The Journal of Volume 35 Number 3 September 2001 Published by The Raptor Research Foundation, Inc THE RAPTOR RESEARCH EOUNDATION, INC. (Founded 1966) OFFICERS PRESIDENT; Michael N. Kochert SECRETARY: Patricia A. Hall VICE-PRESIDENT: Keith L. Bildstein TREASURER: Jim Eitzpatrick BOARD NORTH AMERICAN DIRECTOR #1: Philip Detrich NORTH AMERICAN DIRECTOR #2; Laurie J. Goodrich NORTH AMERICAN DIRECTOR #3: Robert Lehman INTERNATIONAL DIRECTOR #1: Eduardo Inigo-Elias INTERNATIONAL DIRECTOR #2: Ricardo Rodriquez-Estrella OF DIRECTORS INTERNATIONAL DIRECTOR #3: Beatriz Arroyo DIRECTOR AT LARGE #1: Jemima ParryJones DIREGTOR AT LARGE #2: Petra Bohall Wood DIRECTOR AT lAJRGE #3; Michael W. Collopy DIRECTOR AT LARGE #4: Miguel Ferrer DIRECTOR AT LARGE #5: Robert N. Rosenfield DIRECTOR AT LARGE #6; Brian A. Millsap EDITORIAL STAFF EDITORS: James C. Bednarz, Department of Biological Sciences, P.O. Box 599, Arkansas State University, State University, AR 72467 U.S.A. MarcJ. Bechard, Department of Biology, Boise State University, Boise, ID 83725 U.S.A. ASSOCIATE EDITORS Juan Jose Negro Cole Crocker-Bedford Ian G. Wa rken ttn Clint W. Boal Marco Restani BOOK REVIEW EDITOR: Jeffrey S. Marks, Montana Cooperative Research Unit, University of Montana, Missoula, MT 59812 U.S.A. SPANISH EDITOR: Cesar Marquez Rew:s, Instituto Humboldt, Colombia, AA. 094766, Bogota 8, Colombia EDITORIAL ASSISTANTS; Joan Clark, Rebecca S. Maul The Journal of Raptor Research is distributed quarterly to all current members. Original manuscripts dealing with the biology and conservation of diurnal and nocturnal birds of prey are welcomed from throughout the world, but must be written in English. Submissions can be in the form of research articles, letters to the editor, thesis abstracts and book reviews. Contributors should submit a typewritten original and three copies to the Editor. All submissions must be typewritten and double-spaced on one side of 216 X 278 mm (81^ X 11 in.) or standard international, white, bond paper, with 25 mm (1 in.) margins. The cover page should contain a tide, the author’s full name(s) and address(es). Name and address should be centered on the cover page. If the current address is different, indicate this via a footnote. A short version of the 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 1990). Refer to a recent issue of the journal for details in format. Explicit instructions and publication policy are outlined in “Information for contributors,”/. Raptor Res., Vol. 34(4), and are available from the editor. Submit manuscripts to J. Bednarz at the address listed above. COVER: Adult Bonelli’s Eagle {Hieraaetus fasciatus). Painting by Hans Peeters. Contents Long-Term Changes of Raptor Populations in Northern Cameroon. jean-Marc Thiollay 173 Sexing Bonelu’s Eagle Nestlings: Morphometrics Versus Molecular Techniques. Luis Palma, Sara Mira, Pedro Cardia, Pedro Beja, Thomas Guillemaud, Nuno Ferrand, M. Leonor Cancela, and Luis Cancela da Fonseca 187 Embryonic Development of the American Kestrel {Falco sparvebius) : External Criteria For Staging. J.M. Pisenti, G.M. Santolo,J.T. Yamamoto, and A.A. Morzenti 194 Anal^is of Bald Eagle Spatial Use of Linear Habitat. Alan R. Harmata and George j. Montopoli 207 Habitat Use, Population Density, and Home Range of Eit Owls {Micrathene WHETNEYI) AT SaNTA AnA NATIONAL WILDLIFE ReFUGE, TeXAS. Christopher M. Gamel and Timothy Brush 214 Diets of Northern Barred Owls and Northern Spotted Owls in an Area of Sym- PATRY. Thomas E. Hamer, David L. Hays, Clyde M. Senger, and Eric D. Forsman 221 Factors Influencing Length of the Post-Fledging Period and Timing of Disper- sal IN Bonelli’s Eagle (Hieraaetus fasciatus) in Southwestern Spain. Eduardo Minguez, Elena Angulo, and Vivian Siebering 228 Dispersion, Habitat Use, Hunting Behavior, Vocalizations, and Conservation Status of the New Guinea Harpy Eagle {Harpyopsis novaeguineae) . Mark Watson and Smith Asoyama 235 Seasonal and Geographic Differences in the Diet of the Barn Owl in an Agro- Ecosystem IN Northern Italy. Michela Bose and Franca Guidali 240 Potential Negative Effects of Collisions with Transmission Lines on a Bonelli’s Eagle Population. Santi Mahosa and Joan Real 247 Short Communications Agonistic Behavior of Cooper’s Hawks. Clint W. Boal 253 Ground-Nesting Osprevs in Utah. Clark S. Monson 257 Diet and Breeding Success of Eagle Owl in Southeastern Spain: Effect of Rabbit Haemorrhagic Disease. Jose E. Martinez and Jose E Calvo 259 Letter First Record of Tandem Flying in the King Vltlture (Sarcojramphus papa) . Marsha A. Schlee 263 Book Reviews. Edited by Jeffrey S. Marks 265 The Raptor Research Foundation, Inc. gratefully acknowledges funds and logistical support provided by Arkansas State University to ^sist in the publication of the journal. THE JOURNAL OF RAPTOR RESEARCH A QUARTERLY PUBLICATION OL THE RAPTOR RESEARCH FOUNDATION, INC. VoL. 35 September 2001 No. 3 /. Raptor R£s. 35 (3) ; 1 73-186 © 2001 The Raptor Research Foundation, Inc. LONG-TERM CHANGES OE RAPTOR POPULATIONS IN NORTHERN CAMEROON Jean-Marc Thiollay Laboratoire d’Ecologie, E.N.S. 46, rue d’Ulm, 75230 Paris cedex 05, France Abstract. — Comparative mid-dry-season counts of diurnal raptors (42 species) were performed in sim- ilar conditions along nine road transects (total = 1359 km) in 1973 and 2000 in representative areas of northern Cameroon to assess changes in abundance of species over the last three decades. The transects covered Sudanian and southern Sahelian woodlands, savannas, cultivated areas, and wetlands. Additionally, in 2000, three protected woodlands and one new artificial wetland were censused to doc- ument the raptor community of these habitats. All four broad categories of raptors (vultures, other resident species, Afrotropical, and Palearctic migrants) exhibited a decline in numbers, and only two species increased significantly, against 14 which decreased and 24 which did not show population changes. Overall, vultures declined by 67%, notably the formerly abundant Hooded {Necrosyrtes monachus) and White-backed (Gyps africanus) Vultures and Ruppell’s Griffons (Gyps rueppellii) . Their decrease was striking in every habitat and area. All large eagles (seven residents and three Palearctic migrants) also decreased significantly (83%). Only the two smaller migratory Wahlberg’s (Aquila wahlhergi) and Booted Eagles (Hieraaetus pennatus) remained fairly stable or increased. Nine of the 11 smaller African residents did not change markedly; only the Black-shoul- dered Kite (Elanus caeruleus) increased three-fold and the Dark Chanting Goshawk (Melierax metabates) decreased significantly (62%). The five medium-size African migrants either decreased or increased, but not significantly, especially when differences in ecological conditions between years were taken into account. Among nine Palearctic migrants, kestrels decreased dramatically, whereas harriers remained at least stable. The new Maga Lake and associated rice fields were decidedly poorer in most Palearctic migrants than were the former natural humid grasslands which they replaced. Habitat modification due to the rapidly increasing human pressure explained most changes. However, the collapse of vulture populations cannot be accounted for by these changes or by persecution. Key Words: Raptors; population changes; savanna; Cameroon. Cambios poblacionales de rapaces a largo plazo en el norte de Camerun Resumen. — Comparamos los conteos de rapaces diurnas a mediados de la estacion seca (42 especies) los cuales fueron realizados en condiciones similares a lo largo de nueve transectos de carretera (total = 1.359 km) en 1973 y en el 2000; en areas representativas del norte de Camerun, con el fin de evaluar los cambios de abundancia de las especies en las ultimas tres decadas. Los transectos cubrian bosques Sudaneses y sur-sahelianos, sabanas, areas cultivadas y humedales. Adicionalmente en el 2000, se hici- eron censos en bosques perturbados protegidos y en un nuevo humedal artificial con el fin de docu- mentar la comunidad de aves rapaces de estos habitats. Las cuatro categorias de rapaces (buitres, otras especies residentes y las migratorias afrotropicales y palearcticas) mostraron una disminucion en sus numeros, solo dos especies tuvieron un incremento significativo en contraposicion de las 14 que dis- minuyeron y 24 que no mostraron cambios de poblacion. En general los buitres disminuyeron en un 67%, de los cuales sobresalen notablemente los anteriormente abundantes Necrosyrtes monachus. Gyps rueppellii. Su disminucion fue alarmante en cada area de habitat. Todas las grandes aguilas (siete resi- dentes y tres migratorias palearcticas) tambien disminuyeron significativamente (83%). Solo dos pe- quenas rapaces migratorias Aquila wahlbergiy Hieraaetus pennatus permanecieron estables o aumentaron. 173 174 Thiollay VoL. 35, No. 3 Nueve de la 1 1 especies residentes africanas no cambiaron significativamente; solo Elanus caeruleus in- cremento tres veces mientras que Melierax metabates disminuyo significativamente (62%). Las cinco mig- ratorias africanas de tamano mediano disminuyeron o aumentaron pero no en forma drastica, espe- cialmente cnando las diferencias en las condiciones ecologicas entre anos fueron tenidas en cuenta. Entre las nueve migratorias palearcticas, los cernicalos disminuyeron dramaticamente, mientras que los aguiluchos permanecieron por lo menos estables. La nueva represa Maga y los cultivos asociados de arroz fueron definitivamente mas pobres (en cuanto a migratorias palearticas respecta), que los pasti- zales humedos que reemplazaron. La modificacion de habitats debido al rapido aumento de la presion humana explica la mayoria de los cambios. Sin embargo el colapso de las poblaciones de buitres no puede ser atribuido a estos cambios o a la persecucion humana. [Traduccion de Cesar Marquez] Raptors are usually conspicuous, at the top of food chains and sensitive to contaminants and dis- turbance (Newton 1979). They have often substan- tial ecological requirements in terms of foraging habitats, nest sites, and abundance, or diversity of food resources. Therefore, they may be suitable in- dicators of the quality of ecosystems, e.g., large scale degradation, changes in land use, or insidious environmental contamination. When such changes come from climate change or human population growth, they may occur slowly and become appar- ent after long periods of time. Long-term monitor- ing of raptor communities is rare in Africa (Thiol- lay 1998) or involves only eagles (Brown 1952, 1955, Gargett 1990). 1 had the opportunity to do extensive raptor surveys at a 27-yr interval in an area of West Africa where substantial landscape changes occurred under a combination of human population growth and large-scale development programs. Such an evolution is common to many tropical countries and these results may be indic- ative of a widespread situation. The aim of this study was to assess the long-term changes, or sta- bility, in abundance of diurnal raptors (Falconifor- mes), both African species and Eurasian migrants, along the same transects recensused in similar con- ditions (observer, methods, season) nearly three decades after the first survey. The area harbored a rich raptor community and was an important win- tering zone for Palearctic species (Thiollay 1978a). These broad counts over large areas (6036 raptors on 2636 km) give a general picture likely to be representative of regional trends, but they do not provide information about mechanisms and un- derlying factors involved such as habitat changes, shifts in habitat selection, or dynamic of local pop- ulations. Study Area During the last 30 years, i.e., between our two censuses, the human population of Cameroon has tripled, from about 5 to 15 million. The northern part of the country has always been the most densely populated and, as a result, a region with a high deforestation rate and the highest proportion of cultivated lands (Encyclopaedia Universalis 1974, Encyclopaedia Britannica 1998). Defor- estation data are inaccurate for northern Cameroon be- cause the natural vegetation is not moist forest but open canopy deciduous woodlands that are often heavily de- graded rather than cleared, and cultivated areas may re- tain numerous large trees (WCMC 1994, Igemonger et al. 1997). The study area extends across the transition from the Sudanian to the Sahelian woodlands (9°10-11°30'N), along a south-north gradient of decreasing annual rain- fall (900-700 mm). Accordingly, the length of the rainy season decreases from May-October to July-September Mean annual temperature reaches 28°C, the hottest months being March-May and the lowest mean temper- atures in January ranging 19-13°C from south to north. The main vegetation zones (see Letouzey 1968, Louet- te 1981) include the Benue river plains in the south, the Mandara mountains along Nigeria (northernmost exten- sion of the Cameroon mountains ridge, highest peak = 1442 m), the Sahel in the central northern part and the inundation zone of the Logone river along the Chad bor- der (Fig. 1). The southwestern corner of our study area (west and north of Garoua, mostly transect A, Fig. 1) was the northernmost extension of the rather dense, now fragmented, Sudanian woodlands. Elsewhere, the first three protected areas surveyed (see below) were the only significant, forest-like patches within a mostly cultivated landscape. Cultivated areas are now a dominant feature of most Figure 1. Road counts and survey areas in northern Cameroon. Dashed lines are international borders, black dots are main towns and dotted areas are protected areas surveyed. Solid lines are roads where raptor counts have been done and letters along them refer to the transect identification (see Study Area and Appendix) , September 2001 Long-ierm Changes of Raptors in Cameroon 175 t N 'i I WDJAMENA ] NIGERIA ^ N r / C- J WaZidy-^. y h 5 p WAZA^.) ■p; s [j/I^OZOGO Mokoloj gokoro c I K GuirvidigPou^ * magaN? \\ ) ( CHAD \ (TQ O nYagoua \ KALFOU \ \ •N CHAD JM a y 0 K e b i iA 30km 176 Thiollay VoL. 35, No. 3 landscapes. Main crops are millet (rainy season) , cotton, sorghum (dry season in wet lowlands) and locally maize or onions (irrigated). Traditional fields are dotted with isolated trees (e.g., Adansonia, Bombax, Butyrospermum, Khaya, Acacia) , often regularly pruned for fodder. Trees tend to disappear from large cropland plantations (cot- ton, sorghum) . The original vegetation was dry wood- land, so-called savanna woodland because of its high grass cover, dominated by broadleaf trees in the Sudani- an zone and thorny Acacia or Balanites in the Sahel. Nat- ural woodlands are now much reduced, fragmented, and severely degraded by fires, cattle grazing, and wood cut- ting, even in protected areas. Additionally, hunting pres- sure has largely depleted, if not eliminated, large and medium size wild mammals, and game birds to a lesser extent (Ground Hornbill \_Bucorvus abyssinicus] , bustards [Eupodotis sp. and Neotis sp.], Guineafowl [Numida melea- gns]). Road Transects (see Fig. 1). (A) Garoua area. From the main town of Garoua (9°18'N-13°24'E), two tracts were censused, twice each, by vehicle and by foot: to the northwest (Gashiga-Demsa) a hilly area with cliffs, dis- turbed Sudanian woodlands and millet-cotton fields, and to the east (Pitoa-Mayo Kebi) , a plain whose former sea- sonally flooded zones are now mostly cultivated (sor- ghum) and where the last shallow wetlands are over- grazed by cattle. (B) Garoua-Maroua (10°36'N-14°20'E): flat or hilly degraded savanna woodlands, fragmented by many fields and villages. (G) Garoua-Rhumsiki— Mokolo. From the Benue Plain, the dirt road enters the Mandara Mountains up to 800- 900 m, through valleys and plateaus that are densely cul- tivated with few remaining patches of woodlands. Only the numerous Isoberlinia seedlings in fallow fields suggest that the area must have been covered with Sudanian dry forest. Between Rhumsiki and Mogode, the picturesque volcanic plateau landscape of Kapsiki is dominated by iso- lated rocky towers, but is otherwise densely cultivated. (D) Mokolo (10°45'N-13°48'E)-Koza-Mora (11° 03'N-14°09'E). From the northeastern rocky slopes and terraced fields of the Mandara Mountains, down to the Sahelian Plain of the lower Logone Catchment. The few non-cultivated areas are lightly wooded, grazed, and rep- resent degraded Acacia savanna (with Guiera, Commiphora or Salvadora bushes) . (E) Mora-Maroua-Lara-Yagoua (10°20'N-15°14'E) . Patches of northern Sudanian dry woodlands remain within a largely cultivated plain, dominated by occasional rocky outcrops (Mindif) . (F) Yagoua-Pouss (10°5rN-15°03'E)-Guirvidig. This area changed drastically between the two surveys. The track follows the River Logone along which former sea- sonally inundated grasslands (yaeres) are now densely populated and almost completely cultivated (rice) after an embankment had been built all along the river to pre- vent floods. A large dam has also turned the lowest part into Lake Maga. Palm stands {Borassus, Hyphaene) around villages are a typical feature of the landscape. (G) Maroua-Bogo-Guirvidig (10°53'N-14°50'E). Sah- elo-Sudanian savanna woodland, degraded, grazed, and now mostly cultivated. The first half was artificially refor- ested (exotic trees) in the 1980s. (H) Mora-Waza (11°25'N-13°34'E). Cultivated fields were still the dominant habitat along the first half of this road transect, these fields then gave way to the woodland of the Waza National Park fringes and finally to the tree- less inundation zone, south-west of the village of Waza, where an additional foot sample transect was done. Pools along the paved road probably attracted some raptors. (I) Waza National Park. This 170 000 ha protected area was formerly one of the richest in West Africa for large mammals, and also for raptors (Thiollay 1978a). The building of a dam and dikes, to canalize the Logone floods, dramatically changed the vegetation of the for- mer grasslands (yaeres) in the early 1980s, resulting in a crash of the ungulate populations. Later on, the seasonal flooding was partly restored, thereby improving the eco- logical situation. However, the adjacent Acacia seyalviood- lands and little flooded plains were overgrown by an un- palatable spiny annual plant (Monecma ciliata) that further impoverished both mammal and bird popula- tions. Following good rains, pools were still numerous in February 2000 all over the park, contrary to the drought that affected the area in 1973. Protected Areas. Mayo Louti Forest Reserve (10°42'N- 13°44'E), between Mogode and Mokolo, is the last for- ested remnant of significant size (about 3000 ha) on the Kapsiki Plateau (800-850 m) in the Mandara Mountains It is a dry Sudanian broad leaved woodland (including Isoberlinia, Piliostigma, Parkia, Combretum, Cussonia) with few Acacia (but many Balanites). The area is entirely grazed by cattle and regularly burned. Rocky outcrops and a sandy riverbed lined with large trees {Khaya, Bom- bax) provide additional habitat diversity and nest sites for eagles. During the survey (20— 22 January 2000) 112 non- raptor bird species were identified. The 1400 ha Mozogo-Gokoro National Park (10°58'N- 13°55'E), between Koza and Mora, is a dense sahelian forest isolated within largely-cultivated areas of the dense- ly-inhabited northern plain (^ 400 m). It is crossed by the dry sandy bed of a river. Acacia, Erythrina, and Bal- anites are dominant trees with a few scattered sudanian trees. Temporary pools become grassy sites in the dry season. Poaching and woodcutting probably have nega- tive impacts. During the survey (23-25 January 2000), 94 non-raptor bird species were also found. Kalfou-Yagoua forest reserve (10°20'N-15°14'E) is a 4000 ha disturbed sudano-sahelian woodland in the low plain of the Logone (300 m), and is also completely sur- rounded by cultivated areas and severely affected by woodcutting and cattle grazing throughout. Numerous flowering Erythrina trees and grassy dry pools ringed by Acacia seyal woodlands were typical features of the land- scape. During the survey (29-31 January 2000), 104 non- raptor bird species were recorded. The Lake Maga area (10°50'N-15°00'E), formerly a grassy inundation zone, has been turned, since 1979-80, into a large artificial lake providing water for 12 000 ha of irrigated rice fields where reed beds ( Typha) and bush- es {Acacia, Mitragyna) along ditches are the only wild veg- etation, together with some grassy fallow areas and palms around villages. Most fields were dry and harvested in February. The lake itself (13 000 ha) has a continuous belt of Typha or locally Acacia, except along the 22 km earth dam wall. September 2001 Long-term Changes of Raptors in Cameroon 177 Methods Daily raptor counts were performed along the main dirt and tarred roads of northern Cameroon from 1-20 February 1973 and from 17 January-6 February 2000. Additional counts in April 1973 and September 1975 were only referred to for the presence of additional wet- season migrants. In 2000, four specific areas, not previ- ously investigated, were surveyed because they had been tentatively identihed as potential Important Bird Areas to be included in the African network being designed by Birdlife International (M. Languy pers. comm.). All counts took place in the middle of the dry season, which extends from October to May. At that time, the level of the water courses was already low, pools were beginning to dry, most trees were losing their leaves, many Acacia trees were flowering, the grass cover was at its minimum, millet fields were bare, whereas cotton was being har- vested and sorghum was still green and ripening. How- ever, the two years were different: 1973 followed one of the lowest rainy seasons on record, while 2000 came after an above-average rainy season. So, there was more green vegetation, more water available (especially pools) and probably more insects and rodents in 2000 than in 1973. Roadside Counts. Counting raptors from a vehicle along roads is a widespread method (Johnson 1978, Ful- ler and Mosher 1981, Bibby et al. 1992). It provides an abundance index that is sensitive to the behavior of the species involved, the habitat, speed of the vehicle, mete- orological conditions, hour, season, and number and ex- perience of observers. It allows one to cover large areas and to record a significant number of individuals, but it affords little comparability between species where their detectability may be different. Intraspecific comparisons, however, can be made between years because of the sim- ilarity of observer, season, and method. Local differences m habitats between the two periods (e.g., more cultiva- tion) were in fact a critical component of the changes to be assessed. All individuals seen perched or flying within <1 km on either side of the road were counted. Two experienced observers, always including the author, looked carefully for raptors when driving slowly (<50 km/h on average) and during frequent stops, from after sunrise to before sunset, avoiding cloudy, windy, or cool weather. The time spent on any transect and year was always ^3 hr/100 km. Binoculars (10 X 42) were used during stops to search for birds systematically. The length of each transect was measured by the vehicle odometer. Village crossings were included, but not urban areas >1 km wide. All abun- dances were expressed in birds/100 km. Several transects (A, G, H) or part of them (D, I) were each censused twice in both years, on different days and/ or hours, using the maximum number of birds of each species in either count. The minor and unavoidable sources of variation between years that may have been due to differences in conditions (hours, weather, time spent) on some road segments were minimized, counterbalanced between ar- eas and averaged over such long distances. Therefore, I was reasonably confident that overall numbers of raptors recorded were representative and comparable between years. Local Surveys. The censuses performed in 4 additional areas in 2000 were used to assess the local abundance of raptors in particular habitats, but not for comparisons with any roadside counts. They cannot be used as density estimates because areas were unequally surveyed and in- dividual raptors also ranged outside the limits of these areas. Three protected forest reserves (see study areas) were surveyed by foot, during 3 days each. Slow, full-day walks covered all their main areas with multiple stops m all large gaps and vantage points to search for flying birds. The time was too short to detect every pair of rap- tors, let alone to estimate actual densities. So, the abun- dance index used (Table 2) was the maximum number of most likely different individuals of each species seen within any single day. The last area. Lake Maga, included the reservoir, its reed belt, and associated large irrigated fields. The lake itself was censused by boat, moving along the swampy shores and by walking and searching around three bays of the lake. Careful searches of raptors were also done along >50 kilometers of line transects by vehicle, at slow speeds, on dikes and along ditches through two different areas of the rice fields, with frequent stops. Because hab- itats and surveys were distinct, I summed the minimum number of different individuals of each species seen m each zone. A few double counts could not be ruled out, but, on the other hand, many birds were probably missed. As a result, the figures obtained are likely to be conservative estimates of the overall number of migrants wintering in the region. Data Analysis. To allow for comparison between tran- sects, the number of birds recorded was expressed as the mean per 100 km (Table 1). Even when it was counted twice, each yearly transect gave a single value. Thus, there was no possibility to assess statistically the between-year difference within each transect separately. Therefore, I used for each species a nonparametric Wilcoxon signed- ranks test over all matched pairs of counts to assess the global between-year difference expressed in the last col- umn of Table 1 by the sum of individuals seen over all transects (Siegel and Castellan 1988). Actual numbers that were recorded on each transect were used for com- puting the Wilcoxon test except for transect I, where mean values/ 100 km were used because of unequal tran- sect length between years. Resuits Forty-one species of diurnal raptors were iden- tified during these counts (Table 2 and Appendix) , but only seven of them were seen, at least one year, on all nine transects and 16 other species on 5—8 transects. Several species were also recorded out- side the actual counts. The Eurasian Honey-buz- zard {Pernis apivorus) and Hobby {Falco subbuteo) were only Palearctic passage migrants observed in April-September 1973-75. Some Lesser Spotted Eagles {Aquila pomarina) in 1973, were misidenti- fied (from photographs reexamined) and none were recognized in recent surveys. One Levant Sparrowhawk {Accipiter brevipes) was tentatively identified at Waza and 2 Ovampo Sparrowhawks {Accipiter ovampensis) near Yagoua in 2000. The 178 Thiollay VoL. 35, No. 3 Table 1. General trends in raptor populations, from similar dry-season roadside counts in northern Cameroon conducted at a 27-yr interval. See Appendix for specific data. No. OF Individuals Recorded Increasinsr No. OF Species Decreasing (=^) Not Significant 1973 2000 O (*) Vultures (resident) 1300 433 0 3 3 Other resident species^ 526 323 1 6 11 Afrotropical migrants 1562 981 0 1 4 Palearctic migrants’^ 571 340 1 4 6 ® From large eagles to small falcons, including Red-necked Buzzard. ’’ No Black Kite, Egyptian Vulture, or Peregrine Falcon was included. Also includes Lesser Kestrel, together with Common Kestrel. * at < 0.10 level. Beaudouin’s Snake-Eagle {Circaetus beaudouini), an African wet season migrant, was positively identi- fied in the Waza grasslands only in April 1973. The Western Banded Snake-eagle ( Circaetus cinerascens ) , a riparian forest specialist, was seen on a nearby transect in 1973 and possibly in Mozogo forest re- serve in 2000. Other African migrants (Shikra, Grasshopper, and Red-necked Buzzards) were more abundant in the 1973-75 wet seasons than they were in February of either year because most of their populations spend the dry season at more southern latitudes (Thiollay 1978b). Species were grouped into categories according to their taxonomy and the similarity of their pop- ulation trends between years as suggested by com- parative road counts (See Appendix and Table 1). All together, vultures declined by 67% over all counts. The three uncommon species, Egyptian, Lappet-faced, and White-headed Vultures were too rare and local for any significant trend to stand out. Moreover, Egyptian Vultures may have includ- ed unknown proportions of both local breeding birds and European migrants. Conversely, the mostly urban Hooded Vulture and the widespread White-backed Vulture and Ruppell’s Griffon de- clined, respectively, by 67, 60, and 87%. This was further confirmed by the lack, or small number, of vultures seen around carcasses occasionally en- countered. An even more striking evidence was the much lower number of Hooded Vultures recorded during stops in villages and towns (outside road counts) between 1973 and 2000, often by an order of magnitude. In the Waza National Park, a strong- hold of vultures, but where ungulates decreased dramatically after 1980, I counted 141 vultures on 158 km (89/100 km) in late January 2000, com- pared to 395 on 240 km (165/100 km) in early February 1973 and 1257 on 334 km (376/100 km) in April 1973 during a severe drought (i.e., abun- dance of carcasses). The seven large resident ea- gles (Fish-, Brown Snake-, Bateleur, Tawny, Hawk-, Martial and Long-crested Eagles) all declined sig- nificantly (83%, P = 0.025) . Only the smaller Wahl- berg’s Eagle remained stable. For 9 of the 11 smaller African resident species, there was no significant change between years, or at least their sample size was too small for any con- clusive evidence: Lizard Buzzard (here on the northern edge of its distribution), Harrier-hawk (also at its northern dry-season limit) , Gabar Gos- hawk (secretive, but still fairly common), Red- necked Buzzard (a migrant, rare elsewhere in the Sudanian zone during the dry season, Thiollay 1978b), Fox Kestrel (highest known population is in the Mandara Mountains, Thiollay 1977), Grey Kestrel (inconspicuous, local species). Red-necked and Lanner Falcons (bird predators often associ- ated with palms and cliffs respectively as nesting sites) and Peregrine Falcons (including both the local breeding, F. p. minor, and the Palearctic win- tering, F. p. calidus, subspecies) . The three-fold overall increase of Black-shoul- dered Kites in 2000 may have been explained by an increased area under intensive cultivation, to which the species is well adapted, but also, per- haps, by an increasing availability of rodents, fol- lowing a much better rainy season than in 1973. The significant decrease (62%) of the Dark Chant- ing Goshawk, a generalist predator of small terres- trial vertebrates, may be an indicator of the general prey impoverishment of heavily exploited and de- graded natural habitats. September 2001 Long-term Changes of Raptors in Cameroon 179 Table 2. Raptor surveys of specific areas in 2000. The first three columns are forest reserve where the maximum number of different individuals seen within any full-day spent walking through the reserve is given. Maga includes both the lake and the extended rice fields to the north. Here, the total number of birds seen in the whole area is given (see Methods). Mayo Louti Mozogo Gokoro Kalfou Yagoua Lake Maga Osprey** 5 Black-shouldered Kite 1 2 21 African Swallow-tailed Kite* 1 4 12 Black Kite* 2 4 150 African Fish Eagle 1 Hooded Vulture 1 1 1 67 African White-backed Vulture 26 Short-toed Snake-eagle** 1 1 Brown Snake-eagle 2 1 2 1 Bateleur 1 Western Marsh Harrier** 258 Pallid Harrier** 3 Montagu’s Harrier** 1 2 37 African Harrier-hawk 1 Lizard Buzzard 3 Dark Chanting Goshawk 3 1 6 11 Gabar Goshawk 1 2 5 3 Shikra* 2 3 Grasshopper Buzzard* 1 8 3 Long-legged Buzzard** 3 Red-necked Buzzard* 1 Greater Spotted Eagle** 2 Tawny Eagle 1 Steppe Eagle** 4 Wahlberg’s Eagle* 1 2 2 1 African Hawk-Eagle 2 Booted Eagle** 1 1 9 Martial Eagle 1 Long-crested Eagle 1 Common Kestrel** 1 2 1 Lesser Kestrel** 1 Fox Kestrel 1 Grey Kestrel 2 1 Red-necked Falcon 2 2 5 Lanner Falcon 1 2 2 1 * African migrants. ** Palearctic migrants. The main African migrants either decreased (Black Kite and Swallow-tailed Kite) or increased (Shikra, Grasshopper Buzzard) by about 40-50%, but the trend was significant only for the Black Kite. The differences in rainfall between years may have resulted in different proportions of the pop- ulations staying in February at this northern edge of their dry-season range (Thiollay 1978b). Such shifts may obscure any change in population size. Moreover, most Black Kites were of the African subspecies (M. m. parasitus), but a small, albeit un- known, proportion of them were Palearctic winter- ing birds, M. m. migrans (at least 7% in 1973). Among the Palearctic migrants, the Short-toed Snake-eagle and the Steppe Eagles decreased sig- nificantly (84%, Table 1), and probably also the 180 Thiollay Voi. 35, No. 3 Greater Spotted Eagle and the Long-legged Buz- zard, whose numbers were too low for the trend to be signihcant. The Common + Lesser Kestrel com- plex, formerly widespread and easy to see, declined even more strongly (92%). The actual proportion of Lesser Kestrels and of African residents vs. Pa- learctic migrants among Common Kestrels could not be determined, for a lack of time to look care- fully at every individual. In 1973, a sample of 80 birds well identified included at least IS K nau- manni, A F. t. rufescens, and 58 F. t. tinnunculm. From occasional sightings in 2000, there was no evidence that the locally breeding rufescens (always uncommon) had declined, nor that the proportion of Lesser Kestrels had changed. So, most of the decline was ascribable to the European birds. The Saker Falcon is too rare a migrant to show a sig- nificant trend. Surprisingly, the Booted Eagle increased strik- ingly (314%), although numbers were low. The harriers altogether increased by 69%, but, species by species, not significantly {P = 0.182). These win- tering birds tended to use specific habitats: western Marsh Harrier in wetlands, Montagu’s Harrier in humid lowlands and plateaus, Pallid Harrier in dri- er grasslands, and Booted Eagle often in hilly and rocky, cultivated, or open woodlands. Yet, they all may have benefited from the consequences of bet- ter rains in 2000 than in 1973, on prey availability (large insects and small vertebrates). This would further emphasize the decline of Common and/ or Lesser Kestrels, which should have also been fa- vored by these improved conditions. Therefore, the massive population decrease is more likely to have occurred on the kestrels’ breeding grounds. Other Surveys. The three forest reserves studied (Mayo Louti, Mozogo, and Kalfou; Table 2) are the only protected areas between the latitudes of Gar- oua and Waza (225 km) and the last significant patches of woodlands in that part of northern Cameroon. So, they should be representative of a formerly widespread raptor community. They in- cluded not only forest, but patches of open wood- lands, grasslands, seasonal wetlands, and some cul- tivated areas. In these areas, however, only 27 raptor species were recorded (64% of the regional pool). Only six of them were seen in all three reserves and 17 were found in a single site. The species abundance distribution matches that of roadside counts. Es- pecially striking was the lack of vultures, except an occasional Hooded Vulture coming from a nearby village, and also the rarity of eagles, except the Brown Snake-eagle and the smaller Wahlherg’s and Booted Eagles. Both vultures and eagles are con- spicuous birds that could not be missed. The low number of kites, harriers, and kestrels may be due to unsuitable habitats for species mostly associated with fields or large grasslands. The abundance of Chanting, and Gabar Goshawks, Shikras, Red- necked, and Lanner Falcons was indicative of still healthy populations, as already suggested by road counts. It was interesting to see how raptors were using the lake Maga area, a new, artificial wetland, inten- sively cultivated. During a survey of about two- thirds of the lakeshores, river banks, villages and irrigated fields, at least 625 individuals of 23 spe- cies were counted, including eight species not re- corded in any forest reserve. However, 61% of the species and 78% of all individuals were either Eu- ropean (10) or African (4) migrants, 41% of which were Marsh Harriers. In 1973, when the area was still covered with seasonally inundated grasslands, I did only a linear transect through the area and not a census. Yet, along this transect (F in Appen- dix) , the number of species recorded in 2000 had decreased by 38% and the total number of raptors by 65%. This supported the impression that the new habitat was decidedly poorer and less suitable for many species than the former natural grass- lands. While harriers and a few uncommon species remained stable, the vulture numbers declined from 360 to 101, large eagles from 130 to 8, kes- trels from 131 to 0 and even kites (except the black-shouldered) from 765 to 291. Although ex- ternal factors may be partly responsible, habitat degradation must be involved in these dramatic de- clines, through prey reduction, either large insects (kestrels, kites) , large mammals (vultures) or small- er vertebrates (eagles) . A lack of large trees around the lake may explain the rarity of the Fish Eagle, in spite of the introduction of fishes (Tilapid). Discussion We must be cautious in interpreting such counts in terms of population dynamics. Roadside count data are highly sensitive not only to weather, visi- bility, hour and season, but also to number, expe- rience or attentiveness of observers, and to speed, type of vehicle, and frequency of stops. However, in several respects, the counts of 2000 were done under better conditions than in 1973, which may suggest the magnitude of the decline as more se- September 2001 Lonctterm Ceianges of Raptors in Cameroon 181 vere than indicated by the data. Additionally, in 1973, because of the drought, both Afrotropical and Palearctic migrants may have shifted their dry season ranges southward (Thiollay 1977, 1978a) and the breeding success of some resident species may have been reduced, except for the vultures. Therefore, the density of a number of taxa in the study area was likely to be lower than usual in 1973. It is delicate to infer a general trend from only two points along such a time period without inter- mediate data. Nevertheless, most changes in raptor populations recorded in 2000 were consistent with what was expected from the general modification of ecosystems during the past decades. Even if many specific results were not significant because of low sample size and variability between transects, the overall picture (Table 1) is that of a net de- crease in raw numbers in all four broad categories, with only two species increasing, against 14 species decreasing and 24 species apparently stable. Habitat Changes. There are no accurate regional data about the modification of vegetation cover or areas under cultivation. From the comparison of my descriptions of habitats along road transects, from the huge increase in human population size, and from the observed current pressure on natural resources, trends toward ecosystem impoverish- ment were all too obvious. Not only has much of the forest been cleared during recent decades, but most remaining woodlands were heavily degraded by widespread firewood cutting and charcoal pro- duction. The area under cultivation had increased locally, but also there were more areas, previously under a system of long fallows, that were now un- der permanent cultivation. Sorghum planting on wetlands had also expanded dramatically at the ex- pense of former grasslands. However, the most drastic change was the damming of the Logone River and the conversion of >1000 km^ of grass- lands, teeming with wildlife, into rice fields, dense- ly populated by humans, and a large reservoir. Meanwhile, cattle have not noticeably changed and overgrazing of woodlands and savannas was appar- ent almost everywhere. Moreover, the last suppos- edly protected forests were heavily exploited for firewood and hunted without effective control. In the countryside, even small rodents and birds are routinely killed for food, and hunting may have played a role in the disappearance of larger game species. Among other possible factors, pesticide use, that was extremely limited 30 years ago, is now very widespread and sometimes massive, especially m cotton fields and against locusts. There are no data documenting the impact of contaminants on non- target species. Locust control operations also occur on a large scale in years when locust density is high and they have probably reduced the average max- imum abundance of this critical and seasonal source of food. Resident Species. Increasing human pressure can affect predators through prey and nest site availability and disturbance. Only small raptors adapted to cultivated and human inhabited areas were found stable (e.g., Gabar, Shikra, resident kes- trels, Tanner) or even increasing (Black-shoul- dered Kite) . These were rodent, bird, or lizard spe- cialists, favored by clearings, open ground, and croplands. Larger, albeit generalist species (e.g., Black Kite, Chanting Goshawk) were significantly decreasing but the underlying factors were unclear. The consistent decline of all resident eagles may be explained by the probable decrease of their food supply and of large nest trees on over- grazed and cleared natural habitats, as well as by increased disturbance. Forest reserves are too small and scattered to allow, by themselves, the maintenance of viable eagle populations. Yet, a majority of eagles were recorded in or around such reserves. The long-term fate of these eagle species is uncertain. Vulture Decline. The significant decrease of vul- ture populations has no obvious explanation. Vul- tures are traditionally not molested by humans. Cattle, goats, sheep, and donkeys were still abun- dant in 2000. Methods of cattle raising and season- al movements had changed little and poisoning carcasses for predator control was reported to be rare. Urban sanitation, however, may have in- creased a little, but refuse and open-air slaughter- houses still were plentiful and food availability alone could not explain why many Hooded Vul- tures had vanished from towns. The dramatic de- cline of wild ungulates in the Waza National Park area may have affected the number of vultures lo- cally, but not outside where domestic ungulates had long been the main source of carcasses. How- ever, the much better rainy season, prior to the year 2000 counts, than in 1973 may have not fa- vored the vultures as it probably did for predators. The spectacular collapse of Southeast Asian, and more recently, Indian vultures, is suggestive of a widespread disease (Rahmani 1998, Watson 2000). In this context, I report the following observation 182 Thiollay VoL. 35, No. 3 During the two days in the Waza National Park, a former stronghold of vultures, 93% of the vultures recorded were seen sitting near a single pool where there was no carcass. Vultures, all six species together, with empty crops, were gathered on the ground, in spite of an abundance of trees around where a few birds perched. They were not seen flying, even in the afternoon. They were immobile, often crouched, reluctant to move ahead of the vehicle, and then just flying a short distance to sit again on the ground. Some of them were drinking in the afternoon, but none were seen bathing. The weather was clear and moderately hot and windy, i.e., good flight conditions. Such an unusual gath- ering and behavior were not observed elsewhere. Palearctic and African Migrants. Large Palearctic eagles (Snake-, Steppe or Spotted) were too rare or local to convincingly indicate a decline. How- ever, this trend was consistent with that of their African counterparts and with the known change in their status in Europe (Tucker and Heath 1994) . Conversely, the stability (if not increase) of the small African (Wahlberg’s) and European (Boot- ed) migrant eagles was well established. Locusts (and orthopteras in general) are known to be a major source of food for many mi- grant raptors and to be fairly reliable indicators of the abundance of other prey (large insects, small vertebrates) of these raptors (Thiollay 1978b). The average level of locust populations probably decreased both because of habitat changes and of recurrent chemical control oper- ations (use of Dieldrin and organophosphates) . Nevertheless, among species mostly dependent on this prey base, the Eurasian harriers (notably Montagu’s) and small African migrants (especially Grasshopper Buzzard) did not decrease apprecia- bly during recent decades, whereas the Common and Lesser Kestrels suffered a very significant de- cline. Therefore, this decline can be attributed primarily to factors affecting the population on their breeding grounds rather than to drought, habitat degradation, or pesticide use in the win- tering areas. The decline of Lesser Kestrel popu- lations has been widely documented, especially in Europe (Tucker and Heath 1994) and this species may have made a substantial proportion of the kestrels in the 1973 counts. Acknowledgments The second year counts were performed during a Bird- life International survey of Important Bird Areas in northern Cameroon, with the support, funding, and help of the Cameroon Ornithological Club, the local partner of Birdlife International. I warmly thank the following for their help and participation: Roger Fotso, Marc Languy, Serge Bobo Kadiri, Kevin Yana Njabo, Francis Njie, and also the driver Emmanuel Yamgo who patiently stopped and waited so often, and my wife Frangoise for her keen eyesight and typing. Comments of R. Davies, M. Louette, and R Mundy improved the first draft of this paper. Literature Cited Bibby, C.J., N.D. Burgess, and D.A Hill. 1992. Bird cen- sus techniques. Academic Press, London, U.K. Brown, L.H. 1952. On the biology of large birds of prey of Embu district, Kenya colony. Ibis 94:577-620. . 1953. On the biology of large birds of prey of Embu district, Kenya colony. Ibis 95:74-114. . 1955. On the biology of large birds of prey of Embu district, Kenya colony. Ibis 97:38-64, 183-221 Encyclopaedia Britannica. 1998. 1998 Britannica book of the year. Encyclopaedia Britannica Inc., Chicago, IL U.S.A. Encyclopaedia Universalis. 1974. Vol. 3. Encyclopaedia Universalis France, Paris, France. Fuller, M.R. and J.A. Mosher. 1981. Methods of detect- ing and counting raptors; a review. Pages 235-246 in C.J. Ralph andJ.M. Scott [Eds.], Estimating numbers of terrestrial birds. Studies in Avian Biology. Vol. 6 Cooper Ornithol. Soc. Gargett, V. 1990. The Black Eagle. Acorn Books, Rand- burg, South Africa. Igemonger, S., C. Ravilious, and T. Quinton. [Eds.]. 1997. A global view of forest conservation. CD-ROM. V7CMC and CIFOR, Cambridge, U.K. Johnson, D.R. 1978. The study of raptor populations. University Press of Idaho, Moscow, ID U.S.A. Letouzey, R. 1968. Etude phytogeographique du Came- roun. Encyclopedic Biologique 69. Lechevalier, Pans. Louette, M. 1981. The birds of Cameroon. An annotat- ed checklist. Academic Wetenschappen letteren en Schone Kunsten, Brussels, Belgium. Newton, L 1979. Population ecology of raptors. T. & A.D Poyser, Berkhamsted, U.K. Rahmani, a. 1998. A possible decline of vultures in India. Bull. Orient. Bird Club 28:40—42. Siegel, S. and N.J. Castellan. 1988. Non parametric sta- tistics for the behavioral sciences, 2nd Ed. McGraw- Hill Book Company, New York, NY U.S.A. Thiollay, J.M. 1977. Distribution saisonniere desrapaces diurnes en Afrique Occidentale. Oiseau R.F.O. 47:253- 294. . 1978a. Les plaines du Nord Cameroun, centre d’hivernage de rapaces palearctiques. Atowrfa 46:319- 326. . 1978b. Les migrations de rapaces en Afrique Oc- cidentale: adaptations ecologiques aux fluctuations September 2001 Long-term Changes of Raptors in Cameroon 183 saisonnieres de production des ecosystemes. Terre Vie 32:89-133. . 1998. Long-term dynamics of a tropical savanna bird community. Biodiversity Conserv. 7:1291-1312. Tucker, G.M. and M.F. Heath. 1994. Birds in Europe: their conservation status. Birdlife International, Cam- bridge, U.K. Watson, R. 2000. Vultures in crisis. Peregrine Fund News- letter 31:20-21. W.C.M.C. 1994. Biodiversity data source book. World Conservation Monitoring Centre, Cambridge, U.K. Received 29 July 2000; accepted 3 January 2001 Appendix. Comparative dry-season, roadside counts of diurnal raptors in northern Cameroon m 1973 (first value) and 2000 (second value). Mean number of individuals recorded per 100 km from transects A to I. 184 Thiollay VoL. 35, No. 3 Q g u iO 00 r- to 00 1> u f-H O c/5 o 1-H CM c/5 C/5 r- CM CO C/5 c/5 CO PL| o o ^ 'Z o o o 2 o o O 2 2 2 2 d d d d d d d d d C' 05 05 cO 05 CM 1 o 1-H CM o 00 1-H o ■^H t-H GO g 00 Tfl 1 1 CO 1 CM 1 t-H 1 GO 1 CM 1 lO 1 1-H 1 iTi 1 ^H 1 ^H 1 t-H 1 t-H 1 1 H o 1 1-H r-( CM 1 1 00 00 1 1 to 1 CM 1 o 1 GO 1 CM 1 GO 1 O 1 m 1 50 1 to 1 50 1 GO H i> 00 J> (05 GO to (M (05 e- ^H to 1 — 1 CO to CM CM 1-H q 00 lO CO q q q T ( q q CO 1-H 00 q d CM CM q q O d 05 q CM to 1 — 1 I-H d CM CM d GO t-H t-H GO d d t-H t-H rH d \ \ \ \ \ \ \ \ \ \ o CSf CM lO GO t-H CM 00 q t-H CM ^H I-H 05 00 ^H q CM CM d CM d d t-H CM d CM CM 1> d CM d d to 1-H lO e- ^H t-H CM ^H CO GO ^H q o d (d P q q q (05 o o O GO CM 68 CO 1-H to o CM d t-H CM d d d t-H d 00 \ \ \ \ q ^H 1> q q q q q q Th i> CM CM d d d CM d CM d CM CM 00 r-H r~ 05 I-H t-H G] CO CM q GNT t-H d d 00 q o o 00 I-H 1-H CM CM d d d d d CM \ \ kD lO o 05 o 00 q q ?-H I-H d d 00 d d 1-H T—i to q 1-.^ lO q CM q CNT o d GO o O O to to o d q t-H o 00 00 O CM d d d d d d GO ^H CM d UO \ \ \ \ \ ■\ \ \ I-H a\ lO i> q q 'sf to GO GO 00 q oq 00 GO 1-H 1> d CM d d d 7—i t-H d d CM t-H d CM t-H 00 CM CM I-H 1-H CM Th t-H o q q cq o 05 iC> q q q iO CM 1-H CM d d CM d t-H d CM o \ CM o q q o 05 q o q o q 1-H d d d d d d CM d GO GO i-H \ GO GO O d GO d o d q d 00 GO e~ d q d t-H CM GO C/3 2 : GO I CM q d q d q q 50 q q o 1-H t-H I-H t-H CM d d d d \ \ CM q ^H q q 00 GO q q q 1-H 00 1-H to d d 1-H ^H T^ t-H CM CM q q o q ^H o o o O nH d GO CM d d d 05 \ \ \ \ t-H 1-H to 50 o 00 q q I-H CM d d d d d GO GO CO q q 00 q q CM o q t-H q q q Oi 1-H 1-H CM GO d I-H t-H 1-H GO CM CM o \ \ \ \ \ \ \ r — ( in q o 50 q q 00 q q q q CM 1-H d d CM d 1-H GO September 2001 Long-term Changes of Raptors in Cameroon 185 C o CM o m I GO C/D cn 1-H ,J. 1 c/!} (-n o ^ O « ^ -S •x3 e fl cn o OD I> Oi I I ^ CM m (VD 2: Tfi I cn I cn CM 4. i> CM CM O O ' — I o d cn 2 CM m I 00 m CO I t/D O I CM CM 1 — I O d o> CM 4 - % s -T3 U !-i a a 1 -Id cn O o Td c5 N N ^ d3 fe P 3 ~e <3 ^ Isii O ts ■J3 §0 ^ -f s ^ ^ p ^ o G C3 J=! ■ ^ in % ^ s S ^ PQ 2 rt )h d rt rCd *3 C/D O ;s 03 T3 Li * S a. N % ^ S CQ ill bO ^ ^ ^ % V ds Ed S _L cq cJ P< ci "O i U ^ s * t, ■2 53 S s lli ■«5, o d CM I CM OJ =n !0 .8 i ■§■ .. S bc „ ... „ ■S c 3 -S c 3 ■+S ^ P r.i ^ S T5 S ID J> O o d q q r-d q o q q q o CO o o q 00 I — < 1 — ^ d d d d hH d rH d d CO \ \ \ \ \ \ \ q q □0 q o q 73^ CO o in rL rH d d CO in CM d d d d d t-H CM q C75 00 o o o CT! cn o q J> CM CO d d d in CM d in \ \ \ 1— 1 CO q C) q q q q 7t< d CO CM J> 73* rH d in 00 nL I-H m q q q q o o q 1 — 1 CM d rH d d CM \ \ \ 00 q o 00 00 q d d d d d cd o CO q q CO o m CD j> o o l-H I-H t-H d CM d in d d ■\ \ \ \ q GO CO 00 GO I-H CO q q q OCD T— 1 d CO T— 1 in Tfi I-H 1 — 1 in CM nH 00 o in CT) o m q q d d CM d d d t-H \ \ \ \ q q q m o m q d d CM d d d in q q q q d CO CO rH t-H \ \ \ \ q q q l> j> nH I-H CO T ^ T— H rH m 7^^ |> q q q q q q q CM d CM d CM nH d d rH d CM \ \ q o q q q q o in m in 00 in d d CO CM hH d d d d d r-H m o I-H m o d d hH d d q m q m q CO d in d CM 'Ch q o q o o q q q d I^ I-H d I— H d d i-H T 1 t-H \ \ \ \ \ \ q q q q C] q q q o q jq in t-H O 1-H !-H I-H rH d rH o6 C/D C/D 2 2 qj p 7^ TC 2 5 P 5 Td ® « -2 o o p3 o d \ CO o o d d O) Ol CM CM o d \ <£) d CO n- o d m LO j> d iri q d td 1/3 * -2 8 s s s s • ?<• 00 q lO C/D 2 C 3 •0 N,> s o 5 h U if) --SS ■§^ o a e o ’“-M CIh d 0 R O fen h 4 U o to OJ Eh M O d \ CO o d q CM q CM CO C/D 2 X dS o q t-H q I — ( q CM CO d \ CO q CO j> q CM \ CO Ed o - S ,55S S t>H 3 5d Eh s ^ IS HH /) ?3 ‘y O «o u (IJ C) rH ^ ,53 MH r.^ C/D 2 CO c\r 1 44-5 9-2 1 rH 1 CO 1 1 cn 00 CM 1 I> 1 iO rH 1 CO q r-H \ d m !-H m o o q d S Pd Ed o u 'n tL Eh tU Ed Ed C3 E-! 8 ' o 4 S S E. o .« 5J O VJ , Co q u fen -X fen C3 in * •s 186 Thiollay u 3 a ‘■3 3 O a >< • ^ T3 3 V a Q. << 05 00 CJ> GC 00 Plh 2 o o d d j> Xl i> U 03 o ? 0 1 CO 1 CM 1 1 i> 1 CM 1 03 H 3^ m 03 00 00 m CO 00 t-M 03 »— < CM HH \ \ \ o CO in 00 t3 CO 1> CM in O) C-H 1> ffi 00 ,-H CM t£) o O o CM CM -cjd O 00 0 t-H 00 CM \ \ 1— M m o t-H 00 1-M o o m -c3 00 d CM cO U-^ m \ \ dO CM CO d CO J> O) in in CD m d t-M CM w o \ \ CM in O d 00 o 03 Q o rH 00 t3 O CM nH dO CO dO O 1— H CM 00 u \ CM Od CM nM i-H CM CO T ^ o> 03 O CO CM O) ■\ \. CM in rM 00 i> O CM << CM nH o \ \ I-M CO CM CO ^M s " c« H 3 U W c« X H * % S a H O z: w 3 0 C- 3 33 o o u S G \ 13 3 c« r 3 13 3 3 ^3 s c/3 u * rH T3 ;s T3 3 d CiO 3 8 e u u &, ;-! 3 - u V •3 ^ 3 ■Sf S •S g 3 3 I .a 0 -o fG fl C« .S G O ■9 ^ 3 OJ r> o S ^ 3 w ri U ^ ^ rs c« CTS g ■ < "3 ^ I i I 'G S C c3 S gn 8 I- r'3 3 tti w crt 3 u H 3 OJ dj 3 o cn d 4.0 0 0.0 3 21.4 Total 25 14 D2 <-2.0 7 25.9 0 0.0 ]-2.0, -1.0] 7 25.9 0 0.0 ]-1.0, 0.0] 11 40.7 0 0.0 ]0.0, 1.0] 1 3.7 2 13.3 ]1.0, 2.0] 1 3.7 6 40.0 ]2.0, 3.0] 0 0.0 3 20.0 >3.0 0 0.0 4 26.7 Total 27 15 two individuals to the correct sex (overall success = 96%), where values of > 0 identified males and values < 0 identified females (Table 2). In some circumstances data on body mass may not be available (e.g., carcasses); therefore, we repeated the stepwise discriminant analysis excluding this variable. The resulting discriminant function (D 2 = 25.624 - 1.072HCL - 0.239DT + 0.295age) in- cluded claw length, tarsus diameter (DT) , and age and it classified all but one bird correctly in both cross- and external-validation (overall success = 96%). The frequency distribution of discriminant scores indicated that in most cases males were well separated from females by the linear combinations of variables Dj, although not as well by Dg (Table 3) . The separation was much smaller when the ef- fects of age were not accounted for, particularly in the case of Dg. The single sample for which there was disagreement among molecular techniques was clearly classified as a female irrespective of the morphometric criterion used, thus supporting the results of Ml. Discussion Molecular Sexing. The lower success in sexing nestlings with M2 (one putative error and six blanks) than with Ml (one blank) could be par- tially attributed to the lower quality of the samples used, which had been frozen and thawed several times before DNA extraction and amplification. If good-quality samples were used, a higher success rate could have been achieved, probably similar to 192 Palma et al. VoL. 35, No. 3 that obtained by Ml. However, the difference in success rate between the two methods may also have resulted from the techniques themselves, thus indicating that Ml may be more robust than M2. For one sample, the M2 method apparently gave a false result (a male instead of a female pattern). One explanation for this could be that the M2 method is expected to allow the detection of single point mutations, while Ml can only detect differ- ences in length of about 5-10 bp between homol- ogous fragments. This difference in sensitivity to detect sequence variation could explain the diffi- culty in the interpretation of results and the dis- crepancy between methods, as M2 would produce an unknown migration profile for each variant, in- cluding potential false female or false male pat- terns. Therefore, M2 is apparently less robust than Ml because of the lower success rate of interpret- able migration profiles and probably is also less re- liable than Ml according to the false result ob- tained. An important factor is the amount of effort (quantity of products, money, and time) necessary for each method. The results can be obtained with- in 24 hr using M2 and 48 hr with Ml, and the cost of the products is higher for the latter method (e g., radioactive labeling, large sequencing gel, and autoradiography exposure). Thus, M2 is glob- ally quicker and less costly but also less efficient. This balance strengthens the need for a simple morphological way to determine the sex of Bo- nelli’s Eagle nestlings. Morphometries. Our results clearly demonstrat- ed a marked sexual size dimorphism for most ex- ternal body measurements in Bonelli’s Eagle nest- lings from age 35-50 d. The main exceptions were the lengths of the seventh primary and the central tail feather, which were remarkably similar between males and females at any given age. These two measurements have been used in the age estima- tion of nestlings (Mahosa et al. 1995) on the as- sumption that feather growth shows small varia- tions between sexes, as is typical of other raptor species (Poole 1989, Sodhi 1992). Although this issue was not addressed directly in this study, our results do support this assumption, and thus the aging method proposed by Mahosa et al. (1995). Body mass, either alone or combined with other variables, provided the most consistent cue for sex- mg the nestlings. However, the use of this param- eter should be regarded with some caution, for mass is highly variable, even within a 24-hr period, and depends on growth rate, degree of hydration. amount and time of the most recent meal, among other factors. Nevertheless, our results suggest that the differences in mass between sexes tend to pre- vail over the background of natural variability that may be present. There was a single individual that could not be classified by any discriminant func- tion including mass, and this was an extremely lightweighted female, with a severe infection caused by the protozoan parasite Trichomonas gal- linae (trichomoniasis). This disease induces large, fibrous lesions in the oesophagus and oropharynx preventing birds from swallowing food, and even- tually leading to death by starvation (Hofle et al. 2000). Care should thus be taken when sexing nestlings on body mass criteria, if signs of severe trichomoniasis and emaciation are apparent. This same bird could, however, be sexed with the linear discriminant function D 2 , suggesting that even dis- eased birds can be sexed on the basis of morpho- metric criteria. In these circumstances, however, molecular sexing techniques are likely to provide more reliable results. The estimated age of nestlings was included in both discriminant functions, though this variable did not improve the correct assignment of sex of the sampled individuals. However, when age was forced into the equations, the separation between groups along the discriminant axis was always in- creased (expressed by the Wilk’s lambda). Al- though the growth of most body structures levels off at about 35 d, they tend to continue growing at slow rates almost until fledging (Mahosa et al. 1995), making the difference between the two sex- es more evident when nestlings of the same esti- mated ages are compared. Therefore, by consid- ering the age of nestlings, we achieved more robust discriminant functions, providing more confidence to the classifications obtained with these models. In conclusion, the results of this study demon- strated that external morphometry may be used for the determination of sex in nestling Bonelli’s Eagles from 35-50 d. The discriminant measure- ments needed to use our method are easy to obtain in the field, allowing an immediate and about 96% accurate determination of sex. Difficulties may arise, however, in the case of undernourished or diseased birds, for which our equation should not be used. In general, we recommend that both discriminant equations should be computed for each bird to assess the internal coherence of the sexing results. Whenever possible and logistically feasible, molecular sexing should be used along September 2001 Sex Identification of Nestling Bonelli’s Eagles 193 with morphometries to reduce the overall error rates, particularly in those instances where it is es- sential to know, with absolute precision, the sex of every individual handled. Acknowledgments This work was supported partially by a grant from the Portuguese Science Foundation (PRAXIS/BTA/1 32/96). Nestlings were handled under permission from the Por- tuguese Nature Conservation Agency (Instituto da Con- servagao da Natureza). T. Guillemaud was the recipient of a postdoctoral fellowship (PRAXIS XXI/BDP/dd/O/ 96) from the Portuguese Science Foundation. P. Cardia was supported by an ICETA grant. We thank the critical reading of the manuscript by A. Helbig, F. Moreira, J. Real, E. Casado, and an anonymous referee. Literature Cited Cardia, R, B. Fraguas, M. Pais, T. Guillemaud, L. Pal- ma, M.L. Canceia, N. Ferrand, and M. Wink. 2000. Preliminary genetic analysis of some western Palaearc- tic populations of Bonelli’s Eagle, Hieraaetus fasciatus. Pages 845-851 in R.D. Chancellor and B.-U. Meyburg [Eds.], Raptors at risk. WWGBP/Hancock House, London, U.K. Cortes, O., A. Barroso, and S. Dunner. 1999. Avian sex- ing: an optimized protocol using polymerase chain re- action-single-strand conformation polymorphism. /. Vet Diagn. Invest 11:297-299. Clutton-Brock, T.H. 1986. Sex ratio variation in birds. Ms 128:317-329. Ellegren, H. and B. Shelton. 1997. New tools for sex identification and the study of sex allocation in birds. Trends Ecol. df Evol. 12:255-259. Ferrer, M. and C. de le Court. 1992. Sex identification in the Spanish Imperial Eagle./. Field Ornithol. 63:359- 364. Eridolfsson, A.K., H. Cheng, N.G. Copeland, N.A. Jen- kins, H.C. Liu, T. Raudsepp, T. Woodage, B. Chow- dhary, J. Halverson, and H. Ellegren. 1998. Evolu- tion of the avian sex chromosomes from an ancestral pair of autosomes. Proc. Natl. Acad. Sci. U.S.A. 95: 8147-8152. Grier, J.W. and R.W. Fyfe. 1987. Preventing research and management disturbance. Pages 173-182 in B.A. Gi- ron Pendleton, B.A. Millsap, K.W. Cline, and D.M. Bird [Eds.], Raptor management techniques manual. Sci. Tech. Ser. 10, Natl. Wildl. Fed., Washington, DC U.S.A. Griffiths, R. and B. Tiwari. 1995. Sex of the last Spix’s Macaw. Nature 375:454. , M.C. Double, K. Orr, and RJ. Dawson. 1998 A DNA test to sex most birds. Mol. Ecol. 7:1071-1075 Hofle, U., J.M. Blanco, L. Palma, and P. Melo. 2000. Trichomoniasis in Bonelli’s Eagle {Hieraaetus fasciatus) nestlings in southwest Portugal. Pages 45-52 in J T. Lumeij, J.D. Remple, P.T. Redig, M. Lierz, and J E Cooper [Eds.], Raptor Biomedicine III. Zoological Education Network, Inc., Lake Worth, FL U.S.A. Huberty, C.J. 1994. Applied discriminant analysis, Wiley, New York, NY U.S.A. King, J.R. and R. Griffiths. 1994. Sexual dimorphism of plumage and morphology in the Coal Tit Parus ater Bird Study 41:7—14. Manosa, S., j. Real, and J. Codina. 1995. Age estimation and growth patterns in nestling Bonelli’s Eagles J. Raptor. Res. 29:273-275. Morrison, J.L. and M. Maitbie. 1999. Methods for gen- der determination in Crested Caracaras./. Raptor Res 33:128-133. Palma, L. 1994. Nidificacion de aguilas perdiceras sobre arboles en Portugal, (/ucrem 98:11-12. Pareijada, X. 1984. Variacio del plomatge i identificacio de I’aliga cuabarrada {Hieraaetus fasciatus fasciatus). Pages 70-79 in O. Alamany, A. de Juan, X. Parellada, J. Ramon, and J. Tico [Eds.], Rapinyaires Mediterran- is II. Centre de Recerca i Protecio de Rapinyaires, Bar- celona, Spain. Poole, K.G. 1989. Determining age and sex of nestling Gyrfalcons. /. Raptor Res. 23:1-11. Rice, W.R. 1989. Analyzing tables of statistical tests. Evo- lution 43:223-225. Rocamora, G. 1994. Bonelli’s Eagle Hieraaetus fasciatus Pages 184—185 in G.M. Tucker and M.F. Heath [Eds ] , Birds in Europe, their conservation status. BirdLife International, BirdLife Cons. Ser. 3, Cambridge, U.K. Seutin, G., B.N. White, and T. Boag. 1991. Preservation of avian blood and tissue samples for DNA analyses Can. J. TLool. 69:82—90. SoDHi, N.S. 1992. Growth of nestling Merlins, Falco col- umbarius. Can. Field-Nat. 1206:387—389. SPSS, Inc. 1998. SPSS for Windows release 9.0— user’s guide. SPSS Inc., Chicago, IL U.S.A. Zar, J.H. 1996. Biostatistical analysis, 3rd Ed. Prentice- Hall International, London, U.K. Received 30 December 2000; accepted 15 May 2001 J Raptor Res. 35(3):194-206 © 2001 The Raptor Research Foundation, Inc. EMBRYONIC DEVELOPMENT OF THE AMERICAN KESTREL {FALCO SPARVERIUS ) : EXTERNAL CRITERIA FOR STAGING J.M. PiSENTI Department of Animal Science, University of California, One Shields Avenue, Davis, CA 95616 U.S.A. G.M. Santolo^ CH2M HILL, Inc., 2483 Natomas Park Drive, Suite 600, Sacramento, CA 93853 U.S.A. J.T. Yamamoto and A.A. Morzenti Department of Animal Science, University of California, One Shields Avenue, Davis, CA 95616 U.S.A. Abstract. — Descriptions of embryonic development exist for a handful of bird species. Such standard information is essential for the evaluation of species-specific features and detecting abnormal mor- phology. The American Kestrel {Falco sparverius) is a common North American raptor that is frequently used in experimental studies as a model raptor species. We described the normal progression of em- bryonic development in the American Kestrel. This provides a standard for assessing American Kestrel embryos, and potentially those of other raptors. During the first half of incubation, the developmental progression of American Kestrel embryos corresponded closely to developmental stages established in the chicken {Callus domesticus). Morphological parameters that we measured were correlated signifi- cantly with incubation day. These qualitative and quantitative descriptions provide useful benchmarks for determining age and identifying abnormalities of experimentally-treated embryos or embryos of unknown history. Key Words; American kestrel; Falco sparverius; avian embryology; embryo development; embryo staging. Desarrollo embrionario del cernicalo americano {Falco sparverius) criterios externos para su definicion Resumen. — Las descripciones del desarrollo embrionario existen para unas pocas especies. Esta infor- macion standarizada es esencial para la evaluacion de las caracteristicas especificas de una especie y para detectar morfologias anormales. El cernicalo americano {Falco sparverius) es una rapaz comiin de Norteamerica la cual es frecuentemente utilizada en estudios experimentales como una especie substi- tuto de ave rapaz. Describimos la progresion normal del desarrollo embrionario del cernicalo ameri- cano. Esto provee un estandar para evaluar los embriones de esta especie y potencialmente de otras rapaces. Durante la primera mitad de la incubacion la progresion del desarrollo de los embriones del cernicalo americano correspondio a las etapas del desarrollo establecidas para Callus domesticus. Los parametros morfologicos medidos fueron significativamente correlacionados con el dia de la incuba- cion. Estas descripciones cualitativas y cuantitativas representan un punto de referencia para determinar la edad e identificar las anormalidades de los embriones experimentalmente tratados o cuya historia es desconocida. [Traduccion de Cesar Marquez] The ability to age embryos accurately and assess normal development in birds is critical to many areas of biological study. Some of the more impor- tant applications include monitoring for environ- mental contaminant effects and determining nu- ' Corresponding author; Gary M. Santolo, CH2M HILL, Inc., 2485 Natomas Park Dr., Suite 600, Sacramento, CA 95833. tritional requirements for breeding birds. Currently, the most complete and detailed descrip- tion of avian embryonic development is that done for the domestic chicken. Callus domesticus (Ham- burger and Hamilton 1951, Hamilton 1952, Bel- lairs and Osmond 1998). Developmental progres- sions have also been described for other precocial birds including Ring-necked Pheasant {Phasianus colchicus; Hermes and Woodard 1987, Labisky and Opsahl 1958), Mallard {Anas platyrynchos; Caldwell 194 September 2001 Development of the American Kestrel 195 and Snart 1974), Bobwhite Quail (Colinus virgini- anus] Roseberry and Klimstra 1965), domestic chicken, turkey {Meleagris gallopavo), and Japanese quail {Coturnix japonica) (Abbott 1967) , and Adelie Penguin (Pygoscelis adeliae, Herbert 1967), as well as a handful of altricial species (Daniel 1956, Bird et al. 1984, Abbott et al. 1991, Hanbidge and Fox 1996) including the American Kestrel (Falco sparv- erius). In studies requiring a finely detailed assess- ment of development, it is desirable to have spe- cies-specific data on which to base comparisons. While gross or extreme embryonic deformities and stunting are generally distinguishable in the ab- sence of a reference, more subtle morphological changes may be overlooked without a normal stan- dard for comparison. With the exception of Bird et al. (1984), no spe- cies from the Falconiformes have been described during embryonic development. This is in spite of the numerous potential applications of such data, including comparative studies, aging of field col- lected embryos, assessment of abnormal develop- ment, and in captive breeding efforts for rare spe- cies. In this paper we provide, through measurements, qualitative description, and identi- fication of specific aging criteria, a detailed normal developmental progression of the American Kes- trel throughout the incubation period. Methods Animals and Treatments. We obtained captive-bred adult male and female American Kestrels from the Avian Science and Conservation Centre of McGill University (Montreal, Canada). Pedigrees for these birds were known, although the number of generations that were traceable varied among birds (1-4 generations). This group was supplemented by additional male and female American Kestrels we obtained from wild populations in California (Yolo and Solano counties; 38°N, 121°W). We cared for American Kestrels according to animal care protocols approved by the Office of the Campus Veteri- narian at UCD. Prior to the breeding period (early March), we paired 40 each of male and female American Kestrels and placed them in individual breeding pens (approximately 2 m X 2 m X 1.75 m). Most pairs had been together for the previous one or two breeding seasons, and had suc- cessfully produced fertile eggs. We maintained breeding pens at ambient temperature (range 0 to 37°C) in a large screen-sided building that provided protection from di- rect sunlight and rain. The building was equipped with supplemental lighting above the breeding pens, which was controlled by a timer set to coincide with the natural photoperiod. Each pen was equipped with a shelf and a rope perch, and a wooden nest box (entrance hole ap- proximately 7 cm diameter) containing autoclaved pine shavings to a depth of about 5 cm. We maintained birds on a nutritionally complete commercial raptor diet (Ne- braska Bird of Prey Diet, Central Nebraska Packing, North Platte, Nebraska, USA), supplemented with a pow- dered multivitamin additive (Vionate; ARC Laboratories, Atlanta, Georgia, USA). Each pair was provided with about 100 gm of fresh food daily, and water was provided ad libitum. Egg Collection and Incubation. We observed pairs daily for normal appearance and behavior and we checked nest boxes daily in the late afternoon for occupation by the male and/ or female American Kestrel and for newly laid eggs. Due to this collection schedule, some eggs may have been incubated by parent birds for up to 24 hr. However, such a prolonged incubation was unlikely as American Kestrels do not typically incubate until the clutch is nearly complete (typical clutch size was 4—5 eggs). We labeled new eggs using waterproof ink with pen number and Julian date, weighed them, then placed them in sterilized fiber chicken egg flats for immediate transport to cold storage (Heck and Konkel 1991). Eggs were held in cold storage ( 12.5-1 3. 0°C) for 3-4 d. Just before the start of incubation, we placed the eggs in plas- tic chicken flats (up to 30 eggs per flat), and fumigated them for one h with formaldehyde gas, followed by one hr in a formaldehyde-neutralizing compound (ammonia gas). During this procedure, the fumigation chamber was heated to 30°C, then cooled to 13°C. After fumigation, we placed the eggs in an egg storage cold box at 13°C for an additional 6 hr (until about 1800 PST on that day), then weighed them again and allowed them to warm at room temperature (22‘"C) for about 30 min before the start of incubation. We set eggs, air cell up, in plastic set trays designed for pheasant eggs and placed them m a Natureform NOM-125 incubator (Natureform Hatchery Systems, Jacksonville, Florida, USA) at 37.5°C and 55% relative humidity. In a previous study (Santolo et al. 1999), these conditions were shown to promote normal embryonic development and successful hatching of eggs from this colony. Eggs were automatically turned through 90° (45° right to 45° left, etc.) every 15 min. We candled eggs daily, using a variable intensity candler (Lyon’s Elec- tric, San Diego, CA, USA), modified with a 2.5 cm di- ameter black rubber hose taped to the candling mask. This modification served to move the egg further from the heat of the light source. Eggs were candled from both ends, and a record was made of the candling appearance for each egg. This information: (1) allowed early identi- fication of infertile eggs and early dead embryos, and (2) made it possible to identify pre-incubated eggs (those dis- playing development more advanced than other eggs in the age class). On Day 24 of incubation, we moved all eggs with live embryos into individual hatching baskets (sterilized plastic one pint produce baskets) in a table- top, forced draft incubator (Lyon’s Electric, San Diego, CA, USA), set at 37.5°C, 70-75% relative humidity. Breakout Examination. We broke out and examined any eggs that showed no sign of embryonic development after 7 d to determine fertility. We also opened eggs if the embryo appeared dead. During the first half of in- cubation, we selected live embryos for collection based on candling appearance. We usually collected more ad- vanced embryos (i.e., those likely to have been pre-in- cubated by their parents) during the late growth phase 196 PlSENTI ET AL. VoL. 35, No. 3 Figure 1. Line drawings showing embryo measurements: A, young embryo: T1 to T2, trunk length; LI to L2, leg length; W1 to W2, wing length {elbow to tip); El to E2, eye diameter; MBl to MB2, midbrain diameter (across widest section). B, older embryo: LI to L2, tarsus; W1 to W2, forearm; W2 to W3, alula; W2 to W4, manus; B1 to B2, lower beak; B3 to B4, culmen; E3, approximate eye-to-eye measurement; AM, approximate point of ear-to-ear measurement (width of skull at auditory meatus). of embryonic development, in order to minimize the ef- fect of pre-incubation on early embryo assessments. We weighed all eggs prior to opening them. For embryo col- lection, we measured the air cell diameter and then cut the shell open over the air cell and emptied the contents while submerging the egg in deionized water. We re- moved extraembryonic membranes, then weighed the embryo and placed it in a small dish of clean deionized water. We measured diameter of the air cell to the nearest 0 5 mm with calipers, and measured yolk and albumen mass (to nearest 0.001 g; Mettler Instruments, Hights- town, NJ, USA; model HRIAR), and volume (water dis- placement). We obtained yolk sac and albumen measure- ments when these could be reliably isolated from the surrounding water (i.e., no yolk or albumen measure- ments were made during the first week of incubation, as both materials tended to be difficult to isolate during this period). We made measurements of the head, trunk, and limbs to the nearest 0.5 mm using calipers and a metric ruler (see Fig. 1 for diagrams showing specific measure- ments) . For each age class of embryo, we selected struc- tures based on the degree to which they could be reliably measured and easily identified landmarks. We staged all embryos using standard chicken embryo criteria (Ham- burger and Hamilton 1951). Statistical Analyses. We used simple and second-order polynomial regression to develop equations for predict- ing embryo age from measured parameters and ANOVA to measure the quality of the models (SAS Institute 1998). Results Nonviable Eggs. Infertile eggs and early embryo mortality are described below and examples are presented in Appendix 1. Early dead. Very early failures in development will produce a mottled white membrane of varying size with a highly irregular outline spreading across the yolk. The white coloration tends to be most intense at the edge of membrane. If failure occurs slightly later, blood islets, and often parts of the embryo itself, will form. These early dead embryos may ap- pear as a darker region in the center of the mem- brane. Infertile. The infertile germ spot is mottled and whitish, with an irregular circular outline, sur- rounded by a slightly darker band of yolk (width September 2001 Development oe the American Kestrel 197 Table 1. Range of embryo stages found and the number of embryos examined during each day of incubation. Incubation Day H & H Stage^ N Range 1 2 2 2-3 2 4 3 3-7 3 8 3 6-10 4 13 3 12-16 5 18 3 18-19 6 20 3 17-21 7 24 4 23-25 8 25 4 25 9 27 4 25-29 10 28 3 27-29 11 30 4 30-31 12 32 4 30-33 13 33 4 32-34 14 35 3 35 15 36 3 36 16 36 + 3 36-37 17 37 3 37 18 38 3 38 19 38 3 37-38 20 39 4 38-40 21 40 4 37-40 22 42 4 40-44 23 44 2 44 24 44 4 44 25 44 + 1 44 + 26 45 - 3 45 - 27 45 3 45-46 28 46 3 46 Stage of embryonic development as described for the chicken by Hamburger and Hamilton (1951). of the darker band is 0.5-1. 0 mm). The appear- ance is very similar to that of the Day 0 blastoderm. Normal Developmental Stages of the American Kestrel. The text below describes the embryonic development of the American Kestrel, emphasizing key diagnostic features for assessing embryo age and (morphologic) normalcy. Along with descrip- tive text, each daily account includes the approxi- mate stage (median stage with observed range of stages in parentheses) of the embryo, based on normal stages of the domestic chick (Hamburger and Hamilton 1951; see Table 1). Selected daily accounts correspond to embryo photographs in Appendix 1-4. A minority of the descriptions for some aspects of development are clearly only rel- evant to living embryos and therefore are not use- ful in assessing dead embryos. In addition it should by noted by users that quantitative measurements of embryo features are provided as approximations based on measures of a limited number of embry- os. Variability around these values is to be expect- ed, however, it is anticipated that they will provide useful benchmarks for embryo aging. Descriptions of embryos that died very early in development and infertile eggs are also provided to assist in dis- tinguishing these eggs from viable eggs, either un- incubated or partially incubated. In determining fertility, it may be most useful to compare the de- scriptions for infertile eggs and Day 0 (unincubat- ed fertile) eggs below. Days 0 to 4. Key diagnostic criteria include qual- itative characteristics and diameter of the blasto- disc and yolk sac membrane. Also observed during this period are the appearance of somites, head process, and heart, and establishment of the em- bryonic axis. Day 0 Blastoderm appears as a solid white disc, 1.5 mm in diameter, having distinct edges surrounded by a darker region of yolk. Yolk sac may be 2-5 mm in diameter in pre-incubated eggs. Day 1 Stage 2 (2-3) . Blastoderm appears as a dis- tinct white ring, 1.5 mm in diameter. Yolk sac may be 13 mm in diameter in pre-in- cubated eggs. Day 2 Stage 4 (3-7). Primitive streak is distinct. The entire area pellucida/area opaca ap- pears as a raised-domed structure protrud- ing above the surrounding yolk. Day 3 Stage 8 (6-10). Head-fold is visible at an- terior end of embryo. At least four pair (4- 9) of somites are visible. Area pellucida is 3 mm long and pear-shaped. Day 4 Stage 13 (12-16). Blood islets surround embryo. Amniotic fold covers head to hindbrain. Otic pits are visible just above first somite. Head is turning over onto left side. Heart tube is beginning to loop. Days 5 and Onward. Aging is based primarily on development of the head, limbs and tail, and po- sition of the embryo on the yolk. Days 8-12 focus on eye, eyelid, brain size, limb length, neck length, and trunk length. Main diagnostic features from Day 25 to hatch are sloughing of the periderm, keratinization and length of nails and beak, eye diameter, etc. Day 5 Stage 18 (18-19). Eye: Eye unpigmented, 198 PiSENTI ET AL. VOL. 35, No. 3 lens present, midbrain same size as eye. Yolk: Blood moving in yolk vasculature in response to heart tube contractions. Heart: Heart looped in full circle. Limbs: Wing and leg primordia are just visible. Amnion: Amnion may be closed but contains very little fluid. Flexures: Cervical flexure 90 de- grees from trunk to midbrain (L-shaped). Embryo has turned onto left side from head down to mid-torso. Tail: Tail bud is cone-shaped and extends along main body axis. Day 6 Stage 20 (17-21). Eye: Eye faintly pigment- ed, with choroid fissure clearly visible. Vis- ceral arches: Otic vessicle just dorsal to 2nd and 3rd visceral clefts. Midbrain slightly larger than eye. Nasal placodes have deep- ened into pits. Limbs: Both wings and legs are wider than long. Amnion: Amniotic sac sealed but contains very little fluid. Flex- ures: Completely turned onto left side. Tail: Tail bud is perpendicular to main body axis. Allantois: Allantoic sac just visible be- hind right leg bud. Day 7 Stage 23 (23-25) . Eye: Eye heavily pig- mented and lens clearly visible. Mandible is % length of the maxillary process. Limbs: Wings and legs are as long as they are wide. Amnion Flexures: Embryo curved in a C- shape around the heart. Tail: Tail bud has a distinct ventral “hook” at the tip. Allan- tois: Allantoic sac highly vascularized, may cover the eye and forebrain. Day 8 Stage 25. Beak: Sides of the beak still sep- arated from the tip by the nasal groove. Visceral Arches: Otic vesicle about same size as lens. Collar at the base of the neck is distinct and raised. Limbs: Elbow and knee joints distinct on limbs. Wing tip and foot area flattened into paddles but no digits are visible. Both wings and legs longer than wide. Allantois: Allantois covers head and part of the body. Day 9 Stage 27 (25-29). Beak: Tip of upper beak is a square protrusion. Limbs: Connective tissue just visible for tibia/ fibula and radi- us/ulna. Wing middle digit longer than outer two with a slightly fan-shaped digital plate. Distinct grooves between toe pri- mordia. Five toes visible. Amnion: Yolk vas- culature at or approaching the albumen in the small end of the egg. Contractions of the amnion moderate and frequent. Am- nion filling with fluid. Allantois: Allantois/ Chorioallantoic membrane (CAM) covers embryo except along spine. Day 10 Stage 28 (27-29). Beak: Mandible is about V 2 the length of the maxilla and very square when viewed from front. Distinct falcon’s notch (i.e., tomial tooth) is visible just anterior to nasal groove. Limbs: Slight grooves visible between digits on wings. Al- ula is distinctly the shortest digit. Feathers: No distinct primordia, but scapulary tract obviously raised. Day 11 Stage 30 (30—31). Eye: Nictitating mem- brane just visible at anterior corner of eye. Upper and lower eyelid folds just visible. Eye appears about the same size as mid- brain. Beak: Mandible has distinct bend at midpoint and is about same length as max- illary process. Visceral Arches: Nares may be visible at top of nasal groove. Limbs: Legs now longer than the tail bud. Wing slightly bent at wrist. Amnion: Amnion mildly con- tractile. Allantois: CAM now extends over about %Q of yolk sac vascular region. Allan- tois filled with mostly clear fluid. Day 12 Stage 32 (30-33). Eye: Eyelids covering about Vs of eye. Approximately 2-8 scleral papillae visible. Beak: Distinct falcon’s tooth. Lower mandible is wider than upper mandible from frontal view. Egg tooth may be visible on top of beak. Limbs: Alula sep- arated from wing tip. Fifth toe may be gone. Feathers: Two distinct rows of feather primordia on either side of spine. Allantois Lobes of CAM starting to surround albu- men at small end of egg. Day 13 Stage 32-34. Eye: 14 scleral papillae. Nicti- tating membrane about V 2 way across eye- ball toward the scleral papillae. Visceral Arches: Nasal groove fused. Collar at the base of the neck flattened or gone. Beak. Distinct egg tooth. Limbs: Web well re- gressed between digits 2 and 3 on the foot. Amnion: Amnion is inflated with clear flu- id. Feathers: Feather buds visible along spine, neck and scapular tracts and tail rec- trices, pectoral tracts. Day 14 Stage 35. Eye: Nictitating membrane and upper eyelid almost to scleral papillae. Limbs: Toes well separated. Primary toe pads just visible. Ammon; Albumen starting September 2001 Development of the American Kestrel 199 to enter amniotic sac. Tail: Tail almost par- allel with spine. Feathers: Feather buds visi- ble on top of head, eyelids, thighs and back and along ulna. Gonads: Male and fe- male gonads can be differentiated. Male gonads are bean-shaped. Left female go- nad is flattened and translucent, and dis- tinctly larger than right gonad. Day 15 Stage 36. Eye: Eyelid opening is flattened ellipse with lower lid at edge of cornea. Feathers: Primary feather buds just visible on manus. Feather buds just visible around ear opening. Day 16 Stage 36-37. Eye: Nictitating membrane at edge of cornea. Limbs: Primary toe pads well defined. First three scutate scales are on top of foot. Cornification just begin- ning on dorsal side of toenail. Feathers: Pri- maries and secondaries are longer than wide. Allantois: GAM sticks tightly to shell. Day 17 Stage 37. Beak: Groove at tip of mandible just visible. Limbs: Slight ventral curve to toenails. Legs tend to be crossed in front of body. Allantois: CAM may be closed over albumen. Allantoic fluid may be cloudy with precipitate. Day 18 Stage 38. Beak: Upper beak, but not lower beak, starting to cornify around egg tooth. Under side of lower beak (“chin”) is dis- tinctly rounded. Limbs: Scale primordia covering tops of tarsus and tops of toes, not yet overlapping. Nail bed has distinct ridge at base of toenails. Feathers: Feather buds around ear. Two rows of eyelash feather buds. Cloaca: Cloaca distinctly raised and oval. Day 19 Stage 38. Eye: Lens partially covered by eye- lids. Beak: Beak and face may be covered by a lobe of the yolk. Cross-shaped corni- fication centered around egg tooth. Small cornification on lower beak at tip of man- dible. Amnion: Coagulated albumen stick- ing to embryo. Limbs: Scales starting to overlap along the front of the tarsus. Scales appearing along the back of tarsus and on primary toepads. Secondary toepads well- defined. Toenails strongly flexed on hal- lux. Feathers: Feather buds visible on cere. Allantois: Precipitate throughout allantoic sac. Day 20 Stages 38-40. Eye: Eye is almost closed. Beak: Periderm visible on beak. Limbs: Toe- nails flexed at a 90° angle to toe. Amnion' Yolk is in two distinct lobes on either side of the embryo. Only small amount of am- niotic fluid remains. Amnion not contrac- tile. Allantois: Allantoic fluid may be clear, but with precipitate. Day 21 Stage 40. Eye: Eyelids completely closed. Beak: Flole in periderm over egg tooth. Periderm may be starting to separate from cere. Limbs: Scales overlapping on the back of the tarsus. Scales on secondary toepads. Day 22 Stages 40-44. Eye: Eye is fully closed. Beak' Bony tubercule visible in nares. Tip of mandible is even with falcon’s tooth. Peri- derm may be starting to separate from cere. Hatching muscle: Hatching muscle starting to swell. Day 23 Stage 44. Beak: Beak cornification may be complete. Periderm is separating from the cere. Scalloping on side of mandible is di- minished. Tip of mandible extends beyond falcon’s tooth. Amnion: Trace amount of al- bumen in small end of egg; most is in am- niotic sac or sticking to the feathers. Eeath- ers: Eyelash feathers much longer than wide. Feathers over entire body are fila- mentous and white. Day 24 Stage 44. Head: Head near air cell, may be under right wing, but not usually pipped. Beak: Periderm has sloughed about halfway from upper and lower beaks. Egg tooth may have started to wear through the CAM. Limbs: Nails are completely kerati- nized. Amnion: CAM easily separates from shell. Allantois: Most of allantoic fluid is gone. Hatching muscle: Maximum edema of hatch muscle, which may extend into the shoulder area. Cloaca: Cloaca is flattened oval, just raised above the surrounding skin. Day 25 Stage 44+ . Amnion: Yolk just starting to en- ter abdomen. No fluid in amnion. Flexures: Head is under right wing. Other: First crack in shell near the equator. Day 26 Stage 45 — . Beak: Periderm has sloughed. Entire beak appears shiny. Allantois: CAM often does not close completely over the albumen. Very little fluid in allantoic sac, just yellowish strings of urate precipitates. Almost no albumen remaining in the small end of the egg. Yolk: Sac is 14 into abdom- VoL. 35, No. 3 200 PiSENTI ET AL. Figure 2. Egg mass loss as a percent of fresh egg mass (± 1 SE) of American Kestrel eggs developing to hatch (H), American Kestrel eggs failing before hatch (D), and normal chicken eggs (CK). inal cavity. Other: Chick may have pipped into the air cell and may be vocalizing. Day 27 Stage 45. Beak: Beak is in the air cell. Al- lantois: Most of the CAM blood vessels are empty. Yolk: Yolk may be completely inside of abdominal cavity. Typically, yolk will be orange-yellow in color. Other: Chick is peeping loudly and persistently. Day 28 Stage 46. Hatched chick. Yolk fully inter- nal. Egg Mass Loss and Embryo Measurements. American Kestrel eggs lost less mass as a percent of initial (i.e., fresh) egg mass at a slower pace than chicken eggs at comparable developmental stages (Fig. 2). When the embryo died during develop- ment, egg mass loss slowed (Fig. 2). A number of embryo parameters were correlat- ed with incubation day (Table 2). With the excep- tion of yolk and albumen measures, relationships were positive. For each parameter, the ranges of incubation days over which the parameter-incuba- tion day relationship was analyzed are shown in Ta- ble 2, and in general reflect the time period over which the parameter could be accurately mea- sured. Discussion This paper presents the first description of the daily embryonic development for a raptor species and provides a potentially useful tool for experi- mental and field assessments of the development of American Kestrels and possibly other raptors. Use of a species-specific guide is particularly im- portant for identifying morphological abnormali- ties in embryos, which would not otherwise be apparent if comparisons were made to a taxonom- ically distant species (e.g., chicken). Such abnor- malities may be indicators of embryo exposure to pathogens, genetic mutations, physical, thermal or nutritional stresses, or toxic concentrations of some chemicals (Romanoff and Romanoff 1972). In addition to assessment of normalcy, the external criteria described in this study can be used for es- timating age in days, or equivalent Hamburger and Hamilton (1951) stage, in embryos that have been incubated and/or dead for unknown periods, such as those collected in the field. Clearly, some de- scriptions and measured parameters will be more practical than others depending on the condition of the embryo. In embryos that have been dead for some period, dehydration and decay may limit the utility of some of the visual and measured cri- September 2001 Development of the American Kestrel 201 Table 2. Significant relationships observed between incubation day and egg and embryo measurements over the incubation period (P < 0.001). For each parameter, columns show range of incubation days over which measures were taken (Incubation Days), sample size (N), regression equation describing relationship with incubation day, X, value, and F statistic. Parameter (X) Incubation DA'ffi N Equation F F Air cell 0-26 68 -21.03 + 2.03 (X) 0.61 101.3 Eye diameter 5-28 79 5.56 + 0.72 (X) + 0.106 (X)2 0.90 344.7 Brain vessicles/ midbrain 5-13 27 3.67 + 1.28 (X) 0.89 208.6 Wing length 6-11 16 4.79 + 1.44 (X) 0.79 54.5 Mandible 7-28 58 4.09 + 2.33 (X) 0.95 1086.8 Trunk 7-28 63 3.59 -E 0.76 (X) 0.96 1456.6 Yolk sac mass 8-28 47 31.23 — 4.45 (X) + 0.211 (X)2 0.83 104.3 Yolk sac volume 8-28 48 32.17 — 4.97 (X) + 0.267 (X)2 0.84 117.5 Albumen mass 8-24 37 24.10 — 3.45 (X) 0.78 122.9 Albumen volume 8-24 40 24.22 — 3.82 (X) 0.79 139.1 Embryo mass 9-28 53 9.65 + 3.42 (X) — 0.174 (X)2 0.96 567.1 Eye-to-eye 9-28 62 6.81 + 0.09 (X) + 0.050 (X)2 0.79 114.2 Third toe length 10-28 53 5.96 + 2.28 (X) 0.86 318.9 Alula 11-28 54 5.51 + 3.65 (X) 0.85 239.9 Forearm 11-28 56 10.12 + 0.36 (X) + 0.073 (X)2 0.94 429.8 Manus 11-28 56 10.26 + 0.34 (X) + 0.073 (X)2 0.92 304.2 Culmen 11-28 56 6.52 + 2.94 (X) 0.94 793.5 Tarsus 13-28 49 8.43 + 1.43 (X) 0.94 739.8 Tibia 13-28 49 7.83 + 1.04 (X) 0.96 989.7 Ear-to-ear 13-28 49 12.04 — 0.28 (X) + 0.084 (X)2 0.87 149.8 teria described here. Parameters that may remain most useful for postmortem evaluation include limb and beak measurements, eye pigmentation, leg and foot development (e.g., scales, toenail ke- ratinization) , and feathering. In a previous study, the embryonic development of the American Kestrel was described for selected days during incubation using naturally incubated eggs (Bird et al. 1984). Findings reported here for artificially incubated embryos compare favorably with the prior study with respect to timing of ap- pearance of various external features (e.g., eye pig- mentation, allantoic sac, development of toe digits, toenails, and down) . A potential confounding fac- tor in artificial incubation studies of embryos is the possible pre-incubation of eggs by parent birds be- fore collection, which could result in an apparent advancement of embryo maturation, particularly at early stages of development. We controlled for this effect as much as possible by carrying out frequent egg collections, and excluding eggs that were rel- atively advanced, based on candling appearance, from early development examinations. Compari- son of our results with those of the Bird et al. (1984) study suggest that American Kestrel embry- os developed at “normal” rates under the condi- tions of artificial storage and incubation we used. The appearance of the germinal disc at the time of oviposition or before the onset of incubation can be used to determine fertility or early death. However, normal variations in size, shape, and col- or patterns may make it difficult to differentiate reliably between an infertile germinal disc and a developing blastoderm. Such variations have been described in great detail for the domestic turkey, (Bakst et al. 1998). Although precocial, the turkey shares several characteristics with the American Kestrel: an incubation period of 28 days, a blasto- disc that typically appears as a circular, uniformly white structure (see Appendix 1; in contrast to the chicken blastodisc, in which the white area opaca is typically seen as a ring around the darker area pellucida) , and the presence of small vacuoles sur- rounding the unfertilized germinal disc, which may closely resemble the fertile blastoderm in size, shape, and color. Thus, the morphological classes described in detail by Bakst et al. (1998) are a po- tentially useful guide to assessing fertility and ab- 202 PiSENTI ET AL. VoL. 35, No. 3 normalities of American Kestrel germs at this very early period in development. The chicken has long been used as a develop- mental standard for aging avian species (Ham- burger and Hamilton 1951), and it is considered to be accurate for both precocial and altricial spe- cies up to stage 42 (about % of the way through normal incubation; Ricklefs and Starck 1998). However, the chicken staging charts are not partic- ularly useful in aging altricial embryos during the last Vs of incubation, which is characterized by a rapid increase in size. Also, despite the utility of the chicken model, differences in morphology of key embryonic structures between the chicken and specialized altricial species, such as the American Kestrel, can make it difficult to assess incubation age accurately. Acknowledgments Funding for this research was provided by the U.S. Bu- reau of Reclamation under contract number 9-CS- 20S00440 for the Kesterson Reservoir biological moni- toring program. Wild American Kestrels were trapped under California Department of Fish and Game Scientif- ic Collecting Permit No. 801023-02. We thank the Center for Ecological Flealth Research at the University of Cali- fornia at Davis. We particularly thank S. Miller for graph- ics assistance, the UCD undergraduate students that helped with data recording, and F. Bradley for use of laboratory facilities. Literature Cited Abbott, U.K. 1967. Avian developmental genetics. Pages 12-52 in F.W. Will and N.K. Wessels [Eds.], Methods in developmental biology. Thomas Y. Crowell Co., New York, NY U.S.A. , B. Cutler, A.T. Brice, and J.R. Millam. 1991. Development of the Cockatiel embryo./. Assoc. Avian Vet. 5:207-209. Bird, D.M., J. Gautier, and V. Montpetit. 1984. Embry- onic growth of American Kestrels. Auk 101:392-396. Bakst, M.R., S.K. Gupta, W. Popts, V. Akuffo. 1998. Gross appearance of the Turkey blastoderm at ovi- position. Poult. Sd. 77:1228-1233. Bellairs, R. and M. Osmond. 1998. The atlas of chick development. Academic Press, San Diego, CA U.S.A. Caldwell, PJ- AND A.E. Snart. 1974. A photographic in- dex for aging Mallard embryos. J. Wildl. Manage. 38; 298-301. Daniel, J.C. 1956. A comparative study of the embryo- logical development of the domestic fowl and the Red-winged Blackbird. Ph.D. dissertation. University Colorado, Boulder, CO U.S.A. Hamburger, V. and H.L. Hamilton. 1951. A series of normal stages in the development of the chick em- bryo. J. Morphol. 88:49-92. Hamilton, H.L. 1952. Lillie’s development of the chick. Henry Holt and Co., Inc., New York, NY U.S.A. Hanbidge, B.A. and G.A. Eox. 1996. Egg characteristics, growth and developmental landmarks of known-age embryos of Double-crested Cormorants from Mani- toba. Colon. Waterbirds 19:139-142. Heck, W.R. and D. Konkel. 1991. Incubation and rear- ing. Pages 34—76 mJ.D. Weaver and T.J. Cade [Eds ], Ealcon propagation: a manual on captive breeding. The Peregrine Fund, Boise, ID U.S.A. Herbert, C. 1967. Timed series of embryonic develop- mental stages of the Adelie Penguin (Pygoscelis adeliae) from Signy Island, South Orkney Islands. Br. Antarct Surv. Bull. 14:45-67. Hermes, J.C. and A.E. Woodard. 1987. Development of the pheasant embryo: A record of daily visual chang- es. University of California, Department of Avian Sci- ences. lm-pr-8/87-LE/ALS, Davis, CA U.S.A. Labisky, R.F. and J.F. Opsahl. 1958. A guide to aging of pheasant embryos. Illinois Natural History Survey Di- vision. Biological Notes No. 39, Urbana, IE U.S.A. Ricklefs, R.E. and J.M. Starck. 1998. Embryonic growth and development. Pages 31-55 mJ.M. Starck and R E Ricklef [Eds.], Avian growth and development. Ox- ford University Press, New York, NY U.S.A. Romanoff, A.L. and AJ. Romanoff. 1972. Pathogenesis of the avian embryo. Wiley-Interscience, New York, NY U.S.A. Roseberry, J.L. AND W.D. Klimstra. 1965. A guide to age determination of Bobwhite Quail embryos. Illinois Natural History Survey Division. Biological Notes No. 55. Urbana, IL U.S.A. SAS Institute. 1998. StatView reference. Release 5 0. SAS Institute, Cary, NC U.S. A. Santolo, G.M., J.T. Yamamoto, J.M. Pisenti, and B.W Wilson. 1999. Selenium accumulation and effects on reproduction in captive American Kestrels fed sele- nomethionine. /. Wildl. Manage. 63:502-511. Received 23 June 2000; accepted 23 April 2001 September 2001 Development of the American Kestrel 203 Appendix 1. American Kestrel egg contents and embryos from 0-4 d. 204 PiSENTI ET AL. VoL. 35, No. 3 Appendix 2. American Kestrel embryos from 5-12 d. September 2001 Development of the American Kestrel 205 Appendix 3. American Kestrel embryos from 13-22 d. 206 PiSENTI ET AL. VoL. 35, No. 3 Appendix 4. American Kestrel embryos from 23-28 d. /. Raptor Res. 35(3):207-213 © 2001 The Raptor Research Foundation, Inc. ANALYSIS OF BALD EAGLE SPATIAL USE OF LINEAR HABITAT Alan R. Harmata Fish & Wildlife Program, Department of Ecology, Montana State University, Bozeman, MT 5971 7 US. A. George J. Montopoli Department of Mathematics, Arizona Western College, Yuma, AZ 85364 US. A. Abstract. — Several techniques are available for areal analysis of animal locations but few are applicable to those that use linear (i.e., riparian) habitats. Bald Eagles {Haliaeetus leucocephalu.!) often are associated with rivers and concentrate perch sites near shorelines. Distribution of cumulative proportion of perches by distance from most recently active nest sites determined by radio tracking were used to compare spatial use among five adult Bald Eagles breeding along the Snake River, Wyoming. Spatial Use Indi- cators (SUIs) were developed from logistic regression parameters in attempts to: (1) understand and model underlying processes from which the data may have emerged, (2) compare with simple descrip- tive statistical techniques to evaluate utility for presenting a clear, accurate representation of spatial use differences among eagles, and, (3) relate measures of eagle spatial use with long-term productivity of breeding areas. Distance Indicator (DI) was the distance from the nest including 50% of all detected perches used by a radio-tagged eagle and was representative of the size of the range. Slope Indicator (SI) was the slope of the fitted logistic regression curve at the DI (inflection point) . SI was an indicator of linear dispersion of perch sites within the breeding area. Bald Eagles associated with more productive (>0.77 young per occupied nest over 11 years) breeding areas perched closer to nest sites (similar DIs) than eagles of their respective gender in a breeding areas of low productivity (<0.77 young per occupied nest) . Male Bald Eagles in highly productive breeding areas dispersed perch sites more evenly through- out the breeding area (flat SI) than a male in a low production breeding area, while the opposite was true for females. Spatial use profiles derived from analysis of mean and confidence intervals and median and Interquartile Ranges were not as descriptive or illustrative of individual or group similarities or differences as SUIs. Logistic analysis suggested Zone II (primary foraging zone) limits recommended in regional Bald Eagle management plans may need to be extended to maintain performance of highly- productive pairs nesting along rivers. SUIs derived from logistic regression models of distance of loca- tions from important habitat components may be indirect indicators of habitat quality and useful tools for describing and comparing spatial use of linear habitats of other species. Key Words: Greater Yellowstone Ecosystem; Haliaeetus leucocephalus; logistic regression; spatial use indicators; linear habitat; radio-tracking. Analisis espacial del uso de habitat lienar del aquila calva Resumen. — Existen varias tecnicas disponibles para el analisis de areas y de ubicacion de animales. Pocos son aplicables a aquellos que utilizan habitat lineares (i.e., habitats riberenos) . Las aguilas calvas {Hal- iaeetus leucocephalus) usualmente estan asociadas a rios y se concentran en sitios de perchas cerca de las lineas costeras. La distribucion acumulada de la proporcion de percha por la distancia del sitio del nido mas recientemente activo, (determinado por telemetria), fue utilizada para comparar el uso espacial de cinco aguilas calvas adultas que anidaron a lo largo del Rio Snake, Wyoming. Los indicadores de uso espacial (lUE) fueron desarrollados a partir de parametros de regresion logistica con el fin de: (1) Comprender y modelar los procesos subyacentes de los cuales los datos hayan emergido, (2) Comparar con tecnicas de estadistica descriptiva simple y evaluar la utilidad de presentar una clara y acertada representacion del uso espacial y sus diferencia entre aguilas, (3) Relacionar las medidas del uso espacial de las aguilas con la productividad a largo plazo de las areas de reproduccion. El Indicador de distancia (ID) fue la distancia desde el nido incluyendo el 50% de todas las perchas utilizadas por un aguila con transmisor y que fuera representativa del tamano del rango. El indicador de pendiente (IP) fue la pendiente de la curva de la regresion logistica en el ID (punto de inflexion). El IP fue un indicador de la dispersion linear de los sitios de percha dentro del area reproduccion. Se estudiaron las aguilas calvas asociadas a una mayor productividad (>0.77 juveniles por nido ocupado de mas de 11 anos), 207 208 Harmata and Montopoli VoL. 35, No. 3 areas de anidacion y de perchas cercanas a los sitios del nido (con ID similares), en contraposicion de aguilas de su respective sexo en areas de reproduccion de baja productividad (<0.77 juveniles por nido ocupado). Los machos de aguilas calvas en areas de reproduccion altamente productivas con sitios de perchas dispersas en forma similar a lo largo del area de reproduccion (IP piano) en contraposicion de un macho en un area de reproduccion con un productividad baja, se encontro que lo opuesto ocurrio con las hembras. Los perfiles de uso espacial que se derivaron del analisis de la media y de los intervalos de confianza y de la mediana y los intervalos entre rangos no fueron tan descriptivos ni ilustrativos de las similaridades individuales o de grupo o de las diferencias como los IVE. El analisis logistico sugirio que los limites de la zona II (zona de forrajeo primario) recomendada en los planes regionales de manejo de aguila calva necesita ser extendida para mantener su ocupacion por parte de las parejas altamente productivas que anidan a los largo de los rios. Los IVE derivados de los modelos de regresion logistica de las distancias de localidades provenientes de componentes de habitat impor- tantes pueden ser indicadores indirectos de la calidad de habitat y como herramientas valiosas para describir y comparar el uso espacial de los habitats lineares de otras especies. [Traduccion de Cesar Marquez] Several techniques are available to analyze spa- tial distribution of animal locations (Mohr 1947, Worden 1989, Andries et al. 1994, Bogel et al. 1995, Kie et al. 1996, Marzluff et al. 1997, Buchan- an 1997). Most involve construction of two-dimen- sional polygons or kernel ellipses that may include increasing proportions of total animal locations as descriptors of spatial use. Such methods are appro- priate for animals that distribute their activities somewhat uniformly around activity centers but may be misrepresentative when applied to animals that distribute locations linearly along habitat cor- ridors such as rivers. Here, we present a method for comparative analysis of spatial use of Bald Ea- gles (Haliaeetus leucocephalus) using a mostly linear, riparian corridor and contrast the utility for de- scribing linear spatial use with standard statistical techniques. Analysis was confined strictly to one di- mensional spatial distribution of perches within the breeding area and quantified relative to an im- portant component of the habitat, the nest site. Characteristics of spatial use among eagles were compared and related to number of young pro- duced in associated breeding areas, long-term. Mrriious From 1985 through 1989, movements of nesting Bald Eagles were investigated in the Greater Yellowstone Eco- system of northwestern Wyoming. Bald Eagles were resi- dent at nest sites along the —108 km free-flowing portion of the Snake River in Teton County and Grand Teton National Park (see Swenson et al. 1986 for description of study area and Harmata et al. 1999 for description of Bald Eagle population). Data were collected primarily for development of nest site management plans advocated in the Greater Yellowstone Bald Eagle Management Plan (Greater Yellowstone Bald Eagle Working Group 1996). Resident eagles were randomly chosen for study but ul- timately selected based on capture success (some chosen were never caught) . Adult Bald Eagles were captured and radio-tagged with tail-mount and solar backpack trans- mitters. Gender of four eagles was determined by posi- tion during copulation post-release and measurements (Bortolotti 1984, Garcelon et al. 1985) and size relation- ship with the mate for another. Primary function of te- lemetry was to facilitate continuous visual monitoring. Transmitters assisted in locating marked eagles at the ini- tiation of an observation period and aided relocation when eagles moved out of sight. Observation periods var- ied from 1 to 4 hr. Monitoring schedule was designed to provide an even distribution of effort over all hour pe- riods in each week as possible. Bald Eagle perches located during monitoring were plotted on U.S. Geological Survey 7.5-min topographical maps. A perch was defined as any nonflying, diurnal ea- gle location detected, either visually or triangulated by telemetry. Perches were located on trees and logs, in wa- ter, on the ground, or on man-made structures. Trian- gulated locations were easily distinguished as perched or flying by signal characteristics. Triangulated locations were ± 100 m of actual as determined by locations of test transmitters. Activity (hunting, loafing, territorial signal- ing, sentinel, etc.) of eagles at perch sites was not evalu- ated. Night roosts were not included, nor were perches chosen immediately subsequent to infrequent {N < 5) observer-induced movements. Perch locations were analyzed in relation to proximity to the most recently active nest within the radio-tagged eagle’s breeding area. The most recently active nest was defined as an elevated structure composed of sticks, sit- uated in a coniferous or deciduous tree, at which resi- dent adult Bald Eagles were last known to be engaged in reproductive activity. Reproductive activity included nest repair, copulation, incubation, brooding, feeding, or fledging young. Perch and nest locations of each radio-tagged eagle were assigned coordinates based on the Universal Trans- verse Mercator system within the GSDIG subprogram of GEOSGAN (Montana Department of Fish, Wildlife and Parks 1984), a computer-based system of geographic in- formation programs that related animal locations with geographic/habitat data. Straight-line distances between perch and nest locations were calculated in meters by September 2001 Bald Eagle Spatial Use Analysis 209 GEOSCAN subprogram MDPP (Minimum Distance Point to Point). Data were compiled for analysis by tabulating distance of each perch detected from the nest site, nearest to far- thest for each eagle tracked. Because plots of the cu- mulative number of perches at each distance displayed sigmoid shape typical of logistic curves, a logistic regres- sion model (Cox 1972, Hosmer and Lemeshow 1989) was developed for each radio-tagged Bald Eagle. Use of lo- gistic regression was an attempt to understand and model underlying processes from which the data may have emerged. The dependent variable {p) for the logistic model; gPo+Pl(Dx) ^ 1 -f gPo+Pl(Dx)’ was defined as the cumulative proportion of perches de- tected. The independent variable (D^) was distance from the most recently active nest site. Model parameters Pq ( intercept) and Pj (slope) were estimated for each dis- tribution using iteratively reweighted least squares tech- niques (Montgomery and Peck 1982) so that the models could be compared among eagles. Two parameters associated with the logistic model were identified as spatial use indicators (SUls): (1) the dis- tance from the nest including 50% of all perches or dis- tance indicator (DI) and, (2) the slope of the regression curve at the DI (inflection point) or slope indicator (SI) . DI was chosen because it reflected the relative size of the breeding area, tended to minimize effect of outliers and mathematically, was the point on the logistic curve where slope was the steepest. SI was considered an indicator of perch site dispersion within the breeding area. Steep slope (large coefficient) at the inflection point indicated that perches were clustered around the DI. A flatter slope (small coefficient) was indicative of a “lazy-S” curve, in- dicating perches were more evenly distributed through- out the breeding area. DI was calculated by solving for D^ in the logistic equa- tion with p = 0.5. Therefore, DI = — Pq/Pi- SI was cal- culated by differentiating the logistic equation {dp/ dx = Px/i(I — p)] with p = 0.5, resulting in SI = Pi/4. Construction of confidence intervals of DI and SI for among eagle comparisons required estimates of variance (Var) of both Pq and p^. Regression analysis provided es- timates of variance of Po [Var (Po)], variance of Pj [Var (Pi)], and their covariance [Covar (Po, Pi)] which were used to construct confidence intervals. Variance of the Sis (Pi/4) was calculated by Var(pi)/I6. However, DIs involved a ratio of two parameters ( — po/pi) and variance was estimated using the delta method (Bishop et al. 1975): Var Var(Po) pi" - 2|^Covar(P„, p,) Pi + (^) Var(p,). Data also were analyzed by simple mean, median, and post-hoc tests. Results were compared to SUIs to evaluate which methods more accurately depicted Bald Eagle spa- tial use. P value accepted as significant was adjusted for experiment-wise error using Bonferroni criteria (P < a Table I. Perch sites and events detected during moni- toring of adult Bald Eagles along the Snake River, Wyo- ming, 1985-89. Breeding Area Gender Monitored Perches Detected Hours Days Sites Events Butler $ 169 21 16 75 Cabin Creek 364 77 46 187 Oxbow ? 448 118 22 412 Sheep Gulch c? 274 74 26 63 Schwabacher (3 463 164 184 433 = 0.05/10 = 0.005). Differences in spatial use profiles among eagles were indicated by nonoverlap of Inter- quartile Ranges (IQRs) and 95% confidence intervals of DIs, Sis, and means. SUI of radio-tagged eagles was related to productivity of the breeding area and gender. Average number of young produced per occupied breeding area between 1979 and 1990 (Greater Yellowstone Bald Eagle Working Group 1996) was used to classify productivity as high or low for comparisons. This long-term productivity was con- sidered more representative of resident eagles’ perfor- mance than productivity coincident with period of radio- tracking because severity of early spring weather affected productivity of Bald Eagles in the Greater Yellowstone Ecosystem (Harmata and Oakleafl992) and weather var- ied from mild to severe between 1985 and 1989. Addi- tionally, some eagles in sample breeding areas were res- ident up to 11 years (Harmata et al. 1999). Because Sprunt et al. (1973) indicated at least 0.77 young per occupied breeding area was required for Bald Eagle pop- ulations to maintain stability, high productivity was con- sidered >0.77 young produced per occupied breeding area between 1979 through 1990. Low productivity was considered ^0.77 young per occupied breeding area for the same period. Results Five radio-tagged adult Bald Eagles were moni- tored an average of 343.6 hr (SD = 123.4) over 90.8 d (SD = 53.5) (Table 1). Monitoring covered at least two-thirds of the nesting period (mid Feb- mid Jul) for all eagles. Mean number of perch events and perch sites detected per eagle was 234 (SD = 179) and 59 (SD = 71), respectively. Perch events detected were correlated with hr (r = 0.917, P < 0.028) and d (r = 0.888, P = 0.044) moni- tored. However, number of perch sites detected was not correlated with either hr {P > 0.296) or d (P > 0.117) monitored. Linear breeding range varied from 2.2 km to 10.7 km from the nest site (Fig. 1). Logistic re- gression models for each eagle fit the data well. 210 Harmata and Montopoli VoL. 35, No. 3 Figure 1 . Logistic regression curves of cumulative proportion of perches by distance from most recently active nest site for five adult Bald Eagles monitored along the Snake River, Wyoming, 1985-89. “H” or “L” before gender icon indicates high or low reproductive history of the associated breeding area (see text). i.e., all > 0.90 (Table 2). Two males and one female exhibited similar DIs but Sis differed among all eagles (Fig. 2). Two males and a female were associated with high productivity breeding ar- eas and one male and a female were associated with low productivity breeding areas (Table 2). DI of the Sheep Gulch male (low productivity breeding area) was twice as far as DIs of both the Cabin Creek and Schwabacher males’ (high pro- ductivity breeding areas) (Fig. 2) . DI of the Oxbow female (low productivity breeding area) was nearly three times farther than that of the Butler Creek Table 2. Logistic regression parameters of cumulative perch site distribution, Spatial Use Indicators, and productivity of five radio-tagged adult Bald Eagles monitored along the Snake River, Wyoming, 1985-89. Pfor all regressions < 0 001 . Bald Eagle (Gender) Po Pi Spatial Use Indicator^ Distance (m)2 Slope Productivity Butler Creek ( ? ) -1.4290 0.2341 0.95 610 0.0585 High (1.73)3 Oxbow ( $ ) -1.6987 0.0958 0.98 1774 0.0239 Low (0.58) Cabin Creek ( S ) -1.3122 0.0708 0.97 1854 0.0177 High (1.62) Sheep Gulch ( S ) -5.7644 0.1300 0.90 4434 0.0325 Low (0.11) Schwabacher ( 6 ) -0.7724 0.0389 0.90 1986 0.0097 High (1.23) * At 50th percentile (see text) . ^ From most recently active nest. Young per occupied breeding area recorded annually 1979-90. September 2001 Bai.d Eagle Spatial Use Anai.ysts 211 5000 r 4000 • 3000 . _ 2000 ■ Q 1000 ■ O ‘ LOW PRODUCTIVITY. HIGH i OXBOW SCHWABACHER SHEEP GULCH cT BALD EAGLE CABIN CRK cf BUTLER CRK 0.08 - 0.07 - 0.06 - 0.05 ’ 0.04 ■ - 0.03 (J) - 0.02 - 0.01 j 0.00 Figure 2. Confidence interval (95%) estimates of Distance Indicator (DI) and Slope Indicator (SI) derived by logistic regression of cumulative proportion of perches by distance from most recently active nest. Estimates plotted in relation to long-term productivity of the breeding area of radio-tagged adult Bald Eagles monitored along the Snake River, Wyoming, 1985-89. female (high productivity breeding area) (Table 2). Males associated with more highly productive breeding areas also had flatter Sis than the male associated with a breeding area of low productivity (Fig. 1). However, SI of the female associated with a highly productive breeding area was twice as steep as that of the female associated with a low productive breeding area (Table 2). Pairwise comparisons of median perch distance revealed similar relationships among eagle gender- productivity groups as did DIs, but differences were not as pronounced (cf., Fig. 2, 3). Median perch distance was greater for the highly productive fe- male than a female exhibiting low productivity {U = 4656.5, P < 0.005) but the low production male’s median perch distance was different (U = 3737, P < 0.001) only from the male with the high- est production (Fig. 3). While Sis were different among all eagles, IQRs of perch distances suggest- ed dispersion of perches around median distance was not substantially different among eagles, ex- cept for the Sheep Gulch Male (Fig. 3). Means and confidence intervals of perch distances mirrored gender-productivity relationships illustrated by DI results (cf., Fig. 2, 3). Discussion Monitoring effort affected the number of perch events but not number of sites detected, indicating virtually all sites (distance categories) within eagle ranges were detected. Increased monitoring prob- ably would have detected increased use of previ- ously recorded sites only, not expansion of the breeding range or changes in SUIs. SUIs of all Bald Eagles were different but simi- larities among groups were evident. Bald Eagles associated with more productive breeding areas tended to perch closer to the nest site (proximate DIs) than eagles of their respective gender in breeding areas with a history of low productivity. Breeding areas that permit more even distribution of foraging opportunities over a larger area for males (flat Sis) yet allow females to concentrate activities close to the nest (steep Sis) also may favor higher productivity. Because the male is the pri- mary provider during incubation and early to mid- nestling phases of the breeding cycle (Stalmaster 1987), localized, temporal (daily, seasonal) disrup- tions in resource patch availability (e.g., human recreational presence) would be less severe in breeding areas with more dispersed or larger re- source patches than in breeding areas where re- source patches were few and concentrated. Close proximity of the attendant female would facilitate quick access to the nest, permitting more effective, timely defense against predators (e.g., Corvids), sheltering from inclement weather, and frequent 212 Harmata and Montopoli VoL. 35, No. 3 (0 LU O q; L- lU o CO Cl I o a: m 0. 6000 r 5000 - 4000 - 3000 p: 2000 1000 LOW^ PRODUCTIVITY 0 9 t * ! HIGH • MEAN, 95% Cl □ MEDIAN, 25%-75% Quartiles SHEEP GULCH d' SCHWABACHER d" BALD EAGLE CABIN CRK d BUTLER CRK Figure 3. Measures of central tendency and dispersion of perch distance from the nest and associated long-term productivity of the breeding area recorded for five radio-tagged adult Bald Eagles monitored along the Snake River, Wyoming, 1985-89. feedings. Higher productivity would be manifest in pairs with increased female attentiveness and more effective provisioning by the male. Results imply more productive Bald Eagle breed- ing areas along rivers may require a larger primary use area (i.e., Zone II), at least along shorelines, than recommended in regional Bald Eagle man- agement plans (e.g., Montana Bald Eagle Working Group 1994, Greater Yellowstone Bald Eagle Work- ing Group 1996). Maintenance of reproductive performance of highly productive pairs may re- quire extension of Zone II up to 2 km from the active nest to include 50% of male perches and 4 km to include 75% of male perches (Fig. I). Hall et al. (1997) defined habitat quality as “the ability of the environment to provide (for) . . . pop- ulation persistence.’’ Comparative SUIs of resident adults therefore may be indirect measures of hab- itat quality within Bald Eagle breeding areas along rivers and predictive of relative productivity among pioneering pairs or among pairs with unknown re- productive history. Further, temporal changes in SUIs may be indicative of changing habitat quality. Analysis of use of linear habitat by Bald Eagles with parametric tests (means & confidence inter- vals) may be appropriate when data are few, ap- proximately normally distributed, and statistical testing is desired. However, results of such tests were inconsistent with DI results (cf. Fig. 2, 3). Me- dian analysis may be appropriate when data are sparse and no statistical testing for concentration is needed. When data are adequate {N> 25), com- parisons of SUIs derived from logistic regression analysis may be more descriptive of linear habitat use than simple comparisons of medians and means and their respective measures of variability. SUI analysis provided opportunities for objective statistical testing, possibly modeled underlying pro- cesses and was less labor intensive and costly than GIS analysis. Comparisons of SUIs derived from logistic re- gression may have applicability to other species as- sociated with linear habitats (e.g., escarpments, streams, power line and open space corridors) . However, relationships among DIs, Sis, long-term productivity, and habitat quality presented here are admittedly tenuous due to small sample size and require further investigation for confirmation. Acknowi .edgments Bob Oakleaf, Wyoming Game & Fish Dept., originally alerted us to the management value of the 50% level in our analysis and has applied the theory to management of Bald Eagles in Wyoming. Studies from which this pa- per was derived were funded by Wyoming Game & Fish Dept.; Univ. of Wyoming-National Park Service Coop. Research Unit; Resource Management Div., Grand Teton National Park; U.S. Bur. of Reclamation, Minadoka Proj- ect; U.S. Forest Service, Bridger-Teton National Forest, Montana State Univ. Research and Creativity Program, September 2001 Bald Eagle Spatial Use Analysis 213 University of Wyoming, Basic Research Grants; and Dept, of Veterans vAffairs. Radio trackers included A1 Bath, Matt Erickson, Mitch Finnegan, Barb Franklin, Karen Kozie, Carl Mitchell, Jeff Rautio, Pete Piatek, and M. Whitfield. We thank John Marzluff, Mike Morrison, and Marco Res- tani for valuable comments on earlier drafts of this man- uscript. Fiterature Cited Andries, A.M., H. Gulinck, and M. Herremans. 1994. Spatial modeling of the Barn Owl Tyto alba habitat using landscape characteristics derived from SPOT data. Ecography 17:278-287. Bishop, Y, S. Feinberg, and P. Holland. 1975. Discrete multivariate analysis: theory and practice. Massachu- setts Institute of Technology Press. Cambridge, MA U.S.A. Bogel, R., W. d’Oleire-Oltmanns, and H. Franz. 1995. An integrated system for resource inventory, wildlife monitoring, and management using GIS, GPS, and ADF-telemetry techniques. Pages 551-555 m J.A. Bis- sonette and P.R. Krausman [Eds.], Integrating people and wildlife for a sustainable future. Proceedings of the first International Wildlife Management Con- gress. The Wildlife Society, Bethesda, MD U.S.A. Bortolotti, G.R. 1984. Criteria for determining age and sex of nestling Bald Eagles. J. Field OrnithoL 55:524— 542. Buchanan, J.T. 1997. A spatial analysis of the burrowing owl {Speotyto cunicularia) population in Santa Clara County, California, using a geographic information system. Pages. 90-96 in J.R. Duncan, D.H. Johnson, and T.H. Nicholls [Eds.], Biology and conservation of owls of the northern hemisphere; second internation- al symposium. February 5-9, 1997. Winnipeg, Mani- toba, Canada. U.S. Department of Agriculture. Forest Service. North Central Forest Experiment Station, St. Paul, MN U.S.A. Cox, D.R. 1972. Regression models and life tables. J. K Stat. Soc. Series B. 34:187-220. Garcelon, D.K., M.S. Martell, P.T. Redig, and F.C. Buoen. 1985. Morphometric, karyotypic, and laparo- scopic techniques for determining sex in Bald Eagles. J. Wildl. Manage. 49:595-599. Greater Yellowslone Bald Eagle Working Group. 1996. Bald Eagle management plan for Greater Yel- lowstone: 1995 Update. Greater Yellowstone Bald Ea- gle Working Group. Wyoming Game & Fish Depart- ment, Lander WY U.S.A. Harmata, a. and B. Oakleaf. 1992. Bald Eagles of the Greater Yellowstone Ecosystem: An ecological study with emphasis on the Snake River, Wyoming. Wyo- ming Game 8c Fish Department, Cheyenne, WY U.S.A. Harmata, A.R., G.J. Montopoli, B. Oakleae, P.J. Har- mata, and M. Restani. 1999. Movements and survival of Bald Eagles banded in the Greater Yellowstone Eco- system./. Wildl. Manage. 63:781-793. Hall, L.S., P.R. Krausman, and M.L, Morrison. 1997. The habitat concept and a plea for standard termi- nology. Wildl. Soc. Bull. 25:173-182. Hosmer, D.W. and S. Lemeshow. 1989. Applied logistic regression. John Wiley 8c Sons. New York, NY U.S.A. Kie, J.G., A. Baldwin, and J.C. Evans. 1996. CALHOME a program for estimating home ranges. Wildl. Soc Bull. 24:342-344. Marzluff, J.M., S.T Knick, M.S. Vekasy, L.S. Shueck, AND T.J. Zarriello. 1997. Spatial use and habitat se- lection of golden eagles in southwestern Idaho. Auk 114:673-687. Mohr, C.O. 1947. Table of equivalent populations of North American small mammals. Am. Midi. Nat. 37. 233-249. Montana Bald Eagle Working Group. 1994. Montana Bald Eagle management plan July 1994. USDI, Bu- reau of Reclamation, Montana Projects Office, Bill- ings, MT U.S.A. Montana Department of Fish, Wildlife & Parks. 1984. GEOSCAN users manual. Montana Department of Fish, Wildlife 8c Parks Research Bureau, Box 5, Boze- man, MT U.S.A. Montgomery, D.C. and E.A. Peck. 1982. Introduction to linear regression analysis. John Wiley and Sons. New York, NY U.S.A. Sprunt, a. IV, W.B. Robertson Jr., S. Postupalsky, RJ Hensel, C.E. Knoder, and F.L. Ligas. 1973. Compar- ative productivity of six Bald Eagle populations. Trans N. Am. Wildl. Nat. Resour. Conf. 38:96-105. Stalmaster, M.V. 1987. The Bald Eagle. Universe Books New York, NY U.S.A. StatSoft, Inc. 1999. STATISTICA for Windows [Com- puter program manual]. StatSoft, Inc., Tulsa, OK U.S.A. Swenson, J.E., K.L. Alt, and R.L. Eng. 1986. The ecology of the Bald Eagle in the Greater Yellowstone Ecosys- tem. Wildl. Monogr. 95:1—46. Worden, BJ- 1989. Kernel method for estimating the uti- lization distribution in home range studies. Ecology 70; 164-168. Received 13 May 2000; accepted 13 April 2001 J. Raptor Res. 35(3):214-220 © 2001 The Raptor Research Foundation, Inc. HABITAT USE, POPULATION DENSITY, AND HOME RANGE OE ELE OWLS (MICRATHENE WHITNEYI) AT SANTA ANA NATIONAL WILDLIFE REFUGE, TEXAS Christopher M. Gamely and Timothy Brush Department of Biology, University of Texas— Pan American, Edinburg, TX 78539 U.S.A. Abstract, — We collected data on the habitat use, home range size, and population density of the Elf Owl {Micrathene whitneyi idonea) in the Santa Ana National Wildlife refuge (SANWR) Lower Rio Grande Valley, Texas. Eourteen nocturnal surveys in 1995 and 1996 indicated that Elf Owls used chaparral habitat (92%) more than riparian woodlands (8%). Habitats used had high foliage densities (greatest density = 2-5-3.0 m) with partial canopies (x = 3.8 ± 0.36 m [±SE] and semi-open understories (greatest density <1 m). Unused chaparral habitat lacked high canopy coverage and had a denser understory, while riparian woodlands had greater canopy heights (x = 5.25 m) and open understories. Home range size determined by radio telemetry averaged 1.05 ± 0.33 ha (range = 0.24—2.60, N — 9). We estimated the maximum potential population size in SANWR to be 802 Elf Owls, assuming a home range size of 1.05 ha per breeding pair and saturation of preferred habitat. Key Words: Elf Owl; Micrathene whitneyi; home range, habitat selection; chaparral; Santa Ana National Wildlife Refuge, Texas. Uso de habitat, densidad poblacional y range de hogar de Micrathene whitneyi en el Refugio Nacional de Vida Silvestre de Santa Ana, Texas Resumen. — Colectamos dates sobre el uso de habitat, tamaho del range de hogar y densidad poblacional de Micrathene whitneyi idonea en el Refugio Nacional de Vida Silvestre de Santa Ana en el valle bajo del Rio Grande en Texas. (RNVSSA) Catorce monitoreos nocturnos hechos en 1995 y 1996 indicaron que Micrathene whitneyi utilizo el habitat de chaparral (92%) mas que los bosques riberehos (8%). Los habitat utilizados tenian densidades de follaje mayores (mayores densidades = 2 5-3.0 m, con doseles parciales (x = 3.8 ± 0.36 m ± SE) y con sotobosques semiabiertos (con una densidad de < 1 m) . El habitat de chaparral no utilizado no tuvo una cobertura de dosel alta con una densidad mayor en el sotobosque, mientras que los bosques riberehos tuvieron una mayor altura del dosel (x = 5.25 m) y un sotobosque abierto. El tamaho del rango de hogar fue determinado por el promedio de los resultados de telemetria 1.05 ± 0.33 (rango = 0.24-2.60 ha, N = 9) . Estimamos un tamaho potencial maximo de poblacion en el RNVSSA de 802 buhos enanos, asumiendo un rango de hogar de 1.05 ha por pareja reproductiva y una saturacion del habitat preferido. [Traduccion de Gesar Marquez] The Elf Owl {Micrathene whitneyi) extends from the southwestern United States to southern Mexi- co. The subspecies M. w. whitneyi is the most exten- sively studied of the four subspecies breeding in southern Arizona (Ligon 1968, Goad and Mannan 1987), extreme southeastern California (Rosen- berg et al. 1991), western New Mexico, the Big Bend region of Texas (Van Tyne and Sutton 1937), and into western Mexico (Ligon 1968, Stacey et al. ^ Present address: Marine Mammal Physiology Laborato- ry, 5007 Ave. U, Texas A &: M University, Fort Crockett, Galveston, TX 77551 U.S.A. 1983). M. w. idonea breeds along the Texas-Mexico border (Wauer 1971) and patchily in Tamaulipas (F. Gchlbach pers. comm.). M. w. whitneyi occurs in a variety of habitat types including evergreen woodlands, Sinaloan deciduous forests (Ligon 1968), and Sonoran desert in association with sa- guaro cactus {Cereus giganteus, Goad and Mannan 1987), Because the habitats in southern Texas dif- fer drastically, habitat requirements of M. w. idonea are probably different. The Elf Owl is a breeding resident in the Lower Rio Grande Valley (LRGV; Cameron, Willacy, Hi- dalgo, and Starr counties) of south Texas (Ober- holser 1974, Gehlbach 1987). The LRGV delta is 214 September 2001 Habitat Use of Elf Owls 215 the alluvial floodplain of the Rio Grande as it flows toward the Gulf of Mexico. Much of the vegetation occurring in the area is adapted to the annual flooding of the river system. Since its completion in 1953, Falcon Dam has affected the flora and fauna in the floodplain. Today, seasonal flooding of the Rio Grande rarely occurs and there has been a gradual change from riparian woodlands to chap- arral communities, locally known as Tamaulipan thornscrub (Vora 1990). The Elf Owl was first doc- umented in the LRGV in 1889 (Sennett 1889). For the next 70 yr, it was believed to be extirpated from the LRGV, until it was rediscovered in 1960 by James and Hayse (1963) who found it in the west- ern portion of the LRGV in Starr and western Hi- dalgo counties. Oberholser (1974) found two fledglings in Santa Ana National Wildlife Refuge (SANWR) in 1965 and Gehlbach (1987) found them nesting at the edge of evergreen forests in SANWR from 1973-78 suggesting that the species was spreading east. Since 1920 there has been an overall shift in the habitat composition at SANWR so that it now con- sists mainly of chaparral (Vora 1990). Today, ripar- ian woodlands cover only about 25% of SANWR (E. Hopson pers. comm.). The change in habitat has been directly attributed to the absence of sea- sonal flooding of the Rio Grande (Ramirez 1986, Vora 1990), but few studies have looked at subse- quent changes in the plant and bird communities. The objectives of this study were to describe pat- terns of habitat use, home range size, and popu- lation density of the M. w. idonea subspecies of the Elf Owl in the wildlife refuge. Study Area and Methods The study area consisted of 842 ha of native chaparral and other wooded habitat in southern Hidalgo County, TX. The refuge encompasses the largest remaining tracts of chaparral and riparian woodland in the LRGV (Vora 1990). A tour road, 10 km in length, and an extensive trail system cut through both vegetative types and provid- ed a route for surveying the area for Elf Owls. We conducted nocturnal surveys during the breeding seasons (late March-early August) of 1995 and 1996 to determine relative owl densities and second-order habitat use. The survey route had 7.26 km of chaparral (67%) and 3.74 km of riparian woodland (33%) in the 1995 field season. The survey route was increased in 1996 in- creasing the amount of chaparral and riparian woodland to 7.32 km (60%) and 4.88 km (40%), respectively. Sur- veys covered 68% of SANWR. Surveys were made on foot while listening for owl re- sponses to recorded playbacks. They were conducted once every three weeks to determine if there was seasonal variation in the density and habitat use of the owls. The direction of searches was reversed on four occasions to reduce potential biases. Along the routes, we broadcast tape-recorded calls of Elf Owls every 300 m and listened for 1 min for sponta- neously calling Elf Owls, following which the recorded call was played two more times for about 15 sec (6 calls) Each presentation of calls was followed by 1 min of lis- tening and a search of the immediate vicinity with a hand-held flashlight for owls. The estimated locations of all vocalizing owls were recorded on 7.5 min topographic maps of the area (Gamel 1997). We captured Elf Owls in mist nets. Two to four, 3 X 13 m mist nets were set up shortly after sunset in areas where Elf Owls had been heard calling. Continuous broadcasting of a recorded Elf Owl call was used to lure owls into the mist nets. Owls that did not appear to be overly stressed were fitted with 1. 2-1.4 g radio transmit- ters (wildlife Materials, Inc., Carbondale, IL U.S.A.). In 1995, four transmitters were attached with a backpack style cross-chest harness. In 1996, five transmitters were directly attached to the dorsal feathers along the spine, as per Warnock and Warnock (1993). Owl locations were determined by triangulation with a hand-held receiver (TRX-IOS, Wildlife Materials, Inc , Carbondale, IL U.S.A.) and a three element Yagi anten- na. Bearings were taken by a single individual, with read- ings taken from pre-established stations. Tracking usually began within one hour of sunset and ranged from 1-5 hr in length. Tracking was repeated several times over a month period until either 30 locations were recorded for each owl, the transmitter battery failed, or the owl left the area. Area/ observation curves indicated 30 locations were adequate to show movement patterns of the Elf Owls over a month-long portion of the breeding season Individuals were located at 30 min intervals and all bear- ings for a single location were collected within a 10-mm period to reduce error. Home range sizes were calculated with the TELEM88 program (Coleman 1989) using a 100% minimum convex polygon. No individuals were fol- lowed in both 1995 and 1996, as no owls were recaptured We identified three habitat categories: chaparral habitat utilized by Elf Owls (CWO), chaparral habitat not utilized by Elf Owls (CWNO), and riparian woodlands. An analysis of the vegetation in each habitat was conducted using a point-quarter sampling method (Brower et al. 1977). Sam- ple points were placed at 10-m intervals along a 100-m tran- sect. Within each habitat, three transects were established yielding a total of 33 sample points. In the CWNO and m the riparian woodlands, starting locations and directions for all transects were chosen randomly. Within the CWO, tran- sects were established within the home range of three ran- domly picked, radio-tracked owls. At each sample point, measurements were taken on the nearest woody species S2 m in height in each quadrant. For each individual woody species, point-to-plant distance, diameter at breast height (DBH) , and individual tree canopy coverage (at the widest point of coverage) were recorded. Canopy height was mea- sured and % canopy cover was estimated with a densiome- ter. Foliage density was quantified using a modified version of Mills et al. (1991) by marking a 3-m extending pole at 0.5 m intervals along the entire length. When fully extended and held at arm’s lengtli above the head, total height cov- 216 Gamel and Brush VoL. 35, No. 3 Table 1. Comparison of the mean structural composition of the vegetation among habitats in the Santa Ana National Wildlife Refuge, Texas. Characteristics included are tree density, individual tree cover, diameter at breast height (DBH), % canopy cover, and canopy height. An asterisk (*) between two habitats indicates a significant different (P < 0.05). Characteristic Riparian Woodland Chaparral with Owls Chaparrai. with NO Owls Tree density (m^) 0.32 + 0.10 0.41 + 0.07 0.53 + 0.09 Tree cover (m) 4.07 + 0.27 * 3.35 + 0.22 * 2.61 + 0.15 DBH (cm) 12.12 + 1.16 13.04 + 1.55 11.96 + 1.59 % Canopy cover 67.70 + 0.06 55.20 + 0.06 40.00 + 0.06 Canopy height (m) 4.96 + 0.28 * 3.81 + 0.36 2.82 + 0.30 ered was 5.5 m. At each sample point, the pole was erected vertically and the number of plant touches were recorded per 0.5 m segment. Records included touches by all vege- tative matter. We used SPSS-X (SPSS, Inc., Chicago, IL U.S.A.) for statistical analyses. Results were considered significant when P ^ 0.05. Means are accompanied by standard er- rors on all measurements. A chi-square test was used to determine differences in habitat use between the 1995 and 1996 field seasons. For each survey, the expected distribution of Elf Owls reflected the area covered in ri- parian woodlands and chaparral. A Sorensen’s coeffi- cient of community similarity was calculated to compare the similarity of species between different habitat types. A one-way analysis of variance (ANOVA) was used to ver- ify any observed differences between the three habitats for the variables quantified in the vegetation transects. In cases where a significant difference was found, a Tukey HSD test was used to identify which habitats differed sig- nificantly. In all cases, CWO was compared to both CWNO and riparian woodlands. Resui.ts We recorded a total of 145 Elf Owl locations. Significantly more locations were in chaparral {N — 134) than in riparian (N — 11) habitats in 1995 (X^ = 52.408, df = 1, P < 0.001) and 1996 (x^ = 51.946, df = 1, P < 0.001). The three transects within riparian woodlands included 13 tree species, of which anacua {Ehretia anacua) , sugar hackberry {Celtis laevigata), and cedar elm {Ulmus crassifolia) dominated. CWO included 12 species of trees, with la coma {Bumelia celastrina), Texas persimmon {Di- ospyros texana) , and spiny hackberry ( Celtis pallida) the dominant species. CWNO included 16 tree spe- cies, with the dominant species being spiny hack- berry, la coma, and guayacan ( Guaiacum angustifol- lum). Despite the difference in species dominance, there was overlap in species presence between the different habitat types. A Sorensen’s coefficient of community similarity showed riparian woodlands and CWO to be 67% similar, CWO and CWNO 79% similar, and riparian woodlands and CWNO 64% similar. Sorensen’s coefficients comparing the transects in the CWO to each other indicated with- in-habitat dissimilarity to be as great as 26%. Riparian woodlands had significantly greater canopy heights (P = 10.219, df = 2, P < 0.001; Table 1) and individual tree covers (P= 11.425, df — 2, P < 0.001) than CWO. Differences in tree density, DBH, and % canopy cover were not sig- nificant (P> 0.05). CWO had significantly greater individual tree cover (P = 11.425, df = 2, P < 0.001) than CWNO. Tree density, DBH, % canopy cover, and canopy height were not significantly dif- ferent (P > 0.05) . The vegetation in CWO was sig- nificantly denser at 2.50-3.00 m (P = 3.281, df = 2, P = 0.042; Fig. 1) than in riparian woodlands but, in riparian woodlands, the vegetation was sig- nificantly taller 5.00-5.50 m (P = 6.275, df = 2, P = 0.003). The vegetation in CWNO was denser than CWO (P = 8.955, df = 2, P < 0.001; Fig. 2) but the vegetation in CWO was significantly taller >5.50 m (P - 14.145, df - 2, P < 0.001). Ten Elf Owls were captured in 1995 and seven m 1996 over the course of nearly 52 hr of mist net trapping. Of these, nine Elf Owls were instrumented with radio transmitters (4 in 1995 and 5 in 1996). Home range size ranged from 0.24-2.60 ha (x = 1.05 ± 0.33 ha). There was no correlation between number of locations per owl and home range size (r = -0.292, P = 0.20) but an owl with 30 locations had the smallest home range (0.24 ha) . We did not find evidence of home range overlap. Discussion Elf Owls made significantly greater use of chap- arral habitat than riparian woodlands at SANWR. A total of 21 species of trees and shrubs were found along the nine vegetation transects and there was September 2001 Habitat Use of Elf Owls 217 (/) 0 ) 0 . 00 - 0.50 0 . 50 - 1.00 1 . 00 - 1.50 1 . 50 - 2.00 2 . 00 - 2.50 * 2 . 50 - 3.00 3 . 00 - 3.50 3 . 50 - 4.00 4 . 00 - 4.50 4 . 50 - 5.00 * 5 . 00 - 5.50 > 5.50 0 10 20 30 Riparian Woodlands Icwo Percentage of Total Canopy Volume Figure 1. Comparison of the understory vegetation density in riparian woodlands and chaparral habitats used by Elf Owls. Percent of total canopy volume is calculated from the mean number of the pole touches by vegetation for each of the three transects at different heights. An asterisk (*) next to the pole height indicates a significant difference m density at that level. A = 33 in riparian woodlands and chaparral habitats used by Elf Owls. a large degree of overlap in species among the transects. Still, Elf Owls occurred in one habitat more than the other. The diversity of habitats in which Elf Owls are known to reside (Ligon 1968, Schaeffer and Ehlers 1979, Goad and Mannan 1987) suggests that the presence, or absence, of particular woody plant species plays a role in hab- itat use. The suitability of individual tree species for ex- cavation by primary cavity nesters could also affect the distribution of secondary cavity nesters like the Elf Owl. This has been shown to be the case with M. w. whitneyi which tends to favor saguaro cacti as nest sites due to the abundance of Gila Woodpeck- er {Melanerpes uropygialis) and Gilded Flicker {Co- laptes chrysoides) excavations (Goad and Mannan 1987). A similar dependence on primary cavity nesters has been documented in Flammulated Owls {Otus flammeolus, McCallum and Gehlbach 1988), Eastern Screech-Owls {Otus asio, Belthoff and Ritchison 1990), and Northern Spotted Owls (Strix occidentalis occidentalis, Bias and Gutierrez 1992). While availability of cavities is often a limiting factor for secondary cavity nesters (Brawn and Bai- da 1988, Petty et al. 1994), such is likely not the case at SANWR, where primary cavity nesters are common (Carter 1986, Gehlbach 1994, Brush and Cantu 1998) and cavities (rot and woodpecker-ex- cavated) are abundant throughout all three vege- tative communities (T. Brush pers. obs.). An analysis of the structural features of each community showed canopy height, individual tree cover, and understory density were important in the use of habitats by Elf Owls. Canopy height dif- fered significantly between riparian woodlands and CWO with Elf Owls using the intermediate canopy heights found in CWO. The combination of a 218 Gamel and Brush VoL. 35, No. 3 (0 0 ) 0) * ir * 0.00-0.50 0.50-1.00 1.00- 1.50 1.50- 2.00 2.00- 2.50 2.50- 3.00 3.00- 3.50 3.50- 4.00 4.00- 4.50 4.50- 5.0C 5.00- 5.5C >5.50 CWNO CWO 0 10 20 30 Percentage of Total Canopy Volume Figure 2. Comparison of the understory vegetation density in chaparral habitat used by Elf Owls and chaparral habitat not used by Elf Owls. Percentage of total canopy volume is a calculated from the mean number of the pole touches by vegetation for each of the three transects at different heights. An asterisk (*) next to the pole height indicates a significant difference in density at that level. N = 33 in chaparral habitat used by Elf Owls and chaparral habitat not used by Elf Owls. higher canopy and large individual tree coverage causes a distinct, but partial, canopy to form in the CWO which is lacking in the CWNO. Elf Owls in SANWR use habitat that consists of a distinct, but partial, canopy layer at about 4 m and a semi-open understory which is most dense at 2.50—3.00 m. They do not use areas that have no understory or have very dense understories and very high or low canopies. Flammulated Owls (McCallum and Gehlbach 1988) and Boreal Owls {Aegolius funereus, Norberg 1970), are known to dive steeply when leaving a perch, then level off and fly 1-2 m above the ground. Similar behavior occurs in the Elf Owl (F. Gehlbach pers. com.), and may thereby explain why it does not use chap- arral habitat which contains a very dense understo- ry. At the same time, some understory appears nec- essary, possibly to provide protection from predators or to maintain appropriate habitat for prey items. This might explain why they do not use relatively-open riparian woodlands. Alternately, Gehlbach (1987) suggested that competition might occur between the Elf Owl and Eastern Screech- Owl, a resident of riparian woodlands at SANWR. Ligon (1968) mapped Elf Owl home ranges based on the locations of calling males estimated the size to be 0.3 ha. Our radio-telemetry data sup- ports the observation that Elf Owls occupy small home ranges. Although our home range size of 1.05 ha was larger, six of the nine owls tracked used areas <0.6 ha. Of the three that occupied larger home ranges, one moved back and forth between two patches of chaparral habitat that were separat- ed by a band of riparian woodland. While this ri- parian strip was included in the convex polygon, we observed no use of the area except as a corri- dor. Even the closest ecological equivalent owl spe- cies, the Flammulated Owl, maintains an average September 2001 Habitat Use of Elf Owls 219 home range size of 14.1 ha (Howie and Ritcey 1987), over 13 times as large. Prey abundance has been shown to strongly influence home range size of Eastern Screech-Owls (Belthoff et al. 1993) and Northern Spotted Owls (Carey et al. 1992). Eastern Screech-Owls in suburban settings maintain small home ranges of only 4-6 ha (Gehlbach 1994) com- pared to sizes ranging from 11.9—108 ha in other parts of North America (Johnsgard 1988, Belthoff et al. 1993, Gehlbach 1994). The small home range size we observed in Elf Owls may have been due to the insectivorous diet of the Elf Owls and the large numbers of insect prey in chaparral habitat. Several authors have indicated that Elf Owls are territorial (Ligon 1968, Johnsgard 1988). The data we collected supported the notion that this owl ac- tively defends territories. During nocturnal sur- veys, responding males tended to be spaced out rather then clumped close to each other. Also, no instrumented owls maintained overlapping home ranges. In contrast, both Ligon (1968) and Gehl- bach (pers. comm.) observed home range overlap m Elf Owls in Arizona. Based on our findings, we estimated the maxi- mum population of Elf Owls that can be supported in SANWR. In making this estimate, we assumed a home range size of 1 .05 ha, all of the available hab- itat is occupied, all home ranges are occupied by a breeding pair, and owls do not use habitat that is not suitable for them. Based on our estimation that there was 421 ha of suitable habitat in SANWR, we extrapolated that a maximum of 401 pairs, or 802 owls, could be supported in SANWR indicating that that current population is substan- tially smaller than the maximum that can be sup- ported. Wildlife management policies in the LRGV are currently focused on reducing the negative impact that halting of annual flooding has had on riparian woodlands. Recent programs at SANWR involving artificial flooding have the potential of halting, even reversing, the habitat transition from riparian woodlands to chaparral that has been underway since the completion of Falcon Dam in 1953. Some species have apparently benefitted from the in- creased availability of chaparral communities. The Elf Owl, in particular, has gone from a state of pos- sible near-extirpation to maintaining a substantial population since 1963. Acknowledgments We gratefully acknowledge E. Hopson, F. Judd, and R. Lonard for their guidance throughout this study. Field assistance was provided by N. Adame, D. Ake, C. Castillo, C. Rupert, J. Rupert, V. Varela, and R. Zamora. G. Proud- foot provided technical advice. We also thank C. Crocker- Bedford, F. Gehlbach, S. Henry, R.W. Mannan, and P Stacey for suggestions on an earlier draft of the manu- script. Funds were provided by the University of Texas- Pan American Faculty Research Grant number 119123 All captures and banding were conducted under federal bird-banding permit number 22544. Literature Cited Bias, M.A. and R.J. Gutierrez. 1992. Habitat associations of California spotted owls in the central Sierra Neva- da. J! Wildl. Manage. 56:584—595. Belthoff, J.R. and G. Ritchison. 1990. Nest-site selec- tion by Eastern Screech-Owls in central Kentucky. Condor 92:982-990. , E.J. Sparks, and G. Ritchison. 1993. Home rang- es of adult and juvenile Eastern Screech-Owls: size, seasonal variation and extent of overlap. J. Raptor Res 27:8-15. Brawn, J.D. and R.P. Baida. 1988. Population biology of cavity nesters in northern Arizona: do nest sites limit breeding densities? Condor 90:61-7 1 . Brower, J.E., J.H. Zar, and C.N. von Ende. 1977. Field and laboratory methods for general ecology. William C. Brown Publishers, Dubuque, lA U.S.A. Brush, T. and A. Cantu. 1998. Changes in the breeding community of subtropical evergreen forest in the Lower Rio Grande Valley of Texas, 1970s-1990s. Texas J. Sci. 50:123-132. Carey, A.B., S.P. Horton, and B.L. Biswell. 1992. North- ern Spotted Owls: influence of prey base and land- scape character. Ecol. Monogr. 62:223—250. Carter, M.D. 1986. The parasitic behavior of the Bronzed Cowbird in South Texas. Condor SS:ll-25 Coleman, J.S. 1989. Program documentation to TF T E M- a computer system for analyzing radio-telemetry data. Virginia Polytechnic Institute and State University, Blacksburg, VA U.S.A. Gamel, C.M. 1997. Habitat selection, population density, and home range of the Elf Owl, Micrathene whitneyi, at Santa Ana National Wildlife Refuge, Texas. M.S. the- sis, Univ. Texas-Pan American, Edinburg, TX U.S.A Gehlbach, F.R. 1987. Natural history sketches, densities, and biomass of breeding birds in evergreen forests of the Rio Grande, Texas and Rio Corona, Tamaulipas, Mexico. Texas J. Sci. 39:241-251. . 1994. The Eastern Screech-Owl: life history, ecol- ogy, and behavior in the suburbs and countryside Texas A&M Univ. Press, College Station, TX U.S.A Goad, M.S. and R.W. Mannan. 1987. Nest site selection by Elf Owls in Saguaro National Monument, Arizona. Condor 89:659-662. Howie, R. and R. Ritcey. 1987. Distribution, habitat se- lection and densities of Flammulated Owls in British Columbia. Pages 249-254 in R.W Nero, R.J. Clark, R.J. Knapton, and R.H. Hamre [Eds.], Biology and con- 220 Gamel and Brush VoL. 35, No. 3 servation of northern forest owls: symposium pro- ceedings. USDA For. Ser. Gen. Tech. Rep. RM-142. James, P. and A. Hayse. 1963. Elf Owl rediscovered in Lower Rio Grande Delta of Texas. Wilson Bull. '75:179- 182. JoHNSGARD, P.A. 1988. North American Owls, Smithsoni- an Institution Press, Washington, DC, U.S.A. Ligon, J.D. 1968. The biology of the Elf Owl, Micrathene whitneyi. Museum of Zoology, Univ. Michigan, No. 136. McCallum, D.A. AND F.R. Gehlbach. 1988. Nest-site pref- erences of Flammulated Owls in western New Mexico. Con^7or 90:653-661. Mills, G.S., J.B. Dunning, Jr., andJ.M. Bates. 1991. The relationship between breeding bird density and veg- etation volume. Wilson Bull. 103:468-479. Norberg, R.A. 1970. Hunting techniques of Tengmalm’s Owls Aegolius funereus. Ornis Scand. 1:51-64. Oberholser, H.C. 1974. The bird life of Texas. Univ. Texas Press, Austin, TX U.S.A. Petty, S J., G.S. Geoff, and D.I.K. Anderson. 1994. Value of nest boxes for population studies and conservation of owls in coniferous forests in Britain. J. Raptor Res. 28:134-142. Ramirez, P. 1986. Water development projects in the Rio Grande and their relationship to the Santa Ana and Rio Grande Valley National Wildlife Refuges, Texas. Spec. Rep. Santa Ana and Rio Grande Refuges, Ala- mo, TX U.S.A. Rosenberg, K.V., R.D. Ohmart, W.C. Hunter, and B.W. Anderson. 1991. Birds of the lower Colorado River valley. Univ. Arizona Press, Tucson, AZ U.S.A. Schaeffer, P. and S. Ehlers [Eds.]. 1979. Owls of the west: their ecology and conservation. National Audu- bon Society and Western Education Center, Tiburon, CA U.S.A Sennett, G.B. 1889. Micropallas whitneyi, Elf Owl, taken in Texas. Auk 6:276. Stacey, P.B., D. Rosetta, T. Edwards, and N. Joste. 1983. Northeastern extension of the breeding range of the Elf Owl in New Mexico. Southwest. Nat. 28:99- 100 . Van Tyne, J. and G.M. Sutton. 1937. The birds of Brew- ster county, Texas. Misc. Publ. Mus. Zool. Univ. Mich , No. 37. VoRA, R.S. 1990. Plant communities of the Santa Ana Na- tional Wildlife Refuge, Texas. Texas J. Sci. 42:115-128. Warnock, N. and S. Warnock. 1993. Attachment of ra- dio-transmitters to sandpipers: review and methods. Wader Study Group Bull. 70:28-30. Wauer, R.H. 1971. Ecological distribution of birds of the Chisos Mountains, Texas. Southwest. Nat. 16:1-29. Received 23 June 2000; accepted 20 April 2001 / Raptor Res. 35(3):221-227 © 2001 The Raptor Research Foundation, Inc. DIETS OF NORTHERN BARRED OWLS AND NORTHERN SPOTTED OWLS IN AN AREA OF SYMPATRY Thomas E. Hamer Hamer Environmental, 19997 Hwy. 9, ML Vernon, WA 98274 U.S.A. David L. Hays Washington Department of Wildlife, 600 N. Capitol Way, Olympia, WA 98501-1091 U.S.A. Clyde M. Senger^ Department of Biology, Western Washington University, Bellingham, WA 98225 U.S.A. Eric D. Forsman USD A Forest Serxnce, Pacific Northwest Research Station, 3200 Jefferson Way, Corvallis, OR 97331 U.S.A. Abstract. — We compared diets of Northern Barred Owls {Strix varia varia) and Northern Spotted Owls (Strix occidentalis caurina) in western Washington during 1985-89. Diets of both species were dominated by nocturnal mammals, but diets of Barred Owls included a more diverse and more even distribution of prey. Estimated dietary overlap between the two species based on the Pianka Index was 76%. Barred Owl diets included more terrestrial mammals, more birds, more diurnal prey, and more prey that were associated with riparian areas, including fish, amphibians, and snails. The snowshoe hare {Lepus ameri- canus) comprised 35% of prey biomass in the diet of Barred Owls. The diet of Spotted Owls was dom- inated by the northern flying squirrel {Glaucomys sabrinus), which comprised 51% of prey numbers and 57% of prey biomass. We speculate that Barred Owls and Spotted Owls compete for food because their diets overlap considerably, their food appears to be limiting in many years, and Barred Owls are gradually invading territories historically occupied by Spotted Owls. Key Words: Northern Barred Owl; Strix varia varia; Northern Spotted Owl; Strix occidentalis caurina; diet, competition; predation; Washington. Dietas de Strix varia varia y Strix occidentalis caurina en un area simpatrica. Resumen. — Comparamos las dietas de Strix varia varia y Strix occidentalis caurina en el oeste de Washington durante 1985-89. En las dietas de ambas especies predominaron los mamfferos nocturnos sin embargo la dieta de Strix varia fue mas diversa y con una distribucion mas uniforme de presas. Estimamos el traslape de las dietas entre las dos especies con base en el Indice de Pianka (76%). La dieta de Strix varia incluyo mas mamfferos terrestres, mas aves, mas presas diurnas y mas presas asociados con areas riberehas inclu- yendo peces, anfibios y caracoles. Lepus americanus constituyo el 35% de la biomasa de presas de la dieta de Strix varia. En la dieta Strix occidentalis predomino Glaucomys sabrinus con un 51% del mimero de presas y un 57% de la biomasa de presas. Especulamos que Strix varia y Strix occidentalis compiten por comida debido a que sus dietas se traslapan considerableraente, su alimento parece limitarse por ahos y debido a que Strix varia gradualmente esta invadiendo los territorios historicamente ocupados por Strix occidentalis. [Traduccion de Cesar Marquez] During the last century, the Northern Barred Owl {Strix varia varia) has gradually expanded its range westward across Canada and south into the Pacific Northwest and northern California (Grant 1966, Campbell 1973, Shea 1974, Taylor and Fors- ^ Present address: 1103 Yew St., Bellingham, WA 98225 U.S.A. man 1976, Boxall and Stepney 1982, American Or- nithologists’ Union 1983, Dunbar et al. 1991, Ham- er 1988, Hamer et al. 1994, Dark et al. 1998, Wright and Hayward 1998). As a result, the range of the Northern Barred Owl now almost completely over- laps the range of the Northern Spotted Owl (Stnx occidentalis caurina) (Dark et al. 1998, del Hoyo et al. 1999). In some areas in British Columbia and 221 222 Hamer et al. VoL. 35, No. 3 Washington state, Barred Owls are now so abun- dant they outnumber Spotted Owls (Hamer 1988, Dunbar et al. 1991). Although Barred Owls are generally believed to compete with Spotted Owls for space (Dunbar et al. 1991, Leskiw and Gutierrez 1998), the extent to which they also compete for food is unknown. Di- ets of Northern Spotted Owls have been described in many areas (Barrows 1980, Forsman et al. 1984, Richards 1989, Ward 1990), but no data have been published on diets of Northern Barred Owls in the area of range overlap with the Northern Spotted Owl. During radio-telemetry studies of Barred Owls and Spotted Owls in Washington in 1985-89, regurgitated pellets were collected in roosts and nest areas to determine composition of the diets of the owls. We analyzed the pellet samples, and herein describe and compare the diets of these two closely related species in an area of sympatry. Study Area and Methods We collected pellets from 12 Barred Owl terri- tories and 28 Spotted Owl territories in western Washington. All data from Barred Owls came from the 317 km^ study area surrounding Baker Lake on the west slope of the Cascade Range in northern Washington (Fig. 1). Most data from Spotted Owls (87% of prey items) came from 14 territories in the Baker Lake study area and nine territories in adjacent areas within the Mt. Baker-Snoqualmie National Forest (Fig. 1). Small numbers of prey from Spotted Owls were also included from four territories on the Gifford Pinchot National Forest in the southern Washington Cascades {N — 14), and one territory on the Olympic Peninsula (N ~ 9). Barred Owls and Spotted Owls are sympatric throughout this entire region. The Baker Lake study area was characterized by steep-sided valleys with elevations ranging from 244 m on the valley floor to 1800 m at the upper limits of the forested zone on the slopes of Mount Baker. Mean annual precipitation was 254 cm, most of which was rain during winter (Franklin and Dyr- ness 1973). The study area was largely forested, ex- cept for small areas of talus, meadow, marsh, and recent clearcuts. The dominant vegetation was for- ests of western hemlock {Tsuga heterophylla) and Douglas-fir {Pseudotsuga menziesii) at lower eleva- tions, and forests of mountain hemlock {T. merten- siana) and Pacific silver fir {Abies amabilis) at higher elevations (Franklin and Dyrness 1973). Forest age varied from young stands on recent clearcuts to forests >200 yr old. Diets were estimated primarily from prey re- mains in regurgitated pellets, but the sample also included a few freshly-killed prey remains found in roosts or nests. Pellets from Barred Owls were col- lected in 1985-89 and pellets from Spotted Owls were collected in 1986-89. Most pellets were col- lected during the spring and summer (April-Au- gust) , and were of recent origin as indicated by the fact that they had not been washed apart by rain or snow. Thus, our analysis primarily reflects the diet during spring and summer. Some of the owls that we studied were breeding, but the breeding status of many of the owls was not known in each year. We estimated the number of prey in pellets by counting skulls, jaws, or bones of the appendicular skeleton, whichever gave the highest count. Num- bers of insects were estimated from fragments of exoskeletons. Biomass was estimated by multiply- ing the estimated number of individuals of each species by the estimated mean mass of each spe- cies, or by individually estimating the mass of each prey based on comparisons of skeletal remains with specimens of known age and mass. Estimates of mean mass were obtained from a variety of sources, including Maser et al. (1981), Chapman and Feld- hamer (1982), Steenhof (1983), and Forsman et al. (1984) . All comparisons were based on the com- bined sample for all owls, because samples were too small to estimate average diets for individual territories. We used the modified Simpson’s Index (Odum 1975, Simpson 1949) to estimate dietary diversity, and the modified Hill Ratio (Hill 1973, Alatalo 1981) to estimate evenness of prey in the diet. These indices range from 0-1, with larger values indicating greater diversity or evenness. If all prey were taken in equal numbers, then dietary even- ness would be 1. We used Pianka’s Index (Pianka 1973) to compare dietary overlap; this index yields values from 0-1 (no overlap to complete overlap, respectively) . All estimates of dietary diversity, evenness and overlap were based on prey numbers. To evaluate differences in timing of foraging and habitats used for foraging, we grouped prey based on their primary period of activity (nocturnal, di- urnal, both), primary habitat association (forest, ri- parian, meadow, talus) , and primary behavior type (arboreal, semiarboreal, terrestrial/aquatic). We then used tests to compare the relative proper- September 2001 Diets of Barred and Spotted Owls 223 ^ Whatcom \ County Figure 1. Locations where pellets were collected from Northern Barred Owls and Northern Spotted Owls on the Mt. Baker-Snoqualmie National Forest (stippled area) in northwestern Washington, 1985—89. 224 Hamer et al. VoL. 35, No. 3 Table 1. Percent numbers (Num) and biomass (Bio) of prey in samples of pellets collected from Northern Barred Owls and Northern Spotted Owls in western Washington, 1985—89. Prey Barred Owls Spotted Owls LL % Num % Bio*^ % Num % Biot’ Mammals 202 76.1 74.5 285 96.2 98.6 Sorex spp. 26 9.8 0.4 11 3.7 0.2 Neurotrichus gibbsii 20 7.6 0.6 1 0.3 tr'^ Scapanus spp. 17 6.4 2.9 — — — Ochotona princeps — — — 9 3.0 4.8 Lepus americanus 22 8.3 35.0 9 3.0 13.4 Tamias townsendii 2 0.7 0.5 — — — Tamiasciurus douglasii 22 8.3 14.1 5 1.7 3.6 Glaucomys sabrinus 53 20.0 18.4 150 50.7 58.1 Thomomys talpoides — — — 1 0.3 0.3 Peromyscus maniculatus 18 6.8 1.2 61 20.6 4.5 Neotoma cinerea — — — 13 4.4 11.6 Clethrionomys gapperi 3 1.1 0.2 20 6.8 1.6 Microtus spp. 2 0.7 0.2 — — — Microtus oregoni 16 6.0 0.9 2 0.7 0.1 Zapus trinotatus 1 0.4 0.1 1 0.3 0.1 Mustela erminea — — — 2 0.7 0.3 Birds 29 11.0 19.4 8 2.8 1.4 Bonasa umbellus 6 2.3 11.3 — — — Otus kennicottii 1 0.4 0.4 — — — Corvus spp. 3 1.1 4.2 — — — Unident, small bird 19 7.2 3.5 8 2.8 1.4 Miscellaneous 34 12.9 6.1 3 1.0 Fish (small salmonids) 7 2.6 4.7 — — — Amphibians (frogs) 15 5.7 1.4 — - — — Molluscs (snails) 2 0.8 — — — Insects 10 3.8 trc 3 1.0 tr^ Totals 265 100.0 100.0 296 100.0 100.0 “* Indicates number of individual prey identified in pellets. Total biomass was 32 745 g for Barred Owls and 29 154 g for Spotted Owls. ^ tr < 0.05%. tion of prey biomass in different categories. Each species was assigned to only one activity period, habitat association, and behavior type, with the ex- ception of snowshoe hares {Lepus americanus) which were split evenly between forest and riparian habitats. Although snowshoe hares were sometimes active during the day, we labeled them as nocturnal because they are most active at night (Keith 1964). Results We identified 265 prey items from 13 Barred Owl territories and 296 prey items from 26 Spotted Owl territories (Table 1 ) . Diets of Barred Owls in- cluded ^24 species, and diets of Spotted Owls in- cluded 17 species. The sample of prey from Spot- ted Owls was dominated by mammals, which comprised 96.2% of prey numbers, and 98.6% of prey biomass. In contrast, Barred Owl diets includ- ed only 76.1% mammals, with the balance made up of birds, fish, frogs, snails, and insects. In terms of total biomass, the most important prey in diets of Spotted Owls and Barred Owls were northern flying squirrels ( Glaucomys sabrinus) and snowshoe hares, respectively. Mean mass of individual prey captured by Barred Owls and Spotted Owls was 123.6 and 98.5 g, respectively. Both species captured prey in a broad range of size categories, up to and including snowshoe hares weighing about 1000 g each. How- ever, diets of Spotted Owls were dominated by prey September 2001 Diets of Barred and Spotted Owls 225 60 1 -10 11 -40 41 -80 81 -160 161 -300 >300 Prey size classes based on body mass (g) Figure 2. Distribution of prey by size class in diets of Northern Spotted Owls and Northern Barred Owls in western Washington, 1985-89. in the 81-160 g range, whereas prey taken hy Barred Owls were more evenly distributed across all prey-size categories (Fig. 2). The diet of Barred Owls was considerably more diverse than the diet of Spotted Owls (Simpson In- dex = 0.917 vs. 0.699). The distribution of prey in the diet of Barred Owls was also more even than in the diet of Spotted Owls (Hill Index = 0.814 vs. 0.533). That is, Barred Owls captured many differ- ent kinds of prey in similar numbers, whereas Spot- ted Owls tended to concentrate on a few kinds of prey. The Pianka Index of dietary overlap between the two species was 0.76. Diets of both species were dominated by noctur- nal animals, but Barred Owl diets included more diurnal prey (Fig. 3, x ^2 ~ 64.7, P< 0.001). Barred Owl diets were also dominated by terrestrial spe- cies, whereas Spotted Owl diets were dominated hy arboreal or semiarboreal species, especially north- ern flying squirrels and bushy-tailed woodrats {Neo- toma cinered) (Fig. 3, Table 1, x ^2 “ 72.4, P < 0.001). Diets of both species were dominated by animals associated with forest habitats, but diets of Barred Owls included more species associated with riparian areas, swamps, or meadows, and fewer spe- cies associated with talus (Fig. 3, ~ 62.3, P < 0.001). Diets of Spotted Owls included 16.4% bio- mass from mammals associated with rock outcrops or talus, such as bushy-tailed woodrats and pikas ( Ochotona princeps ) , whereas none of the prey cap- tured by Barred Owls were typically associated with rock outcrops or talus (Fig. 3). Both-(-^r Diumai I % of prey biomass Figure 3. Diets (% of total biomass) of Northern Spot- ted Owls and Northern Barred Owls in western Washing- ton, subdivided based on habitat associations, behavior types, and primary activity periods of prey. Discussion Our results indicated that Northern Spotted Owls preyed on a fairly broad range of prey, but primarily focused on a few species of mammals. That is, the distribution of prey in the diet was very uneven. In contrast, the Barred Owl was more of a generalist, preying on a broader range of species at lower frequencies. This finding agrees with pre- vious studies in which diets of Northern Spotted Owls were generally dominated by a few types of arboreal or semiarboreal forest mammals (Barrows 1980, Forsman et al. 1984, 2001, Ward 1990), whereas diets of Barred Owls typically include a diverse mixture of prey (Bent 1938, Hodges 1947, Smith 1952, Sweeny 1959, Korschgen and Stuart 1972, Rhodes 1974). Although both species hunted primarily in for- ests, the composition of the diet suggested that Barred Owls made greater use of meadows and ri- parian areas than did Spotted Owls. During the study, we often observed Barred Owls perched at the edge of marshes and ponds, or directly above small streams (unpubl. data). A sample of pellets collected from two Barred Owl territories in western Montana included mostly microtines associated with meadows and riparian areas (Marks et al. 1984). Potential sources of variation in our data includ- ed differences among years, territories, and breed- 226 Hamer et al. VoL. 35, No. 3 ing status. We did not have enough data to stratify the samples and evaluate these effects. However, Spotted Owl territories in the Baker Lake study area often overlapped two or more Barred Owl ter- ritories and the habitat composition of Barred and Spotted Owl home ranges showed few differences (unpubl. data), indicating both species of owls were likely feeding on similar prey populations. Similarly, Herter and Hicks (2000) found consid- erable overlap in the kinds of forests occupied by Barred Owls and Spotted Owls in the central Cas- cades of Washington. Therefore, we believe that the differences we observed were primarily the re- sult of differences in prey selection as opposed to differences in prey availability. The high proportion of diurnal animals in Barred Owl pellets suggested that these owls were more active during the day than were Spotted Owls. This was confirmed from observations of ra- dio-marked Barred Owls on the Baker Lake study area, which moved, on average, 131 m/hr during the day and 260 m/hr at night (unpubl. data). Spotted Owls studied by Sovern et al. (1994) in the Washington Cascades only moved 20.6 m/hr on average during the day. Although the distribution of prey species in pel- lets of Barred Owls was more diverse and more even than in pellets of Spotted Owls, the 76% over- lap in the two samples suggested that the two spe- cies may compete for food, especially if prey be- comes limiting. That prey is limiting for Spotted Owls and many other species of owls in the north- ern hemisphere seems likely, given that they do not breed in many years (Southern 1970, Adamcik et al. 1978, Mikkola 1983, Forsman et al. 1984, 1996, Hayward and Carton 1988). Some previous comparisons of dietary composi- tion in owls have shown clear differences between species, but others have demonstrated consider- able overlap between some species (e.g., Korsch- gen and Stuart 1972, Marti 1974, Herrera and Her- aldo 1976, Hayward and Carton 1988). However, high dietary overlap does not necessarily prove that species compete for food because they can still avoid direct competition by foraging in different areas or at different times. Our data suggested some species-specihc differences in timing and lo- cation of foraging, but for the most part. Barred Owls and Spotted Owls were similar in that they were primarily nocturnal and foraged primarily in forests. Therefore, we suggest that the above men- tioned approaches to avoiding competition were not particularly effective in this case, especially in view of the fact that Barred Owls have recently in- vaded and taken up residence in areas traditionally occupied by Spotted Owls. Acknowledgments S. Seim with the USDA Forest Service, Pacific Northwest Research Station, helped collect pellets, trap owls, and track owls. This study was funded by the USDA Forest Service, Pacific Northwest Research Station. Literature Cited Adamcik, R.S., A.W. Todd, and L.B. Keith. 1978. De- mographic and dietary responses of Great FLorned Owls during a snowshoe hare fluctuation. Can. Field- Nat. 92:156-166. Alatalo, R.V. 1981. Problems in the measurement of evenness in ecology. Oikos 37:19-204. American Ornithologists’ Union. 1983. Check-list of North American birds. Allen Press, Inc., Lawrence, KS U.S.A. Barrows, C.C. 1980. Feeding ecology of the Spotted Owl in California. Raptor Res. 14:73-77. Bent, A.C. 1938. Life histories of North American birds of prey. Part 2. Smithsonian Inst., U.S. Natl. Mus. Bull 170. Boxall, PC. AND P.R. Stepney. 1982. The distribution and status of the Barred Owl in Alberta. Can. Field- Nat. 6:46-50. Campbell, R.W. 1973. Coastal records of the Barred Owl for British Columbia. Murrelet 54:25. Chapman, J. A. and G.A. Feldhamer. 1982. Wild mammals of North America. John Hopkins Press, Baltimore, MD U.S.A. Dark, S.J., RJ. Gutierrez, and G.I. Gould, Jr. 1998. The Barred Owl (Strix varia) invasion in California. Auk 115:50-56. DEL Hoyo, J., A. Elliott and J. Sargatal [Eds.], 1999 Handbook of the birds of the world, Vol. 5. Barn Owls to hummingbirds. Lynx Edicions, Barcelona, Spain. Dunbar, D.L., B.P. Booth, E.D. Forsman, A.E. Hether- INGTON, and D.J. Wilson. 1991. Status of the Spotted Owl, Strix occidentalis, and Barred Owl, Strix varia, m southwestern British Columbia. Can. Field-Nat. 105 464-468. Forsman, E.D., E.C. Mesi.ow, and H.M. Wight. 1984. Dis- tribution and biology of the Spotted Owl in Oregon. Wildl. Monogr. 87. , S. DeStefano, M.G. Rai^hael, and R.J. Gutierrez [Eds.], 1996. Demography of the Northern Spotted Owl. Stud. Avian Biol. 17. , LA. Otto, S.G. Sovern, M. Taylor, D.W. Hays, H. Ai.len, S.L. Roberts, and D.E. Seaman. 2001. Re- gional, local and temporal variation in the diet of the Spotted Owl in Washington./. Raptor Res. 35:141-150. Eranklin, J.F. AND C.T. Dyrness. 1973. Natural vegetation September 2001 Diets of Barred and Spotted Owls 227 of Oregon and Washington. USDA For. Serv. Gen. Tech. Rep. PNW-8, Portland, OR U.S.A. Grant, J. 1966. The Barred Owl in British Columbia. Murrelet 47:39-45 Hamer, TE. 1988. Home range size of the Northern Barred Owl and the Northern Spotted Owl in western Washington. M.S. thesis, W. Washington Univ., Bel- lingham, WA U.S.A. , E.D. Forsman, A.D. Fuchs, and M.L. Waiters. 1994. Hybridization between Barred and Spotted Owls. Auk 111:487-492. Hayward, G.D. and E.D. Garton. 1988. Resource parti- tioning among forest owls in the River of No Return Wilderness, Idaho. Oecologica 75:253-265. Herrera, C.M. and E. Heraldo. 1976. Food-niche and trophic relationships among European owls. Ornis Scand. 7:29-41. Herter, D.R. and L.L. Hicks. 2000. Barred Owl and Spot- ted Owl populations in central Washington./. Raptor Res. 34:279-286. Hill, M.O. 1973. Diversity and evenness: a unifying no- tation and its consequences. Ecology 54:427-432. Hodges, J. 1947. Barred Owl eating fish. Iowa Bird Life 17:36. Keith, L.B. 1964. Daily activity pattern of snowshoe hares. J. Mammal. 45:626-627. Korschgen, J.R. and H B. Stuart. 1972. Twenty years of avian predator-small mammal relationships in Missou- ri. /. Wild! Manage. 36:269-282. Leskiw, T. and R. j. Gutierrez. 1998. Possible predation of a Spotted Owl by a Barred Owl. West, Birds 29:225- 226. Marks, J.S., D.P. Hendricks, and V.S. Marks. 1984. Win- ter food habits of Barred Owls in western Montana. Murrelet 65:27-28. Marti, C.D. 1974. Feeding ecology of four sympatric owls. Condor 76:45— 61. Maser, C., B.R. Mate, J.F. Franklin, and C.T. Dyrness. 1981. Natural history of Oregon coast mammals. USDA For. Serv. Gen. Tech. Rep. PNW-133, Portland, OR U.S.A. Mikkola, H. 1983. Owls of Europe. Buteo Books, Ver- million, SD U.S.A. Odum, E.P. 1975. Ecology. Holt, Rinehart and Winston, New York, NY U.S.A. PiANKA, E.R. 1973. The structure of lizard communities. Annu. Rev. Ecol, Syst. 4:53-74. Rhodes, L. 1974. Largemouth bass caught by a Barred Owl. Migrant 45:68-69. Richards, J.E. 1989. Spotted Owl food habits and prey availability on the east slope of the Washington Cas- cades. M.S. thesis, Colorado State Univ., Fort Collins, CO U.S.A. Shea, D.S. 1974. Barred Owl records in western Montana Condor 76:222. Simpson, E.H. 1949. Measurement of diversity. Nature 163:688. Smith, E.R. 1952. Barred Owl preys upon trout. Chat 16 99. Southern, H.N. 1970. The natural control of a popula- tion of Tawny Owls {Strix aluco) . J. Zool. {Land.) 62. 197-285. SovERN, S.G., E.D. Eorsman, B.L. Bisweli., D.N. Rolph, AND M. Tayi.or. 1994. Diurnal behavior of the Spotted Owl in Washington. Corec^or 96:200-202. Steenhof, K. 1983. Prey weights for computing percent biomass in raptor diets. Raptor Res. 17:15-27. Sweeny, S. 1959. Barred Owl hshing. Chat 23:66. Taylor, A.L. and E.D. Forsman. 1976. Recent range ex- tensions of the Barred Owl in western North America, including the first records for Oregon. Condor. 78' 560-561. Ward, J.P. 1990. Spotted Owl reproduction, diet and prey abundance in northwest California. M.S. thesis, Hum- boldt State Univ., Areata, CA U.S.A. Wright, A.L. and G.D. Hayward. 1998. Barred Owl range expansion into the central Idaho wilderness./ Raptor Res. 32:77-81. Received 23 June 2000; accepted 6 April 2001 J Raptor Res. 35(3) :228-234 © 2001 The Raptor Research Foundation, Inc. FACTORS INFLUENCING LENGTH OF THE POST-FLEDGING PERIOD AND TIMING OE DISPERSAL IN BONELLFS EAGLE (HIERAAETUS FASCIATUS) IN SOUTHWESTERN SPAIN Eduardo Minguez^ Estacion Biologica de Donana (CSIC), Apdo. 1056, 41080 Sevilla, Spain Elena Angulo Estacion Biologica de Donana (CSIC), Apdo. 1056, 41080 Sevilla, Spain and Instituto de Investigacion en Recursos CAnegeticos, Apdo. 535, Ciudad Real, Spain Vivian Siebering^ Estacion Biologica de Donana (CSIC), Apdo. 1056, 41080 Sevilla, Spain Abstract. — ^We studied factors influencing the length of the post-fledging period (from fledging to the start of dispersal) of Bonelli’s Eagle {Hieraaetus fasciatus) nestlings in southwestern Spain, using 13 nesdings equipped with radiotransmitters. The age at fledging was negatively correlated with hatching date, but the duration of the post-fledging period was direcdy related to hatching date. This pattern could be explained by seasonal changes in prey abundance, especially that of wild rabbits {Oryctolagus cuniculus). Young in- creased their mobility throughout the post-fledging period, with a significant increase in the middle of the period. Dispersal began suddenly. The direction of dispersal was random, but most of the areas first used were located <25 km away and at lower altitudes than the nesting area. Key Words: Bonelli’s Eagle, Hieraaetus fasciatus; juvenile dispersal behavior, post-fledging period'. Red-legged partridge, wild rabbits', Oryctolagus cuniculus; Alectoris rufa. Factores que influyen en el periodo de post-emplumamiento y el comienzo de la dispersion en jovenes aguilas perdiceras Hieraaetus fasciatus. Resumen. — Estudiamos los factores que influyen en la duracion del periodo de post-emplumamiento, periodo comprendido entre el primer vuelo del joven hasta el comienzo de la dispersion, del aguila perdicera {Hieraaetus fasciatus) en el sudoeste de Espana, median te el radio-seguimiento de 13 aguilas jovenes. Pese a que hubo una correlacion negativa entre la fecha de eclosion y la duracion del periodo de emplumamiento, la duracion del periodo de post-emplumamiento estuvo relacionada directamente con la fecha de eclosion. Este patron podria ser explicado por los cambios estacionales en la abundancia de presas, especialmente de conejo silvestre {Oryctolagus cuniculus). Las jovenes aguilas fueron aumen- tando su movilidad a lo largo del periodo de post-emplumamiento, con un incremento drastico hacia la mitad de este periodo. El comienzo de la dispersion fue repentino. La direccion en la que ocurrio la dispersion fue al azar, pero la mayoria de las areas utilizadas por vez primera estaban situadas a <25 km y a menores altitudes que las areas de nidificacion. [Traduccion de autores] In most birds, young are dependent on their parents for some time after leaving the nest. For birds of prey, young usually stay within the natal ' Present address: Conselleria de Medio Ambiente Gen- eralitat Valenciana, Delegacion de Alicante, Cburruca 29, 03071 Alicante, Spain. ^ Present address: De Vlinderstichting, Postbus 506, 6700 AM Wageningen, Holland. territory until initiation of juvenile dispersal. Waser (1985) suggested that later dispersing individuals would probably have less competitive abilities. However, the length of the period from fledging to the start of dispersal, referred to as post-fledging period, and factors that influence the onset of dis- persal vary considerably (Donazar and Ceballos 1990, Ferrer 1992, Bustamante and Hiraldo 1993). The length of the post-fledging period might be 228 September 2001 Post-fledging in Bonelli’s Eagle 229 influenced by food availability, be a result of par- ent-offspring conflict (Trivers 1974), or a decision made by the young themselves. When food is abun- dant, young may stay longer within the natal ter- ritory (Walker 1988, Kennedy and Ward 1995) but, when food is scarce, they may leave earlier (Ken- ward et al. 1993). Some studies have suggested that when there is abundant food in the territory, young reach dispersal or migration condition soon- er and, therefore, leave the natal territory earlier (Bustamante 1994a, Wood et al. 1998). Factors influencing the timing of dispersal may be either environmental, endogenous, or a com- bination of both (Howard 1960). Young develop their flying and hunting skills prior to indepen- dence (Ferrer 1992, Real et al. 1998) and the en- dogenous component is reflected in increased ex- ploratory behavior near the time of dispersal (Holekamp 1986). Here, we evaluate the factors affecting the length of the post-fledging period in the Bonelli’s Eagle, especially the influence of food availability, and describe the first movements involved in ju- venile dispersal. For Bonelli’s Eagles, an Endan- gered Species (Tucker and Heath 1994), this in- formation is limited to two studies conducted by Morvan and Dobchies (1990) and Real et al. (1998). In order to discriminate whether young Bonelli’s Eagles leave the parental territory as soon as possible or whether environmental factors de- termine the length of the post-fledging period, we related territory quality and temporal food avail- ability with movement patterns of young. If individ- uals that disperse sooner have greater competitive abilities (Waser 1985), and environmental factors were not involved, we expected that young that fledged later or with good physical condition had a shorter post-fledging period. Likewise, young with good physical condition would move farther from nests at the end of the post-fledging period. Study Area and Methods We conducted the study in the provinces of Huelva, Cadiz, Malaga, and Sevilla, in southwestern Spain (Fig. 1). The study area is very heteroge- neous comprising mountains, lowlands, and plains. The more mountainous areas supported Mediter- ranean woodlands and scrublands, while agricul- tural fields were prevalent in the lowlands. We studied the post-fledging period of Bonelli’s Eagle, from fledging to start of dispersion. We an- alyzed the length of the post-fledging period in re- V f Mediterranean Sea Figure 1. Study area in southwestern Spain, 1998. Grey circles represent territories of Bonelli’s Eagles and black squares are the sites of prey availability counts. lation to food availability, body condition, hatching date, and fledging duration. During early February 1998, we searched known Bonelli’s Eagle territories for accessible nests oc- cupied by incubating adults. All nests were situated on cliffs, usually surrounded by Mediterranean woodlands {Quercus ilex and Quercus suber). The landscape in breeding areas was a mosaic of cleared woods, Mediterranean scrublands, and ol- ive orchards. After egg laying, nests were checked every 10-20 d to estimate hatching dates. We esti- mated age of young based on morphometric mea- surements (Torres et al. 1981). In May-June 1998, when nestlings were 40-50 d old, we instrumented 14 young from 10 broods with solar {N = 8) and battery-powered {N = 6) transmitters, attached with backpack harness (Kenward 1987). One solar transmitter failed after the young eagle was re- leased and was excluded from the analysis. Five un- marked siblings (three from a brood of two and two from a brood of three) flew from their nests during capture and were observed flying with their marked sibling during the post-fledging period. When radio-tagging birds (on average 12 d be- fore fledging, range = 1-22 d) , measurements and blood extraction were performed to estimate body condition. We assessed body condition as the resid- uals of a linear regression of body mass against forewing length (Reist 1985, Krebs and Singleton 1993, Jacob et al. 1996). We used the concentration of urea in the blood to evaluate nutritional con- dition (Ferrer 1990, 1993) by extracting 2 ml of 230 Minguez ET Al VoL. 35, No. 3 blood from the radial vein of each marked nes- tling. Blood samples were collected in lithium-hep- arin tubes and were centrifuged (15 min at 3000 rpm) <12 hr after samples were drawn. Both the cellular fraction and the plasma samples were fro- zen. Analyses were carried out four months later using a Hitachi 705 multichannel automatic ana- lyzer. To minimize circadian variations of the blood parameters, we extracted all blood samples be- tween 1300-1800 H, except for one brood with two young which were not included in the analysis be- cause blood was drawn later. We used the cellular fraction of the blood sample to sex young with primers 2945F, cfR and 3224R (Ellegren 1996), Of the 13 nestlings, seven were females and six were males. Therefore, our results were not biased by sex. We considered marked young to have fledged if they were observed flying or seen on a perch in- accessible from the nest. We searched for young in natal territories once every week using two vehicles equipped with roof-mounted, omnidirectional an- tennae to detect young and directional antennae to triangulate their loeation. Although the estimat- ed territory size of Bonelli’s Eagles in our study area averaged 3 km around nests (del Junco 1984, Gil et al. 1996, Minguez unpubl. data), we consid- ered young to have left their natal territories only when they roosted >4.5 km from their nests. This was the greatest distance we recorded young roost- ing and returning to nests a few days later. Dates of fledging and initiation of dispersal were consid- ered as the middle of the interval between two vis- its. Because visits were made once a week, we as- sumed an error of ±4 d. To determine if seasonal changes of prey avail- ability could affect fledging date and length of post-fledging period, we used data from a study on European rabbits {Oryctolagus cuniculus) and Red- legged Partridges {Alectoris rufa ) , the main prey of Bonelli’s Eagles (Real 1996, Ontiveros and Plegue- zuelos 2000). From January-November 1998, we conducted monthly surveys of rabbits and partridg- es in six different sites (Fig. 1). Vehicle surveys were conducted along permanent roadside tran- sects (10-14 km long) 1 hr before dusk at 10 km/ hr (Villafuerte et al. 1993). Monthly indexes of rab- bit and partridge abundance were computed for the entire study area as the mean number of rab- bits and partridges seen per kilometer in each area. Prey abundance for each young eagle was considered as the mean of the monthly index of prey abundance for the months of the fledging or post-fledging period. Thus, an average of rabbit and partridge abundance for the months during which fledging or post-fledging occurred corre- sponded to each young. Although these surveyed sites were not within Bonelli’s Eagle territories, we related the post-fledging period with an average of prey abundance of all the surveyed sites, thus rep- resenting the general seasonal pattern of prey availability throughout the study area. For the analyses, we used mean values of the length of fledging and post-fledging period for each brood, except for analyses regarding dispersal movements, when we considered young as inde- pendent observations. Because sample sizes were small, we used nonparametric tests (Norusis 1992). Spearman correlation coefficients were used to test the relationship between the length of fledging and post-fledging period with the different factors considered in the study (body condition, hatching date, fledging duration, and prey abundance) and to test the relationship between body condition and dispersal distance. We used comparisons be- tween means for circular statistics to analyze the direction of dispersal from nests to the first known roosts, as defined by Batschelet (1981, cited in Up- ton and Fingleton 1989). We analyzed differences in the distance moved throughout the post-fledg- ing period (dividing this period in groups of 20 d) with Levene’s test for equality of variances, Wilcox- on paired tests and Friedman tests. Statistics were condueted using SPSS software. Results Post-fledging Period. Young fledged between 15 May-22 June (x = 27 May) and were between 44— 69 d old at fledging (x = 59 ± 8.2 d, ±SD, N = 10). The post-fledging period ranged between 43- 131 d (x = 90 ± 23.4, N — 10). Age at fledging correlated negatively with hatching date (r^ = — 0.65, P = 0.04, N — 10). However, length of the post-fledging period correlated positively with hatching and fledging dates (r^ = 0.75, P — 0.01, V — 10 and = 0.79, P — 0.006, N = 10, respec- tively). No statistically significant relationship was found between age of fledging and availability of prey (r^ = —0.15, P = 0.67, N — 10 and r, = 0.15, P = 0.67, N — 10, for rabbits and Red-legged Par- tridges, respectively). Length of the post-fledging period correlated negatively with seasonal rabbit abundance and positively with seasonal partridge abundance (Fig. 2) . During the year of the study, September 2001 Post-fledging in Bonelli’s Eagle 231 o c (0 13 c 3 Xt < 40 60 80 100 120 140 Length of Post-fledging Period (days) Figure 2. Relationship between prey abundance index (average of monthly prey abundance for the months of the post-fledging period) and length of post-fledging pe- riod of Bonelli’s Eagles (number of days from fledging to the day that young roosted >4.5 km from the nest) in southwestern Spain, 1998. Grey circles represent rabbit abundance (r = —0.93, P < 0.001, N = 13) while black squares represent partridge abundance (r = 0.76, P = 0.003, N= 13). rabbit abundance reached its highest level in June and July and declined thereafter, while partridge abundance reached its highest level in late sum- mer. Peaks for maximum and minimum prey avail- ability were in the same months throughout the study area, although the level of abundance was different in each surveyed site (Fig. 3). This strong- ly suggested that all of the study area had similar patterns of prey availability. Neither urea concentration in the blood nor size corrected by body mass before fledging, showed a correlation with the length of the post-fledging pe- riod (Spearman correlation, = —0.15, P — 0.65, N = 11 and = —0.04, P = 0.89, N = 13, respec- tively) . During the post-fledging period, movements of young increased until about 80 d after fledging (Fig. 4, Levene’s test for equality of variances, F = 6.6, P = 0.001). The development of mobility in- creased throughout the post-fledging period (Friedman test, x ^4 — 14.84, P = 0.005, N = b). Mean distance young moved from their nests in- creased significantly at about the middle of the post-fledging period (Wilcoxon paired test, 40 vs. 60 d after fledging, Z = —2.20, P = 0.028, N = 8). Distance moved increased gradually prior to the middle of the post-fledging period (Wilcoxon paired test, 20 vs. 40 d after fledging, Z = —1.63, P = 0.10, N = 10) and remained the same there- after (Friedman test, comparing groups 60, 80, and 100 d after fledging = 1.14, P = 0.56, N = 7). 20 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Figure 3. Monthly prey abundance of rabbits (black bars) and Red-legged Partridges (grey bars) for the six sites surveyed during the post-fledging period. The Onset of Dispersal. All 13 instrumented young survived the post-fledging period, and the onset of dispersal occurred from 27 June-2 Octo- ber (x = 21 August ± 25 d). Young from broods of two young {N = 3) did not start dispersal at the same time but with a delay of 16, 16, and 38 d between siblings. The distance from the nest to the first-known roost site outside the territory averaged 33 km ± 25.7 (N — 13). However, the sample differed from a normal distribution and the median distance was 23.8 km (Fig. 5). Distance to the first-known roost outside the territory was not related to condition of young (urea levels or size corrected by body mass; Spearman correlation, — —0.07, P = 0.85, Figure 4. Mean distances to nest (±SE) moved by young Bonelli’s Eagles by age during the post-fledging period in southwestern Spain, 1998. Data obtained by di- rect observations and short distance triangulations. Num- ber of fledglings in parentheses. Sample sizes decrease by days as young started dispersal. 232 Minguez et al. VoL. 35, No. 3 o> 5 bi i E Distance of Dispersal (km) Figfure 5. Distribution of dispersal distances for 13 young Bonelli’s Eagles based on the first-known roost outside the natal territory in southwestern Spain, 1998. N = 11 and — —0.03, P = 0.92, N = 13, respec- tively) . Eleven of the 13 instrumented young dispersed to lower elevations. Mean elevation of the first- known roost sites was 281 m lower than nest sites (Wilcoxon paired test, Z = —2.56, P = 0.011, N = 13). Mean direction from the nest to first-known roosts outside the territory was 300.16° (var = 1.21). This direction did not differ from a uniform distribution as determined by the Rayleigh test (^RAY “ 0.13), modified by Wilkie (1983). Discussion The length of the post-fledging period of Bo- nelli’s Eagles was related to hatching date. This pe- riod was longer for young that hatched later, which contradicted results of studies on other species (Donazar and Ceballos 1990, Ferrer 1992). For some migratory species, length of the post-fledging period in later-fledged young might be reduced by the need to migrate (Bustamante and Hiraldo 1989, Donazar and Ceballos 1990, Bustamante 1994b). For a nonmigratory species such as the Spanish Imperial Eagle {Aquila adalberti), Ferrer (1992) suggested that in addition to physical con- dition of young, the physical condition of parents may determine the end of the post-fledging peri- od. Hatching date is generally considered to be a good indicator of territory quality and probably of high food availability (Korpimaki 1987, Cichon and Linden 1995). Therefore, it seems unlikely that young Bonelli’s Eagles that hatched early spent a shorter time in the natal territory as a con- sequence of food scarcity. The main prey item, both in numbers and biomass, for Bonelli's Eagles in Spain is rabbits followed by Red-legged Partridg- es (Leiva et al. 1994, Martinez et al. 1994, Real 1996, Ontiveros and Pleguezuelos 2000). As in oth- er Mediterranean ecosystems (Soriguer and Rogers 1981, Beltran 1991, Villafuerte et al. 1997), our re- sults showed that rabbits and partridges varied sea- sonally in abundance, with the highest abundance in late spring and late summer. These peaks in rab- bit and partridge abundance coincided with the post-fledging period and the initial onset of dis- persal of Bonelli’s Eagles. We found a strong negative correlation between temporal abundance of rabbits and length of the post-fledging period. We also found that young that hatched earlier took longer to fledge and had shorter post-fledging periods. This pattern may have been due to the fact that pairs that laid eggs early experienced low rabbit abundance during the brood-rearing period but high abundance dur- ing the post-fledging period. In contrast, pairs that laid eggs later in the season experienced high rab- bit abundance during the brood-rearing period, but low rabbit abundance during the post-fledging period. Scarcity of rabbits during the latter period could have resulted in young needing more time to reach the necessary body condition for dispers- al. Length of the post-fledging period was directly related to partridge abundance. Therefore, young that fledged later and experienced a scarcity of rabbits would have had access to greater numbers of partridges. These young might have switched from rabbits to partridges thus resulting in the same total length of their post-fledging period as young that fledged earlier. The differences we ob- served could be explained by the preferences of Bonelli’s Eagles for rabbits rather than partridges (Jordano 1981, Leiva et al. 1994, Gil et al. 1994, Ontiveros and Pleguezuelos 2000) that were ob- served as uneaten remains at nests. The lower bio- mass of partridges could also have been a factor, but because we assumed the same temporal prey abundance for all sites, we did not know what the real availability of prey was in each territory. There is need for more information on the effects of prey availability and the importance of each prey spe- cies on Bonelli’s Eagles. Young Bonelli’s Eagles spent most of their time within 3000 m of their nests during the post-fledg- ing period, which is within the estimates of the September 2001 POST-FLEDGlNG IN BONELLl’S EaGLE 233 mean radius of the natal territory (del Junco 1984, Gil et al. 1996). Distances moved by young from nests increased with age. In the middle of the post- fledging period, these distances increased signifi- cantly. Spanish Imperial Eagles exhibit a similar pattern with a significant increase in mobility mid- way during the post-fledging period when soaring flight starts to occur (Ferrer 1992). The increase in distance moved from the nest as young aged, and the fact that young suddenly be- gan to disperse, was caused by an increase in hunt- ing effort by the young, as shown by Real et al. (1998), rather than by exploratory behavior or a behavior caused by the parents. Further, we ob- served no young returning to their natal territory during the first days of dispersal. Siblings seemed to leave territories in an independent manner, sug- gesting the existence of an endogenous factor to start dispersal. However, our data suggested that the time of dispersal was determined mainly by rabbit availability. Acknowitdgments We thank many people for their collaboration in the fieldwork, especially C. Aguilar, C. Fernandez, S. Moran, R. Fernandez, and J. Ayala. G. Janss and E. Revilla were indispensable for their support. G. Pino arranged a house in the field. F. Redo performed plasma analysis in the Valme Hospital in Sevilla. Special thanks go to R. Villafuerte and G. Jordan. We thank F.M. Jaksic, J.E. Jimenez, and an anonymous referee for their useful com- ments. This study was directed by M. Ferrer and it was supported by the regional Government (Junta de Anda- luda). V. Siebering was funded by the Wageningen Uni- versity, while E. Minguez and E. Angulo were funded by the Estacion Biologica de Dofiana (CSIC) . Literature Cited Batschelet, E. 1981. Circular statistics in biology. Aca- demic Press, London, U.K. BeitrAn, J.P. 1991. Temporal abundance pattern of the wild rabbit in Donana, SW Spain. Mammalia 55:591- 599. Bustamante, J. 1994a. Family break-up in Black and Red Kites Milvus migrans and M. milvus-. is time of inde- pendence an offspring decision? Ibis 136:176-184. . 1994b. Behavior of colonial Common Kestrels (Falco tinnunculus) during the post-fledging depen- dence period in southwestern Spain./. Raptor Res. 28: 79-83. AND F. Hiraldo. 1989. Post-fledging dependence period and maturation of flight skills in the Black Kite Milvus migrans. Bird Study 36:199—204. AND . 1993. The function of aggressive chases by breeding Black and Red Kites Milvus mig- rans and M. milvus during the post-fledging depen- dence period. Ibis 135:139-147. CiCHON, M. and M. Linden. 1995. The timing of breed- ing and offspring size in Great Tits Parus major. Ibis 137:364-370. del Junco, O. 1984. Estudio sobre una poblacion de aguilas perdiceras {Hieraaetus fasciatus). Rapinyaires Mediterranis 11:80—85. Donazar, J.A. and O. Ceballos. 1990. Post-fledging de- pendence period and development of flight and for- aging behaviour in the Egyptian Vulture Neophron perc- nopterus. Ardea 78:387-394. Ellegren, H. 1996. First gene on the avian W chromo- some (CHD) provides a tag for universal sexing of non-ratite birds. Proc. R. Soc. Lond. B. 263:1635-1641 Ferrer, M. 1990. Hematological studies in birds. Condor 92:1085-1086. . 1992. Regulation of the period of postfledging dependence in the Spanish Imperial Eagle Aquila ad- alberti. Ibis 134:128—133. . 1993. Ontogeny of dispersal distances in young Spanish Imperial Eagles. Behav. Ecol. Sociobiol. 32:259- 263. Gil, J.M., F.M. Mouno, and G. Valenzuela. 1994. Para- metros reproductivos y alimentacion del aguila real {Aquila chrysaetos) y del aguila perdicera {Hieraaetus fasciatus) en la Provincia de Granada. Aegypius 12:47- 51. , F.M. Moling, and G. Vaienzuela. 1996. Selec- cion de habitat de nidificacion por el aguila perdicera {Hieraaetus fasciatus) en Granada (SE de Espana). Ar- deola 43:189—197. Holekamp, K.E. 1986. Proximal causes of natal dispersal in Belding's ground squirrels {Spermophilus beldingi) Ecol. Monogr. 56:365-391. Howard, W.E. 1960. Innate and environmental dispersal of individual vertebrates. Am. Midi. Nat. 63:152-161 Jacob, E.M., S.D. Marshall, and G.W. Uetz. 1996. Esti- mating fitness: a comparison of body condition indi- ces. Oikos 77:61-67. Jordano, P. 1981. Relaciones interespecificas y coexisten- cia entre el aguila real {Aquila chrysaetos) y el aguila perdicera {Hieraaetus fasciatus) en Sierra Morena cen- tral. Ardeola 28:67-88. Kennedy, P.L. and J.M. Ward. 1995. Postfledging move- ments of the Northern Goshawk in northcentral New Mexico. J. Raptor Res. 29:43. Kenward, R.E. 1987. Wildlife radio tagging. Academic Press, London, U.K. , V. Marcstrom, and M. Karlbom. 1993. Post-nes- tling behaviour in goshawks, Accipiter gentilis: the caus- es of dispersal. Anim. Behav. 46:365-370. Korpimaki, E. 1987. Timing of breeding of Tengmalm’s Owl Aegolius funereus in relation to vole dynamics m western Einland. Ibis 129:58-68. Krebs, J.R. and G.R. Singleton. 1993. Indices of condi- tion for small mammals. Aust. J Zool. 41:317-323. 234 Minguez et al. VoL. 35, No. 3 Leiva, a., G. Pareja, and J. Aragones. 1994. Alimenta- cion del aguila perdicera {Hieradetus fasciatus) en la Provincia de Cordoba. Aegypius 12:15-21. Martinez, J.E., M.A. Sanchez, D. Carmona, and J.A. Sanchez. 1994. Regime alimentaire de I’Aigle de Bo- nelli Hieraatus fasciatus durant la periode de I’elevage des jeunes (Murcia, Espagne). Alauda 62:53-58. Morvan, R. and F. Dobchies. 1990. Dependence de jeu- nes aigles de Bonelli Hieraaetus fasciatus, apres 1’ envoi: variations individuelles. Alauda 58:150-162. Norusis, M.J. 1992. SPSS for Windows. Base system user’s guide. Release 5.0, Chicago, IL U.S.A. Ontiveros, D. and J.M. Pleguezuelos. 2000. Influence of prey densities in the distribution and breeding suc- cess of Bonelli’s Eagles {Hieraaetus fasciatus): manage- ment implications. Biol. Conserv. 93:19-25. Real, J. 1996. Biases in diet study methods in the Bo- nelli’s Eagle. J. Wildl. Manage. 60:632-638. , S. Manosa, and j. Codina. 1998. Post-nestling dependence period in the Bonelli’s Eagle Hieraaetus fasciatus. Ornis Fenn. 75:1—9. Reist, J.D. 1985. An empirical evaluation of several uni- variate methods that adjust for size variation in mor- phometric data. Can. J. Zool. 62:1429-1439. Soriguer, R.C. and P.M. Rogers. 1981. The European wild rabbit in Mediterranean Spain. Pages 600-613 in K. Myers and C.D. Macinnes [Eds.], Proceedings of the world lagomorph conference. University of Guelph, Guelph, Ontario, Canada. Torres, J.A., P. Jordano, and A. Leon. 1981. Aves de presa diurnas de la provincia de Cordoba. Publica- ciones del Monte de Piedad y Caja de Ahorros de Cordoba, Madrid, Spain. Trtvers, R.L. 1974. Parent-offspring conflict. Am. Zool 14:249-264. Tucker, G.M. and M.E Heath. 1994. Birds in Europe, their conservation status. Cambridge Univ. Press, Cambridge, U.K. Upton, G. and B. Fingleton. 1989. Spatial data analysis by example. Vol. 2. Categorical and directional data John Wiley and Sons Ltd., Chichester, U.K. Villafuerte R., M.B. Kufner, M. Delibes, and S. More- no. 1993. Environmental factors influencing the sea- sonal daily activity of the European rabbit {Oryctolagus cuniculus) in a Mediterranean area. Mammalia 347. , A. Lazo, and S. Moreno. 1997. Influence of food abundance and quality on rabbit fluctuations: conser- vation and management implications in Dohana Na- tional Park (SW Spain). Rev. Ecol. (Terre Vie) 52:345- 356. Walker, D.G. 1988. The behaviour and movements of a juvenile Golden Eagle Aquila chrysaetos in England m 1986. Ibis 130:564-565. Waser, P.M. 1985. Does competition drive dispersal? Ecol- ogy 66:1171-1175. Wilkie, D, 1983. Rayleigh test for randomness of circular data. Appl. Stat. 32:311-312. Wood, P.B., M.W. Collopy, and C.M. Sekerak. 1998 Postfledging nest dependence period for Bald Eagles in Florida. J. Wildl. Manage. 62:333-339. Received 27 July 2000; accepted 7 April 2001 J Raptor Res. 35(3):235-239 © 2001 The Raptor Research Foundation, Inc. DISPERSION, HABITAT USE, HUNTING BEHAVIOR, VOCALIZATIONS, AND CONSERVATION STATUS OF THE NEW GUINEA HARPY EAGLE {HARPYOPSIS NOVAEGUINEAE) Mark Watson^ The Peregrine Fund, 566 W. Flying Hawk Lane, Boise, ID 83709 U.S.A. Smith Asoyama Herowana Village, c/o Research and Conservation Foundation, Box 1261, Goroka, Eastern Highlands Province, Papua New Guinea Abstract. — ^We studied the dispersion, habitat use, hunting behavior, vocalizations, and conservation status of the New Guinea Harpy Eagle {Harpyopsis novaeguineae) from December 1998-October 1999 in Crater Mountain Wildlife Management Area (CMWMA), Eastern Highlands Province, Papua New Guinea. Based on territory mapping, we estimated that the mean home range size was 13.0 ± 3.9 km^ (±SD, N= 5). One pair we followed for 42 d over a 4 mo period used an area of only 0.25 km^. We followed the male hunting in this area for 6 d (510 min). A small sample of prey items included ground-dwelling species such as forest wallaby {Dorcopsulus sp.), juvenile Dwarf Cassowary (Casuarius bennetti), New Guinea Megapode {Megapodius decollatus) , and an arboreal marsupial. Eagles called mainly during daylight hours, mosdy near sunup. Spec- trogram analysis indicated there were two main calls. A continuous, low frequency, far-carrying call that was used to advertise territories and for contact between mates over distances <2 km and a higher frequency, chicken-like call that was used in interactions between individuals that were close to each otlier and during hunting, perhaps as a stimulus or lure for prey. In contrast to the rest of the Highlands, eagles were protected inside CMWMA under agreements between villagers and international conservation organizations. Key Words; New Guinea Harpy Eagle, Harpyopsis novaeguineae; Papua New Guinea; rainforest, vocalizations. Dispersion, uso de habitat comportamiento de caza, vocalizaciones y estado de conservacion del aguila Harpia de Nueva Guinea {Harpyopsis novaeguinea) Resumen. — Estudiamos la dispersion, uso de habitat, comportamiento de caza, vocalizaciones y estado de conservacion del aguila harpia de Nueva Guinea {Harpyopsis novaeguineae) desde Diciembre de 1998 — Oc- tubre de 1999 en Area de Manejo de Vida Silvestre de Crater Mountain (AMVSCM) Provincia de Eastern Highlands, Papua Nueva Guinea. Con base en un mapa del territorio, estimamos el tamano promedio del rango del hogar: 13.0 ± 3.9 km^ (±SD, N = 5). Una pareja seguida por 42 dias en un periodo de 4 meses utilizo un area de 0.25 km^, seguimos al macho cazando en esta area por 6 dias (510 minutos). Una muestra pequeha de items incluyo especies del sotobosque como wallaby de bosque {Dorcopsulus spp.), casuarius enanos juveniles ( Casuarius bennetti) , megapodos de Nueva Guinea {Megapodius decollatus) , y un marsupial arboneo. Las aguilas vocalizaron principalmente durante el dia, hacia el amanecer. Los analisis del espectro- grama indicaron que hubo dos vocalizaciones principales. Una continua, con baja frecuencia, que se podia escuchar lejos utilizada para marcar el territorio y contactar las parejas a distancias de <2 km y una con una frecuencia alta, parecida a la de una gallina que fue utilizada en interacciones entre individuos que estaban cerca el uno del otro durante la caza, quizas utilizada como estimulo o como senuelo para las presas. En contraste al resto de las Highlands, las aguilas estaban protegidas dentro de AMVSCM bajo acuerdos entre los pobladores y organizaciones internacionales de conservacion. [Traduccion de Cesar Marquez] The New Guinea Harpy Eagle {Harpyopsis novae- guineae) is a poorly-known, forest eagle endemic to ^ Present address: The Game Conservancy Trust, Eord- ingbridge, Hampshire, SP6 lEH, U.K. Papua New Guinea and Irian Jaya. It is widespread but uncommon throughout undisturbed forests but only two short notes have been published on its ecology, one on its vocalizations (Shulz 1987), and one on its hunting behavior (Beehler et al. 235 236 Watson and Asoyama VoL. 35, No. 3 1992). Little is known about range-restricted rap- tors in tropical forests, yet they are among the most threatened species and habitats in the world (Bild- stein et al. 1998). The need for more information on New Guinea Harpy Eagle led us to undertake this study. Study Area and Methods We studied New Guinea Harpy Eagles from De- cember 1998-October 1999 in Papua New Guinea. The main study area was in Crater Mountain Wild- life Management Area (CMWMA), which extends for approximately 2700 km^ on the south side of the New Guinea Cordillera, in Chimbu and East- ern Highlands Provinces (06°40'S, 145°00'E). CMWMA is approximately 85% undisturbed mon- tane forest and 15% villages and cultivation. Of the forested area, 60% is used for hunting of bush- meat. During our study, there was no seasonal pat- tern of rainfall. The area has been the subject of combined conservadon and sustainable develop- ment initiatives since 1994 (Johnson 1997). Field- work was carried out in forest owned by Gimi vil- lagers at elevations between 200-3000 m. Additional field trips were made to the upper Jimi Valley (05'’34'S, 144°39'E) and the northeast edge of the Kubor Range (05°53'S, 144°22'E) in Western Highlands Province and Mount Giluwe (06°02'S, 144°00'E) in Southern Highlands Province. There, we interviewed people from Imbongu, Melpa, and Jiwaka groups who had different customs and atti- tudes toward the eagle than the Gimi. This allowed us to assess the impact of hunting on eagle num- bers across regions. We searched for eagles in all suitable habitats. Eagles were usually located by their distinctive, far- carrying calls and were then observed and fol- lowed for as long as possible. Locations of eagles were derived using compass bearings from known points established with a GPS unit and an estimate of the distance to eagle perches. When one eagle was followed or seen several times on the same day, this was counted as one sighting. These data were then plotted with basic topographical information using ArcView 3.0 (ESRI Inc. 1994). Recordings of vocalizations were made using a Sony TC-D5 Pro II and Sennheiser ME 66 micro- phone. Spectrograms were made to illustrate each type of call using Canary 1.5.1 software (Cornell Laboratory of Ornithology 1997). All spectrograms used a sampling rate of 12 kHz on a Hanning win- dow of 256 pts with 75% overlap. Legend □ call ■ sighting Figure 1 . (A) Dispersion of New Guinea Harpy Eagles in Crater Mountain Wildlife Management Area based on sightings and calls. (B) Location of study area in Papua New Guinea. Prey remains were collected from a recently used nest and beneath perches. Prey identification was based on comparisons vdth skins and skeletons in the University of Papua New Guinea, Port Moresby, Papua New Guinea. A 3-yr-old female eagle in the raptor collection of The Rainforest Habitat Centre, Lae, Papua New Guinea was measured and weighed to calculate its wing loading (Kerlinger 1989). This was compared to the wing loading of other species of eagles (Brown 1976). Results and Discussion Dispersion. We only found New Guinea Harpy Eagles in CMWMA. Over 212 d, we heard eagles calling on 120 occasions and actually observed them on 24 occasions for 1002 min. Based on clus- tering of points where eagles were seen or heard calling (Fig. 1), we estimated that the area con- September 2001 New Guinea Harpy Eagle 237 Table 1. Prey species of the New Guinea Harpy Eagle in Crater Mountain Wildlife Management Area {N = 10). Common Name Scientific Name Number How Identified Forest wallaby Dorcopsulus sp. 6 bones in nest, pellet Ringtail possum Pseudocheiridae 1 observation Dwarf Cassowary Casuarius bennetti 1 bones in and under nest New Guinea Megapode Megapodius decollatus 1 bones in and under nest Fruit-Dove Ptilinopus sp. 1 bones in nest tained a minimum of 5 pairs of eagles with an av- erage home range of 13.0 ± 3.9 km^ (±SD, N = 5) . The habitat was not continuous in GMWMA be- cause of areas of cultivation around villages, areas where suitable prey had been hunted out, and ar- eas above approximately 2800 m where scrub re- placed forest. Accordingly, we estimated the overall density to be one pair per 150 km^ or a total of 10-20 pairs in the GMWMA. Habitat Use. For one pair of eagles, all sightings and vocalizations were within a 0.25 km^ area. They were followed for 42 d in May, July, August, and September. During this period, they were not de- tected for 8 d which coincided with a prolonged period of heavy rain. It was not clear whether the eagles were still in the area and stayed silent or had flown to a different area beyond hearing range. It was remarkable that this pair used such a small area for such a long period. The male eagle used the area to hunt and was seen carrying a ringtail possum (Pseudocheiridae) at 1610 H on 2 September 1999. Hunting Behavior. Prey items identified from one pellet and other prey remains were mostly for- est wallaby {Dorcopsulus sp.) which agreed with de- scriptions of the diet given by indigenous people (Rand and Gilliard 1967, Majnep and Bulmer 1977; Table 1). Prey were probably taken both on the ground and within the forest canopy. The ring- tail possum we observed being carried during the day suggested that, like other nocturnal, arboreal species in this family, possums were taken from their roosting places during the day. We also ob- served eagles making systematic searches of suit- able roosting places for mammals in the crowns of trees. Seven hunters we interviewed described ea- gles flushing prey from epiphytes or holes by hang- ing from their legs and beating their wings against the vegetation. Although we did not observe this behavior, it has been described twice in the litera- ture, and is comparable to techniques used by Af- rican Harrier Hawks {Polyboroides typus) and Crane Hawks (Geranospiza caerulescens) (Majnep and Bul- mer 1977, Osborne and Osborne 1992). Pooling all 40 flights of the six different individ- ual eagles observed, only 4 were >100 m. Similar short hunting flights have been described as “short-stay perched-hunting” in Northern Gos- hawks {Accipiter gentilis) in more open habitats (Kenward (1982). However, one flight made by a female eagle was >1.5 km across a ravine system indicating that New Guinea Harpy Eagles can trav- el for long distances across the forest. We never observed eagles soaring, which was contrary to de- scriptions by earlier authors (Rand and Gilliard 1967, Brown and Amadon 1968, Diamond 1972, Peckover and Filewood 1976). The wing loading of the single captive eagle we measured was 0.91 g/ cm^ which was 1.3 times greater than values re- corded for other species of soaring eagles (Brown 1976) suggesting that it is unlikely that New Guinea Harpy Eagles soar. Soaring is unlikely to offer se- lective advantage in locating prey since the canopy restricts visibility from the air. However, the sym- patric but morphologically different Gurney’s Ea- gle {Aquila gurneyi) was seen soaring 80% of the time we observed it during our study. Unlike most raptors that use an aerial display flight in pair bonding and territory defense, the New Guinea Harpy Eagles appear to circumvent this by using an unusual repertoire of calls. Vocalizations. Eagles called mainly during the day (Fig. 2) and this agreed with the pattern of vocalizations described from September-Decem- ber 1986 on Mt. Missim, Morobe Province, Papua New Guinea (Shulz 1987). This finding suggested that most calling coincides with crepuscular and daytime activities, including hunting. The most fre- quently-heard call was a continuous, low frequency (<500 Hz) note that has been described as “like a plucked bowstring” (Fig. 3a, Diamond 1972). Two other calls had a much higher frequency (1400- 1600 Hz) and sounded like a variable “cbuck chuck” (Fig. 3b, 3c) . Both of these calls were heard 238 Watson and Asoyama VoL. 35, No. 3 30 1 3 5 7 9 11 13 15 17 19 21 23 UnBofdEy Figure 2. Frequency distribution of New Guinea Harpy Eagle calls by time of day, all individuals combined {N = 120 ). only from one pair with the first call given by the female and the second given by the male. A fourth call was a combination of the “plucked bowstring” and “chuck chuck” calls (Fig. 3d). The low frequency call was audible up to 2 km away. Low frequency, continuous calls propagate much more effectively through foliage, so this call may have been used as a territorial advertisement. Strategies of resource pardtioning and reproduc- tion differ between temperate and tropical birds. In the tropics, where resources are more stable, territories are often defended year-round and pair bonds are more permanent. As a result, both sexes call year-round to defend food resources and main- tain pair bonds (Moreton 1996). The high fre- quency call was only used by the male eagle when hunting. This call may possibly be used as a stim- ulus to flush prey from roosting places in the can- opy or to lure prey in a manner similar to that used by Northern Shrikes (Lanius excubitor, Atkinson 1997). Conservation Status. There was a sharp contrast in the attitudes of villagers toward the New Guinea Harpy Eagle between CMWMA and other areas in Southern and Western Highlands Provinces. In CMWMA, eagles were not hunted and were pro- tected under agreements linked to sustainable de- velopment initiatives made with conservation or- ganizations. However, in the other areas, eagles were still hunted for their feathers which are used as symbols of rank and for personal decoration at ceremonies. In one Melpa village, feathers of four eagles shot within the preceding 18 mo were dis- played and, in another two Imbongu villages, reli- able accounts were given of eagles being killed us- ing slingshots or shotguns. Fourteen hunters who 3 kHz 2 .! 0 200 400 600 800 0 200 400 600 800 ms 0 200 400 600 BOO ms 4 3 1 kHz 2 •< 0 400 800 1200 1600 2000 ms Figure 3. Spectrograms of four New Guinea Harpy Ea- gle calls from a minimum of 5 individuals: (A) N = 63; (B) N= 23; (C) N= 16; (D) N= 13. were interviewed at Mt. Giluwe, the upper Jimi Val- ley, and the Kubor Range reported that New Guin- ea Harpy Eagles were rare in their forest. A study of the use of bird plumes among indigenous cul- tures showed a decline in the frequency of feather trading (Healey 1990). While our results did not allow us to assess directly the effects of hunting on the status of the eagle populations in these areas, we feel that a reduction in hunting pressure and traditional resource extraction through conserva- tion agreements such as those already in place in CMWMA are essential to the conservation of the New Guinea Harpy Eagle. Acknowledgments We are grateful to The Peregrine Fund, Zoological So- ciety of San Diego, Chevron Company, and The Walt Dis- September 2001 New Guinea Harpy Eagle 239 ney Company Foimdation for generous financial sup- port. We also thank The Research and Conservation Foundation of Papua New Guinea and the staff of the Library of Natural Sounds, Cornell University. Literature Cited Atkinson, E.C. 1997. Singing for your supper: acoustical luring of avian prey by Northern Shrikes. Condor 99: 203-206. Beehler, B.M., W. Crill, B. Jefferies, and M. Jefferies. 1992. New Guinea Harpy Eagle attempts to capture a monitor lizard. Emu 92:246-247. Bii.dstein, K.L., W. Schelsky, J. Zalles, and S. Ellis. 1998. Conservation status of tropical raptors./. Raptor Res. 32:3-18. Brown, L. 1976. Eagles of the world. David and Charles, Newton Abbot, Devon, U.K. AND D. Amadon. 1968. Eagles, hawks and falcons of the World, Vol 2. Hamlyn, London, U.K. Cornell Laboratory of Ornithology. 1997. Canary acoustic analysis software, version 1.5.1. Cornell I^ab- oratory of Ornithology, Ithaca, NYU.S.A. Diamond, J.M. 1972. Avifauna of the eastern highlands of New Guinea. Nuttall Ornithological Club, Cam- bridge, MA U.S.A. ESRl Inc. 1994. Arcview, version 3.0. Esri Inc., Redlands, CA U.S.A. Healey, C. 1990. Maring hunters and traders: production and exchange in the Papua New Guinea Highlands. Univ. California Press, Berkeley, CA U.S.A. Johnson, A. 1997. Processes for effecting community par- ticipation in the establishment of protected areas: a case study of the Crater Mountain Wildlife Manage- ment Area. Pages 391-428 in C. Filer [Ed.], The po- litical economy of forest management in Papua New Guinea. National Research Institute, IIED, Hong Kong, China. Kenward, R.E. 1982. Goshawk hunting behavior and range size as a function of food and habitat availabil- ity. /. Anim. Ecol. 51:69-80. KerI-INGER, P. 1989. Flight strategies of migrating hawks Univ. Chicago Press, Chicago, IT U.S.A. Majnep, I.S. and R. Bulmer. 1977. Birds of my Kalarn Country. Auckland Univ. Press, Auckland, New Zea- land. Moreton, E.S. 1996. A comparison of vocal behavior among tropical and temperate passerine birds. Pages 258-268 in D.E. Kroodsma and E.H. Miller [Eds ] , Ecology and evolution of acoustic communication m birds. Cornell Univ. Press, Ithaca, NY U.S.A. Osborne, T. and L. Osborne. 1992. New Guinea Haipy Eagle and Gurney’s Eagle Aquila gurneyi. Aust. Raptor Assoc. News 13:14. Peckover, W.S. and L.W.C. Filewood. 1976. Birds of New Guinea and tropical Australia. Reed, Sydney, Aus- tralia. Rand, A.L. and E.T. Gilliard. 1967. Handbook of New Guinea birds. Wiedenfeld and Nicholson, London, U.K. Shulz, M. 1987. Temporal variation in the frequency of vocalisations of the New Guinea Harpy EaOe. Emu 87 257-258. Received 27 October 2000; accepted 19 April 2001 J Raptor Res. 35(3):240-246 © 2001 The Raptor Research Foundation, Inc. SEASONAL AND GEOGRAPHIC DIFFERENCES IN THE DIET OF THE BARN OWL IN AN AGRO-ECOSYSTEM IN NORTHERN ITALY Michela Bose and Franca Guidali Universitd degli Studi di Milano, Dipartimento di Biologia, Sezione Ecologia, via Celoria, 26-20133 Milano, Italy Abstract. — ^We studied the dietary niche breadth of the Barn Owl ( Tyto alba) in the Po plain of northern Italy. A total of 1266 pellets was collected during 2 yr in an agricultural area located between Mantova and Brescia. Bones of 4455 prey items were identified with small mammals, in particular rodents and Insectiv- ora, the most important dietary component and birds (Passeriformes) comprising most of the remainder of the diet. Arthropods were occasionally eaten and amphibians and bats were rare in the diet. Two families of rodents (Microtidae and Muridae) were the most important dietary components. Members of the order Insectivora were complementary prey, showing an increase in winter when voles declined. In summer, a significant increase of birds was seen. Dietary composition and niche overlap were tested on diets of two pairs of owls. One pair ate mostly voles and the other pair ate birds more frequently. Nevertheless, niche overlap was high because of the similarity between the nest sites used by the two pairs. Key Words: Barn Owl; Tyto alba; diet; seasonal variation; niche breadth. Diferencias estacionales y geograficas en la dieta de Tyto alba en un agroecosistemas del norte de Italia. Resumen. — Estudiamos la amplitud del nicho alimenticio de Tyto alba en el piano del Rio Po en el norte de Italia. Un total de 1266 egagropilas fueron recolectadas durante 2 anos en el area agricola localizada entre Mantova y Brescia. Los huesos de 4455 presas fueron identificados como pequenos mamiferos en particular los roedores e insectos fueron los componentes mas importantes es la dieta. Las aves (Pas- seriformes) constituyeron el resto de esta. Los artropodos fueron ocasionalmente ingeridos y los anfibios y murcielagos rara ve. Dos familias de roedores (Microtidae y Muridae) fueron los componentes mas importantes. Los miembros del orden Insectivora constituyeron una presa complementaria mostrando un incremento en el invierno cuando los roedores declinaron. En el verano hubo un incremento de aves. La composicion de la dieta y el traslape del nicho fueron probados en la dietas de dos parejas de lechuzas, una igirio mas que todo ratones, mientras que la otra comio aves con mas frecuencia. Sin embargo el traslape del nicho fue mayor debido a estas dos parejas. The Barn Owl {Tyto alba) is a widely-distributed owl with euryphagous feeding habits (Lovari et al. 1976, Contoli 1981, Galeotti 1992). Because 95% of the variability in its diet is due to environmental fac- tors (Spitz 1981), analysis of its diet can provide in- formation on the availability of small mammal prey species in a particular area, even when few pellets are available (Contoli 1981). Its specialization on mammal species decreases when the abundance of its main prey decreases (Herrera 1974) and it in- creases the diversity of its diet to new prey items of lower energy content. Consequently, changes in its diet can be used to reflect real changes in the small mammal fauna available to the owl (Marti 1986). Moreover, analysis of Barn Owl diets can be used as a tool in the research for the management and protection of important habitats. Indeed, small la similaridad de los sitios del nido utilizados por [Traduccion de Cesar Marquez] mammals may represent a biological indicator of biocoenosis in terrestrial ecosystems because they occupy various trophic levels, and then their pres- ence or absence gives information on degree of biotope alteration (Contoli 1975). Many studies have examined the diet of Barn Owls in Italy (Contoli 1980, 1981, Contoli et al. 1983, Torre 1983, 1987, Pandolfi and Santolini 1987, 1988, Boldreghini et al. 1988, Bigini and Tur- in! 1992, 1995), but few data are available on the diet of the species in the Po plain in northern Italy (Groppali 1987, Vicini and Malaguzzi 1988). The aim of our study was to provide new information on the niche breadth of the Barn Owl {Tyto alba) and to contribute to the knowledge of the pres- ence and the distribution of its prey species in an area in the Po plain. 240 September 2001 Barn Owl Diet in Northern Italy 241 Study Area and Methods The study was conducted in a 3500 ha area lo- cated between Mantova and Brescia. Approximate- ly 90-95% of the area was intensively cultivated with few patches of the original mixed woodland habitat (Querco-Carpinetum) remaining (Sestini 1963, Ingegnoli 1993, Pignatti 1994). Agricultural fields consisted of cereals, corn, sunflower (40%), vegetables and fodder (35%), and wheat (5%). The area also included the Chiese River and many canals, as well as woody areas, grasses, permanent pastures, buildings, and abandoned fields. The area was subdivided in three zones, 3-5 km apart, representing three different vegetation types. Several buildings such as farms or complexes of houses were also present. It contained two Barn Owl nest sites and we collected pellets at six col- lection sites (two silos and four hay lofts) in the two nest areas. Zone 1 (2 km^) was characterized by cultivated fields (corn, sunflower, wheat), waste herbaceous fields {Convolvulus arvensis, Calystegia septum, Malva spp.), Platanus acerifolia and Sambu- cus ebulus on the edges of the ditches, and small woodlots of Robinia pseudoacacia. Wetlands were abundant along the Chiese River, many ditches, mole drains, and an artificial lake. This zone con- tained one nest site and two roosts. Zone 2 (1 km^) was surrounded by sunflower fields, woodlands of Robinia pseudoacacia, Populus alba and Crataegus monogyna and included two roosts 300-400 m away from the river. Zone 3 (1 km^) had more homo- geneous vegetation with corn fields and meadows and it included one nest site. Zones 1 and 3, within the municipalities of Remedello and Acquafredda, differed in their vegetation. Zone 1 had grassland and pastures and the habitat was more varied. Zone 3 consisted almost completely of corn fields (ISTAT 1990). We visited the sites every 2 wk, at the middle and at the end of every month, from January 1993- December 1994. A total of 1266 pellets was col- lected. Pellets were analyzed using the dry method (Contoli 1980) and bones of prey species were identified by using dichotomous keys for mammals (Chaline et al. 1974, Pucek 1981, Amori et al. 1984, 1993) and birds (Moreno 1985, 1986, 1987). Prey were enumerated based on the minimum number of individuals (Southern 1954). Pelvis fragments were also used to estimate the quantity of prey, be- cause skulls were missing in some cases. A com- plete determination of avian remains was not al- ways possible because diagnostic bones were lost and skulls were fragmented. Biomass was calculat- ed based on average weights of prey obtained from the literature (Toschi and Lanza 1959, Toschi 1965, 1986, Brichetti 1976). For Rattus, we used the average mass given by Di Palma and Massa (1981). We estimated the percent frequency and percent biomass for each prey species based on the total number of specimens found and the biomass value of each specimen. To determine seasonal variation in the diet, we calculated the frequency of the fam- ilies and classes of prey in each season and tested the differences with a chi-square test using the av- erage frequency values between the two years. Fi- nally, we tested the correlation (Spearman corre- lation coefficient) between the seasonal trends of prey frequencies. To determine niche breadth, we first tested the relative importance of the prey categories and the diversity of the diet using a Kruuk and Parish’s di- agram (1981) . It is based on the comparison of the percent frequency on the X-axis and percent bio- mass on the Y-axis. The relation (X*Y/100) equaled the bulk of each prey category in the diet and all points with equal X*Y values were con- nected by a set of isopleths. For both frequency and biomass, the average values between the 2 yr were used. To have another measure of the trophic diversity, we also calculated the Shannon index (H') (Margalef 1968) and the Levins’ normalized (B) index (Levins 1968). When comparing diets, we considered only the samples collected at the two sites that were located 4200 m apart in two areas with different vegetation. A total of 460 pellets was collected at Site 1 {N = 1756) and 449 were collected at Site 2 {N — 1579). Niche overlap between the two nest sites was tested using a Percentage similarity index (Pso, Sehoener 1970) based on the equations: H' = -2 pi In A B = (1/R) 2 A" i Pso = 1 - 0 . 5(2 IAi “ A 2 I) where p-, = n-JN and R = dietary categories. Results and Discussion We identified a total of 4455 prey items with a total biomass of 91021.02 g (Table 1). Mammals (Rodentia, Insectivora, and Chiroptera) were the most important prey species in the diet (F = 90.46, B = 79 550.8 g, %B - 87.40). The two principal prey species (Sorex araneus and Apodemus sylvaticus) 242 Bose and Guidali VoL. 35, No. 3 Table 1. Number of specimens (N) and frequency (F) of prey in Barn Owl diets by season in the Po plain, Italy (pooled values for 1993 and 1994). Dec. -Feb. Mar.- -May June-Aug. Sept. -Nov. Prey N F (%) N F (%) N F (%) N F (%) Mammals Talpa europaea 0 0.00 1 0.20 2 0.08 2 0.27 Soricinae 125 14.99 78 15.73 212* 8.92* 106 14.17 Sorex araneus 118 14.15 69 13.91 175* 7.36* 91 12.17 Crocidurinae 249** 29.86** 128 25.81 484 20.36 148 19.79 Crocidura leucodon 84 10.07 41 8.27 226 9.51 69 9.22 Crocidura suaveolens 161** 19.30** 78 15.73 245 10.31 77 10.29 Microtidae 180* 21.58* 134 27.02 702 29.53 207 27.67 Microtus arvalis 119* 14.27* 102 20.56 483 20.32 138 18.45 Microtus savii 46 5.52 25 5.04 156 6.56 54 7.22 Muridae 196 23.50 113 22.78 631 26.55 196 26.20 Apodemus sylvaticus 125 14.99 76 15.32 456 19.18 118 15.78 Mus domesticus 9 1.08 6 1.21 30 1.26 13 1.74 Micromys minutus 33 3.96 20 4.03 81 3.41 33 4.41 Rattus sp. 14 1.68 3 0.60 15 0.63 13 1.74 Other mammals 64 7.67 50 10.05 235 9.88 74 9.89 Birds 61 7.31 23 4.64 252** 10.60** 58 7.75 Passer domesticus 31 3.72 13 2.62 4.80** 28 3.74 Passer montanus 8 0.96 1 0.20 25** 1.0.5** 11 1.47 Other birds 2 0.24 1 0.20 17 71 2 0.26 Unidentified birds 20 2.40 8 1.61 96** 4.04** 17 2.27 Rana sp. 0 0.00 1 0.20 0 0.00 0 0.00 Arthropoda 0 0.00 1 0.20 21 89 8 1.07 Frequencies lower than expected. Frequencies higher than expected. were not habitat specialists. However, some prey species such as Crocidura suaveolens, Crocidura leu- codon, Micromys minutus and Microtus arvalis were associated with open and cultivated habitats while others including Neomys fodiens, Neomys anomalus and Arvicola terrestris and Mus musculus and Rattus spp. were associated with canals and ditches and human habitations, respectively. Burrowing and ar- boreal species such as Talpa europaea and Muscar- dinus avellananus were taken rarely, while forest specialists such as Clethrionomys glareolus and Apo- demus flavicollis were not found in the diet. Sur- prisingly, we had a record of a Suncus etruscus, a species which is seldom reported in Lombardy (Ot- lolini and Aceto 1996). Our sample included species related not only to various ecosystems but also to various trophic levels with rodents being hrst-order consumers and shrews, moles, and rats being higher-level consum- ers. Moreover, we observed a high diversity and abundance of insectivorous mammals which indi- cated that the study area had not been affected by chemical fertilizers (Contoli 1981). Every species that occurred in the study area occurred in the diet indicating that Barn Owls had large hunting territories and had both crepuscular or nocturnal foraging habits. Birds represented a minor component in the diet (F - 8.84, B = 11 396.22 g, %B = 12.52). The principal prey species were Passer domesticus and Passer montanus which were very common in the study area. Arthropods (Classes Chilopoda and In- secta) and amphibians were only occasionally found. Barn Owls specialize in eating small mammals throughout their range and the variety of small prey taken appears to vary with availability and en- vironmental and geographical features. In North America, Clark and Bunck (1991) found an inverse correlation between shrew abundance and precip- itation. In wetter habitats, shrews were more com- mon in the diet while, in drier areas, rodents were September 2001 Barn Owl Diet in Northern Italy 243 more common. Unlike Herrera (1974) who found a low diversity of mammalian prey in the diet of Barn Owls in the southern Mediterranean, we found an inverse correlation between latitude and diet diversity. In other studies in Africa (Heim de Balsac 1965, Goodman 1986) and South America (Jaksic and Yahez 1979), rodents were the prin- cipal component of the diet of the Barn Owl. Apparently, the presence of shrews in the diet is variable and depends on environmental and geo- graphical features. Birds, amphibians, arthropods, fish, and bats have been reported as prey in some areas when preferred small mammalian prey spe- cies declined. Very narrow Barn Owl diets have been reported by Lenton (1984) for Malaysia and by Morton and Martin (1979) for Australia. In both cases, the Barn Owls colonized areas where the small mammal fauna consisted of few species such as Rattus spp. and Mus musculus. Mammals were the most important component of the diet in all seasons, always exceeding 80%. In spring, they reached their greatest frequency. There were significant seasonal differences of prey frequencies (x^i 5 = 113.79, P < 0.01). In winter, Crocidurinae, mainly Crocidura suaveolens, oc- curred significantly more often and Microtidae, mainly Microtus arvalis, occurred less than expect- ed. In summer, Soricinae decreased in frequency and birds, mostly Pawer spp., increased significant- ly. Other mammals found in the diet included Neo- mys fodiens, Neomys anomalus, Suncus etruscus, Micro- tus multiplex, Arvicola terrestris, Muscardinus avellanarius and Pipistrellus pipistrellus. Other birds found in <3% of the sample included Hirundo rus- tica, Riparia riparia, Melanocorypha calandra, Alauda arvensis, Sturnus vulgaris, Motacilla flava, Hippolais polyglotta, Carduelis carduelis and Carduelis chloris. The predation on small mammals that we ob- served may have been a reflection the reproductive cycles of the prey species (Torre 1983, Boldreghini et al. 1988, Bigini and Turini 1992, 1995). We re- corded the highest predation rates when juvenile dispersal occurred (February— March for Insectiv- ora, summer for Microtidae, and autumn for Mu- ridae). It is possible that the increase in summer vegetation may have also contributed to the de- cline in small-sized insectivores which became dif- ficult to find. Our analysis confirmed the inverse trends for frequency of Crocidurinae and Muridae (r = —0.84, P < 0.001). As for birds, our data con- firmed that their capture fluctuated in accordance to the availability of mammals. Frequency of prey categories Figure 1. Kruuk and Parish’s diagram of the estimated bulk of prey categories in the diet of Barn Owls in the Po plain, Italy. Isopleths connect points of equal relative bulk in the Barn Owl diet. Our analysis of niche breadth confirmed the eu- ryphagous habits and the wide trophic niche of the Barn Owl. The Kruuk and Parish diagram (Fig. I) revealed that Microtidae and Muridae were prin- cipal prey (hyperbola of 10%), while all the other prey categories had only a marginal role (hyper- bola of 1%) and none of them represented more than 50% of the diet. Both the indexes of niche breadth had high values (H' = 1.648, B = 0.58). The Shannon index was higher than previously re- ported by other authors in other regions (e.g., 1.13 [Lovari et al. 1976]; 1.37 [Amori and Pasqualucci 1987]; 1.39 [Petretti 1977]). Our index was was very similar to indexes of 1.67 reported by Grop- pali (1987) and 1.63 reported by Vicini and Mala- guzzi (1988) in the Po plain and to the average value of 1.7 calculated for Italy by Contoli (1988). These results confirmed the presence of a diversi- fied small mammal population with various trophic levels represented which could reflect a complex trophic relationship and an apparently mature and stable ecosystem (Margaleff 1975). At site 1, Microtidae, in particular Microtus ar- valis, were twice as abundant as in Site 2 and Mu- ridae and birds were more frequent in the diet than expected in Site 2 (Table 2). These differenc- es could have been due to the differences in veg- etation between the two sites. Site 1 was located in an area with various kinds of agriculture with fields and woodlands on the edges of canals. The area surrounding Site 2 had less habitat diversity which may explain the absence of Microtidae in the diet and the occurrence of Muridae in the diet. At Site 2, the decrease of Microtidae forced Barn Owls to 244 Bose and Guidali VoL. 35, No. 3 Table 2. Number of specimens (AO and frequency (F) of prey in the diets of Barn Owls at two nests sites (pooled values for 1993 and 1994), Site 1 Site 2 N F N F Mammals 1678* 95.56%* 1310* 82.96%* Insectivora 581 33.09% 523 33.12% Talpidae 1 0.06% 3 0.19% Soricinae 192 10.93% 170 10.77% Crocidurinae 376 21.41% 336 21.28% Rodentia 1096 62.41% 786 49.78% Microtidae 638* 36.33%* 303* 19.19%* Microtus arvalis 529* 30.13%* 142* 8.99%* Muridae 413* 23.52%* 449* 28.44%* Apodemus sylvaticus 283* 16.12%* 321* 20.33%* Gliridae 4 0.23% 1 0.06% Chiroptera 0 0.00% 1 0.06% Birds 65* 3.70%* 260* 16.47%* Passer domesticus 27* 1.54%* 125* 7.92%* Passer montanus 8* 0.46%* 25* 1.58%* Other birds 2 0.12% 15 0.94% Unidentified birds 28* 1.59%* 95* 6.02%* * Frequencies significantly different between the two sites. increase their predation on the other available spe- cies, in particular birds (x\ - 233.4, P < 0.01). The increase in predation on complementary prey categories was confirmed by both indexes of niche breadth, that were higher in Site 2 (H' = 1.686, B = 0.623) than in Site 1 (H' = 1.51, B = 0.29). The Percentage similarity index (Pso = 0.73) in- dicated a high niche overlap in the diets of Barn Owl at the two sites. This finding was expected be- cause the two sites had similar climatic, environ- mental, and anthropogenic features. The diversity of birds and Microtidae in the diet at both sites was a reflection of the adaptability of Barn Owls to dif- fering prey availabilities. A(..KN OWl.EDGMENTS We wish to thank D. Fontaneto and S. Scali for com- ments and suggestions on improving the manuscript and G Barbieri and S. Bandera for information concerning the study area. Lii’Erature Cited Amori, G., M. Cristaldi, and L. Contoli. 1984. Sui Rod- itori (Gliridae, Arvicolidae, Muridae) dell’Italia pen- insulare ed insulare in rapporto all’ambiente biocli- matico mediterraneo. Animalia. Catania. 11:217-269. AND F. Pasqualuccl 1987. Elementi di valutazio- ne ambientale di tre siti nellTtalia centrale, tramite I’analisi dei rigetti di Tyto alba. Acqua-Aria. 9:1085- 1089. , M. Angelici, S, Frugis, G. Gandolfi, R. Grop- PALi, B. Lanza, G. Relini, and G. Vicini. 1993. Check- list delle specie della fauna d’ltalia. 110. Vertebrata Ministero dell’Ambiente e Comitato Scientifico per la Fauna d’ltalia. Calderini, Italy. Bigini, I. and R. Turini. 1992. La micromammalofauna predata da Tyto alba (Scop. 1769) nella fascia costiera di S. Rossore (Pisa). Atti Soc. Toscana Sci. Nat. Mem., Ser. B. 99:61-74. and . 1995. Nicchia trofica del barbagianm Tyto alba (Scop. 1769) e impatto antropico: dati rela- tivi alia fascia planiziale del comprensorio dei Monti Pisani e al Subappennino Lucchese. Atti Mus. Civ. Stor Nat. Trieste. 46:133-143. Boldreguini, R, L. Casini, and R. Santolinl 1988. Dif- ferenze stazionali della dieta di Tyto alba nell’area del- le Valli di Comacchio. Naturalista Sicil. S. lY, XII (suppl.):155-158. Brichetti, P. 1976. Atlante ornitologico Italiano. Ediz. Flli Scalvi, Brescia, Italy. Chaline, J., H. Baudvin, D. Jammot, and M. Saint Gi- rons. 1974. Les proies des rapaces. Petits mammiferes et leur environnement. Doin, Paris, France. ClARK, D.R., Jr. and C.M. Bunck. 1991. Trends in North American small mammals found in Common Barn- Owl {Tyto alba) dietary studies. Can. J. Zool. 69:3093— 3102. September 2001 Barn Owl Diet in Northern Italy 245 CoNTOLl, L. 1975. Micro-mammals and environment in Central Italy: data from Tyto alba (Scop.) pellets. Boll. Zool. 42:223-229. . 1980. Borre di Strigiformi e ricerca teriologica in Italia. Natura e Montagna 3:73-94. . 1981. Ruolo dei Micromammiferi nella nicchia trofica del barbagianni {Tyto alba) nell’Italia Centro Meridionale. Avocetta 5:49-64. , 1988. La nicchia trofica di allocco (Strix aluco) e barbagianni {Tyto alba) in Italia: acquisizioni e prob- lemi. Naturalista Sicil. S. IV, XII (suppl.):129— 143. , F. Agostini, G. Aloise, and A. Testa. 1983. Sul rapporto trofico tra I micromammiferi terragnoli ed il barbagianni {Tyto alba Scopoli) nei monti della Tol- fa (Lazio). Quad. Acc. Naz. Lincei. 256:183-228. Galeotti, P. 1992. Barbagianni. Pages 180-181 m L. For- nasari, L. Bottoni, R. Massa, P. Brichetti, and V. Vi- gorita [Eds.], Atlante degli uccelli svernanti in Lom- bardia Regione Lombardia. Milano, Italy. Goodman, S.M. 1986. The prey of Barn Owls {Tyto alba) inhabiting the ancient temple complex of Karnak, Egypt. (95^7/157:109-112. Groppali, R. 1987. Scelta delle prede da parte del bar- bagianni ed allocco present! nella medesima localita del parco dell’Adda sud. Pianura 1:83-88. Heim de Balsac, H. 1965. Quelques enseignements d’ordre faunistique tires de I’etude du regime alimen- taire de Tyto alba dans Eouest de I’Afrique. Alauda 33: 309-322. Herrera, C.M. 1974. Trophic diversity of the Barn Owl Tyto alba in continental western europe. Ornis Scand. 5:181-191. Ingegnoli, V. 1993. Fondamenti di ecologia del paesag- gio. Citta Studi, Milano, Italy. ISTAT. 1990. Caratteristiche strutturali delle aziende agri- cole. Fascicolo Provincia MN. 4°. Censimento Gener- ale Agricoltura, Mantova, Italy. Jaksic, F.M. and J.L. Yanez. 1979. The diet of the Barn Owl in central Chile and its relation to the availability of prey. Auk 96:619-621. Kruuk, H. and T. Parish. 1981. Feeding specialization of european badger {Meles meles) in Scotland. J. Anim. Ecol. 50:773-788. Lenton, G.M. 1984. The feeding and breeding ecology of Barn Owls Tyto alba in peninsular Malaysia. ftA 126: 551-575. Levins, R. 1968. Evolution in changing environments. Princeton Univ. Press, Princeton, NJ U.S.A. Lovari, S., a. Renzoni, and R. Eondi. 1976. The preda- tory habits of the Barn Owl ( Tyto alba Scopoli) in re- lation to vegetation cover. Boll. Zool. 43:173-191. Margalef, R. 1968. Perspectives in ecological theory. Chicago Univ. Press, Chicago, IL U.S.A. . 1975. Diversity, stability and maturity in natural ecosystems. In W.H. Van Dobben and R.H. Lowe- McConnell [Eds.], Unifying concepts in ecology. Dr. W. Junk B.V. Publishers, The Hague, Netherlands. Marti, C.D. 1986. Barn Owl diet includes mammal spe- cies new to the island fauna of the Great Salt Lake Great Basin Nat. 46:307-309. Moreno, E. 1985. Clave osteologica para la identificacion de los Passeriformes Ibericos. I. Aegithalidae, Remi- zidae, Paridae, Emberizidae, Passeridae, Fringillidae, Alaudidae. Ardeola 32:295-377. . 1986. Clave osteologica para la identificacion de los Passeriformes Ibericos. II. Hirundinidae, Prunel- lidae, Sittidae, Certhiidae, Troglodytidae, Cinclidae, Laniidae, Oriolidae, Corvidae, Sturnidae, Motacilli- dae. Ardeola 33:69—129. . 1987. Clave osteologica para la identificacion de los Passeriformes Ibericos. III. Muscicapidae. Ardeola 34:243-273. Morton, S.R. and A.A. Martin. 1979. Feeding ecology of the Barn Owl, Tyto alba, in arid southern Australia. Aust. Wildl. Res. 6:191-204. Ottolini, E. and E. Aceto. 1996. La microteriofauna nel- le riserve natural! della provincia di Cremona. Pianura 8:45-67. Pandolfi, M. and R. Santolini. 1987. Primi dad sulla micro teriocenosi della valle del Eoglia (Pesaro) attrav- erso la dieta di Tyto alba. Riv. Ital. Ornitol. Milano. 57 165-172. and . 1988. Variazioni stagionali nella dieta di Tyto alba. Naturalista Sicil, S. IV, XII (suppl.) 1988 159-163. Petretti, F, 1977. Seasonal food habits of the Barn Owl {Tyto alba) in an area of central Italy. Gerfaut QlJ 233. Ptgnatti, S. 1994. Ecologia del paesaggio. UTET, Torino, Italy. PUGEK, Z. 1981. Key to the vertebrates of Poland mam- mals. Vol. III. Polish Scientific Publishers, Warszawa, Poland. Schoener, T.W. 1970. Size patterns in West Indian Anohs lizards. II. Correlations with the size of particular sym- patric species — displacement and convergence. Am Nat. 104:155-174. Sestini, a. 1963. Conosci ITtalia. Il paesaggio. Vol. VII Touring Club Italiano, Milano, Italy. Spitz, F. 1981. Un mode de traitement mathematique des donnees numeriques concernant les petits Mammi- feres en France. C.R. Acad. Sci., Paris 293:189—194. Southern, H. 1954. Tawny Owls and their prey. Ibis 96. 384-410. Torre, A. 1983. Variazione stagionale dell’alimentazione del barbagianni Tyto alba ernesti nel nord ovest della Sardegna. Avocetta 7:85-94. . 1987. Variazioni stagionali deH’alimentazione del barbagianni {Tyto alba) in una zona rurale della Sar- degna nord-occidentale. Pages 265—278 in N. Baccetti and M. Spagnesi [Eds.], Atti IV colloquio internazion- ale sui rapaci mediterranei. Instituto Nazionale Biol- ogia della Selvaggina, Bologna, Italy. Toschi, a. 1965. Eauna dTtalia, Mammalia. Lagomorpha, 246 Bose and Guidali VoL. 35, No. 3 Rodentia, Carnivora, Ungulata, Cetacea. Calderini, Bologna, Italy. . 1986. Avifauna Italiana. Edtoriale Olimpia, Firen- ze, Italy. AND B. Lanza. 1959. Fauna d’ltalia, Mammalia. Generalita, Insectivora, Chiroptera. Calderini, Bolo- gna, Italy. ViciNi, G. AND G. Malaguzzi. 1988. Alimentazione (Aut.- Inv.) del barbagianni {Tyto alba Scopoli) in un’area golenale del Po casalasco (Cremona), ed dementi di valutazione ambientale. Pianura 2:21-30. Received 30 September 1999; accepted 15 April 2001 J. Raptor Res. 35(3);247-252 © 2001 The Raptor Research Foundation, Inc. POTENTIAL NEGATIVE EEEECTS OE COLLISIONS WITH TRANSMISSION LINES ON A BONELLFS EAGLE POPULATION Santi Manosa and Joan Rfai. Departament de Biologia Animal, Universitat de Barcelona, Facultat de Biologia, Avinguda Diagonal, 645, 08028 Barcelona, Catalonia, Spain Abstract. — The Bonelli’s Eagle {Hieraaetus fasciatus) population decline in Europe has mainly been attributed to high levels of anthropogenic mortality. We evaluated the potential negative effects of collisions with transmission lines on a breeding population of Bonelli’s Eagles in Catalonia, Spain. Between 1990-97, two of the 12 recorded deaths of breeding Bonelli’s Eagles were caused by collisions with transmission lines. All transmission lines within a 5-km radius of 47 eagle nests were classified into two collision risk categories (low or high) , depending on their locations and habitats. Pairs having high risk lines within 1 km of nesting territories had turnover rates twice as high as pairs with no such lines (0.16 ± 0.11 [±SD] vs. 0.08 ± 0.10). Greatest turnover rates were observed when high-risk lines oc- curred within 100 m of nesting territories. Our results suggest that transmission lines near Bonelli’s Eagle nesting territories constitute a risk for eagles due to the danger of collisions. New transmission lines should avoid crossing areas near nesting territories and, as a precaution, those that are <1 km from eagle nests should be marked in some way. Keywords: Bonelli’s Eagle, Hieraaetus fasciatus; collision] conservation; endangered species', poiverlines. Posible efecto negativo de la colision con lineas de transporte electrico en una poblacion de aguila perdicera Resumen. — El declive del aguila perdicera {Hieraaetus fasciatus) en Europa ha sido atribuido en parte a una elevada mortalidad de origen antropico. Evaluamos el posible efecto negativo de la colision con tendidos de transporte electrico en la poblacion de aguilas en Cataluha (Espaha). Entre 1990-97, dos de las 12 muertes de aguilas sobre las que se obtuvo informacion fueron causadas por la colision con estos tendidos. Cada tramo de tendido situado en un radio de 5-km alrededor de 47 nidos se clasifico en dos categorias de peligrosidad (baja o alta), dependiendo de su localizacion y de las caracteristicas del habitat. Las parejas con tramos de alta peligrosidad a <1 km del nido presen taron tasas de recambio dos veces mas altas que el resto de las parejas (0.16 ± 0.11 [±SD] vs. 0.08 ± 0.10). El efecto mayor se observe cuando las lineas transcurrian a <100 m de los nidos. Los resultados sugieren que la presencia de estas lineas cerca de los nidos podria constituir un riesgo para las aguilas, quizas asociado al peligro de colision. Los nuevos trazados deberian evitar las zonas cercanas a los nidos y los tendidos ya existentes a <1 km de los mismos deberian sehalizarse. [Traduccion de Autores] Casualties caused by powerlines result in the deaths of thousands of birds in the world each year (Bayle 1999). Electrocution on pylons appears to be the major cause of mortalities, but collisions with wires also cause many bird deaths (APLIC 1994, Bevanger 1994, 1998, Bevanger and Over- skaug 1998). While electrocutions have been stud- ied (Olendorff et al. 1981, Negro et al. 1989, Fer- rer et al. 1991, APLIC 1994), the complex nature of collisions, as well as methodological and practi- cal constrains, have limited our understanding of this problem (Bevanger 1994, Henderson et al. 1996). Although bird collisions are common on distribution lines that carry <66 kV (Brown and Drewien 1995, Fernandez-Garcia 1998), they seem to occur mostly on the larger >110 kV transmis- sion lines and are probably related to the greater number of conductors, presence of earth wires, higher tower heights, and larger distances between poles (Fernandez-Garcia 1998). For these reasons, most of the research on bird collisions has focused on these high voltage transmission lines (Beaulau- rier 1981, Alonso et al. 1994, Savereno et al. 1996). Studies evaluating the impacts of collisions on bird populations are most often conducted by means of systematic searches under wires (Alonso 247 248 Manosa and Real VoL. 35, No. 3 et al. 1994, Savereno et al. 1996, Janss and Ferrer 1998). Most of these studies have found high num- bers of waterbirds, gamebirds, storks, and cranes dead under transmission lines, but very few birds of prey have been reported (Alonso et al. 1994, Bevanger 1995, 1998, Savereno et al. 1996, Janss and Ferrer 1998, Bayle 1999). As a result, raptors are generally considered less susceptible to colli- sions with overhead wires than other groups of birds (APLIC 1994, Alonso et al. 1994, Fernandez- Garcia 1998, Alonso and Alonso 1999). The de- creased susceptibility of raptors to collisions has been attributed to their acute vision and flight per- formance as well as their solitary habits and low population densities (R.E.E. 1993, APLIC 1994, Be- vanger 1994). While the number of collision acci- dents involving birds of prey might be low, the con- servation significance of such accidents can be high, especially when a species is endangered. In such a situation, reporting of a few deaths may be relevant enough to justify research on the effects of collisions at the population level (Bevanger 1998, Alonso and Alonso 1999) and the implemen- tation of mitigation actions (APLIC 1994, Bevanger and Overskaug 1998). The decline of the Bonelli’s Eagle (Hieraaetusfas- aatus) in Europe (del Hoyo et al. 1994, Real et al. 1994, Rocamora 1994) has been attributed to sev- eral factors including habitat loss and high adult and juvenile mortality. Demographic studies indi- cate that adult mortality (3.93-16.09% annual mor- tality rate) is the main cause of the population de- cline (Real and Manosa 1997). Depending on the region, powerlines have been reported as respon- sible for 8-100% of the deaths of breeding eagles (Real and Manosa 1997). Collisions of Bonelli’s Ea- gles with powerlines have been reported by several authors (Real and Manosa 1997, B. Arroyo and V. Garza unpuhl. data), which suggests that the pres- ence of transmission lines near nesting territories poses a potential danger to breeding Bonelli’s Ea- gles. This danger has never been thoroughly eval- uated because of difficulties associated with the es- timation of casualty rates (Bevanger 1998) and the lack of demographic data concerning wild popu- lations (Henderson et al. 1996). For Bonelli’s Ea- gles, systematic searches under wires have been considered impractical due to the terrain and thick vegetation found in their nesting territories. To overcome such constraints, we undertook a differ- ent approach to evaluate the effects of powerline collisions on the population dynamics of the spe- cies. Our aim was to compile existing information on collisions of Bonelli’s Eagles with transmission lines in Catalonia as a qualitative indication of the importance of collision casualties and to analyze the relationship between the presence of transmis- sion lines in Bonelli’s Eagle nesting territories and site-specific turnover rates. Study Area and Methods The study was carried out in the littoral and pre- littoral mountain ranges in Catalonia, northeastern Spain, where an estimated 70 Bonelli’s Eagle pairs, approximately 10% of the European breeding pop- ulation, are scattered over a 8000 km^ area (Man- osa et al. 1998). Information was compiled on the causes of death (i.e., electrocution, collision with transmission lines, shooting, other, unknown) of Bonelli’s Eagles that either we found dead or were received by rehabilitation centers and wildlife agencies in the area between 1990—97. Mortality was classified into two categories (collision with transmission lines and other) and birds into two status categories (breeding birds [adult or subadult birds found within a breeding area] and nonbreed- ing birds [all juvenile or immature birds, and sub- adult or adult birds found in a nonbreeding area]). During the last decade, the Catalan popu- lation of Bonelli’s Eagles has been monitored to estimate site-specific turnover rates. Between 1990- 97, nest sites have heen checked every year for the presence of breeders and breeding individuals have been classified as young (<1 yr old), imma- ture (1-2 yr old), subadult (2-3 yr old), and adult (>3 yr old), according to the plumage criteria (Parellada 1986). An estimate of the site-specific turnover rate of breeding birds was computed for 47 nesting territories in which at least 5 yr of mon- itoring data were available. The general computa- tion procedure for turnover rate followed the method used by Real and Manosa (1997) to esti- mate survival rates. At the start of every breeding season, we checked every territory and recorded if a breeding bird had disappeared or had been re- placed by a younger bird. Birds were not marked and turnover was judged based on plumage char- acteristics which changed with age (Parellada 1986). This method gave a minimum estimate of turnover rate, since some changes in adult plum- age would not have been detected. Because Bo- nelli’s Eagles have high mate and site fidelity (Cheylan 1972, Cramp and Simmons 1980), we as- sumed that replacement or disappearance of September 2001 Transmission Lines and Bonelli’s Eagles 249 YES YES HIGH RISK concealed h’ topogi^? LOW RISK Figure 1. Decision rules followed to assign collision risk indexes to sections of transmission lines in the study area breeders in territories was due to mortality within the territory, although some might have been caused by desertions or movement to other areas. To reduce the possibility of overestimating turn- over rates in territories that were deserted, and to eliminate autocorrelation between high site-specif- ic turnover rates and territory desertion, disap- pearances that resulted in site desertions were not coiisidei ed in the calculations of site-specific turn- over rates. Turnover rate for a given site was com- puted as the total number of individuals that were replaced or disappeared in relation to the total number of individuals and years considered. Breeding areas were classified as having low turnover rates (i.e., only one change recorded dur- ing the study period or, if more, a turnover rate ^0.10) or high turnover rates (i.e., more than one change recorded during the study period and a turnover rate >0.10). Collision risks on a transmission line depends on the habitat characteristics and topographical fea- tures crossed by the line in relation to the behavior and habitat requirements of the target species (APLIC 1994). We visited every known Bonelli’s Eagle breeding territory and plotted them on T 50 000 topographic maps showing transmission lines (110-400 kV) within a 5-km radius of every ilcst. Based on onr knuwledgc of eagle bejiavior, we estimated the main flight paths of eagles and predicted which sections of transmission lines were more likely to be crossed by eagles regularly and posed the greatest collision threats. A total of 960 km of transmission lines were plotted, which were subsequently divided into segments of variable length, each one being classed as a High Collision Risk segment or Low Collision Risk segment by means of a decision tree (Fig. 1) which took into account whether (1) a given section of line crossed 250 Manosa and Real VoL. 35, No. 3 a “flight path” (Bevanger 1994) that eagles pre- sumably followed in their nesting territories on a regular basis, (2) if a section of line crossed areas suitable for eagle hunting, and (3) if a section was concealed by topographic features. Based on our experience, eagles typically fly along lines that follow nest cliffs and slopes, and cross mountain passes. We classified habitats under lines as either good for hunting (i.e., open forest, ecotones, bush and shrubs, traditional nonirrigat- ed farmland) or not good for hunting (i.e., exten- sive woodland areas, compact urban areas, irrigat- ed land, intensive irrigated farmland) (del Hoyo et al. 1994). Finally, sections of lines found in good hunting habitats were considered as being con- cealed by topographic features when they were be- hind mountain ridges or mountain passes, or in places where they did not stand out against the background. No consideration was made about powerline design because all of them had earth wires which are considered to be the main source of bird collisions in transmission lines (APLIC 1994). The following variables of collision risk were measured within 0-1 km, 1-3 km, and 3-5 km radii of every nesting site: LROl — km of Low Collision Risk lines < 1 km away, LRl 3 — km of Low Collision Risk lines 1-3 km away, LR35 — km of Low Collision Risk lines 3-5 km away, HROl — km of High Colli- sion Risk lines <1 km away, HR13 — km of High Collision Risk lines 1-3 km away, HR35 — km of High Collision Risk lines 3—5 km of nest, TOl — total km of transmission lines within <1 km of nesting territory, T13 — total km of transmission lines 1-3 km of nesting territory, and T35 — total km of transmission lines 3-5 km of nesting terri- tory. Turnover rates were correlated to collision-risk variables using Spearman rank correlation coeh- cients. Bonferroni adjustment for multiple test comparisons (Rice 1989) were conducted in mul- tiple Spearman rank correlation tables (Table 1). Corrections where made considering three test families of size = 3 and adopting a global experi- ment-wise error of 0.1, following the indications of Chandler (1995). As a consequence, we set critical T-values for individual tests at 0.03. Mann-Whitney U tests were used to compare turnover rates be- tween sites having transmission lines and those without them. When these comparisons were made for distances of 1-3 km and 3-5 km, sites having powerlines at shorter distances were excluded. Table 1. Spearman rank correlation coefficient between turnover rates and the amount of Low Collision Risk transmission lines (LR) , High Collision Risk transmission lines (HR), and Total transmission lines (T) present at increasing radii from Bonelli’s Eagle nesting territories {N = 47 sites). b P LROl 0.11 0.47 LRl 3 0.03 0.83 LR35 0.16 0.27 HROl 0.33 0.02 HR13 0.02 0.91 HR35 0.02 0.87 TOl 0.26 0.08 T13 0.03 0.81 T35 0.16 0.28 Kruskal- Wallis tests were used to compare average turnover rates between sites having High Collision Risk transmission lines at increasing distances from nests. Results During the 7-yr study, a total of 502 individuals were monitored in 47 nesting territories. We re- corded a minimum of 44 changes in breeding in- dividuals, or turnovers of adults, at 27 nesting ter- ritories. At three of these territories, where 36 individuals had been monitored and 7 changes were observed, we actually found three dead ea- gles. Nine more eagles were reported dead in the study area, that could not be related to specific nesting territories. Of the 12 eagles found dead, collision with transmission lines was the cause of death of 2 (17%). Six (50%) died by electrocution, 3 (25%) were shot, and one died of unknown caus- es. Both Bonelli’s Eagles that collided with trans- mission lines were males. One was found in March 1993, freshly dead under a 110-kV transmission line with its jaw and wing broken 900 m from a nest. The second dead eagle was found under a 400-kV transmission line in November 1994 with a broken wing, only 100 m from its nest. Neither bird appeared to be shot and lead pellets were not found in their bodies. We classified the section of transmission line where the first mortality occurred as Low Collision Risk and the second mortality oc- curred on a section of transmission line classified as High Collision Risk. September 2001 Transmission Lines and Bonelli’s Eagles 251 Annual turnover rates at nesting territories ranged from 0.00-0.40 (x = 0.10 ± 0.11, ±SD, N — 47). Median turnover rate was 0.07 and 25% of territories showed turnover rates >0.14. Site-spe- cific turnover rates were positively correlated with the length of High Collision Risk transmission lines <1 km from the territory (Table 1). Territories having the greatest amount of High Collision Risk lines within a radius of 1 km showed significantly higher turnover rates (0.16 ± 0.11, N= 10) than those not having High Collision Risk lines (0.08 ± 0.10, 37) (Mann-Whitney test, C = 100.0, P = 0.02). For territories having High Collision Risk lines 0—1 km (turnover rate = 0.16 ± 0.11, N — 10), 1-3 km (turnover rate = 0.09 ± 0.13, N — 10), and 3-5 km (turnover rate = 0.04 ± 0.07, N = 10) away from nesting territories, turnover rates decreased significantly as lines occurred farther away (Kruskal-Wallis tests, ^ 8.05, df = 2, P = 0.018). For the 10 nesting territories with High Collision Risk lines <1 km away, all of the sites with High Collision Risk transmission lines <100 m away had high turnover rates (x = 0.28 ± 0.02, N — 4) . Only one of the six territories with transmis- sion lines >100 m away had a high turnover rate (x = 0.11 ± 0.09, N — 6) (Fisher exact test, P — 0.024). Discussion Although much more emphasis has been placed on the importance of electrocution on the mortal- ity of birds of prey (Bevanger 1994, 1998), colli- sions between raptors and powerlines are known to cause high mortalities in such large and endan- gered species as the California Condor ( Gymnogyps californianus) (Snyder and Snyder 1989). Our re- sults show that, after electrocution and shooting, collisions by breeding eagles with transmission lines is the third most important cause of nonnat- ural mortality in Bonelli’s Eagles in Catalonia caus- ing as much as 17% of deaths in the population. Assuming an average annual mortality rate of 10% (Real and Manosa 1997) , collisions with powerlines causes 1.7% of the annual mortality in the popu- lation. In view of the fact that the annual adult mortality rate must not exceed 2-6% for the pop- ulation to remain at equilibrium (Real and Manosa 1997), our estimate indicates that collisions with powerlines poses a serious threat to the popula- tion. Transmission lines near nests were associated with high site-specific turnover rates, which may in- dicate that they cause the deaths of adults when they collide with wires. Other studies have shown similar results for birds in the vicinity of powerlines (Bevanger 1995). Although eagles may become fa- miliar with transmission lines within their territo- ries (APLIC 1994), the need for them to increas- ingly cross powerlines near nests during the nesting season apparently increases their risk of collisions, particularly in bad weather or while hunting (APLIC 1994). Other potential causes of mortality associated with transmission lines but not caused by collisions, such as the presence of access roads, increased hu- man disturbance, shooting, habitat humanization, and reduced prey availability, did not appear to be important because the transmission lines typically crossed areas that were inaccessible to people and not inhabited. The fact that turnover rates were related to the amount and proximity of High-Col- lision Risk transmission lines, but not the total length of transmission lines within the territory, in- dicated that turnover was the result of a direct col- lision danger to eagles rather than to reduced hab- itat quality or higher interferences in areas crossed by powerlines. Given the danger of transmission lines near Bo- nelli’s Eagle nesting territories, new transmission lines should be constructed so that they avoid crossing <1 km of an eagle territory. In addition, all existing transmission lines within a radius of 5 km of a Bonelli’s Eagle nesting territory should be carefully checked, evaluated for risk, and adequate- ly marked if they interfere with flight paths of breeding eagles. Site-specific monitoring should follow these mitigation actions to confirm any po- tential reduction in mortality. More research is needed on the habitat use and behavior of Bo- nelli’s Eagles to better evaluate the risks posed by specific sections of transmission lines, both near nests and in hunting areas, where collisions may also occur. Acknowledgments Red Electrica de Espaiia (R.E.E.) gave financial sup- port to this study. We are grateful to J. Roig, V. Navazo, and X. Parellada for their interest and help. Long-term monitoring of the Bonelli’s Eagle population in Catalo- nia is sponsored by Fundacio Miquel Torres. J.M. Grande helped in obtaining data on mortality. We also thank the Museu de Zoologia de Barcelona and the Serve! de Ges- tio i Proteccio de la Fauna (Departament d’Agricultura Ramaderia i Pesca, Generalitat de Catalunya) for permis- sion to use their specimens and files. Comments of B 252 Manosa and Real VoL. 35, No. 3 Arroyo and an anonymous referee improved an early ver- sion of the manuscript. Literature Cited Alonso, J.A. and J.C. Alonso. 1999. Colision de aves con Hneas de transporte de energia electrica en Espaha. Pages 61-88 in M. Ferrer and G. Janss [Eds.], Aves y Hneas electricas. Colision, electrocucion y nidifica- cion. Quercus, Madrid, Spain. Ai.onso, J.C., J.A. Ai.onso and R. Munoz-Pulido. 1994. Mitigation of bird collision with transmission lines through groundwire marking. Biol. Conser. 67:129— 134. APLIC. 1994. Mitigating bird collisions with powerlines: the state of the art in 1994. Edison Electric Institute, Washington, DC U.S.A. Bayle, P 1999. Preventing birds of prey problems at transmission lines in western Europe./. Raptor Res. 33: 43-48. Beaulaurier, D.L. 1981. Mitigation of bird collisions with transmision lines. Bonneville Power Admin., U.S. De- partment of Energy, Portland, OR U.S.A. Bevanger, K. 1994. Bird interactions with utility struc- tures: collision and electrocution, causes and mitiga- tion measures. Ibis 136:412-425. . 1995. Estimates and population consequences of tetraonid mortality caused by collisions with high ten- sion power lines in Norway. / Appl. Ecol. 32:745-753. . 1998. Biological and conservation aspects of bird mortality caused by electricity power lines: a review. Biol. Conserv. 86:67-76. AND K. OVERSKAUG. 1998. Utility structures as a mortality factor for raptors and owls in Norway. Pages 381-392 in R.D. Chancellor, B.-U. Meyburg, and J.J. Ferrero [Eds.], Holarctic birds of prey. ADENEX- WWGBP, Calamonte, Spain. Brown, W.M. and R.C. Drewien. 1995. Evaluation of two power line markers to reduce crane and waterfowl collision mortality. Wildl. Soc. Bull. 23:217-227. Chandler, C.R. 1995. Practical considerations in the use of simoultaneous inference for multiple tests. Anini. Behav. 49:524-527. Cheyi-AN, G. 1972. Le cycle annuel d’un couple d’aigle de Bonelli Hieraaetus fasciatus (Vieillot). Alauda 11: 214-234. Cramp, S. and K. Simmons. 1980. The birds of the west- ern Palearctic. Vol. 2. Oxford Univ. Press, Oxford, U.K. DEL Hoyo, J., A. Elliott, and J. Sargatal. [Eds.] 1994. Handbook of the Birds of the World. Vol 2. New World Vultures to Guineafowl. Lynx Edicions, Barce- lona, Spain. Fernandez-GarciA, J.M. 1998. Relationship between mor- tality in electric power lines and avian abundance in a locality of Leon (northwest of Spain) . Ardeola 45:63- 67. Ferrer, M., M. De La Riva, and J. Castrovtejo. 1991. Electrocution of raptors on powerlines in Southern Spain. /. Field Ornithol. 62:54-69. Henderson, EG., R.H.W. Langston, and N.A. Clark. 1996. The response of Common Terns Sterna hirundo to power lines: an assessment of risk in relation to breeding commitment, age and wind speed. Biol. Con- serv. 77:185-192. Janss, G. and M. Ferrer. 1998. Rate of bird collision with power lines: effects of conductor marking and static wire marking./. Field Ornithol. 69:8-17. Manosa, S., J. Real, and J. Godina. 1998. Selection of settlement areas by juvenile Bonelli’s Eagles in Cata- lonia./. Raptor Res. 32:208-214. Negro, J.J., M. Ferrer, C, Santosa, and S. Regidor 1989. Eficacia de dos metodos para prevenir electro- cuciones de aves en Hneas electricas de distribucion. Ardeola 36:201-206. Olendorff, R.R., A.D. Miller, and R.N. Lehman. 1981. Suggested practices for raptor protection on power lines: the state of the art in 1981. Raptor Research Report No. 4. Raptor Research Foundation, St. Paul, MN U.S.A. Parellada, X. 1986. Variacio del plomatge i identihcacio de I’aliga cuabarrada Hieraaetus fasciatus. Pages 97-96 in O. Alamany, A. de Juan, X. Parellada, J.R. Tied, and J. Tied [Eds.], Rapinyaires Mediterranis II. Centre de Recerca i Proteccio de Rapinyaires, Barcelona, Spain Real, J. and S. Manosa. 1997. Demography and conser- vation of western European Bonelli’s Eagle Hieraaetus fasciatus populations. Biol. Conserv. 79:59-66. , , and J. Codina. 1994. Estatus, demografia y conservacion del Aguila perdicera {Hieraaetus fascia- tus) en el Mediterraneo. Pages 83-89 inj. Muntaner and J. Mayol [Eds.], Biologia y conservacion de las rapaces mediterraneas. Monografias S.E.O., Madrid, Spain. , L. Pai.ma, and G. Rocamora. 1997. Bonelli’s Ea- gle. Pages 174-175 in W.J.M. Hagemeijer and M.J. Blair [Eds.], Atlas of European breeding birds. T. and A.D. Poyser, London, U.K. R.E.E. 1993. Sehalizacion de Hneas de alta tension para la proteccion de la avifauna. Red Electrica de Espaha, S.A., Madrid, Spain. Rice, W.R. 1989. Analyzing tables of statistical tests. Evo- lution 43:223-225. Rocamora, G. 1994. Bonelli’s Y.digle Hieraaetus fasciatus. Pages 184—185 in G.M. Tucker and M.F. Heath [Eds.], Birds in Europe: their conservation status. BirdLife Conservation Ser. No. 3, BirdLife International, Cam- bridge, U.K. Savereno, A.J., L.A. Savereno, R. Boettcher, and S.M Haig. 1996. Avian behavior and mortality at power lines in coastal South Carolina. Wildl. Soc. Bull. 24* 636-648. Snyder, N.F.R. and H. Snyder. 1989. Biology and conser- vation of the California Condor. Curr. Ornithol. 6:175- 267. Received 3 September 2000; accepted 30 April 2001 Short Communications J. Raptor Res. 35(3);253-256 © 2001 The Raptor Research Foundation, Inc. Agonistic Behavior of Cooper’s Hawks Clint W. Boat* School of Renewable Natural Resources, University of Arizona, Tucson, A7. 85721 U.S.A. Key Words: Cooper's Hawk, Accipiter cooperii; agnostic behavior, nest defense, extra-pair copulation. Most Falconiformes defend breeding or hunting areas from other conspecific and heterospecific raptors (New- ton 1979). Such defense behaviors may be relatively non- aggressive (e.g., posturing, flight displays; Jamieson and Seymour 1983, Bildstein and Collopy 1985), aggressive (e.g., chases, physical contact; Jamieson and Seymour 1983, Sodhi 1991, Bustamante and Hiraldo 1993, Fernan- dez and Azkona 1994), or combinations of displays and aggression (Dawson and Mannan 1991). Most descrip- tions of agonistic behavior among birds of prey are for open country species (Jamieson and Seymour 1983, Bild- stein and Collopy 1985, Dawson and Mannan 1991, Sodhi 1991, Fernandez and Azkona 1994). This is probably due to the relative ease in making observations in open land- scapes compared to forests and woodlands. For example, aside from using Great Horned Owls (Bubo virginianus) as lures at traps (Bloom 1987), there are few accounts of agonistic behavior among woodland raptors. The respon- siveness of woodland raptors to broadcasts of conspecific and heterospecific calls (Bosakowski and Smith 1998) and a few anecdotal accounts (Meng 1951, Crannell and DeStefano 1992), however, suggest agonistic interactions may be relatively common. As is the case for most woodland raptors, there is little information available on the behavior of Cooper’s Hawks {Accipiter cooperii) toward conspecific and heterospecific in- truders in nest areas (Meng 1951, Rosenfield and Papp 1988). The relatively open landscape of the urban envi- ronment of Tucson, Arizona, and the approachability of urban nesting Cooper’s Hawks (Boal and Mannan 1999), made it possible to observe agonistic interactions between territory holders and nest-area intruders. Here, I describe agonistic behaviors of Cooper’s Hawks responding to con- specific and heterospecific intruders during the prelaying and incubation portions of the breeding season. ' Present address: U.S. Geological Survey, Texas Coop- erative Fish and Wildlife Research Unit, Texas Tech Uni- versity, Uubbock, TX 79409-2120 U.S.A. Methods These data were collected during a study of Cooper’s Hawks in the greater Tucson metropolitan area (32°12'N, 110°57'W) in southeastern Arizona, 1994-96. The area encompasses approximately 70 000 ha with an estimated human population of about 800 000. Nesting Cooper’s Hawks were located through standardized surveys and were being closely monitored (Boal and Mannan 1999) so the stage of the breeding cycle was known for all ob- servations of agonistic interactions. When a nest intru- sion was observed, 1 recorded which member (s) of the breeding pair was present, which one(s) responded to the intruder, and the age (adult or subadult) and sex of conspecific intruders. Plumage and sexual size dimor- phism facilitated the visual determination of age and sex of Cooper’s Hawks. Heterospecific intruders were iden- tified to species, age, and, when possible, sex. I catego- rized responses as chases, or chases with strikes (e g., chase with observed physical contact), and noted when either a resident or intruding Cooper’s Hawk vocalized during an interaction. Observations of nest area intrusions were made during the course of routine field activities (e.g., nest checks, radiotracking) so I did not calculate a rate for intrusions per time of observation. Rather, this is a compilation of observations that, taken together, may help elucidate an understanding of the defense behavior of Cooper’s Hawks. Results I observed 19 breeding season encounters between known, marked, breeding Cooper’s Hawks and intruding conspecifics. Fourteen of the intrusions occurred during the pre-laying stage (73.7%), fewer occurred during the incubation stage (26.3%), and none were observed dur- ing the nestling stage. Fifteen of the 19 intrusions elicited aggressive responses by breeding Cooper’s Hawks (Table 1). Both members of breeding pairs were present during 14 (73.7%) of the 19 intrusions, but both members pur- sued intruders on only two (14.3%) of those 14 occa- sions. Males were more likely to respond to conspecific intruders (x^i = 11.1, P = 0.0008), engaging 85.7% of the intruders, whereas females engaged only 20.0%. However, sex-related differences in responsiveness of breeding Cooper’s Hawks appeared to be more closely associated with the sex of the intruder (Table 1). Four- 253 254 Short Communications VoL. 35, No. 3 Table 1. Incidents of aggressive response {N = 15) and no response (N = 4) by breeding Cooper’s Hawks to conspecific nest area intruders during the pre-laying and incubation periods, Tucson, Arizona, 1994-96. Prelaying Intruder Incubation Intruder Responding Resident d 9 Unk. d 9 Total Aggressive response 5 7 I 0 3 0 11 5 0 2 0 0 0 2 0 I 1 0 0 2 No response d 0 0 0 0 0 0 9 0 0 0 3 0 3 d9 1 0 0 0 0 1 teen (73.7%) of the intruders were males (10 adult, four subadult), four were subadult females (21.0%), and one was not identified to sex or age (5.3%; Table 1). Breeding Cooper’s Hawks were more likely to aggressively respond to conspecific intruders of their own sex (Fisher exact test; P = 0.022; Table 1). Of the 15 aggressive interac- tions, 10 (66.7%) intruders were chased from the nest area and five (33.3%) were physically struck by breeding Cooper’s Hawks. Cooper’s Hawks also tended to be silent (77%) during aggressive responses to conspecific intrud- ers. Two particularly physical interactions showed the de- gree of aggression and potential for injury that can ac- company agonistic encounters between Cooper’s Hawks. In one situation, an adult male chased an intruding adult male out of the nest stand and pursued it for approxi- mately 500 m, neither bird ever flying >20 m above the ground. On three occasions during the chase the hawks faced each other while hovering and repeatedly made stabbing strikes at each other with their feet. Contact was Table 2. Incidents of aggressive response by breeding Cooper’s Hawks to heterospecific nest area intruders, Tucson, Arizona, 1994—96. Number of individual intru- sions followed by total number of intruding birds in pa- renthesis (e.g., one intrusion by Common Ravens in- volved two intruding ravens. CORA“ CHOW« RTHA^ HAHA^ TUVU^ d 1 (2) 0 3 (4) 1 (1) 2(4) 9 0 0 1 (1) 0 0 d9 0 2(4) 1 (1) 0 0 •* CORA = Common Raven ( Corvus corax) , GROW = Great Horned Owl {Bubo xjirginianus) , RTHA == Red-tailed Hawk (Buteo jamaicensis) , HAHA = Harris’ Hawk {Parabuteo unicinctus), and TUVU = Turkey Vulture ( Cathartes aura) . made several times, but neither male maintained a hold on the other. In the second situation, an adult female rose up to engage a subadult female circling above the nest stand. The adult female made several passing strikes at the intruder which rolled and extended its talons to- ward the aggressor. The subadult was struck solidly at least three times before leaving the area. Four of the 19 intrusions did not elicit aggressive re- sponses (Table 1). In the hrst incident, a known, marked pair of Cooper’s Hawks and an unbanded subadult male were perched in plain view and within 100 m of each other. The adults appeared to ignore the subadult, which flew away from the nest area after about 15 min of ob- servation. All nestlings of the marked pair had been banded during the previous year, so the subadult could not have been their offspring. The other three occasions occurred when intruding males from adjacent territories perched in nest trees or adjacent trees while resident fe- males were incubating. In each case, the intruding and resident males were radio tagged; telemetry indicated resident males were away from the nest areas. One nonaggressive interaction was observed that did not include a breeding area intrusion. Rather, the inci- dent appeared to be a border display between the known breeding females from adjacent nests. This incident in- volved both females approaching each other, followed by continuous circling and altitude gain near each other, but not overlapping, until both were lost from view. One of the females was observed diving back toward its nest stand a few minutes later. No evidence of aggression was observed during the encounter. 1 observed 11 intrusions by other species that elicited responses by Cooper’s Hawks (Table 2). Some intrusions consisted of more than one intruding individual. Males engaged intruders during seven (63.6%), females during one (9.1%), and both members during three (27.3%) of these intrusions. Although nest-area defense by Cooper’s September 2001 Short Communications 255 Hawks was successful in most situations. Cooper’s Hawks were unsuccessful twice in driving Great Horned Owls away and, in both cases, the owls usurped the Cooper’s Hawk nests. Dtscussion Energetic demands in the production of sperm are probably negligible in comparison to that required for the production of energy-rich eggs. This disparity has lead to suggestions that females contribute more to the reproductive effort than males, at least up to the time of fertilization (Dawkins 1976). In contrast, Beissinger (1987) suggested Snail Kite {Rostrhamus sociabilis) males, and Rosenfield and Bielefeldt (1991) suggested Cooper’s Hawk males make greater investments in reproduction than females because they provide most of the food, do most of the nest building, and chase potential nest pred- ators more frequently than females. I found that breed- ing Cooper’s Hawks were more responsive to conspecific intruders of their own sex. Thus, it appears caution should be taken with using nest defense as a measure of reproductive investment. While defense against hetero- specific nest intruders and predators may be a valid in- dication of investment, nest defense against conspecifics may be a poor measure. Sperm competition (Birkhead 1988) and food re- source competition (Temeles 1989) are two hypotheses that are not mutually exclusive and may both partially account for the high degree of nest defense by male and comparative lack of defense by female Cooper’s Hawks. The majority of conspecific intrusions were by male Coo- per’s Hawks and occurred during the pre-laying (i.e., fer- tile) stage. Resident male Cooper’s Hawks always drove away intruding males and resident female Cooper’s Hawks engaged in conspecific nest defense only when the intruder was a female. Likewise, Meng (1951) re- ported the male of a breeding pair of Cooper’s Hawks making repeated strikes on a captive male tethered near the nest, but the female of the pair was comparatively unresponsive. It is possible that, in order to avoid extra- pair fertilization of the female, male members of Coo- per’s Hawk pairs are most diligent in driving away in- truding males. In contrast, female members of breeding pairs may not avoid extra-pair copulations, but may be more inclined to drive away intruding females that could compete for food resources provided by the male. For example, in this study three intruding males failed to elic- it a response by the resident females when resident males were away from the nest. Other factors may also contribute to the agonistic be- havior of Cooper’s Hawks. Rosenfield and Papp (1988) suspected an intruding subadult female Cooper’s Hawk killed and cannibalized nestling Cooper’s Hawks. If such behaviors occur, they could lead to breeding hawks de- fending nest areas against conspecifics in addition to oth- er nest predators. However, in this study I never observed intruding Cooper’s Hawks acting aggressively toward nes- tlings, and noted one example of a male Cooper’s Hawk acting as a helper at one nest throughout the breeding cycle (Boal and Spaulding 2000). It is possible that the predominance of males among intruding Cooper’s Hawks was a result of sexual differ- ences in detectability. Such an explanation suggests that some behavior (s) of male Cooper’s Hawks makes them more susceptible to detection than the physically larger females. I did not observe any behaviors among intrud- ing or resident Cooper’s Hawks that led me to believe sex-related behavioral differences influenced detectabili- ty. However, stage of the nesting cycle may influence in- trusion rates. While males remain mobile during the courtship and incubation periods, movements of the nesting female become necessarily more restricted to the nest area. It seems plausible that both paired and un- paired male Cooper’s Hawks may invade nest areas to seek copulations with resident females. Likewise, an un- paired female may also invade a nest area seeking a po- tential mate. It seems unreasonable, however, that a res- ident paired female would as readily leave her nest area to seek extra-pair copulations. While this line of reason- ing is speculative, it appears reasonable that intrusion rates would favor males. I suggest male Cooper’s Hawks may be aggressive to- ward intruding male conspecifics to avoid extra-pair fer- tilization of their mates, and that female aggression to- ward intruding females may be food resource related. To address agonistic behavior quantitatively among Cooper’s Hawks, detailed observations at nests should be conduct- ed. The study should include calculations of copulation rates, correspondence of copulations to prey delivery, in- trusion rates, and incidences of extra-pair copulations Difficulties would include the probable low frequency of nest intrusions, the chance of observing such intrusions when they occur, and the inherent difficulty of making such observations in wooded habitat. The density of the study population and number of floaters would also in- fluence nest intrusion rates; lower densities would likely have lower incidence of nest intrusions, and hence, lower probability of observing the intrusions. During 679 hr of observations at Merlin {Falco columbarius) nests, Sodhi (1991) observed only 28 conspecific nest intrusions and five extra-pair copulations/ attempts. It is difficult to draw statistical significance from so few data points, yet the study was conducted under close to ideal conditions: the population was the densest ever reported for Merlins (So- dhi 1991) and was in an urban area where observations are often more easily conducted than in natural habitats Such observations are even less likely under typical con- ditions. Thus, a quantitative investigation of agonistic be- havior of Cooper’s Hawks will require a substantial time commitment in terms of nest observation and would be facilitated by a study area with a dense population, such as one of the several urban areas where the species is now known to nest. 256 Short Communications VoL. 35, No. 3 Resumen. — Examine los patrones de respuesta de Accip- iter cooperii en reproduccion en relacion a las intrusiones al area del nido producidos por individuos de la misma especie y de especies diferentes. Quince de las 19 intru- siones produjeron respuestas agresivas por parte de los gavilanes residentes. Los Accipiter cooperii residentes ten- dian a responder agresivamente a sus congeneres intru- sos del mismo sexo. Los machos de los gavilanes se in- volucraron mas con intrusos de otras especies que las hembras. Los patrones de defensa del area del nido por parte de Accipiter cooperii fueron ocasionados por difer- entes factores tales como la competencia de esperma, competencia de recursos y el escenario del ciclo de reproduccion. [Traduccion de Cesar Marquez] Acknowledgments R L. Spaulding and B.D. Bibles assisted with field obser- vations. D. Stahlecker and two anonymous reviewers pro- vided suggestions that improved the clarity of the man- uscript. Literature Cited Beissinger, S.R. 1987. Anisogamy overcome: female strat- egies in Snail Kites. Am. Nat. 129:486-500. Bildstein, K.L. AND M.W. CoLLOPY. 1985. Escorting flight and agonistic interactions in wintering Northern Har- riers. Ccmrfor 87:398-401. Birjkhead, T.R. 1988. Behavioral aspects of sperm com- petition in birds. Adv. Stud. Behav. 18:35-72. Bloom, P.H. 1987. Capturing and handling raptors. Pag- es 99-123 in B.A. Giron Pendleton, B.A. Millsap, K.W. Kline, and D.M. Bird [Eds.], Raptor management techniques manual. Nat. Wildl. Fed., Washington, DC U.S.A. Boai., C.W. AND R.W. Mannan. 1999. Comparative breed- ing ecology of Cooper’s Hawks in urban and exurban areas of southeast Arizona. /. Wildl. Manage. 63:77-84. AND R.L. Spaulding. 2000. Helping at a Cooper’s Hawk nest. Wilson Bull. 112:275-277. Bosakowski, T. AND D.G. Smith. 1998. Response of a for- est raptor community to broadcasts of heterospecific and conspecific calls during the breeding season. Can Field-Nat. 112:198-203. Bustamante, J. and F. Hiraldo. 1993. The function of aggressive chases by breeding Black and Red Kites Milvus migrans and M. Milvus during the post-fledging dependence period. Ibis 135:139-47. Crannell, D. AND S. DeSteeano. 1992. An aggressive in- teraction between a Northern Goshawk and a Red- tailed Hawk. J. Raptor Res. 26:269-270. Dawkins, R. 1976. The selfish gene. Oxford Univ. Press, Oxford, U.K. Dawson, J.D. and R.W. Mannan. 1991. The role of ter- ritoriality in the social organization of Harris’ Hawks. Auk 108:661-672. EernAndez, C. and P. Azkona. 1994. Sexual differences in conspecific territorial defense of Marsh Harriers {Circus aeruginosus) . J. Raptor Res. 28:23-26. Jamieson, LG. and N.R. Seymour. 1983. Inter- and intra- specific agonistic behavior of Osprey {Pandion haliae- tus) near their nest sites. Can. J. Zool. 61:2199-2202. Meng, H.K. 1951. Cooper’s Hawk, Accipiter cooperii (Bon- aparte). Ph.D. dissertation, Cornell Univ., Ithaca, NY U.S.A.. Newton, I. 1979. Population ecology of raptors. Buteo Books, Vermillion, SD U.S.A. Rosenfield, R.N. and j. Bielefeldt. 1991. Reproductive investment and anti-predator behavior in Cooper’s Hawks during the pre-laying period. J. Raptor Res. 25 113-115. AND J.M. Papp. 1988. Subadult intrusion and prob- able infanticide at a Cooper’s Hawk nest. Wilson Bull 100:506-507. SODHI, N.S. 1991. Pair copulations, extra-pair copula- tions, and intraspecific nest intrusions in Merlins. Con- dor 93:433-437. Temeles, E.J. 1989. The effect of prey consumption on territorial defence by harriers: differential responses to neighbors versus floaters. Behav. Ecol. Sociobiol. 24; 239-243. Received 2 Dec. 2000; accepted 15 April 2001 September 2001 Short Communications 257 J Raptor Res. 35(3):257-258 © 2001 The Raptor Research Foundation, Inc. Ground-nesting Ospreys in Utah Clark S. Monson Department of Geography, Brigham Young University, Provo, UT 84602 U.S.A. Key Words: Osprey, Pandion haliaetus; ground nesting, predation-, Utah. Ospreys {Pandion haliaetus) utilize an array of both nat- ural and artificial nesting sites. Frequently-chosen nesting sites include tree tops, rock pinnacles, utility poles, and elevated nesting platforms (Bent 1937). Ground nests are also common, but are restricted to small coastal islands where mammalian predators are absent (Poole 1989). Ospreys do not nest on the ground on the North Amer- ican mainland where the likelihood of mammalian pre- dation is high (Palmer 1988). In 1999, however, I wit- nessed a pair of ground-nesting Ospreys in Utah. The following account is perhaps the first recorded incident of ground-nesting Ospreys in the interior of North America. Ospreys are historically rare breeders in Utah, where only three to four lakes and reservoirs traditionally sup- port nesting pairs (Hayward et al. 1976). Beginning in 1995, they have colonized numerous additional water bodies, particularly in the northern portion of the state (Monson 1996). On 5 May 1999, while viewing a pair of Ospreys on a nesting platform at Deer Creek Reservoir, Wasatch coun- ty (Fig. 1), I noticed a second pair of Ospreys standing on an extensive mud flat 30 m from the water’s edge at the upper end of the reservoir. Several days later, the two birds were seen again in the same location. While I ob- served them, the male flew a short distance over the res- ervoir, collected some floating plant material, deposited it on the ground near the female, and copulated with her. On 9 May, the female appeared to be in an incuba- tion posture, yet I could not see a nest structure from an elevated road <300 m away. On 10 May, I found the nest. The female flushed when I approached within 100 m and was clearly not handicapped in any way. The nest con- sisted of approximately 30 sticks encircling a thin mat of dried grass. It had apparently been depredated since only a large fragment of eggshell remained. On 12 May, the female was again seen incubating. A return visit to the nest revealed a newly laid, intact egg (Fig. 2) . Several days later, however, the Ospreys moved to a new ground site 60 m away, presumably in response to persistent predator molestation. Bailing twine was the only nesting material they brought to the new nest site. An egg was laid on 18 May, but on 19 May it was missing. Nesting activity ceased with the loss of this third and final egg, but the adults remained in the immediate area, continuing to stand on the mud flat near their former nest sites rather than perching in nearby trees. While Opreys nest successfully on the ground on coast- al islands, the ground-nesting pair in Utah was ultimately destined to fail. Mammalian predators, especially striped skunks {Mephitis mephitis), raccoons {Procyon lotor), and red foxes {Vulpes vulpes) are abundant near the reservoir Moreover, the nesting area is flooded in late May when Figure 1 . Map that shows the location of Deer Creek Res- ervoir. The circle at the northern tip of the reservoir iden- tifies the location of the Osprey ground nest. The nest site was inundated shordy after the cessation of egg laying. 258 Short Communications VoL. 35, No. 3 Figure 2. Osprey ground nest located at Deer Creek Reservoir, Utah. The large forked limb in the photo is grounded drift material that was already present when the Ospreys selected the nest site. spring runoff elevates the water level of Deer Creek Res- ervoir. At full capacity, which is normally attained in early June, Deer Creek Reservoir is approximately 3 m deep where the Ospreys were nesting (Fig. 2). Interestingly, the ground nest sites were selected over a vacant artificial nest platform 400 m away, utility poles, and a suitable snag <1 km away which was used by a pair of Ospreys that appeared suddenly in late June. The choice of ground nests over potential elevated sites in an area high- ly populated with mammalian predators and sparsely in- habited by other Ospreys is highly unusual for the species and constitutes a singular and most confounding nesting event. Resumen. — Las aguilas pescadoras {Pandion haliaetus) que anidan en el suelo generalmente se encuentran res- tringidas a islas fibres de depredadores. Sin embargo en 1999 una pareja de aguilas pescadoras anidaron en el suelo en el reservorio de Deer Creek, Utah, en donde los depredadores mamiferos son abundantes. La depre- dacion del sitio del nido durante la postura ocasiono su fracaso. [Traduccion de Cesar Marquez] Acknowi.edgments I thank C.S. Houston, M. Martell, and an anonymous referee for comments on the manuscript and J. Bird for producing the map. Literature Cited Bent, A.C. 1937. Life histories of North American birds of prey. Part 1. Smithsonian Inst., U.S. Natl. Mus. Bull 170, Washington, DC U.S.A. Hayward, C. L., C. Cottam, A. M. Woodbury, and H. H Frost. 1976. Birds of Utah. Great Basin Nat. Memoirs No. 1. Monson, C.S. 1996. A geographical review of the histor- ical and current status of Ospreys {Pandion haliaetus) in Utah. Great Basin Nat. 56:150-156. Palmer, R. 1988. Handbook of North American birds, Vol.4, diurnal raptors. Part 1. Yale Univ. Press, New Haven, CT U.S.A. Poole, A. 1989. Ospreys: a natural and unnatural history Cambridge Univ. Press, Cambridge, U.K. Received 2 December 2000; accepted 15 April 2001 September 2001 Short Communications 259 J. Raptor Res. 35(3):259-262 © 2001 The Raptor Research Foundation, Inc. Diet and Breeding Success of Eagle Owl in Southeastern Spain: Effect of Rabbit Haemorrhagic Disease Jose E. Martinez and Jose F. Calvo Departamento de Ecologia e Hidrologia, Facultad de Biologia, Universidad de Murcia, Campus de Espinardo, 301 00 Espinardo (Murcia), Spain Key Words: Eagle Owl, Bubo bubo; rabbit disease', Spain', abandoned territories', breeding success', diet. Food is one of the main factors influencing the breed- ing biology and population ecology of raptors (Newton 1979). Several studies have documented numerical and functional responses in breeding populations resulting from changes in prey abundance (Korpimaki and Norr- dahl 1991, Rohner 1996, Redpath and Thirgood 1999, Nielsen 1999). In generalist raptor species, individuals can respond to prey declines by remaining in nesting territories without reproducing (Southern 1970, Smith et al. 1981, Korpimaki et al. 1990), shifting their diet com- position (Steenhof and Kochert 1988), or reducing their brood size (Vihuela and Veiga 1992, Houston and Schmutz 1995, Steenhof et al. 1997). Rabbits (Oryctolagus cuniculus) are the main prey for many predators in European Mediterranean ecosystems including Eagle Owls (Bubo bubo) (Delibes and Hiraldo 1981, Rogers et al. 1994), but how changes in rabbit pop- ulations affect the diet and reproductive success of pred- ators has been poorly studied (Vihuela and Veiga 1992). In areas where rabbits are scarce. Eagle Owls have been reported to prey on small mammals, especially rodents (Donazar et al. 1989). Since 1988, a new viral disease, rabbit haemorrhagic disease (RHD), has affected rabbit populations in the Mediterranean area (Villafuerte et al. 1995). While the epizootic has been shown to affect the breeding success of Golden Eagles {Aquila chrysaetos) (Fernandez 1993) and red foxes (Vulpes vulpes) (Villafuerte et al. 1996), lit- tle information is available on its effects on Eagle Owls. In this paper, we present evidence for the effect of RHD on the diet and breeding success of Eagle Owls in south- eastern Spain. Study Area and Methods Our study was conducted in a mountainous area (0— 828 m) in the south of the Murcia Region (southeastern Spain) from 1987-91. The study area covered about 1300 km^ and was characterized by an arid and semiarid Med- iterranean climate (annual rainfall <350 mm). The Ea- gle Owl is fairly common in Murcia Region and the Ibe- rian Peninsula (Diaz et al. 1996, Sanchez-Zapata 1999). In southern Murcia, rabbit is its staple food, representing up to 50% of the prey items (Martinez et al. 1992). RHD was first recorded in the study area in autumn 1988 (Rog- ers et al. 1994), when it reduced the rabbit population by 75% (Etisa 1990, M.A. Sanchez pers. comm.). A sim- ilar mortality rate was also observed in the first outbreak of the disease in other localities of southern Spain (Peiro and Seva 1990). Since then, there have been no other detectable changes in land use or food supply in the area, although other factors such as overhunting may have contributed to the decline in rabbit numbers (Fernandez 1993). RHD reappears annually, mainly during spring and winter, but these outbreaks have not caused as great a mortality as the initial outbreak in 1988 (Etisa 1990, Villafuerte et al. 1995). From 1987-91, we located occupied owl territories by searching for suitable owl nest sites, listening for elicited vocalizations using recorded calls, and listening to spon- taneous vocalizations. We considered a territory to be va- cant when no owls were sighted or heard, no occupied nests were located, or no fresh prey remains, pellets, or droppings were observed. When pairs were not recorded in traditional sites, we searched suitable breeding habitats within a radius of 3 km of the tradition site to exclude the possibility that they moved to a new breeding site To estimate the reproductive success of breeding pairs of owls, a minimum of three visits were made to each site to confirm egg-laying, successful reproduction, brood size at fledging (fledglings/successful pair), and produc- tivity (number of young fledged/breeding pair). The dietary analysis of Eagle Owls was based on a sam- ple of 2026 prey items (1340 prey items before RHD and 686 prey items after RHD). Pellet contents were identi- fied by macroscopic comparison with skeletal and skin reference collections. Prey remains and loose bones were omitted because they overestimate the occurrence of larger prey (Mersmann et al. 1992, Real 1996). We iden- tified and counted each prey item using the most fre- quently found bone or feather to calculate the minimum number of individuals present (Olsson 1979). Pellets were obtained from nests after young fledged and from perch sites outside the breeding season. During the pe- riod December-March, no collections were made be- cause females were incubating. To compare yearly breeding parameters with the owl diet before and after the RHD outbreak, Kruskal-Walhs tests were used (Sokal and Rohlf 1969). We used a Chi- square test to assess differences in nest site occupancy before and after RHD. Statistical analyses were perfomed with STATISTIX (Analytical Software 1992). 260 Short Communications VoL. 35, No. 3 Table 1. Annual changes in breeding success of the Eagle Owl before (1987-88) and after (1989-91) rabbit haem- orrhagic disease (RHD). Year Pairs Vacant Territories 1.AYING Pairs Successful Pairs Producuvity Fledging Success Before RHD 1987 19 0 12 11 1.57 2.72 1988 19 0 17 14 2.36 3.21 After RHD 1989 17 0 7 7 0.94 2.28 1990 8 11 5 5 1.12 1.80 1991 9 10 8 8 1.55 1.75 Results and Discussion After the RHD epidemic, both the brood size and young fledged per breeding pair decreased significantly {H = 0.5942, P = 0.0034 and H = 25.6471, P < 0.001, respectively) (Table 1). The number of laying pairs and the number of successful pairs after RHD did not show significant variation. Before the occurrence of RHD, all the Eagle Owl nesting territories were regularly occupied whereas, after RHD, there was a significant decrease in occupancy (x^ = 18.74, P< 0.001). Rabbits were the main prey consumed by Eagle Owls during the overall period analyzed, but after the RHD outbreak, the proportion of rabbits in the diet decreased slightly (from 55.97% to 53.64%) (Table 2). Also, the proportion of alternative prey (e.g., Rattus spp.) in the diet before and after RHD did not differ significantly (from 15.75% to 24.49%). However, after RHD the pro- portion of pigeons (Columba spp.) and other mammals in the diet decreased significantly {H = 4.0102, P = 0.0441 and H = 12.1708, P < 0.001, respectively). After the outbreak of RHD, rabbits remained the principal prey of Eagle Owls despite the crash in rabbit densities and the fact that consumption of alternative prey species did not increase. Our findings differed from those of Fernandez (1993) and Manosa (1994) who found dietary shifts following the decrease in the rabbit population caused by the viral haemorrhagic dis- ease. The owls in our study may not have shown a func- tional response switching prey species because there was an increase in the availability of sick rabbits in spring and summer and/or alternative prey species were not available. Several authors have suggested that diseases such as RHD and myxomatosis facilitate the capture of rabbits by predators (Vinuela and Veiga 1992, Fernan- Table 2. Comparison of the diet of Eagle Owls before and after rabbit haemorrhagic disease (RHD). The number (A) and proportions (%) of prey found in pellets as well as the Kruskal-Wallis statistic {H) are indicated. Taxon Before RHD (1987-88) After RHD (1989-91) H N % N % Mammals 1078 80.4 592 86.3 Rabbits 750 56.0 368 53.6 0.0494 (NS)i Rats 211 15.7 168 24.5 2.4214 (NS) Small mammals 24 1.8 8 1.2 2.2171 (NS) Hedgehogs 79 5.9 47 6.8 3.1598 (NS) Other mammals 14 1.0 1 0.1 12.1708*** Birds 262 19.5 94 13.7 Galliformes 41 3.1 17 2.5 0.6753 (NS) Columbidae 41 3.1 10 1.5 4.0102* Corvidae 27 2.0 7 1.0 1.7138 (NS) Birds of prey 40 3.0 19 2.8 0.5346 (NS) Other birds 113 8.4 41 6.0 1.8100 (NS) Total 1340 686 ' NS = not significant; * F < 0.05; *** P < 0.001. September 2001 Short Communications 261 dez 1993, Villafuerte et al. 1996, Villafuerte and Vinuela 1999). Nesting productivity decreased after RHD indicating that the crash in the rabbit population negatively af- fected the breeding success of Eagle Owls. Sharp de- clines in food resources during the breeding season have a marked negative influence on the breeding suc- cess of predators (Steenhof and Kochert 1988, Fernan- dez 1993, Villafuerte et al. 1996, Steenhof et al. 1997), especially if the predator cannot find alternative prey (Korpimaki et al. 1990). After RHD, the number of ter- ritorial pairs laying eggs and the number of laying pairs that were successful were lower indicating that the de- crease in rabbits negatively affected the fecundity of Ea- gle Owls. Several authors have suggested that rabbit availability determines the number of pairs of Eagle Owls that begin breeding (Olsson 1979, Mikkola 1983, Donazar 1990, Serrano 2001); nevertheless, it was clear from our findings that this species can still breed when Its main prey decreases. However, we are unclear as to the minimum rabbit density which causes Eagle Owls to cease breeding. We found that the number of occupied nesting ter- ritories after RHD decreased by 50%, probably due to the virtual disapperance of rabhits. Our findings con- curred with previous studies that have found rabbit scar- city to have caused the extinction of Eagle Owls in Med- iterranean localities (Donazar and Ceballos 1984, Serrano 1998). Nevertheless, they contrasted with nu- merous other studies which found that most species of raptors remain on nesting territories but do not lay eggs during periods of low prey abundance (Southern 1970, Saurola 1989, Fernandez 1993, Steenhof et al. 1997). Because pairs can continue to occupy territories but not start breeding during periods of low prey density mak- ing their presence difficult to detect, we may have failed to locate some pairs that continued to occupy nesting territories after RHD. Consequently, our comments should be taken cautiously. Resumen, — Estudiamos el efecto de la neumonia vfrica del Conejo (NHV) sobre la dieta y el exito reproductor en una poblacion de Buho Real {Bubo bubo) del sureste de Espana. El conejo {Oryctolagiis cuniculus) y las ratas {Rattus spp.) fueron las presas mas importantes en la die- ta. Despues de la neumonia hemorragica del conejo, no se observaron cambios en el consumo de conejo y ratas. La proporcion de conejo en la dieta del Buho Real no fue afectada por el cambio de densidad de conejo, quizas debido a la mayor accesibilidad de los individuos enfer- mos durante la epizootia. La drastica reduccion de las poblaciones de conejo condujo a una fuerte disminucion de la productividad y la tasa de vuelo del Buho Real. El numero de parejas reproductoras disminuyo drastica- mente despues de la NHV, y la mayorfa de las parejas no fueron detectadas. Sospechamos que este hecho pudiera estar relacionado con el abandono de las zonas de nidi- ficacion por parte del Buho Real, debido a la fuerte dis- minucion de la poblacion de conejo y a una simultanea escasez de presas alternativas. [Traduccion de autores] Acknowledgments J.A. Martinez, A. Izquierdo, M. Carrete, L.L. Vizcaino, V. Penteriani, and S. Fabrizio kindly made many valuable recommendations for which we are most grateful. laTERATURE ClTED Anai.yticai. Software. 1992. Statistix version 4.0. user’s manual. St. Paul, MN U.S.A. Delibes, M. and F. Hiraldo. 1981. The rabbit as prey in the Iberian Mediterranean ecosystem. Pages 614—622 in K. Myers and C.D. Mclnnes [Eds.], Proc. world lagomorph conf. (1979). Univ. Guelph, Guelph, ON Canada. DiAZ, M., B. Asensio, and J.L. TellerIa. 1996. Aves ib- ericas. Vol. I. No Passeriformes. Editorial J.M. Reyero, Madrid, Spain. Donazar, J.A. 1990. Geographic variation in clutch and brood size of the Eagle Owl Buho bubo in the Western Palearctic. /. Ornilhol. 131:439-443. and O. Ceballos. 1984. Algunos datos sobre sta- tus, distribucion y alimentacion del buho real {Bubo bubo) en Navarra. Rapinyaires Mediterranis 2:246-254. , F. Hiraldo, M. Delibes, and R.R. Estrella. 1989. Comparative food habits of the Eagle Owl Bubo bubo and the Great Horned Owl Bubo virginianus in six Pa- learctic and Nearctic biomes. Ornis Scand. 20:298-306. Etisa, S.L. 1990. Plan de aprovechamiento cinegetico de la Region de Murcia. Agenda Regional para el Medio Ambiente y la Naturaleza, Murcia, Spain. Fernandez, C. 1993. Effect of the viral haemorrhagic pneumonia of the wild rabbit on the diet and breed- ing success of the Golden Eagle Aquila chrysdetos (L.). Rev. EcoL Terre Vie 48:323-329. Houston, C.S. andJ.K. Schmutz. 1995. Declining repro- duction among Swainson’s Hawks in prairie Canada. /. Raptor Res. 29:198-201. Korpimaki, E. and K. Norrdahl. 1991. Numerical and functional responses of kestrels. Short-eared Owls, and Long-eared Owls to vole densities. Ecology 72:814- 826. , K. Huhtala, and S. Sulkava, 1990. Does the year-to-year variation in the diet of Eagle and Ural Owls support the alternative prey hypothesis?. Otkos 58:47-54. Manosa, S. 1994. Goshawk diet in a Mediterranean area of northeastern Spain./. Raptor Res. 28:84—92. Martinez, J.E., M.A. Sanchez, D. Carmona, J.A. Sanchez, A. Ortuno, and R. MartInez. 1992. Ecology and con- servation of the Eagle Owl Bubo bubo in Murcia, south- east Spain. Pages 84-88 in C.A. Galbraith, l.R. Taylor, and S. Percival [Eds.], The ecology and conservation of European owls. U.K. Nature Conservation No. 5 262 Short Communications VoL. 35, No. 3 Peterborough Joint Nature Conservation Committee, Peterborough, U.K. Mersmann, T.J., D.A. Buehler, J.D. Fraser, and J.K.D. Seegar. 1992. Assessing biases in studies of Bald Eagle food habits./. Wildl. Manage. 56:73-78. Mikkola, H. 1983. Owls of Europe. T. and A.D. Poyser Ltd., Calton, U.K. Newton, 1. 1979. Population ecology of raptors. T. and A.D. Poyser Ltd., Calton, U.K. Nielsen, O.K. 1999. Gryfalcon predation on ptarmigan: numerical and functional responses. J. Anim. Ecol. 68: 1034-1050. Olsson, V. 1979. Studies on a population of Eagle Owls. Viltrevy 11:1-99. Peiro, V. AND E. Seva. 1990. Le lapin de garenne dans la province d’Alicante (sud-est de I’Espagne). Bull. Mens. Off. Nat. Chase 182:189-196. Real, J. 1996. Biases in diet study methods in the Bo- nelli's Eagle./. Wildl. Manage. 60:632-638. Redpath, S.M. and S.J. Thirgood. 1999. Numerical and functional responses in generalist predators: Hen Harriers and peregrines on Scottish grouse moors./. Anim. Ecol. 68:879-892. Rogers, P.M., C.P. Arthur, and R.C. Soriguer. 1994. The rabbit in continental Europe. Pages 22-63 in H.V. Thompson and C.M. King [Eds.], The European rab- bit: history and biology of a successful colonizer. Ox- ford Univ. Press, Oxford, U.K. Rohner, C. 1996. The numerical response of Great Horned Owls to the snowshoe hare cycle: conse- quences of non-territorial floaters on demography./ Anim. Ecol 65:359-370. Sanchez-Zapata, J.A. 1999. Las aves rapaces y su relacion con la estructura del paisaje en ambientes mediter- raneos semiaridos. Tesis Doctoral, Universidad de Murcia, Spain. Sauroia, P. 1989. Breeding strategy of the Ural Owl Strix uralensis. Pages 235-240 in B.-U. Meyburg and R.D. Ghancellor [Eds.], Raptors in the modern world. World Working Group for Birds of Prey, Berlin, Ger- many. Serrano, D. 1998. Diferencias interhabitat en la alimen- tacion del Buho Real {Bubo bubo) en el valle medio del Ebro (NE de Espaha) : efecto de la disponibilidad de conejo {Oryctolagus cuniculus). Ardeola 45:35-46. . 2001. Parametros reproductivos y fecha de puesta del Buho Real {Bubo bubo) en el Valle del Ebro. Rocin, Anuario Ornitologico de Aragon 1998—1999. In press. Smith, D.G., J.R. Murphy, and N.D. Wofeinden. 1981. Relationships between jackrabbit abundance and Fer- ruginous Hawk reproduction. Condor 83:52-56. SoKAL, R.R. and F.J. Rohlf. 1969. Biometry. W.H. Free- man and Co., San Francisco, CA U.S.A. Southern, H.N. 1970. The natural control of a popula- tion of Tawny Owls {Strix aluco).J. Zool. {Lond.) 162: 197-285. Steenhof, K. and M.N. Kochert. 1988. Dietary respons- es of three raptor species to changing prey densities in a natural environment./. Anim. Ecol. 57:37-48. , , and T.L. McDonald. 1997. Interactive effects of prey and weather on Golden Eagle repro- duction. / Anim. Ecol. 66:350-362. Villafuerte, R. and j. Vinuela. 1999. Size of rabbits con- sumed by Black Kites increased after a rabbit epizo- otic. Mammal Rev. 29:261-264. , C. Calvete, J.C. Blanco, and J. Lucientes. 1995 Incidence of viral haemorrhagic disease in wild rabbit populations in Spain. Mammalia 59:651-659. , D.F. Lugo, C. Gortazar, and J.C. Bianco. 1996. Effect on red fox litter size and diet after rabbit haem- orrhagic disease in north-eastern Spain. /. Zool. {Lond.) 240:764-767. Vinuela, J. and J.P. Veiga. 1992. Importance of rabbits in the diet and reproductive success of Black Kites in southwestern Spain. Ornis Scand. 23:132-138. Received 30 December 2000; accepted 19 May 2001 Letter J Raptor Res. 35(3):263-264 © 2001 The Raptor Research Foundation, Inc. First Record of Tandem Flying in the King Vulture (Sarcoramphus papa) Tandem flight is a distinct type of flight formation in which two birds soar for several seconds with one immediately above the other (Mundy et al. 1992, The vultures of Africa, Acorn Books, Randburg, South Africa). It has been described in several griffon vulture species (Pennycuick 1972, Ibis 114:178-218; Vernon et al. 1982, Vulture News'! A'l , Mouze and Bagnolini 1995, Can J. Zool. 73:2144-2153). In Eurasian Griffons {Gyps fulvus), tandem flights most frequently occur during the breeding season and most often they involve mates, either currently or formerly paired, with the female occupying the upper position (Mouze and Bagnolini 1995). However, in Cape Griffons (G. coprotheres) , tandem flight does not necessarily involve breeding and it frequently takes place when birds return from foraging (Vernon et al. 1982, Mundy et al. 1992). Several functions of tandem flight have been proposed, including pair bond strengthening (Mouze and Bagnolini 1995), synchronization of breeding, and the establishment of social dominance (Mundy et al. 1992). Tandem flight has not previously been reported to occur in cathartid vultures. My observations of tandem flight in King Vultures {Sarcoramphus papa) were made during a long-term study con- ducted at Hato Las Nieves (6°35'N, 66°12'W), Sabana Nueva, Estado Bolivar, Venezuela between 1994—2000. The valley of Las Nieves is about 20 km long and 8-9 km wide and is dominated by lowland shrub savanna at elevations ranging from 220-260 m. Bordering mountains are covered with undisturbed primary forest and reach elevations of 1600 m on the west and 1880 m on the north. All observations were made at this site on two different occasions using 10 X 50 binoculars. On 13 July 1995 at 1005 H on a clear day, a juvenile King Vulture, probably 3-4 yr old judging from its white underparts (Clinton Eitniear 1996,/. Raptor Res. 30:35-38), was sighted approaching Las Nieves from the southwest An adult was higher and several body lengths behind, having come from a more westerly direction. As the two birds reached the ridge, the adult swooped down very close to the juvenile positioning itself for about 2-3 sec immediately above the young vulture. Approximately 1.5 km from the ridge, the adult lowered its feet, swooped down a second time, and held a close tandem position for about 2 sec. As the two separated, the adult soared eastward over the valley and the juvenile rapidly lost altitude in a gliding descent across the foothills, and then continued to glide above the gallery forest for >1 km until it disappeared from view. On 17 August 1997 at 1222 H on a slightly hazy day, two pairs of King Vultures soared into the valley from the southwest. The two pairs flew in parallel separated by 4-5 wingspans (8-10 m) with one bird in each pair positioned about 1.5 wingspans above and slightly behind its partner. The vultures maintained strict formation while gliding and turning abruptly, often clockwise, at right angles relative to the ground. Their flight, which took place over the foothills, was similar to that described for African vultures gliding into a landing by flying upwind, crosswind, and downwind (Tucker 1991, Ibis 108:1-7). After briefly circling on a thermal at the edge of the savanna, the four King Vultures paired again and took up the previous formation. The pairs then went east for about 1.5 km. During this segment of the flight, the upper bird in the pair nearest me suddenly lowered its feet, swooped down, and flew in close tandem with its partner for 2-3 sec before regaining the formation. Shortly afterwards all four vultures abruptly turned and went south for about 1 km, turned again, and went southwest. During the latter segment, the upper bird in the pair nearest me suddenly lowered its feet and again dropped down to a close tandem position for 2-3 sec. It resumed flight formation and the two pairs continued straight ahead crossing the ridge near their point of arrival. Since the flrst King Vulture tandem flights took place at considerable distance, I could not determine if they involved aggressive interactions. Nevertheless, aggressive aerial pursuits are common in Andean Condors {Vultur gryphus) with adults diving vertically on immature birds trying to strike them on their backs with their feet (McGahan 1972, Biology and ecology of the Andean Condor, Ph.D. dissertation, Univ. Wisconsin, Madison, W1 U.S.A.). In fact, the upper bird in Cape Griffon tandems often tries to strike the lower one with its feet and may even knock it off course (Mundy et al. 1992). On the other hand, tandem flights in Eurasian Griffons are often terminated when one bird deviates from its rectilinear path even in the absence of physical contact (Mouze and Bagnolini 1995). The strict intra- and inter-pair formation maintained between the two pairs of vultures during the second obser- vation led me to believe that the flight behavior was an aerial display, possibly linked to breeding. Twice during the same period (June-August), I saw an adult King Vulture briefly take up the typical courtship posture (Schlee 1987, C R. Acad. Sci., Paris, Ser. Ill, 304:207-212). The two tandem episodes within the display did not appear to be ag- 263 264 Letter VoL. 35, No. 3 gressive, although it cannot be ruled out that the upper bird was trying to strike the lower bird with its feet and thus exert social dominance. It is plausible that the two tandem flights may have been performed for aerodynamic pur- poses to allow the upper bird to keep the spacing or the speed necessary for maintaining strict flight formation. I would like to thank I. de Angelis and the late Y. Carbonell for their hospitality at Las Nieves and permission to conduct the King Vulture study on their property. — Marsha A. Schlee, Museum National d’Histoire Naturelle, Me- nagerie du Jardin des Plantes, 57 rue Cuvier, 75005 Paris, France. BOOK REVIEWS Edited by Jeffrey S. Marks Buteo Books is pleased to sponsor the Book Review section of the Journal of Raptor Research. Buteo Books stocks a comprehensive selection of ornithology books, both new and used. J. Raptor Res. 35(3):265— 266 © 2001 The Raptor Research Foundation, Inc. Island Eagles: 20 Years of Observing Golden Ea- gles on the Isle of Skye. By Ken Crane and Kate Nellist. 1999. Cartwheeling Press, Glenbrittle, Isle of Skye, U.K. 142 pp., color cover, 26 drawings, 3 fig- ures. ISBN 0-9536033-0-X. Paper, £10 (ca. $16.00).— Modern ecology is a world of mathematical models and statistical inferences that by and large can be pretty dull fare. Thus, it is with great pleasure that I introduce this little volume, which is a treasure trove of previously undescribed Golden Eagle {Aq- uila chrysaetos) behavior. Although the book is well organized in its presentation of topics, they are wo- ven so naturally into the fabric of the authors’ ex- periences that the book reads like a novel. We are drawn into the story as the old male at one territory slowly succumbs to avian tuberculosis, or as a wid- owed female chooses a new mate. So much here is new to science that this volume almost eclipses even the father of Golden Eagle naturalistic observations, Seton Gordon himself. The value of this book as a source of behavioral observations is enhanced by the authors not having restricted their observations to the breeding season. Most of what is available on Golden Eagle behavior and ecology comes from visits to occupied nests, and little comes from nonbreeding seasons that are much more difficult to study. Here, the authors em- phasize subjects such as pair formation, responses to intruding eagles in winter, and interactions with mammalian predators. Not only do the authors pro- vide a description of what eagles did, they give con- text so that we learn much about the function and rarity of unusual behavior. The experienced field biologist will detect, from the way each story is re- lated, that the observers are objective in their de- scriptions and conservative in their interpretations. What’s more, many of the anecdotes are illustrated with charming pencil drawings by the authors. These illustrations are not detailed but appear to be drawn from photographs, so they provide an accu- rate record of behavioral positions. Ironically, the only failing of the book is also one of its strengths. The authors appear to be largely unaware of the scientific literature on the Golden Eagle, especially work not done in western Europe. Although the practice of planning a study from the most up-to-date literature has merit, something can also be said for tackling a scientific study with a novel perspective, unencumbered by preconcep- tions that come from a thorough review of the lit- erature. Indeed, this little volume is testimony to the value of low-tech in that we learn what can be accomplished by two people with binoculars, note- books, a telescope, and little else. Little else, but incredible will. No doubt this volume will be most highly revered by those who understand what it means to stay the watch when a storm glides in and covers the mountain and the rain turns to sleet and then snow. To have persisted in such a study on those soggy moors for five years is remarkable; to have continued for 20 years has few parallels in the world of raptor ecology. So, what do these rain-soaked highlanders have to offer? In addition to a well-written and most in- teresting text, the book presents the following nov- el observations: (1) ground nesting (without a nest); (2) novel plumage features (e.g., a juvenile with a tail white to the tip, and a bird with asym- metrical plumage — lighter on one side than the other); (3) juvenile Golden Eagles playing in their first autumn snow; (4) talon grappling with White- tailed Eagles (Haliaeetus albicilla) , and the occasion- al displacement of Golden Eagles by this species; (5) using the sound of sea breeze and surf to cam- ouflage an eagle’s attack; (6) both adults lying down, tail-to-tail, apparently resting; (7) behavior of newly formed pairs, (8) role reversal in sexual solicitation; (9) observations of Golden Eagles kleptoparasitizing foxes; and (10) more and better observations of the roller coaster courtship flights 265 266 Book Reviews VoL. 35, No. 3 (undulations) than can be had in any other pub- lication. The authors have drawn from more than 500 undulation bouts and describe an extreme bout of 143 undulations. Furthermore, they dia- gram circular and figure-of-eight undulation pat- terns and undulation flights in two contexts, court- ship and territorial defense. The long duration of their study allowed them to document the survival of some eagles in excess of 20 yr. Although few students of Aquila eagles have seen more than a dozen kills, these authors report details from 26 prey captures, including observations of killing as an apparent displacement activity with no attempt to feed on the prey. Some philosophical points are also worth men- tion and emulation. The authors appropriately with- hold details of eyrie locations to protect the privacy (and the survival) of the birds. They also repeatedly encourage watching (gathering data) from far enough away to avoid disturbance. Unfortunately, they do not specify safe distances. From my own ob- servations, I know that it is best to begin open field observations from about 1 km (farther if the birds have been subject to shooting) and to move closer only after the birds show little attention to the ob- server and occasionally drift closer themselves. Louis Agassiz, the great 19th century glaciologist and ichthyologist, put it this way: “If a man studies nature from books alone, he will not know her when he meets her in the woods and fields.” Those of you with extensive knowledge of raptors from persistent fieldwork will recognize the value of this little book immediately, and those of you without such experiences will perhaps be inspired after reading this book to put down the “mouse” and turn off the computer. — David H. Ellis, USGS Pa- tuxent Wildlife Research Center, Laurel, MD 20708 U.S.A. J Raptor Res. 35(3):266-267 © 2001 The Raptor Research Foundation, Inc. Raptor Watch: A Global Directory of Raptor Mi- gration Sites. Edited by Jorje I. Zalles and Keith L. Bildstein. 2000. BirdLife Conservation Series No. 9, BirdLife International, Cambridge, U.K. xviii + 419 pp., 22 black-and-white photos, 21 tables, 21 figures, 3 appendices. ISBN 0-946888-38-8. Cloth, $58.00. — In 1988, Hawk Mountain Sanctuary launched a global conservation initiative entitled “Hawks Aloft Worldwide.” In response to this ini- tiative, more than 800 raptor biologists from around the world submitted information on poten- tial and ongoing raptor migration watchsites. This book is a compilation of the information submitted by those biologists and serves as a companion to the Raptor Migration Watch-Site Manual published by the Hawk Mountain Sanctuary Association in 1995 (see J. Raptor Res. 30:52, 1996) . Raptor Watch presents a tremendous amount of information that is skillfully summarized in an ac- cessible format. Introductory chapters contain ta- bles that list, among other things, the continental distribution and migration status (complete vs. par- tial vs. irruptive migrants) of all migratory species, species of global conservation concern according to BirdLife International, the regional origin of breeding populations of migratory species, and the countries of occurrence of all taxa that are of con- servation concern. One can turn to the “Global Analyses” chapter and instantly learn that 183 of the world’s 292 species of diurnal raptors are mi- gratory; that the highest numbers of migratory spe- cies occur in Asia (66 species) and Africa (61); and that 388 raptor watchsites have been identified around the world, 252 of which occur on protected lands. Following this information is a table that lists each watchsite that has reported at least 10,000 mi- grating raptors annually. I could go on and on. Suf- fice it to say that if you wish to know where to go to see migrating raptors, or where certain species are most likely to occur, this book will either pro- vide the answer directly, or give you the informa- tion to answer your question with little additional effort. The bulk of the book, some 312 pages, consists of descriptions of the 388 watchsites organized by country. In most cases, a “country description” also provides information on the size (km^) of the country, the length of its coastlines, human popu- lation size and population growth rate, per capita GNP, the names of bordering countries, and major land uses. The watchsite descriptions themselves run from about one-half to one and one-half pag- es, furnishing information on location by latitude and longitude, elevation, a site description, land tenure and protection, land use, threats, monitor- ing activity, main periods of migration, raptor spe- cies present, research and conservation activities, a September 2001 Book Reviews 267 list of contacts for the site, and the criteria that resulted in inclusion of the site in the global direc- tory. The book contains no index, which is a minor inconvenience when searching for specific sites in the United States and Canada (emphasis on “mi- nor”). Aside from that, I found nothing to criti- cize, because this book is excellent, both in content and in production. Raptor Watch represents an as- tounding achievement that will be a valuable con- tribution to conservation efforts for migrating fal- coniforms. Moreover, it is an extremely handy reference on the distribution and conservation sta- tus of the world’s diurnal raptors. The editors, compilers, BirdLife International, Hawk Mountain Sanctuary, and the 800-plus contributors to this book should be genuinely proud of their accom- plishment. — Jeff Marks, Montana Cooperative Wildlife Research Unit, University of Montana, Missoula, MT 59812 U.S,A. BUTEO BOOKS The following Birds of North America Species Accounts are available through Buteo Books, 3130 Laurel Road, Shipman, VA 22971. TOLL-FREE ORDERING: 1-800-722-2460; FAX: (804) 263-4842. E-mail: alien® buteobooks.com Barn Owl (1). Carl D. Marti. 1992. 16 pp. Boreal Owl (63). G.D. Hayward and P.H. Hayward. 1993. 20 pp. Broad-winged Hawk. (218). L.J. Goodrich, S.C, Crocoll and S.E, Senner. 1996. 28 pp. Burrowing Owl (61), E.A. Haug, BA. Millsap and M.S. Martell. 1993. 20 pp. Common Black-Hawk (122). Jay H. Schnell. 1994. 20 pp. Cooper’s Hawk (75). R.N. Rosenfield and J. Bielefeldt. 1993. 24 pp. Crested Caracara (249). Joan L. Morrison. 1996. 28 pp. Eastern Screech-owl (165). Frederick R. Gehlbach. 1995. 24 pp. Ferruginous Hawk (172). MarcJ. Bechard and Josef K. Schmutz. 1995. 20 pp. Flammulated Owl (93). D. Archibald McCallum. 1994. 24 pp. Great Gray Owl (41). Evelyn L. Bull and James R. Duncan. 1993. 16 pp. Great Horned Owl (372). C. Stuart Houston, Dwight G. Smith, and Christoph Rohner. 1998. 28 pp. Gyrfalcon (114). Nancy J. Clum and Tom J. Cade. 1994. 28 pp. Harris’ Hawk (146). James C. Bednarz. 1995. 24 pp. Long-eared Owl (133). J.S. Marks, D.L. Evans and D.W. Holt. 1994. 24 pp. Merlin (44). N.S. Sodhi, L. Oliphant, P. James and I. Warkentin. 1993. 20 pp. Mississippi Kite (402). James W. Parker. 1999. 28 pp. Northern Saw-whet Owl (42). Richard J. Cannings. 1993. 20 pp. Northern Goshawk (298). John R. Squires and Richard T. Reynolds. 1997. 32 pp. Northern Harrier (210). R. Bruce MacWhirter and Keith L. Bildstein. 1996. 32 pp. Northern Hawk Owl (356). James R. Duncan and Patricia A. Duncan. 1998. 28 pp. Red-shouldered Hawk (107). Scott T. Crocoll. 1994. 20 pp. Red-tailed Hawk (52). C.R. Preston and R.D. Beane. 1993. 24 pp. Short-eared Owl (62). D.W. Holt and S.M. Leasure. 1993. 24 pp. Snail Kite (171). P.W. Sykes, Jr., J.A. Rodgers, Jr. and R.E. Bennetts. 1995. 32 pp. Snowy Owl (10). David F. Parmelee. 1992. 20 pp. Spotted Owl (179). R.J. Gutierrez, A.B. Franklin and W.S. Lahaye. 1995. 28 pp. Swainson’s Hawk (265). A. Sidney England, MarcJ. Bechard and C. Stuart Houston. 1997. 28 pp. Swallow-tailed Kite (138). Kenneth D. Meyer. 1995. 24 pp. Turkey Vulture (339). David A. Kirk and Michael J. Mossman. 1998. 32 pp. White-tailed Hawk (30). C. Craig Farquhar. 1992. 20 pp. White-tailed Kite (178). Jeffrey R. Dunk. 1995. 16 pp. Buteo Books stocks all published species accounts, not only those covering raptors. The current list in taxo- nomic order may be viewed at; http://www.buteobooks.com Buteo Books stocks the Handbook of the Birds of the World. The first five volumes of this projected 12-volume work have been published including: Volume 2: New World Vultures to Cuineafowl (1994) covering the diurnal raptors and Volume 5: Barn Owls to Hummingbirds (1999) covering owls. These volumes are priced at $185 each plus shipping and handling. Usually available from Buteo Books, the classic reference on diurnal birds of prey: Brown, Leslie and Dean Amadon. Eagles, Hawks and Falcons of the World. Country Life Books, 1968. Two volumes. First English edition in brown cloth. Fine in slipcase. $300.00 and other editions at lesser prices. A Telemetry Receiver Designed with The Researcher in Mind What you've been waiting for! Finally, a highly sensitive 999 channel synthesized telcmetiy receiver that v»elghs less than 13 ounces, is completely user programmable and offers variable scan rates over all frequencies. For each animal being tracked, the large LCD display provides not only the frequency (to lOOHz) and channel numb^, but also a 7 character alphanumeric comment field and a digital signal strength meter. Stop carrying receivers that are the size of a lunch box or cost over SI 500. The features and performance of the new R-1000 pocket sized telemetry receiver will impress you, and the price will convince you. Other Features Include: • Factory tuned to any 4MHz wide segment in the 148174MHz Band • Very high sensitivity of -148dBm to •ISOdBm • Illuminated dispUy and keypad for use in low light or darkness • User selectable scan rates from 1-30 seconds in 1 second steps • Rechargeable batteries operate the receiver for 12 hours and can be replaced vvith standard AA Alkaline batteries In the held. Both 12vdc and llOvac chargers are included. • 6.r (15.5cm) high. 2.6“ (6.6cm) wide, 1.5“ (3.8cm) deep. • 3 year warranty • 1 day delivery S695.00 Please specify desired 4MHz wide segment in the 148-174MHZ band Visit our website for complete speciheations, operating manual and information on the R‘1000 or call our tolUfree number to order your receiver now. Try the New R-1000 and You'll Be Impressed! COMMUNICATIONS SPECIALISTS, INC. i26 W»il T,(t Awnue • Orangt. CA 92B6S-A296 • 1-714-998-3021 • fax 1-7 14-974- J420 Entire U.S.A. (800) 854-0547 • Fax (800) 850-0547 • http://www.com-spec.com 2001 ANNUAL MEETING The Raptor Research Foundation, Inc. 2001 annual meeting will be held on 25-30 October in Winnipeg, Manitoba, Canada. For information about the meeting contact Jim Duncan, Biodiversity program. Wildlife Branch, Manitoba Natural Resources, Box 24, 200 Saulteaux Crescent, Winnipeg, MB R3J 3W3 Canada. Email jduncan@nr.gov.mb.ca. Persons interested in predatory birds are invited to join The Raptor Research Foundation, Inc. Send requests for information concerning membership, subscriptions, special publications, or change of address to OSNA, P.O. Box 1897, Lawrence, KS 6604T8897, U.S.A. The Journal of Raptor Research (ISSN 0892-1016) is published quarterly and available to individuals for $33.00 per year and to libraries and institutions for $50.00 per year from The Raptor Research Foundation, Inc., 14377 117th Street South, Hastings, Minnesota 55033, U.S.A. (Add $3 for destinations outside of the continental United States ) Periodicals postage paid at Hastings, Minnesota, and additional mailing offices. POSTMASTER: Send address changes to The Journal of Raptor Research, OSNA, P.O. Box 1897, Lawrence, KS 66044-8897, U.S.A. Printed by Allen Press, Inc., Lawrence, Kansas, U.S.A. Copyright 2001 hy The Raptor Research Foundation, Inc. Printed in U.S.A. © This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). Raptor Research Foundation, Inc., Awards Recognition for Significant Contributions^ The Dean Amadon Award recognizes an individual who has made significant contributions in the field of systematics or distribution of raptors. Contact: Dr. Clayton White, 161 WIDE, Department of Zoology, Brigham Young University, Provo, UT 84602 U.SA.. Deadline August 15. The Tom Cade Award recognizes an individual who has made significant advances in the area of captive propagation and reintroduction of raptors. Contact: Dr. Brian Walton, Predatory Bird Research Group, Lower Quarry, University of California, Santa Cruz, CA 95064 U.S.A. Deadline: August 15. The Fran and Frederick Hamerstrom Award recognizes an individual who has contributed significantly to the understanding of raptor ecology and natural history. Contact: Dr. David E. Andersen, Department of Fisheries and Wildlife, 200 Hodson HaU, 1980 FolweU Avenue, University of Minnesota, St. Paul, MN 55108 U.S.A. Deadline: August 15. Recognition and Travel Assistance The James R. Kophn Travel Award is given to a student who is the senior author of the paper to be presented at the meeting for which travel funds are requested. Contact: Patricia A. Hall, 5937 E. Abbey Road, Flagstaff, AZ 86004 U.S.A. The WUliam C. Andersen Memorial Award is given to the student who presents the best paper at the annual Raptor Research Foundation Meeting. Contact: Ms. Laurie Goodrich, Hawk Mountain Sanctuary, Rural Route 2, Box 191, Kempton, PA 19529-9449 U.S.A. Deadline: Deadline established for meeting paper abstracts. Grants^ The Stephen R. TuUy Memorial Grant for $500 is given to support research, management and conservation of raptors, especially to students and amateurs with limited access to alternative funding. Contact: Dr. Kimberly Titus, Alaska Division of Wildlife Conservation, P.O. Box 20, Douglas, AK 99824 U.S.A. Dead- line: September 10. The Leslie Brown Memorial Grant for $500-$ 1,000 is given to support research and/or the dissemination of information on raptors, especially to individuals carrying out work in Africa. Contact: Dr. Jeffrey L. Lincer, 1220 Rosecrans St. #315, San Diego, CA 92106 U.S.A. Deadline: September 15. ^Nominations should include: (1) the name, title and address of both nominee and nominator, (2) the names of three persons qualified to evaluate the nominee’s scientific contribution, (3) a brief (one page) summary of the scientific contribution of the nominee. ^Send 5 copies of a proposal (<5 pages) describing the applicant’s background, study goals and methods, anticipated budget, and other funding.