrial del area, los niveles de contaminantes en las especies presa consumidos por el aguila pescadora fu- eron bajos, y la especie no fue perseguida por los pob- ladores locales. [Traduccion de Cesar Marquez] ACKNO Wr .EDGMENTS This work was possible through grants received from Funda^ao O Boticario de Protegao a Natureza, MacAr- thur Foundation and Ultrafertil S.A. to both authors. We are very grateful for their timely support over the years. We also thank Maria do Carmo Amaral (Cubatao Zoo, Cotia-Para Ecological Park) for saving the Osprey speci- men now in the MZUSP. Joel Escobar and the personnel at Nautica da Ilha provided invaluable support when we needed to deal with boats and outboard motors. We wish to thank Mark Martell, Alan E. Poole, and Richard O. Bierregaard, Jr. for comments and suggestions on im- proving the manuscript. Lu erature Cited Boshoff, A.E. and N.G. Palmer. 1983. Aspects of the bi- ology and ecology of the osprey in the Cape Province, South Africa. Ostrich 54:189-204. Companhia Estadual de Tecnologia e Saneamento Am- biental. 1991. Avaliayao do estado de degradagao dos ecossistemas da Baixada Santista. Sao Paulo, Brazil. . 2001. Levantamento da contaminagao ambiental do sistema estuarino Santos/Sao Vicente. CETESB, Sao Paulo, Brazil. Dean, W.R.J. and W.R. Tarboton. 1983. Ospreys breed- ing records in South Africa. Ostrich 54:238-239. Gutberlet, J. 1996. Cubatao: desenvolvimento, exclusao social, degrada^ao ambiental. Edusp/Fapesp, Sao Paulo, Brazil. del FIoyo, J., a. Elliott, and J. Sargatal (Eds.). 1994. Handbook of the birds of the world. Vol. 2. New world vultures to guineafowl. Lynx Edicions, Barcelona, Spain. Haverschmidt, E. and G.F. Mees. 1994. Birds of Surina- me. Vaco Press, Paramaribo, Suriname. IuAMPARELLI, C.C. 1999. Mapeamento dos ecossistemas costeiros do estado de Sao Paulo. Secretaria do Meio Ambiente/CETESB, Sao Paulo, Brazil. Luederwaldt, H. 1919. Os manguesaes de Santos. Rev Mus. Paulista 11:310-409. Martell, M.S., C.H. Henny, RE. Nye, and M.J. Soi.ensky 2001. Fall migration routes, timing, and wintering sites of North American Ospreys as determined by sat- ellite telemetry. Condor 103:715-724. Martuscelli, P. 1992. Notas sobre aves pouco conheci- das do estado de Sao Paulo. Pages 82-83 in Anais do VI Encontro Nacional de Anilhadores de Aves, EDU- CAT, Pelotas, Brazil. Menezes, N.A. 1983. Guia pratico para conhecimento e identificayao das tainhas e paratis (Pisces, Mugilidae) do litoral Brasileiro. Rev. Brasil. Zool. 2:1-12. Olmos, E. 1989. A avifauna da baixada do polo industrial de Cubatao. Rev. Bras. Biol. 49:373-379. and R. Silva e Silva. 2001. The avifauna of a southeastern Brazilian mangrove swamp. Int. J. Orm- thol. 4:137-207. Poole, A.F. 1989. Ospreys: a natural and unnatural his- tory. Cambridge Univ. Press, Cambridge, U.K. AND B. Agler. 1987. Recoveries of Ospreys band- ed in the United States, 1914-84. y. Wildl. Manage. 51. 148-155. Prevost, Y.A. 1982. The wintering ecology of Ospreys in Senegambia. Ph.D. dissertation, University of Edin- burgh, Scotland. Saggese, M.D., E.R. De Lucca, S.F. Krapovickas, and E.H. Haene. 1996. Presencia del aguila pescadora {Pandion haliaetus) en Argentina y Uruguay. Hornero 14:44-49. Sick, H. 1997. Ornitologia Brasileira. Editora Nova Eron- teira, Rio de Janeiro, Brazil. Wiley, J.W. and F.E. Lohrer. 1973. Additional records of non-fish prey taken by Ospreys. Wilson Bull. 85:468- 470. WiLlARD, D.E. 1985. Comparative feeding ecology of twenty-two tropical piscivores. Ornithol. Monogr. 36. 788-797. Received 8 February 2002; accepted 3 July 2002 / RapiorRes. 36(4) :332-334 © 2002 The Raptor Research Foundation, Tnc. Diet of Breeding Tropical Screech-Owls (Otus choliba) in Southeastern Brazil Jose Carlos MottaJunior' Departamento de Ecologia, Instituto de Biociencias da Universidade de Sdo Paulo, 05508-900, Sdo Paulo, SP, Brazil Key Words: Tropical Screech-Owl; Otus choliba; breeding dieL; prey biomass; Brazil. The Tropical Screech-Owl {Otus choliba) occurs east of the Andes Mountains from Costa Rica to Uruguay and northern Argentina, and is also found throughout much ol Brazil (Meyer de Schauensee 1966, Burton 1992, Sick 1993). It is one of the most common and widespread neotropical owl species inhabiting forest edges, open woodlands, savannas, and other habitats with some ar- boreal cover, including urban areas (Sick 1993, del Hoyo et al. 1999). Despite its commonness and widespread dis- tribution, little ecological information is available on this species, except for some data concerning natural history and breeding (Thomas 1977, Smith 1983). Food habits have been described only qualitatively (e.g., Thomas 1977, Smith 1983, Gallardo and Gallardo 1984). Here, I provide more detailed information about the diets of nestling and adult Tropical Screech-Owls during the breeding season. Data on nest locations and the timing of reproduction in southeastern Brazil are also present- ed Study Area and Methods The study was conducted at Chacara Mattos/Faber-Cas- tell (21°59'S, 47°56'W), located 1 km west of the city of Sao Carlos, Sao Paulo State, Brazil. The 90 ha study area consists primarily of Pinus spp. plantations with some sec- ondary-grassland savanna. A small patch (3 ha) of dis- turbed gallery forest also occurs in this area. The land- scape surrounding the study area is sugar cane plantations and the outskirts of the city of Sao Carlos, fhe climate is a transition between Koppens’s Cwai and Awi, or rainy tropical with a wet (October-March) and a dry (April-September) season (Tolentino 1967). Four samples of pellet debris (representing ca. 30 pel- lets) and six complete pellets were collected from three occupied nest cavities. This material was washed through a line mesh screen (0.2 mm) and oven-dried (5(LC) for 24 hr for storage arid analysis. Prey remains were identi- hed by comparison with a reference collection made from material from the study. I also measured the mass ol prey items collected from the study area. Individuals m the prey remains were counted by pairing mandibles, with the exception of beetles and ants, which were count- ed by the number of heads, and scorpions by the number of stings. The analyzed prey remains presumably were from owlets and possibly adults. I also assumed that ver- tebrate prey were entirely ingested like invertebrates, be- cause crania and other body bones were always present ' E-mail address: mottajr@ib.usp.br in the pellets. Both prey remains and the reference col- lection were deposited at the Departamento de Ecologia, Universidade de Sao Paulo, Brazil. Results and Discussion All three nests were in cavities located at a height of 1.0-1. 5 m in dead Eucalyptus sp. tree trunks, presumably made by woodpeckers originally. Smith (1983) pointed out that Tropical Screech-Owl nests are typically located in tree cavities. In spite of the monthly field excursions to the study area during 1992-93, nests were only found on 25 November 1992 (Nest 1 , with one female and three owlets), 28 October 1993 (Nest 2, with one female and two owlets), and 6 November 1993 (Nest 3, with one fe- male and three owlets). The three adult females inside cavities were captured by hand and weighed with a spring scale. The mean and standard deviation of body mass was 128.3 ± 11.7 g. Subsequently, the adults were placed back into the cavities. Prey remains were collected the first time the nest was found and shortly after owlets fledged. Analysis of pellets and pellet debris revealed at least 34 species of prey consumed. Invertebrates, mostly orthop- terans such as Lutosa brasiliensis, were most frequent in the diet (Table 1). Spiders (Lycosidea) and ants (For- micidea) were also important numerically. In terms of biomass, invertebrates also prevail, however, the few ver- tebrates found, represented a third of the consumed bio- mass (Table 1). The mean body mass of prey consumed by Tropical Screech-Owls was 0.93 ± 2.35 g, ranging from 0.02-28.80 g {N = 309 prey items). Most prey (73.5%) weighed be- tween 0. 1-1.0 g. Tropical Screech-Owls only were observed foraging at night. On two occasions, an individual was observed leav- ing a perch on a tree and, in flight, capturing insects on the leaves of another tree. On another occasion, an in- dividual left a perch on a bush and captured an uniden- tified invertebrate on the ground. Gallardo and Gallardo (1984) reported a similar behavior in Tropical Screech- Owls. 1 have observed these screech-owls catching insects in flight, particularly in the vicinity of artificial light sources, which was also reported by Smith (1983) and Sick (1993). During the period of activity (1800-0600 H) owls were observed on perches waiting for potential prey; therefore, this species probably should be classified as a “sit-and-wait” forager, which is typical for the genus Otus (Jaksic and Carothers 1985). The qualitative studies of Thomas (1977) and Smith (1983) indicated that the diet of Tropical Screech-Owls consisted mostly of insects in Costa Rica and both insects 332 December 2002 Short Communications 333 Table 1. Prey items found in pellets and pellet debris of Tropical Screech-Owls in southeastern Brazil, with their percentages in relation to total number and estimated biomass (g). Activity periods and sites of prey were determined based on field observations and information provided by Manoel M. Dias (pers. comm.). Prey Items Aciivn Y Period Activity Site Pfrcfne Number PFRt:FNT Biomass Rodents Bolomys lasiurus Nocturnal/ crepuscular, Diurnal Ground 0.3 10.1 Calomys tener Nocturnal/crepuscular Ground 0.3 3.6 Oligoryzomys nigripes Nocturnal/ crepuscular Foliage/branches, Ground 0.3 6.0 Opossums Gracilinanus sp. Nocturnal/ crepuscular Foliage/branches, Ground 0.3 5.7 Snakes Unidentified small sp. p ? 0.3 3.3 Amphibians Hylidae (unidentified sp.) Nocturnal/crepuscular Foliage/branches 0.3 5.3 SUBTOTAL VERTEBRATES — — 1.9 34.1 Scorpions Bothriurus spp. Nocturnal/crepuscular Ground 3.6 1.0 Tityius bahiensis Nocturnal/ crepuscnlar Ground 0.3 0.1 Spiders Lycosidae (unidentified sp.) Nocturnal/ crepuscular Ground 11.3 5.9 Unidentified spp. p ? 1.3 0.8 Harvestmen Opiliones (unidentified sp.) p p 0.3 0.1 Insects Blattidae {Parahormetica sp.) Nocturnal/crepuscular Ground 1.6 3.2 Blattidae (unidentified sp.) ? ? 0.3 0.2 Termitidae (workers) Nocturnal/crepuscular, Diurnal Ground 2.6 0.2 Acrididae spp. Diurnal Foliage/branches 2.9 1.4 Tettigoniidae (Copiphorinae) Nocturnal/crepuscular Fo liage / bran ch e s 5.8 6.0 Tettigoniidae (Conocephalinae) Nocturnal/ crepuscular Foliage/branches 0.6 0.2 Gryllacrididae {Lutosa brasiliensis) Nocturnal/crepuscular Ground 41.1 34.2 Gryllidae Nocturnal/ crepuscular Ground 1.6 1.6 Unidentified Orthoptera ? ? 0.6 0.3 Mantidae Nocturnal/crepuscular Foliage/branches 4.2 2.8 Carabidae (small unidentified spp.) Nocturnal/ crepuscular Ground 1.0 0.1 Scarabaeidae (Rutelinae) Nocturnal/ crepuscular Foliage/branches 0.3 0.3 Scarabaeidae (Dynastinae) Nocturnal/ crepuscular Ground 2.6 3.1 Cerambycidae Nocturnal/crepuscular Tree trunks 1.6 2.4 Unidentified adult Coleoptera ? p 1.0 0 4 Unidentified larvae Coleoptera p ? 1.0 0.4 Lepidoptera (unidentified small moth) ? p 0.3 0 1 Lepidoptera (unidentified caterpillar) ? Foliage/branches 1.0 0 3 Formicidae {Atta sexdens queen) Nocturnal/ crepuscular, Diurnal Ground 0.3 0.2 Formicidae {Camponotus Nocturnal/crepuscular, Diurnal Tree trunks, Ground 5.2 0.1 Formicidae (Dorylinae) Nocturnal/ crepuscular, Diurnal Ground 1.9 0 1 Formicidae (unidentified spp.) ? p 3.2 0 1 Other unidentified Insecta ? ? 0.3 02 SUBTOTAF INVERTEBRATES — — 98.1 65 9 334 Short Communications VoL. 36, No. 4 and vertebrates in Venezuela, respectively. Prey taken by these owls include katydids, beetles, cockroaches, small snakes, and rodents (Thomas 1977, Smith 1983, this study). Larger species like Otus asio and O. kennicottii seem to include proportionally more vertebrates in their diets (e.g., Ritchison and Cavanagh 1992, del Hoyo et al. 1999). On the other hand, smaller species such as O. tnchopsis, O. flammeolus, and O. choliba appear to be mostly insectivorous (Ross 1969, del Hoyo et al. 1999, this study) . The frequent consumption of the terrestrial arthro- pods, Lutosa brasiliensis (Lycosidae) and others (68.5% by number and 63.3% by biomass) suggests that prey are often captured on the ground (Table 1). A similar pat- tern in the prey data supports that the Tropical Screech- Owls were essentially nocturnal; the diet consists mainly of night prey (76.8% by number and 81.5% by biomass; Table 1). Resumen. — Se estudio la dieta del Autillo Choliba {Otus choliba) durante el periodo reproductivo, entre los meses de octubre y diciembre de 1992 y 1993, en nna localidad del Sudeste de Brasil. Se identificaron por lo menos 34 especies de presas a partir de egagropilas, directamente colectadas en tres nidos ubicados en cavidades de troncos muertos a 1.0-1. 5 m del suelo. Insectos, en especial Lu- tosa brasiliensis (Gryllacrididae) y otros ortopteros, arahas y escorpiones forraaron la base de la dieta. Aunque los mvertebrados fueron los mas importantes numerica- mente (98.1% del total de 309 individuos), los vertebra- dos tuvieron representacion significativa en terminos de biomasa consumida estimada (34.1% del total de 286.0 g) La inayoria de las presas eran nocturnas y terricolas, mdicando los habitos de caza de esta lechuza. [Traduccion del autor] Acknowledgments 1 thank A.W. Faber-Castell S/A for facilities in the study site. The entomologists Manoel M. Dias and Alejo M. Lar- rambebere helped with invertebrate identihcation. Mar- tha Desmond, F.R. Gehlbach, Gary Ritchison, and Clint W Boal made valuable comments. Diego Queirolo kindly revised the Spanish text. This work was a small part of a Ph.D. dissertation sponsored by Coordenacao De Aper- feigoamento Ce Pessoal De Nivel Superior and World Wildlife Fund/Brazil. Literature Cited Burton, J.A. (Ed.). 1992. Owls of the world: their evo- lution, structure, and ecology, 3rd Ed. Peter Lowe, Wallingford, U.K. DEL Hoyo, J., A. Elliott, and J. Sargatai.. 1999. Hand- book of the birds of the world. Barn Owls to hum- mingbirds. Vol. 5. Lynx Edicions, Barcelona, Spain, Gai.iardo, L.A. AND J.M. Gallardo. 1984. Observaciones realizadas sobre el comportamiento de Otus choliba en liberdad. Comun. Mus. Argent. Cienc. Nat. Bernardino Rivadavia Zool. 4:109-114. JAKSIC, F.M. AND J.H. Carothers. 1985. Ecological, mor- phological, and bioenergetic correlates of hunting mode in hawks and owls. Ornis Scand. 16:165-172. Meyer de Schauensee, R. 1966. The species of birds of South America and their distribution. Livingston, Nar- berth, PA U.S.A. Ritchison, G. and P.M. Cavanagh. 1992. Prey use by Eastern Screech-Owls: seasonal variation in central Kentucky and a review of previous studies. J. Raptor Res. 26:66-73. Ross, A. 1969. Ecological aspects of the food habits of insectivorous screech-owls. Proc. West. Found. Vertebr. Zool 1:301-344. Sick, H. 1993. Birds in Brazil. A natural history. Princeton Univ. Press, Princeton, NJ U.S.A. Smith, S.M. 1983. Otus choliba. Pages 592-593 in D.H Janzen [Ed.], Costa Rican natural history, Univ. of Chicago Press, Chicago, IL U.S.A. Thomas, B.T 1977. Tropical Screech-Owl nest defense and nestling growth rate. Wilson Bull 89:609-612. Tolentino, M. 1967. Estudo critico sobre o clima da re- giao de Sao Carlos. Concurso de Monograhas Muni- cipals, Sao Carlos, Brazil. Received 17 September 2001; accepted 5 July 2002 Associate Editor: Clint W. Boal Letters J. Raptor Res. 36(4):335-336 © 2002 The Raptor Research Foundation, Inc. Comments on the First Nesting Record of the Nest of a Slaty-backed Forest-Falcon {Micrastur mirandollei) in the Ecuadorian Amazon De Vries and Melo (2000,/. Raptor Res. 34:148-150) recently reported the first documented nest of a Slaty-backed Forest-Falcon {Micrastur mirandollei), based on their studies in Yasuni National Park, Ecuadorian Amazon. Certain details they reported differ markedly from the nesting habits and behavior that my colleagues and I observed for the Barred Forest-Falcon (M. ruficollis) (Thorstrom et al. 2000a, Auk 117:781-786) and Collared Forest-Falcon (M. semi- torquatus) (Thorstrom et al. 2000b, Ornithol. Neotrop. 11:1-12) in northern Central America, from 1988-96. These differences are profound enough to suggest that the raptor species was possibly misidentified as a forest-falcon, or that the Slaty-backed Forest-Falcon displays some rather uncommon behavior within the genus of Micrastur. De Vries and Melo (2000) reported on 14 September, and again on 23 October 1997, that the nesting forest- falcons had a changeover, implying an incubation and a brooding switch, respectively. I never recorded an incubation or brooding switch during 8 breeding seasons of nest observations of Barred {N = 70 nesting attempts) and Collared Forest-Falcons {N = 9 nesting attempts). For these two forest-falcons, the males’ role was providing food to the incubating and brooding female and nestlings. During incubation, the male contact-called to the female upon his arrival to the nest vicinity with prey, and the female exited the nest to receive the prey item where she ate it. On rare occasions, the male entered the cavity while the female ate. The male stayed inside for several minutes either incubating, attempting to incubate or to look at the nest contents, and then he exited the cavity prior to the female’s return (Thorstrom 1993, M.S. thesis, Boise State University, Boise, ID U.S.A., Thorstrom et al. 2000a, Thorstrom et al. 2000b). De Vries and Melo (2000) suspected also that the female fed herself away from the nest and returned to take over incubation or brooding from the male. This does not agree with my observations (Thorstrom et al. 2000a, Thorstrom et al. 2000b) and those of Baker et al. (2000, Ornithol. Neotrop. 11:81-82). Only on very rare occasions did the female leave the nest during incubation to feed herself when the male was late with a prey delivery and hunger brought her off the nest (Thorstrom 1993). De Vries and Melo (2000) describe an open nest constructed of small sticks and deep enough to hide the head of the incubating Slaty-backed Forest-Falcon. In contrast, all described nests for the genus Micrastur have been m cavities (Mader 1979, Condor 81:320) and all nesting attempts by both Barred Forest-Falcons and Collared Forest- Falcons observed by my colleagues and me in Central America were in tree cavities (Thorstrom et al. 1990, Condor 90:237—239, Thorstrom et al. 2000a, Thorstrom et al. 2000b) except the observation by Baker et al. (2000) of a pair of M. ruficollis nesting in a cliff pothole below canopy level. There is also a record of a Collared Forest-Falcon nesting in a ruined building (Cobb 1990, cited by Howell and Webb 1995, The birds of Mexico and Central America, Oxford Univ. Press, New York, NYU.S.A.). There is one mention of stick nesting by the Lined Forest-Falcon (M. gilvicollis) in del Hoyo et al. (1994, Handbook of the birds of the world, Lynx Edicions, Barcelona, Spain), but no details were provided. Forest-falcons do appear to have some flexibility in their choice of nest sites, but they seem to prefer a site that simulates a cavity; i.e., trees or cliff potholes surrounded by forests. The Slaty-backed Forest-Falcon has several described calls: one is a 10-14 nasal aah syllables and another a two part series '‘ah, ow, ow, ow, ow, ow, ow, uah, uah, uah, uah, uaK' (Ridgely and Gwynne 1989, A guide to the birds of Panama, Princeton Univ, Press, Princeton, NJ U.S.A.), a chanting series 8-13 nasal, thus contrasting with the 6 calls “kui kui kui kui kui kui' reported by de Vries and Melo (2000). The authors’ example of the Monk Parakeet {Myiopsitta monachus), a cavity nester that builds a stick nest, does not support their suggestion that the Slaty-backed Forest-Falcon is normally a stick nest builder. Also, their basis for suspecting the forest-falcons built the nest is somewhat vague, i.e., “We did not see the nest being built so we did not know if the falcons had taken an old nest made by another species, but we felt that this was unlikely because we did not see the nest on our regular censuses” (de Vries and Melo 2000). In Central America, forest-falcons were never observed constructing nests out of sticks or carrying nesting material (Thorstrom et al. 2000a, Thorstrom et al. 2000b). However, it is quite possible that forest-falcons can occupy a previously-built nest in a situation that replicates a cavity. There are six known species of Micrastur, and the Lined Forest-Falcon may represent two separate species (A. Whittaker pers. comm.), which would make seven. Among these, two (M. ruficollis, M. semitorquatus) are 335 336 Letters VoL. 36, No. 4 known cavity nesters, three (M. plumbeus, M. gilvicollis, and the proposed new species) are suspected cavity nesters, and for one nesting details are unknown (M. huxkl£yi). Thus, the report by de Vries and Melo (2000), stating that M mirandollei is a stick builder and nester, contrasts sharply with the large body of evidence from its congeners. De Vries and Melo (2000) seemed uncertain about their identification of the species they were observing on the nest. They came to the conclusion that the birds they had observed were not Gray-bellied Goshawks {Accipiter polio- gastei) because of the white belly and yellow facial area of the supposed female, and the other bird, the supposed male, had buff below with both birds having long legs and three narrow, dirty white tail bands (de Vries and Melo 2000). I suggest that the characteristics that de Vries and Melo used to identify this nesting pair of raptors were not conclusive and that they have misidentified this species. The Grey-bellied Goshawk has dark back and crown, long yellow legs, white to gray belly and female larger than male (del Hoyo et al. 1994). Accipiters have long legs, faint tail bands, and females are larger than males. The size dimorphism was commented on by the authors “the first falcon was smaller and probably the male of the pair.” Sexual size dimorphism of forest-falcons was very difficult to distinguish in the field (pers. observ.) because they are only slightly to moderately dimorphic (Thorstrom 1993). Size dimorphism is also a characteristic described for the Grey-bellied Goshawk and Bicolored Hawk {Accipiter bicolor) (del Hoyo et al, 1994). The authors did not give any further detailed characteristics of the color of the legs, back, eyes, and crown of the nesting raptors. The nesting behavior and habitat, vocalization, and plumage characteristics suggest lhat the nesting birds described by de Vries and Melo (2000) were possibly accipiters, either the Crrey-bellied Goshawk or Bicolored Hawk. I thank The Peregrine Fund for support and L. Kiff, R. Bierregaard, J. Bednarz, and one anonymous reviewer for their comments on this manuscript. — Russell Thorstrom, The Peregrine Fund, 5668 West Flying Hawk Lane, Boise, ID 83709 U.S.A.; E-mail address: rthorstrom@peregrinefnnd.org Received 21 August 2001; accepted 19 June 2002 /. RaplorRes. 36(4);337 © 2002 The Raptor Research Foundation, Inc. Micrastur or Accipiter, That is the Question . . the more you look the more you see” (Peter Grant 1986, Ecology and evolution of Darwin’s finches. Princeton Univ. Press, Princeton, NJ U.S.A.). The main point made by Thorstrorn (2002, J. Raptor Res. 36:335-336) concerns the behavior of our birds, which he claims is not the behavior of Micrastur (he should perhaps say the behavior of M. ruficollis and M. semilorquatus, as the behavior of the other four species is still unknown) . We can report that the Slaty-backed Forest-Falcon {Micrastur mirandolld) was present in the area of its now defunct 1997 stick nest (de Vries and Melo 2000,/ Raptor Res. 34:148- 150) in March of 1998. Although we cannot be sure that it is the same bird we saw previously at and around the nest, it responded (vocalized, but did not come out into the open) to the species’ call as recorded by John Mooie m his series of bird sounds of eastern Ecuador. So far, we have been unable to locate its new nesting site. The Grey-bellied Goshawk {Accipiter poliogaster) was present at some 5 km distance from the MicrasturncA site in both 1997 and 1998. We observed this species in the more open and bare branches of the canopy, rather than the densely-vegetated, middle canopy layer, where we noted Micrastur mirandollei. We hope that the comment by R. Thorstrorn (2002) stimulates more observations on Micrastur, which are badly needed. In addition, further study on why some avian raptors are so similar in plumage patterns, as is the case with M. mirandollei and Accipiter poliogastca; would be valuable in understanding the potential adaptive benefits of such “mimicry.” In our field experience, the “capped” appearance and tail banding oi Accipiter 'Are diagnostic, as are the round grey head and yellow face that Micrastur features. — Tjitte de Vries and Cristian Melo, Departamento de Biol- ogia, Pontificia Universidad Catolica del Ecuador, Apartado 17-01-2184, Quito, Ecuador; E-mail address: tdevries@ puceuio.puce.edu.ee Received 8 November 2001; accepted 27 June 2002 337 BOOK REVIEWS J Raptor Rf,s. 36(4) :338 © 2002 The Raptor Research Foundation, Inc. Birds of Prey; Health 8c Disease. By John E. Cooper. 2002. 3rd edition. Blackwell, Oxford, U.K. xvii + 345 pp., 13 tables, 56 figures, 28 color plates, 11 appendices. ISBN 0-632-05115-9. Hard- back, £59.50. — The study of raptors is unquestion- ably enhanced by the diversity of perspectives from which it derives, i.e., from biologists, falconers, medical practitioners, and others. Each of these disciplines provides data and references that ad- dress birds of prey in unique and relevant terms. In updating his earlier pioneering work, Vetmnary Aspects of Captive Birds of Prey, now entitled Birds of Prey: Health & Disease, John Cooper continues his intentionally eclectic collection of information and points-of-view. The new work is organized identi- cally to the former but for the addition of three new chapters from contributing authors. Roughly one-third of the information in the new book is repeated verbatim from the earlier one, one-third includes previous information updated by a brief summary of advancements and a list of references for further study, and the remainder is an expan- sion that reflects work that has appeared since the earlier volume. Cooper has long advocated collaboration among all who deal with birds of prey, which is borne out by his approach to the new addition. Cooper is at once observer, historian, and medical practitioner in discussing the various topics. He makes no at- tempt to cover the subjects in depth, but rather provides context and references for the reader’s use. The diverse array of topics, and the somewhat lack of in-depth coverage, may serve as a disap- pointment to some but will be acceptable to others as an interesting and general guide. The value of the book will certainly be defined by the expecta- tions of those who use it. The contributed chapters vary considerably in scope and detail. Paolo Zucca’s chapter on anato- my is generalized with varying degrees of detail. Once again, subjectivity is a factor in some conclu- sions. Ian Newton’s discussion of parasitic diseases IS, by nature, broad, yet is thorough and sound in content. The late David Peakall’s contribution re- garding poisonings in free-living raptors is an in- clusive general overview of chemical toxicity on a worldwide scale. This is an intensely personal work. Cooper states emphatically that he has laced the writing with opinions, personal experiences, and other subjec- tive matter. One of the most difficult challenges in reviewing this work is to assign it to a category for the reader. The book is not a clinical manual, nor a biology text, nor a handbook for falconers. It does, however, contain elements of each along with personal reflections on the experiences and opin- ions of the principal author. Although resisting classification. Birds of Prey: Health & Disease fills a useful and interesting niche in raptor literature for all those involved in the field. — James D. Elliott, South Carolina Center for Birds of Prey, P.O. Box 1247, Charleston, SC 29402 U.S.A. J. Raptor Res. 36(4):338-339 © 2002 The Raptor Research Foundation, Inc. Owls. By Floyd Scholz. 2001. Stackpole Books, Mechanicsburg, PA. xiii + 379 pp., more than 700 color photographs. ISBN 0-8117-102T1. Cloth, $80.00. — In 1993, Floyd Scholz published Birds of Prey, which was a collection of detailed color pho- tographs of 17 species of North American falconi- forms (see /. Raptor Res. 28:278—279, 1994). Team- ing up once more with photographer Tad Merrick, Scholz has expanded his domain to include the strigiforms. The result is an extensive series of col- or photographs of 17 of the 19 species of owls (Western Screech-Owl \_Otus kennicottii] and Whis- kered Screech-Owl [O. trichopsis] omitted) that breed in the United States and Canada. Floyd Scholz is an extraordinary carver and painter of birds, and the main purpose of Owls is to serve as a reference guide for artists. An intro- 338 December 2002 Book Reviews 339 ductory chapter, “What is an Owl?” presents infor- mation on the various morphological adaptations of owls; the treatment enhances the text and is ba- sically the same as that found in any coffee table book or general ornithology text that deals with owls. The species accounts make up the heart of the book. Each includes a page of information on morphology and behavior of the species in ques- tion, details on the size of various body parts (in- cluding line drawings of the species from the back and side), and numerous color photographs (typ- ically 30-40 per species). The photos are taken from nearly every conceivable angle to provide de- tails on plumage, talons, ear tufts, facial ruffs, eyes, bills, nostrils, and napes, to name but a sample of the features depicted. These photos contain fine points that would seldom be noticed without hav- ing a bird in the hand and thus will serve as a valuable reference for people who do not have ac- cess to museum specimens. Most of the photos are of captive birds, the exceptions being a handful of very nice shots of wild Northern Hawk Owls (Sur- nia ulula) , Great Gray Owls {Strix nebulosa), and Boreal Owls {Aegolius funereus) by Ron Austing and Robert Taylor. The photo selection, and the photos themselves, are truly excellent, but two criticisms are worth mentioning. First, the Great Horned Owl {Bubo vir- ginianus) shown in many of the photos can be readily described by two words: pissed off. Tm not sure why an artist would need photos of an owl in this state, nor why someone would continue to bother a bird that so obviously detested whatever treatment it was receiving. In fairness to the author and photographer, this same attitude is displayed by a wild female Great Horned Owl depicted on the cover of the Birds of North America species ac- count. Second, I was surprised that none of the 32 photos of the two species of pygmy-owls contained a decent view of the so-called “false eyes” on the back of the head. This is a minor quibble to be sure, but an opportunity missed nonetheless. The last two chapters are entitled “Techniques for the Artist and Garver” and “Gallery,” the latter consisting of a collection of photos of some of the owls Mr. Scholz has carved and painted. The carv- ings are absolutely gorgeous and point to the enor- mous talent of the author. This book will appeal to anyone with an interest in owls, although the price may place it out of reach of all but the confirmed “owlaholics,” to borrow a phrase coined by Heimo Mikkola. — Jeff Marks, Montana Cooperative Wild- life Research Unit, University of Montana, Missou- la, MT 59812 U.S.A. J Raptor Res. 36(4);340— 343 © 2002 The Raptor Research Foundation, Inc. Journal of Raptor Research INFORMATION FOR CONTRIBUTORS The Journal of Raptor Research (JRR) publishes original research reports and review articles about the biology of diurnal and nocturnal birds of prey. All submissions must be in English, but contribu- tions from anywhere in the world are welcome. Manuscripts are considered with the understand- ing that they have not been published, submitted or accepted for publication elsewhere. Manuscripts are subjected to peer review for evaluation of their significance and soundness, and edited to improve communication between authors and readers. De- cisions of the editor are final. Material is published as feature articles, short communications (usually not longer than four printed pages), and letters (see recent issue of the JRR for examples) . Submissions that adhere closely to the JRR’s format greatly enhance the efficiency and cost of the editorial and publishing processes. Author’s efforts in this regard are deeply appreci- ated by the editorial staff. When submitting scholarly papers, send the orig- inal and three copies, a completed checklist (see below), and a cover letter that includes: (1) a state- ment that the data in the manuscript have not been published or accepted for publication in the same form, and have not been submitted simulta- neously elsewhere, (2) the name and address of the corresponding author (in multiauthored pa- pers) including any temporary addresses where that author will be during the review process (also the phone number and, if possible, a FAX number and e-mail address of the corresponding author), and (3) if applicable, any special instructions. Au- thors may also suggest potential reviewers. If the manuscript submitted was produced on a word processor, also send a diskette (3 1/2") con- taining a single hie that is identical with the print- ed copy. The electronic copy should be supplied as an IBM-compatible text hie (ASCII, Word, or WordPerfect). Include information on the type of computer and word processor used. Optional elec- tronic submissions are accepted (see General In- structions below) . Manuscripts that are accepted upon condition of revision must be returned to the editor within 60 days. Manuscripts held longer will lose their pri- ority and may be treated as new submissions. The editor should be notihed if extenuating circum- stances prevent a timely return of the manuscript. Authors will receive proofs of their articles prior to publication. Proofs must be read carefully to cor- rect any printer errors and returned by the fastest mail within two days of receipt TO THE EDITOR. Changes in typeset text are expensive and authors making extensive changes will be billed for the costs. A reprint order will accompany page proofs to enable authors to buy reprints. Costs of reprints are the author’s responsibility and payment for re- prints ordered must accompany the order form. Both must be sent TO THE EDITOR. Publication is expensive and member dues do not cover the entire cost of producing the JRR. Hence, the Raptor Research Foundation, Inc. ex- pects that authors defray the high costs of publi- cation through payment of page costs (currently $100.00 EI.S. per page). Authors who are not as- sociated with a research institution or simply do not have access to such grants may request the page charges be waived. Such a request can only be approved if the author is a member of the Foun- dation and the article is short. Payments of amounts less than the full page charges will be ac- cepted. Authors of long manuscripts are expected to pay publishing costs. It is unlikely that articles longer than 10 printed pages or 18 typewritten pages including tables and illustrations can be pub- lished without full payment. Authors employed by government agencies, univer.sities, or hrms that will meet reprint and page charges may forward a statement to the editor indicating intent to pay. Upon receipt of such a statement, reprints will be mailed to the author and the agency will be billed with the understanding that payment will be made within 30 days. All checks should be made payable to the Raptor Research Foundation, Inc. All per- sonal payments toward publication costs are tax- deductible in the United States. 340 December 2002 Information for Contributors 341 Journal of Raptor Research CHECKIJST FOR PREPARATION OF MANUSCRIPTS {check items and submit with manuscript) I. General Instructions (Consult recent issues for additional guidance on format) I I Type manuscripts on one side of either 216 X 278 mm (8.5 X IT') or standard international size (210 X 297 mm) good quality paper (do not use erasable or lightweight paper). Word-processor-generated manuscripts must be done with a letter-quality or near- letter-quality printer. DOUBLE SPACE THROUGH- OUT including title, text, tables, hgure legends, and literature cited. □ Optional electronic submission. Submit e-mail mes- sage with two attached file copies of manuscript: 1) pdf hie (must include hgures) and 2) word processing hie (Word Perfect [preferred] or MS Word hies ac- cepted). Follow same format guidelines as for stan- dard mail submission. Letter of transmittal should be included in the e-mail message. E-mail to; jrr@astate. edu □ Give the scientihc name at the hrst mention of a spe- cies, both in the abstract and in the article. Scientihc names of birds should follow the usage of the AOU Check-list of North American Birds (7th. ed. 1998 and subsequent supplements in the Auk) or an authorita- tive source corresponding to other geographic re- gions. Do not give subspecihc identihcation unless it is pertinent. Capitalize hrst letter of words in com- plete common names for birds. Use lower case for all other common names. D Use American spelling and Webster’s Tenth New Colle- giate Dictionary (1996, Merriam-Webster, Inc.) as a spelling authority. □ Leave at least a 25 mm (1") margin on all sides. Avoid hyphens or dashes at ends of lines; do not divide a word at the end of a line. □ Use a nonproportional font of at least elite size (4.7 characters/cm = 12 characters/inch) or 12 point, preferably Courier. DO NOT USE RIGHT JUSTIFI- CATION-LEAVE RIGHT MARGIN RAGGED. D Use italic type for addresses, scientihc names, journal names, and third level headings. □ Type last name(s) of author (s) and page number in upper right-hand corner of page 2 and all following pages. □ Cite each hgure and table in the text. Do not repeat material in two forms (i.e., in text and table, or table and hgure). Organize text, as far as possible, so that tables and hgures are cited in numerical order. □ Use “Figure” only to start a sentence; otherwise “Fig.” if singular, “Figs.” if plural (e.g.. Fig. 1; Figs. 2, 3; Figs, d— 6). □ Use metric units throughout. □ Use these abbreviations without spelling out: hr, min, sec, yr, mo, wk, d, km, cm, mm; designate temperature as 32°C. □ Use “continental” dating (e.g., lOJuly 1993, l-3June, 11 May to 11 June). □ Use 24-hour clock (e.g., 0800 H, 1345-1400 H) □ Write out numbers one to nine unless a measurement (e.g., four birds, 3 km, 40 sites, 6 yr). Use 1000 and 10 000; 0.15 instead of .15; % instead of percent. □ Each reference cited in text must be listed in the Lit- erature Cited section, and vice versa. Double check the accuracy of all entries- — THE EDITORIAI. STAFF CANNOT DO THIS FOR YOU. □ Literature citations in the text are as follows; a. One author-Jones (1993) or (Jones 1993) b. Two authors-Smith and Jones (1991) or (Smith and Jones 1991) c. Three or more authors-Hernandez et al. (1990) or (Hernandez et al. 1990) d. Manuscripts accepted for publication but not yet published-Howard (in press) or (Howard in press) e. Unpublished materials-K. Jacobson (unpubl. data); (K. Jacobson pers. comm.); or K. Jacobson (pers. comm.) -do not place in the Literature Cited section. f. When citing several references within parentheses, separate with commas and put in chronological or- der, olde.st hrst). g. For manuscripts submitted as letters, place cita- tions in text in abbreviated form, e.g., (I.C. Birds 1993, J. Raptor Res. 27:45-50). □ Assemble manuscripts for regular articles in this or- der: (1) title page, (2) abstract page, (3) text, (4) ta- bles, (5) hgure legends, (6) hgures. DO NOT STA- PLE. II. Title Page □ Place full title 6-8 lines below top of page in all capital letters. Below title, center author’s name(s) in all capital letters and address (es) followed by a running title (short title) not to exceed 30 characters. If the author(s) is/are currently at another location from where the work was done, use superscript number (s) fol- lowing author (s) name(s) to indicate current address in footnote at bottom of the page. In multiauthored papers, indicate the author re- sponsible for correspondence and requests for reprints. Give phone number and, if possible, FAX number and e-mail address of the corre- sponding author. III. Abstract/summary □ For regular articles, include an abstract of about 250 words in one paragraph that is completely 342 Information for Contributors VoL. 36, No. 4 without reference to the text. Be concise, in- clude the paper’s purpose, but emphasize the results. Statements like “results will be dis- cussed” are not appropriate. The abstract will also be published in Spanish. Authors fluent in both languages are encouraged to include both versions, otherwise the JRR will provide the Spanish translation. □ Include hve to seven key words for indexing af- ter the abstract. □ Short communications will be printed with a Spanish summary only. Authors must provide an English summary to be translated into Span- ish unless they are fluent in Spanish. □ Avoid citing references in the abstract. If they must be cited, include journal name, volume, pages, and year, all in parentheses. IV. Text □ Follow instructions in section I. □ Main headings are all capital letters and flush with left margin, n Typical main headings for regular articles are: METHODS, RESULTS, and DISCUSSION. An introduction begins the text but does not have a heading. □ Put second-level headings in bold. Use normal indentation and capitalize first letter of each word in the second-level headline except prep- ositions and articles. n Put third-level headings in italics. Capitalize first letter of first word only. □ Short communications and letters may or may not have headings within the text depending upon the need. V. Literature Cited □ Type references in upper and lower case letters, including all authors’ names (e.g., Nemeth, N.M, and J.L. Morrison. 2002. Natal dispersal of the Crested Caracara . . . etc. . . .). □ Verify all entries against original sources includ- ing diacritical marks and spelling in languages other than English. Capitalize all nouns in Ger- man, □ Cite references in alphabetical order by first au- thor’s surname. References by a single author precede multiauthored works by the same se- nior author regardless of date. □ List works by the same author (s) chronologi- cally, beginning with the oldest. □ Use six hyphens when the author is the same as in the preceding citation. □ “In press” citations must have been accepted for publication and must include date, volume number, and the name of the journal or pub- lisher. □ Initials of second, third, and. . . authors precede their surname. □ Abbreviate journal names according to the Se- rial Sources for the BIOSIS Data Base (published annually by the BioSciences Information Ser- vice) . □ Do not list personal communications and un- published reports. VI. Tables (Tables are expensive to print-keep them to a min- imum and put each on a separate page-try to de- sign them to fit a single column.) □ Double space throughout. Assign each table an Arabic number followed by a period. □ Table titles must be complete sentences. □ Use same size of type as in text. □ Indicate footnotes by lowercase superscript let- ters. □ Do not use vertical lines. VII. Figure Legends □ Print all figure legends on one page, double spaced. □ Number using Arabic numbers consecutively in the same order the hgures appear in the text (i.e.. Figure 1., Figure 2., etc.). VIII. Preparation of Illustrations (Illustrations are referred to as figures and include drawings, graphs, and black and white half-tones [photographs] . CONSULT THE EDITOR IN AD- VANCE ABOUT COLOR.) □ Use professional standards in preparing figures, their reproduction in the JRR is virtually iden- tical to what is submitted. Consult issues of JRR for examples and Steps Toward Better Scientific Il- lustrations (Allen Press, PO. Box 368, Lawrence, KS 66044) for more information. □ Plan figures to fit proportions in the JRR, pref- erably for a single column-printed size is 72 mm for single column width, 148 mm for full page width or 195 mm for lengthwise figures. Figures should be submitted no smaller than the final size nor larger than twice the final size (on paper no larger than 216 X 278 mm (8.5 December 2002 Information for Contributors 343 X 11") or standard international (210 X 297 mm). □ Submit drawings as original art (undiluted In- dia ink on heavy-weight, smooth-finish drafting paper) or as photomechanical transfers (PMTs). Submit graphs as mechanical drawings or as high-resolution laser prints. Typed or handwritten text or symbols are not acceptable. Add text and symbols with press-on symbols and letters or mechanical lettering. Review copies of figures can be photographic copies but must ap- proach the quality of the original. □ Figure text must be a plain (sans serif) typeface (e.g., Helvetica), not compressed, and large enough so that it will be as large as the text type (8-10 point) when in print. □ Photographs must be sharp, high-contrast, glossy prints approximately the size that they will appear in print. If several photographs are to be included in one figure, group them butt- ed together with no space between. □ Use the same style of lettering and presentation for all figures. IX. What to Send □ Cover letter. □ Copy of this checklist completed. □ Original and three copies of manuscript and il- lustrations. □ Diskette containing a text file of the manuscript text, tables, and figures (if the manuscript was prepared using a word processor) . D Submit to: James C. Bednarz, Editor Journal of Raptor Research Department of Biological Sciences P.O. Box 599, 117 Caraway Road Arkansas State University State University, AR 72467 U.S.A. More information? Telephone: (870) 972-3082 FAX: (870) 972-2638 E-mail: jrr@astate.edu J RaplorRes. 36(4):M4-350 © 2002 The Raptor Research Foundation, Inc. Index to Volume 36 By Kristina Baker This index includes references to general, species, common names, key words, and authors. Reference is also made to book reviews, letters, and reviewers. Taxa other than raptors are included where referenced by authors. A Abuladze, Alexander and Jevgeni Shergalin, The Golden Eagle in north Caucasia and Transcaucasia, 10-17 (suppl.) Afdpiler, 229-230 (ooperii 229-230 gentilis, 141-143, 229-230, 265-279 sLriaius, 229-230 Aegolius funereus, 2 1 8—2 1 9 Age differences, 115-120 Agostini, Nicolantonio, Luca Baghino, Charles Coleiro, Ferdinando Corbi, and Guido Premuda, Circuitoxis autumn migration in the Short-toed Eagle { Circaetus gallicus), 111-114 Alaska, 50-54 (suppl.) Albinism, 200-202 Alloparental care, 70-73 Ammer, Frank K. and Petra Bohall Wood, Probable breeding of Short-eared Owls in southern West Vir- ginia, 237-238 Andersen, David E., see Warnke, D. Keith Anderson, David L., see Thorstrom, Russell Angulo, Elena, Factors influencing length of the pc^st- fledging period and timing of dispersal in Bonelli’s Eagle {Hieraaetus fasciatus) in southwestern Spain, 157 Anthropogenic food sources, 220-224 Aparicio, Jose, see Cordero, Pedro J. Apodaca, Christine K., see Seavy, Nathaniel E. Aquila dirysaetos, 3-9 (suppl.), 10-17 (suppl.), 18-19 (suppl.), 20-24 (suppl.), 25-28 (suppl.), 29-31 (suppl.), 32-40 (suppl.), 41-49 (suppl.), 50-54 (suppl.), 55-61 (suppl.), 62-69 (suppl.), 70-77 (suppl.) xvahlbergi, 51—57 Argentina, 206-212, 315-319 Asia otus, 73-77 Athene cunicularia Jloridana, 3-10 Avila, Gregorio Lopez, see Whitacre, David F. Avila, Juventino Lopez, see Whitacre, David F. Avery, Michael L., John S. ITiimphrey, Eric A. Tillman, Kimberly O. Phares, and Jane E. Hatcher, Dispersing vulture roosts on communication towers, 45-50 Ayers, Serena, see Thorstrom, Russell B Baghino, Luca, see Agostini, Nicolantonio Bainbridge, Ian R, see McGrady, Michael J. Baja California Peninsula, 3-9 (suppl.) Baker, Aaron, see Thorstrom, Russell Balbontin, Javier and Miguel Ferrer, Plasma chemistry reference values in free-living Bonelli’s Eagle {Hi- eraaetus fasciatus) nestlings, 231-235 Bald and Golden Eagle Protection Act, 29-31 (suppl.) Band, encounters, 97-110 recoveries, 97-110 Barry, Irene M., see Powell, Larkin A. Barton, Nigel W.H., Nicholas C. Fox, Peter F. Surai, and Brian K. Speake, Vitamins E and A, carotenoids, and fatty acids of the raptor egg yolk, 33-38 Bates, John M., see Grifliths, Carole S. Bats, 146-148 Beasom, Sam. L, see Proudfoot, Glenn A. Bechard, MareJ. and Michael J. McGrady, Status and con- servation of Golden Eagles, 2 (suppl.) Bednarz, James C., Working toward excellence, 1-2 Behavior, 77-81, 121-127, 136-139 copulation, 66-70 hunting, 194-199 ranging, 70-77 (suppl.) Beheim, Janne, Katrine Eldegard, Gro Bj0rnstad, Mats Isaksson, Geir Sonerud, Olav Heie, and Helge Klungland, DNA polymorphisms in Boreal Owls {Ae- golius funereus ) , 2 1 8—2 1 9 Bellocq, M. Isabel, Patricio Ranhrez-Llorens, andjulieta Filloy, Recent records of Crowned Eagles {Harpyhal- iaetus coronalus) from Argentina, 1981-2000, 206- 212 Berkelman, James, James D. Fraser, and Richard T. Wat- son, Nesting and yjerching habitat use of the Mada- gascar Fish-Eagle, 287-293 Bertram, Joan and Antoni Margalida, Social organization of a trio ol Bearded Vultures {Gypaetus barbatus): sex- ual and parental roles, 66-70 Bianchi, Edward W., see Hunt, W. Grainger Bias, 11-16 Bildstein, Keith 1.., sec Yosef, Reuven Bj 0 rnstad, Gro, see see Beheim, Janne Blood, 231-235 parasites, 139-141 Bloxton, Thomas D., Audi Rogers, Michael F. Ingraldi, 344 December 2002 Index to Volume 36 345 Steve Rosenstock, John M. Marzluff, and Sean R Finn, Possible choking mortalities of adult Northern Goshawks, 141-143 Boano, Giovanni and Roberto Toffoli, A line transect sur- vey of wintering raptors in the western Po Plain of northern Italy, 128-135 Bogliani, Giuseppe, see Sergio, Fabrizio Bolivia, 146-148 Boto, Alberto, see Sergio, Fabrizio Brazil, 328-331, 332-334 Breeding, 161-169 diet, 332-334 season, 194-199 success, 24-32, 81-84, 224-228 Brendel, Ulrich M., Rolf Eberhardt, and Karen Wies- mann, Conservation of the Golden Eagle (Aquila chrysaetos) in the European Alps — a combination of education, cooperation, and modern techniques, 20-24 (suppl.) Bubo bubo, 11-16 virginianus, 58—65 Buhay, Jennifer E. and Gary Ritchison, Hunting behavior of and space use by Eastern-Screech Owls during the breeding season, 194—199 Buteo buteo, 24-32, 115-120, 128-135, 188-193 lineatus, 152-153 C Calvert, Dan J., see Powell, Larkin A, Canada, 32-40 (suppl.) Canary Islands, 17-23 Cannibalism, 200-202 Capote, Nieves, see Donazar, Jose Antonio Capture, 188-193 techniques, 320-323 Caracara cheriway, 203—206 Carpathian Mountains, 25—28 (suppl.) Carrion, 152-153 Caryospora hutzeri, 84—86 Cathartes aura, 45-50, 144—145 burrovianus, 183-187 melambrotus, 183-187 Caucasia, 10-17 (suppl.) Ceballos, Olga, see Donazar, Jose Antonio Central America, 39-44 Chavez-Ramirez, Felipe, see Proudfoot, Glenn A. Chihuahua, 3-9 (suppl.) Chile, 315-319 Choking, 141-143 Circaetus gallicus, 111-114 Cliffs, 39-44 Coahuila, 3-9 (suppl.) Coccidiosis, 84—86 Coleiro, Charles, see Agostini, Nicolantonio Colonization, 18-19 (suppl.) Colorado, 256-264 Columbretes Islands, 139-141 Common Buzzard, 24—32, 115-120, 128-135, 188-193 Communication tower, 45-50 Conservation, 10-17 (suppl.), 20-24 (suppl.), 29-31 (suppl.), 41-49 (suppl.), 51-57, 206-212 Constraints, 41-49 (suppl.) Control region, 17-23 Cooke, Raylene, Robert Wallis, and John White, Use of vegetative structure by Powerful Owls in outer urban Melbourne, Victoria, Australia — implications for management, 294-299 Coragyps atratus, 45-50 Corbi, Charles, see Agostini, Nicolantonio Cordero, Pedro J., Jose M. Aparicio, and David T. Parkin, Genetic evidence of alloparental care of a female Lesser Kestrel in an alien nest, 70-73 Coulson, Jennifer O., Mississippi Kites use Swallow-tailed Kite nests, 155—156 Crested Caracara, 203-206 Cromrich, Lee A., Denver W. Holt, and Shawne M. Lea- sure, Trophic niche of North American Great Horned Owls, 58-65 Culver, Melanie, see Tingay, Ruth E. Cumulative impacts, 55-61 (suppl.) Cytochrome b, 183-187 D Dams, 245-255 de Vries, Tjitte and Cristian Melo, Micrastur or Accipiter, that is the question, 337 Deforestation, 51-57 Degraaf, Richard M., see Smith, Harvey R. Demography, 3-10 Denali National Park, 50-54 (suppl.) Density, 24—32 Development, 3-10, 77-81 Dho-gaza, 320-323 Diet, 24-32, 58-65, 148-152, 328-331 assessment methods, 11-16 Dispersal, 176-182, 309-314 natal, 203—206 roost, 45-50 Distribution, 10-17 (suppl.) Disturbance, 294—299 DNA, fingerprinting, 280-286 mitochondrial, 17—23 multilocus fingerprinting, 70-73 polymorphisms, 218-219 Donazar, Jose Antonio, Juan Jose Negro, Cesar Javier Pa- lacios, Laura Gangoso,Jose Antonio Godoy, Olga Ce- ballos, Fernando Hiraldo, and Nieves Capote, De- scription of a new subspecies of the Egyptian Vulture (Accipitridae: Neophron percnopterus) from the Canary Islands, 17-23 Driscoll, Daniel E., see Hunt, W. Grainger Durango, 3-9 (suppl.) 346 Index to Volume 36 VoL. 36, No. 4 Dykstra, Cheryl R., Michael W. Meyer, and D. Keith Warn- ke, Bald Eagle reproductive performance following video camera placement, 136-139 Dykstra, Cheryl R., see Warnke, D. Keith E Eagle, Bald, 29-31 (suppL), 121-127, 136-139, 161-169, 245-255, 256-264, 324-327 Bonelli’s, 231-235 Crested, 77-81 Crowned, 206-212 Crowned Hawk-, 300—308 Golden, 3-9 (suppL), 10-17 (suppL), 18-19 (suppl.), 20-24 (suppl.), 25-28 (suppl.), 29-31 (suppl.), 32-40 (suppl.), 41-49 (suppl.), 50-54 (suppl.), 55-61 (suppl.), 62-69 (suppl.), 70-77 (suppl.) Long-crested, 51-57 Madagascar Fish-, 280-286, 287-293, 309-314 Short-toed, 111-114 Wahlberg’s, 51-57 White-tailed Sea, 220-224 Eberhart, Rolf, see Brendel, Ulrich M. Effigy, 45-50 Eggs, 324-327 yolk, 33-38 Eggshell characteristics, 324-327 Elat, 115-120 Eldegard, Katrine, see Beheim, Janne Elliot, James D., A review of Birds of Prey: Health & Dis- ease, by John E. Cooper, 2002, 338 Elhs, David H., Beth Ann Sabo, James K. Fackler, and Brian A. Millsap, Prey of the Peregrine Falcon {Falco peregtinus cassini) in southern Argentina and Chile, 315-319 Ellis, David H., Lynn W. Oliphant, and James K. Fackler, Schizochromism in a Peregrine Falcon from Arizo- na, 200-202 Emergent trees, 300-308 Englund, Judy Voigt, see Martell, Mark S. Environmental education, 20-24 (suppl.) European Alps, 20-24 (suppl.) F Fackler, James K., see Ellis, David H., 200-202 Fackler, James K., see Ellis, David H., 315-319 Fahler, Natalie A. and Lester D. Flake, Nesting of Long- eared Owls along the lower Big Lost River, Idaho: a comparison of 1975-76 and 1996-97, 73-77 Falco deiroleucus, 39-44 eleonorae, 139-141 naumanni, 70-73, 148-152 peregtinus, 176-182, 200-202, 203-217, 315-319 tinnunculus, 81—84, 84-86 Falcon, Eleonora’s, 139-141 Orange-breasted, 39-44 Pallid, 315-319 Peregrine, 176-182, 200-202, 213-217, 315-319 Fatty acids, 33-38 Feeding, 144-145, 152-153 habits, 224-228 Ferrer, Miguel, see Balbontin, Javier Filloy, Julieta, see Bellocq, M. Isabel Finn, Sean R, Daniel E. Varland, and John M. Marzluff, Does Northern Goshawk breeding occupancy vary with nest-stand characteristics on the Olympic Pen- insula, Washington?, 265-279 Finn, Sean R, see Bloxton, Thomas D. Firmanszky, Gabor, The status of the Golden Eagle {Aq- uila chrysaetos) in Hungary, 18-19 (suppl.) Flake, Lester D., see Fahler, Natalie A. Flocking, 111-114 Florida, 3-10, 203-206 Flyways, 97-110 Food, 144-145 -niche breadth, 58-65 Foraging habitat, 220-224 Forestry, 24—32 Fox, Nicholas C., see Barton, Nigel W.H. Fractures, 229-230 Framework, 41—49 (suppl.) Fraser, James D., see Berkelman, James Fraser, James D., see Tingay, Ruth E. French, Thomas W., see Roth, Aaron J. G Gangoso, Laura and Cesar J. Palacios, Endangered Egyp- tian Vulture {Neophron percnopterus) entangled in a power line ground-wire stabilizer, 238-239 Gangoso, Laura, see Donazar, Jose Antonio Genetics, 183-187 Gladdium brasilianum, 170-175 Godoy, Jose Antonio, see Donazar, Jose Antonio Grant, Justin R., see McGrady, Michael J. Great Lakes, 136-139 Griffiths, Carole S. and John M. Bates, Morphology, ge- netics and the value of voucher specimens: an ex- ample with Cathartes vultures, 183-187 Gypaetus barbatus, 66-70 H Habitat, 287-293 management, 55-61 (suppl.) preferences, 224-228 quality models, 20-24 (suppl.) selection, 245-255 use, 51-57 Haliaeetus albicilla, 220-224 leucocephalus, 29-31 (suppl.), 121-127, 136-139, 161- 169, 245-255, 256-264, 324-327 vociferoides, 280-286, 287-293, 309-314 Hallerman, Eric M., see Tingay, Ruth E. December 2002 Index to Volume 36 347 Harmata, Alan R., Vernal migration of Bald Eagles from a southern Colorado wintering area, 256-264 Harpy haliaetus coronatus, 206-212 Hatcher, Jane E., see Avery, Michael L. Hawk, Cooper’s, 229-230 Red-shouldered, 152-153 Sharp-shinned, 229-230 Heie, Olav, see see Beheim, Janne Hieraaetus fasciatus, 231-235 Hiraldo, Fernando, see Donazar, Jose Antonio Hoffman, Stephen W., Jeff R Smith, and Timothy D. Meehan, Breeding grounds, winter ranges, and mi- gratory routes of raptors in the mountain west, 97- 110 Hokkaido, Japan, 220-224 Holt, Denver W., see Cromrich, Lee A. Home range, 70-77 (suppL), 245-255 Homosexual matings, 66-70 Humphrey, John S., see Avery, Michael L. Hungary, 18-19 (suppl.) Hunt, W. Grainger, Ronald E. Jackman, Daniel E. Dris- coll, and Edward W. Bianchi, Foraging ecology of nesting Bald Eagles in Arizona, 245-255 I Idaho, 73-77 Immobilization, 188-193 Ingraldi, Michael R, see Bloxton, Thomas D. Injuries, 229-230 Isaksson, Mats, see Beheim, Janne Italy, 11-16, 24-32, 128-135 J Jackman, Ronald E., see Hunt, W. Grainger Jacobs, Eugene A. and Glenn A. Proudfoot, An elevated net assembly to capture nesting raptors, 320-323 Janovsky, Martin, Thomas Ruf, and Wolfgang Zenker, Oral administration of tiletamine/zolazepam for the immobilization of the Common Buzzard {Buteo bu- teo), 188-193 Jenkins, M. Alan, Steve K. Sherrod, David A. Wiedenfeld, and Donald H. Wolfe, Jr., Florida Bald Eagle {Hal- iaeetus leucocephalus) egg characteristics, 324-327 Jones, Gwilym S., see Roth, Aaron J. K Karasov, William H., see Warnke, D. Keith Kery, Marc, New observations of the Peregrine Falcon {Falco peregrinus) in Peru, 213-217 Kestrel, Eurasian, 81-84, 84-86 Lesser, 70-73, 148-152 Klungland, Helge, see see Beheim, Janne Kochert, Michael N. and Karen Steenhof, Golden Eagles in the G.S. and Canada: status, trends, and conser- vation challenges, 32-40 (suppl.) Kopij, Grzegorz, Food of the l,esser Kestrel {Falco nau- manni) in its winter quarters in South Africa, 148- 152 Krone, Oliver, Fatal Caryospora infection in a free-living Juvenile Eurasian Kestrel {Falco tinnunculus) , 84—86 L land use, 55-61 (suppl.) Landaeta, Carlos A., see Vargas, Julieta Leasure, Shawne M., see Cromrich, Lee A. Leucism, 200-202 Linda, Lomo, see Nunnery, Tony Logistic regression, 265-279 Londei, Tiziano, The Fox Kestrel {Falco alopex) hovers, 236-237 Lophaetus occipitalis, 51-57 M Madagascar, 287-293 Madders, Mike and Dave Walker, Golden Eagles in a mul- tiple land-use environment: a case study in conflict management, 55-61 (suppl.) Malan, Gerard and Susanne Shultz, Nest-site selection of the Crowned Hawk-Eagle in the forests of Kwazulu- Natal, South Africa, and Tai, Ivory Coast, 300-308 Management, 20-24 (suppl.), 294—299 Manganaro, Alberto, see Salvati, Luca Mangrove, 328-331 Marches!, Luigi, Paolo Pedrini, and Fabrizio Sergio, Bi- ases associated with diet study methods in the Eur- asian Eagle Owl, 11-16 Margalida, Antoni, see Bertran, Joan Mark-recapture, 3-10 model, 176-182 Marks, Jeff, A review of Owls, by Floyd Scholz, 2001, 338- 339 Martell, Mark S., Judy Voigt Englund, and Harrison B. Tordoff, An urban Osprey population established by translocation, 91—96 Martinez-Abrain, Alejandro and Gerardo Urios, Absence of blood parasites in nestlings of the Eleonora’s Fal- con {Falco eleonorae), 139-141 Marzluff, John M., a review of The Spanish Imperial Ea- gle, by Miguel Ferrer, 2001, 242-244 Marzluff, John M., see Bloxton, Thomas D. Marzluff, John M., see Finn, Sean P. Mating system, 280-286 Mays, Jody L., see Proudfoot, Glenn A. McAllister, Kelly R., see Watson, James W. McGrady, Michael J., Justin R. Grant, Ian P. Bainbridge, and David R.A. McLeod, A model of Golden Eagle {Aquila chrysaetos) ranging behavior, 62-69 (suppl ) McGrady, Michael J., see Bechard, MarcJ. McGrady, Michael J., see McLeod, David R.A McIntyre, Carol L., Patterns in nesting area occupancy and reproductive success of Golden Eagles {Aquila 348 Index to Volume 36 VOL. 36, No. 4 chrysaetos) in Denali National Park and Preserve, Alaska, 1988-99, 50-54 (suppl.) McLeod, David R.A., D. Philip Whitfield, and Michael J. McGrady, Improving prediction of Golden Eagle {Aquila chrysaetos) ranging in western Scotland using GIS and terrain modeling, 70-77 (suppl.) McI.eod, David R.A., see McGrady, Michael J. Mediterranean areas, 81-84 Meehan, Timothy D., see Hoffman, Stephen W. Melo, Cristian, see de Vries, Tjitte Mexico, 3-9 (suppl.) Meyer, Michael W., see Dykstra, Cheryl R. Meyer, Michael W., see Warnke, D. Keith Microsatellite, 218-219 Migration, 97-110, 111-114 differential, 97-110 spring, 115-120 vernal, 256-264 Millar, Jody Gustitus, The protection of eagles and the Bald and Golden Eagle Protection Act, 29-31 (suppl.) Miller, Richard S., see Smith, Harvey R. Millsap, Brian A., see Ellis, David H. Millsap, Brian A., Survival of Florida Burrowing Owls along an urban-development gradient, 3-10 Minnesota, 91-96 Mizera, Tadeusz, see Waclawek, Krzysztof Monitoring, 32-40 (suppl.) Morphnus guianensis, 77-81 Morrison, Joan L., see Nemeth, Nicole M. Mortality, 32-40 (suppl.), 141-143 Mottajunior, Jose Carlos, Diet of breeding Tropical Screech-Owls (Otus choliba) in southeastern Brazil, 332-334 Movements, 62-69 (suppl.) N Navigation, 111-114 Negro, Juan Jose, see Donazar, Jose Antonio Nemeth, Nicole M. andjoan L. Morrison, Natal dispersal of the Crested Caracara {Caracara cheriway) in Flori- da, 203-206 Neofjhron percnopterus majormsis, 1 7-23 Nest, access, 300-308 helper, 280-286 tree, 287-293 Nest-site selection, 300-308 Nesting, 73-77 biology, 77-81 Nestlings, 77-81, 139-141 free-living, 231-235 Net, drop, 320-323 elevated, 320-323 mist, 320-323 New records, 206-212 Ntnox strenua, 294—299 North America, 58-65, 229-230 western, 97-110 Northern Goshawk, 141-143, 229-230, 265-279 Nunnery, Tony, Lomo Linda, and Mark R. Welford, Barred Forest-Falcon {Micrastur ruficollis) predation on a hummingbird, 239-240 O Olfaction, 144-145 Oliphant, Lynn W., see Ellis, David H. Olmos, Fabio, see Silva e Silva, Robson Oral administration, 188-193 Orientation, 111-114 Osprey, 91-96, 328-331 Otus asio, 194-199 choliba, 332-334 Overstory depth, 265-279 Owens, Thomas E., see Watson, James W. Owl, Barn, 146-148, 224-228 Boreal, 218-219 Burrowing, 3-10 Eastern Screech-, 194-199 Eurasian Eagle-, 11-16 Ferriginous Pygmy-, 170-175 Great Horned, 58-65 Long-eared, 73-77 Mechanical, 320—323 Powerful, 294-299 Tengmalm’s, 218-219 Tropical Screech-, 332-334 P Palacios, Cesar Javier, see Donazar, Jose Antonio Palacios, Cesar J., see Gangoso, Laura Pandion haliaetus, 91-96, 328-331 Parental roles, 66-70 Parkin, David T, see Cordero, Pedro J. Pedrini, Paolo, see Marches!, Luigi Perch time, 194—199 tree, 287-293 Peru, 213-217 Pesticides, 324-327 Phares, Kimberly O., see Avery, Michael L. Piscivory, 245-255 Plasma, 33-38 chemistry, 231-235 Poland, 25-28 (suppl.) Polyandry, 280-286 cooperative, 66-70 Population restoration, 176-182 Post-fledging, length, 157 movement, 220-224 status, 161-169 Powell, Larkin A., Dan J. Calvert, Irene M. Barry, and Lowell Washburn, Post-fledging survival and dispers- Dec;ember 2002 Index to Voeume 36 349 al of Peregrine Falcons during a restoration project, 176-182 Pranty, Bill, Red-shouldered Hawk feeds on carrion, 152- 153 Pre-Alps, 24-32 Premuda, Guido, see Agostini, Nicolantonio Prey, 141-143, 315-319 biomass, 332-334 size, 146-148 Productivity, 91-96, 161-169 Protozoa, 84-86 Proudfoot, Glenn A., Sam L, Beasom, Felipe Chavez-Ra- mirez, and Jody L. Mays, Response distance of Fer- ruginous Pygmy-Owls to broadcasted conspecific calls, 170-175 Proudfoot, Glenn A., see Jacobs, Eugene A. Provisioning, 121-127 R Radio-tracking, 256-264 Radiotelemetry, 245-255 Rafanomezantsoa, Simon, Richard T. Watson, and Russell Thorstrom, Juvenile dispersal of Madagascar Fish-Ea- gles tracked by satellite telemetry, 309-314 Ramirez-Llorens, Patricio, see Bellocq, M. Isabel Ranazzi, Lamberto, see Salvati, Luca Range, 62-69 (suppl.) model, 70-77 (suppl.) use, 194-199 Raptor nutrition, 33-38 Raptors, 128-135 Recovery, 161-169 Repopulation, 25-28 (suppl.) Reproduction, 50-54 (suppl.) Reproductive performance, 136-139 Restani, Marco, A review of Raptors of the World, by James Ferguson-Lees and David A. Christie, 2001, 241-242 Riparian, 73-77 Ritchison, Gary, see Buhay, Jennifer E. Rivers, 245-255 Rodriguez-Estrella, Ricardo, A survey of Golden Eagles in northern Mexico in 1984 and recent records in cen- tral and southern Baja California Peninsula, 3-9 (suppl.) Rogers, Andi, see Bloxton, Thomas D. Rome, 224-228 Rosenstock, Steve, see Bloxton, Thomas D. Roth, Aaron J., Gwilym S. Jones, and Thomas W. French, Incidence of naturally-healed fractures in the pec- toral bones of North American Accipiters, 229-230 Ruf, Thomas, see Janovsky, Martin S Sabo, Beth Ann, see Ellis, David H. Salvati, Luca, Lamberto Ranazzi, and Alberto Manganaro, Habitat preferences, breeding success, and diet of the Barn Owl ( Tyto alba) in Rome: urban versus rural territories, 224-228 Salvati, Luca, Spring weather and breeding success of the Eurasian Kestrel {Falco tinnunculus) in urban Rome, Italy, 81-84 Saskatchewan, 256-264 Satellite telemetry, 309-314 Scandolara, Chiara, see Sergio, Fabrizio Scavenging, 144-145 Schizochromism, 200-202 Scotland, 62-69 (suppl.) Seavy, Nathaniel E. and Christine K. Apodaca, Raptor abundance and habitat use in a highly-disturbed- for- est landscape in western Uganda, 51-57 Sergio, Fabrizio, Alberto Boto, Chiara Scandolara, and Guiseppe Bogliani, Density, nest sites, diet, and pro- ductivity of Common Buzzards {Buteo buteo) in the Italian pre-Alps, 24-32 Sergio, Fabrizio, see Marchesi, Luigi Sex, 231-235 Shergalin, Jevgeni, see Abuladze, Alexander Sherrod, Steve K., see Jenkins, M. Alan Shiraki, Saiko, Post-fledging movements and foraging habitats of immature White-tailed Sea Eagles in the Nemuro Region, Hokkaido, Japan, 220-224 Shoreline, 297-293 Shrub cover, 265-279 Shultz, Susanne, see Malan, Gerard Silva e Silva, Robson, and Fabio Olmos, Osprey ecology in the mangroves of southeastern Brazil, 328-331 Simonetti, Javier A., see Vargas, Julieta Smith, Harvey R., Richard M. DeGraaf, and Richard S. Miller, Exhumation of food by Turkey Vulture, 144- 145 Smith, Jeff R, see Hoffman, Stephen W. Solifugae, 148-152 Sonerud, Geir, see Beheim, Janne Sonora, 3-9 (suppl.) South Africa, 148-152 South America, 213-217 Spain, 66-70 Speake, Brian K., see Barton, Nigel W.H. Status, 32-40 (suppl.), 206-212 Steenhof, Karen, A review of The Raptor Almanac, by Scott Weidensaul, 2000, 87-88 Steenhof, Karen, see Kochert, Michael N. Stephanoaetus coronatus, 300—308 Stinson, Derek, see Watson, James W. Subspecies, 17-23 Surai, Peter E, see Barton, Nigel W.H. Survey, 39-44 broadcast, 170-175 roadside, 51-57 winter, 128-135 Survival, 3-10, 176-182 350 Index to Volume 36 VoL. 36, No. 4 T Tea plantation, 51-57 Temperature, 294-299 Territoriality, 62-69 (suppl.) Territories, 224-228 Thorstrom, Russell, Comments on the first nesting re- cord of the nest of a Slaty-backed Forest-Falcon {Mi- crastur mirandollei) in the Ecuadorian Amazon, 335- 336 Thorstrom, Russell, Richard Watson, Aaron Baker, Ser- ena Ayers, and David L. Anderson, Preliminary ground and aerial surveys for Orange-breasted Fal- cons in Central America, 39-44 Thorstrom, Russell, see Rafanomezantsoa, Simon Tiletamine, 188-193 Tillman, Eric A., see Avery, Michael L. Time budgets, 121-127 Tmgay, Ruth E., Melanie Culver, Eric M. Hallerman, James D. Fraser, and Richard T. Watson, Subordinate males sire offspring in Madagascar Fish-Eagle {Hal- iaeetus vociferoides) polyandrous breeding groups, 280-286 Toffoli, Roberto, see Boano, Giovanni Tordoff, Harrison B., see Martell, Mark S. Transect, 128-135 Translocation, 91-96 Tropical, 77-81 Tryjanowski, Piotr, see Yosef, Reuven Tyto alba, 146-148, 224-228 U Uganda, 51-57 Urban, habitats, 81-84, 224—228 wildlife, 91-96 wildlife management, 3-10 Urbanization, 294—299 Urios, Gerardo, see Martinez-Abrain, Alejandro U.S., 32-40 (suppl.) V Vargas, Julieta, Carlos Landaeta A., and Javier A. Simo- netti. Bats as prey of Barn Owls { Tyto alba) in a trop- ical savanna in Bolivia, 146-148 Variability, 218-219 Varland, Daniel E., see Finn, Seair P. Vegetation structure, 294-299 Video camera, 136-139 Vitamins, 33-38 Voucher specimens, 183-187 Vultures, 45-50 Bearded, 66-70 Egyptian, 17-23 Turkey, 144—145 Yellow-headed, 183-187 W Waclawek, Krzysztof and Tadeusz Mizera, The status of the Golden Eagle {Aquila chrysaelos) in Poland, 25- 28 (suppl.) Wallis, Robert, see Cooke, Raylene Warnke, D. Keith, David E. Andersen, Cheryl R. Dykstra, Michael W. Meyer, and William H. Karasov, Provi- sioning rates and time budgets of adult and nestling Bald Eagles at inland Wisconsin nests, 121-127 Warnke, D. Keith, see Dykstra, Cheryl R. Washburn, Lowell, see Powell, Larkin A. Washington, 161-169, 265-279 Watson, James W., Derek Stinson, Kelly R. McAllister, and Thomas E. Owens, Population status of Bald Eagles breeding in Washington at the end of the 20^’^ cen- tury, 161-169 Watson, Jeff and Philip Whitfield, A conservation frame- work for the Golden Eagle {Aquila chrysaetos) in Scot- land, 41-49 (suppl.) Watson, Richard, see Thorstrom, Russell Watson, Richard T., see Berkelman, James Watson, Richard T., see Rafanomezantsoa, Simon Watson, Richard T., see Tingay, Ruth E. Weather conditions, 81-84 Welford, Mark R., see Nunnery, Tony Whitacre, David R, Juventino Lopez Avila, and Gregorio Lopez Avila, Behavioral and physical development of a nestling Crested Eagle {Morphnus guianensis) , 77— 81 White, John, see Cooke, Raylene Whitheld, D. Philip, see McLeod, David R.A. Whitfield, Philip, see Watson, Jeff Wiedenfeld, David A., see Jenkins, M. Alan Wiesmann, Karen, see Brendel, Ulrich M. Wildlife, habitat relationships, 265-279 management, 300-308 Wind energy, 55-61 (suppl.) Winter quarters, 148-152 Wintering, 256-264 ecology, 328-331 Wisconsin, 121-127 Wolfe, Jr., Donald H., see Jenkins, M. Alan Wood, Petra Bohall, see Ammer, Frank K. Y Yosef, Reuven, Piotr Tryjanowski, and Keith Bildstein, Spring migration of adult and immature buzzards {Buteo buteo) through Elat, Israel; timing and body size, 115-120 Z Zacatecas, 3-9 (suppl.) Zemplen Mountains, 18-19 (suppl.) Zenker, Wolfgang, see Janovsky, Martin Zolazepam, 188-193 Zones, 41-49 (suppl.) THE JOURNAL OF RAPTOR RESEARCH A QUARTERLY PUBLICATION OF THE RAPTOR RESEARCH FOUNDATION, INC. (Founded 1966) EDITOR IN CHIEF James C. Bednarz James R. Belthoff Clint W. Boal Joan L. Morrison Juan Jose Negro ASSOCIATE EDITORS Marco Restani Ian G. Warrentin Troy I. Wellicome BOOK REVIEW EDITOR Jeffrey S. Marks CONTENTS FOR VOLUME 36, 2002 Number 1 Editor’s Page: Working Toward Excellence. James C. Bednarz 1 Survival of Florida Burrowing Owls Along an Urban-Development Gradient. Brian A. Millsap 3 Biases Associated with Diet Study Methods in the Eurasian Eagle-Owl. Luigi Marchesi, Paolo Pedrini, and Fabrizio Sergio 11 Description of a New Subspecies of the Egyptian Vulture (Accipitridae; Neo- phron fercnopterus) FROM THE Canary Islands. Jose Antonio Donazarjuan Jose Negro, Cesar Javier Palacios, Laura Gangoso, Jose Antonio Godoy, Olga Ceballos, Fernando Hiraldo, and Nieves Capote 17 Density, Nest Sites, Diet, and Productivity of Common Buzzards {Bijteo buteo) in the Italian Pre-Alps. Fabrizio Sergio, Alberto Boto, Chiara Scandolara, and Giuseppe Bogliani 24 Vitamins E and A, Carotenoids, and Fatty Acids of the Raptor Egg Yolk. Nigel W.H. Barton, Nicholas C. Fox, Peter F. Surai, and Brian K. Speake 33 Preliminary Ground and Aerial Surveys for Orange-breasted Falcons in Central America. Russell Thorstrom, Richard Watson, Aaron Baker, Serena Ayers, and David L. Anderson 39 Dispersing Vulture Roosts on Communication Towers. Michael l. Avery, John s, Humphrey, Eric A. Tillman, Kimberly O. Phares, and Jane E. Hatcher 45 Raptor Abundance and Habitat Use in a Highly-Disturbed-Forest Landscape in Western Uganda. Nathaniel E. Seavy and Christine K. Apodaca 51 Trophic Niche of North American Great Horned Owix. Lee a. Cromrich, Denver w. Holt, and Shawne M. Leasure 58 Short Communications Social Organization of a Trio Of Bearded Vultures ( Gypaetus barbatus) : Sexual and Parental Roles. Joan Bertran and Antoni Margalida 66 Genetic Evidence of Ali.oparental Care of a Female Lesser Kestrei. in an Alien Nest. Pedro J. Cordero, Jose M. Aparicio, and David T. Parkin 70 NE.STING OF Long-eared Owls Along the Lower Big Lost River, Idaho: A Comparison of 1975-76 and 1996-97. Natalie A. Fahler and Lester D. Flake 73 Behavioral and Ph\sical Development of a Nestling Crested Eagle {Morphnus GuiANENsrs) . David F. Whitacre, Juventino Lopez Avila, and Gregorio Lopez Avila 77 Spring Weather and Breeding Success of the Eurasian Kestrel {Falco riNNUNCULUs) in Urban Rome, Italy. Luca Salvati 81 Eatal Caryospora Infection in a Free-living Juvenile Eurasian Kestrel {Falco riNmmcuLus) .Olivei: Krone 84 Book Review. Edited by Jeffrey S. Marks 87 Manuscript Referees 89 Number 2 An Urban Osprey Population Established by Translocation. Mark s. Marteii, Judy Voigt Englund, and Harrison B. Tordoff 91 Breeding Grounds, Winter Ranges, and Migratory Routes of Raptors in the Mountain West. Stephen W. Hoffman, Jeff P. Smith, and Timothy D. Meehan 97 Circuitous Autumn Migration in the Short-toed Eagle ( Circaetus gallicus) . Nicolantonio Agostini, Luca Baghino, Charles Coleiro, Ferdinando Corbi, and Guido Premuda Ill Spring Migration of Adult and Immature Buzzards {Buteo buteo) through Elat, Israel: Timing and Body Size. Reuven Yosef, Piotr Tryjanowski, and Keith L. Bildstein 115 Provisioning Rates and Time Budgets of Adult and Nestling Bald Eagles AT Inland Wisconsin Nests. D. Keith Wamke, David E. Andersen, Cheryl R. Dykstra, Michael W. Meyer, and William H. Karasov 121 A Line Transect Survey of Wintering Raptors in the Western Po Plain of Northern Italy. Giovanni Boano and Roberto Toffoli 128 Short Communications Bald Eagle Reproductive Performance Following Video Camera Placement. Cheryl R. Dykstra, Michael W. Meyer, and D. Keith Warnke 136 Absence of Bi.ood Parasites in Nestlings of the Eleonora’s Falcon {Falco ejmonoraf). Alejandro Martinez-Abram and Gerardo Urios 139 Possible Choking Mortalities of Adult Northern Goshawks. Thomas D. Bloxton, Andi Rogers, Michael F. Ingraldi, Steve Rosenstock, John M. Marzluff, and Sean P. Finn 141 Exhumation of Food by Turkey Vulture. Harvey R. Smith, Richard M. DeGraaf, and Richard S. Miller 144 Bats as Prey of Barn Owls ( lYro alba) in a Tropical Savanna in Bolivia. Julieta Vargas, Carlos Landaeta A., and Javier A. Simonetti 146 Food of the Lesser Kestrel (Falco naumanni) in its Winter Quarters in South Africa. Grzegorz Kopij 148 Red-shouldered Hawk Feeds on Carrion. Bill Pranty 152 Letters First Repiacement Clutch by a Polvandrous Trio of Bearded Vultures (Gypaetus baebatus) in the Spanish Pyrenees. Antoni Margalida and Joan Bertran 154 Mississippi Kites Use Swallow-tailed Kite Nests. Jennifer O. Coulson 155 Erratum 157 Number 3 Population Status of Breeding Bald Eagles in Washington at the end of the 20th Century. James W. Watson, Derek Stinson, Kelly R. McAllister, and Thomas E. Owens 161 Response Distance of Ferruginous Pygmy-Owls to Broadcasted Conspecific Calls. Glenn a. Proudfoot, Sam L. Beasom, Felipe Chavez-Ramirez, and Jody L. Mays 170 POST-FLEDGING SURVIVAL AND DISPERSAL OF PEREGRINE FALCONS DURING A RESTORATION Project. LarkinA. Powell, DanJ. Calvert, Irene M. Barry, and Lowell Washburn 176 Morphology, Genetics, and the Value of Voucher Specimens: An Example with CATT/ARTES Vultures. Carole S. Griffiths and John M. Bates 183 Oral Administration of Tiletamine/Zolazepam for the Immobilization of the Common Buzzard {BUTEO BUTEO) . Martinjanovsky, Thomas Ruf, and Wolfgang Zenker 188 Hunting Behavior of and Space Use by Eastern Screech-Owls during the Breeding Season. Jennifer E. Buhay and Gary Ritchison 194 Short Communications ScHizocHROMiSM IN A PEREGRINE Falcon erom ARIZONA. David H. Ellis, Lynn W. Oliphant, and James K. Fackler 200 Natal Dispersal of the Crested Caracara (Caracara cherjway) in Florida. Nicole M. Nemeth and Joan L. Morrison 203 Recent Records of Crowned Eagles {Harpyhaliaetus coronatus) from Argentina, 1981-2000. M. Isabel Bellocq, Patricio Ramirez-Llorens, and Julieta Filloy 206 New Observations of the Peregrine Falcon {Falco peregrinus) in Peru. Marc Kery 213 DNA Polymorphisms in Boreai. Owls {Aegolius funereus) . Janne Beheim, Katrine Eldegard, Gro Bj0rnstad, Mats Isaksson, Geir Sonerud, Olav Heie, and Helge Klungland 218 Post-fledging Movements and Foraging Habitats of Immature White-tailed Sea Eagi.es in the Nemuro Region, Hokkaido, Japan. Saiko Shiraki 220 Habitat Preferences, Breeding Success, and Diet of the Barn Owl (Tyto alba) in Rome; Urban versus Rural Territories. Luca Salvati, Lamberto Ranazzi, and Alberto Manganaro 224 Incidence of Naturally-healed Fractures in the Pectoral Bones of North American Accipiters. Aaron J. Roth, Gwilym S. Jones, and Thomas W. French 229 Piasma Chemistry Reference Values in Free-living Bonelli’s Fagle (Hieraaetus fasciatus) Nestlings. Javier Balbontin and Miguel Ferrer 231 Letters The Fox Kestrel {Falco alopex) Hoytrs. Tiziano Londei 236 Probable Breeding of Short-eared Owls in Southern West Virginia. Frank K. Ammer and Petra Bohall Wood 237 Endangered Egyptian Vulture {Neophron percnopierus) Entangled in Powerline Ground-wire Stabilizer. Laura Gangoso and Cesar J. Palacios 238 Barred Forest-Falcon {Miciustur ruficollis) Predation on a Hummingbird. Tony Nunnery and Mark R. Welford 239 Book Reviews. Edited byjeffrey S. Marks 241 Number 4 Foraging Ecology of Nesting Bald Eagles in Arizona, w. Grainger Hunt, Ronald e. Jackman, Daniel E. Driscoll, and Edward W. Bianchi 245 Vernal Migration of Bai.d Eagles from a Southern Colorado Wintering Area. Alan R. Harmata 256 Does Northern Goshawk Breeding Occupancy Vary with Nest-stand Character- istics on the Olympic PeNINSUIA, Washington? Sean P. Finn, Daniel E. Varland, and John M. Marzluff 265 Subordinate Males Sire Offspring in Madagascar Fish-Eagle {Haliaeetus vociee- ROWES) POLYVNDROUS BREEDING GROUPS. Ruth E. Tingay, Melanie Culver, Eric M. Hallerman, James D. Fraser, and Richard T. Watson 280 Nesting and Perching Habitat Use of the Madagascar Fish-Eagle. James Berkeiman, James D. Fraser, and Richard T. Watson 287 Use of Vegetative Structure by Powerful Owls in Outer Urban Melbourne, Victoria, Austraija — Implications for Management. Rayiene Cooke, Robert WaiUs, and John White 294 Nest-Site Selection of the Crowned Hawk-Eagle in the Eorests of Kwazulu- NATAL, South Africa, and TAI, Ivory Coast. Gerard Malan and Susanne Shultz 300 Short Communications Juvenile Dispersal of Madagascar Fish-Eagles Tracked by Satellite Telemetry. Simon Rafanomezantsoa, Richard T. Watson, and Russell Thorstrom 309 Prey OF the Peregrine Falcon {Falco Peregrinus cassini) in Southern Argentina and Chile. David H. Ellis, Beth Ann Sabo, James K. Fackler, and Brian A. Millsap 315 An Elevated Net Assembly to Capture Nesting Raptors. Eugene A. Jacobs and Glenn A. Proudfoot 320 Florida Bald Eagle {Haliaeetus teucocephalus) Egg Characteristics. M. Alan Jenkins, Steve K. Sherrod, David A. Wiedenfeld, and Donald H. Wolfe, Jr. 324 Osprey Ecology IN THE Mangroves OF Southeastern Brazil. Robson Silva e Silva, Fabio Olmos .... 328 Diet of Breeding Tropical Screech-Owls ( Otus choliba) in Southeastern Brazil. Jose Carlos Motta-Junior 332 Letters Comments of the First Nesting Record of the Nest of a Slaty-backed Forest Falcon {Micrastur MIRANDOLLEI) IN THE ECUADORIAN AMAZON. Russell Thorstrom 335 Micrastur or Accipiter, That is the Question. Tjitte de Vries and Cristian Melo Book Reviews. Edited by Jeffery S. Marks Information For Contributors 337 338 340 Index to Volume 36 344 A Telemetry Receiver Designed with The Researcher in Mind What you've been waiting fori fui.illv, ,1 h.Kjhly th.ifine! telemetry receiver th.it wrtqhs less than 13 ounces, is (ompleteiy user proqr.imm.ible