(ISSN 08924016) y The Journal OF TOR Research I Volume 32 June 1998 Number 2 Contents Barred Owl Range Expansion into the Central Idaho Wilderness. Anthony L. Wright and Gregory D. Hayward 77 Estimating Core Ranges: A Comparison of Techniques Using the Common Buzzard {BUTEO BUTEO) . Kathy H. Hodder, Robert E. Kenward, Sean S. Walls and Ralph T. Clarke 82 Home Range Size and Habitat Requirements of Peregrine Falcons on the Cape Peninsula, South Africa. Andrew R. Jenkins and Grant a. Benn 90 Seasonal Patterns of Habitat Use by Snail Kites in Florida. Patricia l. Valentine-Darby, Robert E. Bennetts and Wiley M. Kitchens 98 Characteristics of Spotted Owl Habitat in Landscapes Disturbed by Timber Harvest in Northwestern California, rj. Gutierrez, John e. Hunter, Gilbert© Chavez-Leon and John Price 104 Food Habits and Hunting Ranges of Short-eared Owls {Asio flammeus) in Agricultural Landscapes of Southern Chile. David r. Martinez, Ricardo a. Figueroa, Carmen L. Ocampo and Fabian M. Jaksic Ill Winter Roost Sites of Northern Harriers and Short-eared Owls on Illinois Grasslands. Jeffery w. Walk 116 Size Variation of Migrant Bald Eagles at Glacier National Park, Montana. B. Riley McClelland, David S. Shea, Patricia T. McClelland and David A. Patterson 120 A Possible New Subspecies of the Philippine Hawk-Eagle (Spizaetus PHILIPPENSIS) AND ITS FUTURE PROSPECTS. Monika Preleuthner and Anita Gamauf 126 Patterns of Egg and Clutch Size Variation in the Montagu’s Harrier. Beatriz Arroyo, Alain Leroux and Vincent Bretagnolle 136 Organochlorine Pesticides, PCBs and Mercury in Hawk, Falcon, Eagle AND Owl Eggs from the Lipetsk, Voronezh, Novgorod and Saratov Regions, Russia, 1992 — 1993 . C.J. Henny, V.M. Galushin, P.l. Dudin, A.V. Khmstov, A.L. Mischenko, V.N. Moseikin, VS. Sarychev and V.G. Turchin 143 Review of Hazards to Raptors from Pest Control in Sahelian Africa. James O. Keith and Richard L. Bruggers 151 Contents continued on outside back cover The Raptor Research Foundation, Inc. gratefully acknowledges a grant and logistical support provided by Boise State University to assist in the publication of the journal. Persons interested in predatory birds are invited to join The Raptor Research Foundation, Inc. Send requests for information concerning membership, subscriptions, special publications, or change of address to OSNA, P.O. Box 1897, Lawrence, KS 66044-8897, U.S.A. The Journal of Raptor Research (ISSN 0892-1016) is published quarterly and available to individuals for $33.00 per year and to libraries and institutions for $50.00 per year from The Raptor Research Foundation, Inc., 14377 ll7th 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.SA. Printed by Allen Press, Inc., I^wrence, Kansas, U.S.A. Copyright 1998 by The Raptor Research Foundation, Inc. Printed in U.S.A. 0 This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). THE JOURNAL OF RAPTOR RESEARCH A QUARTERLY PUBLICATION OF THE RAPTOR RESEARCH FOUNDATION, INC. VoL. 32 June 1998 No. 2 J. Raptor Res. 32(2):V7-81 © 1998 The Raptor Research Foundation, Inc. BARRED OWL RANGE EXPANSION INTO THE CENTRAL IDAHO WILDERNESS Anthony L. Wright Hornocker Wildlife Institute, HC-83 Running Creek Ranch, Cascade, ID 83611 U.S.A. Gregory D. Hayward Department of Zoology and Physiology, University of Wyoming, Laramie, WY 82071 U.S.A. Abstract. — During the past century the geographic range of the Barred Owl (Strix varia) has expanded in western North America beginning in Alberta and British Columbia and moving southward in the United States. The pattern of range expansion remains poorly documented and several untested ex- planatory hypotheses have been proposed. We surveyed for owls in central Idaho from 1980—95 and recorded the occurrence of Barred Owls (1980-93 surveys did not employ playback calls of Barred Owls) . Despite numerous surveys for owls across a broad range of habitats, we did not encounter Barred Owls until 1985. Thereafter, we heard them regularly at several sites in central Idaho. We encountered Barred Owls most frequently in upland mature and old-growth mixed conifer forests. We suggest that expanding Barred Owl populations reached central Idaho very recently, possibly in the early to mid- 1980s. Although earlier reports of expanding Barred Owl populations in western North America focused mainly on managed forest lands with extensive timber harvest, forest conditions within designated wil- derness also apparently support expanding Barred Owl populations. Range expansion by this species in western North America is especially surprising in light of the limited range of a close congener, the Spotted Owl {Strix occidentalis) , in the region. Key Words: Barred Owl; Strix varia; range expansion; habitat use, movement, wilderness. La expansion del rango de Strix varia hacia el centro de Idaho Resumen. — Durante el siglo pasado el rango geografico del buho Strix varia ha aumentado en el oeste de Norte America, comenzando desde Alberta y British Columbia, Canada, y continuando hacia el sur de los Estados Unidos. Extensiones de hordes geograficos, documentados, son escazas de datos esen- ciales y mas recientes, varias hipotesis explanatorias han sido propuestas sin ninguna prueba. Hicimos conteos de buhos en Idaho central durante los ahos 1980—95, la presencia de S. varia fue anotada utilizanda grabaciones de cantos repetitivos (entre 1980-93 no fue utilazada la repeticion de cantos grabados) . A pesar de que se realizaron conteos numerosos a traves de un rango amplio de habitats, S. varia no fue registrado hasta el ano 1985. Desde ese entonces, fueron registrados con frequencia en varios sitios. 5. varia fue encontrado con mas frequencia en bosques maduros de coniferas mixtos. Sugerimos que hace pocos anos llegaron las poblaciones de S. varia a Idaho central, posiblemente a principle o a mediados de la decada 1980. Aunque los reportes preliminarios sobre la expansion de poblaciones de S. varia en el oeste de Norte America se enfocaron en bosques administrados con aprovechamiento extensive de madera, los bosques de areas pristinas tambien sostienen poblaciones en expansion. La expansion del rango de esta especie es aun mas sorprendente en comparacion al rango limitado de su congener Strix occidentalis en la region. [Traduccion de David L. Anderson] 77 78 Wright and Hayward VoL. 32, No. 2 The Barred Owl {Strix varia) has been reported regularly in the northern Rocky Mountains of the U.S. since the late 1960s (Shea 1974, Taylor and Forsman 1976), expanding its range into the area from the adjacent mountains of British Columbia and Alberta (Grant 1966, Jones 1987). From 1912- 49, the Barred Owl was reported only five times in Alberta (Boxall and Stepney 1982). Early sightings were in mixed coniferous-deciduous boreal forests, but after 1950 the number of reports increased substantially and included breeding records from coniferous forests of the northern Rocky Moun- tains. The first Barred Owl was reported from Brit- ish Columbia in 1943, but by 1966 the species was considered a “common resident of a large part of the interior of the province” (Grant 1966). Prior to 1965, there were no records for Barred Owls from Washington, Oregon, or Idaho although there were several records from extreme northern Montana (e.g., Wright 1976). The expansion of Barred Owls in western North America, south of Canada, was documented by Taylor and Forsman (1976), Forsman (1988), Sharp (1989), and Ste- phens and Sturts (1991). Ellis et al. (1987) de- scribed the concomitant range expansion in Mon- tana as far south as the Bitterroot Valley. However, Stephens and Sturts (1991) reported it only as a transient (nonbreeding records) in central Idaho. We document the occurrence of Barred Owls at several dispersed sites in central Idaho since 1985 with strong evidence of breeding. Two nonexclusive hypotheses have been pro- posed to explain the range expansion of the Barred Owl into southwestern Canada and north- western U.S. Grant (1966) concluded that an eco- logical barrier, perhaps the Rockies, but more like- ly something farther east, was recently bridged. Boxall and Stepney (1982) attributed range expan- sion and increasing numbers in part to an “in- creased tolerance of coniferous forest.” In addi- tion to these hypotheses, Dunbar et al. (1991) in- dicated that Barred Owls have successfully colo- nized a wide range of habitats including old-growth and mature forests used by Spotted Owls {Strix oc- cidentalis ) . They went on to say that it was not clear if Spotted Owls were absent from particular areas logged in the early 1900s, “because of the removal of older forests or the presence of Barred Owls, or a combination of both factors.” We consider the two hypotheses and the observations of Dunbar et al. (1991) as they relate to our observations of Barred Owls in extensive wilderness areas. Study Area and Methods Our study area consisted of four geographically dis- persed sites in the 1.5 million ha of contiguous, federally designated wilderness that comprise the Selway-Bitter- root (SBW) and Frank Church — River of No Return (RNR) wilderness of central Idaho. Our study sites, which sampled the full range of elevations across this region, were Chamberlain Basin, Cold Meadows, and the Big Creek drainage in the RNR, and the Selway River and its tributaries from Moose Creek to Whitecap Creek in the SBW. Elevations at these sites range from 750-3000 m and topography varies from rolling plateau at Chamberlain Basin to high peaks and deep, rocky canyons in the Big Creek and Selway River drainages. The vegetation of cen- tral Idaho reflects elevation, topography, and moisture (which increases from south to north) gradients and is strongly influenced by fire history. Most of the landscape is a mosaic of conifer forests dominated by ponderosa pine {Pinus ponderosa) , grand fir {Abies grandis), Douglas- fir {Pseudotsuga menziesii) , lodgepole pine {Pinus contarta ) , Engelmann spruce {Picea engelmannii) and subalpine fir {Abies lasiocarpa). Open areas include brush lands, steep slopes of grasses and forbs, and wet meadows. Some ri- parian areas support diverse, deciduous vegetation. Fink- lin (1988) and Hayward et al. (1993) describe the vege- tation and climate of the region in more detail. From 1980 to the present, we conducted a variety of field studies in the wilderness of central Idaho (e.g., Hay- ward and Carton 1988, Quigley et al. 1989, Hayward et al. 1993, Wright and Kelsey 1997) some of which were directed at owls and some of which provided opportu- nities to observe owls incidental to other objectives. Dur- ing 1980-81, we surveyed for owls from mid January through May in the Big Creek drainage and at Cham- berlain Basin. We alternated between sites on a rotating 10-14-d period (Hayward and Carton 1988). Survey routes were selected to allow sampling of major forest habitats and topographic positions at each study site. During surveys, we broadcast recorded songs of several owl species including Creat Horned Owl {Bubo virgini- anus) but not Barred Owl. Calls of one to three owl spe- cies were broadcast at stations 0.3-0. 6 km apart on tran- sects traveled by walking or skiing. Approximately 80 night-time surveys were conducted during which seven species of owls were observed (see Hayward and Carton 1988). During January through April 1984-87, we surveyed for owls at Chamberlain Basin and Cold Meadows (Hay- ward et al. 1993). These surveys were associated with studies of the Boreal Owl {Aegolius funereus). Beginning when the first stars became visible, we played tape re- cordings of the staccato song of the Boreal Owl at 0.5-1- km intervals along trails and ridge lines. We remained at each calling station 10-12 min playing three series of staccato song with 2 min of silence after each series. In addition to time spent listening for owls at each calling station, we paused for 1 min at least once between sta- tions. Approximately 160 nocturnal surveys were con- ducted and nine species of owls were observed during these studies. During 1986 and January-March 1987, we recorded Barred Owl contacts made in the Big Creek drainage dur- June 1998 Barred Owl Range Expansion 79 Table 1. Barred Owl observations made in the Frank Church — River of No Return (RNRW) and Selway — Bitterroot Wildernesses (SBW) during surveys conducted 1980-95, Date Location Area Lat./Long. Elev. (m) Forest Type Description 4 Feb. 1985 Chamberlain Cr. RNRW 45°21'40"N 115°10'30'W 1830 Mixed conifer Pair, sang 28 March 1985 Chamberlain Cr. RNRW 45°21'40"N 115°10'30'W 1850 Mixed conifer Pair, sang 19 April 1986 Chamberlain Cr. RNRW 45°21'40"N 115°10'30'TV 1850 Mixed conifer Sang 8 June 1986 Rush Cr. RNRW 45°04'21"N 115°58'33'W 1850 Mixed conifer Adult seen 22 March 1987 No Name Cr. RNRW 45°20'50"N 115°13'30'W 2160 Mixed conifer Sang 1 April 1987 Chamberlain Cr. RNRW 45°21'40"N 115°10'30'W 1850 Mixed conifer Sang 7 August 1989 Goat Cr. SBW 45°58'00"N 114°53T3'W 1600 Mixed conifer Fledgling seen 6 June 1994 Crow Cr. SBW 45°59T1"N 114°46'06'W 1650 Mixed conifer Responded to song 18 June 1994 Elk Cr. SBW 45°59T1"N 114°51'38'W 1100 Mixed conifer Responded to song 4 July 1994 Eagle Cr. SBW 45°52'34"N 114“52'47"W 1300 Mixed conifer Responded to song 25 October 1994 Ditch Cr. SBW 45°59'02"N 114°59'08"W 2050 Spruce/Fir Responded to Boreal Owl song 30 September 1995 Gardiner Fork SBW 45°58'54"N 1500 Mixed conifer Sang ing unrelated field work. During November 1988-May 1994, we recorded Barred Owls observed incidentally in SBW. Most time was spent in early successional, low ele- vation (<1525 m) forests except during July and August when higher areas were visited. During June and July 1994, we gave vocal imitations of Barred Owl songs both morning and evening at six locations in the SBW in Douglas-fir/ grand fir/ponderosa pine forest (mixed co- nifer). During June and July 1995, we repeated this pro- cedure at seven more locations in the SBW in high ele- vation forests of lodgepole pine, spruce-fir, or whitebark pine {Pinus albicaulis) . Results Despite numerous surveys across a broad range of habitats, we did not encounter Barred Owls in the RNR in 1980, 1981, or 1984. We first heard a pair of Barred Owls on 4 February 1985 at a site in Chamberlain Basin which had been visited in earlier years (45°22'N, 115°07'W). We encountered Barred Owls at this same site again in 1986 and 1987 and at two other sites in Chamberlain Basin (Table 1). In 1986, we saw a Barred Owl southwest of Rush Creek Lookout about 36 km southeast of the earliest location in Chamberlain Basin. During surveys in the SBW targeting Barred Owls, they re- sponded to our vocalizations at 3 of 6 mixed-co- nifer sites in 1994 and 0 of 7 high elevation forest sites in 1995. The high elevation surveys may have sampled sites of lower habitat quality. We encountered Barred Owls most frequently (9 of 10 sites) in upland, mature to old growth, mixed-conifer forests (Table 1). Stands were most often dominated by an overstory of Douglas-fir and/ or grand fir although, on one site, ponderosa pine dominated the overstory. On many sites, dense patches of pole-size (diameter at breast height <25 cm) Douglas-fir occurred in the un- derstory. Several sites had patches (up to 1 ha) of dead, mature Douglas-fir that may have been killed by root rot. Discussion The pattern of Barred Owl detections during surveys from 1980-95 suggests that the southward expansion of this species into central Idaho may have occurred sometime in the early to mid-1980s. This conclusion must be accepted with caution, as Barred Owls may have occurred in our study areas 80 Wright and Hayward VoL. 32, No. 2 but were not detected during the first 3 yr of field surveys. Breeding populations of Barred Owls were not reported in central Idaho prior to our obser- vations. For instance, Burleigh (1972) did not list Barred Owls for the state although he noted rec- ords for more elusive species such as Boreal and Hawk (Surnia ulula) Owls. Although the other eight resident owl species were located during sur- veys and fieldwork of 1980, 1981, and 1984, we did not encounter Barred Owls until 1985. After 1985, we encountered Barred Owls each year that we conducted fieldwork in the RNR. When we searched for them in mixed conifer forest in the SBW in 1994, we found them to be widespread. Our observations of singing, multi-year persistence on sites, and a locally produced fledgling provided convincing evidence of Barred Owl breeding in central Idaho. What factors are responsible for the recent ex- pansion of Barred Owls in the Rocky Mountains? Changes in climate and/or forest cover may have permitted the westward expansion of Barred Owl populations through the mixed deciduous-conif- erous forests of northern Alberta into the Cana- dian Rockies as documented by Boxall and Stepney (1982) . This process parallels that proposed for the expansion of the Blue Jay ( Cyanocitta cristata) and other birds across the great plains along the ex- panding forests of the Platte and other river sys- tems (Knopf 1994). Once Barred Owls reached the Rockies, various mechanisms may have influenced the rate and pat- tern of spread through the region. By the mid- 1950s, less than 10 yr after the first reports in the region (Grant 1966), the species had become wide- spread in interior British Columbia. This seems a very short time frame to accommodate the genetic adaptation to coniferous forest suggested by Boxall and Stepney (1982). Because the initial coloniza- tion and much of the subsequent expansion of the Barred Owl has occurred outside the range of the Spotted Owl (Campbell and Campbell 1984, Fors- man et al. 1984), any hypothesis regarding habitat mediated changes in competition between the two species is not relevant east of the Cascades. The rate and extent of range expansion by Barred Owls is particularly striking in light of the absence of Spotted Owls from much of this region. Why didn’t this close congener, already endemic to the region and adapted to a variety of forest types across its range (e.g., Gutierrez and Carey 1985, Gutierrez et al. 1995, Verner et al. 1992), exploit forests of the northern Rockies first? One possible, untested hy- pothesis is that recent anthropogenic change cre- ated a niche for an owl of this type which the Barred Owl was better preadapted to exploit. If such anthropogenic changes in habitat have en- hanced the southward movement of Barred Owls, our observations suggest that such changes oc- curred both inside and outside of wilderness. Beginning about 1935, efficient fire suppression led to major changes in forest structure both in central Idaho (Habeck 1976, Steele et al, 1986, Barrett 1988) and elsewhere in the region (sum- mary in Morgan 1994). The extent and homoge- neity of mixed-conifer forests increased as ponder- osa pine stands were invaded by fir. As a conse- quence of fire suppression, closed canopy forests developed, dense patches of young conifers be- came common beneath mature trees, and tree mortality due to disease and insects increased. Based on this pattern of vegetation change and our observations of Barred Owl occurrence in central Idaho, we pose our untested hypotheses regarding the spread of Barred Owls in the region. We have no direct evidence that human-induced changes in forest structure facilitated the spread of the Barred Owl, but the changes are typical of sites where we observed owls. Compared to pre-1935 forests, per- haps these stands offer increased protection from Great Horned Owls and large Acdpiters, more suit- able dead tree nest sites, or higher prey densities. Present knowledge of Barred Owl habitat associa- tions in the region is too limited to allow an eval- uation of this scenario. Referring to Barred Owl, Holt and Hillis (1987) note only the importance of mature and old growth trees, particularly west- ern larch {Larix ocddentalis) , in western Montana. However, Dunbar et al. (1991) state that Barred Owls were most common in broad riparian corri- dors in southwestern British Columbia; these ri- parian forests may have similar, closed canopy structure to Barred Owl sites we observed in cen- tral Idaho. More refined data on the movement of Barred Owls into new habitat both within and outside the range of the Spotted Owl will be necessary to un- derstand the ecological factors most important in determining the range expansion of the former in western North America. Understanding character- istics of Barred Owl dispersal behavior will further aid interpretation. Our observations provide in- sight into the timing of this expansion, habitats be- ing occupied, and use of landscapes by Barred June 1998 Barred Owl Range Expansion 81 Owls in relatively unmanaged forest within desig- nated wilderness. Acknowledgments We thank S. Anderson, E. Forsman, J. Duncan, R. Gu- tierrez, and D, Holt for reviewing earlier versions of this manuscript, P.L. Wright for sharing useful insights, and M.G. Hornocker for making this paper possible. The U.S. Forest Service Rocky Mountain and Intermountain For- est and Range Experiment Stations, Idaho Department of Fish and Game, and University of Idaho Wilderness Research Center, also provided funding and support; we thank each of them. Literature Cited Barrett, S.W. 1988, Fire suppression’s effects on forest succession within a central Idaho wilderness. West. J. Appl. For. 3:76—80. Boxall, RC. and RH.R. Stepney. 1982. The distribution and status of the Barred Owl in Alberta. Can. Field- Nat. 96:46-50. Burleigh, T.D. 1972. Birds of Idaho. Caxton Printers, Ltd., Caldwell, ID U.S.A. Campbell, E.C. and R.W. Campbell. 1984. Status report on the Spotted Owl (Strix ocddentalis caurina) in Can- ada, 1983. Report to the Committee on the Status of Endangered Wildlife in Canada. Canadian Nature Federation, Ottawa, Ontario, Canada. Dunbar, D.L., B.P. Booth, E.D. Forsman, A.E. Hether- ington and D.J. Wilson. 1991. Status of the Spotted Owl Strix ocddentalis and Barred Owl Strix varia in southwestern British Columbia. Can. Field-Nat. 105: 464-468. Ellis, D.H., D.G. Smith and P.L. Wright. 1987. Barred Owl specimen records for Montana. Western Birds 18: 217-218. Finklin, A.I. 1988. Climate of the Frank Church, River of No Return Wilderness, central Idaho. USDA For. Serv. Gen. Tech. Rep. INT-240. Forsman, E.D. 1988. A survey of Spotted Owls in young forests in the northern Coast Range of Oregon. Murrelet 69:65-68. , E.C. Meslow and H.M. Wight. 1984. Distribu- tion and biology of the Spotted Owl in Oregon. Wildl. Monogr. 87:1-64. Grant, J. 1966. The Barred Owl in British Columbia. Murrelet 47:39-45. Gutierrez, R.J. AND B. Carey [Eds.]. 1985. Ecology and management of the Spotted Owl in the Pacific North- west. USDA For. Serv. Gen. Tech. Rep. PNW-185. , A.B. Franklin and W.S. LaHaye. 1995. Spotted Owl {Strix ocddentalis). Pages 1-28 in A. Poole and F. Gill [Eds.], The birds of North America, No. 179. Academy of Nat. Sci., Philadelphia, PA and Am. Or- nithol. Union, Washington, DC U.S.A. Habeck, J.R. 1976. Forests, fuels, and fire in the Selway- Bitterroot Wilderness, Idaho. Proc. Tall Timbers Fire Ecology Conf. 14:305-352. Hayward, G.H. and E.O. Carton. 1988. Resource parti- tioning among forest owls in the River of No Return Wilderness, Idaho. Oecologia (Berlin) 75:253-265. , P.H. Hayward and E.O. Carton. 1993. Ecology of Boreal Owls in the northern Rocky Mountains, USA. Wildl. Monogr. 124:1-59. Holt, D.W. and J.M. Hillis. 1987. Current status and habitat associations of forest owls in western Montana. Pages 281-288 in R.W. Nero, RJ. Clark, C.R. Knapton and RJ. Hamre [Eds.], Biology and conservation of northern forest owls: symposium proceedings. USDA For. Serv. Gen. Tech. Rep. RM-142. Jones, E.T. 1987. Early observations of Barred Owl in Al- berta. Blue Jay 45:31-32. Knopf, F.L. 1994. Avian assemblages on altered grass- lands. Stud. Avian Biol. 15:247—257. Morgan, P. 1994. Dynamics of ponderosa and Jeffrey pine forests. Pp. 47-73 in G.D. Hayward and J. Verner [Eds.], Flammulated, Boreal, and Great Gray Owls in the United States: a technical conservation assess- ment. USDA For. Serv. Gen. Tech. Rep. RM-253. Quigley, H.B., G.M. Koehler and M.G. Hornocker. 1989. Dynamics of a mountain lion population in cen- tral Idaho over a 20-year period (abstract). Page 54 in R.H. Smith [Ed.], Proc. of 3rd mountain lion work- shop, Ariz. Chapter The Wildlife Society, Phoenix, AZ U.S.A. Sharp, D.U. 1989. Range extension of the Barred Owl in western Washington and first breeding record on the Olympic Peninsula. J. Raptor Res. 23:179—180. Shea, D.S. 1974. Barred Owl records in western Montana Condor 76:222. Steele, R., S.F. Arno and K. Geier-Hayes. 1986. Wildfire patterns change in central Idaho’s ponderosa pine- Douglas-fir forests. West.]. Appl. For. 1:16—18. Stephens, D.A. and S.H. Sturts. 1991. Idaho bird distri- bution. Spec. Pub. No. 11, Idaho Museum of Nat. Hist., Pocatello, ID U.S.A. Taylor, A.L., Jr. and E.D. Forsman. 1976. Recent range extensions of the Barred Owl in western North Amer- ica, including the first records for Oregon. Condor'll: 560-561. Verner, J., K.S. McKelvey, B.R. Noon, RJ. Gutierrez, G.I. Could, jR- and T.W. Beck [Technical Coordi- nators]. 1992. The California Spotted Owl: a tech- nical assessment of its current status. USDA For. Serv. Gen. Tech. Rep. PSW-GTR-133. Wright, A.L. and R.G. Kelsey. 1997. Effects of spotted knapweed on a cervid winter-spring range in Idaho. J. Range Manage, (in press) . Wright, P.L. 1976. Further hird records from western Montana. Condor 78:418-420. Received 10 April 1997; accepted 4 February 1998 J. Raptor Res. 32(2):82-89 © 1998 The Raptor Research Foundation, Inc. ESTIMATING CORE RANGES: A COMPARISON OF TECHNIQUES USING THE COMMON BUZZARD {BUTEO BUTEO) Kathy H. Hodder and Robert E. Kenward Institute of Terrestrial Ecology, Furzebrook Research Station, Wareham, Dorset, BH20 3 AS, U.K. Sean S. Walls Biotrack, 52 Furzebrook Road, Wareham, Dorset, BH20 5 AX, U.K. Ralph T. Clarke Institute of Terrestrial Ecology, Furzebrook Research Station, Wareham, Dorset, BH20 5 AS, U.K. Abstract. — The need to describe the relative intensity with which an animal uses different parts of its home range has been recognized for at least half a century. Such descriptions are particularly important for wide-ranging raptors with home ranges covering a variety of habitats. In studies of many taxa, the description of internal range structure is addressed by describing a core range of most intensive use. However, there is still no broadly accepted definition of a core or method of objectively estimating core ranges. Here, we propose that a core range can be usefully defined by the exclusion of excursive activity with the assumption that behavior differs between core and excursive activities. Two methods of ex- cluding excursive activity are presented for winter ranges of the Common Buzzard (Buteo buteo) in lowland U.K. The first involves subjective exclusion of outlying locations, using the outermost discon- tinuity in the utilization distribution (UD). Incremental Cluster Polygons are used to produce the UD because this method provides the closest spatial relationship to the animal locations and the most clearly defined discontinuities. The potential for error or bias in this subjective method may often be unac- ceptable, particularly for home ranges which do not have well-defined core areas. The second method is a new application of incremental cluster analysis that objectively excludes excursive locations. The objective and subjective approaches are compared, and implications of core range definition in habitat and sociality analysis of raptors are explored in the context of published analyses on raptors and other taxa. Key Words: home range, core range, radiotelemetry; incremental cluster analysis; Buteo buteo. Estimacion de rangos c«„iiLiale^: Una comparacion de tecnicas utilizando a Buteo buteo Resumen. — La necesidad de describir la intensidad relativa en que un animal utiliza las diferentes partes de su rango de hogar ha sido reconocida por lo menos desde hace medio siglo. Estas descripciones son particularmente importantes para las aves de presa que tiene rangos amplios con una gran variedad de habitats. En los estudios de muchos taxones, la descripcion de la estructura del rango interno es abor- dada mediante la descripcion del rango central como el mas utilizado. No obstante, no existe aun una definicion aceptada del centro o de un metodo para estimar en forma objetiva los rangos centrales. Aqui proponemos que un rango central puede ser definido en forma util mediante la exclusion de las actividades de incursion. Dos metodos para excluir la actividad de incursion son presentados para los rangos de invierno de Buteo buteo en el Reino Unido. [Traduccion de Cesar Marquez] Radiotelemetry has been used to study raptors for nearly three decades, providing data for many aspects of ecological research (Kenward 1985a). However, improvement in the collection and anal- ysis of data has been slower than the technical de- velopments in radio-tracking (Lance and Watson 1980, Harris et al. 1990, Kenward 1992). The de- scription of the intensity with which animals use different parts of their home ranges presents a fun- damental problem (Hayne 1949). Animals gener- ally live in a spatially heterogeneous environment in terms of food availability, nest and roost sites, density of competitors, and other factors. There- fore, they tend to have one or more core areas of 82 June 1998 Core Range Estimates for Common Buzzards 83 intensive use in their home ranges (Kaufman 1962) and it is likely that their behavior will differ in the core and outer areas of the home range. Many raptors are wide-ranging and their outer home-range boundaries may enclose habitat known to be avoided (e.g., Stahlecker and Smith 1993) . Raptors may also make excursions from their usual range. Common buzzards {Buteo buteo), for instance, often make brief movements of up to 20 km during their first year (Walls and Kenward 1994) , before returning to ranges typically less than 1.1 km in diameter (Walls and Kenward 1995) . Therefore, an analysis method for core ranges should exclude excursive locations (Burt 1943) and areas within an outer home-range boundary that are avoided (White and Garrott 1990). A number of methods have been proposed for estimating range cores, but none are widely ac- cepted. Therefore, comparisons between studies is generally not possible. The process of finding a core range can be split into three stages: (1) the description of the inter- nal range structure, giving nominal cores; (2) the choice of a percentage inclusion of radio locations that selects a biologically meaningful core for each individual; and (3) setting a standard core range (in terms of a standard percentage of locations) for a sample of animals to allow statistical compar- ison within the sample. Here, we present a com- parison of techniques used to describe the core ranges of raptors. Collection of Data Common Buzzards in southern Dorset, U.K. were instrumented with radio-tags at the nest just before fledging. Each radio-tag weighed 30 g and was mounted as a backpack with 6-mm-wide Teflon ribbon (Bio track, 52 Furzebrook Rd., Wareham, U.K.). These tags had a life of up to 4 yr and a maximum range of 40 km from the ground and 80 km from the air. To avoid disturbance of the study animals, buzzard locations were determined by triangulation from roadsides. Error associated with the locations was estimated at about 100 m. Standard 30 location home ranges (three locations per d for 10 d [Kenward 1987]) were recorded for 122 buzzards from 1990-96. These included buz- zards aged between 1 and 4 yr. Data were collected in the nonbreeding season after the main dispersal period, when the buzzards had settled in relatively stable ranges. Data were analyzed using a modified version of RANGES V (Kenward and Hodder 1996). Description of Range Structure Several different methods have been used for de- scribing internal range structure. The efficacy of these methods can be judged by their conforma- tion to the locations, including their ability to con- form to multinuclear cores. A further major re- quirement is efficiency. To give time for data col- lection on a sufficient number of animals, most projects require an analysis method that can esti- mate the range structure from a minimum number of locations per animal. The grid cell approach (Siniff and Tester 1965, Abies 1969, Voigt and Tin- line 1980, Samuel et al. 1985, Samuel and Green 1988) conforms to location distribution. However, this method may require more than 150 locations to calculate a stable home range that does not in- crease in size as more locations are added (Don- caster and Macdonald 1991). The tessellation tech- nique proposed by Wray et al. (1992a) also re- quires a large number of locations. Contouring methods (Dixon and Chapman 1980, Worton 1989) stabilize with fewer than 50 locations. How- ever, their accuracy of fit to the locations is influ- enced by dependency on an arbitrary grid and the use of parametric estimation functions (Spencer and Barrett 1984, Kenward 1992, Wray et al. 1992b). Ellipses (Jennrich and Turner 1969) give stability with even fewer locations but conform poorly to the locations and can only provide one nucleus. Polygon-based techniques have stable out- er edges with 30 locations (Kenward 1982, Parish and Kruuk 1982, Kenward 1987); however, peeled convex polygons (Kenward 1985b, 1987) provide a poor fit to multinuclear or curved ranges (White and Garrott 1990). These problems may be avoid- ed with Incremental Cluster Polygons (ICP) (Ken- ward 1987). ICP analysis is based on forming groups of locations and separating outliers. Convex polygons drawn around the clusters provide a range outline that is not influenced by a grid or the position of outlying locations (Kenward 1987). Also, the outlines produced by elimination of out- lying locations stabilize at less than 50 locations (Kenward 1992) . Therefore, we adopted ICP as the method for estimating internal range structure in this paper. Selection of a Biologically Meaningful Core Size The outline methods discussed can provide nominal core areas at any percentage inclusion of 84 Hodder et al. VoL. 32, No. 2 locations. The second problem is to select a core that has biological significance. In the literature, the selection of core areas of most intense use has been largely subjective (Kenward 1985b, Harris et al. 1990, Wauters and Dhont 1992), or even arbi- trary (Mohr and Stumpf 1966, Anderson 1982, Wray et al. 1992b, Hohmann 1994). For instance, the core range has been defined as a 50% contour (Heikkila et al. 1996) or a 95% ICP core (Hulbert et al. 1996) but there is no biological basis for this. The subjective approach commonly uses the utili- zation distribution (UD) which is the polygon or contour area plotted against the percentage inclu- sion of locations (Van Winkle 1975). Identification of a discontinuity in this plot indicates the point where oudying fixes are excluded (Kenward 1985b, 1987, Harris et al. 1990) . ICP is particularly suitable for this method because it produces a stepped UD, unlike contour methods for which the plot tends to be smooth (Kenward 1987). An objective method for estimating a core range has been proposed by Samuel et al. (1985). How- ever, to achieve good conformity to the locations, their method depends on a large sample size of locations. In this paper, we analyze ICP ranges for buzzards to compare the subjective method (using the UD) with a new objective method for exclud- ing outlying (excursive) locations. Selection of Percentage Locations to be Included IN Core Ranges Subjective Exclusion of Exclusive Locations. When locations recorded for an animal include relatively long distance excursions (Fig. la), the outermost discontinuity on the slope of a utiliza- tion distribution can be very clearly defined (Fig. lb). In such cases, it is easy to visually select the point on the UD curve that indicates the size of the Excursion-Excluded Core (EEC). However, other individuals may have less well-defined cores (Fig. Ic and Id). In this case, it is difficult to decide whether there is a core at 95% or 100% inclusion of locations (Fig. Id). EEC ranges were subjectively estimated from the UD curve for the ICP winter ranges of each of 122 buzzards we studied. Objective Exclusion of Excursive Locations. The separation of excursions from core ranges is based on the assumption that behavior differs between excursive and core activity. For example, buzzards and Northern Goshawks {Acdpiter gentilis) tend to fly for much of the time during excursions but make shorter, less frequent flights during more typ- ical foraging (Kenward 1977, Walls and Kenward 1995). Therefore, if sufficient locations were re- corded, it is likely that the nearest neighbor (NN) distances between locations would form two fre- quency distributions representing core and excur- sive activity. In these hypothetical distributions, the mean distance between NN locations in the core would be expected to be smaller than the mean of NN distances where at least one of the locations is excursive. In most animal location data sets excur- sions are relatively rare. Therefore, the frequency distribution of NN distances would be expected to be negatively skewed with the positive tail indicat- ing excursive activity. The frequency distribution of NN distances was examined in a subsample of 10 buzzard ranges. For each of the ranges, the NN distance was calculated for all the locations {N = 30). To compensate for differences in the frequency distributions of NN distances between ranges, the distances were stan- d ardized by divi ding by h; for range; where h; = V (^{a^xi + and Uy; are the standard deviations of the location coordinates for range; in the X and y directions (Worton 1989) . As expected, the pooled frequency distribution of the standard- ized distances for all 10 ranges was highly negative- ly skewed, with many short distances and fewer lon- ger distances (Fig. 2). A transformation was sought that would high- light the excursive locations as outliers. There were only a small number of NN distances in each range; therefore, it was not possible to use tests of normality to seek an optimal transformation. In- stead, we sought the transformation that mini- mized the coefficient of variation (SD/^c) for the frequency distribution of NN distances. The effect of a number of logarithmic, reciprocal, and expo- nential transformations was examined on each of the 10 ranges. The transformation which most fre- quently minimized the coefficient of variation was generally the negative exponential transformation to q, where q == EXP (—0.5 (NN/h;)^). This is the Gaussian kernel function of Worton (1989). This transformation was therefore applied to all 122 buzzard ranges which were used to estimate cores with the subjective (UD) method. In order to sep- arate the excursive locations in each range, mini- mum excursion distances (MED) were estimated. These were the lower P = 0.05 (MED 0.05) and P = 0.01 (MED 0.01) percentiles of the normal dis- tribution fitted to the transformed NN distances. Clusters were then formed incrementally until the June 1998 Core Range Estimates for Common Buzzards 85 a. c. b. d. 95% Percent inclusion of fixes Percent inclusion of fixes Figure 1. (a) Range areas and fixes of a juvenile male Common Buzzard (Buteo (JM906) . Hatched line indi- cates the 100% Minimum Convex Polygon, solid line shows the 90% Incremental Cluster Polygons, (b) Utilization distribution for incremental cluster analysis (ICP) of the winter range of male buzzard (JM906). There is a clearly defined discontinuity in the curve at 90% inclusion of fixes, (c) Range areas and fixes of a juvenile male Common Buzzard (JM939). Hatched line indicates the 100% Minimum Convex Polygon, solid line shows the 90% Incremental Cluster Polygons, (d) Utilization distribution for incremental cluster analysis (ICP) of the winter range of juvenile male buzzard (JM939) . There is no clearly defined discontinuity in the curve. NN distance of the next fix to be added would have exceeded the MED. Any locations beyond this level were treated as excursive. The area of the convex polygon around the fixes in each cluster was summed to estimate the core range area. Core percentages and areas obtained with the two probability levels of the MED method were compared with those estimated by the subjective (UD) method. Results obtained by the UD and the MED methods were similar when MED = 0.05 and often larger when using MED = 0.01. For instance, buzzard JM939 had a 96% core at MED = 0.05, similar to the tentative core assigned with the UD method (Fig. Id) but had a 100% core at MED = 0.01. Buzzard JM906 had a core at 90% inclusion of locations with the UD method (Fig. lb) and the MED method (0.05 and 0.01). Core areas of the 122 buzzard ranges estimated by the UD method correlated well with core areas estimated using MED = 0.05 (r = 0.82, N = 122), and MED = 0.01 (r = 0.89, N = 122). However, the distribution of the points around the 1:1 line showed that with 86 Hodder et al. VoL. 32, No. 2 Distance of each fix from its nearest neighbor (NN) / hi (standardizing factor) Figure 2. Frequency distribution of nearest neighbor (NN) interfix distances pooled from locations (N = 296) from 10 winter ranges of Common Buzzards. NN dis- tances for each range are divided by a smoothing factor (h) . Note that the y-scale is truncated. a. Core areas estimated using utilization distribution curves (ha) MED = 0.05 all except one range were close to the line, whereas with MED = 0.01 the majority of core areas were larger than those estimated subjectively (Fig. 3a and 3b) . Therefore, MED = 0.05 was con- sidered appropriate for this sample of animals. This comparison showed that the MED method can give results which correlate well with the UD method, with the advantage that it is automated and does not involve subjective assessment for each animal. Setting a Standard Core Range Size for a Sample of Animals If a core is required with a standard percentage of locations for a sample of animals, it is advisable to select a percentage at which cores have been estimated for most of the ranges. We recommend setting a standard where 95% of the ranges have been cored, because this permits one range in 20 to have many excursive locations without a dispro- portionate reduction in the sample core size. The resulting standard core percentage will only be larger than the core ranges of 5% of the sample. Therefore, few standard core ranges will include excursive locations. This is important because in- clusion of excursive locations greatly increases the core range area, and thus the variance of areas in the sample. A larger proportion of the cores may be larger than the standard percentage. However, this has a small effect on sample variance, because b. Core areas estimated using utilization distribution curves (ha) Figure 3. Comparison of core areas of 122 winter rang- es of Common Buzzards obtained using subjective (UD) and objective (MED) methods, (a) with MED P = 0.05 and (b) with MED P = 0.01. The dotted line is the 1:1 line. the removal of peripheral locations from these cores reduces the area much less than the removal of excursive locations. This process was applied to the sample of 122 buzzard ranges. The number of ranges that had been cored increased as a function of the percent- June 1998 Core Range Estimates for Common Buzzards 87 a. Percentage of fixes included in the Excursion Excluded Core (EEC) b. T I I I T 100 95 90 85 80 Percentage of fixes included in the Excursion deluded Core (EEC) Figure 4. The core range size (percentage inclusion of locations) at which excursive activity is excluded by (a) subjective appraisal of the Incremental Cluster Polygon utilization distribution for buzzard home ranges (N = 122) and (b) by the Minimum Excursion Distance (0.05) for buzzard home ranges (A^ = 122). age of core locations, for the objective (MED = 0.05) and subjective (UD) coring methods (Fig. 4a and 4b). In each case, cores had been estimated for 95% of the buzzard ranges when 15% of the locations were removed (i.e., the excursion exclud- ed core contained 85% of the locations). With both methods, all the ranges reached cores that included at least 80% of the locations. Discussion In analyses of radio-tracking data, the choice of method for estimating home range size and struc- ture depends on the goal of the research (Voigt and Tinline 1980, Kenward and Walls 1994). Incre- mental Cluster Polygon analysis is particularly use- ful for identifying frequently used areas, as well as producing range structure statistics. Subjective (UD) and objective (MED) methods can be used to select the core ICP by excluding excursive lo- cations. However, the subjective choice of core from the UD may be difficult, particularly for rang- es that do not have a clearly defined core, and this has the potential to introduce error or even bias. For example, interpretation of the utilization dis- tribution could be influenced by prior knowledge of a typical percentage inclusion for cores in the sample of animals. In contrast the MED method provides a means to objectively plot a boundary that delimits the usual area of the study animal. The transformation applied to the frequency dis- tribution of NN distances and the choice of prob- ability level of the MED need to be tested with oth- er data sets. In the future, improvement in core range delineation might be achieved by plotting restricted edge polygons (Stickel 1954, Harvey and Barbour 1965, Voigt and Tinline 1980, Wolton 1985) rather than convex polygons around the clusters. Further work is also desirable to test the efficacy of this approach for data other than the standard 30-location range, especially data includ- ing variable numbers of locations or collected dur- ing the breeding season. We suggest that excursion-excluded core ranges will provide important insights in behavioral ecol- ogy, especially in studies of sociality and habitat se- lection. In analyses of habitat use, a core range es- timator allows data to be viewed at three spatial levels: overall availability, familiar area, and the usual area sensu Burt (1943). The study area (usu- ally arbitrarily defined) can be used for overall availability. A range outline including all the loca- tions such as the Minimum Convex Polygon (MCP) (Mohr 1947) or a probabilistic contour (e.g., Worton 1989) can delineate the familiar area of the animal. Finally, a biologically meaningful core can reveal the usual area. The importance of including internal range structure in studies of habitat use has been demonstrated in analyses us- ing ICP cores. For example, in a study of Tawny Owls {Strix aluco) in woodland patches, Redpath 88 Hodder et al. VoL. 32, No. 2 (1995) showed that the owls had much larger MCP ranges where woodland was fragmented in com- parison to owls in continuous woodland. However, IGF cores (multinuclear polygons), were of a sim- ilar size in the different classes of woodland. The estimation of a core range may also be ex- tremely important in studies of behavioral inter- actions. Least overlap with conspecifics has been used to define core range boundaries (Auffenberg 1978, Christian et al. 1986). However, a core range that can show if individuals regularly use the same areas, is more ecologically informative. For in- stance, when resources are very concentrated, in- dividuals may have overlapping home ranges, such as Northern Goshawks hunting near pheasant re- lease pens (Kenward and Walls 1994). Conspecifics that have overlapping outer ranges may show avoidance in their core (Kenward 1985b, Samuel et al. 1985, Harris et al. 1990). Intraspecific differ- ences in space use may also be masked by outer home range boundaries but revealed by the cores (Harris et al. 1990). ICP cores selected by the MED for the buzzard give a biologically useful estimate of the core range. We anticipate that the method presented here will prove useful for raptors and for other taxa. Acknowledgments We are grateful to the landowners on our study area for allowing us access. We also thank Amber Budden, David Hall, Adam Kelly, Maarit Pahkala, and Mark Wil- liams for their help in locating and climbing to nests. Biotrack provided a landrover for radio-tracking, Nigel Webb kindly commented on a draft of the manuscript and the suggestions made by three referees were very helpful. Literature Cited Ables, E.D, 1969, Home range studies of red foxes Vulpes vulpes. J. Mammal, 50:108—120. Anderson, D.J. 1982. The home range: a new nonpara- metric estimation technique. Ecology 63:103-112. Auffenberg, W. 1978. Social and feeding behavior in Var- anus komodoensis. Pages 77—491 in N. Greenberg and P.D. MacLean [Eds.], Behavior and neurology of liz- ards. DHEW, New York, NY U.S.A. Burt, W.H. 1943. Territoriality and home range as ap- plied to mammals, y. Mammal. 24:346-352. Christian, K., W.P. Porter and C. Tracey. 1986. Core areas within the home ranges of land iguanas Cono- lophus palidus. J. Herpetol. 20:272-276. Dixon, K.R. and J.A. Chapman. 1980. Harmonic mean measure of animal activity measures. Ecology 61:1040- 1044. Doncaster, C.P. and D.W. Macdonald. 1991. Drifting territoriality in the red fox Vulpes vulpes. J. Anim. Ecol. 60:423-439. Harris, S,, W.J. Cresswell, P.G. Forde, W.J. Trewhelia, T. Woollard and S. Wray. 1990. Home-range analysis using radio-tracking data — a review of problems and techniques particularly as applied to the study of mammals. Mammal Rev. 20:97-123. Harvey, M.J. and R.W. Barbour. 1965. Home range of Microtus ochrogaster as determined by a modified min- imum area method./. Mammal. 46:398-402. Hayne, D.W. 1949. Calculation of size of home range./ Mammal. 30:1-18. Heikkila, R., K. Nygren, S. Harkonen and A. Mykkanen. 1996. Characteristics of habitats used by female moose in the managed forest area. Acta Theriologica 41:321- 326. Hohmann, U. 1994. Status specific habitat use in the Common Buzzard Buteo buteo. Pages 359-366 in B.-U. Meyburg and R.D. Chancellor [Eds.], Raptor conser- vation today. Proceedings of the IV world conference on birds of prey and owls, Berlin, Germany. Pica Press, East Sussex, U.K. Hulbert, I.A.R., G.R. Iason, D.A. Elston and P.A. Racey. 1996. Home-range sizes in a stratified upland land- scape of two lagomorphs with different feeding strat- egies./ Appl. Ecol. 33:1479-1488. Jennrich, R.I. and EB. Turner. 1969. Measurement of a noncircular home range, /. Theor. Biol. 22:227—237. Kaufman, J.H. 1962. Ecology and social behaviour of the coati Nasua nirica on Barro Colorado island, Panama. Univ. Calif. Publ. Zool. 60:95-222. Kenward, R.E. 1977. Predation on released pheasants Phasianus colchicus by goshawks Acdpiter gentilis in cen- tral Sweden. Swed. Game Res. 10:79-112. . 1982. Techniques for monitoring the behaviour of grey squirrels by radio. Pages 175—196 in C.L. Cheeseman and R.B. Mitson [Eds.], Telemetric stud- ies of vertebrates. Academic Pre.ss, London, U.K. . 1985a. Raptor radio-tracking and telemetry. Pages 409-420 in I. Newton and R.D. Chancellor [Eds.], Conservation studies of raptors. ICBP, Cam- bridge, U.K. . 1985b. Ranging behaviour and population dy- namics in grey squirrels. Pages 319-330 in R.M. Sibly and R.H. Smith [Eds,], Behavioural ecology: ecologi- cal consequences of adaptive behaviour. Blackwell Sci- entific, Oxford, U.K. . 1987. Wildlife radio-tagging equipment, field techniques and data analysis. Academic Press, Lon- don, U.K. . 1992. Quantity versus quality: programmed col- lection and analysis of radio-tracking data. Pages 231- 244 in LG. Priede and S.M. Swift [Eds.], Wildlife te- lemetry: remote monitoring and tracking of animals. Ellis Horwood, London, U.K. and K.H. Hodder. 1996. Ranges V: an analysis June 1998 Core Range Estimates for Common Buzzards 89 system for biological location data. Natural Environ- ment Research Council, U.K. and S.S. Walls. 1994. The systematic study of ra- dio-tagged raptors: 1. survival, home-range and habi- tat-use. Pages 303-316 in B.-U. Meyburg and R.D. Chancellor [Eds.], Raptor conservation today. Pro- ceedings of the IV world conference on birds of prey and owls, Berlin, Germany. Pica Press, East Sussex, U.K. Lance, A.N. and A. Watson. 1980. A comment on the use of radio-tracking in ecological research. Pages 355-359 in C.J. Amlaner and D.W. Macdonald [Eds.], A handbook on biotelemetry and radio-tracking. Per- gamon Press, Oxford, U.K. Mohr, C.O. 1947. Table of equivalent populations of North American small mammals. Am. Midi. Nat. 37: 223-249. AND W.A. Stumpf. 1966. Comparison of methods for calculating areas of animal activity. J. Wildl. Man- age. 30:293-304. Parish, T. and H. Kruuk. 1982. The uses of radio-track- ing combined with other techniques in studies of badger ecology in Scotland. Pages 291-299 in C.L. Cheeseman and R.B. Mitson [Eds.], Telemetric stud- ies of vertebrates. Academic Press, London, U.K. Redpath, S.M. 1995. Habitat fragmentation and the in- dividual: Tawny Owls Strix aluco in woodland patches. J. Anim. Ecol. 64:652—661. Samuel, M.D., D.J. Pierce and E.O. Garton. 1985. Iden- tifying areas of concentrated use within the home range./. Anim. Ecol. 54:711-719. AND R.E. Green. 1988. A revised test procedure for identifying cores within the home range. J. Anim. Ecol. 57:1067-1068. SiNiFF, D.B. AND J.R. Tester. 1965. Aspects of animal movement and home range data obtained by telem- etry. Trans. N. Am. Wildl. Nat. Res. Conf. 30:379-392. Spencer, W.D. and R.H. Barrett. 1984. An evaluation of the harmonic mean measure for determining carni- vore activity measures. Acta Zool. Fenn. 171:255-259. Stahlecker, D.W. AND T.G. Smith. 1993. A comparison of home range estimates for a Bald Eagle wintering in New Mexico./. Raptor Res. 27(1):42— 45. Stickel, L.F. 1954. A comparison of certain methods of measuring ranges of small mammals. /. Mammal. 35. 1-15. Van Winkle, W. 1975. Comparison of several probabilis- tic home range models./. Wildl. Manage. 39:118-123 Voigt, D.R. and R.R, Tinline. 1980. Strategies for ana- lysing radio-tracking data. Pages 387-404 in C.J. Am- laner and D.W. Macdonald [Eds.], A handbook on biotelemetry and radio-tracking, Pergamon Press, Ox- ford, U.K. Walls, S.S. and R.E. Kenward. 1994. The systematic study of radio-tagged raptors: II. sociality and dispers- al. Pages 317-324 in B.-U. Meyburg and R.D. Chan- cellor [Eds.], Raptor conservation today. Proceedings of the rV world conference on birds of prey and owls, Berlin, Germany. Pica Press, East Sussex, U.K. AND . 1995. Movements of radio-tagged Common Buzzards Buteo buteo in their first year. Ibis 137:177-182. Wauters, L. AND A.A. Dhont. 1992. Spacing behaviour of red squirrels, Sciurus vulgaris: variation between habitats and the sexes. Anim. Behav. 43:297-311. White, G.C. and R.A. Garrott. 1990. Analysis of wildlife radio tracking data. Academic Press, San Diego, CA U.S.A. Worton, B.J. 1989. Kernel methods for estimating the utilization distribution in home range studies. Ecology 70:164-168. WOLTON, RJ. 1985. The ranging and nesting behaviour of wood mice Apodemus sylvatica (Rodentia: Muridae), as revealed by radio-tracking. /. ZooZ. 206:203-224. Wray, S., WJ. Cresswell and D. Rogers. 1992a. Dirichlet tesselations: a new nonparametric approach to home range analysis. Pages 247-255 in LG. Priede and S.M. Swift [Eds.], Wildlife telemetry: remote monitoring and tracking of animals. Ellis Horwood, London, U.K. WJ. Cresswell, P.C.L. White and S. Harris. 1992b. What, if anything is a core area? An analysis of the problems of describing internal range configura- tions. Pages 256-271 in LG. Priede and S.M. Swift [Eds.], Wildlife telemetry: remote monitoring and tracking of animals. Ellis Horwood, London, U.K. Received 1 July 1997; accepted 12 February 1998 J Raptor Res. 32(2):90-97 © 1998 The Raptor Research Foundation, Inc. HOME RANGE SIZE AND HABITAT REQUIREMENTS OF PEREGRINE FALCONS ON THE CAPE PENINSULA, SOUTH AFRICA Andrew R. Jenkins and Grant A. Benn^ Percy FitzPatrick Institute of African Ornithology, University of Cape Town, Rondebosch 7700, South Africa Abstract. — ^Two adult male and two adult female Peregrine Falcons (Falco peregrinus) were radiotracked in adjacent territories on the Cape Peninsula, South Africa between 1989-94. The falcons were tracked from two fixed stations over periods of 2-3 wk. The objective of the study was to determine the spatial and habitat requirements of peregrines and their ranging behavior. Males occupied larger home ranges than females, and females were more sedentary, spending over 50% of their time at the nest cliff. The home ranges of neighboring birds overlapped by about 20%, but neighbors tended not to forage in the same area on the same day. Areas with cliffs and ridges were preferred to slopes and flats, but other habitat and land-use categories were used randomly. Key Words: peregrine falcon; Falco peregrinus; Cape Peninsula; home range; habitat use. Tamano del rango de hogar y requerimientos de habitat de Halcon Peregrino en la Peninsula del Cabo, Sur^rica. Resumen. — Dos machos adultos y dos hembras adultas de Halcon Peregrino {Falco peregrinus) fueron estudiados mediante telemetria en territories de la Peninsula del Cabo, Sur^rica, entre 1989-94, El seguimiento fue hecho desde dos estaciones fijas por periodos de dos a tres semanas. El objetivo de este estudio era el determinar los requerimientos espaciales y de habitat de los Peregrines y sus habitos de forrajeo. Los machos ocuparon ranges de hogar mas grandes que las hembras, las hembras fueron mas sedentarias y emplearon el 50% del tiempo en el risco de anidacion. Los ranges de hogar de los halcones vecinos se traslaparon en un 20%, pero tendian a no forrajear en las mismas areas ni en los mismos dias. Las areas con riscos y acantilados fueron preferidas a las montanas y planicies, otras categorias de habitat y uses del suelo fueron utilizados al azar. [Traduccion de Cesar Marquez] Radiotracking of raptors can provide useful esti- mates of the extent and speed of their movements (Kenward 1987). We analyzed radio telemetry data to investigate the ranging behavior of Peregrine Falcons {Falco peregrinus) on the Cape Peninsula, South Africa. Few studies have directly measured the extent and distribution of peregrine home ranges, and the spatial and habitat requirements of African peregrines {F. p. minor) are poorly known. The objectives of tbis study were to esti- mate the size of peregrine home ranges and the degree of overlap between neighboring territories, to estimate the frequency, speed and distance of falcon movements and to assess habitat use in re- lation to its availability. Because of practical con- straints, we obtained only a moderate number of ^ Present address: KWA-Zulu National Conservation Ser- vice, P.O. Box 662, Pietermaritzburg 3200, South Africa. reliable locations for a small sample of birds, and therefore adopted a conservative approach to the analysis and interpretation of our data. Methods The study area was located in a central, discontinuous range of mountains and rocky ridges extending south- ward from Table Mountain on the Cape Peninsula (Fig. 1 ) . The area has a mosaic of low-growing, heath-like fyn- bos vegetation types, interspersed with patches of forest and woodland. The suburbs of Cape Town extend along the east and west sides of Table Mountain and across the Cape Flats to the southeast. The east coast of the pen- insula is built up along its northern half and the west coast features scattered settlements extending south to Scarborough. A band of suburban development connects Fish Hoek on the east coast with Noordhoek on the west. Altitude ranges from sea level to about 1100 m. Annual rainfall varies locally from about 40-200 cm, and falls mostly during winter (May-September) (Cowling et al. 1996). Temperature ranges from an average winter min- imum of about 9°C to an average summer maximum of about 25°C. 90 June 1998 Peregrine Falcon Home Ranges 91 Figure 1. The Cape Peninsula south of Table Mountain, showing the home ranges (100% MCPs) of the four radio- tagged Peregrine Falcons in relation to the peninsula mountain chain and the main urban centers. Four territorial adults were radio tracked during two study periods. Female 1 (FI) occupied a cliff on the southwest coast of the peninsula (Fig. 1) and was trapped and instrumented on 6 September 1989. This bird was tracked for 7 d between 21 September and 13 October for a total of 55.5 h. The earliest tracking began was at 0755 H and the latest it ended was at 1822 H. Female 2 (F2), male 1 (Ml) and male 2 (M2) were from three separate but adjacent territories on the central eastern side of the peninsula. They were instrumented between 21 July-15 August 1994, and were tracked simultaneously on 9 days between 1-26 August for a total of 66.8 h. The earliest tracking of these birds began at 0700 H and the latest ended at 1830 H. Ml occupied a territory adjacent to FI in 1994, but this territory was not occupied when FI was tracked in 1989. We attached the Biotrack transmitters (frequency range 150-151 MHZ) with both main and ground plane antennae using tail mounts (Kenward 1978), except that the attachment threads were sewn through the vane of a central rectrix on FI only, and were tied around the vanes and sealed with epoxy on the other three birds. This was a precaution against feather damage and re- duced the time and amount of manipulation required to fit each transmitter. Because they were larger, FI and F2 were instrumented with heavier 12 g (1.3% body mass) and 7.5 g (0.9% body mass) transmitters, respectively, to obtain more power. These transmitters had an expected life of five and two months, respectively. Males were each instrumented with lighter transmitters (5 g, <1% of body mass) with an expected life of 10-14 d. All of the trans- mitters had nominal line-of-sight reception ranges of at least 20 km. Tracking was carried out from two fixed stations, po- sitioned at high points overlooking the territories. FI was tracked from stations 4 km apart, and 2 km and 3.5 km 92 Jenkins and Benn VoL. 32 , No. 2 from the nest cliff (Fig. 1). F2 and the two males were tracked from stations 8 km apart, and 0.2 km and 11 km from each nest cliff. Locations of falcons were obtained using Yaesu FT-290R II transceivers and paired, in phase, five-element Yagi antennae, connected through a null- peak switch box to improve location accuracy (Kenward 1987). Directional fixes on each transmitter were taken synchronously from each station at prearranged intervals of 10-30 min. Bearings were recorded for each fix using a Recta DP 2 hand-held compass. Data from the two sta- tions were combined later, and locations for each pair of fixes were plotted manually by triangulation on 1:50 000 topographic maps. Fix bearing error (Lee et al. 1985) was measured at three of the four stations used by tracking a transmitter carried by car along a route through the study area not known by the observers at each station. Periodically, the car was stopped, the driver recorded the exact position and time, and held the transmitter as high as possible above the vehicle, turning it around to modulate the sig- nal, simulating a moving bird. Observers at the tracking stations took fixes on the transmitter signal and recorded a time and compass bearing for each fix. Using this tech- nique, bias and precision (Lee et al. 1985) at the three stations were calculated as 2.7° ± 2.2° (N = 9 fixes), 4.1° ± 3.7° (N= 10 fixes) and 2.4° ± 2.1° (N = 8 fixes), with an overall bias of 3.1° for the study, which resulted in a location error of about 500 m over a tracking distance of 10 km. This figure may be higher than the actual bias incurred because the test transmitters were closer to the ground, and therefore more susceptible to signal bounce and interference than transmitters carried on promi- nently perched or flying birds. We used the home range analysis program CALHOME (U.S. Forest Service, California Dept, of Fish and Game 1994) to determine the size and distribution of falcon home ranges. We converted our latitude-longitude loca- tions to a metric format compatible with CALHOME using Its associated utility program, LATLONG. Our location sample sizes were limited, so we used the nonpredictive minimum convex polygon method (Mohr 1947) to esti- mate range sizes. We produced 100% minimum convex polygon (MCP) home ranges for each bird for the entire study period and daily ranges for each tracking day with >6 hr of accumulated tracking time. Daily ranges were generated primarily to examine the frequency with which neighboring birds foraged in the same areas on the same day. The combined data for each bird were analyzed us- ing the adaptive kernel method (Worton 1989) to illus- trate patterns of home range use. Locational data were imported into the GIS ARC/INFO ver. 6.1.1 (Environ- mental Systems Research Institute, Redlands, California) to calculate the area of daily and home range overlap between the tracked birds. Using ARC/INEO, we isolated the locations which placed each of the falcons at their respective nest cliffs. Time elapsed over each series of consecutive nest cliff locations was summed to derive an estimate of cliff atten- dance for each bird, expressed as a percentage of the time over which the bird was tracked. Each series of con- secutive locations away from the nest cliff was considered to be an excursion or ranging flight. ARC/INFO facili- tated the calculation of distances between successive lo- cations allowing us to estimate the minimum distance covered by each falcon per hour of tracking time on each excursion away from its nest cliff and during the entire tracking period. Comparable average and maximum speed indices were calculated simply by dividing interlo- cation distance by interlocation interval for all cases where the interlocation distance exceeded zero and the interval between successive locations was 10-15 min. We overlayed our telemetry data on previously com- piled GIS coverages of the distribution of land use and vegetation (see Trinder-Smith et al. 1996 for descriptions of these databases) on the Cape Peninsula. These cov- erages were simplified into fewer, broader categories for our analysis by combining like habitats (e.g., five types of proteoid fynbos were considered as one habitat catego- ry). Using ARC/INFO, we calculated habitat availability (A) for each bird as the relative proportion of each hab- itat type present in its home range. In order to account for the level of bias affecting the accuracy of telemetry data, each location was buffered with a circle (diameter 500 m) and the area of each habitat within these buffers was calculated. Habitat use ( U) was expressed as the total area of each habitat type contained within the total area of the buffered points. Using information compiled by Forestek (CSIR 1995), a 500 X 500 m grid covering the Cape Peninsula was generated, containing information on the location of steep ridges and cliffs. A was measured in terms of the presence/absence of cliffs in the grid cells comprising each home range. U was determined as the proportion of each location which fell in cells with or without cliffs. Ivlev’s (1961) electivity index E, which provides a sim- ple measure of “preference” or “avoidance,” was calcu- lated for each bird for each habitat type where E = U — A/U + A. Nonparametric Kruskal-Wallis (land use and vegetation) or Wilcoxon paired signed-rank (cliffs) tests (e.g., Kenward and Walls 1994) were used to determine whether patterns of habitat use differed significantly from random. Peregrines generally spend fairly long periods of the day perched at their nest cliffs (Jenkins 1987, 1995). In order to examine habitat use by ranging or foraging birds more closely, the above analyses were repeated with the nest cliff locations for each individual excluded. Home ranges were also overlaid on a simplified coverage of land tenure and nature reserves on the Cape Peninsula (Cowl- ing et al. 1996, Trinder-Smith et al. 1996), to determine the extent to which their spatial requirements are met by the existing or possible future reserve networks. Results None of the transmitters malfunctioned; how- ever, three of the four peregrines molted their rec- trices prematurely shedding their transmitters. An additional male and female that were instrument- ed lost their transmitters within lid and were not tracked. FI lost its transmitter after 40 d but before the completion of scheduled tracking. Ml and M2 molted their tail feathers after 31 and 88 d respec- tively, and F2 was recovered injured with its trans- mitter still attached and operating 56 d after it was June 1998 Peregrine Falcon Home Ranges 93 Table 1. Radiotelemetry data obtained for four Pere- grine Falcons on the Cape Peninsula. Bird Davs Tracked Hours (-1-1^ davs) Tracked Number of Locations Average Location Interval (min) Female 1 5 (+2) 54.4 307 10.6 Female 2 5 (+4) 63.1 161 23.5 Male 1 4 (+5) 62.3 153 24.4 Male 2 6 ( + 1) 57.7 115 30.1 Overall 20 (-H12) 237.5 736 22.2 instrumented. The two untracked birds which shed their transmitters quickly and Ml did not grow new central rectrices during molt in the subsequent sea- son suggesting that the transmitters may have caused damage to the feather follicles. On average, only 59% of the locations recorded at each of the tracking stations were acceptable. In 9% of cases, signal distortion or observer error produced divergent fix pairs which could not be triangulated. Thirty-two percent of locations were unpaired because of signal interruption at one of the two tracking stations. We were able, however, to compile about 20 d (240 h) of telemetry data describing the move- ments of the four peregrines we tracked (Table 1 ) . Range size and habitat use of each individual were assessed in terms of the distribution of at least 100 locations, with an average interlocation interval of about 22 min. The two females occupied smaller home ranges than the two males, although the average daily ranges of the four birds were similar in size (Table 2) . Home range areas calculated using the adaptive kernel method showed the relatively sedentary na- ture of the peregrines, with up to 95% of the lo- cations sufficiently densely packed to account for only about 30% of the total home range predicted for each falcon. Each of the home ranges of the three peregrines tracked in 1994 overlapped sub- stantially with at least one other falcon (Fig. 1). F2 shared an area of 15.2 km^ with Ml and 52.6 km^ with M2. The home ranges of the two males over- lapped by 6.1 km^. There was less overlap in the daily ranges, with F2 overlapping on 1 d with Ml by 2.7 km^, and on 3 d with M2 by <1-2.3 km^. The daily ranges of the two males did not overlap. The female peregrines spent over half of the tracking time at their respective nest cliffs (FI = 54%, F2 = 53%), whereas the two males were at their nest cliffs for only about 20% of the total tracking time (Ml = 15%, M2 = 29%). Nearly 80 excursions from the nest cliff were recorded (Table 3). These ranging flights were between <1 km and >80 km in length (Table 3). Males moved less fre- quently but covered longer distances (Fig. 2). Over the entire tracking period FI covered a minimum total distance of 268 km and ranged up to 9.2 km from its nest cliff, and F2 covered 228 km and flew up to 9.2 km from its nest cliff; Ml covered 350 km and ranged out to 11.1 km, while M2 covered 346 km and ranged out to 16.4 km. Males also flew farther per tracking hour than females, and achieved higher maximum speeds between consec- utive locations (Table 3). F2 and M2 occupied territories in the northern half of the peninsula (Fig. 1 ) . They frequently flew out over the Cape Flats so their home ranges in- cluded tracts of urban and suburban land, and fea- tured asteracious fynbos, the dominant natural veg- etation in these low relief, coastal areas (Table 4). Table 2. Estimates of home range size and average daily range size (100% MCPs) for four Peregrine Falcons on the Cape Peninsula based on telemetry data, with additional home range size estimates calculated using the adaptive kernel method for the densest percentage of the locations obtained for each bird. Daily ranges were generated from 34-67 locations for FI, 12-27 locations for F2, 19-25 locations for Ml and 9-24 locations for M2. Bird Home Range ( km2) Average Daily Range ( km2) Adaptive Kernel (km^) 50% 75% 95% 100% Female 1 89.7 25.9 0.2 4.0 52.6 172.2 Female 2 94.7 20.2 0.1 9.8 67.7 269.2 Male 1 115.2 22.8 4.5 21.1 84.7 159.2 Male 2 192.1 22.3 13.8 37.6 140.4 419.4 Average 123.0 22.8 4.7 18.1 86.3 255.0 94 Jenkins and Benn VoL. 32, No. 2 Table 3. Number of excursions or ranging flights away from nest cliffs, average excursion distance, average distance covered per hour of tracking time, and maximum speed between consecutive locations recorded for four Peregrine Falcons radiotracked on the Cape Peninsula. Bird Number of Excursions Average Distance ( km) Average km per Tracking Hour Maximum Speed ( km/h) Female 1 26 10.3 (0.2-43.5) 5.1 40.7 Female 2 18 12.7 (1.4-49.6) 4.1 44.5 Male 1 16 21.9 (0.6-83.5) 6.1 59.4 Male 2 16 21.7 (0.5-75.0) 6.7 63.0 Overall 76 16.7 (0.2-83.5) 5.5 51.9 FI and Ml occupied territories farther south on the peninsula, and ranged over largely unused land, covered mosdy by proteoid fynbos. Hence, the habitat composition of each of the home rang- es seemed to be largely dependent on the location of the territory. There were no consistent patterns in the habitat electivity indices of the four birds (Table 5) . This suggested that habitat use, defined in terms of land use (including nest cliff locations, Kruskal-Wallis H = 4.46, P = 0.35, excluding nest cliff locations, H = 4.16, P — 0.38) or vegetation (including nest cliff locations, H = 7.82, P = 0.35, excluding nest cliff locations, H = 5.78, P = 0.57) was random. However, when habitat use was de- fined by topography, peregrines on the Cape Pen- insula significantly favored cliffs and ridges over slopes and flats (Table 5, Wilcoxon signed-rank test, P = 0.045, for electivity indices both including and excluding nest cliff locations) . About 45% (range = 33-57%) of each of the four peregrine home ranges fell within state-owned land, manz^ed by the Western Cape Provincial Ad- ministration or by local municipalities. The re- maining 55% is privately owned. On average, about 11% of each of the four home ranges was within the existing reserve network of the peninsula (FI = 3%, F2 = 26%, Ml = 12%, M2 = 4%). Discussion The transmitters we used in this study were con- siderably lighter than the recommended upper limit of 3% of body mass (Kenward 1987), and each was attached carefully and with as little ma- nipulation of the tail feathers as possible. Despite this, five of six birds shed their transmitters pre- maturely. We suspect that the bulk or shape of the transmitters, rather than their mass, caused this high incidence of premature molting. In particu- lar, the ground plane antenna of each transmitter protruded above the otherwise streamlined profile 0> c .2 '■P (0 o o 0 ) .Q E 3 20 15 10 5 3-4 km 4-5 km 5-6 km 6-7 km 7-8 km 8-9 km 9-1 0 km > 1 0 km Distance from the nest cliff ■ Female 1 . Female 2 ■ Male 1 n Male 2 Figure 2. The number of locations recorded for a range of distance intervals away from the nest cliff for each of the four Peregrine Falcons radiotracked on the Cape Peninsula. June 1998 Peregrine Falcon Home Ranges 95 Table 4. Habitat composition of the home ranges of four Peregrine Falcons radiotracked on the Cape Peninsula. Female 1 Female 2 Male 1 Male 2 Habitat Category Area^ Area % Area % Area % Land use Waterbodies 8.0 9.0 1.9 2.0 20.8 18.0 63.0 32.8 Unused vegetated 75.6 84.2 24.0 25.3 81.7 70.9 56.9 29.6 Unused open 0.4 0.5 18.3 19.3 0.8 0.7 20.6 10.7 Suburban 3.2 3.6 26.8 28.2 6.3 5.5 28.2 14.7 Urban 2.5 2.7 23.9 25.2 5.7 5.0 23.3 12.1 Vegetation Restioid fynbos 0.6 0.7 0.0 0.0 2.7 2.3 0.0 0.0 Asteracious fynbos 4.6 5.1 63.9 67.4 7.1 6.2 121.2 63.1 Ericacious fynbos 1.0 1.1 0.0 0.0 0.1 0.1 0.0 0.0 Proteoid fynbos 72.0 80.3 29.1 30.7 83.0 72.0 7.6 4.0 Cliff communities 0.3 0.3 0.1 0.1 0.1 0.1 0.0 0.0 Forest and thicket 0.4 0.5 0.5 0.5 0.2 0.2 0.2 0.1 Wetlands 1.5 1.7 1.3 1.4 0.7 0.6 8.3 4.3 Open sea 9.3 10.4 0.0 0.0 21.4 18.6 54.7 28.5 ikm^. ^ % of home range area. Table 5. Ivlev’s electivity indices describing habitat utilization by four Peregrine Falcons on the Cape Peninsula, based on the distribution of radiotelemetry locations. Category Female 1 Female 2 Male 1 Male Land use excluding nest cliff locations Waterbodies -0.62 -0.54 0.18 -0.51 Unused vegetated 0.06 0.30 -0.20 0.04 Unused open -0.67 -0.02 0.36 -0.26 Suburban -0.47 -0.26 0.41 0.20 Urban -0.42 -0.24 0.42 0.26 Vegetation excluding nest cliff locations Restioid fynbos 0.12 na -0.48 na Asteracious fynbos -0.59 -0.24 0.56 -0.01 Ericacious fynbos -0.37 na 0.60 na Proteoid fynbos 0.07 0.31 -0.21 0.75 Cliff communities 0.68 0.50 0.00 na Forest and thicket -0.11 -1.00 -1.00 0.88 Wetlands -0.26 -0.40 0.33 0.00 Open sea -0.72 na 0.20 -0.71 Topography including nest cliff locations Cliffs and ridges 0.73 0.61 0.64 0.85 Slopes and flats -0.16 -0.16 -0.13 -0.15 Topography excluding nest cliff locations Cliffs and ridges 0.75 0.42 0.32 0.70 Slopes and flats -0.17 -0.07 -0.03 -0.05 96 Jenkins and Benn VoL. 32, No. 2 of the flying bird, and may have caused sufficient turbulence during high speed maneuvers to desta- bilize and ultimately dislodge the rectrix. Activities of each of the peregrines were cen- tered on nest cliffs, and only a small number of distant locations accounted for most of their home ranges. Sample sizes (tracking time, number of lo- cations) may not have been sufficient to adequate- ly represent these irregular movements causing us to underestimate the actual size of the home rang- es. Nevertheless, our home range estimates were similar to those recorded for peregrines in previ- ous studies (e.g., Mearns 1985, White and Nelson 1991) . The Cape Peninsula mountain chain (and its ad- jacent coastal flats) occupies an area of about 440 km^ and supports at least 12 pairs of Peregrine Fal- cons. If these resident pairs maintained exclusive home ranges, each range would comprise about 37 km^. Telemetry data suggest that peninsula pere- grines require over three times this area, and that the home ranges of adjacent pairs overlap by about 20%. A study of urban-breeding Merlins {E colum- harius) recorded a similar degree of overlap be- tween neighboring territories (Sohdi and Oliphant 1992) . It may be significant that the daily ranges of peninsula peregrines overlapped less than their aggregate home ranges, suggesting that falcons from different territories tended not to forage over the same area at the same time. Although only two birds of each sex were tracked, differences in the ranging behavior of males and females were evident. Tracking took place immediately before the onset of breeding and each of the instrumented peregrines and their mates had commenced nuptial displays, copula- tion, and courtship feeding. Hence, the males ranged more extensively and covered ground more rapidly in order to provide food for their females, whereas females were relatively sedentary (e.g., Newton 1979, Ratcliffe 1993). More comprehen- sive estimates of the foraging ranges of the females could have been obtained by tracking them much later in the breeding season or in the nonbreeding season (Enderson and Kirven 1983, Mearns 1985, Kimsey and Marzluff 1993). Our average maximum speeds (41-52 km/h) were similar to predicted (56 km/h) and observed (46 km/h) cruising speeds of African peregrines (Jenkins 1995), and probably represent occasions when peregrines traveled most directly between consecutive locations. The absence of significant land-use or vegetation preferences by these peregrines may indicate that the scale, accuracy or number of telemetry loca- tions obtained were inadequate to show prefer- ences. Certainly, foraging conditions for peregrines are likely to vary in the different habitat categories we defined and we expected differential habitat use as a result. However, the relative importance of habitats as sources of prey may not be reflected in the distribution of the locations. Peregrines gen- erally hunt from perched or aerial vantage points and make rapid strikes at flying prey (usually birds), often over long distances (Cade 1982). While hunts may be concluded over particular hab- itats where prey are numerous or vulnerable to predation, falcons do not necessarily make very fre- quent visits to these foraging areas, or spend long periods of time in them. Cliffs and ridges were used extensively, probably because they provided elevated sites for perch hunting, and updrafts which facilitate aerial hunting and cross-country flying (Jenkins 1995). The Cape Peninsula is bounded by the city of Cape Town and its extensive suburbs, but much of the mountain chain south of the urban center re- mains undeveloped and comprises a loosely-con- nected network of conservation areas. While the peninsula generally lacks significant populations of local endemic bird species (Hockey et al. 1989, Cowling et al. 1996, Picker and Samways 1996), it supports one of the densest peregrine populations recorded in sub-Saharan Africa (average nearest neighbor distance = 4.6 km, range 1. 3-7.8 km, un- publ. data). Nine of the peregrine nest sites we know of on the peninsula are in existing nature reserves and two of the remaining three sites are located within the Cape Peninsula Protected Nat- ural Environment (CPPNE) (Cowling et al. 1996), which is likely to be incorporated into any exten- sion of the area’s present reserve system (Trinder- Smith et al. 1996, van Niekerk 1996). However, the individuals tracked during this study ranged exten- sively beyond the CPPNE boundary, over noncon- served and privately-owned land that is unlikely to be conserved in the future (Trinder-Smith et al. 1996). Much of this land has already been devel- oped. Thus, while the nest sites and preferred hab- itats (cliffs and ridges) of peregrines on the pen- insula are likely to be preserved, the spatial and habitat requirements of most pairs are not met by existing or possible future conservation areas. Our analysis of habitat use suggests that the spread of June 1998 Peregrine Falcon Home Ranges 97 urban development into presently undeveloped ar- eas of peregrine home ranges may not be detri- mental, and peregrines may benefit from the abun- dance of prey associated with the metropolitan area fringing the mountains. However, their capac- ity to exploit this food base may depend on the nature of the urban environment. For example, peregrines frequently use tall trees or buildings as feeding or hunting perches when foraging in sub- urban areas. Such vantage points are not a feature of low-cost housing schemes, which are currently the dominant form of urbanization on the penin- sula. Hence, the value of certain areas to foraging falcons may decline with the future spread of de- velopment. Ackn o wledgments We would like to thank Dave Allan, Harold and Bridget Jenkins, Greg McBey, Tim Wagner, and particularly Zelda Bate for their invaluable assistance in the held. Nigel Ad- ams provided much needed technical advice. We thank Forestek and Richard Cowling (Institute for Plant Pro- tection) for the use of their existing GIS databases for the Cape Peninsula. James Enderson, Phil Hockey, Ge- rard Malan, and Clayton White provided useful com- ments on earlier drafts of the manuscript. This research was partly funded by the Foundation for Research De- velopment. Literature Cited Cade, T.J. 1982. The falcons of the world. Collins, Lon- don, U.K. Cowling, R.M., I.A.W. Macdonald and M.T. Simmons. 1996. The Cape Peninsula, South Africa: physiograph- ical, biological and historical background to an ex- traordinary hot-spot of biodiversity. Biodivers. Conserv. 5:527-550. Enderson, J.H. and M.N. Kirven. 1983. Flights of nesting Peregrine Falcons recorded by telemetry. Raptor Res. 17:33-37. Hockey, P.A.R., L.G. Underhill, M. Neatherway and P.G. Ryan. 1989. Atlas of the birds of the southwestern Cape. Cape Bird Club, Cape Town, South Africa. Ivlev, V.S. 1961. Experimental ecology of the feeding of fishes. Yale Univ. Press, New Haven, CT U.S.A. Jenkins, A.R. 1987. Notes on the behavior of a pair of Peregrine Falcons in the southwestern Cape. Ostrich 58:161-171. . 1995. Morphometries and flight performance of southern African Peregrine and Lanner Falcons. J. Avian Biol. 26:49-58. Kenward, R.E. 1978. Radio transmitters tail-mounted on hawks. Ornis Scand. 9:220—223. . 1987. Wildlife radio tagging: equipment, field techniques and data analysis. Academic Press, Lon- don, U.K. AND S.S. Walls. 1994. The systematic study of ra- dio-tagged raptors: I. Survival, home-range and habi- tat-use. Pages 303-315 in B.-U. Meyburg and R.D. Chancellor [Eds.], Raptor conservation today WWGBP, Berlin, Germany. Kimsey, B.A. and J.M. Marzluff. 1993. Differential space use by male and female Prairie Falcons (Falco mexican- us): consequences for sampling requirements to esti- mate home ranges./. Raptor Res. 27:75. Lee, J.E., G.C. White, R.A. Garrott, R.M. Bartmann and A.W. Alldredge. 1985. Accessing accuracy of a radio- telemetry system for estimating animal locations. /. Wildl. Manage. 49:658-663. Mearns, R. 1985. The hunting ranges of two female per- egrines towards the end of the breeding season. Rap- tor Res. 19:20-26. Mohr, C.O. 1947. Table of equivalent populations of North American small mammals. Am. Midi. Nat. 37: 223-249. Newton, I. 1979. Population ecology of raptors. T. & A.D Poyser, Berkhamsted, U.K. Picker, M.D. and M.J. Samwavn. 1996. Faunal diversity and endemicity of the Cape Peninsula, South Africa — a first assessment. Biodivers. Conserv. 5:591-606. Ratcliffe, D.A. 1993. The Peregrine Falcon. T. & A.D. Poyser, London, U.K. Sohdi, N.S. and L.W. Oliphant. 1992. Hunting ranges and habitat use and selection of urban-breeding Mer- lins. Condor 94:743-749. Trinder-Smith, T.H., A.T. Lombard and M.D. Picker. 1996. Reserve scenarios for the Cape Peninsula: high-, middle- and low-road options for conserving the re- maining biodiversity. Biodivers. Conserv. 5:649—669. VAN Niekerk, L. 1996. Table Mountain: the ultimate ur- ban open space. Conserva 11:16-18. White, C.M. and R.W. Nelson. 1991. Hunting range and strategies in a tundra breeding Peregrine and Gyrfal- con observed from a helicopter. J. Raptor Res. 25:49- 62. Worton, B.J. 1989. Kernel methods for estimating utili- zation distribution in home-range studies. Ecology 70 164-168. Received 10 April 1997; accepted 4 February 1998 J. Raptor Res. 32(2):98-103 © 1998 The Raptor Research Foundation, Inc. SEASONAL PATTERNS OF HABITAT USE BY SNAIL KITES IN FLORIDA Patricia L. Valentine-Darby^ and Robert E. Bennetts Department of Wildlife Ecology and Conservation, Florida Cooperative Fish & Wildlife Research Unit, University of Florida, P.O. Box 110450, Gainesville, FL 32611-0450 US. A. Wiley M. Kitchens US. Geological Survey /Biological Resources Division, Florida Cooperative Fish Wildlife Research Unit, University of Florida, RO. Box 110450, Gainesville, FL 32611-0450 US. A. Abstract. — ^The movements of 165 adult Snail Kites (Rostrhamus sociabilis) were monitored at biweekly intervals in central and southern Florida using radiotelemetry. Over the 3-yr study period (15 April 1992-15 April 1995), 3361 kite locations were obtained. Snail Kite habitats were classified as graminoid marsh, cypress prairie, northern lake, miscellaneous peripheral (e.g., agricultural retention ponds), or Lake Okeechobee. Kites showed seasonal patterns in habitat use. Use of cypress prairies and miscella- neous peripheral habitats showed strong seasonal fluctuations with these areas used primarily during the nonbreeding season (July-December) . Kite use of graminoid marshes and northern lakes fluctuated to a lesser extent and was highest during the breeding season (January-June). Use of Lake Okeechobee also fluctuated greatly but showed no obvious seasonal pattern. One potential reason for the high use of cypress prairies during the nonbreeding season was the kites’ ability to perch hunt in these habitats, which may offer an energetic advantage over aerial hunting. Birds were less likely to breed in cypress prairies, however, due probably to their greater likelihood of drying down during the breeding season. Snail Kites without transmitters were more difficult to detect in cypress prairie and peripheral habitats due to limited access and dense vegetation. This seasonal use of habitats with lower detectability for kites may have important implications for kite monitoring in Florida. Key Words: Snail Kite, Rostrhamus sociabilis; breeding season; Florida; habitat use, radiotelemetry; seasonal shifts. Patrones estacionales de uso de habitat de Rostrhamus sociabilis en Florida Resumen. — Los movimientos de 165 Rostrhamus sociabilis adultos fueron seguidos en intervalos de 2 semanas en el centro y sur de Florida mediante radiotelemetria. A traves de un estudio de tres anos (15 de Abril 1992—15 de Abril 1995), 3361 localidades fueron obtenidas. Los habitats utilizados por Rostrhamus sociabilis fueron clasificados como pantanos de gramineas, praderas de cipres, lagos del norte y areas perifericas (i.e., pozos de retencion de agua para agricultura) o el Lago de Okeechobee. Ros- trhamus sociabilis utilizo habitats en forma estacional. El uso de las praderas de cipres y habitats perifericos tuvo fluctuaciones estacionales siendo utilizados durante la epoca en que no se reproduce (Julio- Diciembre). El uso de pantanos de gramineas y lagos del norte fluctuo hasta cierto punto y fue mas alto durante la epoca de reproduccion (Enero-Junio). El uso del Lago Okeechobee fluctuo marcada- mente sin ningun patron estacional obvio. Una posible razon para la alta utilizacion de las praderas de cipres durante la epoca en que la especie no se reproduce consiste en la utilizacion de perchas de caza en este habitat, lo cual ofrece una ventaja energetica sobre la caza aerea. Las aves no tendieron a reproducirse en las praderas de cipres quizas debido a que estas areas se secan durante la epoca de reproduccion. Los individuos sin radio fueron mas dificiles de detectar en las praderas de cipres y en los habitats perifericos debido al acceso limitado y la densa vegetacion. El uso estacional de habitats y la poca detectabilidad de la especie puede tener implicaciones importantes para su seguimiento en Florida. [Traduccion de Cesar Marquez] ^ Present address: St.Johns River Water Management District, P.O. Box 1429, Palatka, FL 32178-1429 U.S.A. 98 June 1998 Snail Kite Habitat Shifts 99 In the U.S., the Endangered Snail Kite {Rostrha- mus sociabilis) is found only in central and southern peninsular Florida where it inhabits freshwater wet- lands throughout the area that vary in their phys- ical characteristics and plant communities (Sykes et al. 1995). Almost exclusively, kites feed on the apple snail {Pomacea paludosa) by hunting for them over open-water patches in emergent marshes, shallow lakes, ponds, ephemeral wetlands, shallow banks of rivers, borrow pits, and canals (Sykes et al. 1995). The Snail Kite has been described as a nomadic species (Stieglitz and Thompson 1967, Sykes 1979, Bennetts 1993), but until recendy no major re- search effort has monitored individual kite move- ments throughout the year in Florida. Bennetts and Kitchens (1997) found that the birds routinely move among wedands within their range. Here, we present an analysis of habitats used by Snail Kites that were radio-tracked over a 3-yr period. Methods Movements of Snail Kites were monitored from 15 April 1992-15 April 1995 in central and southern Florida using radiotelemetry. We restricted our study to adult kites because juveniles do not choose their natal habitat and may remain in these areas for as long as 240 d (66%) of their first year of life (Bennetts and Kitchens 1997). We captured adult kites at their nests with a net gun (Mechlin and Shaiffer 1979). We attached 15-gm radio- transmitters to birds using backpack harnesses. Trans- mitters had a battery-life of 9-18 mo, so there was some overlap of monitored birds from year to year. We attempted to locate each instrumented kite once every 2 wk. For the 3-yr study period, the average time between consecutive kite locations was 13.5 d (SD = 7.9). Tracking was done primarily from a fixed-wing aircraft but also occasionally from an airboat or levee. Locations obtained from the ground were usually accompanied by a visual confirmation of the bird’s identity from uniquely- numbered anodized aluminum leg bands. Two 2.5-4 hr flights were made each week to cover the large study area. To track birds on a biweekly basis, we generally vis- ited specific wetland systems once every 2 wk. Once a bird’s radio signal was received on a flight, we circled the vicinity several times to determine its exact location. From an altitude of 330 m, with the bird below the plane, we recorded coordinates from the aircraft’s GPS and notes on the wetland habitat. If the kite was in an area near the border of two wetland systems (e.g.. Big Cypress National Preserve and western WGA-3A), we spent additional time isolating the signal to determine which wetland the bird was using. For the current anal- ysis, we only included Snail Kite locations for which we could identify the habitat type; the one exception to this was Lake Okeechobee, which is discussed below. Snail Kite locations were assigned to one of five broad Figure 1. Major wetland systems used by Snail Kites in central and southern Florida. Habitat types are grami- noid marsh (G), northern lake (N), Okeechobee (O), and cypress prairie (C). Peripheral habitats were scat- tered throughout the entire range and are not shown. habitat categories: (1) graminoid marsh, (2) cypress prai- rie, (3) northern lake, (4) miscellaneous peripheral, and (5) Lake Okeechobee. All or parts of these habitats are described in greater detail by Loveless (1959), Gunder- son and Loftus (1993), Gunderson (1994), and Bennetts and Kitchens (1997). Graminoid marshes were generally slough, sawgrass (Cladium jamaicense) marsh, or wet prai- rie communities (Loveless 1959, Gunderson and Loftus 1993, Gunderson 1994). The dominant emergent vege- tation was usually comprised of sawgrass, spike rush {Eleo- charis spp.), or maidencane (Panicum spp.) with scattered patches of woody vegetation. Most of these habitats were found in the Everglades, including Everglades National Park and a series of Water Conservation Areas (WCAs) , the Loxahatchee Slough, and the headwater marshes of the St. Johns River (Fig. 1 ) . A distinguishing feature of cypress prairies was a sparse overstory of cypress {Taxo- dium ascendens) with an understory of wet prairie (Duever et al. 1986, Gunderson and Loftus 1993). Cypress trees in our study area usually had a stunted growth form, but taller circular domes or linear strands interspersed with wet prairies also were common. This habitat occurred primarily in the western region of WCA-3A, Big Cypress National Preserve, and portions of the Loxahatchee Slough. The northern lake habitat type primarily consist- 100 Valentine-Darby et al. VoL. 32, No. 2 ed of lakes within the Kissimmee Chain-of-Lakes, but also included a few lakes along the Lake Wales Ridge. In con- trast to Lake Okeechobee, this habitat type consisted of a narrow littoral zone, often <100 m, usually dominated by maidencane interspersed with patches of bulrush (Scir- pus spp.) or cattail {Typha spp.). The miscellaneous pe- ripheral habitat type was mostly comprised of agricultural areas and ephemeral wetlands scattered throughout the kite’s range. These included citrus grove retention ponds, agricultural ditches, and other, usually highly dis- turbed habitats, including larger nonagricultural canals. We assigned locations at Lake Okeechobee to their own habitat type. The littoral zone of the lake is an extensive system of diverse marsh habitats, and it consequently had elements of at least three of the other habitat types (i.e., graminoid marsh, northern lake, and peripheral). Be- cause of this local habitat diversity, it would have been extremely difficult to assign locations to a particular type without extensive ground verification. Furthermore, birds often used more than one of these habitat types within a given day due to their close physical proximity. We used log-linear models (SAS Institute 1988, PROC CATMOD) to explore the interaction of habitats, time (year, month, and seasonal effects), and sex on locations of instrumented adult Snail Kites. The model selection procedure described by Hosmer and Lemeshow (1989) for logistic regression was used. We first explored individ- ual first-order interactions (i.e., two-way interactions) us- ing a likelihood ratio test of saturated models (for those effects being evaluated) and the same model without the interaction being tested. At this stage of the analysis, we used a liberal rejection criteria of a = 0.25 because of the potential for some interaction effects to be masked. We then constructed a model including all effects meet- ing the above criteria. At this and all subsequent steps of the analysis, we used a rejection criteria of ot = 0.05. Al- though log-linear models are intended to assess interac- tions, we retained all main effects for any term with a significant interaction in order to account for marginal totals (Everitt 1992). We then used a combination of like- lihood ratio tests and Akaike’s Information Criteria (AIC) (Akaike 1973, Burnham and Anderson 1992) to determine the most parsimonious model based on all combinations of effects indicated from our preliminary exploration. We evaluated the influence of time using monthly, an- nual, and seasonal (i.e., breeding season/nonbreeding season) effects. Monthly effects were determined using calendar months. Annual differences, however, were compared based on a study year from 15 April to 14 April of consecutive years (Bennetts and Kitchens 1997). The reason for this was that adults were trapped during spring, and we usually had a sufficient sample for survival analyses (another aspect of our study) by mid-April. We used this same criterion here because it enabled analysis of three complete study years, rather than two complete (1993 and 1994) and two partial (1992 and 1995) cal- endar years. Because we suspected that some seasonal differences were attributable to breeding status, we also evaluated the effect of whether or not it was the primary breeding season, which we refer to as the breeding sea- son and define as January-June. Although Snail Kites are capable of breeding throughout the year in Florida, 90% of the breeding occurs during this 6-mo period (Bennetts and Kitchens 1997). We designate the remainder of the year (i.e., July-December) as the nonbreeding season, even though occasional nesting may occur. We conducted 343 hr of foraging observations in 1993 and 1994. Each snail capture was assigned to one of two Snail Kite foraging methods: aerial hunting (hunting by low flight over the marsh) , or perch hunting. We did not attempt to assess foraging energetics among habitat types because foraging is highly dependent upon temperature, season (e.g., fall vs. winter), location, and specific vege- tation (Cary 1985, Sykes 1987), and we had insufficient data to partition out these potentially confounding ef- fects. Results We instrumented 165 adult Snail Kites (83 fe- males and 82 males). Forty-five kites were captured in 1992, 60 in 1993, and 60 in 1994. From these kites, we obtained 3361 locations in which the hab- itat type was known. Our preliminary analysis indicated a strong hab- itat*month interaction (x^ = 526.39, df = 44, P < 0.001); however, AIC (AIC = 39613.60 and 26082.35 for models with month and season, re- spectively) indicated that a model using seasonal (breeding season/nonbreeding season), rather than monthly differences, was more parsimonious (i.e., a fully saturated model with habitat, year, and month effects had 180 parameters compared to 27 using an analogous model with season) . Our pre- liminary analysis also supported using a year effect regardless of whether monthly or seasonal within- year effects were used. We found evidence of a hab- itat*sex interaction (x^ = 160.68 , df = 4, P < 0.001 ) in our preliminary analysis. However, this term was dropped from all subsequent models af- ter we accounted for temporal variation; this was based on both likelihood ratio criteria (P > 0.05) and AIC (AIC was substantially higher for all mod- els including a sex term). Thus, our final, most parsimonious, model included annual and season- al effects on habitat use (Table I ) . This model did not include a two-way interaction of season*year, but it did include a three-way interaction for hab- itat*season*year. Snail Kites showed seasonal patterns in their use of habitats. The use of cypress prairies and periph- eral habitats, in particular, showed strong seasonal (breeding season/nonbreeding season) fluctua- tions (Fig. 2). Generally, the use of these areas peaked between September-November of each year, except for a peak in February of 1993 for peripheral habitats and a peak in January of 1993 June 1998 Snail Kite Habitat Shifts 101 Table 1. Source terms and their corresponding contri- bution to our final (most parsimonious) log-linear model of the interactions between habitat (HAB), season (SEAS), and study year (SYR) on locations of radio- tracked adult Snail Kites. Source df Prob > SEAS 1 48.64 <0.001 HAB 4 522.83 <0.001 SYR 2 81.56 <0.001 SEAS * HAB 4 292.10 <0.001 HAB * SYR 8 95.15 <0.001 SEAS * HAB * SYR 8 67.65 <0.001 Likelihood Ratio ( Goodness-of-Fi t) ^ 2 2.70 0.260 ® Based on a likelihood ratio test between the model with all terms listed above and a fully saturated model. The null hypoth- esis of a Likelihood Ratio goodness-of-fit test is that the model fits the data (i.e., a failure to reject indicates fit). for cypress prairies. These two habitat types were used most extensively in the nonbreeding season. During the times that cypress prairie and periph- eral habitats were used the most, graminoid marsh and northern lake habitats were used the least. Snail Kite use of Lake Okeechobee fluctuated greatly but showed no obvious seasonal pattern (Fig. 2). Use of the lake’s wedands was relatively high for the first half of the study, then dropped to very low in September 1993, and increased again to moderate use in early 1994. The period of low kite use in the fall of 1993 coincided with a period of low water levels at Lake Okeechobee (Bennetts 2 ind Kitchens 1997). We recorded 814 prey captures during our for- aging observations. Aerial hunting by Snail Kites accounted for 671 (82%) of the captures, while perch hunting accounted for the remaining 143 (18%) captures. The proportional use of the two foraging methods was highly dependent upon hab- itat type (x^ = 249.78, df = 3, P < 0.001), with perch hunting mostly used in cypress prairie hab- itats (Fig. 3). 0) S* c 0) o O) (0 o 0> Q. O) (0 o 50% of ground cover comprised of forbs, grass, rock, soil, and woody plants <2.5 cm DBH. Brush <25% CC, any woody plants 2.5-15.2 cm DBH. <30% CC, >50% of ground cover comprised of brush, conifer, and hardwood species 2.5-12.6 cm DBH. Pole and medium conifer >25% CC, >50% of overstory domi- nated by 15.2-61.0 cm DBH coni- fers. 5:30% CC, >50% of conifer basal area comprised of trees ranging from 12.7-53.2 cm DBH. Mature and old-growth >25% CC, >50% of overstory domi- nated by conifers ^61.0 cm DBH. >30% CC, >50% of conifer basal area comprised of trees 5:53.3 cm DBH. Hardwood >25% CC, overstory dominated by hardwoods >15.2 cm DBH. >30% CC, >80% of basal area com- prised of hardwood trees >12.6 cm DBH. roost stands. We first delineated Spotted Owl nest or roost stands using aerial photographs (scale 1:12 000) as the sum of all contiguous mature/old-growth forest (in- cluding previously-logged forest if residual old trees re- mained) adjacent to roosts and nests. We then estab- lished plots at five random locations within every nest and roost stand to estimate used habitat characteristics. In order to adequately sample habitat variation within larger nest and roost stands, we randomly located one additional plot for each 20-ha increase in stand area. Out- side of nest and roost stands we established one plot at a random distance along each of eight 1500-m lines (ap- proximating the radius of an owl home range) extending outward from the edge of each nest or roost stand. All random plots were located in forests. These lines extend- ed from each stand to the north, northeast, east, south- east, south, southwest, west, and northwest. Data were combined for all used and available plots at each owl site such that each site was represented by a single average owl sample and a single average random sample. We measured 67 microhabitat variables at plots follow- ing procedures described by Solis (1983), LaHaye (1988), and Chavez-Leon (1989). At each plot, we used a 20 basal area factor Bitterlich angle to determine which trees would be included in each variable radius plot (Dil- worth 1981:259). For each tree in the plot we recorded species, diameter at breast height (DBH), and growth condition. We recorded the height of four trees in every one of the five DBH class intervals; 10.1-12.4, 12.5-27.4, 27.5-52.4, 52.5-89.9, and ^90 cm.). We estimated canopy closure at plot center by averaging four readings of a spherical densiometer. We measured ground cover along a 22.6 m line oriented north-south, and centered at the middle of each sample plot; the length of the line inter- cept represented the diameter of a 0.04 ha circular plot. We measured the intercept of trees ^10 cm in DBH, shrubs, herbs, litter, and coarse woody debris along this line. We reduced 67 potential variables for analysis by first eliminating those variables that potentially would not oc- cur on all plots. We then examined all possible correla- tions among variables and eliminated one member of a pair of highly correlated variables, retaining the one that was most biologically interpretable. We included some variables in the final a priori selection that have been used to characterize Spotted Owl habitat in other studies even though it may have been correlated with another habitat variable. We then conducted a two-group multivariate analysis of variance (MANOVA, Dixon et al. 1990) using the 10 microhabitat variables resulting from the above filtering process. We used Hotelling-Lawley Trace to test the sig- nificance of the MANOVA. Following a significant MAN- OVA, we used post-hoc f-tests to test for differences in in- dividual variables between owl and random plots (Stevens 1986:122-125). Landscape Structure. We mapped habitat within the ARA using interpretation of 1:12 000 color aerial photo- graphs from 1988 and 1992. We used the Wildlife Habitat Relationships system (Mayer and Laudenslayer 1988) to describe habitat types. We characterized habitat types by successional stage of coniferous forest or other broad vegetation type (Table 2). We estimated the proportion of each habitat type within the ARA by measuring habitat type polygons on 1:24000 topographic maps using a pla- nimeter after the boundaries of habitat types were trans- ferred from the air photos to topographic maps. We mapped habitat types within the WCSA using 1990 Landsat Thematic Mapper (TM) digital imagery. Hunter et al. (1995) provided detailed information on the meth- ods used to map and assess the accuracy of habitat clas- sifications within the WCSA. Successional stages were classified according to the criteria listed in Table 2. Given the spatial resolution (625-m^ grid-cells) of Landsat TM data, we were unable to map most areas of water. We estimated amounts of habitat types using the IDRISI geo- graphic information system (Eastman 1992). We used two different methods for spatial habitat as- sessments because we did not have access to Landsat im- agery for the ARA. In order to test that the two methods June 1998 Northern Spotted Owl Habitat 107 Table 3. Spotted Owl habitat characteristics within used and available habitats in the Areata Resource Area, north- western California. Variable Used Habitats {N= 14) Mean SE Available Habitats {N= 11) Mean SE t dP P-Value*’ Live tree basal area (m^/ha) 221.70 46.47 184.53 11.78 2.77 16 0.014 Conifer basal area (m^/ha) 102.80 14.87 79.28 17.74 1.02 21 0.321 Hardwood basal area (m^/ha) 118.57 13.06 105.25 11.32 0.77 23 0.449 Snag basal area (m^/ha) 12.60 1.98 10.58 2.34 0.66 21 0.518 Tree cover (%) (<10 cm DBH) 4.01 0.81 9.35 3.71 -1.40 11 0.188 Woody debris (%)*" 1.80 1.05 0.61 0.25 1.10 14 0.288 Canopy closure (%) 94.82 0.89 85.45 3.09 2.91 12 0.013 Mean DBH (cm) 66.72 5.05 51.01 3.72 2.50 22 0.020 Variance of DBH 1040.58 142.62 536.84 103.22 2.86 22 0.009 Mean height (m) 82.25 4.44 65.69 6.03 2.21 19 0.039 ® t-tests were for unequal (separate) variances. Thus, degrees of freedom were approximated and may vary among different variables (see Dixon et al. 1990). ’’ Results are from univariate post-hoc <-tests. A two group MANOVA (Zar 1984) also resulted in significant differences between habitat characteristics in stands used by owls and available stands {F ~ 5.28, df = 10,14, P = 0.002). Woody debris is coarse debris with the large end diameter greater than 30 cm. of habitat assessment did not influence the outcome of the between study area comparison, we randomly select- ed eight (25% of the sample) WCSA owl territories from among those we used in the between study comparisons. We then classified habitats on those eight territories us- ing air photos. Finally, we compared the Landsat classi- fication with the air photo classification. In no case did we find significant differences in the area of the habitat types estimated using the different methods. For exam- ple, for the key variable of mature/old-growth forest area we estimated an average of 107.72 ha (SE = 10.05) and 100.23 ha (SE = 11.72) within 800 m circles using air photos and Landsat TM data, respectively {t = 0.2354; P = 0.818). Since the habitat classes we compared were broad, the two methods were likely to produce similar classifications. Consequently, we felt it was justified to use the results of the different methods in the analysis. We measured all landscape variables within 800 m (200 ha) circular plots around owl locations on the ARA and WCSA. We chose this plot size because it represented one-half the mean nearest-neighbor distance (1579 m) between 1990 Spotted Owl territory centers at WCSA (Hunter et al. 1995), and, therefore, represented an eco- logically derived plot. This plot size also reduced overlap between adjacent plots. If a nest was located for a partic- ular territory, the corresponding plot was centered on that nest. If only a roost was located for a particular ter- ritory, the corresponding plot was centered on that roost. If more than one nest or roost was detected for a terri- tory, one location was randomly chosen from among those available. We selected a random subset of nest and roost locations from the available WCSA owl locations that was equal to the total number of nest and roost lo- cations found in the ARA. Within the ARA, we located 800-m circular plots at the geometric center of random areas that were not occupied by owls. We measured the proportion of each habitat type within each 800-m cir- cular plot and used these proportions to calculate Simp- son’s (1949) heterogeneity index, which was a measure of the heterogeneity of successional stage vegetation. We compared landscape characteristics around owl sites in the ARA and the WCSA with Mann-Whitney (MW) tests (Zar 1984). We also compared landscape characteristics at used and unused areas in the ARA with MW tests. Results We found 29 owl territories in 44 separate areas (66%) within the ARA and 50 territories within the WCSA. In the ARA, we sampled microhabitat in 14 Spotted Owl nest or roost stands; random plots were located around 1 1 of these stands to estimate available habitat characteristics. We did not estab- lish random plots at all occupied areas due to lack of access to adjacent private lands. Therefore, we only sampled microhabitat characteristics in the 14 stands because we could not achieve reasonably equal samples of used and available habitats. We used 10 of the microhabitat variables mea- sured for the two-group MANOVA (Table 3). Ho- telling-Lawley Trace indicated there was a signifi- cant difference between the characteristics of hab- itats used by owls and those available (Test Value = 3.77, F = 5.28, df = 10,14, P = 0.002). Habitats used by owls had higher values for all 10 habitat features except for small tree cover, indicating that owls used habitats characterized by greater struc- tural diversity. The higher mean value for small tree cover in available habitats probably reflected 108 Gutierrez et al. VoL. 32, No. 2 Table 4. Landscape characteristics within 800 m radius (200 ha) plots around Spotted Owl sites at the Areata Resource Area (ARA) and the Willow Creek Study Area (WCSA) , northwestern California. ARA WCSA {N = 29) {N = 29) Variable Mean SD Mean SD P-Value Habitat type (%) Herbaceous 6.1 6.2 4.2 3.0 0.40 0.692 Brush 24.4 19.2 9.7 8.6 3.06 0.002 Pole and medium conifer 7.1 13.0 14.1 6.3 4.08 <0.001 Mature and old-growth 31.8 20.3 49.6 14.0 3.17 0.002 Hardwood 30.5 20.4 22.4 8.2 1.73 0.083 Landscape index Habitat heterogeneity'’ 0.6 0.1 0.6 0.1 1.56 0.118 Mann-Whitney test statistic (Zar 1984). ’'Estimated using Simpson’s (1949) index. conifer regeneration and/or hardwood establish- ment following logging. Five vegetation features were significantly different between used and avail- able habitats (Table 3). Of the occupied areas on the ARA, landscape plots were centered on nine nest locations and 20 daytime roost locations. For comparison of land- scape characteristics, we centered landscape plots on nine nest sites and 20 roost sites randomly se- lected from the 50 territories on the WCSA. Owl sites at WCSA had less brush and hardwood, more pole/ medium-sized conifer, and more mature/ old- growth than did owl sites at ARA but amounts of herbaceous habitat types and habitat heterogeneity were not different between ARA and WCSA (Table 4). Sites used by Spotted Owls at ARA had less brush, more mature and old-growth, and lower habitat heterogeneity than unused sites (Table 5) . Amounts of herbaceous, pole/medium-sized coni- fer, hardwood, and habitat heterogeneity were not different between used and unused sites at ARA although hardwood comprised 10% more of the area at used sites. Discussion Previous studies at the WCSA (LaHaye 1988, So- lis and Gutierrez 1990, Hunter et al. 1995) and in coastal redwood forest (Folliard et al. 1993) showed that owls used habitats with greater amounts of mature and old-growth forest and more complex forest structure than available sites. Within both landscapes we studied, sites used by owls had more mature/ old-growth forest than Table 5. Landscape characteristics within 800 m radius plots (200 ha) within areas used and unused by Spotted Owls at the Areata Resource Area, northwestern California. LfsED Areas Unused Areas (A =29) (A= 15) Variable Mean SD Mean SD 2? P-VtLUE Habitat type (%) Herbaceous 6.1 6.2 6.3 8.8 0.68 0.496 Brush 24.4 3.7 40.0 17.6 2.71 0.007 Pole/medium conifer 7.1 13.0 10.8 12.6 1.45 0.148 Mature/old-growth 31.8 20.3 22.2 11.4 1.67 0.095 Hardwood 30.5 20.4 20.7 15.1 1.51 0.131 Landscape index Habitat heterogeneity 0.60 0.09 0.65 0.11 1.82 0.07 ^ Mann-Whitney test statistic (Zar 1984). June 1998 Northern Spotted Owl Habitat 109 available sites, suggesting that owls select sites with more older forest. While owls established territo- ries with comparatively low amounts of old forest on the more disturbed landscape, it was not clear if this influenced owl fitness. However, Chavez- Leon (1989) found that the number of young fledged per owl pair during a 2-yr (1987-88) pe- riod was significantly lower on the ARA (x = 0.47) than on the WCSA {x = 0.65; t = 5.61, df = 135, P < 0.001). We were unable to expand this com- parison because survey effort subsequent to 1988 was different between the two areas. Our results indicate that owls select habitats dif- ferentially within their territories at both micro- habitat and landscape scales. While the amount and condition of older forest stands was important, the presence of younger stands or brush stands, which also provide habitat for owl prey, could po- tentially offset the influence of reduced amounts of nesting and roosting habitat (Zabel et al. 1993, Hunter et al. 1995, Franklin 1997). However, the difference in the amount of brush between used and unused areas within the ARA suggests that at some point the amount of brush may have a neg- ative influence on site occupancy by Spotted Owls. Had our surveys at the ARA included areas we clas- sified as unsuitable for Spotted Owls due to the absence of mature and old-growth forest, the dif- ferences we observed in the amounts of brush be- tween used and unused areas would have been even more pronounced. Spotted Owl fragmenta- tion threshold tolerances also have been suggested in both field (Johnson 1993) and theoretical stud- ies (Lande 1987). Some redwood forests harvested for timber have high densities of Spotted Owls (Thomas et al. 1990). These areas are promoted as evidence of the adaptability of Spotted Owls to logging distur- bance not only in redwood forests but other forests as well (USDI 1994). Because redwood forests con- stitute <7% of the range of the Northern Spotted Owl (Thomas et al. 1990), have different climates and productivity, and a third of our survey sites in Douglas-fir forests did not support owls, land man- agers and policy makers should use caution in ap- plying results from owl studies in logging-disturbed redwood forests to the much larger Douglas-fir re- gion inhabited by Northern Spotted Owls. Ackn o wledgments We thank R. Bloom, M. Dossett, D. Copeland, A. Frank- lin, S. Hoover, E. Hopson, D. Leslie, C. Moen, A. Padilla, R. Quick, and J. Thrailkill for fieldwork. P. Carlson, S. Dark, Z. Peery, M. Seamans, and B. Twedt read drafts of this paper. S. Hawks was especially helpful for his long- term support of this project. Funding was provided by the Bureau of Land Management and the U.S. Forest Service. The Instituto Nacional de Investigaciones Fores- tales y Agropecuarias, Mexico provided a graduate stud- ies grant to GCL. Literature Cited 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. Chavez-Leon, G. 1989. Characteristics of fragmented habitat used by the Northern Spotted Owl {Strix oca- dentalis caurina) in northwestern California. M.S. the- sis, Humboldt State Univ., Areata, CA U.S. A. Dilworth, T. 1981. Log scaling and timber cruising. Or- egon State Univ. Press, Corvallis, OR U.S.A. Dixon, W.J., M.B. Brown, L. Engelman and R.l. Jenn- RICH. 1990. BMDP Statistical software manual. Vol. 2 Univ. Calif. Press, Berkeley, CA U.S.A. Eastman, J.R. 1992. IDRISI user’s guide, version 4 0 Clark Univ., Graduate School of Geography, Worces- ter, MA U.S.A. Easterbrook, G. 1994. The birds. The New Republic March:22-29. Folliard, L.B., L.V. Diller and K.P. Reese. 1993. Occur- rence and nesting habitat of Northern Spotted Owls in managed young-growth forests in northwestern California./. Raptor Res. 27:69. Forsman, E.D., E.C. Meslow and MJ. Strub. 1977. Spot- ted Owl abundance in young versus old-growth for- ests. Oreg. Wildl. Soc. Bull. 5:43-47. , E.C. Meslow and H.M. Wight. 1984. Distribu- tion and biology of the Spotted Owl in Oregon. Wildl. Monogr. 87:1-64. Franklin, A.B. 1997. Factors affecting temporal and spa- tial variation in Northern Spotted Owl populations in northwest California. Ph.D. dissertation. Colorado State Univ., Ft. Collins, CO U.S. A. , D.R. Anderson, E.D. Forsman, K.P. Burnham AND F.W. Wagner. 1996a. Methods for collecting and analyzing demographic data on the Northern Spotted Owl. Stud. Avian Biol. 17:12-20. , RJ. Gutierrez, B. Noon and J.P. Ward. 1996b. Demographic characteristics and trends of Northern Spotted Owl populations in northwestern California. Stud. Avian Biol. 17:83—91. Gutierrez, R.J., A.B. Franklin and W.S. LaHaye. 1995. Spotted Owl {Strix occidentalis) . Pages 1-27 m A. Poole and F. Gill [Eds.], The birds of North America, No. 179. Acad. Nat. Sci., Philadelphia, PA and Am. Orni- thol. Union, Washington, DC U.S.A. , E.D. Forsman, A.B. Franklin and E.C. Meslow. 1996. History of demographic studies in the manage- ment of the Northern Spotted Owl. Stud. Avian Biol 17:6-11. 110 Gutierrez et al. VoL. 32, No. 2 Hunter, J.E., R.J. Gutierrez and A.B. Franklin. 1995. Habitat configuration around Spotted Owl sites in northwestern California. Connor 97:684-693. Johnson, D.H. 1993. Spotted Owls, Great Horned Owls, and forest fragmentation in the central Oregon Cas- cades. M.S. thesis, Oregon State Univ., Corvallis, OR U.S.A. LaHaye, W.S. 1988. Nest-site selection and nesting habitat of the Northern Spotted Owl {Strix occidentalis cauri- na) in northwestern California. M.S. thesis, Humboldt State Univ., Areata, CA U.S.A. Lande, R. 1987. Extinction thresholds in demographic models of territorial populations. Am. Nat. 130:624— 635. Lehmkuhl, J.F. AND M.G. Raphael. 1993. Habitat pattern around Northern Spotted Owl locations on the Olym- pic Peninsula, Washington. J. Wildl. Manage. 57:302- 315. Mayer, K.E. and W.F. Laudenslayer. 1988. A guide to wildlife habitats of California. USDA Forest Service, Sacramento, CA U.S.A. Noon, B.R. and K.S. McKelvey. 1996. Management of the Spotted Owl: a case history of conservation biol- ogy. Annu. Rev. Ecol. Syst. 27:135-162. Ripple, D.H., D.H. Johnson, K.T. Hersheyand E.C. Mes- LOW. 1991. Old-growth and mature forests near Spot- ted Owl nests in western Oregon./. Wildl. Manage. 55: 316-318. Simpson, E.H. 1949. Measurement of diversity. Nature 163:688. Solis, D.M. 1983. Summer habitat ecology of Spotted Owls in northwestern California. M.S. thesis, Hum- boldt State Univ., Areata, CA U.S.A. Solis, D.M., Jr. and RJ. Gutierrez. 1990. Summer hab- itat ecology of Northern Spotted Owls in northwest- ern California. ConeJor 92:739-748. Stevens, J. 1986. Applied multivariate statistics for the social sciences. Lawrence Erlbaum Assoc., Hillsdale, NJ U.S.A. Thomas, J.W., E.D. Forsman, J.B. Lint, E.C. Meslow, B.R. Noon and J. Verner. 1990. A conservation strat- egy for the Northern Spotted Owl. U.S. Gov. Printing Off., Washington, DC U.S.A. USDl. 1994. Endangered and threatened wildlife and plants; 90-day finding on a petition to remove the Northern Spotted Owl in California from the list of threatened and endangered Species. Federal Re^ster 59:44958-44959. Wells, J.V. and M.E. Richmond. 1995. Populations, meta- populations, and species populations: what are they and who should care? Wildl. Soc. Bull. 23:458-462. Zabel, C.J., H.F. Sakai and J.R. Waters. 1993. Associa- tions between prey abundance, forest structure, and habitat use patterns of Spotted Owls in California. J. Raptor Res. 27:58-59. Zar, J.H. 1984. Biostatistical analysis. Prentice Hall, En- glewood Cliffs, NJ U.S.A. Received 21 February 1997; accepted 15 February 1998 J. Raptor Res. 32 (2) :1 11-1 15 © 1998 The Raptor Research Foundation, Inc. FOOD HABITS AND HUNTING RANGES OF SHORT-EARED OWLS {ASIO FLAMMEUS) IN AGRICULTURAL LANDSCAPES OE SOUTHERN CHILE David R. Martinez, Ricardo A. Figueroa and Carmen L. Ocampo Laboratorio de Ecologia, Departamento de Ciencias Bdsicas, Universidad de Los Lagos, Casilla 933, Osorno, Chile Fabian M. Jaksic Departamento de Lcologia, Pontificia Universidad Catolica de Chile, Casilla 1 1 4-D, Santiago, Chile Abstract. — ^The diet of the Short-eared Owl {Asia flamnieus) was quantihed by analyzing 400 pellets collected in two agricultural landscapes of southern Chile (Osorno and Chahuilco) . Diet composition fluctuated seasonally and included several species of small mammals, birds, and insects. Almost 80% of the annual biomass consumed was from two rodent species (Akodon olivaceus and Rattus norvegicus) and of a bird {Vanellus chilensis) . No differences in the composition of the winter diets were detected between the two study sites, and the latter were similar in landscape structure and use by humans. Also, the size of hunting ranges used by Short-eared Owls was similar between the two sites (ca. 250 ha), with a prevalence of landscape elements such as meadows, wetlands, and agricultural fields. Nevertheless, Short-eared Owls concentrated their hunting activity in areas with little human disturbance, such as vegetation fringes along roadsides, ungrazed meadows, and untilled lands. Even though Short-eared Owls perched on the ground, they also used posts along roads and between properties as perches. Although suboptimal or marginal habitat for most other raptors, these human-dominated landscapes appeared to be valuable for the survival and persistence of Short-eared Owl populations, as long as their food and shelter remained unaffected. Key Words: Short-eared Owl, Asio flammeus; prey selection', hunting range, Chile, temperate agroecosystems. Habitos alimentarios y ambitos de caza de nucos {Asio flammeus) en agroecosistemas del sur de Chile Resumen. — Se cuantihco la dieta del nuco {Asio flammeus) analizando 400 egagropilas recolectadas en dos agroecosistemas del sur de Chile (Osorno y Chahuilco). La composicion de la dieta fluctuo esta- cionalmente e incluyo varias especies de mamiferos, aves e insectos. Sin embargo, casi el 80% de la biomasa total consumida estuvo constitmda por solo dos especies de roedores {Akodon olivaceus y Rattus norvegicus) y de un ave {Vanellus chilensis). No existieron diferencias en la composicion de la dieta invernal entre los dos sitios de estudio, y ambos agroecosistemas fueron similares en cuanto a estructura y a uso por humanos. De igual modo, el tamano estimado del area de caza utilizada por los nucos fue similar entre los dos sitios (ca. 250 ha), predominando en ambos elementos del paisaje tales como praderas, humedales y cultivos. Sin embargo, los nucos concentraron su actividad de caza en areas con nula o escasa intervencion humana, tales como franjas vegetadas a orillas de camino, juncales, praderas no pastoreadas y terrenos baldfos. Si bien los nucos se posaban en el suelo, tambien utilizaban como perchas los postes de cercos a orillas de caminos y los existentes entre predios. Concluimos que estos ambientes antropicos, inadecuados para muchas rapaces, permitirian la subsistencia y residencia de parejas de nucos, siempre y cuando sus presas y sitios de crianza no fuesen afectados. [Traduccion Autores] The Short-eared Owl {Asio flammeus) occurs on all continents except Australia and Antarctica. De- spite being extensively distributed throughout South America (Clark 1975), literature available on the natural history of this species is chiefly anec- dotal (e.g., Housse 1945, Borrero 1962). In Chile, Short-eared Owls are regarded as poorly studied (Glade 1988). Populations are declining in most of the country but are increasing in southernmost Chile (Jaksic and Jimenez 1986). Rau et al. (1992) reported the diet of Short-eared Owls in mainland Chile based on 53 pellets and Fuentes et al. (1993) reported the diet of a population inhabiting Juan Fernandez archipelago based on 20 pellets. Thus, 111 112 Martinez et al. VoL. 32, No. 2 this species is one of the least known owls in Chile (see review in Jaksic 1997). Short-eared Owls in Chile occupy a variety of open habitats. Urbanization has had negative im- pacts on many Chilean raptors, which are decreas- ing due to illegal hunting, habitat alteration, and prey reduction (Jaksic and Jimenez 1986). Al- though the habitat and prey requirements of most raptors are generally not met in urban environ- ments (Martinez and Jaksic 1996, Petty 1996), the Short-eared Owl is an exception. At least in south- ern Chile, this owl opportunistically uses suburban environments such as airports, pasturelands, and golf courses. Here, we report the prey identified in 400 pellets of Short-eared Owls collected in two agricultural areas of southern Chile, and provide preliminary data on the size and landscape fea- tures of their hunting grounds. Study Area and Methods We collected 336 Short-eared Owl pellets on a seasonal basis from April (autumn) 1995-February (summer) 1996 in an area located on the outskirts of the city of Osorno (40°35'S, 73°05'W), in southern Chile. The study area included a golf course, a web of fallow vegetation strips between agricultural fields or along roads, an apple orchard, a sedge-rush {Carex-Juncus marsh, pasture- lands (some abandoned), an airport, and lawns sur- rounding the main campus of Universidad de Los Lagos. From June (winter) 1995-February 1996, we also collect- ed 64 Short-eared Owl pellets at Chahuilco (40°42'S, 73°09'W), an area that included pasturelands, a marsh, fallow vegetation along roads and agricultural fields, as well as berry farms. The climate of these two study areas IS within the oceanic region with mediterranean influ- ence of di Castri (1975), which is characterized by heavy rainfall (200-300 cm yearly), mostly during winter and decreasing in summer. Only pellets with identifiable prey remains were con- sidered. From the Osorno sample (336 pellets), we mea- sured and weighed 241 intact pellets. We identified and quantified most vertebrates in the pellets on the basis of skulls, beaks or dentary pairs (Reise 1973), which gave the highest count. For remains such as hair and feathers, we used reference collections and quantified these prey assuming the smallest possible number of individuals (e.g., hair or feathers of a given species were deemed as representing only one individual). For insect identifica- tion, we followed Pena (1986) and quantified these prey by counting head capsules and mandibles. We identified prey items to the finest possible taxonomic category in all cases, as recommended by Marti (1987). In Osorno, we evaluated the relative abundance of small mammals from May-July 1995 (autumn to winter) by live-trapping at an abandoned pasture located inside the owls’ hunting area. The mass of most prey species was determined by weighing individuals captured in Osorno. Some mass estimates for mammals were ob- tained from Pearson (1983) and Martinez (1993). Masses of birds were obtained from Morgado et al. (1987). We estimated the total biomass of each prey species in the diet by multiplying the number of individuals in the pel- lets by the mean body mass of each species. We assumed that masses of unidentified prey were similar to the mean mass of the most closely related identified taxon. We an- alyzed the Osorno diet on a seasonal and year-round ba- sis, and compared the results obtained for winter 1995 with those from the sample from Chahuilco with Pianka’s symmetrical niche overlap index (Oj^), using programs listed in Krebs (1989). Concomitantly with pellet collections, we estimated the hunting area used by a pair of Short-eared Owls in each of the two study areas. To the best of our knowledge, these were the only pairs present in each site. With a hand-held global positioning system (GPS) receiver (Gar- min GPS 38), we determined the location of each pre- viously known roosting place, as evidenced both by sight- ing and pellet collection at roosts. To determine the size of the area used by Short-eared Owls, the data were ex- pressed in UTM coordinates and analyzed using the min- imum convex polygon method (Jenrich and Turner 1969). For Osorno and Chahuilco, we pooled the spatial data available for autumn and winter 1995. The percent- age of landscape elements of each range was estimated from 1:30 000 aerial photographs, which were resized to 1:15 000 with a scanner. Ranges of Short-eared Owls, re- sized accordingly, were overlaid and the area of each cov- er category included was estimated using a Placom EP- SON digital planimeter. Statistical significance was set at P < 0.05 for all tests unless otherwise stated. Results and Discussion The 241 whole pellets we measured averaged 41.6 ± 0.072 mm X 21.4 ± 0.043 mm and had a mean dry mass of 2.8 ± 0.080 g (x ± SE). All three measurements were slightly lower than those re- ported by Holt et al. (1987) for Short-eared Owl pellets in North Aanerica. The 336 pellets from Osorno yielded 812 prey items (Table 1), of which small mammals were nu- merically the most frequent, followed by insects and birds. Olivaceous field-mice {Akodon olivaceus) were the most frequent prey in the diet year-round, although they were somewhat more frequent dur- ing winter. This is in close agreement with the au- tumn-winter peak of these vole-like mice in prai- rie-scrublands of southern Chile (Murua and Gon- zalez 1986). Long-tailed rice rats {Oryzomys longi- caudatus) were also eaten relatively frequently. Although this species is as abundant as the oliva- ceous field-mouse in southern Chile (Meserve et al. 1991), only its winter consumption by Short- eared Owls coincided with their autumn-winter peak abundance (Murua and Gonz^ez 1986). Per- haps, the skewed consumption of long-tailed rice rats reflects their high vagility, as well as their pro- June 1998 Ecology of Chilean Short-eared Owl 113 Table 1. Food habits of Short-eared Owls in agricultural landscapes around Osorno, Chile. B% is percent by biomass and N is prey by number. Mass ^ Autumn Winter Spring Summer Total Prey Species ( g ) B% (AO B% ( N ) B% (iV) B% ( N ) B% {N) Mammals Akodon olivaceus 23 24.3 (21) 23.8 (154) 10.1 (62) 26.3 (38) 18.5 (275) Auliscomys micropus 58 20.5 (7) 7.8 (20) 3.3 (8) 10.5 (6) 7.0 (41) Geoxus valdivianus 25 0.0 (0) 0.7 (4) 0.5 (3) 0.8 (1) 0.6 (8) Oryzomys longicaudatus 26 13.1 (10) 11.0 (63) 2.4 (13) 0.8 (1) 6.6 (87) Mus musculus 21 0.0 (0) 0.1 (1) 0.0 (0) 0.0 (0) 0.1 (1) Rattus norvegicus 201 10.1 (1) 39.2 (29) 5.7 (4) 24.2 (4) 22.3 (38) Unidentified rodents 59 17.7 (6) 7.5 (19) 3.4 (8) 1.7 (1) 5.8 (34) Subtotal mammals 85.7 (45) 90.1 (290) 25.4 (98) 64.3 (51) 60.9 (484) Birds Vanellus chilensis 270 13.6 (1) 9.1 (5) 73.2 (38) 32.5 (4) 37.9 (48) Sturnella loyca 78 0.0 (0) 0.5 (1) 0.0 (0) 2.4 (1) 0.4 (2) Unidentified passerines 20 0.0 (0) 0.1 (1) 0.1 (1) 0.0 (0) 0.1 (2) Subtotal birds 13.6 (1) 9.7 (7) 73.3 (39) 34.9 (5) 38.4 (52) Insects Carabidae 0.84 0.0 (0) 0.0 (0) 1.0 (172) 0.0 (0) 0.4 (172) Scarabaeidae 0.48 0.2 (3) <0.1 (31) <0.1 (5) 0.5 (31) 0.1 (70) Gryllacridiidae 0.69 0.2 (1) <0.1 (3) <0.1 (2) 0.0 (0) <0.1 (6) Unidentified insects 0.67 0.3 (6) <0.1 (5) <0.1 (2) 0.3 (15) <0.1 (28) Subtotal insects 0.7 (10) 0.2 (39) 1.3 (181) 0.8 (46) 0.7 (276) Total prey items (No.) Total biomass (g) 56 1980.2 336 14879.4 318 14 008.6 102 3318.9 812 34186.9 Total pellets (No.) 30 181 89 36 336 ® Masses of mammals were obtained from Pearson (1983) for Auliscomys, and Martinez (1993) and D.R. Martinez (unpubl. data) for the remaining taxa. Masses of birds were obtained from Morgado et al. (1987). Masses of insects were the mean of representatTve members of each family collected at the study site (D.R. Martinez unpubl. data). nounced population fluctuations (Murua et al. 1986). On a numerical basis, austral greater mice {Auliscomys micropus) were the third most frequent prey, but their biomass contribution was higher than that of long-tailed rice rats. Norway rats {Rat- tus norvegicus) were also eaten and, by biomass, they were the staple mammalian food of Short- eared Owls during winter. Other mammal species eaten were the Valdivian mole-mouse {Geoxus val- divianus) and house mouse {Mus musculus), but their number and biomass were minimal. In 594 trap-nights, we caught 95 small mammals (recaptures not included) . Their number and spe- cies composition were as follows: 78 (82.1%) oli- vaceous field-mice, six (6.3%) long-tailed rice rats, six (6.3%) house mice, and five (5.3%) Norway rats. All species, except for Norway rats, were re- ported by Rau et al. (1992) as prey of Short-eared Owls, in a similar frequency ranking. The only dif- ferences were the long-haired field-mice {Akodon longipilis) which was neither consumed nor trapped in Osorno, but preyed upon at Chahuilco, black rats {Rattus rattus), and Darwin’s leaf-eared mice {Phyllotis danoini) . These differences may have been attributable to the pooled sample used by Rau et al. (1992), which comprised pellets gath- ered at sites located as far apart as 100 km. In Co- lombia, although no quantitative data were provid- ed, Borrero (1962) reported that Short-eared Owls ate mostly Norway, black, and cotton rats {Sigmodon hispidus ) . By number, birds were unimportant as prey year- round, but by biomass and during spring and par- tially in summer. Southern Lapwings {Vanellus chi- lensis) were the main food item for Short-eared Owls in our study. Other birds taken were Red- breasted Meadowlarks (Sturndla loyca) and uniden- tified Passeriformes, but their biomass contribu- tion to the diet was minimal. Although we did not observe direct predation on Southern Lapwings, we found eight carcasses on the ground under perches used by owls. Also, on 6 September 1995, 114 Martinez et al. VoL. 32, No. 2 Table 2. Area (in ha) and percent cover of landscape features in two hunting ranges (Osorno and Chahuilco) used by Short-eared Owls in Chile. Also included, linear km of six-strand barbed-wire fences, number of posts, and posts suitable as perches for owls. Landscape Features Osorno Chahuilco Area (%) Area (%) Roads 5.3 (1.8) 1.6 (0.7) Roadsides 21.1 (7.4) 2.7 (1.3) Buildings 3.0 (1.0) 0.3 (0.1) Orchards 8.4 (3.0) 39.5 (18.0) Water bodies 8.2 (2.9) 3.6 (1.6) Meadows 25.7 (9.0) 55.3 (25.1) Pasturelands 214.3 (74.9) 117.0 (53.2) Total range 286.0 (100.0) 220.0 (100.0) Fences (km) 17.0 12.3 Number of posts 8500 4940 Number of suitable posts 212 123 we flushed a Short-eared Owl that was eating a freshly-killed fledgling lapwing (230 g) on a ground perch. Rau et al. (1992) did not report birds as prey of Short-eared Owls, but Fuentes et al. (1993) found that in Juan Fernandez archipelago, they preyed secondarily on birds, particularly on adults and eggs of petrels {Pterodroma spp.). In the northern hemisphere, Clark (1975) reported a generally low consumption of birds, chiefly of Western Mead- owlarks {Sturnella neglecta), but apparently there is higher bird predation among coastal and insular Short-eared Owls (Holt and Leasure 1993) . In Eu- rope, Glue (1977) reported that birds were the main source of food during winter for Short-eared Owls inhabiting Great Britain and Ireland. Insects outnumbered small mammals in the diet during spring, and birds during all seasons. Nev- ertheless, their biomass contribution was irrelevant on a yearly basis. Coleopterans, particularly Carab- idae and Scarabaeidae, were the most frequent items, followed by gryllids and unidentified insect remains. The 64 pellets from Chahuilco yielded 121 prey items: 34.8% olivaceous field-mouse, 25.6% Nor- way rat, 11.1% austral greater mouse, 8.9% long- tailed rice rat, 1.7% long-haired field-mouse, 17.8% unidentified rodents, and 0.1% Gryllacridi- idae. This showed that the diets of Short-eared Owls were similar between Osorno and Chahuilco during the winter of 1995 (symmetrical niche over- lap = 91.8%). In both, olivaceous field-mouse was the most frequent and Norway rat the highest bio- mass-contributor of mammalian prey (Table 1 ) . Osorno and Chahuilco are 17.5 km apart, but hunting range sizes and landscape features were similar for both populations of Short-eared Owls (Table 2). The similarity in landscape features was likely the result of human colonization of southern Chile in the late 19th century, which resulted in extensive burning of rainforests to clear the land for agricultural use (Martinez andjaksic 1996). In both areas, pasturelands and sedge-rush marshes were the predominant landscape features, followed by orchards (apple trees), berry farms, water bod- ies (rivers, streams, ponds) , fallow vegetation along roads and fences, paved or gravel roads, and some interspersed old southern beeches {Nothofagus dom- beyi) in addition to buildings. In our two study areas, most of the hunting ac- tivity of Short-eared Owls was performed near farmland and road borders, where agriculture was not intensive or nonexistent. Fence posts were the most prominent perches for these owls in such a flat landscape, although some saplings and old southern beech trees were used as roosting sites. Owing to widespread construction practices, the number of wooden split posts used per km of fence erected is almost the same everywhere in southern Chile (400-500 posts/km). However, only 2.5% of the posts in both study sites had flat tops and were suitable as perches. Most of the posts were sharp- ened on their tops. Because of this, suitable posts were easily seen in the field. Either an owl was June 1998 Ecology of Chilean Short-eared Owl 115 perched on them, or white wash, prey remains or scattered pellets were on the ground around them. Grazed, plowed or mowed areas, although inside their range, were not used for hunting or breeding by Short-eared Owls. Apparently, overgrazing and grass mowing removed cover needed by potential prey. Most of the time, the owls only crossed over these areas, flying straight and steady at about 4 m altitude, in the direction of their hunting grounds. Nesting areas apparently were located in marshy areas where sedge and rush were dense. Due to the high cover, it was difficult to find any nests, but in January (summer) 1997, we found four discarded eggshells of Short-eared Owls on a tiny island with dry ground and tall grasses. In both study areas, marshes were not used by humans, and hunting was prohibited by landowners. Although human- dominated habitats are suboptimal or marginal for many raptors, they seemed to be valuable for the survival and persistence of Short-eared Owls in our study. Acknowledgments DRM would like to thank S. Chaplin for sponsoring him as a member of RRF, and T.J. Cohn for a Wilson Bulletin subscription. This study was supported by grant FONDECYT 194-1256, from Chile’s Fondo Nacional de Investigacion Cientifica y Tecnologica. Denver Holt, Carl Marti, and an anonymous reviewer, made suggestions that greatly improved rendition of this report. Literature Cited Borrero, H.J.I. 1962. Notas varias sobre Asio flammeus bogotensis Colombia. Rev. Biol. Trop. 10:45-49. Clark, R.J. 1975. A field study of the Short-eared Owl, Asio flammeus, in North America. Wildl. Monogr. 47:1- 67. Di Castri, F. 1975. Esbozo ecologico de Chile. Ministerio de Educacion, Santiago, Chile. Fuentes, M.A., J.A. SiMONETTi, M.S. Sepulveda and RA. Acevedo. 1993. Diet of the Red-backed Buzzard {Bu- teo polyosoma exsul) and the Short-eared Owl {Asio flam- meus suinda) in the Juan Fernandez Archipelago off Chile./. Raptor Res. 27:167-169. Glade, A.A. [Ed.]. 1988. Red list of Chilean terrestrial vertebrates. Corporacion Nacional Forestal, Santiago, Chile. Glue, D.E. 1977. Feeding ecology of the Short-eared Owl in Britain and Ireland. Bird Study 24:70-78. Holt, D.W. and S.M. Leasure. 1993. Short-eared Owl (Asio flammeus) . Pages 1-24 in A. Poole and F. Gill [Eds.], The birds of North America, No. 62. Acad. Nat. Sci., Philadelphia, PA and Am. Ornithol. Union, Washington, DC U.S.A. , L.J. Lyon and R. Hale. 1987. Techniques for dif- ferentiating pellets of Short-eared Owls and Northern Harriers. Condor 89:929-931. Housse, R. 1945. Las aves de Chile en su clasificacion moderna: su vida y sus costumbres. Ediciones Uni- versidad de Chile, Santiago, Chile. Jaksic, F.M. 1997. Ecologia de los vertebrados de Chile Ediciones Universidad Catolica de Chile, Santiago, Chile. AND J.E. Jimenez. 1986. The conservation status of raptors in Chile. Birds Prey Bull. 3:95—104. Jennrich, R.I. AND F.B. Turner. 1969. Measurement of a noncircular home range. J. Theor. Biol. 22:221— 231 . Krebs, C.J. 1989. Ecological methodology. Harper and Collins Publ., New York, NYU.S.A. Marti, C.D. 1987. Raptor food habits studies. Pages 67- 80 in B.A. Giron Pendleton, B.A. Millsap, K.W. Kline and D.M. Bird [Eds.], Raptor management tech- niques manual. Nat. Wildl. Fed., Washington, DC U.S.A. Martinez, D.R. 1993. Food habits of the Rnfous-legged Owl (Strix rufipes) in temperate rainforests of south- ern Chile./. Raptor Res. 27:214—216. AND F.M. Jaksic. 1996. Habitat, relative abun- dance, and diet of Rufous-legged Owls (Strix rufipes King) in temperate forest remnants of southern Chile. Ecoscience 3:259-263. Meserve, P.L., B.K. Lang, R. Murua, A. Munoz-Pedreros AND L.A. Gonzalez. 1991. Characteristics of a terres- trial small mammal assemblage in a temperate rain- forest in Chile. Rev. Chil. Hist. Nat. 64:157-169. Morgado, E., B. Gunther and U. GonzAlez. 1987. On the allometry of wings. Rev. Chil. Hist. Nat. 60:71-79. Murua, R. and L.A. GonzAlez. 1986. Regulation of num- bers in two Neotropical rodent species in southern Chile. Rev. Chil. Hist. Nat. 59:193-200. , L.A. GonzAlez and P.L, Meserve. 1986. Popula- tion ecology of Oryzomys hngicaudatus philippii (Ro- dentia: Cricetidae) in southern Chile. / Anim. Ecol. 55:281-293. Pearson, O.P. 1983. Characteristics of a mammalian fau- na from forests in Patagonia, southern Argentina / Mammal. 64:476-492. Pena, L.E. 1986. Introduccion a los insectos de Chile. Editorial Universitaria, Santiago, Chile. Petty, S.J. 1996. Adaptations of raptors to man-made spruce forests in the uplands of Britain. Pages 201- 214 in D. Bird, D. Varland and J. Negro [Eds.], Rap- tors in human landscapes: adaptations to built and cultivated environments. Academic Press, London, U.K. Rau,J.R., M.C. Villagra, M.L. Mora, D.R. Martinez and M.S. Tilleria. 1992. Food habits of the Short-eared Owl (Asio flammeus) in southern South America / Raptor Res. 26:35-36. Reise, D. 1973. Clave para la determinacion de los cra- neos de marsupiales y roedores chilenos. Gayana: Zool. 27:1-20. Received 17 May 1997; accepted 8 February 1998 J Raptor Res. 32(2):116-119 © 1998 The Raptor Research Foundation, Inc. WINTER ROOST SITES OF NORTHERN HARRIERS AND SHORT-EARED OWLS ON ILLINOIS GRASSLANDS Jeffery W. Walk^ Department of Zoology, Eastern Illinois University, Charleston, IL 61920 U.S.A. Abstract. — Characteristics of the diets and roosting habitats used by Northern Harriers (Circus cyaneus) and Short-eared Owls (Asio flammeus) on grassland sanctuaries in Jasper and Marion counties, south- eastern Illinois, were described for December 1993 and January 1994. Northern Harrier communal roosts contained from 21-43 individuals and were only observed in undisturbed cool-season grasses. Short-eared Owls also roosted communally (3-17 individuals) but the two species were not observed to roost within the same fields. Short-eared Owls roosted in shorter cover than Northern Harriers, but there was no evidence of selection for any grass or management type by Short-eared Owls. Both species frequently fed on southern bog lemmings (Synaptomys cooperi) and avian remains were found in less than 5% of all pellets. Key Words; Northern Harrier, Circus cyaneus; Short-eared Owl, Asio flammeus; communal roosts-, roost hab- itat, diet. Perchas de invierno de Circus cyaneus y Asio flammeus en las praderas de Illinois Resumen. — Las caracteristicas de las dietas y de las perchas utilizadas por Circus cyaneus y Asio flammeus en los santuarios de praderas de los condados de Jasper y Marion en el sur de Illinois fueron descritas en Diciembre de 1993 y Enero de 1994. Las perchas comunales de Circus cyaneus estaban compuestas por 21-43 individuos y fueron detectadas en pastizales estacionales no perturbados. Asio flammeus utilizo perchas comunales (3—17 individuos), las dos especies no utilizaron perchas en el mismo campo. Asio flammeus utilizo perchas en sitios de cobertura menos altos que los utilizados por Circus cyaneus. No hubo evidencia que Asio flammeus tuviera alguna preferencia por algun tipo de pastizal o area de manejo. Las dos especies se alimentaron frecuentemente de en menos del 5% del total de egagropilas. Both the Northern Harrier (Circus cyaneus) 2 Ci\d Short-eared, Owl (Asio flammeus) are noted for their communal winter roosts in the Midwest (e.g., Wel- ler et al. 1955, Mumford and Danner 1974). Little has been reported, however, on the characteristics of habitats that are selected for roosting. Because both species are listed as Endangered in Illinois (Herkert 1992), information regarding the specific habitats these species use would help in their con- servation. Most studies of these two raptor species have re- ported that they capture similar prey, particularly small rodents (Colvin and Spaulding 1983, Clark and Ward 1974, Craighead and Craighead 1956). Therefore, they may overlap in diet and, as a result. ^ Present address; Department of Natural Resources and Environmental Sciences, University of Illinois at Urbana- Champaign, Urbana, Illinois 61801 U.S.A. Synaptomys cooperi. Restos de aves fueron encontrados [Traduccion de Cesar Marquez] competition for food may influence their habitat use. The purpose of this study was to document roost site characteristics of Northern Harriers and Short-eared Owls on a 800-ha grassland sanctuary system in southeastern Illinois. I also collected pel- lets to examine the diets and possible competition between these two species. Methods The study was conducted at the Prairie Ridge State Natural Area in Jasper and Marion counties, approxi- mately 800 ha of restored grassland tracts managed par- ticularly for Threatened and Endangered wildlife. The Jasper and Marion county units are approximately 60 km apart. A total of 13 tracts (seven in Jasper county and six in Marion county), ranging in size from 7-120 ha, were examined. Within each county, tracts averaged 0.8 km from their nearest-neighbor tract. The maximum dis- tance between any two tracts was about 7 km. Tracts were further subdivided into management units averaging about 3 ha (range 0.5-15 ha). The areas studied were about 75% introduced cool-season grasses such as redtop 116 June 1998 Winter Roosting of Grassland Raptors 117 bentgrass (Agrostis alba) , timothy {Phleum pratense) , smooth brome {Bromus inermis), and bluegrass (Poaspp.), and 25% native warm-season grasses such as switchgrass {Panicum virgatum), big bluestem {Andropogon gerardii), indian grass {Sorgastrum nutans), and little bluestem {Schi- zachyrium scoparium) . Management measures include mowing, haying, limited grazing, and undisturbed fields (Simpson and Esker 1997). Northern Harriers and Short-eared Owls were ob- served from 26 December 1993-12 January 1994 to de- termine habitats used for roosting. Observations were made from a stationary vehicle and were continuous from 1.5 hr before sunset until dark. Each tract was sur- veyed for one evening. Two small (<20 ha), adjacent tracts were surveyed simultaneously once and three large tracts were divided and observed from two locations when topography and size prevented a complete survey from a single location. A total of 13 tracts were surveyed from 15 points, each with an observation radius of <500 m. Every available tract was observed for the same amount of time regardless of the level of raptor activity in that tract. The locations of Northern Harriers landing in communal roosts and of Short-eared Owls coming off of communal roosts were recorded on detailed site maps. Population estimates of Short-eared Owls were obtained by flushing birds off of all known communal roost sites on 30 December (Jasper county) and 6 January (Marion county) at midday (1000-1400 H). The population of Northern Harriers was estimated by summing the num- ber of birds observed roosting on each tract. Visual ob- struction measurements (Robel et al. 1970) were made to characterize vegetation height and density at 10 North- ern Harrier and six Short-eared Owl communal roost sites during the day in January 1994. Vegetation type and management history were provided for each manage- ment unit within grassland tracts by the grassland man- agers, confirmed in the field, and recorded for the roost- ing site of every individual observed of both species. Pellets were collected from the roost sites of each spe- cies. Pellet identification was further confirmed based on characters described by Holt et al. (1987). Pellets were carefully separated, teeth and/or skulls were identified when possible according to Hoffmeister (1989), and non- mammalian remains were noted. Results The wintering population of Northern Harriers was estimated to be at legist 102 birds, or 12.3 har- riers/100 ha. The Short-eared Owl population was estimated to be at least 34 birds, or 4.1 owls/ 100 ha. Three Northern Harrier and five Short-eared Owl communal roosts were documented. The number of harriers at each communal roost ranged from 21-43 birds (x = 34 harriers). Indi- vidually roosting Northern Harriers were recorded five times. Only a single roost site occurred in Mar- ion county. Double-counting of Northern Harriers between days on the two Jasper county communal roost sites was possible, but did not alter popula- tion estimates by more than about 5%. A site man- ager observed one roost site (which I had previ- ously surveyed) on the same evening I surveyed the second Jasper county roost. Only 2 more har- riers were counted than I had previously counted at the roost. Short-eared Owl communal roosts contained from 3-17 individuals (x = 6.8 owls). One Short-eared Owl was observed roosting indi- vidually. Flush counts of Short-eared Owl commu- nal roosts were the same (four of five roosts) or larger (observation estimate = 15 owls and flush count = 17 owls) than counts derived from ob- serving owls leaving the roosts in the evenings. Northern Harrier communal roosts were only observed in undisturbed areas of cool-season grass- es. Due to the small samples of these roosts, the results were not significant (Table 1). Short-eared Owls were not found to select a particular cover type (Table 1). However, 32 of 34 total roost ob- servations were in shorter vegetation of mowed grasses (15 observations) and new grass seedings (17 observations). Average visual obstruction height (Robel et al. 1970) at harrier roosts (x = 25 cm) was significantly higher than at owl roosts (x = 12 cm; t — 2.64, df = 5, < 0.025, separate variance Kest). Short-eared Owls tended to roost at the base of a dense clump of grass within a field of much thinner cover, while Northern Harriers roosted in relatively even stands of grass. Rodents accounted for 100% of the mammalian remains found in 65 Northern Harrier and 52 Short-eared Owl pellets. The only rodent species identified was the southern bog lemming (Synap- tomys cooperi). Every pellet analyzed contained ro- dent remains. A few also contained unidentified passerine bird remains (4.6% of the harrier pellets and 3.8% of the owl pellets). Discussion Northern Harriers and Short-eared Owls have used the Prairie Ridge State Natural Area for at least 25 years as a wintering ground (R. Weste- meier, pers. comm.). Since 1990, winter raptor cen- suses have reported 100-150 harriers and 25—50 owls annually (S. Simpson pers. comm.). This con- sistent level of winter activity seems unique to the region, as Bildstein (1979) shows high variation in harrier activity in Ohio and Weller et al. (1955) remarked on the ephemeral occurrence of both species in Missouri. Habitat of harrier communal roosts has been de- scribed as often-damp wheat stubble fields over- 118 Walk VoL. 32, No. 2 Table 1. Relative availability of cover types and numbers of Northern Harrier and Short-eared Owl winter communal roosts located within each cover type in Illinois. % Of Avatt art e # Of Northern Harrier Communal Roost Sites # Of Short-eared Owl Communal Roost Sites Cover Type^ Grassland Observed Expected"* Observed Expected Undisturbed cool-season grass 42.5 3 roosts 1.3 roosts (99 harriers) 2 roosts 2.1 roosts (19 owls^) Mowed cool-season grass 33.7 0 1.0 roost (1 harrier) 2 roosts 1.7 roosts (10 owls) Undisturbed warm-season grass 15.4 0 0.5 roost 0 0.8 roost Mowed warm-season grass 8.4 0 0.3 roost (2 harriers) = 4.19 P = 0.24 1 roost 0.4 roost (5 owls) = 4.28 P= 0.23 ® Expected values for number of roost sites per cover type were calculated as the total number of observed communal roost sites for a species multiplied by the proportion of each available cover type. ^ 17 of 19 of these observations were from one young (<1.5 yr) Bromus inermis seeding. This field was atypical of undisturbed cool- season grasses, being shorter with grass in scattered clumps. grown with ragweed {Ambrosia artemisaefolia) to a height of 60-90 cm (Weller et al. 1955), stubble fields, prairie grasses, and fescue 60-110 cm tall with up to “several inches” of water (Mumford and Danner 1974), and weedy old fields (Colvin and Spaulding 1983). In this study, harriers roosted in fields with dense cover up to a height of about 45 cm and thinner screening cover to a height of 1 .0- 1.2 m. Fields of undisturbed redtop bentgrass and timothy and undisturbed smooth brome fit these characteristics. One communal roost was situated in a wet field, with the others in well-drained areas. Short-eared Owl communal roosts have been even less frequently described, but Weller et al. (1955) found the roosts in dense grass less than 30 cm high, frequently in a tuft of grass. My observa- tions closely agree, with communal roost sites often found in fields mowed to 30-40 cm. Seventeen of the 19 observations of Short-eared Owls roosting in undisturbed cool-season grasses came from a single field which had been sowed to smooth brome 15-16 mo previous to this study. That field was atypical of undisturbed cool-season grass fields, being quite short (under 50 cm versus the typical 100-120 cm) and the grass in isolated clumps as opposed to an even stand. This type of structure was more typical of mowed areas. Available fields of undisturbed warm-season grass may have been too tall or too dense to be suitable for roosting for either species. Bildstein (1979) showed that Northern Harrier communal roost sites were centrally located within a foraging area. I likewise observed birds leave and arrive at roost sites from all directions. Individual harriers were followed for over 7 km from foraging areas in the surrounding agriculture landscape to the roosting area, and birds were frequently seen foraging up to 16 km from the roost site. No other roosts were found or reported locally. Short-eared Owls were observed foraging on grasslands and ad- jacent agricultural lands, but the distance they trav- eled to foraging areas was not known. Other studies of the winter dietary habits of Northern Harriers and Short-eared Owls have shown that they have similar diets and utilize any abundant mid-sized prey available. Rodents are normally the primary prey (e.g., Weller et al. 1955, Craighead and Craighead 1956, and Colvin and Spaulding 1983, Holt and Leasure 1993). Only oc- casionally are birds taken in large numbers (e.g., Collopy and Bildstein 1987, Holt and Leasure 1993). My results were similar to these studies, with rodents being the primary winter food used by both Northern Harriers and Short-eared Owls. Long-term small mammal data are unavailable for this site, but current research indicates rodent pop- ulations are relatively high (E. Heske pers. comm.). This site is the last in Illinois to host the Endan- gered Greater Prairie-Chicken {Tympanuchus cupi- do). No evidence was found indicating that these raptors kill prairie-chickens despite the high raptor density. Northern Harriers do, however, harass prairie-chickens displaying on leks in both the June 1998 Winter Roosting of Grassland Raptors 119 spring and autumn (S. Simpson and R. Westemeier pers. comm.; J. Walk pers. obs.), although the largely-migrant harrier population is much re- duced by the time prairie-chicken breeding begins. Therefore, it appears that maintaining prairie- chickens and grassland raptors are compatible management objectives. These two open country raptors exhibit a great deal of sympatry and habitat and dietary overlap (Clark and Ward 1974), potentially contributing to interspecific competition. Schoener (1983) dem- onstrated that overlap of macrohabitat usually does not lead to competition, whereas overlap of micro- habitat and food usually does. This study suggests that Northern Harriers and Short-eared Owls uti- lize separate roosting habitat. The diets of these species were essentially identical. However, niche overlap does not necessarily translate into compe- tition (Holt 1987). Rodents may have provided a superabundant food source, foraging habitat may have differed, or foraging densities may have been so low as to make exploitation competition levels insignificant. Further isolating the species with re- spect to foraging behavior is timing. Harriers are apparently restricted to hunting during daylight hours. Short-eared Owls are not limited in this sense, but are primarily crepuscular (Clark and Ward 1974, Holt and Leasure 1993). Some level of interference competition exists between the spe- cies, which manifested itself during this study in the form of aerial sparring in the mornings and evenings as one species became active and the oth- er returned to communal roost sites. In part due to occasional kleptoparasitism by Northern Harri- ers, Clark (1975) concluded that the Short-eared Owl is a fugitive species, which accounts for occa- sional, temporary food and habitat overlap. Acknowledgments This paper is based on an undergraduate internship completed through the Environmental Biology program of Eastern Illinois University’s Zoology Department. I thank T.L. Esker and S.A. Simpson for field assistance and sharing site information. J.R. Herkert, K.C. Kruse, and R.E. Warner provided advise and review of manu- scripts and J.M. Fair, D.W. Holt, N. JenksJay, and an anonymous referee reviewed this manuscript. I gratefully acknowledge the Illinois Department of Natural Re- sources, Illinois Nature Preserves Commission, and The Nature Conservancy for access to properties and field equipment. Fieldwork was supported by funds from the Undergraduate Research Council, Eastern Illinois Uni- versity. Literature Cited Bildstein, K.L. 1979. Fluctuations in the number of Northern Harriers Circus cyaneus hudsonius at com- munal roosts in south central Ohio. Raptor Res. 13:40- 46. Clark, RJ. 1975. A field study of the Short-eared Owl {Asio flammeus) in North America. Wildl. Monogr. 47- 1-67. andJ.G. Ward. 1974. Interspecific competition in two species of open country raptors Circus cyaneus and Asio flammeus. Proc. PA. Acad. Sci. 48:79-87. COLLOPY, M.W. AND K.L. BiLDSTElN. 1987. Foraging be- havior of Northern Harriers wintering in southeast- ern salt and freshwater marshes. Auk 104:11-16. Colvin, B.A. and S.R. Spaulding. 1983. Winter foraging behavior of Short-eared Owls {Asio flammeus) in Ohio. Am. Midi. Nat. 110:124-128. Craighead, JJ- AND F.C. Craighead. 1956. Hawks, owls and wildlife. Stackpole Co., Harrisburg, PA U.S.A. Herkert, J.R. [Ed.]. 1992. Endangered and threatened species of Illinois: status and distribution. Vol. 2. Illi- nois Endangered Species Protection Board, Spring- field, IT U.S.A. Hoffmeister, D.F. 1989. Mammals of Illinois. Univ. Illi- nois, Urbana, IL U.S.A. Holt, D.W. and S.M. Leasure. 1993. Short-eared Owl {Asio flammeus). In the Birds of North America, No 62, A. Poole and F. Gill [Eds.]. Acad. Nat. Sci., Phil- adelphia, PA U.S.A. and Am. Ornithol. Union, Wash- ington, DC U.S.A. , L.J. Lyon and R. Hale. 1987. Techniques for dif- ferentiating pellets of Short-eared Owls and Northern Harriers. Co/irfor 89:929— 931. Holt, R.D. 1987. On the relationship between niche overlap and competition: the effect of incommensu- rable niche overlap. Oikos 48:110-114. Mumeord, R.E. AND C.R. Danner. 1974. An Indiana marsh hawk roost. Indiana Audubon Quarterly 52:96- 98. Robel, R.J., J.N. Briggs, A.D. Dayton and L.C. Hurl- BERT. 1970. Relationships between visual obstruction measurements and weight of grassland vegetation. J. Range Manage. 23:295-297. Schoener, T.W. 1983. Field experiments on interspecific competition. Am. Nat. 122:240-285. Simpson, S.A. and T.L. Esker. 1997. Prairie Ridge State Natural Area habitat plan. Illinois Department of Nat- ural Resources, Springfield, IL U.S.A Weller, M.W., I.C. Adams, Jr. and B.J. Rose. 1955. Win- ter roosts of marsh hawks and Short-eared Owls in central Missouri. Wilson Bull. 67:189-193. Received 8 March 1997; accepted 2 February 1998 J Raptor Res. 32 (2): 120-1 25 © 1998 The Raptor Research Foundation, Inc. SIZE VARIATION OF MIGRANT BALD EAGLES AT GLACIER NATIONAL PARK, MONTANA B. Riley McClelland, ^ David S. Shea,^ and Patricia T. McClelland^ School of Forestry, University of Montana, Missoula, MT 59812 U.S.A. David A. Patterson Department of Mathematics, University of Montana, Missoula, MT 59812 U.S.A. Abstract. — ^We measured morphological variables on 303 migrating Bald Eagles {Haliaeetus leuco- cephalus) at Glacier National Park, Montana (GNP), during 1977-88. Based on results of a concurrent migration study, most of the eagles we measured were from summering areas in the Northwest Terri- tories, Canada. Eagles were classified by plumage as juvenile, subadult, or adult. Feather lengths differed among plumage classes in all variables, but only the eighth and ninth primaries and tail differed (de- creased with age) among all three plumage classes in both sexes. We found no differences in beak depth among the three plumage classes of either sex. Masses did not differ among age classes of either sex of eagles with empty crops. The length of the exposed culmen did not differ between adult eagles at GNP and a southern Colorado wintering area. Culmen length, beak depth, and length of hallux did not differ between adult eagles at GNP and museum study skins from Canada, Alaska, and the northern U.S. Key Words: Haliaeetus leucocephalus; Bald Eagle, morphology, measurements-, migration. Variacion de tamano de Aguilas Calvas migratorias en el Parque Nacional Glacier, Montana Resumen. — Medimos las variables morfologicas de 303 Aguilas Calvas {Haliaetus leucocephalus) en el Parque Nacional Glacier, Montana, durante 1977-88. Con base en los resultados del estudio sobre su migracion encontramos que la mayoria de las aguilas medidas eran de territorios de reproduccion del noroeste de Canada. De acuerdo al plumaje, las aguilas fueron clasificadas como juveniles, subadultas o adultas. La longitud de las plumas y variables analizadas fue diferente entre clases, pero solo la octava y novena remige primaria y la cola disminuyeron su longitud con la edad en las tres clases de plum^ye y en los dos sexos. No se encontro diferencia en la profundidad del pico entre clases y sexos. La masa corporal no fue diferente entre edades y sexos sin contenido estomacal. La longitud del culmen, pro- fundidad del pico y longitud del halux no difirio entre %uilas adultas del Parque Nacional Glacier y pieles de museo de Canada, Alaska y el norte de Estados Unidos. [Traduccion de Cesar Marquez] Morphology of Bald Eagles {Haliaeetus leucocephalus) is known to vary with age and sex (Bortolotti 1984a). Females generally are larger than males from similar latitudes (Stalmaster 1987). Bald Eagles of northern natal origin are, on average, larger and heavier than those from the south (Palmer 1988). Measurements of Bald Eagles have been described from western and northern North America by Imler and Kalmbach (1955), Harmata (1984), Bortolotti (1984a), Gar- celon et al. (1985), and others. The purpose of ^ Present address: Box 366, West Glacier, MT 59936 U.S.A. Present address: Box 90, Babb, MT 59411 U.S.A. our paper is to report morphological and mass measurements of migrant Bald Eagles captured during autumns 1977-88 in Glacier National Park, Montana (GNP). Size measurements of Bald Eagles from the Mackenzie-lntermountain Flyway (McClelland et al. 1994) have not previ- ously been reported. Study Area and Methods Our study was conducted at GNP (approximately 48°30'N, 114°00'W) in northwestern Montana. Spawning kokanee salmon ( Oncorhynchus nerka) attracted migrating eagles to the site each fall during the project. During a concurrent migration study, 30 of 31 transmitter- equipped eagles summered in the northeastern half of Canada’s Mackenzie River Basin, primarily in the North- 120 June 1998 Bald Eagle Size 121 west Territories (latitudes of eagle locations ranged from 54°17'-65°35'N, averaging about 62°) (McClelland et al. 1994). This was the probable natal region of the eagles we measured. We used McCollough’s (1989) three broad categories to classify eagles; juvenile (juvenal plumage), subadult (basic plumages 1, 11, and 111), and adult (basic IV and definitive plumages). We measured the following vari- ables (methods in Bortolotti [1984a, 1984b] and Garce- lon et al. [1985]): wingspan (WSpan), unflattened wing chord (WnCh); eighth (EPr), ninth (NPr), and tenth (TPr) primary feathers; tail length measured on a central tail feather (Tail) ; length of exposed culmen without the cere (ClLn); beak depth at the leading edge of the cere (BDp); narrowest tarsal thickness frontal (NTTF) and lat- eral (NTTL); length of hallux claw (HalCl) (we report both left and right measurements because both were used in our sex identification model) ; outer claw (OCl) , middle claw (MCI), inner claw (ICl); and mass (Mass). Only undamaged and apparently fully emerged feathers were included in our analyses. We palpated the crop area on each eagle and qualitatively assessed each as empty (no food detected), full (crop distended and firm), or partially full (conditions not fitting the previous two cat- egories). We included only eagles with empty crops in analysis of mass. Sexes were identified with a formula (using ClLn and both HalCls) derived from Bortolotti’s (1984a) and Garcelon et al.’s (1985) models. This derivation was explained in detail in McClelland et al. (1994). We tested for differences among the three plumage classes within each gender using ANOVA, followed by Fisher’s Least Significant Difference method (a = 0.05) for in- dividual comparisons when the overall ANOVA was sig- nificant. Power (1 — P) of ANOVAs was estimated at a = 0.05, where observed differences were the alterna- tives to hypothesized differences of zero. The ANOVAs should be treated with some caution since the mea- surements were used to classify by sex before age class comparisons were made. With some variables, the pow- er to detect differences was low because of a relatively small sample size of adults. Stalmaster (1987) and others have summarized mor- phological measurements of Bald Eagles from other geographic areas. We restrict comparisons to migrant adult eagles that were primarily from northern natal areas (studies by Bortolotti [1984a] and Harmata [1984] ) . Means from our study are compared to means from those studies using two-sample unpooled t-tests. Immature plumage classes have been grouped various- ly in research projects; therefore, we did not attempt inter-study comparisons with our two immature age classes. Results and Discussion We measured 303 Bald Eagles; 201 juveniles (123 males, 77 females, and one of unknown sex), 77 subadults (37 males, 40 females), and 25 adults (9 males, 15 females, and one of unknown sex). In each plumage class, variables tended to be larger in females than in males (Table 1). Pat- terns generally were consistent with Bortolotti’s (1984a) previous descriptions. Feather lengths decreased from Juvenile to adult in both males and females (x = 9.8 and 8.7%, respectively). Differences in EPr, NPr, and Tail lengths were significant among all three plumage classes in both sexes (Table 1 ) . Bill, tarsus, and talon mea- surements tended to increase with age, but less markedly than the feather length decreases. For example, Tail mean in males decreased by 16.1% from juvenile to adult, whereas ClLn and BDp increased by 1.8%. BDp in females increased only 1.4%. Bortolotti (1984a) and Garcelon et al. (1985) both used BDp in their sex determination models. We found no differences in BDp among the three plumage classes of either sex. However, the power to detect differences was relatively low because of the small sample size in the adult age class (Table 1). We recorded full crops in 16% of subadults and adults, and 12% of juveniles. Overall, some food was detected in 33% of crops. Considering only eagles with empty crops, mean masses tend- ed to be greater for adults than for juveniles (4% in males and 5% in females). However, differ- ences were not significant (Table 1). Stalmaster (1987) attributed generally lower masses in young eagles to incomplete bone calcification and muscle development. Some adults may have larger masses due to omental fat (Harmata pers. comm.) . ClLn did not differ between adult eagles at GNP and a southern Colorado wintering area (Harmata 1984) (Table 2). ClLn, BDp, and HalCl did not differ (a = 0.05) between adult eagles at GNP and museum study skins from Canada, Alaska, and the northern U.S. (Bortolotti 1984a) (Table 2). Tail and WSpan of male and female adult Bald Eagles at Harmata’s (1984) study area (37°30'N, 1375 km south of GNP), were significantly longer than our values. Har- mata tracked several of his eagles to nest areas near 55°13'N, roughly 1100 km south of the far- thest north nest site we documented for a GNP adult eagle. Based on latitudinal variation in size (Brown and Amadon 1968), we might have ex- pected feather measurements taken by Harmata to have been shorter than ours. Some differ- ences in size measurements among studies may be attributable to subtle variations in measure- ment techniques. For example, the amount of stretch exerted during wingspan measurement is 122 McClelland et al. VoL. 32, No. 2 Table 1. Sex and plumage class variation in size of Bald Eagles at Glacier National Park, Montana. Mass is in kg and all other measurements are mm. Males Plumage Power Variable Class N x(±SD)® Range F P> F (a = 0.05) WSpan Ad 9 1952 (88.00) (AB)i^ 1800-2100 4.96 0.008 0.804 Sub 35 1966 (61.13) (B) 1840-2090 Juv 124 2007 (85.44) (A) 1810-2370 WnCh Ad 7 569 (20.64) (A) 540-596 20.41 <0.001 0.999 Sub 33 583 (18.00) (A) 560-617 Juv 121 601 (17.87) (B) 570-654 EPr Ad 7 430 (16.86) (A) 409-457 23.37 <0.001 0.999 Sub 30 445 (17.41) (B) 411-481 Juv 121 461 (17.07) (C) 405-520 NPr Ad 7 400 (11.77) (A) 387-417 18.13 <0.001 0.999 Sub 31 417 (15.94) (B) 377-446 Juv 121 429 (15.40) (C) 402-496 TPr Ad 7 306 (11.76) (A) 289-322 13.22 <0.001 0.997 Sub 32 334 (26.52) (B) 297-450 Juv 121 338 (17.06) (C) 295-405 Tail Ad 9 284 (14.75) (A) 265-307 90.91 <0.001 0.999 Sub 36 306 (22.16) (B) 274-380 Juv 124 339 (14.70) (C) 305-393 ClLn Ad 9 51.0 (1.63) (A) 47.7-53.0 1.05 0.354 0.231 Sub 37 50.2 (1.69) (A) 47.7-54.7 Juv 124 50.1 (1.90) (A) 46.0-54.6 BDp Ad 2 32.8 (0.92) (A) 32.1-33.4 0.21 0.814 0.08 Sub 3 32.5 (1.40) (A) 31.0-33.8 Juv 20 32.2 (1.31) (A) 30.0-35.1 NTTF Ad 7 17.4 (1.26) (A) 15.1-19.0 1.23 0.296 0.264 Sub 27 16.6 (1.08) (A) 15.1-19.3 Juv 100 16.8 (1.14) (A) 13.5-17.9 NTTL Ad 8 15.0 (1.53) (A) 13.5-18.4 0.91 0.403 0.206 Sub 32 14.4 (1.03) (A) 13.0-17.3 Juv 122 14.6 (1-22) (A) 11.1-17.9 HalClleft Ad 9 40.1 (1.49) (AB) 37.7-42.0 3.71 0.027 0.674 Sub 36 39.8 (1.21) (B) 37.6-42.4 Juv 124 39.1 (1.61) (A) 36.0-44.2 HalClright Ad 9 40.2 (1.37) (A) 38.5-42.1 3.58 0.030 0.657 Sub 35 39.8 (1.46) (A) 37.0-43.0 Juv 124 39.1 (1.67) (B) 35.8-45.0 OCl Ad 7 28.4 (1.00) (A) 27.3-29.7 2.76 0.066 0.539 Sub 36 28.1 (1.33) (A) 23.3-31.3 Juv 124 27.6 (1.29) (A) 24.5-34.2 MCI Ad 9 32.9 (1.29) (A) 30.0-34.0 9.45 <0.001 0.978 Sub 36 32.9 (1.23) (A) 31.0-36.1 Juv 124 31.9 (1.31) (B) 29.3-36.0 ICl Ad 9 39.5 (1.71) (A) 36.0-42.3 1.11 0.333 0.242 Sub 36 39.1 (1.80) (A) 33.2-42.5 Juv 123 38.8 (1.84) (A) 33.4-46.4 Mass Ad 6 4.30 (0.37) (A) 3.8-4.8 0.66 0.516 0.159 Sub 21 4.22 (0.31) (A) 3.8-5.2 Juv 90 4.14 (0.43) (A) 3.2-5. 1 “ Within a given variable, means followed by the same letter are not different, based on Fisher’s LSD (a = 0.05). ^ With some variables, the power to detect differences was relatively low because of small sample size; e.g., in this case juveniles are different from subadults but not from adults. June 1998 Bald Eagle Size 123 Table 1. Continued. Females N x(±SD)=* Range F P> F Power ( ot = 0.05) 15 2047 (58.39) (A) 1940-2130 11.14 <0.001 0.991 39 2071 (63.61) (A) 1930-2220 75 2118 (68.31) (B) 1940-2290 15 599 (7.81) (A) 583-613 37.36 <0.001 0,999 36 615 (17.80) (B) 580-665 70 631 (12.98) (C) 605-662 13 447 (10.18) (A) 433-462 30.73 <0.001 0.999 34 464 (19.90) (B) 429-546 69 479 (13.24) (C) 447-510 15 420 (8.97) (A) 409-434 21.48 <0.001 0.999 36 434 (22.02) (B) 385-495 70 447 (13.13) (C) 415-479 14 331 (14.44) (A) 312-369 9.54 <0.001 0.978 34 348 (20.62) (B) 313-403 70 351 (12.78) (B) 320-384 15 300 (12.14) (A) 281-334 121.10 <0.001 0.999 40 318 (18.28) (B) 285-358 76 353 (13.82) (C) 308-380 15 56.0 (2.19) (A) 51.0-59.9 8.42 <0.001 0.961 40 55.2 (1.58) (A) 51.7-59.0 76 54.2 (1.71) (B) 50.3-59.2 5 35.7 (1.96) (A) 34.0-39.1 0.38 0.690 0.102 3 35.8 (1.29) (A) 34.7-37.2 13 35.2 (0.94) (A) 33.2-36.6 13 19.6 (1.18) (A) 17.7-21.4 5.17 0.007 0.817 29 18.5 (1.17) (B) 15.6-21.0 60 18.5 (1.06) (B) 16.2-20.7 15 17.0 (1.27) (A) 13.8-18.9 2.28 0.107 0.455 37 16.3 (1.15) (B) 12.2-17.9 74 16.4 (1.00) (AB) 14.0-18.7 14 45.2 (1.63) (A) 42.3-48.0 15.30 <0.001 0.999 40 44.5 (1.42) (A) 41.5-47.8 76 43.3 (1.39) (B) 40.4-46.5 15 44.7 (2.33) (A) 38.6-48.3 11.08 <0.001 0.991 40 44.5 (1.64) (A) 40.5-49.0 76 43.1 (1.64) (B) 39.4-46.8 15 31.5 (1.31) (A) 29.5-34.1 13.05 <0.001 0.997 38 31.3 (1.25) (A) 29.0-34.2 76 30.2 (1.28) (B) 27.5-32.9 15 36.3 (1.63) (A) 33.9-39.1 17.32 <0.001 0.999 40 35.7 (1.93) (A) 27.0-39.1 76 34.4 (1.71) (B) 28.5-38.8 15 44.1 (1.13) (A) 42.3-46.6 22.25 <0.001 0.999 40 43.5 (1.42) (A) 38.2-46.0 76 42.2 (1.22) (B) 39.6-45.6 12 5.14 (0.36) (A) 4.5-5.S 2.28 0.109 0.451 23 5.05 (0.45) (A) 4.3-6.2 49 4.90 (0.41) (A) 4.0-5.8 124 McClelland et al. VoL. 32, No. 2 Table 2. Size measurements of migrant adult Bald Eagles measured at Glacier National Park, Montana (GNP), compared with study skins from Canada, Alaska, and the northern U.S. (Bortolotti 1984a) (B in table), and migrant eagles at San Luis Valley, Colorado (Harmata 1984) (H in table). Mass is in kg; all other units are mm. Vari- able Sex GNP B H N x(± SD) N x(±SD)^ N x(±SD)^ WSpan M 9 1952 (88.00) 17 2010 (54.0)* F 15 2047 (58.39) 17 2140 (57.0)*** WnCh M 7 569 (20.64) 21 570 (12.89) 17 574 (14.0) F 15 599 (7.81) 14 629 (6.87)*** 17 605 (14.0) Tail M 9 284 (14.75) 20 255 (10.67)*** 17 301 (11.0)** F 15 300 (12.14) 14 289 (16.07) 17 317 (11.0)*** EPr M 7 430 (16.86) 21 407 (15.18)** F 13 447 (10.18) 13 452 (16.18) ClLn M 9 51.0 (1.63) 21 50.8 (1.42) 17 50.4 (2.0) F 15 56.0 (2.19) 14 57.2 (1.41)* 17 55.4 (2.0) BDp M 2 32.8 (0.92) 18 32.2 (1.07) F 5 35.7 (1.96) 12 36.9 (1.63) HalCl M 9 40.1 (1.49) 20 39.8 (1.42) F 14 45.2 (1.63) 13 45.7 (1.93) Mass M 6 4.30 (0.37) 17 4.74 (0.7)* F 12 5.14 (0.36) 17 5.32 (0.7) ^ Differences from GNP based on unpooled t-tests: a = 0.1’*'; 0.05**; 0.001***. subjective and may influence the result by several cm. Although Bortolotti (1984b) and Garcelon et al.(1985) found high levels of confidence in the repeatability of measurements within their studies, there may be greater variation between studies. About 25% of the eagles we measured had feath- ers that were incompletely emerged. Molt in north- ern Bald Eagles generally is limited to late spring, summer, and early fall (McCollough 1989). Borto- lotti and Honeyman (1985) suggested that molt in adult Bald Eagles in Saskatchewan continues well into the autumn. Harmata captured migrant eagles 2-5 mo later than we did. We believe this added time for growth of new feathers could have par- tially accounted for the longer feathers in Colora- do. Wear between winter and spring might then produce slightly shorter feather measurements on summer territories. Alternatively, one might argue that wear between fall and winter should have pro- duced shorter measurements in the Colorado win- tering area compared to our fall measurements. Some differences in feather lengths among studies may be partly an artifact of time of year of mea- surements. Acknowledgments Funding was provided by the National Park Service and the Montana Cooperative Wildlife Research Unit, University of Montana. E. Spettigue, L. Young, J. Cren- shaw, H. Allen, R. Williams, R. Yates, E. Caton, M. Mc- Fadzen, V. Wright, R. Bennetts, T.M. McClelland, S. Gniadek, R. Keating, B. Zinn, and R. Bown, helped with field-work. R. Bennetts helped with statistical analyses. R. Bennetts, G. Bortolotti, A. Harmata, J. Smith, and K. Steenhof provided helpful suggestions on an earlier draft of this manuscript. Literature Cited Bortolotti, G.R. 1984a. Sexual size dimorphism and age-related size variation in Bald Eagles. J. Wildl Manage. 48:72-81. . 1984b. Criteria for determining age and sex of nestling Bald Eagles./. Field Ornithol. 55:467-481. AND V. Honeyman. 1985. Flight feather molt in breeding Bald Eagles in Saskatchewan. Pages 166- 178 mJ.M. Gerrard and T.N. Ingram [Eds.], The Bald Eagle in Canada. Proc. of Bald Eagle Days, 1983. White Horse Publishing, Winnipeg, Manito- ba, Canada. Brown, L.H. and D. Amadon. 1968. Eagles, hawks and falcons of the world. Country Life Books, Feltham, Middlesex, U.K. Garcelon, D.K., M.S. Martell, P.T. Redig and L.C. Buoen. 1985. Morphometric, karyotypic, and lapa- roscopic techniques for determining sex in Bald Eagles./. Wildl. Manage. 49:595-599. Harmata, A.R. 1984. Bald Eagles of the San Luis Val- ley, Colorado: their winter ecology and spring mi- June 1998 Bald Eagle Size 125 gration. Ph.D. dissertation, Montana State Univ., Bozeman, MT U.S.A. Imler, R.H. and E.R. Kalmbach. 1955. The Bald Eagle and its economic status. USDI, Fish and Wildl. Serv., Circular No. 30. Washington, DC U.S.A. McClelland, B.R., L.S. Young, P.T. McClelland, J.G. Crenshaw, H.L. Allen and D.S. Shea. 1994. Migra- tion ecology of Bald Eagles from autumn concen- trations in Glacier National Park, Montana. Wildl. Monogr. 125:1-61. McCollough, M.A. 1989. Molting sequence and aging of Bald Eagles. Wilson Bull. 101:1-10. Palmer, R.S. [Ed.]. 1988. Handbook of North Ameri- can birds. Vol. 4. Diurnal raptors. Part 1. Yale Univ Vail-Ballou Press, Binghamton, NY U.S.A. Stalmaster, M.V. 1987. The Bald Eagle. Universe Books, New York, NY U.S.A. Received 10 April 1997; accepted 4 February 1998 J Raptor Res. 32(2): 126-135 © 1998 The Raptor Research Foundation, Inc. A POSSIBLE NEW SUBSPECIES OF THE PHILIPPINE UAS/m-EAGl.E(SPIZAETUS PHILIPPENSIS) AND ITS FUTURE PROSPECTS Monika Preleuthner and Anita Gamauf^ Konrad Lorenz-Institute for Comparative Ethology, Austrian Academy of Sciences, Savoyenstrafie. la, A ll 60 Vienna, Austria Abstract. — On the basis of 19 study skins from eight different museum collections, five captive birds, and 67 field observations (cumulative observation time 6.8 hr), we here describe a possible new sub- species of the Philippine Hawk-Eagle {Spizarius philippensis pinskeri) . This new subspecies is restricted to the rain forests of the southern part of the Philippine Archipelago. Its altitudinal range reaches from 0-1900 m and it occurs at least on Mindanao, Samar, and Negros. The population size of S. p. philippensis is estimated to be 200-220 pairs, that of A p. pinskeri does not exceed 320-340 pairs. We emphasize the current treats to the entire species due to the ongoing destruction of its natural rain forest habitat. Key Words: Philippine Hawk-Eagle, Spizaetus philippensis; new subspecies-, morphology, distribution-, conser- vation; Philippines. Una posible nueva subespecie de Spizaetus philippensis y su futuro Resumen. — Con base en 19 pieles de estudio de 8 colecciones diferentes, 5 aves en cautiverio y 67 observaciones de campo (tiempo de observacion acumulado 6.8 horas), describimos una posible nueva subespecie de Spizaetus philipensis (5. p. pinskeri). Esta nueva subespecie esta restringida a los bosques de Iluvia del sur del archipielago de las Pilipinas. Su rango altitudinal es de 0-1900 m y se encuentra en Mindanao, Samar y Negros. El tamano estimado de la poblacion de S. p. philippensis es de 200-220 pares, el de A p. pinskeri no excede a los 320-340 pares. Enfatizamos las actuales amenazas para la especie entera debido a la continuada destruccion del habitat natural, el bosque de Iluvia. [Traduccion de Cesar Marquez] Taxonomically, birds of prey are well studied. Although only a few new species have been re- cently discovered, the number of accepted spe- cies has increased because some subspecies have been raised to species status (Peters 1931, Brown and Amadon 1969, Mayr and Cottrell 1979, Weick 1980, Amadon and Bull 1988, Sibley and Monroe 1990). According to the recent compi- lation of del Hoyo et al. (1994), the order com- prises 302 species and 702 subspecies. Within the genus Spizaetus (Vieillot 1816), 10 species (in- cluding 20 subspecies) have been described. Five ^ Present address: Museum of Natural History Vienna, 1. Zoological Division — Bird Collection, Burgring 7, A-1014 Vienna, Austria. ^ Since Peters (1931) the authorship for A philippensis usually was credited to Gould. The specific name togeth- er with a description satisfying the criteria for availability was published verbatim as proposed by Gurney, who ac- cording to Art. 50(a) ICZN (1985) should be cited as the author. new subspecies have been added during this cen- tury, three from Southeast Asia and two from Central and South America. In 1993-94, we assessed the habitat use and behavior of 21 species of raptors in the Philip- pines. As part of this ecomorphological study (Gamauf et al. 1998), we measured specimens in 13 museum collections worldwide. One part of the project focused on the endemic Philippine Hawk-Eagle {Spizaetus philippensis Gurney in Gould^ (1863)). Because it is a forest-dwelling species with low population densities, the hawk- eagle is difficult to study and information on its population size and habitat use is rather limited (McGregor 1909, Delacour and Mayr 1946, Ama- don 1953, Brown and Amadon 1969, Mallari 1992, Danielsen et al. 1993). We searched in var- ious museum collections (17 were contacted by mail, 13 visited personally), but preserved spec- imens were available in only eight collections. Al- though the Philippine Hawk-Eagle has been con- 126 June 1998 New Phiuppine Hawk-EIagle Subspecies 127 Figure 1. Top> — holotype of Spizaetns philippensis pinskeri (USNM, Catalog No. 578113). Bottom — ^Typical adult rej>- resentative of Spizaetus philippensis philippensis (AMNH, Catalog No. 459 063) . sidered to be monotypic, we found a striking di- chotomy in the 19 museum specimens we ex- amined. The differences in size, plumage patterns, and coloration were found to be cor- related with the geographic origin of the speci- mens whether they were of northern or southern portions of the Philippines. The distinction was consistently observed in skins, captive birds, and individuals in the wild. On the basis of our com- parisons, we concluded that there are two mor- phologically different and geographically sepa- rated populations of hawk-eagles in the Philip- pines. Although the morphological differentiation appeared clear-cut, the typical disyllabic calls are very similar and there is no proof that there is any reproductive barrier between the two morphs. Since the degree of genetic isolation is unknown, we propose to recognize the northern and southern populations as two distinct subspe- cies rather than separate species. The type spec- imen (The Natural History Museum — Tring, BM 1955.6.N.20.424, type location substituted Luzon by Swann and Wetmore 1945) is herein recog- nized as a representative of the northern nomi- notypical subspecies S. p. philippensis Gurney in 128 Preleuthner and Gamauf VoL. 32, No. 2 Table 1. Holotype and paratype measurements of S. p. pinskeri subsp. nov. in comparison to the nominotypical form dependent on age and sex. Number of study skins ** = P < 0.01 (Mest). are given in parentheses. F = female, M = male. * = P < 0.05, SpIZAETUS PHILIPPENSIS PINSKERI SUBSP. NOV. Age Class Sex Specimens Adult F (3)^ Adult M (2)^ Juvenile Measurement (mm) X ± SD Range X ± SD Range M (1) F (1) Body length 609.0 ± 1.0 608.0-610.0 543.5 ± 41.5 502.0-585.0 550.0 — Wing length 365.5 ± 15.5 350.0-381.0 326.0 — 336.0 378.0 Kipp’s distance 92.5 ± 7.8 87.0-98.0 77.0 ± 7.1 72.0-82.0 87.0 93.0 Number of notched primaries 8.7 ± 0.6 8.0-9.0 8.0 — 8 9 Length of central tail feather 244.0 ± 5.7 234.0-248.0 211.0 ± 5.0 208.0-215.0 216.0 232.0 Length of outermost tail feather 253.0 ± 5.0 248.0-258.0 214.0 ± 15.6 203.0-225.0 221.0 240.0 Length of hind toe 26.9 ± 3.2 24.5-30.5 22.9 ± 1.2 22.0-23.7 23.5 30.0 Length of middle toe 50.9 ± 2.7 49.3-54.0 44.5 ± 3.5 42.0-47.0 49.5 44.2 Length of hind claw 36.9 ± 1.5 35.6-38.5 30.7 ± 1.0 30.0-31.4 34.6 36.2 Length of middle claw 24.7 ± 1.1 23.9-26.0 21.8 ± 1.1 21.0-22.5 23.0 25.0 Tarsus length 82.5 ± 1.3 81.5-84.0 74.7 ± 6.4 70.2-79.2 68.0 88.0 Bill length with cere 40.9 ± 2.1 38.7-42.9 36.5 ± 1.0 35.8-37.2 40.8 41.5 Bill width of distal edge of cere 29.4 ± 1.0 28.3-30.0 28.6 ± 2.6 26.7-30.4 29.5 31.8 Bill depth 24.8 ± 0.5 24.3-25.3 20.5 ± 0.4 20.2-20.8 — 23.8 ^ Holotype included. ^ Subadult included. Gould (1863). We propose that the southern subspecies be designated a new subspecies Spi- zaHus philippinsis pinskeri subsp. nov. Description Holotype. It is an adult female that weighed 1281.2 g and was collected on 16 May 1963 by D.S. Rabor in the CanCan-Mad-Lan Area, Surigao del Sur, Mindanao, Philippines (elevation 330—700 m). Originally, this specimen was in the collection of Silliman University Natural History Museum (Cat- alog No. 35020), but it is now in the U.S. National Museum, Smithsonian Institution (USNM, Catalog No. 578113, Fig. 1). Description of Holotype. The holotype of 5. p. pinskeri is very colorful (Fig. 1). The head and crown show a pale olive buff (SIO: YIO, MOO) (no- menclature taken from Ridgway [1912], color codes from Kiippers [1984]) with a background color of blackish streaks which are more bold on the crown. The head contrasts with the deep brownish olive (S80: Y80, M30) back. The long, prominent crest consists of 4—5 black feathers of unequal length (longest 7 cm) . The throat is white, divided by a bold, black median stripe and bor- dered by black, lateral moustache stripes com- posed of fine black streaks. The plumage on the ventral side is divided into three regions showing different colors. The upper breast is white with bold longitudinal, black streaks edged with tawny olive (S30; Y99, M50). The adjoining area is clearly separated from the upper breast and carries a taw- ny olive (S30: Y99, M50) band with some white, spotted feathers. The lower belly as well as the feathered legs and the undertail coverts are con- trastingly barred, from dark clove brown (Y99: M70, C99) to blackish and white. The wings are short and roundish with tips extending less than halfway to the tailtip. Like the lower belly, the un- derwing coverts are finely barred clove brown (^99; M70, C99) to blackish and white. The up- perparts beginning at the hind neck are uniform brownish olive (S80: Y80, M 30) . The long tail has the same color. A broad black subterminal bar is separated by a broader unmarked zone from five narrower bars. The cere and bill are sooty black (S90: Y20, MOO) and the toes yellow (faded in the specimen) . Measurements of the holotype are as follows: body length 608 mm, wing length 350 mm, Kipp’s distance (primary projection) 87 mm, number of notched primaries 9, length of central tail feather 240 mm, length of outermost tail feather 248 mm, length of hind toe 25.6 mm, length of middle toe June 1998 New Philippine Hawk-Eagle Subspecies 129 Table 1. Extended. SFIZAETUS PHIUPPENSIS FHILIPPENSIS Gvrnev m GoULD (1863) Adult F (4)^ Adult M (6)^ Juvenile M (2) X ± SD Range X ± SD Range X ± SD Range 618.0 ± 16.6 601-634 563.3 ± 22 543-587 — — 399.0 ± 1.0** 398-400 358.0 ± 7.3** 352-372 358.0 — 104.0 ± 5.5 98-108 92.8 ± 5.5 86-99 87.0 — 8.3 ± 0.6 8-9 8.2 ± 0.4 8-9 8.5 ± 0.1 8-9 246.8 ± 11.5 236-249 222.3 ± 11.6 208-236 247.0 — 257.7 ± 5.9 251-262 232.7 ± 7.9 222-245 249.0 — 29.3 ± 1.7 27.8-31.6 25.0 ± 1.4 23-27 26.9 ± 0.8 26.3-27.4 50.6 ± 3.4 47-55 46.0 ± 3.3 43.0-49.5 47.6 ± 1.1 46.8-48.3 38.7 ± 2.0 37.0-40.7 34.6 ± 0.7** 33.8-35.7 32.5 ± 1.3 31.5-33.4 26.8 ± 1.2* 25.2-28.0 25.4 ± 1.0** 24-27 24.3 ± 0.4 24.0-24.5 86.8 ± 5.6 81-94 81.7 ± 2.0 79-84 — 80.1 45.6 ± 2.1 44.3-48.1 40.7 ± 1.4* 39.5-42.5 39.3 ± 1.9 37.9-40.6 30.3 ± 4.0 29.0-35.6 27.1 ± 1.0 25.5-28.3 28.4 ± 0.1 28.3-28.4 25.0 ± 1 23.6-25.9 21.2 ± 1.0 20.3-22.4 20.5 — 49.3 mm, length of hind claw 35.6 mm, length of middle claw 24.4 mm, tarsus length 84.0 mm, bill length with cere 41 . 1 mm, bill width of distal edge of cere 30.0 mm, and bill depth 24.3 mm. Measurements from Museum Specimens. We ob- tained measurements on 42 morphometric char- acters from 19 museum specimens of Philippine Hawk-Eagles, 14 of which were prominent and dis- criminative characters (Table 1). Despite the pro- nounced sexual dimorphism, the measurements of S. p. pinskeri give clearly smaller values than those of S. p. philippensis. This finding was corroborated by discriminant function analysis (Fig. 2) which separated both species in both sexes. The clearest discriminating feature found was the size of the feet and bill, although the distinction held even when wing measurements were included. All spec- imens were correcdy classified and there was no multivariate overlapping of the groups. Discriminant function 1 (DF 1) concerned char- acters related to the mode of handling and killing the prey as well as indirectly to prey size. The length of the bill was loaded the highest on axis 1 , followed by the length of hind claw and the length of the hind toe. DF 2 was correlated with charac- ters describing the killing apparatus but also with the type of prey (especially birds vs. mammals) . It was dominated by the length of the claws of middle toe and hind toe and secondarily by the length and width of the bill (Table 2). Bird hunters typically have longer claws on their middle toes (positive correlation) and shorter hind claws, their bill is shorter and the cross-section of the bill (reflected in the width of the bill) is more roundish com- pared to the mammal hunters. This suggests that members of the bigger S. p. philippensis may have a higher proportion of mammals in their diet whereas S. p. pinskeri may preferentially feed on birds (Rochon-Duvigneaud 1952, Wattel 1973, Brown 1976, Hertel 1995, Gamauf et al. 1998). Etymology Scientific Name. The proposed new subspecies is named in honor of Prof. Dr. Wilhelm Pinsker, Institute for Medical Biology, University of Vienna, for his excellent scientific work as well as his emi- nent skill in the guidance of his students as a teach- er. We wish to emphasize our gratitude for the in- valuable help he has given to both of us. English (German) Names. According to the geo- graphical distribution of the two subspecies, we propose the name Southern Philippine Hawk-Ea- gle (Siidlicher Philippinenhaubenadler) for 5. p. pinskeri and Northern Philippine Hawk-Eagle (Nordlicher Philippinenhaubenadler) for the nominotypical subspecies S. p. philippensis. Paratypes. The three specimens from Mindanao are herein designated paratypes: 130 Preleuthner and Gamauf VoL. 32, No. 2 Figure 2. Separation of S. p. philippensis {N = 12, round symbols) and S. p. pinskeri {N = 7, square symbols) according to discriminant function analysis of 5 morphological variables (bill length, bill width, length of the hind toe, hind claw and middle claw). 1. Adult male collected on 30 January 1964 by D.S. Rabor at Tucay, Mt. Matutum, South Cotabato, Mindanao (Museum of Natural History, Univer- sity of the Philippines at Los Banos [UPLB], Philippines, Catalog No. 108). 2. Subadult male collected on 12 March 1953 by F. Solomonsen at Talacogon, Agusan del Sur, Mindanao (University Museum of Zoology [UMZC], Copenhagen, Denmark, Catalog No. 936). 3. Immature female (we determined it to be a male) collected on 30 October 1946 by D.S. Ra- bor at Maduum Tagum, Davao del Norte, Min- danao (Field Museum of Natural History [FMNH], Chicago, IL U.S.A., Catalog No. 1247). Table 2. Canonical disciminant analysis of four hawk-eagle groups (5. p. philippensis, S. p. pinskeri, males and females) . Shown are loadings on discriminant function axes and results of univariate F-tests. Character Axis 1 Length of hind toe 0.255 Length of hind claw 0.430 Length of middle claw -0.119 Bill length with cere 0.508 Bill width of distal edge of cere 0.175 Percent variance explained 62.1 Axis 2 F P -0.067 4.610 <0.05 -1.294 17.512 <0.00001 1.819 10.136 <0.001 0.459 16.199 <0.0001 0.289 2.208 n.s. 37.2 June 1998 New Philippine Hawk-Eagle Subspecies 131 The three remaining specimens from the south- ern population were collected in Samar and Ne- gros. One was an adult female collected by D.S. Rabor on 6 April 1957 at Matuguinao, Samar (FMNH, Catalog No. 247 377) . The second was an adult male (we determined it to be a female) col- lected on 21 December 1955 by D.S. Rabor at Bas- ey, Bayawan, Negros Oriental (UPLB, Catalog No. 107) and the third specimen was an immature fe- male collected on 1 August 1871 by A. Everett at Nueva Valencia, Negros Oriental (Tweeddale Col- lection, The Natural History Museum [BMNH], Tring, Herts, U.K., Catalog No. 87.11.1.333). Description. In adults of S. p. pinskeri, the colors of the head and nape vary from ivory (SOO: YIO, MIO) to pale olive grey (S20: YOO, MOO) or dark olive buff (S30: Y50, M20) with more or less fine black shaft streaks. These shaft streaks can become very bold so that the crown appears black such as the specimen from Samar. The black occipital crest has a maximum length of 8 cm. The throat is al- ways completely white. The upper breast has a white ground color with pronounced black streaks, in some cases especially at the distal side of the body edged with tawny olive (S30: Y99, M50) . The color of the lower breast and flanks is more vari- able, ranging from yellow ocher (S20: Y80, M20) to cinnamon (S20: Y70, M40), clay color (S30: Y60, M50) or tawny olive (S30: Y99, M50) with more or less pronounced bars. This pattern appears only in adult birds, which are at least in their fourth cal- endar year. The lower belly and the legs are in- variably barred clove brown (Y99; M70, C99) to black and white. Back varies between raw umber (Y99: M70, C80), brownish olive (S80: Y80, M30) or sepia (S80: Y99, M40). The cere and bill are sooty black (S90: Y20, MOO) in study skins and liv- ing birds. The unfeathered toes are apricot yellow (SOO: Y60, M20) to lemon crome (SOO: Y90, M30) in living birds, bright lemon crome are also the eyes. Since juvenile hawk-eagles are difficult to distin- guish from other species, we give a rather detailed account of the plumage pattern. The juvenile plumage is white at the ventral side, also on head and neck, except the long black crest feathers. Only one young bird, observed and photographed at Mt. Kitanglad, Mindanao, had dark grey flanks. The feathered legs are white with fine warm buffy (S20: Y60, M20) bars on the tibiotarsus. There is a gradual change in color from the white head to the dark back. The broadly pale-edged feathers vary between mouse grey (S70: YIO, MIO) to chae- tura drab (S80: Y30, M30). The uppertail coverts are also white. The cere and bill are sooty black (S90: Y20, MOO) and together with the black lores they form a dark mask. The toes in living birds are yellow as in adults, whereas the eye coloration changes from dark grey in juveniles to bright lem- on crome (SOO: Y90, M30) in subadults (at least in Basic III) and adults. The median and lesser wing coverts are white and form a broad band on the upper side of the wing. The secondaries are deep mouse grey (Y40; M30, C50) , the primaries blackish. Both are heavily barred with 7-9 relatively fine bars. In soaring birds, a narrow white sickle-like panel is seen in backlighting along the base of the primaries. On the mouse grey (S70: YIO, MIO) to chaetura drab (S80: Y30, M30) colored tail, 6—7 bars are regularly spaced or one broad subterminal bar is discern- ible. Specimens Examined. Including the holotype, seven study skins of S. p. pinskeri were available. In addition, two captive birds were examined at the Breeding Center of the Philippine Eagle Conser- vation Program Foundation in Toril, Davao. Sixty individuals were observed in the field on the island of Mindanao. For comparisons, measurements were taken from 12 skins of representatives of S. p. philippensis (five including the holotype from BMNH — Tring, three from the DMNH, one from the CMNH, one from the RNMS, one from the AMNH, and one from the FMNH). Ten of these 12 birds come from Luzon but the origins of the other two specimens could not be traced. We also studied three captive birds, one in the Manila Zoo- logical Garden and two at the Wildlife Research Center in Manila. In Luzon, we observed the nom- inotypical subspecies eight times. The cumulative observation time for both subspecies in their nat- ural habitat was 6.8 hr during three field trips to the Philippines (January-April 1993, November 1993-February 1994, and March-July 1994). The study of plumage changes in captive birds was es- sential for age determination. Diagnosis. The proposed new subspecies can be distinguished from S. p. philippensis by the smaller size of both sexes (Table 1), and by the different plumage coloration and plumage patterns on the head, breast and belly. In contrast to S. p. pinsken, the head of S. p. philippensis is raw sienna (S30: Y80, M50) to antique brown (S40: Y99, M50) with broad blackish streaks and the throat is mainly fine ivory 132 Preleuthner and Gamaue VoL. 32, No. 2 to warm buffy (S20: Y60, M20) with fine dark streaks. The breast is ochraceous-tawny (S30: Y90, M50) and the belly somewhat darker, antique brown (S40: Y99, M50) to cinnamon-brown (S40: Y90, M50). The bold black streaks on the breast are less contrastful. Individuals of S. p. philippensis have fine whitish bars on the cinnamon-brown (S40; Y90, M50) to clove brown (Y99: M70, C99) feathered legs and undertail coverts. Almost all il- lustrations of Philippine Hawk-Eagles found in the literature (e.g., Walden 1875, Brown and Amadon 1969, Weick 1980, del Hoyo et al. 1994) depict rep- resentatives of S. p. philippensis. One exception is the individual shown in duPont (1971), which matches with 5. p. pinskeri in the most important characters. To our knowledge, only a single pho- tograph by M.C. Witmer published in Gonzales and Rees (1988) shows an adult S. p. pinskeri (cap- tive bird at the Philippine Eagle captive breeding center, Barracatan, Toril, Davao City; R.S. Kennedy pers. comm.). The morphological traits and differences in col- or patterns do not vary clinally from north to south. Our data indicate that the boundary be- tween the subspecies runs along the San Bernar- dino Channel which separates Luzon and Samar by a distance of less than 20 km. Within the Phil- ippines, this line is also known to separate taxa of other vertebrates (birds, mammals, reptiles) with limited ability to disperse across saltwater channels (Heaney 1986, ICBP 1992). Effects of the Pleistocene history, with repeated land bridge connections between many of the is- lands, may serve as an explanation for the present distribution pattern (Hauge et al. 1986). Growth and recession of continental glaciers during the Pleistocene were associated on a global basis with changes in sea level and temperature. In the late middle Pleistocene (about 160 000 yr ago), sea lev- el was about 160-180 m below the present level. The shallow (140 m) San Bernardino Strait be- tween southern Luzon and northern Samar may have been dry during this period, allowing free ex- change throughout much of the archipelago (Hea- ney 1985) and also to the large Visayan islands like Negros. At the end of the Pleistocene, about 18 000 yr ago, sea level had risen to 120 m below the pres- ent coast line covering the channel with a hody of water 20 m depth that cut off the connection be- tween the islands. Thus, the separation of the two subspecies must have happened after the geo- graphical isolation, which cannot be dated accu- rately. Due to particular morphological features, especially the short, hroad and round wings, the Philippine Hawk-Eagle is well adapted for rainfor- est habitats. These apparent adaptations do not al- low long-distance flights across open habitats as it is actually the case between Luzon and Samar (Ga- mauf et al. 1998). Therefore, disruption of the gene flow by such a geographic barrier could have led to genetic isolation between the northern and southern populations and subsequently to the di- vergence into separate subspecies. This isolation process is enhanced by further subdivision of the extant populations through the fragmentation of the habitat. To explain the size differences between 5. p. pin- skeri and S. p. philippensis we have to take into con- sideration interactions and competition with other species. One reason for the smaller size of S. p. pinskeri could be niche separation with respect to other sympatric eagles also specialized for the hunting of large mammals and birds. In the south, two larger eagle species exist, the Changeable Hawk-Eagle {Spizaetus cirrhatus) and the Philippine Eagle {Pithecophaga jefferyi). In contrast, on the northern island Luzon, only the Philippine Eagle is larger than the Philippine Hawk-Eagle. The Ru- fous-bellied Eagle {Hieraaetus kienerii) comes next in size, a species which has been recorded from the whole Philippine Archipelago. Thus, both sub- species fit approximately into the respective size gap between their food competitors. The size dif- ference to its larger competitor is greater in S. p. philippensis than in 5. p. pinskeri. The wider niche reflects the larger body size and the more pro- nounced sexual size dimorphism in 5. p. philippen- sis (Fig. 2). Actual Situation and Future Prospects Habitat and Distribution. We found the Philip- pine Hawk-Eagle in large, continuous areas of dip- terocarp rainforests. It definitely prefers extensive primary, or well-structured, old secondary forests which were selectively logged 20-30 yr ago. Occa- sionally even transitional stages to semi-open hab- itats are used. With respect to altitude, 5. p. pinsken was observed from almost sea level up to 1900 m. Of the 11 islands from which the species has been recorded so far (Dickinson et al. 1991, Brooks et al. 1992, Evans et al. 1993a, 1993b), only the main island Luzon is inhabited by S. p. philip- pensis with certainty. Mindanao, Negros, and Samar are doubtlessly populated by S. p. pinskeri as docu- June 1998 New Philippine Hawk-Eagle Subspecies 133 merited by study skins. Individuals recorded from Leyte and probably also from Biliran, Basilan, Si- quijor, and more recently from Bohol (Hornskov 1995) presumably belong to S. p. pinskeri. No un- equivocal evidence for the occurence of either sub- species is known so far from Palawan, The only museum specimen labeled as “Philippine Hawk- Eagle” from the locality Palawan was a misidenti- fied Changeable Hawk-Eagle (Staatliches Naturhis- torisches Museum Braunschweig, Catalog No. 14158, collector Dr. C. Platen). In general, infor- mation on the distribution of the Philippine Hawk- Eagle (and other bird species) is still insufficient and further studies are needed. Conservation. Most current threats to raptors in the tropical forest belt are related to habitat de- struction (Kennedy 1986, Thiollay 1994). Owing to the construction of new roads for timber harvest- ing, often in a very damaging way, the final destruc- tion is conducted by shifting cultivators who enter the region illegally. An additional threat arises from the fact that the Philippines are one of the most densely populated countries in Southeast Asia, with more than 65 million people on 300 000 km^. Dickinson et al. (1991) have presented data on the extent of forests remaining on the major island groups. They found that only the major islands possess sufficiently large forested areas. According to Collins et al. (1991), the deforestation rate (de- termined for the period 1986-90) is 1380 km^/yr for the whole Philippines. In 1988, the forested area with closed canopy cover on the island of Lu- zon was determined as 7621 km^. Assuming a uni- form rate of deforestation over the Philippines, the remaining forest habitat is roughly estimated to be 5000 km2. Knowing the inhabitable area and the present population density, we attempted to gauge the ac- tual population number. In six study areas, we de- termined the number of pairs using two methods: census from exposed points (cliffs, clearings, prominent trees) and line transects. These two methods have been used successfully in studies on tropical rain forest raptors (Thiollay 1989, Whita- cre et al. 1992) . Point censuses proved very suitable to map the locations of hawk-eagles since they are year-round residents that often fly above the can- opy. In each study area, the total time of observa- tion was at least 2 wk. The density of the Philippine Hawk-Eagle in Luzon was determined in two study areas in the Sierra Madre mountain range in the south of Quirino province. Around Asaclat (35 km^, 500-900 m), we found two pairs (5.7 pairs/ 100 km^) and in Don Mariano Perez (32 km^, 400— 1100 m) one pair (3.1 pairs/100 km^). Based on this survey, we estimated the size of the S. p. philip- pensis population on Luzon to be about 200—220 pairs. For S. p. pinskeri, we surveyed four study areas in Mindanao. In the lowland forests of PRI (former PICOP) (Surigao del Sur, 58 km^, 90-180 m), we found 3-4 pairs (5. 2-6. 9 pairs/100 km^) and in Carmen-Cantilan (Surigao del Sur, 27 km^, 80-540 m) one pair (3.7 pairs/100 km^). At the gentle slope on the Dalwangan side of Mt. Kitanglad (Bu- kidnon, 38 km^, 900-1800 m), we found three pairs (7.9 pairs/100 km^) and on the steep slope on the Barracatan side of Mt. Apo (Davao City, 25 km^, 950-1800 m) we found one pair (4 pairs/ 100 km^) . Thus, in the total 7500 km^ of closed canopy forest available, we estimated the number of pairs may be about 320—340. In the world list of threatened birds (Collar et al. 1994), the Philippine Hawk-Eagle is listed as Vulnerable with a high risk of extinction in the wild within the medium-term future. Based on our recent investigations, the estimated maximal num- ber of mature individuals is <1200, a population size clearly below the criterion of <2500 set by BirdLife International for endangered species. Ac- cording to this criterion the classification of the Philippine Hawk-Eagle as an Endangered species seems appropriate. Acknowledgments The work was supported by the Austrian Science Foun- dation (FWF-project P-8889-Bio) and the IWJ, University of Agriculture, Vienna. We want to express our sincere gratitude to the Department of Environment and Natural Ressources (DENR) of the Republic of the Philippines, the Philippine Eagle Conservation Program Foundation, the Haribon Foundation, Green Mindanao, the industry companies in Carmen (Puyat Logging) and Bislig (PRI) as well as the Technical Aid Agency of the Federal Re- public of Germany (GTZ) and our local guides for their excellent cooperation. The authors are very much obliged to the curators of the following museum collections for access to specimens in their care: The Natural History Museum (BMNH), Bird Group (Tring, U.K.), Royal National Museum of Scotland (RNMS, Edinburgh, U.K.), Universitets Zoolo- giske Museum (UMZC, K0benhavn, DK), Rijksmuseum van Natuurlijke Historic (RMNH, Leiden, NL), Zoolo- gisches Museum der Humboldt Universitat Berlin (ZMB, Berlin, D), Staatliches Naturhistorisches Museum Braun- schweig (SNMB, Braunschweig, D), Naturhistorisches Museum Wien (NMW, Wien, A), American Museum of Natural History (AMNH, New York, NYU.S.A), Smith- 134 Preleuthner and Gamauf VoL. 32, No. 2 sonian Institution (USNM, Washington DC, U.S.A.), Field Museum of Natural History (FMNH, Chicago, IL U.S.A.), Cincinnati Museum of Natural History (CMNH, Cincinnati, OH U.S.A.), Delaware Museum of Natural History (DMNH, Wilmington, DE U.S.A.), National Mu- seum of the Philippines (PNM, Manila, PH), University of the Philippines at Los Banos (UPLB, Los Banos, PH), and Zoological Garden Manila (Manila, PH). We are es- pecially grateful to R. Prysjones, M. Walters and P. Col- ston, R. McGowan, J. Fjeldsa, C. Smeenck and R. Dekker, B. Stephan, G. Boenig, E. Bauernfeind, G.F. Barrow- dough and P. Sweet, D. Zusi and P. Angle, D. Willard and P. Baker, R.S. Kennedy and J. Brown, G. Hess, P.C. Gonzales, A. Dans and R. A. Andres for their helpful co- operation. We are very much indebted to A. Schuster, S. Tebbich and M. Zeiler for their assistance in the field. We are also obliged to H. Winkler for valuable suggestions, and to E. Bauernfeind, G. Bortolotti, R.S. Kennedy and an anony- mous reviewer for critical comments on the manuscript. Literature Cited Amadon, D. 1953. Remarks on the Asiatic Hawk-Eagles of the genus Spizaeetus. Ibis 95:492-500. AND J. Bull. 1988. Hawks and owls of the world: a distributional and taxonomic list. West. Found. Vertebr. Zool. 3:297-357. Brooks, T.M., T.D. Evans, G.C.L. Dutson, G.Q.A. An- derson, D.C. Asane, R.J. Timmins and A.G. Toledo. 1992. The conservation status of the birds of Negros, Philippines. Bird Conserv. Int. 2:273-302. Brown, L.H. 1976. Birds of prey, their biology and ecol- ogy. Hamlyn Publishing Group, Feltham, U.K. AND D. Amadon. 1969. Eagles, hawks and falcons of the world. Country Life Books, London, U.K. Collar, N.J., M.J. Crosby and A.J. Stattersfield. 1994. Birds to watch 2: the world list of threatened birds. BirdLife International, BirdLife Conservation Series No. 4, Cambridge, U.K. Collins, N.M.,J.A. Sayer and T.C. Whitmore. 1991. The conservation atlas of tropical forests — ^Asia and the Pa- cific. Macmillan Press, London, U.K. Danielsen, E, D.S. Balete, T.D. Christensen, M. Hee- GAARD, O.F. Jacobsen, A. Jensen, T. Lund and M.K. PouLSEN. 1993. Conservation of biological diversity in the Sierra Madre Mountains of Isabela and southern Cagayan Province, The Philippines. DENR, ICBP, ZMCU & Danish Ornith. Soc., Manila, The Philip- pines. Delacour, J. and E. Mayr. 1946. Birds of the Philippines. The Macmillan Company, New York, NY U.S.A. Dickinson, E.C., R.S. Kennedy and K.C. Parkes. 1991. The birds of the Philippines: an annotated check-list. British Ornithologists’ Union, Check-list No. 12, Tring, U.K. DEL Hoyo, j., a. Elliot and J. Sargatal. 1994, Hand- book of the birds of the world. Vol. 2. New World vultures to guineafowl. Lynx Edicions, Barcelona, Spain. DUPONT, J.E. 1971. Philippine birds. Monograph Series No. 2. Delaware Museum of Natural History, Green- ville, DE U.S.A. Evans, T.D., G.C.L. Dutson and T.M. Brooks. [Eds.]. 1993a. Cambridge Philippines rainforest project 1991. final report (study report no. 54). Birdlife Interna- tional, Cambridge, U.K. , P. Magsalay, G.C.L. Dutson and T.M. Brooks. 1993b. The conservation status of the forest birds of Siquijor, Philippines. Forktail 8:89-96. Gamauf, A., M. Preleuthner and H. Winkler. 1998 Philippine birds of prey: interrelations among habitat, morphology, and behavior. Auk 115:in press. Gonzales, P.C. and C.P. Rees. 1988. Birds of the Philip- pines. Haribon Foundation, Manila, The Philippines Gould, J. 1863, The Birds of Asia. Part XV. Publ. by the author, London, U.K. Gurney, J.H. 1863. Spizaetus philippensis. Text to plate Spi- zaetus alboniger m J. Gould, The Birds of Asia. Part XV Publ. by the author, London, U.K. Hauge, R, j. Terborgh, B. Winter and J. Parkinson 1986. Conservation priorities in the Philippine Archi- pelago. Forktail 2:83-91. Heaney, L.R. 1985. Zoogeographic evidence for middle and late Pleistocene land bridges to the Philippine Islands. Mod. Quat. Res. SE" Aiza 9:127-143. . 1986, Biogeography of mammals in SE Asia: Es- timates of rates of colonization, extinction and speci- ation. Biol. J. Linn. Soc. 28:127-165. Hertel, F. 1995. Ecomorphological indicators of feeding behavior in recent and fossil raptors. Auk 112:890- 903. Hornskov, j. 1995. Recent observations of birds in the Philippine Archipelago. Forktail 11:1-10. ICBP. 1992. Putting biodiversity on the map: priority ar- eas for global conservation. International Council for Bird Preservation, Cambridge, U.K. ICZN. 1985. International code of zoological nomencla- ture. Third edition. Univ. Calif Press, London, U.K Kennedy, R.S. 1986. Raptors in the tropics — the next 50 years. Raptor Res. Rep. No. 5:17-25. Kuppers, H. 1984. DuMonts Farben-Atlas. Third Edition DuMont Buchverlag, Cologne, Germany. Mallari, N.A.D. 1992. Bird species diversity along eleva- tional gradients in different sites in the Sierra Madre Mountains in Isabela and Cagayan. Masters thesis, UPLB, The Philippines. Mayr, E. and G.W. Cottrell. 1979. Check-list of the birds of the world. Vol.l, second edition. Mus. Comp. Zoology Cambridge, MA U.S.A. McGregor, R.C. 1909. A manual of Philippine birds. Bu- reau of Science, Manila, The Philippines. Peters, J.L. 1931. Check-list of birds of the world. Vol. 1 Harvard Univ. Press, Cambridge, MA U.S.A. Ridgway, R. 1912. Color standards and color nomencla- ture. United States National Museum, Washington, DC U.S.A. June 1998 New Philippine Hawk-Eagle Subspecies 135 Rochon-Duvigneaud, a. 1952. Les armes des rapaces. Oiseau Rev. Fr. Ornithol. 22:91-106. Sibley, C.G. and B.L. Monroe. 1990. Distribution and taxonomy of birds of the world. Yale Univ. Press, New Haven, CT U.S.A. Swann, H.K. and A. Wetmore. 1945. A monograph of the birds of prey (Order Accipitres). Vol. II. Wheldon 8c Wesley, London, U.K. Thiollay, J.-M. 1989. Censusing of diurnal raptors in a primary rainforest: comparative methods and species detectability./. Raptor Res. 23:72-84. . 1994. A world review of tropical forest raptors — current trends, research objectives and conservation strategy. Pages 231-239 in B.-U. Meyburg and R.D. Chancellor [Eds.], Raptor conservation today. World Working Group on Birds of Prey, Pica Press, Mount- field, U.K. ViEiLLOT, L.P. 1816. Analyse d’une nouvelle Ornithologie elementaire. Paris, France. Walden, A.V. 1875. II. A list of the birds known to inhabit the Philippine Archipelago. Trans. Zool. Soc. Lond. 9' 125-252. Wattel, J. 1973. Geographic differentiation in the genus Accipiter. Publ. Nuttall Ornith. Club No. 13, Museum of Comparative Zoology, Harvard Univ., Cambridge, MA U.S.A. Weick, F. 1980. Die Greifvdgel der Welt. Parey Verlag, Hamburg, Germany. Whitacre, D.A., L.E. Jones andJ. Sutter. 1992. Census- ing raptors and other birds in tropical forest: further refinements of methodology. Pages 39-52 in D.A. Whitacre and R.K. Thorstrom [Eds.], Maya Project The Peregrine Fund, Boise, ID U.S.A. Received 6 June 1997; accepted 9 February 1998 J. Raptor Res. 32(2);136-142 © 1998 The Raptor Research Foundation, Inc. PATTERNS OF EGG AND CLUTCH SIZE VARIATION IN THE MONTAGU’S HARRIER Beatriz Arroyo Centre National de la Recherche Scientifique, Centre d’Etudes Biologiques de Chize, Villiers en Bois, F-7 9360 France Alain Leroux Cisse, F-86000 France Vincent Bretagnolle Centre National de la Recherche Scientifique, Centre d Etudes Bioloques de Chize, Villiers en Bois, F-7 9360 France Abstract. — ^We describe the phenotypic plasticity of egg and clutch size in the Montagu’s Harrier (Circus pygargus) based on measurements of 1292 eggs from 403 nests and clutch sizes recorded at 579 nests in four study areas in western France and central Spain. Variability of egg size, clutch size and clutch volume was high (coefficients of variation 10, 24, and 28%, respectively). Egg volume and shape (length/ width) were positively correlated and bigger eggs tended to be relatively longer. Shape was not normally distributed, with elongated eggs being more common than rounded ones. Montagu’s Harriers appeared, therefore, to be less constrained to increase egg length than egg width. Egg volume was positively correlated with clutch size. The first and last eggs in each clutch were on average smaller than the intermediate ones, but the differences were not significant. High interannual variation existed in clutch size, but not in egg size. No regional variation was found in egg or clutch size, once interannual differ- ences were taken into account. Overall, clutch size appeared to be a more plastic trait than egg size, although the large variability in the latter suggested that it had the potential, at least, to vary according to environmental or individual conditions. Key Words: Montagu ’.s Harrier, Circus pygargus; clutch siije, egg size, phenotypic plasticity. Patrones de variacion en el tamaho de puesta y de huevos en el Aguilucho Cenizo Resumen. — Describimos la variabilidad fenotipica del tamaho de la puesta y los huevos en el Aguilucho Cenizo (Circus pygargus) basandonos en medidas de 1292 huevos procedentes de 403 nidos y en los tamahos de puesta observados en 579 nidos de 4 zonas de estudio localizadas en Francia occidental y Espaha central. La variabilidad tanto del tamaho de los huevos como sobre todo del tamaho de puesta y del volumen de puesta fueron muy altas (con coeficientes de variacion de 10, 24, y 28%, respectiva- mente). Se encontro una correlacion positiva y significativa entre volumen y forma (longitud/ anchura) de los huevos, es decir que los huevos mas grandes tendian a ser proporcionalmente mas alargados. La variable forma no estaba distribuida de manera normal; los huevos alargados eran mas frequentes que los redondeados. Los Aguiluchos Cenizos parecen por tanto estar menos limitados para aumentar la longitud que la anchura de los huevos. Se encontro tambien una correlacion positiva y significativa entre el volumen de los huevos y el tamaho de puesta. Los primeros y ultimos huevos de cada puesta eran en general mas pequehos que los intermedios, pero las diferencias no fueron significativas. Se encontraron diferencias interanuales significativas en el tamaho de puesta, pero no en el tamaho de huevos. En cambio, las diferencias entre las zonas de estudio no fueron significativas para ninguna de las dos variables. En general, el tamaho de puesta es un caracter mas flexible que el tamaho de los huevos, aunque la gran variabilidad en este ultimo sugiere que tiene al menos el potencial de variar en relacion a diversas caracterfsticas individuales o ambientales. [Traduccion Autores] Among avian life history traits, egg size and clutch size have been traditionally considered as two major components of fitness (Stearns 1992). These traits are direct measures of female invest- ment in reproduction which have been shown to influence reproductive success in raptors (e.g., 136 June 1998 Egg and Clutch Size in Montagu’s Harrier 137 Hakkarainen and Korpimaki 1994, Wiebe and Bor- tolotti 1995). Understanding the patterns of egg and clutch size variation is, therefore, important in order to assess whether these traits are potentially adaptive. Montagu’s Harrier {Circus pygargus) is a medium-sized palearctic raptor that breeds on the ground (Cramp and Simmons 1980), sometimes colonially (e.g., Cramp and Simmons 1980, Leroux 1987, Martelli 1987, Karpov and Berbaev 1990, Ar- royo 1995). These characteristics facilitate sam- pling of a large number of nests, and indeed there are several studies reporting clutch sizes (Perez Chiscano and Fernandez Cruz 1971, Cormier 1985, Underhill-Day 1990, Pandolfi and Giacchini 1991, Martelli and Sandri 1991, Bijlsma et al. 1993, Krogulec and Leroux 1994). Even so, there are few multi-year studies, where among-year differences can be evaluated (Butet and Leroux 1993, Faralli 1994, Arroyo 1995, Castano 1997), and fewer stud- ies which report egg sizes (Studinka 1941, Schon- wetter 1967, Hays 1971, Perez Chiscano and Fer- nandez Cruz 1971). The latter data, when given, refer just to unhatched eggs or are presented only as mean values. The general patterns and distri- bution of egg size variation have never been de- scribed for this species, nor has the relationship between egg size and clutch size. The aims of this paper are to describe the phenotypic plasticity of egg size in the Montagu’s Harrier, its annual and geographical variability, and its relationship with clutch size. We show that egg size is highly variable, but that plasticity of egg volume is less than that of clutch size, with the latter showing a higher annual and regional variation. Methods We systematically searched for nests in four study areas, which cover 200-300 km^ each. Three of them were lo- cated in western France: Marais de Rochefort (45.57°N, 0 55°W, Charente-Maritime region) , for which clutch size and egg size data were available from 1992—97; south of Deux Sevres (46.11°N, 0.28°W, Poitou-Charente region), for which clutch- and egg-size data were available from 1994 - 97 ; and Bale de I’Aiguillon (46.24°N, 1.24°W, De- partement de la Vendee), with data collected from 1995- 97. The fourth study area is northeast of Madrid, Spain (40.38°N, 3.30°W); clutch-size data from this area were collected since 1991 and egg-size data were available mainly for 1994, 1996 and 1997, although some scattered data existed from previous years. Clutch size was known for 579 clutches. Additionally, 1292 eggs from 403 of these clutches were measured in the four study areas. Egg length and width were mea- sured with a vernier calliper to the nearest 0.1 mm. From these two measurements, we derived two additional vari- ables: egg volume, which was calculated from Hoyt’s (1979) equation (volume (cm-J = 0.51 X length X width^/1000) , and egg shape (length/width). To test for differences between measurers, 24 eggs in 1996 were measured by two different people. Measurements be- tween observers were usually within 0.2 mm, and differ- ences were rarely higher than 0.3 mm (<1% of average length or width). Slight, but significant differences were found in measurements of egg length (paired t-test, = 2.30, P = 0.03). No significant differences were found for egg width measurements (paired t-test, ^24 ^ 132, F = 0 . 20 ), or for egg volumes calculated from measure- ments from both observers (paired ttest, ^34 = 1.00, P = 0.33). Thus, all data were pooled regardless of measurer. Mean egg volume and clutch volume (sum of egg vol- umes in a clutch) were calculated for each clutch. In 47 clutches, measurements for some eggs were missing (ei- ther because one or two eggs had already hatched by the visit when egg measurements were taken, c>r because the nest failed after the first visit, when the clutch was incom- plete). In those cases, clutch volume was calculated as mean egg volume (calculated from measured eggs) X clutch size (when hnal clutch size was known). Clutches of more than one egg in which only one egg was mea- sured {N = 63) were excluded from clutch analysis. Eggs were marked for identification with a (nontoxic) pencil when nests were visited. Egg rank was estimated for nests that were visited at least once during laying, according to shell color (eggs freshly laid being tinged blue, Balfour 1962), and degree of cleanliness (old eggs variably covered with earth, Simmons 1994). The rela- tionship between egg rank and egg volume was analyzed for each clutch, grouping eggs in three egg categories (following Jover et al. 1993): first egg in the clutch, last egg in the clutch, and intermediate eggs (for which egg volume was averaged) . Clutches where all eggs could not be assigned to one of these three categories were exclud- ed for egg-rank analysis. Accuracy of ranking method was tested for 107 eggs from 50 nests that were visited during hatching, and proved to be higher than 94%. Annual variation in mean egg size was calculated using clutch data from the four study areas for 1994—97 alone (as data for earlier years were not available or adequate for all study areas). Annual and regional variability in clutch size was analyzed with data from all nests visited, whether eggs were measured or not. All analyses were performed with SAS 6.11 statistical package. Results The distributions of neither egg length, width or volume departed from normality (Table 1). Egg- shape, in contrast, differed significantly from a normal distribution (skewness = 0.67, Fig. 1), with long eggs being more frequent than round ones. Overall variation in egg length, as seen from the coefficient of variation, was slightly higher than that of egg width. The smallest egg was only 37.9% of the volume of the largest. Egg shape and egg volume were positively and significantly correlated 138 Arroyo et al. VoL. 32, No. 2 Table 1. Summary statistics for Montagu’s Harrier eggs and clutches. N = sample size (number of eggs for egg biometrics, number of clutches for clutch size or clutch volume); SD = Standard Deviation; CV = Coefficient of variation, 100 X SD/;^ P = probability for departure from the normal distribution (Wilks-Saphiro test). Measurement N Mean SD Range CV P Egg-length (mm) 1292 41.5 1.96 35.9-49.2 4.7 0.09 Egg-width (mm) 1292 33.1 1.22 27.0-39.1 3.7 0.20 Egg-volume (cm’’) 1292 23.2 2.36 13.3-35.2 10.2 0.98 Egg-shape 1292 1.25 0.06 1.05-1.52 4.8 0.0001 Clutch size 579 4.0 0.97 1-8 24.2 0.0001 Clutch volume (cm^) 313 92.9 25.7 21.5-195.8 27.6 0.07 (r = 0.079, N = 1292, P = 0.004) : larger eggs tend- ed to be relatively more elongated. A one-way ANOVA with all years and sites com- bined showed that 59% of the variance in mean egg volume was due to differences among clutches as opposed to differences within clutches (F 2121002 = 6.82, P < 0.0001). Regarding within clutch vari- ation, there was a tendency for the first and last eggs of a clutch to be smaller than the intermedi- ate ones (Fig. 3), although differences were not significant (one-way ANOVA, F 2 376 = 0.7, P = 0.5). Clutch size and clutch volume were more vari- able than egg size (see CV in Table 1). Most of the variability in clutch volume was due to clutch size (linear regression, F^ 311 = 2451, P = 0.0001, 7 ^ — 89%). Within each clutch size, the coefficient of variation of clutch volume was similar to that of egg volume (Table 2) . After pooling all data, mean egg volume was significantly and positively corre- lated with clutch size (r = 0.157, N = 313, P — 0.005, Fig. 2) . Thus, bigger clutches also had rela- tively larger eggs. Clutch size was also positively and significantly correlated with mean egg width (r — 1 ^ ^ ^ r 1.05 1.15 1.25 1.35 1.45 1.55 Egg shape (lengthAwidth) Figure 1. Distribution of egg shapes (length/width) in the Montagu’s Harrier. 0.152, N — 313, P = 0.006), but not with mean egg length (r = 0.110, N — 313, P = ns). Mean egg shape was not related to clutch size (r == —0.008, N= 313, P = ns). We also pooled data from all years and com- pared the mean values for the different variables among the study areas (Table 3) . A general linear model of clutch size in relation to site, year, and their interaction showed that clutch size varied sig- nificantly among years (F 3 501 = 13.7, P = 0.0001), and that there was a significant interaction be- tween year and site (Fgsei = 8.05, P = 0.0001). However, no significant differences among sites were found when year differences were taken into account (F 3 561 = 1.58, P = 0.2). Mean egg length, width, volume or shape did not vary significantly neither among sites nor among years (MANOVA, F 8 384 0.58, P — 0.79 for site effect; F 4 191 = 1.49, P = 0.21 for year effect, = 1-29, P = 0.25 for the interaction). Outch size Figure 2. Relationship between mean egg volume and clutch size in the Montagu’s Harrier. June 1998 Egg and Clutch Size in Montagu’s Harrier 139 Table 2. Summary statistics of clutch volume in the Montagu’s Harrier according to clutch size. N = sample size (number of clutches); SD = Standard Deviation; CV = Coefficient of variation, 100 X SD/ x. Clutch Size N Mean SD CV 1 2 22.9 2.01 8.8 2 11 44.9 3.41 7.6 3 81 68.1 6.23 9.1 4 130 92.6 8.65 9.4 5 71 117.7 10.37 8.9 6 16 141.1 12.54 9.2 7 1 169.9 — — 8 1 195.8 — — Discussion Mean egg size values found in this study were similar to those reported in other studies (Studin- ka 1941, Schonwetter 1967, Hays 1971, Perez Chis- cano and Fernandez Cruz 1971). Overall variability in egg size in the Montagu’s Harrier was relatively high compared to that found for other nonpasser- ine species (Flint and Sedinger 1992, Swennen and van der Meer 1992, Jover et al. 1993, Meathrel et al. 1993, Hakkarainen and Korpimaki 1994, Rob- ertson 1995, Weidinger 1996), but similar to that found in the American Kestrel {Falco sparverius) (Wiebe and Bortolotti 1995). Of the parameters contributing to egg volume, egg length was more variable than egg width, as seen from the coeffi- cient of variation and the positive skewness of the distribution of egg shapes. Females augmented the volume of their eggs by increasing the lengths rath- er than the widths of their eggs, as reflected by the fact that egg volume and shape were positively cor- related. This is opposite to what is reported for the American Kestrels {Falco sparverius), where the in- crease in egg volume was mainly the result of an increase in egg width (Wiebe and Bortolotti 1995). Similarly, in the Cape Petrel {Daption capense) the relationship between egg shape and volume was negative, with larger eggs being relatively more rounded (Weidinger 1996). As egg shape is deter- mined during its passage through the oviduct (Campbell and Lack 1985), the reasons for these interspecific d iff erences might be related to mor- phological, physiological or genetic differences be- tween species. Up to 59% of the variance in egg size was ex- plained by differences among clutches, proportion similar to that of other nonpasserine species (see Wiebe and Bortolotti 1995 for a summary), but a lower figure than that reported for American Kes- trels (Wiebe and Bortolotti 1995). Between clutch differences in egg size are related to differences between females, which are in turn linked to in- dividual or environmental conditions. We found no significant effect of site or year on mean egg volume, although such effects have been found in other avian species (Flint and Sedinger 1992, Crox- all et al. 1992, Robertson 1995, Weidinger 1996). Egg rank Figure 3. Mean (±SE) egg volume of first, intermediate and last eggs within clutches. Data from 92 clutches where rank order was determined. Volume of “Intermediate” eggs was averaged for each clutch. 140 Arroyo et al. VoL. 32, No. 2 Table 3. Geographical differences in clutch parameters. Mean, SD, and sample size (number of clutches) in brackets. Deux Sevres Rochefort Baie de l’Aiguillon Madrid Clutch size 4.28 ± 1.09 (143) 4.00 ± 1.04 (133) 3.75 ± 0.80 (101) 3.93 ± 0.84 (215) Mean egg volume (cm^) 23.2 ± 2.16 (121) 23.3 ± 2.01 (107) 22.9 ± 1.95 (56) 23.0 ± 2.26 (42) Mean egg length (mm) 41.6 ± 1.9 (121) 41.6 ± 1.6 (107) 41.1 ± 1.6 (56) 41.1 ± 1.6 (42) Mean egg width (mm) 33.0 ± 1.0 (121) 33.1 ± 1.1 (107) 33.0 ± 1.0 (56) 33.1 ± 1.2 (42) This might indicate that phenotypic plasticity in egg size according to environmental conditions is relatively limited in the Montagu’s Harrier, maybe due to a high heritability or genetic component in egg size. Alternatively, the relationship between egg size and environmental conditions might be masked by more evident changes in clutch size, or modulated by changes in within clutch egg-size pat- terns, as has been shown for the African Marsh- Harrier {Circus ranivorus) (Simmons 1994). Within clutch differences in the Montagu’s Harrier, which accounted for 41% of the variance in egg size, showed highly variable patterns, as was reflected by the large standard deviations in average egg sizes for each rank. Variability in clutch size was higher in the Mon- tagu’s Harrier than in other species (Flint and Sed- inger 1992, Wiebe and Bortolotti 1995), and much higher than that of egg size, indicating that clutch size was a trait with higher plasticity than egg size in this species. A higher variation in clutch rather than egg size according to food abundance was found in the American Kestrel (Wiebe and Borto- lotti 1995). The higher flexibility of clutch size to- ward different environmental conditions was also reflected in the strong annual variation of this pa- rameter. In contrast, no significant regional varia- tion was found after annual differences were taken into account. Average clutch size in this study was similar or in the upper range to that reported in Table 4. Clutch size of Montagu’s Harrier in different regions of its breeding range, x ± SD. N = sample size (number of nests). Regression of clutch size on latitude: Fj 17 = 1.91, P= 0.18, = 4.8%. When excluding one-year studies, Fj 14 = 0.50, P = 0.49, = 0%. Study Site Lati- tude Clutch Size N Range Length OF Study Reference The Netherlands 53.30 4.20 ± 0.80 112 2-7 30 Bijlsma et al. 1993 Belarus 53.00 3.42 12 1-5 1 Yasevitch and Vintchewski, 1995 Poland 53.00 3.90-4.40 125 2-5 several Krogulec and Leroux 1994 England 52.30 4.02 ± 0.07 227 2-10 30 Underhill-Day 1990 Morbihan, France 47.44 4.27 ± 0.14 29 3-6 14 Hays 1971 Milly, France 47.10 3.69 ± 0.19 26 2-5 >4 Cormier 1985 Noirmoutier, France 47.01 3.61 ± 0.17 23 2-5 >4 Cormier 1985 Rochefort, France 45.57 2.80 ± 0.40 - 4.05 ± 0.18 162 7 Butet and Leroux 1993 Rochefort, France 45.57 4.00 ± 1.04 133 1-6 6 This study Deux Sevres, France 46.11 4.28 ± 1.09 143 2-8 4 This study B. Aiguillon, France 46.24 3.75 ± 0.80 101 2-6 3 This study N. Apennines, Italy 43.28 3.32 31 5 Earalli 1994 Marche, Italy 43.40 3.80 37 2 Pandolh and Giacchini 1991 Emilia-Romagna, Italy 44.30 3.70 ± 0.80 61 2-5 7 Martelli and Sandri 1991 Madrid, Spain 40.38 3.93 ± 0.84 215 2-6 6 This study Ciudad Real, Spain 38.35 4.30 ± 1.00 120 2-9 7 Castano 1997 Madrigalejo, Spain 39.09 3.48 ± 0.99 98 1-5 5 Corbacho et al. 1997 Caceres, Spain 39.25 3.31 ± 0.65 19 2-5 1 P. Chiscano and E. Cruz 1971 Castro Verde, Portugal 37.42 2.82 ± 0.40 11 2-3 1 Onofre 1992 June 1998 Egg and Clutgh Size in Montagu’s Harrier 141 other areas (Table 4). It has been reported that clutch size in this species increases with latitude (Corbacho et al. 1997). However, our wider analy- ses show that the relationship is not significant, es- pecially if only the studies with data of at least two yr are considered given the potential annual vari- ation of this parameter. This indicates that the pat- tern is more complex than just latitudinal, and probably more related to the the relative food abundance between regions, or to the relative vari- ation in food abundance among years within each site. Life-history theory predicts a trade-off between the number and size of offspring produced (Stearns 1992). Thus, a negative relationship be- tween clutch and egg size is generally expected (Blackburn 1991), and has been found in some species (Potti 1993, Rowe 1994). In the case of the Montagu’s Harrier, this relationship was positive, which suggests that this particular trade-off does not exist in this species. Actually, no relationship between clutch and egg size has been found in some species (e.g., Duncan 1986, Jarvinen 1991, Hakkarainen and Korpimaki 1994), while others show, as the Montagu’s Harrier, a positive relation- ship (Flint and Sedinger 1992). In the American Kestrel, mean egg volume increases with increasing clutch size, except for the largest clutches where it declines (Wiebe and Bortolotti 1995). However, a trade-off will only be evident for a given amount of resources. The positive relationship between clutch and egg size might simply come as a result of pooling years and sites where the available re- sources are very different. Acknowledgments We are grateful to M. Salamolard, C. Pacteau, andJ.T. Garcia for their collaboration with data collection. This research benefited from funds made available to P. Dun- can and V. Bretagnolle by the Region Poitou Charente and the CNRS (Contrat de Plan Etat-Region 1994-1998). Literature Cited Arroyo, B.E. 1995. Breeding ecology and nest dispersion in the Montagu’s Harrier Circus pygargus in central Spain. Ph.D. dissertation, Oxford Univ., Oxford, U.K. Balfour, E. 1962. The nest and eggs of the Hen Harrier in Orkney. Bird Notes 30:69-73. Bijlsma, R.G., A.M. Blomert, W. van Manen and M. Quist. 1993. Ecologische Atlas van de Nederlandse Roofvogels. Schuyt, Harlem, The Netherlands. Blackburn, T.M. 1991. An interspecific relationship be- tween egg size and clutch size in birds. Auk 108:209- 211 . Butet, a. and A.B.A. Leroux. 1993. Effect of prey on a predator’s breeding success. A 7-year study on com- mon vole (Microtus arx>alis) and Montagu’s Harrier {Circus pygargus) in a west France marsh. Acta Oecol. 14:857-865. Campbell, B. and E. Lack. 1985. A dictionary of birds. T. & A.D. Poyser, Calton, U.K. Castano, J.P. 1997. Fenologfa de puesta y parametros re- productivos en una poblacion de Aguilucho cenizo ( Circus pygargus) en el Campo de Montiel. Ardeola 44. 51-60. Corbacho, C., J.M. SAnchez and A. Sanchez. 1997 Breeding biology of Montagu’s Harrier Circus pygargus L. in agricultural environments of southwest Spain; comparison with other populations in the western Pa- learctic. Bird Study 44:166-175. Cormier, J.P. 1985. La reproduction du busard cendre. Circus pygargus L. dans deux sites de I’ouest de la France, Oiseau Rev. Fr. Ornithol. 55:107-114. Cramp, S. and K.E.L. Simmons, [Eds.]. 1980. The birds of the western Palearctic, Vol. 2. Oxford Univ. Press, Oxford, U.K. Croxall, J.P., P. Rotheryand a. Crisp. 1992. The effect of maternal age and experience on egg-size and hatching success in Wandering Albatrosses Diomedea exulans. Ibis 134:219—228. Duncan, D.C. 1986. Variation and heritability in egg size of the northern pintail. Can. J. Zool. 65:992-996. Faralli, U. 1994. Breeding biology, habitat selection and conservation of Montagu’s Harrier Circus pygargus in the Northern Apennines, Italy. Pages 97-101 in B.-U. Meyburg and R.D. Chancellor [Eds.], Raptor conser- vation today. WWGBP/Pica Press, Berlin, Germany. Flint, P.L. and J.S. Sedinger. 1992. Reproductive impli- cations of egg-size variation in the Black Brant. Auk 109:896-903. Hakkarainen H. and E. KorpimAki. 1994. Environmental, parental and adaptive variation in egg size of Teng- malm’s owls under fluctuating food conditions. Oec- ologia 98:362-368. Hays, C. 1971. Essai sur la biologie de reproduction du busard cendre ( Circus pygargus) dans le Morbihan. Ar Vran 4:1-15. Hoyt, D.F. 1979. Practical methods of estimating volume and fresh weight of bird eggs. Auk 96:73-77. JArvinen, a. 1991. Proximate factors affecting egg vol- ume in subarctic hole-nesting passerines. Ornis Fenn. 68:99-104. Jover, L., X. Ruiz and M. Gonzalez-Martin. 1993. Sig- nificance of intraclutch egg size variation in the Pur- ple Heron. Ornis Scand. 24:127—134. Karpov, F.F. and E.Z. Berbaev. 1990. [Colonial nesting of the Montagu’s Harrier {Circus pygargus) in the low- er Hi River]. Ornitolo^ya 24:126-127. Krogulec, J. AND A.B.A. Leroux. 1993. Breeding ecology of Montagu’s Harrier Circus pygargus on natural and reclaimed marshes in Poland and France. In B.-U 142 Arroyo et al. VoL. 32, No. 2 Meyburg and R.D. Chancellor [Eds.], Raptor conser- vation today. Pica Press, Cornwall, U.K. Leroux, A.B.A. 1987. Recensement des busards nicheurs, Circus aeru^nosus L. et Circus pygargus L. et zonage de I’espace dans les marais de I’ouest de la France. Acta Oecol. Oecol. Appl. 8:387-402. Martelli, D. 1987. Datti sull’ecologia riprodutiva dell’albanella minore {Circus pygargus) in Emilia-Ro- magna. Nota preliminare. Suppl. Ric. Biol. Selvaggina 12:125-137. AND V. Sandri. 1991. Status ed ecologia deU’Albanella minore {Circus pygargus) in Emilia Ro- magna. Analisi conclusiva. Pages 49—52 in SROPU [Ed.], Atti V Conv. Ital. Omit. Ric. Biol. Selvag. XVII. Meathrel, C.E., J.S. Bradley, R.D. Wooller and I.J. Ski- RA. 1993. The effect of parental condition on egg-size and reproductive success in Short-tailed Shearwaters Puffinus tenuirostris. Oecologia 93:162-164. Onofre, N. 1992. Bio-ecologia da aguia ca^adeira Circus pygargus (L.) numa area de agricultura cerealifera em Castro Verde. Dados preliminares. I Conferencia Nacional sobre Aves de Rapina, Gaia, Portugal. Pandolfi, M. and P. Giacchini. 1991. Distribuzione e suc- cesso riproduttivo di albanella minore Circus pygargus nell Marche. Riv. Ital. Ornitol. 61:25-32. Perez Chiscano, J.L. and M. Fernandez Cruz. 1971. So- bre Grus grusy Circus pygargus en Extremadura. Ardeola Vol. esp:509-574. POTTI, J. 1993. Environmental, ontogenic and genetic variation in egg size of Pied Flycatchers. Can. f. Zool. 71:1534-1542. Robertson, G.J. 1995. Annual variation in Common Ei- der egg size: effects of temperature, clutch size, laying date and laying sequence. Can. J. Zool. 73:1579-1587. Rowe, J.W. 1994. Reproductive variation and egg size- clutch size trade-off within and among populations of painted turtles ( Chrysemys picta bellii) . Oecologia 99:35- 44. SCHONWETTER, M. 1967. Handbuch der Oologie. Liefu- rung 14:3-64. Akad. Verlag, Berlin, Germany. Simmons, R.E. 1994. Supplemental food alters egg size hierarchies within harrier clutches. Oikos 71:341-348. Stearns, S.C. 1992. The evolution of life histories. Ox- ford Univ. Press, Oxford, U.K. Studinka, L. 1941. The habits and plumages of Monta- gu’s Harrier. Aquila 46/49:247-268. Swennen, C. and j. van der Meer. 1992. Variation in egg size of Common Eiders. Ardea 80:363-373. Underhill-Day, J.C. 1990. The status and breeding bi- ology of Marsh Harrier and Montagu’s Harrier in Brit- ain since 1900. Ph.D. dissertation, CNAA, London, U.K. Weidinger, K. 1996. Egg variability and hatching success in the Cape Petrel Daption capense at Nelson Island, South Shetland Islands, Antarctica. J. Zool. Lond. 239: 755-768. Wiebe, K.L. AND G.R. Bortolotti. 1995. Egg size and clutch size in the reproductive investment of Ameri- can kestrels, y. Zool. Lond. 237:285—301. Received 17 May 1997; accepted 6 February 1998 J. Raptor Res. 32(2) ; 143-1 50 © 1998 The Raptor Research Foundation, Inc. ORGANOCHLORINE PESTICIDES, PCBs AND MERCURY IN HAWK, FALCON, EAGLE AND OWL EGGS FROM THE LIPETSK, VORONEZH, NOVGOROD AND SARATOV REGIONS, RUSSIA, 1992-1993 CJ. Henny U.S. Geological Survey, Forest and Rangeland Ecosystem Science Center, 3080 SE Clearwater Drive, Corvallis, OR 97333 U.S. A. V.M. Galushin, RI. Dudin, A.V. Khrustov, A.L. Mischenko, V.N. Moseikin, V.S. Sarychev and V.G. Turchin Russian Bird Conservation Union, Kibalchicha 6, Moscow 129278 Russia Abstract. — Fifty-two eggs (one per nest) of 12 species of raptors were collected in 1992—93 for contam- inant analysis in three southern European locations in Russia. One Peregrine Falcon (Falco peregrinus) egg was also collected farther northwest in the Novgorod region. A high DDE concentration (2*7. 3 ppm, wet weight [w/w] ) in the Peregrine Falcon egg raised concern for the species in European Russia south of the Arctic Circle. Although a number of organochlorine contaminants were found in eggs of the other species, concentrations were all below known effect levels. Mercury levels were also extremely low. Nesting success in southern Russia in 1992 (only year with follow-up visits) appeared normal. Key Words: Peregrine Falcon-, Falco peregrinus; eagles-, hawks-, falcons-, owls; organochlorine pesticides; PCBs, mercury; Russia. Pesticidas organoclorados, PCBs y mercurio en huevos de gavilanes, halcones, aguilas y buhos de las regiones de Lipetsk, Voronezh, Novgorod y Saratova en Rusia, 1992-93 Resumen. — Cincuenta y dos huevos (uno por nido) de 12 especies de aves rapaces fueron colectados en 1992-93 para el analisis de contaminantes en tres localidades al sur de Europa en Rusia. Un huevo de Halcon Peregrine {Falco peregrinus) fue colectado al noroeste en la region de Novgorod. Una alta concentracion de DDE (27.3 ppm), en este huevo desperto preocupacion para la especie en la Rusia europea al sur del Circulo Artico. Aunque un buen numero de contaminantes organoclorados fueron encontrados en los huevos de otras especies, las concentraciones estuvieron por debajo de niveles pe- ligrosos. Los niveles de Mercurio fueron tambien extremadamante b^os. El exito reproductive en 1992 (unico ano en el que se efectuo seguimiento) aparentemente fue normal. [Traduccion de Cesar Marquez] Concerns over possible pesticide contamination of raptors in Russia were first raised in the 1970s (Galushin 1977), an alarm first sounded by the rapid and continuous decline of the Peregrine Fal- con {Falco peregrinus ) . To assess the extent of chem- ical contamination in Russian birds of prey, a series of egg collections were made. Eggs of the Pere- grine Falcon were specifically collected above the Arctic Circle on the Kola Peninsula in 1991 (Hen- ny et al. 1994). Osprey {Pandion haliaetus) eggs were collected along the Upper Volga River in 1992. This paper is the third in a series that pre- sents the results of contaminant analyses conduct- ed on eggs of species collected in southern Euro- pean Russia in 1992 and 1993. Study Area and Methods In 1992, 13 eggs (one egg per clutch) were collected from the Lipetsk Region which includes the Upper Don River. This study area was located 30 km east of Elets Town and 40 km west of Lipetsk City near the Galichya Gora Nature Reserve. The Liptsk Region is mostly agri- cultural (fall and spring planted wheat, small grains, peas, and row crops) with some natural forests. Another 13 eggs were collected in the Voronezh Region about 300 km east and south of the Lipetsk Region. It consists of agricultural lands separated by planted forest strips. This study area was 10 km south of Talovaya or 120 km south- 143 144 Henny et al. VoL. 32, No. 2 east of Voronezh City. Raptors nested in planted forest strips. Tree planting began 100 yr ago and has continued for the purpose of stopping wind erosion of top soils. Crops were mainly small grains, corn, peas, sunflowers, and other row crops with some fields irrigated. One egg was an addled Peregrine Falcon egg taken in the more northwestern Novgorod Region. There were no nestlings m this nest but a fledgling was observed nearby. In 1993, another 26 eggs were collected southeast of Saratov at the middle Volga River located south and east of the 1992 study areas. Habitat included forest-steppe and steppe. Nests where eggs were collected in 1992 were subsequently observed by local biologists to determine nest productivity. Nests where eggs were collected in the Saratov Region in 1993 could not be monitored. Analytical and Statistical Methods The collected eggs were opened in Russia and their contents were placed in chemically cleaned jars contain- ing Na 2 S 04 for preservation. Egg contents were analyzed at the Geochemical and Environmental Research Group, Texas A&M University, College Station, Texas. Egg sam- ples for organics were extracted by the NOAA Status and Trends Method (MacLeod et al. 1985) with minor revi- sions (Brooks et al. 1989; Wade et al. 1988). Briefly, the egg samples were homogenized with a Teckmar Tissum- izer. A 1—10 g sample w/w was extracted with the Teck- mar Tissumizer by adding surrogate standards, Na 2 S 04 and methylene chloride in a centrifuge tube. The tissue extracts were purified by silica/alumina column chro- matography to isolate the aliphatic and PAH/pesticide/ PCB fractions. The pesticide/PCB fraction was further purified by HPLC in order to remove interfering lipids. Eggs were analyzed for alpha-hexachlorocyclohexane (A- BHC), |3-BHC, A-BHC, lindane, hexachlorobenzene (HCB), heptachlor, heptachlor epoxide (HE), oxychlor- dane, gamma chlordane, alpha chlordane, tram-nona- chlor, raVnonachlor, aldrin, dieldrin, endrin, mirex, p,p - DDE (DDE), p,p'-DDT (DDT), p,p'-DDD (DDD), o,p'-DDE, o,p'-DDT, o,p'-DDD, and total PCBs. The quantitative analyses were performed by capillary gas chromatogra- phy (GC) with electron capture detection for pesticides and PCBs. Total PCBs were quantified by summing the concentrations of homolog groups which were esti- mated by degree of chlorination (Wade et al. 1988). The pesticides and PCBs are initially analyzed on a DB- 5 capillary column. The analyte identity and concen- trations are confirmed on a DB-17 capillary column. We also analyzed for 87 PCB congeners in 1993, but the concentrations were so low in almost all species that congener data were only discussed for the Pere- grine Falcon egg analyzed (species with the highest PCB concentrations). A comparison of estimated total PCBs vs. sum of PCB congeners showed excellent agreement between the two procedures for individual eggs with the highest PCB concentrations. Mercnry was determined by EPA method 245.5 with minor revisions. Tissue samples were homogenized in the original sample containers with a tissumizer and subsam- pled. Mercury was determined by a modification of the method of Hatch and Ott (1968). A portion of the digest solution was placed in a sealed container. To this was added 0.4 ml of 10% (w/w) stannous chloride. Mercury was reduced to the elemental state and aerated from so- lution into an atomic absorption spectrophotometer where its concentration was measured. Residue concentrations in eggs were corrected to an approximate fresh w/w using egg volumes (Shekel et al. 1973); all organochlorines and mercury in eggs were ex- pressed on a fresh w/w basis. The detection limit was considerably below 0.01 ppm w/w, but from a biological perspective, values were not presented unless >0.01 ppm. The General Linear Models Procedures (SAS 1996) was used to perform a one-way analysis of variance. Res- idue data were log^o transformed for all analyses and means were presented as geometric means, and « = 0.05 for statistical significance. Results In general, raptors common to the southern portion of European Russia are found in wood- lands, forest-strips, forest-steppe, isolated groves, semi-open and open steppes (Table 1). Segments of the steppe are now farmed, and many of the buzzards, kestrels, falcons, harriers, goshawks, kites, and owls nest in close association with agri- culture. We considered all species to be migratory in at least portions of their range and, therefore, could accumulate contaminants on wintering grounds which sometimes are great distances from nesting areas. Only eggs collected in 1992 were an- alyzed for mercury so the mercury results were pre- sented separately. Buzzards. DDE, P-BHC and PCBs were found in nearly every buzzard egg (Tables 2 and 3). Com- mon Buzzards {Buteo buteo) from both agricultural areas contained low and not statistically different DDE concentrations (geometric mean [jc] = 0.15 and 0.20 ppm). Likewise, concentrations of DDE in Long-legged Buzzard {Buteo rufinus) eggs (0.13 ppm) from the native steppe were not significantly different from those of the Common Buzzard. (3-BHC (0.01 and 0.02 ppm) and PCBs (0.06 and 0.22 ppm) were low in Common Buzzard eggs from both agricultural sites. When (3-BHC and PCBs in Common Buzzard eggs were combined for the two sites (p-BHC — 0.02 and PCBs = 0.13 ppm) and compared to Long-legged Buzzards, (B- BHC was significantly higher in eggs of the Long- legged Buzzard (P = 0.01), but PCBs were not sig- nificantly different. The highest PCB concentra- tion was 2.20 ppm in a Common Buzzard egg from the Lipetsk Region. Harriers. Five Montagu’s Harrier {Circus pygar- gus) eggs from Voronezh contained 0.45 ppm DDE which was not significantly different from the 0.18 June 1998 Toxic Chemicals in Russian Raptors 145 Table 1. General characteristics of birds of prey studied in European Russia, 1992-93.^ Species Habitat Migratory Winter Areas Long-legged Buzzard Steppe, semidesert Yes E. Africa, S. Asia Common Buzzard Mixed Yes Africa, S. Europe Imperial Eagle Woodland, forest-steppe Yes N.E. Africa, S.W. Asia Steppe Eagle Steppe, semiopen Yes India, S. Africa Booted Eagle Thinly wooded Yes Africa, S. Asia Common Kestrel Groves, semiopen Yes Africa, India, S. Europe Red-footed Falcon Groves, mixed Yes S. Africa Peregrine Falcon*^ Woodland, semiopen Partly S. and C. Europe Montagu’s Harrier Open Yes Africa, S. Asia Black Kite Floodplains, water Yes Africa, S. Asia Northern Goshawk Forest, woodland Partly S. and C. Europe Long-eared Owl Woodland, thickets Partly S. and C. Europe ® Information obtained from Amadon et al. (1988), Cramp (1980), Dementev et al. (1966), Ilychev (1982). Those nesting below 60°N latitude. ppm DDE found in Common Buzzard eggs from Voronezh and Lipetsk. ^-BHC {x — 0.02 ppm) and PCBs (jc = 0.05 ppm) were found in low concen- trations in all Montagu’s Harrier eggs from Voro- nezh and, when compared to buzzard eggs from Voronezh and Lipetsk combined (P-BHC 0.02 and PCBs 0.13 ppm), there was no significant differ- ence between the two species. As with the buzzards, DDD and DDT were sometimes detected at low concentrations in the harrier eggs. Dieldrin and oxychlordane were regularly found in the harrier eggs but at low concentrations and low concentra- tions of heptachlor epoxide were reported in four of five harrier eggs but not in the buzzard eggs. Falcons. The Common Kestrel {Falco tinnuncu- lus) was found in all three study areas so it gave the first opportunity to make a broad comparison of residue concentrations in one species. We found no statistically significant differences in the low DDE concentrations between Voronezh (0.03 ppm), Lipetsk (0.06 ppm), or Saratov (0.03 ppm). DDD and DDT were not detected in any kestrel eggs. PCB concentrations in kestrel eggs at Voro- nezh (0.02 ppm) and Lipetsk (0.05 ppm) were not significandy different, but when the two areas were combined and compared to kestrels in Saratov (0.02 ppm), the small difference was statistically significant (P = 0.01), HCB appeared in low con- centrations in one egg from Lipetsk and two eggs from Saratov. The two Red-footed Falcon {Falco ves- pertinus) eggs from Saratov contained DDE, P- BHC, and PCB concentrations that were nearly identical to the kestrels from the same region. Common Buzzard eggs contained significantly higher concentrations of DDE (0.18 ppm) than kestrel eggs (0.05 ppm) (P = 0.01), significantly higher concentrations of PCBs (0.13 vs. 0.04 ppm) than kestrel eggs (P < 0.05), but P-BHC (0.02 vs. 0.01 ppm) was not significantly different. A com- parison of Montagu’s Harrier eggs from Voronezh with kestrel eggs from Voronezh and Lipetsk com- bined showed DDE was significantly higher in the harrier eggs (0.45 vs. 0.05 ppm) (P< 0.002), while P-BHC and PCBs were not significantly different. The Peregrine Falcon egg from the Novgorod Region contained much higher concentrations of nearly all contaminants (Table 3). Despite the fact that 27.3 ppm DDE was the highest concentration of a contaminant found in any egg collected, one young fledged from the nest. On 15 July 1993 near the same Peregrine Falcon nest, two fledglings were observed with two adults. In 1994, the nest was not visited but, on 30 May 1995, three nestlings and a pipping egg were observed in the nest. This nest site was the only one found in a bog on the ground within a forest zone of European Russia. PCBs and HCB levels were also higher than in any other eggs and DDD, DDT, dieldrin, HE, lindane mirex, ^rawi-nonachlor and oxychlordane were also present. The eggshell thickness was 0.316 mm. The Peregrine Falcon egg was analyzed for 87 PCB congeners of which almost half were reported at concentrations above 0.01 ppm. PCB congeners (lUPAC Numbers) including PCB 138 (2.44 ppm), PCB 153 (4.93 ppm) and PCB 180 (1.59 ppm) were dominant and accounted for 61% of the 14.6 ppm. Other PCB congeners with >0.20 ppm were PCB 74 (0.23 ppm), PCB 194 (0.24 ppm), PCB 196 146 Henny et al. VoL. 52, No. 2 (0.25 ppm), PCB 128 (0.26 ppm), PCB 170 (0.29 ppm), PCB 172 (0.36 ppm), PCB 183 (0.43 ppm), PCB 146 (0.54 ppm), PCB 156/171/202 (0.54 ppm), PCB 187/182/159 (0.66 ppm), and PCB 118/108/149 (0.96 ppm). This list of congeners accounted for 13.72 ppm, or 94% of the PCBs in the Peregrine Falcon egg. The estimated total PCB concentrations, using a different methodology on the same egg, was nearly identical (14.3 vs. 14.6 ppm) (Table 3). Eagles. Seven eagle eggs were collected and the Booted Eagle (Hieraaetus pennatus) egg contained the highest DDE concentration (1.06 ppm) with some DDD and DDT; the Imperial Eagle (Aquila heliaca) egg was intermediate with 0.33 ppm DDE and some DDD and DDT; and the five Steppe Ea- gle (Aquila rapax) eggs were lowest (0.05 ppm) with no DDD or DDT. The list of other contami- nants found in the eagle eggs was fairly long, in- cluding a-BHC and A-BHC in the Steppe Eagle. Steppe Eagle and Common Kestrel eggs from Sa- ratov were compared. DDE concentrations were not significantly different, but Steppe Eagles con- tained significantly higher concentrations of both P-BHC (0.28 vs. 0.02 ppm, P = 0.0001) and PCBs (0.03 vs. 0.02 ppm, P < 0.001). Goshawks and Kites. Two Northern Goshawk (Acdpiter gentilis) eggs from Voronezh and Lipetsk contained nearly identical DDE residue concentra- tions (1.14 and 1.37 ppm) which were higher than reported for all eggs collected except for one Mon- tagu’s Harrier (1.48 ppm) and the Peregrine Fal- con (27.3 ppm) eggs. DDD, DDT, |3-BHC and PCBs were also found in the goshawk eggs. Black Kite (Milvus migrans) eggs from Lipetsk and Saratov contained very similar concentrations of DDE (0.38 to 0.54 ppm), in addition to DDD and DDT, and low concentrations of several others. Long-eared Owl. The Long-eared Owl (Asio otus) egg from Saratov contained extremely low concen- trations of DDE, PCBs and HCB. Mercury Residues in Eggs. No eggs contained mercury concentrations above 0.05 ppm, which is an extremely low concentration (Table 2). All Common Buzzard eggs contained mercury, but showed no significant difference between Voro- nezh and Lipetsk. Only one of eight kestrel and none of the Northern Goshawk nor Montagu’s Harrier eggs had detectable mercury concentra- tions from either Voronezh or Lipetsk. The Black Kite and Booted Eagle eggs also contained low Table 2. Organochlorine pesticides, PCBs, and mercury (ppm, wet weight) in birds of prey eggs (number of eggs in parentheses) from Voronezh and Lipetsk Regions, Russia, 1992. Species Location (N)- Clutch Size (N) b Young Fledged (N)- Common Buzzard Voronezh (4) 3.50(2) 2.33(3) Range 3-4 2-3 Lipetsk (5) 3.25(4) 1.75(4) Range 2-4 1-3 Northern Goshawk Voronezh (1)^ 4.00(1) 2.00(1) Lipetsk (1) 3.00(1) 1.00(1) Common Kestrel Voronezh (3) 5.33(3) 3.00(3) Range 5-6 0-5 Lipetsk (5)^ 5.80(5) 3.00(5) Range 4-7 0-5 Montagu’s Harrier Voronezh (5)s 4.75(4) 2.40(5) Range 4-5 0-4 Black Kite Lipetsk (1) 3.00(1) 2.00(1) Booted Eagle Lipetsk (1)8 2.00(1) 1.00(1) “ Number of eggs analyzed for contaminants. ^ Mean includes egg collected for residue analysis, but excludes nests with fresh eggs (possibly incomplete clutches). Young fledged value excluded from mean calculation if lone egg collected from nest. Note that 1 egg was collected from each nest. Geometric mean used for residue concentrations (0.005 ppm assigned to values <0.01 ppm) since no value presented <0.01; e.g., = none detection (ND). No mean calculated if ^50% of samples <0,01 ppm. HG = mercury, DDE = p,p'-DDE, DDD = p,p'-DDD, DDT = p,p'-DDT, P-BHC = beta-hexachlorocyclohexane, PCBs = poly- chlorinated biphenyls, OXY = oxychlordane, HE = heptachlor epoxide. ^ Goshawk from Voronezh and 1 kestrel from Lipetsk each con- tained 0.01 ppm hexachlorobenzene, K Two Montagu’s Harriers and the Booted Eagle each contained 0.01 ppm trons-nonachlor. concentrations of mercury. The Saratov eggs were not analyzed for mercury. Discussion and Conclusions The Peregrine Falcon egg collected in 1992 from the Novgorod Region (south of St. Peters- June 1998 Toxic Chemicals in Russian Raptors 147 Table 2. Extended. Geometric Mean (ppm)'^ HG^ DDE^ DDD*= DDT^ PCBs^ OXY^ Dieldrin HE<= 0.03 0.15 0.01 0.06 0.02-0.05 0.03-0.30 ND-0.02 ND 0.01-0.05 0.04-0.10 ND ND-0.10 ND 0.04 0.20 — — 0.02 0.22 — — — 0.03-0.05 0.06-0.59 ND-0.04 ND-0.01 ND-0.09 0.07-2.70 ND ND-0.01 ND ND 1.14 0.07 0.16 0.08 0.26 ND ND ND ND 1.37 0.06 0.01 0.04 0.41 ND ND ND 0.03 0.01 0.02 _ ND 0.01-0.11 ND ND 0.01 0.02-0.03 ND ND ND — 0.06 — — 0.01 0.05 — — — ND-0.01 0.03-0.21 ND ND 0.01-0.02 0.01-0.16 ND ND ND _ 0.45 _ _ 0.02 0.05 0.01 0.02 0.02 ND 0.17-1.48 ND-0.01 ND-0.04 0.01-0.07 0.02-0.12 ND-0.02 ND-0.05 ND-0.10 0.03 0.45 0.05 0.01 0.05 0.20 ND 0.03 ND 0.02 1.06 0.03 0.01 0.04 0.31 0.02 0.01 0.02 burg) with 27.3 ppm DDE causes us the most con- cern. This bog-associated site was successful in 1992, 1993, and 1995 but we are not sure if the same female occupied the site each year. The DDE concentration in the egg was much higher than the mean of 3.5 ppm w/w reported for the increas- ing Peregrine Falcon population nesting along the Ponoy Depression (above 66°N) of the Kola Pen- insula, Russia in 1991 (Henny et al. 1994). The fact that young were produced at this nest in 1992 gives reason for some optimism. Peakall et al. (1975) tentatively concluded that 15-20 ppm DDE w/w was associated with the inability of Peregrine Fal- cons to maintain population numbers. Ambrose et al. (1988) believed that 15-20 ppm DDE should not be construed so rigidly as to predict the success or failure of individual eggs, and believed that a gradual reduction in productivity occurs above a DDE level that is not precisely known. In fact, they reported young were fledged from five of six nests in Alaska where randomly-sampled fresh eggs con- tained over 15 ppm DDE. Even though the critical concentration for DDE is not precisely known, 27.3 ppm DDE found in the egg from this study pro- vides a serious reason for continued concern for Peregrine Falcons in the more southern latitudes of European Russia. The shell thickness of the egg collected was 0.316 mm, which represents only 10% shell thinning compared to 0.350 mm for pre- DDT era Swedish Peregrine Falcons (Lindberg 1975) . Pesticides were never used during breeding seasons within the bog-associated hunting area of that Peregrine Falcon pair we studied. Few pere- grines exist in European Russia south of the Arctic Circle, and their specific migration routes and win- tering areas remain unknown. DDE in the eggs of the other species were much lower with values seldom above 1 ppm. Since a dif- ference exists in sensitivity among species to DDE, it is difficult to reach conclusions about its poten- tial effects on the various species we studied. How- ever, two of the most sensitive species to DDE are 148 Henny et al. VoL. 32, No. 2 Table 3, Organochlorine pesticides and PCBs (ppm, wet weight) in birds of prey eggs from the Sarotov Region, 1993 and one egg from Novgorod Region, Rnssia, 1992. Clutch Geometric Mean (ppm)*^ Species { N )^ Size { N )^ DDE^ DDDd DDT^ 3-BHC^ PCBs<^ Common Kestrel (12) 4.80(5) 0.03 — — 0.02 0.02 Range 3-6 0.01-0.07 ND ND 0.01-0.06 0.01-0.03 Steppe Eagle (5) 2.80(5) 0.05 — — 0.28 0.03 Range 2-3 0.03-0.11 ND ND 0.13-0.52 0.03-0.04 Red-footed Falcon (2) — 0.03 — — 0.01 0.02 Range — 0.03-0.04 ND ND 0.01 0.02 Long-legged Buzzard (3) 2.50(2) 0.13 — — 0.11 0.14 Range 2-3 0.07-0.26 ND ND 0.07-0.26 0.08-0.36 Black Kite (2) 2.00(2) 0.45 0.06 0.05 0.33 0.09 Range 2 0.38-0.54 0.06 0.03-0.08 0.27-0.41 0.08-0.10 Imperial Eagle (1) 2.00(1) 0.33 0.25 0.05 0.35 0.10 Long-eared Owl (1) 4.00(1) 0.07 ND ND ND 0.04 Peregrine Falcon Novgorod (1)*^ 27.3 0.28 0.09 0.17 14.3 ® Number of eggs analyzed for contaminants. Mean includes egg collected for residue analysis, but excludes nests with fresh eggs (possibly incomplete clutches) . Geometric mean used for residue concentrations (0.005 ppm assigned to values <0.01 ppm) since no value presented <0.01 ppm; 1 e., = nondetected (ND). No mean calculated if ^50% of samples <0.01 ppm. ^ DDE = p,p'-DDE; DDD = p,p'-DDD; DDT = p,p'-DDT; (J-BHC, A-BHC, A-BHC=beta, alpha and delta hexachlorocyclohexane; PCBs = polychlorinated biphenyls; HCB = hexachlorobenzene; HE = heptachlor epoxide; OXY = oxychlordane. Also contained mirex 0.01 ppm, o,p'-DDT 0.02 ppm, and trans-nonachlor 0.01 ppm. the Brown Pelican {Pelecanus occidentalis) (nearly total reproductive failure at 3.0 ppm w/w in eggs), and the White-faced Ibis (Plegadis chihi) (reduced reproductive success above 4 ppm w/w in eggs) (Blus 1982, Henny and Herron 1989). With the highest DDE egg concentrations <2 ppm, it is doubtful that DDE had an adverse effect on any of the other species. Furthermore, the number of young produced in 1992 appeared normal (e.g., six of eight nests [75%] with fresh or slightly in- cubated eggs at time of collection successful; 15 of 16 nests [94% ] with medium-heavy incubation suc- cessful with fledging young) . Effects of specific concentrations of other organ- ochlorines are not well understood, and again sen- sitivity undoubtably differs among species. HE above 1.5 ppm w/w in American Kestrel (Falco sparverius) eggs resulted in reduced production (Henny et al. 1983) , but Canada Geese (Branta can- adensis) showed no sign of reduced hatchability un- til egg residues were above 10 ppm w/w (Blus et al. 1984). Henny et al. (1983) found no negative impact on productivity of American Kestrels when dieldrin in eggs ranged from 2. 2-3. 9 ppm w/w. None of the HE or dieldrin concentrations report- ed during this study approached the above levels, therefore, it is extremely doubtful that HE or diel- drin were adversely impacting the species studied. Concentrations of all the other organochlorine pesticides were extremely low. PCBs were seldom reported above 0.50 ppm and the buzzard nest with the highest PCB egg concentration (2.7 ppm) had four eggs with one egg collected, and fledged the maximum of three young. The only other high concentration of PCBs was in the Peregrine Falcon egg- Mercury concentrations in eggs were <0.05 ppm, which is lower than the geometric mean of 0.09 ppm w/w for Ospreys on the Upper Volga Riv- er in 1992. Mercury concentrations in eggs below 0.5— 0.6 ppm w/w do not adversely affect raptor re- production (Hakkinen and Hasanen 1980, Wie- meyer et al. 1984, Newton and Haas 1988). A-BHC, A-BHC, lindane, HCB, DDT and its me- tabolites, the chlordanes, and PCBs which were found in many of the eggs were also found in the blubber of a resident seal {Phoca caspica) from the coast of Iran on the Caspian Sea (Vetter et al. 1995). The Caspian Sea has no connection to the oceans and is south of the Saratov study area, and bordered by Russia, Kazakhstan, Turkmenistan, Iran and Azerbaijan. The resident seal which can- June 1998 Toxic Chemicals in Russian Raptors 149 Table 3. Extended. Geometric Mean (ppm) c A-BHC^ A-BHC^ Dieldrin Lindane HCBd HE'i OXYi ND ND ND ND ND-0.09 ND ND 0.01 0.01 0.01 — — — — ND-0.02 ND-0.01 ND-0.03 ND ND-0.01 ND ND ND ND ND ND ND 0.01 0.01 0.01 0.01 ND ND ND ND ND-0.06 ND ND ND-0.01 ND-0.08 ND-0.02 ND-0.01 ND-0.03 ND-0.09 ND-0.02 0.04 0.03 ND 0.01 0.01 0.01 0.01 ND ND ND ND 0.08 ND ND ND ND 0.18 0.01 1.16 0.12 0.12 not migrate, therefore, provides some evidence that many of the contaminants found in the raptor eggs could be accumulated in the region where they nested. Species sharing breeding localities sometimes contained different contaminants, which may be a function of different wintering lo- calities or food habits. Our data are too fragmen- tary to discuss contaminant residue patterns in eggs with respect to the species migratory charac- teristics or wintering ranges and food habits. Acknowledgments We are grateful to Vladimir Flint and Alexander Sor- okin (Russian Institute for Nature Conservation) and Steven Kohl (Office of International Affairs, U.S. Fish and Wildlife Service) for supporting this project. This study was conducted under Area V of the U.S.-Russia En- vironmental Agreement. We also appreciate the assis- tance of Professor V.V. Gusev (Voronezh University) and N.J. Skolznev (Director, Galichya Gora Nature Reserve) . L.J. Blus improved the manuscript with his review and R.A. Grove assisted with the statistical analysis. Literature Cited Amadon, D., J. Bull, J.T. Marshall and B.F. King. 1988. Hawks and owls of the world; a distributional and tax- onomic list. Proc. West. Found. Vertebr. Zool. 3:295-357. Ambrose, R.E., C.J. Henny, R.E. Hunter and J.A. Craw- ford. 1988. Organochlorines in Alaskan Peregrine Falcon eggs and their current impact on productivity. Pages 385-393 in T.J. Cade, J.H. Enderson, C.G. The- lander and C.M. White [Eds.], Peregrine Falcon pop- ulations: their management and recovery. The Pere- grine Fund, Inc., Boise, ID U.S.A. Blus, L.J. 1982. Further interpretation of the relation of organochlorine residues in Brown Pelican eggs to re- productive success. Environ. Pollut. 28:15-33. , CJ. Henny, D.J. Lenhart and T.E. Kaiser. 1984. Effects of heptachlor- and lindane-treated seed on Canada Geese./. Wildl. Manage. 48:1097-1111. Brooks, J.M., T.L Wade, E.L. Atlas, M.C. Kennicutt II, B.J. Presley, R.R. Fay, E.N. Powell and G. Wolff. 1989. Analysis of bivalves and sediments for organic chemicals and trace elements. Third Annual Report for NOAA’s National Status and Trends Program. Contract 50-DGNC-5-00262. Cramp, S. [Ed.]. 1980. Handbook of the birds of Europe, the Middle East and North Africa: the birds of the western Palearctic. Oxford Univ. Press, Oxford, U.K. Dementev, G.P., N.A. Gladkov, E.S. Ptushenko, E.P. Spangenberg and A.M. Sudilovskaya. 1966. Birds of the Soviet Union, Vol. 1. Israel Program Scientific Translations, Jerusalem, Israel. (Translated from Rus- sian) . Galushin, V.M. 1977. Recent changes in the actual and legislative status of birds of prey in the U.S.S.R. Pages 152-159 in Proc. World Conference on Birds of Prey, Vienna 1975. ICBP, London, U.K. Hakkinen, I. AND E. Hasanen. 1980. Mercury in eggs and nestlings of the Osprey {Pandion haliaetus) in Finland and its bioaccumulation from fish. Ann. Zool. Fenn. 17: 131-139. Hatch, W.R. and W.L. Ott. 1968. Determination of sub- microgram quantities of mercury in solution by a flameless atomic absorption technique. Anal. Chem. 40:2085-2087. Henny, C.J. and G.B. Herron. 1989. DDE, selenium. 150 Henny et al. VoL. 32, No. 2 mercury and White-faced Ibis reproduction at Carson Lake, Nevada./. Wildl. Manage. 53:1032-1045. , L.J. Blus and C.J. Stafford. 1983. Effects of hep- tachlor on American Kestrels in the Columbia Basin, Oregon./ Wildl. Manage. 47:1080-1087. , S.A. Ganusevich, F.P. Ward and T.R. Schwartz. 1994. Organochlorine pesticides, chlorinated dioxins and furans, and PCBs in Peregrine Falcon Falco pere- grinus eggs from the Kola Peninsula, Russia. Pages 739-749 in B.-U Meyburg and R.D. Chancellor [Eds.], Raptor Conservation Today. Pica Press, London, U.K. Ilychev, V. D. 1982. Migrations of birds of eastern Eu- rope and northern Asia: Falconiformes— Gruiformes. Nanka Publishing House, Moscow, Russia. Lindberg, P. 1975. Pilgrimsfalken i Sverige. Svenska Na- turskyddsforeningen, Stockholm, Sweden. MacLeod, W.D., D.W. Brown, A.J. Friedman, D.G. Bur- row, O. Mayes, R.W. Pearce, G.A. Wigren and R.G. Bogar. 1985, Standard analytical procedures of the NOAA National Analytical Facility 1985-1986. Extract- able Toxic Organic Compounds. 2nd Ed. U.S. De- partment of Commerce, NOAA/NMFS, NOAA Tech. Memo. NMFS/NWRC-92. Newton, I. and M.B. Haas. 1988. Pollutants and Merlin eggs and their effects on breeding. Br. Birds 81:258- 269. Peakall, D.B., T.J. Cade, C.M. White and J.R. Haugh. 1975. Organochlorine residues in Alaskan peregrines. Pestic. Monit. /. 8:255-260. SAS Institute, Inc. 1996. SAS user’s guide: statistics. Ver- sion 6.12. Ed. SAS Institute, Inc., Cary, NC U.S.A. Stickel, L.F., S.N. Wiemeyer and L.J. Blus. 1973. Pesti- cide residues in eggs of wild birds: adjustment for loss of moisture and lipid. Bull. Environ. Contain. Toxicol. 9‘ 193-196. Vetter, W., C. Natzeck, B. Luckas, G. Heidemann, B. KlABi and M. Karami. 1995. Chlorinated hydrocar- bons in the blubber of a seal {Phoca caspica) from the Caspian Sea. 30:1685-1696. Wade, T.L., E.L. Atlas, J.M. Brooks, M.C, Kennicutt II, R.G. Fox, J. Sericano, B. Garcia and D. Defreitas. 1988. NOAA Gulf of Mexico status and trends pro- gram: trace organic contaminant distribution in sed- iments and oyster. Estuaries II, 171-179. Wiemeyer, S.N., T.G. Lamont, C.M. Bunck, C.R. Sinde- LAR, FJ. Gramlich, J.D. Fraser and M.A. Byrd. 1984. Organochlorine pesticide, polychlorobiphenyl, and mercury residues in Bald Eagle eggs — 1969-79 — and their relationships to shell thinning and reproduc- tion. Arch. Environ. Contain. Toxicol. 13:529-549. Received 9 June 1997; accepted 10 February 1998 J. Raptor Res. 32(2):151-158 © 1998 The Raptor Research Foundation, Inc. REVIEW OF HAZARDS TO RAPTORS FROM PEST CONTROL IN SAHELIAN AFRICA James O. Keith and Richard L. Bruggers USD A, APHIS, National Wildlife Research Center, 1201 Oakridge Drive, Fort Collins, CO 80525 U.S.A. Abstract. — Pesticides applied to control pests sometimes kill raptors. One region receiving frequent and heavy pesticide treatments is the Sahel that lies between the Sahara Desert and tropical forest areas of northern Africa. Plagues of locusts, grasshoppers, birds and rodents periodically develop in the Sahel when heavy rainfall follows periods of drought. Insecticides, avicides and rodenticides are applied over millions of hectares at rates known to kill birds. Observations indicate that many pesticide applications in the Sahel do not cause serious mortality. Only minimal raptor losses were reported following appli- cations of malathion, fenitrothion, chlorpyrifos and other insecticides for control of locusts and grass- hoppers scattered over about 14 million acres in northern Africa. Similarly, applications of zinc phos- phide bait on 430 000 ha in Sudan did not cause known losses of raptors. In contrast, use of fenthion to control Red-billed Quelea (Quelea quelea) in breeding colonies often killed or debilitated nontarget vertebrates, including raptors. A more specific avicide should be developed for quelea control. Until then, nontarget birds should be kept out of colonies for several days following avicide applications and dead quelea should be removed from treated areas. Key Words: Raptors', pesticides', locust control, quelea control, rodent control, Africa. Resumen de amenazas para las aves rapaces a partir del control de plagas en el Sahel, Africa Resumen. — La utilizacion de pesticidas puede ocasionar la muerte de aves rapaces. Una de las regiones en donde se utilizan estas substancias con frecuencia y volumenes importantes es el Sahel entre el desierto de Sahara y los bosques tropicales del norte de Africa. Periodicamente despues de la sequia y con las fuertes Iluvias ocurren erupciones de langostas, saltamontes, aves y roedores. Insecticidas, avi- cidas y rodenticidas son aplicados en millones de hectareas en concentraciones que pueden eliminar a las aves. Algunas observaciones indican que muchas de las aplicaciones de pesticidas en el Sahel no causan una mortalidad importante. Algunos casos de aves rapaces han sido reportados despues de la aplicacion de malathion, fenitrothion, chlorpyrifos y otros insecticidas para el control puntual de lan- gostas y saltamontes en una extension de 14 millones de acres en el norte del Africa. En forma similar la utilizacion de cebos de fosfato de zinc en 430 000 hectareas en Sudan no causo perdidas conocidas en aves rapaces. En contraste, el uso del fenthion para el control de Quelea quelea en sus colonias de anidacion ha ocasionado la muerte y/ o afectado algunos vertebrados incluyendo a las aves rapaces. Se debe evitar la presencia de aves rapaces por varios dias en las colonias despues de aplicar el avicida, como tambien se debe remover las Quelea muertas de estas areas. [Traduccion de Cesar Marquez] The Sahel is a strip of arid country that lies south of the Sahara Desert and north and east of the forests of central ;^rica. The area includes most of the farming and grazing lands across Africa between 10°-17°N latitude and south to about 5°S latitude in East Africa. Soils are poor and annual precipitation is low (10-90 cm), but the area has been farmed and grazed by indigenous peoples for thousands of years. The vegetative cover includes the grasslands of the subdesert, the shrubsteppes, and the thornbush and woodland savannas of the Sahelian and Sudanian bioclimatic zones. After prolonged droughts, rains support vegeta- tive growth, which stimulates population irruptions of the desert locust {Schistocerca gregaria) and the Senegalese grasshopper {Oedaleus senegalensis) (Rowley 1993), as well as the multimammate rat (Praomys natalensis), the Nile rat {Arvicanthis niloti- cus), and other desert species (Fiedler 1994). Red- billed Quelea ( Quelea quelea) also increase with the rains as the birds will breed only when precipita- tion is sufficient to produce the insects and grass seeds they need to raise young (Jones 1989). All of these animals can become serious pests when 151 152 Keith and Bruggers VoL. 32, No. 2 huge populations decimate forage for livestock and cultivated crops. The advent of synthetic pesticides has intro- duced new methods to kill these pests. The use of pesticides often creates hazards to nontarget or- ganisms and there is a potential for undesirable effects on nontarget wildlife such as raptors from pest control programs. Because there is little ob- jective information available on impacts of pesti- cides on raptors in the Sahel, this paper will de- scribe the nature, frequency, and spatial extent of control programs for acridids (locusts and grass- hoppers) , birds and rodents in the Sahel and what is known about their effects on African raptors. Some 93 species of raptors are either resident or seasonal visitors to sub-Saharan Africa. Included are 21 species of eagles, 17 falcons, 14 sparrow- hawks and goshawks, eight vultures, seven buz- zards, six kites, six harriers, the Osprey {Pandion haliaetus), the Secretary-bird {Sagittarius serpentar- m$) and 12 owls (Brown et al. 1982, Pickford et al. 1989). Of those, 17 Palearctic migrants and 60 Af- rotropical species frequent the Sahel. Fifty-eight of the 77 species are uniquely adapted to the open, arid habitats of the Sahel and eastern Africa and many specialize their feeding on the pests that oc- cur there. Control of Locusts and Grasshoppers and Its Effect on Raptors Between 1908-70, locust populations irrupted four times in the Sahel, with each plague lasting from 7—13 yr (Rowley 1993). The Sahel suffered a severe and prolonged drought from 1970-85. Rains began in 1985 and were heavy in 1987. Lo- cust populations responded, became mobile, and invaded 43 countries in northern Africa and Asia comprising more than 20% of the earth’s land sur- face. Plagues were characterized by numerous large swarms of flying adult locusts and bands of hopping immatures scattered thinly throughout in- fested areas. Swarms migrated between seasonal breeding areas by flying with strong winds enabling them to cover up to 1000 km in a week (Symmons 1992). Between outbreaks, scattered locust aggre- gations persisted in small recessional areas of wet- ter habitat. Grasshopper populations increased with rains, but they were not migratory and caused less damage. From 1986-90, applications of insec- ticides for locust control were made to almost 14 million ha in northern Africa (Everts 1990). About one-half of the area treated was in North Africa Table 1. Number of avian species that are major pred- ators of Sahelian pests. ^ Family Desert Locusts Quelea Rodents Accipitridae 8 19 11 Falconidae 5 3 4 Strigidae 8 2 6 Other 80 17 1 “From Brown et al. (1982), Steedman (1988), Pickford et al (1989). from Morocco to Egypt. The rest was in the Sahel from Mauritania and Senegal to Ethiopia. While most applications for locust control were made in relatively small areas (1-12 km^), larger areas (~100 km^) were treated for control of grasshop- per populations. Both organophosphate and car- bamate insecticides were used and persisted for only a few days in the desert environments (Everts 1990). Migrating locusts represent one of the largest biomasses of insects ever to congregate on earth. In Somalia, one plague area was estimated to con- tain 1.6 billion locusts weighing 50000 tons (Row- ley 1993). Such concentrations attract large num- bers of resident and migratory birds that often trav- el with swarms of locusts. Smith and Popov (1953) observed thousands of eagles and falcons feeding for days on a swarm containing tens of millions of locusts. Many raptors feed on locusts (Table 1) and the African Cuckoo Falcon {Aviceda cuculoides), Montagu’s Harrier {Circus pygargus), Red-footed Falcon {Falco vespertinus) , Amur Falcon {Falco amu- rensis) , and Lesser Kestrel {Falco naumanni) are considered locust and grasshopper specialists (Steedman 1988, Pickford et al. 1989). Of the estimated 15 million liters of insecticides used in the recent outbreak of locusts in the Sahel, most was the organophosphate insecticide, fenitro- thion (Symmons 1992). The recommended rate of fenitrothion application (250-300 g/ha) was near that shown to cause mortality in birds (Steedman 1988). Toxicity data (Smith 1987) suggested that at rates of 250 g/ha chlorpyrifos, a widely used car- bamate insecticide, might also pose hazards to birds. Organophosphate and carbamate insecti- cides kill animals by decreasing levels of the neu- rotransmitter acetylcholinesterase (ChE) . Measure- ment of ChE activity in brain and blood can indi- cate the intensity of exposure of animals to the in- June 1998 Pesticides and African Raptors 153 secticides, and reductions of 50% or more in brain levels may cause mortality (Hill and Fleming 1982). Controlled studies of the environmental effects of locust control are not possible during operational applications of insecticides to moving swarms of lo- custs. However, several experimental applications of insecticides to large study plots have been eval- uated. In 1989, experimental applications were made of fenitrothion (485 and 825 g/ha) and chlorpyrifos (270 and 387 g/ha) to four large (6 km'^) plots in a 400 km^ study area located in the savanna of northern Senegal (Mullie and Keith 1993). A few individuals of 18 raptor species were identified in or near the study area. Fourteen species were of Afrotropical origin and four were Palearctic mi- grants. None of the raptors were known to be ex- posed to or affected by treatments. Mortality of other birds was negligible (seven dead, 12 debili- tated) on study plots, but numbers of birds showed significant decreases (up to 43%) on all treated plots. Decreases probably were due to movements from plots following a reduction in the insect food supply of birds. For instance. Singing Bushlarks (Mirafra cantillans) on the high dosage fenitrothion plot ate only about one-half the amount of insects after treatments as larks on the control plot. Ex- perimental applications of the organophosphate insecticides malathion (750 g/ha) and dichlorvos (200 g/ha) were made to six large (3 km^) plots in southern Morocco in 1992 (Keith et al. 1995). These two insecticides were the primary ones used against locusts in Morocco between 1987-89. Three Long-legged Buzzards {Buteo rufinus), two Eastern Marsh Harriers {Circus spilonotus), and 20 Eurasian Kestrels (Falco tinnunculus) were seen on or near the 160 km^ study area, but there was no indication that they or any other birds or mammals were killed or debilitated by the insecticides. Six- teen Little Owls {Athene noctud) were captured and fitted with radios so that their movements and mor- tality could be monitored before and after treat- ments. No owls died and neither their home rang- es (<10 ha), their daily movements (20-460 m), nor their foods (largely beedes) changed following insecticide applications. The absence of heavy bird mortality in Senegal on fenitrothion plots treated with 485 and 825 g/ ha was surprising. Applications to forests in Canada at 300 g/ha and higher most often killed small pas- serines (Bushy et al. 1983) . Mortality and decreases in bird abundance followed applications of 210- 410 g/ha of fenitrothion to rangelands in the west- ern U.S. for grasshopper control (McEwen 1982). During the 1986-89 locust outbreak in the Sahel, raptor mortality was reported in areas operation- ally sprayed with fenitrothion (W. Mullie pers. comm.) and with dichlorvos (A. El Hani pers. comm.), but documented losses were minimal (one to five birds) and infrequent in occurrence (four cases). Keith (1994) discussed possible rea- sons why large numbers of birds were killed in Can- ada but not in the Sahel following fenitrothion ap- plications. We concluded that locust control with most or- ganophosphate and carbamate insecticides did not result in serious mortality of nontarget vertebrates. Experimental studies of fenitrothion, chlorpyrifos and malathion did not show any mortality. Like- wise, there were no reports of excessive avian kills following operational applications of those insec- ticides. Some mortality of vertebrates was reported following dichlorvos applications in Morocco, which was the only country in which it was used. Dichlorvos should not be used for locust control. Its only advantage is a quick knockdown of locusts, but this is offset by its high toxicity to vertebrates. Control of Red-billed Quelea and Its Eefect on Raptors Irruptions in quelea populations are more sub- tle. Quelea are numerous every year and can be found scattered throughout their range where hab- itats are relatively stable, such as near marshes, lakes and irrigated agriculture. Their numbers, however, fluctuate considerably with environmen- tal conditions (Jones 1989). During droughts last- ing for several years, fewer young quelea are pro- duced. When annual rainfall increases to 20-45 cm, conditions improve sufficiently to support breeding (Jaeger et al. 1989). With good rains, quelea populations increase and can become ir- ruptive. The birds can breed two or three times in one year and thereby increase populations to plague proportions in a single breeding season (Jarvis and Mundy 1989). The most common method of quelea control is ground or aerial application of Queletox(r) (60% fenthion) to breeding colonies and night roosts (Elliott and Allan 1989). Queletox(r) is most often applied at a volume of 4 1/ha in a ULV formula- tion, but amounts applied vary (2—8 1/ha), with smaller areas (<5 ha) receiving the highest rates (Meinzingen et al. 1989). Four liters of Quele- 154 Keith and Bruggers VoL. 32, No. 2 Table 2. Nontarget avian species found dead in treated quelea colonies. Avicide and Location No. Raptors Killed Species Other Birds Killed Reference® Parathion South Africa 46 Black Kites 409 1 16 Tawny Eagles Steppe Eagles (Aquila nipalensis) 1 Wahlberg’s Eagle Mali 400 Black Kites — 2 Fenthion Sudan 100 Unknown — 2 Tanzania 1 Goshawk 15 2 South Africa 6 Common Buzzards 8 3 Kenya 5 Tawny Ealges'^ 44 4 1 Pearl-spotted Owlet’’ Kenya 24 8 species >100 5 Botswana 56 Common Buzzards — 6 8 Steppe Eagles 6 Wahlberg’s Eagles ^ 1-Tarboton (1987), 2-Meinzingen et al. (1989), 3-Colahan and Ferreira (1989), 4-Bruggers et al. (1989), 5-Thomsett (1987), 6-R Bruggers (unpubl. data). Seriously debilitated, but fate unknown. tox(r) per ha equals a fenthion rate of 2.4 kg/ha (2.14 Ib/ac). This is the highest application rate of fenthion for any use anywhere in the world. Elliott (1989) estimated the breeding population of quelea in Africa at 753 million birds and the total area of colonies used by 249 million birds was 4145 ha. Proportionately, 753 million birds would use about 12 500 ha. Therefore, if all colonies in Africa were treated at one time, only 125 km^ would be involved. Colonies can develop in most months of the year in response to seasonal rains (Jaeger et al. 1989). This extends the period of fenthion applications, hut reduces the probability of raptors continuously intercepting treated areas containing contaminated habitats and dead que- lea. In a local area, however, raptors will move from colony to colony as they are sequentially treated (Bruggers et al. 1989). Thiollay (1989) listed 80 species of avian preda- tors in 15 families that have been observed in que- lea breeding colonies. He considered 41 species in nine families to be major predators (Table 1). Of those, 24 were raptors and included eight species that specialize in feeding on quelea: the Black Kite {Milvus migrans), Shikra {Accipiter badius), Wahl- berg’s Eagle {Aquila wahlbergi), Tawny Eagle {Aq- uila rapax), Dark Chanting-Goshawk {Melierax me- tabates), Gabar Goshawk {Melierax gabar), African Harrier-Hawk {Polyboroides typus), and Tanner Fal- con {Falco biarmicus) . Predator numbers increase in colonies when quelea eggs hatch, and are greatest when flightless young leave nests to perch on branches several days before fledging. Up to 1200 eagles can gather at a colony at that time (Pienaar 1969). Raptor densities at colonies often can ex- ceed those in the adjacent countryside by 70-500- fold. Small raptors can eat four to eight young per day and larger raptors can eat 15-20 per day. In Mali, Thiollay (1989) estimated 678 and 763 avian predators ate 8000 and 10 000 young quelea from two colonies, respectively. Fenthion applications often cause some nontar- get mortalities of smaller passerines and raptors (Table 2). Fenthion is an organophosphate insec- ticide, and applications kill and contaminate many invertebrates in treated areas. This creates an abundance of dead and dying insects that attracts insectivorous birds to treated sites. Insects collect- ed in treated quelea colonies have contained up to 7.2 ppm of fenthion (Bruggers et al. 1989). Like- wise, up to 90% of quelea die in colonies that con- tain millions of dead and dying quelea, each of June 1998 Pesticides and African Raptors 155 Table 3. ChE levels found in blood of raptors in or near areas sprayed with fenthion"^. Species Pre- treat- ment Unex- POSED Exposed Tawny Eagles 1416 1300 840 1418 1820 583 1437 335 200 136 Pygmy Falcons — 1264 623 888 661 Pale Chanting-Goshawks — — 513 627 785 Gabar Goshawks — — 632 663 Bateleur Eagles — 2260 — 2025 1450 ^ Units are mU/ml of blood, from Bruggers et al, (1989). which can contain as much as 45 to 92 ppm of fenthion (Jaeger and Elliott 1989). The acute oral toxicity of fenthion ranges from 2.0-25 mg/kg for nonraptor species (Smith 1987) and is 1.3 mg/kg for American Kestrels {Falco sparverius) (Schafer 1972), the only raptor against which fenthion is known to have been tested. Therefore, a meal of insects or quelea killed with fenthion contains suf- ficient residues to be lethal to raptors (Bruggers et al. 1989). In two Kenyan colonies sprayed with fenthion (Bruggers et al. 1989), blood from 13 of 22 raptors examined exhibited low ChE activity (Table 3). Five Tawny Eagles were sufficiently debilitated that their survival was of concern; ChE levels in blood were 41, 59, 76, 86, and 90% below normal. No raptors other than Tawny Eagles were tested before spraying to determine normal ChE levels in blood; therefore, the degree of inhibition in other raptors could not be calculated. After spraying, three Pyg- my Falcons {Polihierax semitorquatus) , three Pale Chanting-Goshawks (Melierax canorus), and two Ga- bar Goshawks had low ChE levels. Some of the rap- tors ultimately may have died, but no dead raptors were found within 5 d of treatments. Brain tissue from a moribund Pearl-spotted Owlet {Glaucidium perlatum) contained 6.3 =1 mol/g of ChE activity, which indicated serious fenthion exposure. Three Bateleur Eagles {Terathopius ecaudatus) had high ChE levels after spraying and apparently did not feed on dead or debilitated quelea. Nine of 10 spe- cies of smaller birds found dead or debilitated showed average brain ChE inhibition of 15-76%. Thomsett (1987) observed mortality of raptors near Mt. Kenya following aerial and ground spray- ing of fenthion at high rates on wheatfields and quelea roosts. Water troughs used by quelea also were poisoned with fenthion. Twenty-four raptor carcasses were recovered including those of a Sec- retary-bird, two Black Kites, four Black-shouldered Kites (Elanus caeruleus), four Tawny Eagles, seven Augur Buzzards {Buteo augur), a Gabar Goshawk, three Cape Eagle-owls {Bubo capensis) , and two Ver- reaux’s Eagle-owls {Bubo lacteus). Counts taken on 4800 ha of farmland before and after treatments suggested decreases of 85% in five of the eight spe- cies affected. Surveys conducted along the Nairobi- Naivasha road in 1974 were repeated after fenthi- on applications in 1986. Results indicated decreas- es of 93% in Augur Buzzards, Long-crested Eagles {Lophaetus occipitalis), and Black-shouldered Kites (Thomsett 1987). As fenthion is very toxic to most birds, it should either be replaced with an avicide more specific to quelea or used so as to minimize nontarget haz- ards. In South Africa, exploders are set out to dis- perse raptors and other birds from quelea colonies before they are treated. In addition, quelea car- casses are removed from colonies to prevent sec- ondary poisoning (G.H. Verdoorn pers. comm.). These practices should be followed in the Sahel. Raptors should be razed from colonies before and during a 48-hr period after applications. Bruggers et al. (1989) found that fenthion residues on grass- es and debilitated quelea were high the day after spraying, but decreased 75-80% by the second day after spraying. Thus, after 2 d raptors and other birds should be able to safely return to treated ar- eas. Fenthion is much less toxic to mammals than to birds (Smith 1987), but its effects on mamma- lian predators and scavengers are not known. Control of Rodents and Its Effect on Raptors Rodent populations irrupt when food is abun- dant because they have short reproductive cycles and multiple generations in a year. Fiedler (1994) reported 10 irruptions of the multimammate and Nile rat, each lasting 1-2 yr between 1951-89 in Sudan, East Africa and Zimbabwe. Multimammate and Nile rats live in areas with loam or clay soils 156 Keith and Bruggers VoL. 32, No. 2 and good vegetative cover. They cause damage to crops of groundnuts, sesame, sorghum, fruits and vegetables. In contrast, arid-tolerant species of Jirds {Meriones spp.), Jerboas (Jaculus spp.), and Gerbils {Tatera spp. and Gerbillus spp.) inhabit semidesert country with sandy soils and sparse vegetation, and consume the seeds of millet and sorghum planted by farmers at the beginning of the rainy season. Irruptions of these latter species occurred in 1962, 1977, and 1986-87. Irruptions of rodents seriously affected agricul- tural areas throughout the Sahel in 1986 and 1987. In nine crop areas of Chad, rodent densities were estimated from 360-2290 per ha in 1987 (Fiedler and LaVoie 1992). In the Sudan, the rodent prob- lem was recognized throughout the country in 1986, a state of emergency was declared, and in- ternational donors supplied most of the zinc phos- phide used in the 1987 control program. The pro- gram conducted there from June-August 1987 was the largest ever attempted in a country. About 1300 metric tons of a 1.0% zinc phosphide bait (sor- ghum) was prepared and distributed by the gov- ernment and applied by farmers to about 430 000 ha infested with rodents (Fiedler and LaVoie 1992). Some 21 raptor species in the Sahel rely on ro- dents for much of their diet (Table 1). Eight of those species specialize on rodent prey (Brown et al. 1982, Pickford et al. 1989): the Augur Buzzard, Common Buzzard {Buteo buteo). Black-shouldered Kite, Montagu’s Harrier, Pallid Harrier {Circus ma- crourus), Barn Owl {Tyto alba), African Grass Owl {Tyto capensis), and White-faced Scops Owl {Otus leucotis). Raptors congregate in areas where ro- dents are abundant during irruptions and readily take animals poisoned during control campaigns. As with dead and moribund locusts and quelea, the affected rodents pose potential hazards to raptors and other predators. The critical factors are the susceptibility of predators and their exposure to poisons used to kill the pests. Rodenticides also dif- fer in their toxicity to secondary consumers and exposure of scavengers can be mediated by con- ditions in the field. Rodenticides can kill nontarget animals that eat poisoned bait. In addition, rodents are eaten by numerous predators, and secondary poisoning seems almost unavoidable. Still, the impact of ro- dent control on raptors can vary with the poison used, where baits are placed, whether rodents die in their burrows or on the surface, and how raptors eat the poisoned rodents. The control of rodents in the Sudan in 1987 with zinc phosphide on sorghum baits was a well-con- ducted program with minimal effects on nontarget animals. Training was conducted for all personnel including farmers, and baits were prepared, trans- ported, and in most cases, stored and applied safe- ly. Rodent activity was reduced an estimated 90% throughout the area of infestation. Several goats and chickens died from eating bait, but farmers admitted deaths were due to their failure to isolate animals from stored bait and to the poor place- ment of bait in rodent burrows. Wild rabbits and unknown gallinaceous birds (reported as “wild hens”) were killed, again because bait was poorly placed in burrows. Village leaders corrected these problems when they occurred. Black Kites and Tawny Eagles were seen in large aggregations feeding on poisoned rodents in treat- ed areas. Farmers observed avian predators regur- gitate rat intestines containing zinc phosphide bait, but claimed most birds removed the intestines from rodents before eating them. In our searches, we did not see dead avian scavengers and none were reported to us by villagers or technicians who readily related problems they observed with the control program. Trials to evaluate the secondary toxicity of zinc phosphide have been run on foxes, dogs, cats, mustelids, rats, mongooses, owls, eagles, kestrels, vultures, turtles, alligators and snakes. Results show little potential for secondary poisoning (Johnson and Fagerstone 1994). Zinc phosphide does not accumulate in animal tissue. Tests have shown that 90% of the zinc phosphide on bait eaten by ground squirrels was metabolized before the ani- mal died. In the presence of water and weak acid in the stomach, zinc phosphide releases phosphine gas, which is absorbed into the blood and causes heart, liver, lung and kidney failure. Poisoning of scavengers can occur from ingestion of bait re- maining in the gastrointestinal tracts of rodents. Most scavengers and predators, when given a choice, refused to eat the tracts of poisoned ani- mals (Johnson and Fagerstone 1994). Studies and anecdotal evidence from field observations show zinc phosphide can kill nontarget animals that con- sume baits, but the probability of secondary poi- soning is low. Zinc phosphide is used for rodent control in some areas of Egypt (Mullie and Meininger 1985). June 1998 Pesticides and Aerican Raptors 157 There were reports of mass mortalities of raptors due to zinc phosphide, but no details were pre- sented. Numbers of nine breeding species and 12 wintering species of raptors have decreased in Egypt. Mullie and Meininger (1985) felt this was due to changes in agricultural practices, habitat de- struction and use of pesticides. They considered rodenticides as an important factor preventing re- covery in raptor populations. Raptors have been killed by other rodenticides used in Africa. In 1967, Klerat(r), a wax-block bait containing 0.005% brodifacoum, was used by the Plant Protection Department in Chad to protect maturing rice from rodents. In doing so, a large number of raptors was killed by secondary poison- ing (Fiedler and LaVoie 1992). Thallium sulfate (Mendelssohn 1972) and azodrin (Mendelssohn and Paz 1977, Mullie and Meininger 1985) have killed raptors after their use in rodenticide baits. Conclusions Pesticides used in agriculture often cause raptor mortality (Balcomb 1983, Henny et al. 1985, Smith 1987, Goldstein et al. 1996). Indirect effects on raptors from pesticides are not well-known. In the Sahel, locust insecticides reduced insect food sup- plies and thereby lowered reproductive success in nonrap torial birds. Raptors following locust swarms face food shortages after swarms are sprayed with insecticides. In addition, exposure to pesticides can cause sublethal debilities by adverse- ly affecting physiological processes controlling re- production and longevity. There is the potential for hormonal disruptions in animals exposed to insecticides and other environmental contami- nants (Colborn et al. 1996). On the surface it appears that control of acri- dids, quelea and rodents has a great potential to cause losses of resident and migratory raptors in the Sahel. Every pest killed contains residues that could cause secondary poisoning. Still, studies do not indicate the likelihood of serious primary or secondary poisoning of raptors if control programs are intelligently conducted. Hazards to raptors be- come real if use protocols are not followed, if eco- nomics and expediency are given priority over en- vironmental safety, and if efforts are not made to change existing programs when improvements are clearly needed. Literature Cited Balcomb, R. 1983. Secondary poisoning of Red-shoul- dered Hawks with carbofuran. J. Wildl. Manage. 47; 1129-1132. Brown, L.H., E.K. Urban and K. Newman. 1982. The birds of Africa. Volume 1. Academic Press, London, U.K. Bruggers, R.L., M.M. Jaeger, J.O. Keith, P.L. Hegdal, J.B. Bourassa, A.A. Latigo and J.N. Gillis. 1989. Im- pact of fenthion on nontarget birds during quelea control in Kenya. Wildl. Soc. Bull. 17:149-160. Busby, D.G., P.A. Pearce, N.G. Garrity and L.M. Reyn- olds. 1983. Effects of an organophosphorus insecti- cide on brain cholinesterase activity in White-throated Sparrows exposed to aerial forest spraying. J. Appl Ecol. 20:255-263. Colahan, B.D. and N.A. Ferreira. 1989. Steppe Buzzards poisoned in the course of quelea spraying in the Or- ange Free State, South Africa. GabarA:\l. Colborn, T., D. Dumanoski and J.P Meyers. 1996. Our stolen future. Penguin Books, New York, NY U.S.A. Elliott, C.C.H. 1989. The pest status of the quelea. Pages 17-34 in R.L. Bruggers and C.C.H. Elliott [Eds.], Quelea quelea — ^Africa’s bird pest. Oxford Univ Press, New York, NY U.S.A. and R.G. Allan. 1989. Quelea control strategies in action. Pages 317-326 in R.L. Bruggers and C.C.H. Elliott [Eds.], Quelea quelea — ^Africa’s bird pest, Ox- ford Univ. Press, New York, NY U.S.A. Everts, J.W. [Ed.]. 1990. Environmental effects of chem- ical locust and grasshopper control: a pilot study. Pro- ject Report. Project ECLO/SEN/003/NET. FAO/UN Rome, Italy. Fiedler, L.A. 1994. Rodent pest management in eastern Africa. FAO Plant Protection and Production Paper 123. FAO/UN. Rome, Italy. AND G.K. LaVoie. 1992. Solving rodent pest prob- lems in the Sahel. Pages 349-353 in ITnstitut du Sahel [Ed.], Deuxieme Seminaire sur la Lutte Integree Cen- tre les Ennemis des Cultures Vivrieres dans le Sahel. Jan. 1990. Bamako, Mali. Goldstein, M.I., B. Woodbridge, M.E. Zaccagnini, S.B. Canavelli and A. Lanusse. 1996. An assessment of mortality of Swainson’s Hawks on wintering grounds in Argentina./. Raptor Res. 30:106-107. Henny, C.J., L.J. Blus, EJ. Kolbe and R.E. Fitzner. 1985 Organophosphate insecticide (famphur) topically ap- plied to catde kills magpies and hawks. J. Wildl. Man- age. 49:648-658. Hill, E.F. and W.J. Fleming. 1982. Anticholinesterase poisoning of birds; field monitoring and diagnosis of acute poisoning. Environ. Toxicol. Chem. 1:27-38. Jarvis, M.J.F. and PJ- Mundy. 1989. Current policies and future plans. Pages 142-147 in P.J. Mundy and M.J.F. Jarvis [Eds.], Africa’s feathered locust. Baobab Books, Harare, Zimbabwe. Jaeger, M.E. and C.C.H. Elliott. 1989. Quelea as a re- source. Pages 327-338 in R.L. Bruggers and C.C.H Elliott [Eds.], Quelea quelea — ^Africa’s bird pest. Ox- ford Univ. Press, New York, NY U.S.A. Jaeger, M.M., R.L. Bruggers and W.A. Eric:kson. 1989 158 Keith and Bruggers VoL. 32, No. 2 Formation, sizes, and groupings of quelea nesting col- onies. Pages 181-197 in R.L. Bruggers and C.C.H. El- liott [Eds.], Quelea quelea — ^Africa’s bird pest. Oxford Univ. Press, New York, NY U.S.A. Johnson, G.D. and K.A. Fagerstone. 1994. Primary and secondary hazards of zinc phosphide to nontarget wildlife — a review of the literature. DWRC Res. Rep. 11-55-005. USDA/APHIS/ADC, Washington, DC U.S.A. Jones, P.J. 1989. Factors determining the breeding season and clutch size. Pages 158—180 in R.L. Bruggers and C.C.H. Elliott [Eds.], Quelea quelea — ^Africa’s bird pest. Oxford Univ. Press, New York, NY U.S.A. Keith, J.O. 1994. Insecticides: why are animals killed only some of the time? Pages 4—13 in I.D. Thompson [Ed.], Proc. Internatl. Union Game Biologists XXI Congress. August 1993. Halifax, Nova Scotia, Canada. , R.L. Bruggers, PC. Matteson, A. El Hani, S. Ghaout, L.A. Fielder, E.H. Arroub, J.N. Gillis and R.L. Phillips. 1995. An ecotoxicological assessment of insecticides used for locust control in southern Mo- rocco. DWRC Res. Rep. 11-55-006. USDA/ APHIS/ ADC, Washington, DC U.S.A. McEwen, L.C. 1982. Review of grasshopper pesticides ver- sus rangeland wildlife and habitat. Pages 362-382 in J.M. Peak and P.D. Dalke [Eds.], Proc. Wildlife-live- stock Relationships Symposium. April 1981. Univ. Ida- ho, Moscow, ID U.S.A. Meinzingen, W., E.S.A. Bashir, J.D. Parker, J.-U. Heckel AND C.C.H. Elliott. 1989. Lethal control of quelea. Pages 293-316 in R.L. Bruggers and C.C.H. Elliott [Eds.], Quelea quelea — ^Africa’s bird pest. Oxford Univ. Press, New York, NY U.S.A. Mendelssohn, H. 1972. The impact of pesticides on bird life in Israel. ICBP Bull. 11:75-104. Gland, Switzer- land. AND U. Paz. 1977. Mass mortality of birds of prey caused by Azodrin, an organophosphorus insecticide, Biol. Conserv. 11:163-170. Mullie, W.C. and P.L. Meininger. 1985. The decline of bird of prey populations in Egypt. Pages 61-92 in I. Newton and R.D. Chancellor [Eds.], Conservation studies on raptors. ICBP Tech. Publ. No. 5. Gland, Switzerland. and J.O. Keith. 1993. The effects of aerially aj> plied fenitrothion and chlorpyrifos on birds in the savannah of northern Senegal. J. Appl. Ecol. 30:536- 550. PiCKEORD, R, B. Pickford AND W. Tarboton. 1989. Afri- can birds of prey. Cornell Univ. Press, Ithaca, NY U.S.A. Pienaar, V. de V. 1969. Observations on the nesting hab- its and predators of breeding colonies of Red-billed Quelea, Quelea quelea lathami, in the Kruger National Park. Bokmakierie 21 (Suppl.): 11-15. Rowley, J. 1993. Grasshoppers and locusts — the plague of the Sahel. The Panos Institute, London, U.K. Schafer, E.W., jR. 1972. The acute toxicity of 369 pesti- cidal, pharmaceutical, and other chemicals to wild birds. Toxicol. Appl. Pharmacol. 21:315-330. Smith, G.J. 1987. Pesticide use and toxicity in relation to wildlife: organophosphorus and carbamate com- pounds. Fish Wildl. Serv. Resour. Publ. 170. USDI, Washington, DC U.S.A. Smith, J.D. and G.B. Popov. 1953. On birds attacking desert locust swarms in Eritrea. Entomolo^st 86:3-7. Steedman, a. [Ed.]. 1988. Locust handbook. Overseas Development Natural Resources Institute, London, U.K. Symmons, P. 1992. Strategies to combat desert locust. Crop Prot. 11:206-212. Tarboton, W. 1987. Red-billed Quelea spraying in South Africa. Gabar2:^8-39. Thiollay, J.-M. 1989. Natural predation on quelea. Pages 216-229 in R.L. Bruggers and C.C.H. Elliott [Eds.], Quelea quelea — ^Africa’s bird pest. Oxford Univ. Press, New York, NY U.S.A. Thomsett, S. 1987. Raptor deaths as a result of poisoning quelea in Kenya. Gabar 2:33-38. Received 10 April 1997; accepted 3 February 1998 J. Raptor Res. 32{2):159-162 © 1998 The Raptor Research Foundation, Inc. HAEMOPROTEUS TINNUNCULI IN CRESTED CARACARAS (CARACARA PLANCUS AUDUBONIl) FROM SOUTHCENTRAL FLORIDA Garry W. Foster Department of Pathobiology, College of Veterinary Medicine, University of Florida, Gainesville, FL 32610 US. A. Joan L. Morrison Department of Wildlife Ecology and Conservation, P.O. Box 110430, University of Florida, Gainesville, FL 32611 U.S.A. Christine S. Hartless Department of Statistics, Institute of Food and Agricultural Sciences, University of Florida, Gainesville, FL 32611 U.S.A. Donald J. Forrester Department of Pathobiology, College of Veterinary Medicine, University of Florida, Gainesville, FL 32610 U.S.A. Abstract. — From 1994-96, 223 Crested Caracaras {Caracara plancus audubonii) from eight counties in southcentral Florida were examined for blood parasites. Haemoproteus tinnunculi was the only hemato- zoan seen. The overall prevalence was 28.2%. Prevalence was higher for adults (50.0%) than for nestlings (20.4%, P = 0.0001). There were no differences in prevalence of infection between years {P = 0.8899). The only significant interaction was between age and sampling month {P = 0.0027). The adult birds had a constant prevalence from February-May, while nestlings had an increasing prevalence over the same time period. This is the first report of Haemoproteus tinnunculi in Crested Caracaras. Key Words: Caracara plancus audubonii; Haemoproteus tinnunculi; Crested Caracara-, blood', parasite. Haemoproteus tinnunculi en caranchos comunes ( Caracara plancus audubonii) en el sur y centro de Florida Resumen. — Deste 1994 hasta 1996, 223 caranchos comunes {Caracara plancus audubonii)de: ocho con- dados en el sur y centro de Florida fueron examinados para parasitos sanguineos. Haemoproteus tinnunculi fue el unico hematozoaio encontrado. La prevalencia en general fue 28.2%. La prevalencia en adultos fue mas alta (50.0%) que en las crias (20.4%, P - 0.0001). No hubo differencias en las prevalencias de infeccion entre los anos {P = 0.8899). La unica interaccion significante fue entre la edad y el mes en que se colecto la muestra {P = 0.0027). Los adultos tuvieron una prevalencia constarite de Febrero hasta Mayo, mientras que la prevalencia en las crias fue subiendo durarte el mismo perido. Este es el primer reporte de Haemoproteus tinnunculi en caranchos comunes. [Traduccion de Tania Carnes] The Crested Caracara {Caracara plancus) has a fragmented geographic distribution occurring from northern Mexico to Tierra del Fuego in South America, and in North America in Texas and Arizona, with an isolated population in south- central Florida (Morrison 1996). The blood para- sites of these medium-sized raptors have not been studied thoroughly. We are aware of only three published blood parasite surveys, which included samples from two Crested Caracara subspecies. Carini and Maciel (1916) reported Haemoproteus in a Crested Caracara (C. p. plancus) from Brazil, Renjifo et al. (1952) reported Haemoproteus sp. in three of four Crested Caracaras ( C. p. cheriway) in eastern Colombia, and Gabaldon and Ulloa (1980) reported three undetermined species of Plasmodi- um in two nestling C. p. plancus collected in Vene- zuela. The presence of an unidentified species of Haemoproteus, in Crested Caracaras {C. p. audubon- ii) from Florida, was mentioned in the account by Morrison (1996), but no specific data were given. The impact blood parasites may have on the cara- cara’s survival in Florida is unknown. The objec- tives of this study were to determine the species 159 160 Foster et al. VoL. 32, No. 2 and prevalence of blood parasites in the subspecies of Crested Caracara from Florida. Methods During the January-July breeding season from 1994— 96, we examined 223 Crested Caracaras from eight coun- ties in southcentral Florida for blood parasites. We sam- pled 175 nestlings (5-8 wk old), and 48 breeding adults (3+ yr old). Birds in the age span between nestling and breeding adult were not sampled. Most nestlings were captured by hand from nests in the morning. During the capture process, a few older nesdings jumped from the nest to the ground where they were caught. Adult birds were captured using the methods of Morrison and McGehee (1996). Thin blood films were prepared from peripheral blood, air dried, fixed in absolute methanol, and stained by standard Giemsa technique for 1 hr at pH 7. A mini- mum of 10 000 red cells was examined at 400 X and lOOOX oil immersion to determine the presence of par- asites in each sample. A generalized linear model with logit link and bino- mial errors was fit to the data to test the effects of age class, sample year, and sampling months on the preva- lence of any parasites seen. Age class and sampling year were included in the model as factors with hxed levels, and sampling month was included as a linear covariate. All possible interactions were included, and those with P > 0.20 were deleted from the final model. This model was fit with the GLIMMIX macro described by Littell et al. (1996) for SAS 6.12 (SAS Institute, Cary, North Car- olina, U.S.A.). Because few birds were sampled in Janu- ary, June, and July, only those sampled between Febru- ary-May {N = 205) were included in the statistical anal- ysis. The significance level was set at a = 0.05 for all tests. Representative blood films have been deposited in the collections of The International Reference Centre for Avian Haematozoa, Queensland Museum, South Bris- bane, Queensland, Australia (Accession Nos. G462377- G462431); the U.S. National Parasite Collection, Belts- ville, Maryland, U.S. A. (Accession Nos. 87058-87061); and the Harold W. Manter Collection, University of Ne- braska State Museum, Lincoln, Nebraska, U.S.A. (Acces- sion Nos. 39289-39292). Results and Discussion Birds sampled by month were as follows: January (8 nestlings, one adult), February (17 nesdings, 14 adults), March (58 nesdings, nine adults), April (60 nesdings, 13 adults), May (22 nesdings, 12 adults), June (five nestlings, no adults), July (four nesdings, no adults). Haemoproteus tinnunculi was the only hematozoan seen in blood samples. Immature forms of the par- asite tended to develop in a polar or subpolar po- sition within the erythrocytes. Multiple infections within an erythrocyte were common, with up to six trophozoites in a single cell. The pigment granules were small, randomly scattered, and averaged 21 in Table 1 . Characteristics of infection of Haemoproteus tin- nunculi in 223 Crested Caracaras sampled in Florida be- tween 1994—96. Nestlings Adults All No. NO. No. County N Inf. N Inf. N Inf. DeSoto 8 5 2 1 10 6 Glades 39 2 12 4 51 6 Hendry 3 I 1 1 4 2 Highlands 67 20 19 10 86 30 Indian River 2 0 — — 2 0 Okeechobee 53 12 14 7 67 19 Osceola 2 0 — — 2 0 Polk 1 0 — — 1 0 Totals 175 40 48 23 223 63 mature gametocytes. The average number of pig- ment granules was less than the average of 23 re- ported by Peirce et al. (1990) from a Eurasian Kes- trel (Falco tinnunculus) . This variation could be due to interhost parasite variation. All other parameters were consistent with Haemoproteus tinnunculi. The overall prevalence, disregarding county, year, and age, for the 223 birds sampled was 28.2% (Table 1). For the 205 birds included in the statis- tical analysis, the prevalence was 20.4% for nest- lings and 50.0% for adults. Bird age and sampling month had a significant effect on the prevalence of Haemoproteus tinnunculi (P = 0.0001 and P = 0.0294, respectively); how- ever, there was also a significant interaction be- tween age and sampling month (P = 0.0027). There was no significant change in the prevalence in adult birds during the sampling months (test slope equal to zero, P = 0.5410), and the overall predicted prevalence was 50.1% (Fig. 1). The prev- alence in nestlings increased from February to May (test slope equal to zero, P = 0.0004) (Fig. 1). In February, tbe predicted prevalence for nestlings was 4.6%, and increased to 46.4% in May. Preva- lences of Haemoproteus tinnunculi did not vary an- nually {P = 0.8899). Adult caracaras were more likely to be infected than nestlings (P = 0.0001), which was probably due to the duration of exposure to vectors. Nest- lings were sampled from 5-8 wk of age. All adults sampled were breeding and in adult plumage, so they were at least 3 yr of age (Morrison 1996). These older birds had a much greater opportunity to be infected and probably acquired the infection June 1998 Haemoproteus in Crested Caracaras 161 Predicted g 1994 ^ 1995 ^ 1996 Figure 1. The predicted prevalence and observed an- nual prevalence of Haemoproteus tinnunculi over the sam- pling season for nestling and adult Crested Caracaras. In February of 1994, 1995, and 1996, no sampled nestlings were infected. at a much earlier age. One free-ranging adult car- acara that was tested positive in July 1992 was still positive in April 1996 indicating that the parasitem- ias may be very long-lived, or reinfection may be a common occurrence. The increase in probability of infection in nestlings from February to May could be a function of both an increase in the bit- ing activity of arthropod vectors as the tempera- tures increase in the spring, as well as an increase in the population of arthropods which serve as vec- tors. Atkinson et al. (1988) reported the year- round presence of Culicoides edeni, the vector for Haemoproteus meleagridis in Wild Turkeys {Meleagris gallopavo) , and three other ornithophilic Culicoides species in Glades County. They reported also lower numbers of Culicoides and lower biting activity dur- ing the cooler winter months, with trap collections peaking in March and April. This peak in trapping collection of Culicoides corresponds with the peak in nesting activity for the Crested Caracara in Flor- ida. Bennett et al. (1993) reviewed the pertinent lit- erature and concluded that species of Haemoproteus do not have a major impact on wild bird popula- tions and cause little direct mortality to individual free-ranging birds. However, Peirce (1989) stated that under certain conditions hematozoans could have significant pathogenic effects on hosts that had concurrent infections with other disease agents. Morrison (unpubl. data) found that late season caracara nests have lower fledging rates, and the fledglings from those late nests have lower long-term survival rates, when compared with early season fledglings. It is unclear if Haemoproteus tin- nunculi could be one of the factors in the increased mortality of caracaras that fledge late in the season. Additional late season stresses on these nestlings may include increased temperature and decreased food resources. Roosting at night in cabbage palms (Sabal pal- metto) up to 20 m high and foraging on the ground and perching on fence posts or vegetation (<2 m high) during the day allows the caracara to be ex- posed to insect vectors that may be vertically strat- ified. A distinct vertical distribution of ornithophi- lic Culicoides was reported by Tanner and Turner (1974) in a Virginia forest. Most of the ornitho- philic species of Culicoides they trapped increased in numbers as the trap height increased, with few or none found at ground level. In Florida, howev- er, Haemoproteus tinnunculi transmission to caracar- as took place as high as 7-20 m in the nesting trees, the same trees in which the adult birds roosted during the nonbreeding season. Further study is needed to fully understand the relationship of Hae- moproteus tinnunculi, its insect vectors, and the mode of transmission to Crested Caracaras in Flor- ida. Forrester et al. (1994) reported an overall prev- alence of 26% for Haemoproteus in Falconiformes (excluding Crested Caracaras) in Florida. This is comparable to the 28% prevalence for all the car- acaras we sampled. Haemoproteus tinnunculi has been reported in the American Kestrel {Falco spar- verius) in Florida (Forrester et al. 1994), but tbis is the first report from Crested Caracaras. Acknowledgments We thank Ellis C. Greiner for verifying our identifica- tion of Haemoproteus tinnunculi, and for reviewing an early draft of the manuscript. We also thank M.G. Spalding and M.D. Young for reviewing an early draft of the manu- script, and making helpful suggestions. We acknowledge the MacArthur Agro-Ecology Research Center, Archbold 162 Foster et al. VoL. 32, No. 2 Biological Station, and the Avon Park Air Force Range, and particularly private landowners throughout south* central Florida for permitting us access to the nesting sites. We thank S. McGehee, V. Dreitz, D. Todd, B. Meal- ey, C. Pages, M. McMillian, and J. Arnett for assistance in the field. We thank Tania A. Games for translating the Brazilian literature and for supplying the Spanish abstract for this paper. This research was supported in part by Florida Game and Fresh Water Fish Commission’s Non- Game program, and contracts from the Florida Game and Fresh Water Fish Commission’s Federal Aid to Wild- life Restoration Program, Florida Pittman-Robertson Pro- ject W-41. This is contribution No. 45 from the Mac- Arther Agro-Ecology Research Center of Archbold Bio- logical Station. This is Florida Agricultural Experiment Stations Journal Series No. R-05735. Literature Cited Atkinson, C.T., D.J. Forrester and E.C. Greiner. 1988. Epizootiology of Haemoproteus meleagridis (Protozoa: Haemosporina) in Florida: Seasonal transmission and vector abundance./. Med. Entomol. 25:45-51. Bennett, G.F., M.A. Peirce and R.W. Ashford. 1993. Avi- an Haematozoa: mortality and pathogenicity. /. Nat. Hist. 27:993-1001. Carini, a. and J. Maciel. 1916. Quelques hemoparasites du Bresil. Bull. Soc. Pathol. Exot. 9:247-265. Forrester, D.J., S.R. Telford, Jr., G.W. Foster and G.F. Bennett. 1994. Blood parasites of raptors in Florida. /. Raptor Res. 28:226-231. Gabaldon, a. and G. Ulloa. 1980. Holoendemicity of malaria: an avian model. Trans. R. Soc. Trop. Med. Hyg. 74:501-507. Littell, R.C., G.A. Milliken, W.W. Stroup and R.D. Wolfinger. 1996. SAS System for Mixed Models. SAS Institute Inc., Cary, NC U.S.A. Morrison, J.L. 1996. Crested Caracara (Caracara plan- cus). In A. Poole and F. Gill [Eds.], The birds of North America, No. 249. The Academy of Natural Sciences, Philadelphia, PA and The American Ornithologists’ Union, Washington, DC U.S.A. AND S.M. McGehee. 1996. Capture methods for Crested Caracaras./. Field Ornithol. 67:630-636. Peirce, M.A. 1989. The significance of avian Haematozoa in conservation strategies. Pages 69-76 mJ.E. Cooper [Ed.], Disease and threatened birds. ICBP Technical Bulletin No. 10, Cambridge, MA U.S.A. , G.F. Bennett and M. Bishop. 1990. The hae- moproteids of the avian order Falconiformes. J. Nat. Hist. 24:1091-1100. Renjifo, S., C. Sanmartin and J. de Zulueta. 1952. A survey of the blood parasites of vertebrates in eastern Colombia. Acta Trop. 11:151-169. Tanner, G.D. and E.C. Turner, Jr. 1974. Vertical activi- ties and host preferences of several Culicoides species in a southwestern Virginia forest. Mosq. News 34:66- 70. Received 19 May 1997; accepted 6 February 1998 J. Raptor Res. 32(2):163-169 © 1998 The Raptor Research Foundation, Inc. COMPARATIVE HEMATOLOGY AND PLASMA BIOCHEMISTRY OF RED-TAILED HAWKS AND AMERICAN KESTRELS WINTERING IN CALIFORNIA Robert W. Stein 3115 Calhoun Way, Stockton, CA 95219^3707 U.S.A. Julie T. Yamamoto, D. Michael Fry and Barry W. Wilson Department of Avian Sciences, University of California, Davis, CA 95616 U.S.A. Abstract. — In December and January of the winters of 1990-91, 1991-92, and 1992-93, blood samples were collected from 52 Red-tailed Hawks {Buteo jamaicensis) and 91 American Kestrels {Falco sparverius) . Twenty-five blood parameters were measured, including white blood cell components, plasma enzyme activities, and plasma chemistry concentrations. Interspecific comparisons were made to identify species specific differences pertinent to health assessment. American Kestrels had a higher incidence of he- matozoa infection, higher alkaline phosphatase and acetylcholinesterase activities, and higher plasma cholesterol, blood urea nitrogen, uric acid and sodium concentrations. Red-tailed Hawks had higher white blood cell concentration and eosinophil count estimates, aspartate aminotransferase and butryl- cholinesterase activities, and a higher plasma albumin concentration. Key Words; Red-tailed Hawk, Buteo jamaicensis; American Kestrel-, Falco sparverius; hematology, plasma biochemistry, hematozoa. Hematologia comparativa y bioquimica de plasma de Buteo jamaicensis y Falco sparverius durante el in- vierno en California Resumen. — En Diciembre y Enero de los inviernos de 1990-91, 1991-92, y 1992-93, se tomaron muestras de sangre de 52 Buteo jamaicensis y de 91 Falco sparverius. Veinticinco parametros de sangre fueron medidos incluyendo componentes de celulas de globules blancos, actividad de enzimas del plasma y concentraciones quimicas del plasma. Se efectuaron comparaciones inter-especificas con el fin de iden- tificar diferencias relacionadas con la salud de cada especie. Falco sparverius tuvo una mayor incidencia de infecciones por hematozoarios, mayor fosfatasa alkalina y actividad de acetilcolinesterasa, mayor colesterol en el plasma, nitrogeno de urea en la sangre, acido urico y concentraciones de sodio. Buteo jamaicensis tuvo mayor concentracion de celulas blancas y estimativos de conteo de eosinofilos, aspartato aminotransferosa y actividad de butrycolinesterosa y altas concentraciones de albumina en plasma. [Traduccion de Cesar Marquez] Hematological and plasma biochemistry mea- sures are routinely used in monitoring bird health, and potentially provide information regarding nu- tritional and immunological status, toxicant expo- sure, and other aspects of physiological function. Critical to meaningful interpretation of these val- ues are reference data from wild birds sampled un- der specific criteria. Blood samples from Red-tailed Hawks {Buteo jamaicensis) and American Kestrels {Falco sparverius) wintering in the Central Valley of California were analyzed for hematozoa infection and 25 hematological and biochemical parameters. These data were used to assess the general health of all birds captured, to establish reference ranges for these species in California, and to determine interspecific differences pertinent to health assess- ment. Methods During December and January of the winters of 1990- 91, 1991-92, and 1992-93, Red-tailed Hawks and Amer- ican Kestrels were captured in orchard areas of Stanislaus County (37°35'N, 120°50'W) and nonorchard areas of Yolo County (38°35'N, 121°41'W), California using bal- chatri traps. Within 2 hr of capture, 1-1.5 ml of blood was collected from the metatarsal or brachial vein using a heparinized syringe and a 26-gauge needle. Immediate- ly after collection, whole blood was transferred to a mi- crocentrifuge tube and spun for 10 min at 1000 X g to separate cells from plasma. Plasma was divided into three 163 164 Stein et al. VoL. 32, No. 2 aliquots: 250 |jl1 were stored at 0-4°C for biochemical analyses and two 100 |jl1 fractions were stored initially on dry ice, then at 70°C, for cholinesterase assay. When suf- ficient blood was available, three blood smears were pre- pared and air-dried. A qualitative physical examination was conducted on each bird to check for evidence of burns associated with surviving electrocution, a pro- nounced or “sharp” keel, excessive ectoparasite load, and discharge at the eyes or beak. Each bird was banded with a USGS band and released at its capture site. Hematological and plasma biochemical analyses were performed by Consolidated Veterinary Diagnostics, Inc. (CVD), West Sacramento, CA. Whole blood slides were examined at 40X power magnification to estimate white blood cell concentration, perform differential cell counts, and identify hematozoa (Campbell 1988, Hawkey and Dennet 1989). Hematocrit was not measured due to limited sample volume; therefore, the white blood cell concentration estimates and white cell differential count estimates reported here were not corrected for anemia. Five plasma enzyme activities: lactate dehydrogenase (LDH), alkaline phosphatase (AP), alanine aminotrans- ferase (ALT) , aspartate aminotransferase (AST) , and cre- atine kinase (CK) (AP, Bessey et al. 1946; LDH, Wolf et al. 1972; CK, Szasz et al. 1976; AST, Bergmeyer et al, 1977; ALT, Bergmeyer and Horder 1980) and 11 plasma constituent concentrations: total protein, albumin, glob- ulin (total protein minus albumin), albumin/globulin ra- tio, cholesterol, blood urea nitrogen, uric acid, glucose, sodium, potassium, calcium, and phosphorous (total pro- tein, Layne 1957; glucose, Schmidt 1961; blood urea ni- trogen, Talke and Schubert 1965; albumin, Doumas and Biggs 1972; calcium, Baginski et al. 1973; cholesterol, Al- 1am et al. 1974; phosphorus. Woo and Cannon 1984; uric acid, Merdes et al. 1985) were determined with a Hitachi 736 autoanalyzer. Three additional plasma enzyme activ- ities, total cholinesterase (CHE), acetylcholinesterase (ACHE), and butrylcholinesterase (BCHE) were mea- sured using the methods of Ellman et al. (1961). Intraspecific and interspecific comparisons were per- formed on 23 hematological and biochemical measure- ments using a t-test (f-test procedure, SAS). Based on this number of comparisons, a P-value < 0.002 was used to determine significant differences. A few Red-tailed Hawks and American Kestrels that were captured and sampled were excluded from the analyses presented here. Individuals were excluded based on findings during a qualitative physical examination; individuals showing evidence of burns associated with surviving electrocution, a pronounced or “sharp” keel, excessive ectoparasite load, or discharge from the eyes or beak, were excluded. Individuals were also excluded based on elevated organo- phosphate (OP) residue levels. In conjunction with pes- ticide exposure studies on wintering raptors (Wilson et al. 1993), those Red-tailed Hawks (N = 37) and American Kestrels {N = 30) captured in orchard areas of Stanislaus County were sampled for OP residues. To limit the anal- yses to measures from birds with low and nondetectable OP exposure (see Hooper et al. 1989 for description of OP analyses), we excluded three Red-tailed Hawks and four American Kestrels with more than 5.0 pg and 1.0 pg total OP residues on their feet, respectively. Linear regression was used to detect relationships between these allowed OP levels and hematological and biochemical pa- rameters for each species (regression procedure, SAS). Samples from 15 Red-tailed Hawks and 61 American Kes- trels that were captured in nonorchard areas of Yolo County and that met the physical examination criteria were included in the analyses. OP residues were not mea- sured on birds captured in Yolo County; however, it is unlikely that they would have encountered OP spraying in these nonorchard areas. Individuals are represented in the analyses only once. Results There were no intraspecific differences between Red-tailed Hawk age classes (adult vs. immature) or American Kestrel genders for any of the hema- tological or biochemical parameters (Mest, P < 0.002). Therefore, data from adult and immature Red-tailed Hawks were pooled as were data from male and female American Kestrels. Linear regres- sion yielded no significant relationships between OP levels and any of the hematological or bio- chemical parameters for the 34 Red-tailed Hawks with low (26 of 34) and nondetectable (eight of 34) OP residues, or for the 26 American Kestrels with low (21 of 26) and nondetectable (five of 26) OP residues (regression procedure, P < 0.05 for slope) . Sample sizes were not consistent due to the variable quality of whole blood smears and insuf- ficient plasma to measure all parameters. American Kestrels had a higher incidence of he- matozoa infection than Red-tailed Hawks (45.5% vs. 26.3%). Haemoproteus occurred more frequently than Leukocytozoa in American Kestrels while the inverse was true for Red-tailed Hawks. Of 90 Amer- ican Kestrels, 38 were infected with Haemoproteus (42.2%) and four were infected with Leukocytozoa (4.4%); one bird was infected with both (1.1%). Of 38 Red-tailed Hawks, one was infected with Hae- moproteus (2.6%) and nine were infected with Leu- kocytozoa (23.7%). Several interspecific differences in hematologi- cal and plasma parameters were identified. Amer- ican Kestrels had lower white blood cell concentra- tion estimates and eosinophil count estimates than Red-tailed Hawks (Table 1). However, the mean heterophil/lymphocyte ratio estimate was 2.3 for both Red-tailed Hawks and American Kestrels. Plas- ma enzyme activity differences between Red-tailed hawks and American Kestrels were most notable for AP, AST and ACHE (Table 2) . Red-tailed Hawks had higher plasma concentrations of total protein and albumin than American Kestrels, but had low- er plasma concentrations of cholesterol, blood urea nitrogen and uric acid (Table 3). The mean June 1998 Hematology of California Raptors 165 Table 1. White blood cell concentration estimates and differential cell counts of wild Red-tailed Hawks (RTHA) and American Kestrels (AMKE). White blood cell concentration estimates are uncorrected for anemia and differential cell counts are parameter estimates. Species WBC EST^ 10^ (Cells/ p.1) HETERO 10^ (Cells/ ijlI) LYMPHO 10^ ( Cells/ |jl1) MONO ( Cells/ p.1) EOSINO (Cells/ |jtl) BASO (Cells/ p.1) RTHA Mean 16 9.0 5.6 214 1300 258 SEM 2.0 1.1 0.9 82 256 68 Range 4-48 1.4-26.2 1.0-22.6 0-1710 0-5520 0-1440 N 31 31 31 31 31 31 AMKE Mean 8.6 5.0 3.4 17 31 102 SEM 0.9 0.7 0.6 9.0 15 28 Range 1.7-17 1.1-13.6 0.4-12 0-150 0-300 0-450 N 21 21 21 21 21 21 P-value 0.001 0.004 0.05 0.02 0.0001 0.04 ® WBC EST is white blood cell concentration estimate; HETERO is heterophil; LYMPHO is lymphocite; MONO is monocyte; EOSINO is eosinophil; BASO is basophil. albumin/globulin ratios were 0.56 and 0,43 for Red-tailed Hawks and American Kestrels, respec- tively. American Kestrels had a higher plasma so- dium concentration than Red-tailed Hawks (Table 4). Discussion Cooper (1989) emphasized the importance of health monitoring for migrating raptors due to the stresses of competition, contact with diseased or contaminated prey, fatigue, reduced food intake and metabolism of body reserves. Therefore, he- matological and plasma parameters from wild rap- tors are essential tools for assessing the health and understanding the physiology of these birds. How- ever, detailed data are limited for wild raptors (Hunter and Powers 1980, Gessaman et al. 1986, Gonzalez and Hiraldo 1991, Lavin et al. 1992, Pow- ers et al. 1994). Hematozoa infection rates reported for Ameri- can Kestrels and Red-tailed Hawks from Colorado are nearly identical to those detected in wintering Table 2. Plasma enzyme activities of wild Red-tailed Hawks (RTHA) and American Kestrels (AMKE) . Species LDH^ (lU/L) AP (lU/L) ALT (lU/L) AST (lU/L) CK (lU/L) CHE (p.mol/ ml/ min) ACHE (p,mol/ml/ min) BCHE (p.mol/ml/ in) RTHA Mean 798 27 38 327 1948 0.66 0.14 0.52 SEM 107 1.6 4.1 17 429 0.03 0.01 0.02 Range 176-3284 12-46 2-138 182-761 332-16126 0.28-1.14 0.04-0.37 0.21-1.04 N 45 37 45 45 44 44 44 44 AMKE Mean 798 242 52 98 891 1.9 1.6 0.26 SEM 28 12 3.4 7.6 37 0.05 0.05 0.01 Range 520-1710 57-522 16-171 44-405 286-2078 0.78-3.41 0.61-3.04 0.10-0.49 N 64 60 60 64 64 86 86 86 P-value 0.9 0.0001 0.01 0.0001 0.02 0.0001 0.0001 0.0001 * LDH is lactate dehydrogenase; AP is alkaline phosphatase; ALT is alanine aminotransferase; AST is aspartate aminotransferase; CK is creatine kinase; CHE is total cholinesterase; ACHE is acetylcholinesterase; BCHE is butrylcholinesterase. 166 Stein et al. VoL. 32, No. 2 Table 3. Plasma biochemistry concentrations of wild Red-tailed Hawks (RTHA) and American Kestrels (AMKE). Species Tpa (g/dl) ALB (g/dl) GLOB (g/dl) CHOLEST (mg/dl) BUN (mg/dl) UA (mg/dl) GLUC (mg/dl) RTHA Mean 3.6 1.3 2.4 191 5.0 8.3 424 SEM 0.1 0.04 0.1 6.0 0.4 1.4 9.2 Range 2.2-5.S 0.7-2.0 1. 0-3.9 95-296 1-12 1.5-38.1 306-574 N 45 39 39 45 45 45 45 AMKE Mean 3.1 0.9 2.2 277 7.4 18.5 428 SEM 0.1 0.02 0.1 7.7 0.4 1.1 6.7 Range 2.2-6.2 0.4-1.4 1. 5-5.0 181-569 2-12 3-49.2 329-618 N 64 60 60 64 60 64 64 P-value 0.0002 0.0001 0.2 0.0001 0.0001 0.0001 0.8 ® TP is total protein; ALB is albumin; GLOB is globulin; CHOLEST is cholesterol; BUN is blood urea nitrogen; UA is uric acid; GLUC IS glucose. American Kestrels and Red-tailed Hawks in Cali- fornia (Stabler and Holt 1965). It is not clear whether hematozoa infection is associated with pathogenesis in birds (Sibley and Werner 1984, Ol- sen and Gaunt 1985). However, there were no out- ward indications of disease or emaciation observed in the birds reported on here. Total white blood cell counts or concentration estimates and white cell differential count refer- ence ranges are important to establish, since a shift in the proportions of white cell types, rather than a change in absolute numbers, may be the only consequence of disease (Campbell and Dein Table 4. Plasma trace element concentrations of wild Red-tailed Hawks (RTHA) and American Kestrels (AMKE). Species NA^ (MEQ/L) K (MEQ/L) CA (mg/dl) P (mg/dl) RTHA Mean 151 4.9 8.0 2.6 SEM 0.5 0.6 0.2 0.2 Range 144-158 2.1-15.3 4-10 0.9-5. 7 N 31 35 38 39 AMKE Mean 158 6.9 7.9 2.4 SEM 0.6 0.7 0.1 0.1 Range 155-160 2.4-24.6 6.4-11.7 0.4-5. 6 N 7 60 64 60 P-value 0.0001 0.03 0.7 0.5 ■* NA is sodium; K is potassium; CA is calcium; P is phosphorous. 1984). Response to stressors can also induce a shift in the proportions of white cell types in birds, par- ticularly lymphocytes, which decrease, and hetero- phils, which increase (Gross and Siegel 1983). In our study, the prolonged holding period (1-2 hr) prior to blood sample collection makes it unlikely that the effects of capture stress on heterophil and lymphocyte levels were avoided. This, together with the fact that white blood cell estimates were not corrected for anemia, should be taken into consid- eration in the interpretation of the white blood cell parameters reported here. The predominant leukocytes of wild Red-tailed Hawks and American Kestrels were heterophils and lymphocytes, accounting for 95% or more of the white blood cells, on average. The numbers of monocytes, eosinophils and basophils were low, though variable. The degree of variability observed is not unique to these two raptor species (Powers et al. 1994) or to wild raptors (Hernandez et al. 1990). This variability is most likely indicative of individual response to different immunological challenges and different stages of response. Eosin- ophils play a defensive role against parasites such as worms and protozoa, while monocytes are criti- cal in defense against intracellular parasites such as viruses and certain bacteria. The role of baso- phils is less well understood; however, they are in- volved in the early inflammatory response of aller- gic reaction (Maxwell 1993). The mean total white blood cell count and white cell differential counts reported by Hernandez et al. (1990) for captive June 1998 Hematology of California Raptors 167 Common Buzzard {B. buteo) were consistently low- er than those of wild Red-tailed Hawks, except for eosinophils and basophils. White blood cell esti- mates and heterophil count estimates of wild Red- tailed Hawks were two times higher, while lympho- cyte and monocyte count estimates were four times higher than those of captive Common Buzzards. These differences are consistent with heterophil and lymphocyte shifts associated with capture stress. Wild American Kestrels were more similar in these measures to captive Common Buzzard than were wild Red-tailed Hawks. Enzyme activities of wild Red-tailed Hawks and American Kestrels were similar to values previously reported for captive hawks and falcons. Wild Red- tailed Hawks had similar enzyme activities to cap- tive Red-tailed Hawks for AP, ALT and AST (Kollias and McLeish 1978) and to captive Common Buz- zards for LDH, ALT and AST (Hernandez et al. 1990) . CK activity was higher and more variable in wild Red-tailed Hawks than in captive Common Buzzards; this may he indicative of capture myopa- thy exhibited by the wild Red-tailed Hawks. Dab- bert and Powell (1993) reported elevated CK activ- ity as evidence of capture myopathy in Mallards (Anas platyrhynchos) sampled after a similar time duration as the raptors reported on here. Wild American Kestrels had higher AP activity, but sim- ilar LDH and AST activities compared with captive Peregrine Falcons (F. peregrinus tundrius) (Gee et al. 1981) . For most enzyme activities, wild birds had larger ranges than those reported in the literature for captive raptors. Presumably, this is largely due to natural variability in the condition of wild birds (e.g., due to differences in migratory status, time of capture, activity level, time since last meal and contact with pathogens and other stressors prior to capture) . Plasma concentrations of total protein, albumin, globulin, cholesterol, uric acid, calcium and phos- phorus of wild Red-tailed Hawks were similar to means and ranges reported for captive Common Buzzards (Ferrer et al. 1987, Garcia-Rodriguez et al. 1987b, Hernandez et al. 1990). Plasma concen- trations of total protein, albumin, blood urea ni- trogen, calcium and phosphorus of wild American Kestrels were also similar to means and ranges re- ported for captive Peregrine Falcons (Gee et al. 1981). Glucose concentration was lower for both captive Common Buzzards and Red-tailed Hawks than wild Red-tailed Hawks; similarly, glucose and uric acid concentrations were lower for captive Per- egrine Falcons than wild American Kestrels (Kol- lias and McLeish 1978, Gee et al. 1981, Hernandez et al. 1990). Garcia-Rodriguez et al. (1987a) re- ported increases in the concentrations of blood urea nitrogen, uric acid, glucose and cholesterol during long-term fasting of captive Common Buz- zards. An analogous situation may exist for winter- ing raptors, which experience relatively low prey densities and increased metabolic demands, and may account for the higher levels of these com- ponents found in our study. Comparison of Red-tailed Hawk and American Kestrel hematological parameters, enzyme activi- ties and other plasma constituents indicate impor- tant interspecific differences in hematological pro- files, knowledge of which is critical for use in health assessment. In an attempt to report data from healthy wild raptors, individuals that did not meet qualitative health criteria or were known to carry elevated OP residues were excluded from the analyses. However, since white blood cell estimates were not corrected for anemia and substantial time elapsed between capture and blood sampling, cap- ture stress may have influenced the heterophil and lymphocyte count estimates and CK activity values reported here. Furthermore, determining whether statistically significant differences are biologically important is not straightforward; however, a num- ber of the highly significant differences may reflect differences important for clinical consideration during health assessment, particularly AP, AST, ACHE activities and cholesterol concentration. Clearly, further studies are desirable for expansion and confirmation of values reported here for rap- tors wintering in California. In particular, investi- gations on factors contributing to variability in he- matological values such as sampling time, body condition, diet, season, geographical location and captive vs. wild status, are warranted. Acknowledgments This research was supported by the Almond Board of California, the Center for Ecological Health Research at U.C. Davis, and a National Institute of Environmental Health Sciences training grant (#ESO7059-15). The au- thors thank Dr. James Seiber and Michael McChesney for pesticide residue analyses; Nancy Ottum, Peter Bloom and Ed and Judy Henckel for raptor trapping assistance, and Dr. Scott Newman for helpful comments on the manuscript. Literature Cited Allain, C.C., L.S. PooN, C.S. Chan, W. Richmond and P.C. Fu. 1974. Enzymatic determination of total se- rum cholesterol. Clin. Chem. 20:470-475. 168 Stein et al. VoL. 32, No. 2 Baginski, E.S., S.S. Marie, W.L. Clarke and B. Zak. 1973. Direct microdetermination of serum calcium. Clin. Chim. Acta 46:46-54. Bergmeyer, H.U., G.N. Bowers, Jr., M. Horder and D.W. Moss. 1977. Provisional recommendations on IFCC methods for the measurement of catalytic concentrations of enzymes. Part II. IFCC method for aspartate aminotransferase. Clin. Chem. 23:887- 893. AND M. Horder. 1980. Provisional recommen- dations on IFCC methods for the measurement of catalytic concentrations of enzymes. Part III. IFCC method for alanine aminotransferase,/. Clin. Chem. Clin. Biochem. 18:521-525. Bessey, O.A., O.H. Lowry and M.J. Brock. 1946. A method for the rapid determination of alkaline phosphatase with five cubic millimeters of serum. / Biol. Chem. 164:321-329. Campbell, T.W. 1988. Avian hematology and cytology. Iowa State Univ. Press. Ames, lA U.S.A. AND F.J. Dein. 1984. Avian hematology: the ba- sics. Vet. Clin. North Am. Small Anim. Bract. 14:223- 248. Cooper, J.E. 1989. The importance of health monitor- ing of migrating raptors. Pages 39-42 in B.-U. Mey- burg and R.D. Chancellor [Eds.], Raptors in the modern world: proceedings of the III world con- ference on birds of prey and owls. World Working Group on Birds of Prey and Owls, London, U.K. Dabbert, C.B. and K.C. Powell. 1993. Serum enzymes as indicators of capture myopathy in Mallards {Anas platyrhynchos) . J. Wildl. Dis. 29:304-309. Doumas, B.T. and H.G. Biggs. 1972. Determination of serum albumin. Stand. Methods Clin. Chem. 7:175- 183. Ellman, G.L., K.D. Courtney, V. Andres and R.M. Featherstone. 1961. A new and rapid colorimetric determination of acetylcholinesterase activity. Bio- chem. Pharmacol. 7:88-95. Ferrer, M., T. Garcia-Rodriguez, J.C. Carillo andJ. Castroviejo. 1987. Hematocrit and blood chemis- try values in captive raptors ( Gyps fulvus, Buteo buteo, Milvus migrans, Aquila heliaca). Comp. Biochem. Phys- iol. 87A:1 123-1 127. Garcia-Rodriguez, T., M. Ferrer and P. Fores. 1987a. Metabolic responses of Buteo buteo to long-term fast- ing and refeeding. Comp. Biochem. Physiol. 87A:331— 386. , M. Ferrer, F. Recio and J. Castroviejo. 1987b. Circadian rhythms of determined blood chemistry values in buzzards and eagle owls. Comp. Biochem. Physiol. 88A:663-669. Gessaman, J.A., J.A. Johnson and S.W. Hoffman. 1986. Hematocrits and erythrocyte numbers for Cooper’s and Sharp-shinned Hawks. Condor 88:95-96. Gee, G.,J.W. Carpenter and G.L. Hensler. 1981. Spe- cies differences in hematologic values of captive cranes, geese, raptors, and quail. J. Wildl. Manage. 45:463-483. Gonzalez, J.L. and F. Hiraldo. 1991. Some hemato- logical data from Marsh Harriers {Circus aerugino- sus) in central Spain. Comp. Biochem. Physiol. lOOA: 735-737. Gross, W.B. and H.S. Siegel. 1983. Evaluation of the heterophil/lymphocyte ratio as a measure of stress in chickens. Avian Dis. 27:972-979. Hawkey, C.M. and T.B. Dennet. 1989. A colour atlas of comparative veterinary haematology: normal and abnormal blood cells in mammals, birds and reptiles. Wolfe Medical Publications, London, U.K. Hernandez, M., S. Martin and P. Fores. 1990. Clinical hematology and blood chemistry values for the Common Buzzard. / Raptor Res. 24:113-119. Hooper, M.J., P.J. Detrich, C.P. Weisskopf and B.W. Wilson. 1989. Organophosphorus insecticide ex- posure in hawks inhabiting orchards during winter dormant-spraying. Bull. Environ. Contam. Toxicol. 42: 651-659. Hunter, S.R. and L.R. Powers. 1980. Raptor hemato- crit values. Condor 82:226-227. Kollias, G.V. and I. McLeish. 1978. Effects of ketam- ine hydrochloride in Red-tailed Hawks {Buteo ja- maicensis) . II — biochemical and hematologic. Comp. Biochem. Physiol. 60C:21 1—213. Lavin, S., R. Cuenca, I. Marco, R. Velarde and L. ViNAS. 1992. Hematology and blood chemistry of the Marsh Harrier {Circus aeruginosus) . Comp. Bio- chem. Physiol. 103A:493-495. Layne, E. 1957. Spectrophotometric and turbidimetric methods for measuring proteins. Pages 447-453 in S.P. Colowick and N.O. Kaplan [Eds.], Methods in enzy- mology, Vol. 3. Academic Press, New York, NY U.S.A. Maxwell, M.H. 1993. Avian leucocyte response to stress. WorkPs Poult. Sd.J. 49:34-43. Merdes, H., W. Rittersdorf and W. Werner. 1985. Re- flotron® uric acid: evaluation of a new dry chemistry test for Reflotron.® /. Clin. Chem. Clin. Biochem. 28: 608-609. Olsen, G.H. and S.D. Gaunt. 1985. Effect of hemopro- tozoal infections on rehabilitation of wild raptors. /. Am. Vet. Med. Assoc. 187:1204-1205. Powers, L.V., M. Pokras, K. Rio, C. VrvERETTE and L. Goodrich. 1994. Hematology and occurrence of hemoparasites in migrating Sharp-shinned Hawks {Ac- cipiter striatus) during fall migration. /. Raptor Res. 28; 178-185. SAS Institute, Inc. 1988. SAS/STAT user’s guide: statis- tics. SAS Institute, Inc., Cary, NC U.S.A. Schmidt, F.H. 1961. Die enzymatische Bestimmung von Glucose und Fructose nebeneinander. Klin. Woch- enschr. 39:1244-1247. Sibley, L.D. and J.K. Werner. 1984. Susceptibility of Pe- kin and Muscovy Ducks to Haemoproteus nettionis. J. Wildl. Dis. 20:108-113. June 1998 Hematology of California Raptors 169 Stabler, R.M. and P.A, Holt. 1965. Hematozoa from Colorado birds. II. Falconiformes and Strigiformes. /. Parasitol. 51:927-928. SzASZ, G., W. Gruber and E. Bernt. 1976. Creatine ki- nase in serum: 1. Determination of optimum reaction conditions. Clin. Chem. 22:650—656. Talke, H, and G.E. Schubert. 1965. Enzymatische harn- stoffbestimmung in blut and serum im optischen test nach Warburg. Klin. Wochenschr. 43:174—175. Wilson, B.W., J.N. Seiber and D.M. Fry. 1993. Toxicity of sprays and exposure of Red-tailed Hawks in the Central Valley of California. Report to the California Department of Food and Agriculture. Wolf, RL., D. Williams, T. Tsudaka and L. Acosta. 1972. Pages 234-251 in Methods and techniques in clinical chemistry. Wiley-Interscience, New York, NY U.S.A. Woo, J. AND D.C. Cannon. 1984. Metabolic intermediates and inorganic ions. Pages 133-164 in J.B. Henry [Ed.], Clinical diagnosis and management by labora- tory methods. W.B. Saunders, Philadelphia, PA U.S.A. Received 6 June 1997; accepted 14 February 1998 /. Raptor Res. 32(2):170-174 © 1998 The Raptor Research Foundation, Inc. ACTIVITY OF INCUBATING FEMALE LONG-EARED OWLS AS MEASURED BY FLUCTUATIONS IN NEST TEMPERATURES Davorin Tome Institute of Biology, Vecna 111, 1 000 Ljubljana, Slovenia Abstract. — In 1993 and 1997, activities of incubating Long-eared Owls {Asio otus) were measured in the Ljubljansko barje area (Slovenia). Fluctuations in nest temperature were used to record movements of females in three nests. The owls were about two times less active during the day than during the night. Approximately 10% of the movements resulted from feeding activities and the remainder were due to different movements on the nest. There were two prominent peaks in activity at dawn and dusk with less pronounced bursts in activity every 2-4 hr. These latter bursts of activity may have coincided with increased activity of voles which were the main prey. Key Words: Long-eared Owl; Asio otus; nesting activity. Medicion de la actividad de incubacion de Asio otus a traves de variaciones de temperatura del nido Resumen. — En 1993 y 1997, la actividad de incubacion de Asio otus fue medida en el area de Ljubljansko baije (Slovenia). Las fluctuaciones de temperatura fueron utilizadas para registrar los movimientos de la hembra en tres nidos. Los buhos fueron dos veces menos activos durante el dia que en la noche. Aproximadamente 10% de los movimientos fueron el resultado de actividades de alimentacion y los restantes de otro tipo de movimientos en el nido. Hubo dos picos de actividad notables, al atardecer y al amanecer con sobresaltos menores cada 2-4 horas. Estos sobresaltos de actividad pueden coincidir con el aumento de actividad de los roedores que son su presa principal. [Traduccion de Cesar Marquez] The Long-eared Owl {Asio otus) is considered to be one of the most strictly nocturnal owls (Glutz and Bauer 1980), spending the day hidden in a dense vegetation (Mikkola 1983). Due to its cryptic way of life, little information has been published on its nesting behavior (von Wendland 1957, Arm- strong 1958, Wijnandts 1984, Korpimaki 1987, Craig et al. 1988). Unfortunately, methods of data collection have emphasized movements such as flight and have been unable to distinguish more subtle activities not involving flight. Since owls do not fly during the day, little is known about their daytime activity on their nests. My objective in this study was to estimate the tim- ing and pattern of nocturnal and diurnal activity of incubating female Long-eared Owls using fluc- tuations in nest temperature to indicate diurnal changes in nesting behavior. Methods I used the nest temperature fluctuation recording method (NTFR) to monitor behavioral changes on the nest. The method is based on continual recording of nest temperature. Any change in position by an incubating bird, such as repositioning on the eggs, moving around the nest, or absence from the nest, is recorded as a rapid drop in nest temperature. The duration of movements can be obtained using the time intervals between the be- ginning of a nest temperature drop and the beginning of a rise in nest temperature (Fig. 1, Part A). Likewise, temperature drops during movements can be measured using the difference between temperature before move- ments and the lowest temperature reached during move- ments (Fig. 1, Part B). NTFR also enables collection of data on the duration of incubation between two move- ments (periods of uninterrupted incubation, Fig. 1, Part C) and the timing of the movements. I studied the activity of long-ears during the incubation period at three nests located in the Ljubljansko barje area (south of Ljubljana, Slovenia) in 1993 and 1997. All three nests (Nl, N2, and N3) were in artificial nesting platforms, consisting of wooden, open-topped boxes filled with straw. Platforms were located in spruce {Picea sp.) trees near forest edges. Recording in Nl began on 22 April 1993 after the clutch was complete (A= 7 eggs) and lasted for 9 d until the first young hatched. Record- ing in N2 began on 3 May 1993, after the clutch was completed {N = 4 eggs), and lasted for 15 d until the first young hatched. Recording in N3 began on 30 April 1997, after the clutch was completed {N = 4 eggs) and lasted for 18 d until the first young hatched. Data from the first day of recording (habituation period) were not used for analysis. Recording was interrupted each day at 1100 H for 3 hr in Nl and for 2 hr in N2, to save data 170 June 1998 Activity of the Long-eared Owl 171 A C 45 u o 40 0 35 u Hi -p 30 fd u 25 0 a S 20 0 Eh 15 o o o o o O O o o o o O O o o o o o o eg m ■sT m O 1— H CM CO lO o 1 — t CM CO LO t — 1 1 — 1 \ — 1 \ — \ \ 1 \ — \ eg eg eg eg CM CM m CO CO CO CO CO CNJ CNJ CM csi CM CM CM CM CM CM CM CM CM CM eg CM CM CM hr of day Figure 1. Long-eared Owl movements during incubation as recorded using the nest temperature fluctuation method (NTFR) . Arrows indicate the beginning of movements. A represents the duration of a movement, B represents the temperature drop during a movement, and C represents the period of uninterrupted incubation. on the computer disk. The incubating owl was not dis- turbed during this process. In N3, data were collected once per day without interrupting the collection process. While recording, sunrise and sunset were at about 0445 and 1915 H, respectively. Nest temperatures were measured using a thermis- tor-type temperature probe (diameter = 2 mm, re- sponse time = 0.8 sec, Delta-T Co.) attached to a 30 m cable. A data logger (Delta-T Co.) was used for re- cording the data every 10 sec. The timing and dura- tion of movements obtained with the NTFR method are comparable between nests. Nevertheless, the size of temperature drops in nests differed individually be- cause of the number of eggs in nests, thickness of nest- ing material, influences of atmospheric conditions and microlocation of temperature probes. In Nl, for ex- ample, the owl was considered to move if the temper- ature decreased by at least 0.1°C or more, or when there was a sudden drop of at least 0.3°C. In N2 and N3, temperature drops considered for movement were 0.2 and 0.5°C, respectively. Results and Discussion Male Long-eared Owls either do not incubate and brood young or do so only occasionally and Table 1 . Movements of incubating Long-eared Owls at three nests in Slovenia. Nest 1 Nest 2 Nest 3 Total # days observed 8 14 17 39 # movements 413 690 1081 2184 # movments/24 hr X 52 49 64 56 maximum 73 76 80 80 minimum 42 29 42 29 SD 10.2 11.5 10.4 12.5 X movements/hr (day) 1.5 1.6 2.1 1.7 X movements/hr (night) 3.5 3.3 3.4 3.4 # movements lasting ^3 min/ 24 hr 3.5 6.3 4.1 4.8 # movements lasting ^5 min/24 hr 2.0 3.4 1.7 2.4 duration of longest movement* 10:20 12:20 11:30 duration of longest incubation'’ 04:50 05:22 03:13 max. temp, drop during movement 13.6 27.2 19.3 ® Duration in minutes and seconds. ^ Duration in hours and minutes. Temperature in °C. 172 Tome VoL. 32, No. 2 •H s to -p p: 0) e 0) > o fd > u CD -P -H ip O 1.8 1 . 6 1-4 hr before or after sunset c •H to I — I to ■p a (U g 0) > 0 g I — I (TS > u (U 4-> o "H Mh o * 1.8 hr before or after sunrise Figure 2. Activity pattern of Long-eared Owls during incubation. Data are averaged from three nests. Bars represent movement frequencies in 15 min intervals and lines represent three point moving averages. Data are shifted such that sunrise and sunset activity (denoted by 0) match between individual nests. for very short periods (Glue 1977, Glutz and Bauer 1980, Wijnandts 1984, Marks et al. 1994, Scott 1997). Therefore, I assumed that movements de- tected with the NTFR method represented the ac- tivities of only female owls. Altogether, I recorded a total of 2184 movements at the three nests over 39 days (Table 1). The duration of movements (x^ = 140, df = 22, P < 0.001) and the number of movements per day (F = 7.6, P <; 0.01) were dif- ferent among the nests. Movements per hour were approximately twice as frequent during the night than during the day with movements lasting ^3 min being nearly 6 times more frequent during the night. On average, females made only 2-3 movements in a 24-hr pe- riod that lasted longer than 5 min. Almost all oc- curred during the night with the longest lasting for 12.33 min. Female owls spent on average 170 sec/ hr (67 min per day) not incubating eggs. This was about twice as long as reported by Wijnandts (1984) and was probably due to differences in data collection and inclusion of numerous movements on the nest that I was able to detect using NTFR. The longest uninterrupted incubation period last- ed for 5.37 hr and the greatest temperature drop during a movement was 27.2°C (Table 1). NTFR methods cannot identify the type of be- havior that results in temperature changes in nests unless they are coupled with direct observations. Because von Wendland (1957) and Armstrong (1958) observed incubating female Long-eared Owls leaving their nests for 5-10 min during the June 1998 Activity of the Long-eared Owl 173 evening to hunt and Wijnandts (1984) found that female Long-eared Owl movements during incu- bation consisted of an average of 2.8 flights from nests per night, with each flight lasting on average for 11.5 min, I concluded that 2-3 absences of fe- male owls from nests in my study >5 min in du- ration (about 5% of all movements) were associ- ated with female hunting behavior. Another 5% of female movements apparently in- volved food deliveries hy males to nests. Female owls move an average of 1.8-2. 8 times per night (Wijnandts 1984, Marks et al. 1994) when males deliver prey items. DeLong (1982: in Marks et al. 1994) reported that female owls turn eggs 5—12 times per night. This explains another 10-20% of the movements I recorded using NTFR. The remaining 70-80% of the movements were probably simply due to re- positioning movements of females on their nests. Daily activity patterns were similar at all three nests (x^ 21, df = 190, P > 0.05). There were two prominent peaks in activity approximately 1 hr before sunrise (at about 0400 H, Fig. 1) and 1 hr after sunset (at about 2000 H). Similar peaks in activity have also been reported for Long-eared owls by Wijnandts (1984) and Korpimaki (1987). In N1 and N2, less pronounced peaks appeared about 5 hrs after sunset (at midnight) , while in N3 two peaks were observed 3 hr after sunset (2200- 2300 H) and 3.5 hr before sunrise (0100-0200 H). During the day, I observed less pronounced bursts in activity every 3^ hr. These bursts of ac- tivity occurred in all three nests at about 2-3 and 5-6 hr before sunset and 2-3 and 4—5 hr after sun- rise. A similar activity pattern with peaks at 2-5, 5, and 7.5 hr after sunrise has been shown in diurnal raptors in the Netherlands (Raptor Group 1982). These periods of peak activity coincide with the 2— 4 hr activity periods of voles which are the main prey of Long-eared Owls (Davis 1933, Lehman and Sommersberg 1980, Raptor Group 1982, and Tam- arin 1985). Voles are also the main prey of Long- eared Owls in Slovenia (Tome 1991, 1994). There- fore, vole activity may influence the activity of the owls since hunting is more successful when prey are active (Halle 1993). There are several advantages to using the NTFR method of studying avian behavior as opposed to direct or indirect observations using video systems, radiotelemetry, automatic recording of nest visits using light traps, and systems for weighing nests. Advantages include the fact that NTFR is inexpen- sive, it uses low energy consumption, it is easy and quick with installation within 15 min and minimal disturbance during the operation. In addition, NTFR results include all the movements of females on nests and not just movements occurring when they leave or return. NTFR is also useful at night as well as during the day. Limitations of NTFR in- clude the inability to determine the types of move- ments that cause temperature changes in nests and if and when mates swap incubation duties during the incubation period. Acknowledgments I would like to thank Drs. D.H. Thomas and H. Pie- tiainen for comments on an early version of the manu- script. This study was partially supported by the Slovenian Ministry of Science and Technology (Grant Sl-7858-0105- 96). Literature Cited Armstrong, W.H. 1958. Nesting and food habits of the Long-eared Owl in Michigan. Publ. Mus. Mich. State Univ., Biol. Ser. 1:61-96. Craig, E.H., T.H. Craig and L.R. Powers. 1988. Activity patterns and home-range use of nesting Long-eared Owls. Wilson Bull. 100:204-213. Davis, D.H.S. 1933. Rhythmic activity in short-tailed vole, Microtus. J. Anim. Ecol. 2:232-238. Glue, D.E. 1977. Breeding biology of the Long-eared Owl. Br. Birds 70:318-331. Glutz von Blotzheim, U.N. and K.M. Bauer. 1980 Handbuch der Vogel Mitteleuropas, Vol. 9. Akadem- ische Verlagsgesellschaft, Wiesbaden, Germany. Halle, S. 1993. Diel pattern of predation risk in micro- tine rodents. Oikos 68:510-518. Korpimaki, E. 1987. Dietary shifts, niche relationship and reproductive output of coexisting kestrels and Long- eared Owls. Oecologia 74:277-285. Lehman, U. and C.W. Sommersberg. 1980. Activity pat- terns of the common vole Microtus arvalis — automatic recording of behaviour in an enclosure. Oecologia 47- 61-75. Marks, J.S., D.L. Evans and D.W. Holt. 1994. Long- eared Owl. The birds of North America, No. 133, A. Poole and F. Gill, [Eds.]. The Academy of Natural Sciences, Philadelphia, PA and The American Orni- thologists’ Union, Washington DC U.S.A. Mikkola, H. 1983. Owls of Europe. T. & A.D. Poyser, Staf- fordshire, U.K. Raptor Group. 1982. Timing of vole hunting in aerial predators. Mammal Rev. 12:169-181. Scott, D. 1997. The Long-eared Owl. The Hawk and Owl Trust. London, U.K. Tamarin, R.H. [Ed.]. 1985. Biology of new world Microtus. Am. Soc. Mammal., Spec. Publ. 8. Stillwater, OK U.S.A. 174 Tome VoL. 32, No. 2 Tome, D. 1991. Diet of the Long-eared Owl Asia otus in Yugoslavia. Omis Fenn. 68:114—118. . 1994. Diet composition of the Long-eared Owl in central Slovenia: seasonal variation in prey use. J. Raptor Res. 28:253—258. VON Wendland, V. 1957. Aufzeichnungen uber Brutbio- logie und Verhalten der Waldohreule {Asio otus). J. Omithol. 98:241-261. WijNANDTS, H. 1984. Ecological energetics of the Long- eared Owl {Asio otus). Ardea 72:1—92. Received 23 January 1997; accepted 2 February 1998 Short Communications J. Raptor Res. 32(2):175-177 © 1998 The Raptor Research Foundation, Ine. Breeding Biology and Nestling Diet of the Great Black-Hawk Nathaniel E, Seavy 17142 Lemolo Shr Dr. N.E., Poulsbo, WA 98370 U.S.A. Richard P. Gerhardt 341 N.E. Chestnut St., Madras, OR 97741 U.S.A. Key Words: Great Black-Hawk, Buteogallus urubitinga; breeding biology; diet, Peten; Guatemala. The Great Black-Hawk {Buteogallus urubitinga) rang- es from Mexico south to eastern Bolivia, Paraguay and northern Argentina, inhabiting coastal lowlands and foothills (Brown and Amadon 1968). The few accounts describing its breeding biology have been brief and at times contradictory (Grossman and Hamlet 1964, Smithe 1966, Brown and Amadon 1968, ffrench 1976, Mader 1981). Based primarily on isolated observations of hunt- ing and prey remains collected beneath roosts, a wide variety of prey items has been recorded, including inver- tebrates, fish, frogs, reptiles, birds, mammals and carrion (Dickey and van Rossem 1938, Lowery and Dalquest 1951, Haverschmidt 1962, Wetmore 1965, Mader 1981, Olmos 1990, Lewis and Timm 1991). We studied Great Black-Hawks in Tikal National Park, Peten, Guatemala as part of The Peregrine Fund’s “Maya Project.” Two nests were located and studied during 1991, and information collected on food habits was pub- lished by Gerhardt et al. (1993). We continued to study the species during 1993-94. Here, we present informa- tion on its breeding biology, including nesting phenolo- gy, reproductive success and nest descriptions, and ad- ditional dietary data. Study Area and Methods Tikal National Park covers 576 km^ in N.E. Guatemala (17°13'N, 89°38'W). Elevation averages 200-250 m, to- pography is gently rolling, and the climate is tropical, with annual rainfall of about 135 cm. The rainy season begins in June or July, with the highest rainfall in Sep- tember, and a pronounced dry season occurs from Feb- ruary until June or July. Vegetation, climate and land use patterns of the Tikal area were described by Schulze (1992). Observations of courtship behavior or of hawks carry- ing nest material or prey led to the eventual location of nests. After they were found, nests were checked every 2-3 d to record nesting phenology. During all years we recorded nest size (diameter and depth) and situation, and described nest trees. Observations of prey deliveries to nests were made with binoculars from observation platforms constructed in trees about 35 m from nests. We climbed to nests weekly to weigh and measure nest- lings in 1991; in 1993 and 1994 we avoided climbing to nests, except to verify some clutch sizes, until after fledg- ing. Additional information on clutch size, nesting phe- nology and nests was obtained from egg-set data records from published accounts, the Western Foundation of Ver- tebrate Zoology (WFVZ), and the Delaware Museum of Natural History (DMNH). Results and Discussion The earliest we observed nesting activity of Great Black-Hawks was 23 March 1994 when a pair began to copulate and build a nest. Egg-laying dates were known for two nests, 16-17 April 1991 and 4 May 1994. An in- cubation period of 40 ± 2 d was recorded for one nest Based on this interval, we estimated an additional laying date as 25 March 1994. Known hatch dates were 27 May 1991 and 4 May 1994. One nestling fledged from the nest on 28 June 1994, 55 d after hatching. Another nestling had not yet fledged at the age of 63 d when observations were concluded on 27 July 1991. One nest, found on 6 July 1994, contained a nestling estimated to be 60 d old that had not yet fledged. We know of no other accounts of the duration of incubation or nestling periods for Great Black-Hawks in the wild. For the Common Black- Hawk {Buteogallus anthracinus) , slighdy shorter incuba- tion (37-39 d) and nestling (43-52 d) periods have been reported (Schnell 1994). Substantial geographic varia- tion in Great Black-Hawk breeding phenology may exist. At Tikal young fledged at the onset of the rainy season. 175 176 Short Communications VoL. 32, No. 2 In Surinam (Haverschmidt 1962), Trinidad (ffrench 1976), and Venezuela (Mader 1981), however, the species has been observed incubating during, or at the onset of, the rainy season. Egg-laying may occur as early as Decem- ber in El Salvador (Brown and Amadon 1968) and as late as June in Tamaulipas, Mexico (Martin et al. 1954). Eggs in DMNH and WFVZ were collected in Trinidad and Ven- ezuela during April and May and in Argentina from Au- gust-January. Our observations of one juvenile each from two 1994 nests indicated an extended dependency period during which food was provided by the adults. We observed a prey delivery to one Juvenile almost 8 mo after fledging, and another juvenile was consistently found within 500 m of the nest tree up to an age of 12 mo, when obser- vations were discontinued. Such an extended dependen- cy period is supported by Mader (1981) who reported an immature Great Black-Hawk, ca. 7 mo postfledging, perched next to an adult, begging for food. In contrast, the Common Black-Hawk is believed to reach indepen- dence about 2 mo after fledging (Schnell 1994). Post- fledging dependency periods of from 7 mos to more than a year have been documented for several other neotrop- ical raptors, including Ornate Hawk-Eagles (Spizaetus or- natus; Madrid et al. 1992), Black Hawk-Eagles (S. Tyran- nus] D, Whitacre and J. Lopez pers. comm.). Harpy Ea- gles {Harpia harpyja; E. Alvarez pers. comm,). Crested Ea- gles {Morphnus guianemis; D. Whitacre pers. comm.), and Gray-backed Hawks {Leucopternis occidentalis; Vargas 1995). All nests we located were in “bajo” and “transitional” forest types, which are at the lower end of the topograph- ic gradient. Seasonally flooded, these forest types are characterized by dense understory vegetation and a can- opy height that is generally low, but broken by large, iso- lated, emergent trees (Schulze 1992). Nests were built within live, emergent trees {Swietenia macrophylla, Bucida buceras, Platymiscium spp., Ceiba pentandra and Lonchocar- pus castilloi) , which had minimal crown contact with the surrounding forest canopy. Nest trees {N = 7) had a mean height of 29 m (SD = 4 ni, range = 22-35 m) and a mean DBH of 120 cm (SD = 81 cm, range = 66-300 cm). Nests were at a mean height of 25 m (SD = 5 m, V = 7, range = 20-31 m) and supported by branches and/or vine tangles. Nests were constructed of dry sticks and lined with green, leafy twigs. Throughout incubation and nestling periods, adults added fresh, leafy twigs to nests. Nests were flat throughout the nesting period, with only one having a measurable depression, 10 cm in depth. Smithe (1966), Brown and Amadon (1968), and egg records from Venezuela and Trinidad (DMNH and WFVZ) de- scribed nests as “deeply-cupped,” but our observations of flat nests concur with the description by Grossman and Hamlet (1964). Considering only nests that contained clutches, the mean nest diameter was 83 cm (SD = 21 cm, N = 4, range = 53-100 cm), and mean external height was 56 cm (SD = 12 cm, N = 3, range = 42-65 cm). We found eight nests that Great Black-Hawks were con- structing or embellishing. Four did not contain eggs and have been excluded from analysis because we were un- able to determine whether these pairs nested at alternate sites. Of the four nests in which eggs were laid, two failed during incubation, after 10 and 37 d, respectively. Young hatched at two nests; one was killed by an unknown pred- ator 6.5 wk after hatching and the other fledged. In each of three nests at which we documented clutch size, only one egg was laid. Three additional nests, dis- covered after hatching, contained a single nestling each. We documented 27 Great Black-Hawk egg sets from Ven- ezuela (7 sets), Argentina (18 sets) and Trinidad (2 sets) (Swann 1923, Norris 1926, WFVZ, DMNH). All clutches contained a single egg. To our knowledge the only evidence of a larger clutch size is an observation of a “family group” of four individuals in Mexico (Martin et al. 1954). Including the 106 prey of Gerhardt et al. (1993), stud- ies of Great Black-Hawks in Tikal have identified 126 prey items at least to class. Prey delivered to nestlings have included 41 lizards (32.5% of identified prey), 34 snakes (27.0%), 24 mammals (19.1%), 16 birds (12.7%), eight anurans (6.4%) and three insects (2.4%). Basilisk lizards {Basiliscus vittatus) have made up a large percentage (70.7%) of all lizards delivered. Other lizards in the diet include arboreal genera, including Norops (formerly Ano- lis) . Snakes include both arboreal and terrestrial as well as venomous and nonvenomous species. Boa and Oxybelis have been the snake genera most commonly observed in the diet. Mammals delivered to the nest have included eight bats (6.4% of identified prey), nine rodents (7.1%) and two marsupials (1.6%). Bats are medium-sized spe- cies, about 30 g in mass. Although we were unable to identify bats beyond order during nest observations, Ar- tibeus spp. remains were collected on several occasions from nests. Rodents have included squirrels {Sciurus spp.) and unidentified cricetids. The two small marsupi- als were probably mouse-opossums {Marmosa mexicanus) . Birds in the diet have ranged in size from an oriole {Ic- terus spp.) to a Pale-billed Woodpecker {Campephilus gua- temalensis) . Birds most commonly observed as prey items are medium-sized Columba spp. and Clay-colored Robins ( Turdus grayi) . Anurans and insects were infrequently de- livered to nests. The variety of prey delivered to the nests and the wide range of prey items and foraging habits reported in the literature indicate that Great Black-Hawks are dietary generalists and opportunists, able to exploit a large di- versity of prey types and hunting situations. Midday de- liveries of nocturnal mammalian prey and reports of pre- dation upon nestlings (Lewis and Timm 1991) and eggs (Brown and Amadon 1968, Olmos 1990) suggest that Great Black-Hawks may invest time in searching for vul- nerable and easily captured prey. Insects and anurans June 1998 Short Communications 177 may be captured more frequently than our observations suggest; adults may eat most such small prey upon cap>- ture rather than transporting them to nests. Although fish and crustaceans are often mentioned as important prey items (Grossman and Hamlet 1964, Brown and Amadon 1968), we observed no indication that Great Black-Hawks at Tikal utilized such aquatic prey. Great Black-Hawks exhibit a conservative breeding bi- ology that is characteristic of many large raptors, especial- ly in the tropics (Newton 1979); clutch size is small and young appear to have a protracted dependency period. Additionally, the fact that the species requires several years to obtain full adult plumage (Howell and Webb 1995, R. Gerhardt unpubl. data) suggests that there is a substantial population of nonbreeding subadults. Resumen. — Estudiamos la biologia reproductiva de Buteo- gallus urubitinga en el parque Nadonal Tikal, Peten, Gua- temala, durante 1991, 1993 y 1994. Un solo periodo de incubacion fue de 40 (±2 dias); el pichon emplumo a la edad de 55 dias. Los nidos son construidos con ramas y palos grandes en arboles emergentes. El tamano de la nidada de 3 nidos en Tikal, 2 en Trinidad, 7 en Vene- zuela y 18 en Argentina fue de un huevo por nido. De las 126 presas identificadas en tres nidos la mayoria fue- ron reptiles (59.5%), mamiferos (19.1%) y aves (12.7%). [Traduccion de Cesar Marquez] Acknowledgments This is a contribution of the “Maya Project,” a conser- vation research effort of The Peregrine Eund. Einancial support was provided by Robert Berry, Crystal Channel Foundation, Fanwood Foundation, Gold Family Founda- tion, KENNETECH/U.S. Windpower, the John D. and Catherine T. MacArthur Foundation, Norcross Founda- tion, Hank and Wendy Paulson, Pew Charitable Trusts, Andres Sada, Joe and Flinda Terteling, and the U.S. Agency for International Development. We are grateful to the curatorial staff at the Delaware Museum of Natural History and the Western Foundation of Vertebrate Zo- ology for providing data on egg clutches. Lloyd Kiff, Carl Marti, Dave Whitacre, Rich Glinski, and an anonymous reviewer provided helpful comments on the manuscript. Literature Cited Brown, L. and D. Amadon. 1968. Eagles, hawks, and fal- cons of the world. McGraw-Hill Book Go., New York, NYU.S.A. Dickey, D.R. and A.J. van Rossem. 1938. The birds of El Salvador. Zool. Ser., Vol. 23. Field Mus. Nat. Hist., Publ. No. 406. Chicago, IL U.S.A. ffrench. 1976. A guide to the birds of Trinidad and To- bago. Harrowood, Valley Forge, PA U.S.A. Gerhardt, R.P., P.M. Harris and M.A. Vasquez M. 1993. Food habits of nesting Great Black-Hawks in Tikal Na- tional Park, Guatemala. Biotropica 25:349-352. Grossman, M.L. and J. Hamlet. 1964. Birds of prey of the world. Crown Publishers, New York, NY U.S.A. Haverschmidt, F. 1962. Notes on the feeding habits and food of some hawks in Surinam. Cowrfor 64:154—158 Howell, S.N.G. and S. Webb. 1995. A guide to the birds of Mexico and northern Central America. Oxford Univ. Press, New York, NY U.S.A. Lewis, S.E. and R.M. Timm. 1991. Predation on nestling Bare-throated Tiger-Herons by a Great Black-Hawk Ornitologia Neotropical 2:37. Lowery, G.H., Jr. and W.W. Dalquest. 1951. Birds from the State of Veracruz, Mexico. Univ. of Kansas Publ , Mus. Nat. Hist., Lawrence, KS U.S.A. Mader, W.J. 1981. Notes on nesting raptors of the llanos of Venezuela. Condor 83:48-51 . Madrid M., H.D., R.A. Madrid M., J.R. Cruz E., J.L. Cordova A., M.C. Rivera, W.E. Martinez A. and A.R Caal. 1992. Behavior and breeding biology of the Or- nate Hawk-Eagle. Pages 179-191 mD.F. Whitacre and R.K. Thorstrom [Eds.], Maya project progress report V. The Peregrine Fund, Boise, ID U.S. A. Martin, P.S., C.R. Robbins and W.B. Heed. 1954. Birds and biogeography of the Sierra de Tamaulipas, an iso- lated pine-oak habitat. Wilson Bull. 66:38-57. Newton, I. 1979. Population ecology of raptors. Buteo Books, Vermillion, SD U.S.A. Norris, J.R, Jr. 1926. A catalogue of sets of Accipitres’ eggs in the collection of Joseph Parker Norris, Jr , Philadelphia, PA U.S.A. Oologists’ Rec. 6:25-41. Olmos, F. 1990. Nest predation of Plumbeous Ibis by ca- puchin monkeys and Greater Black-Hawk. Wilson Bull. 102:169-170. Schnell, J.H. 1994, Common Black-Hawk Buteogallus an- thracinus. In A. Poole and F. Gill [Eds.], The birds of North America, No. 122, Acad. Nat. Sci., Philadelphia, PA and Am. Ornithologists’ Union, Washington, DC U.S.A. Schulze, M. 1992. A preliminary description of woody plant communities of Tikal National Park. Pages 53— 62 in D.F. Whitacre and R.K. Thorstrom [Eds.], Maya project progress report V. The Peregrine Fund, Boise, ID U.S.A. Smithe, FB. 1966. The birds of Tikal. The Natural His- tory Press, Garden City, NY U.S.A. Swann, H.K. 1923. Notes on the Gordon collection of eggs of the Accipiters. Oologists’ Rec. 3:25-30. Vargas, H. 1995. Food habits, breeding biology, and sta- tus of the Gray-backed Hawk {Leucopternis occidentalis) in western Ecuador. M.S. thesis, Boise State Univ., Boi- se, ID U.S.A. Wetmore, a. 1965, The birds of the Republic of Panama, Part 1. Smiths. Misc. Coll., Vol. 150. Washington, DC U.S.A. Received 17 May 1997; accepted 8 February 1998 178 Short Communications VoL. 32, No. 2 J Raptor Res. 32(2): 178-1 80 © 1998 The Raptor Research Foundation, Inc. A Comparison of Raptor Use of Reclaimed Surface Mines and Agricultural Habitats IN Pennsylvania Richard H. Yahner and Ronald W. Rohrbaugh, Jr.^ School of Forest Resources, The Pennsylvania State University, University Park, PA 16802 US. A. Key Words: American Kestrel, Red-tailed Hawk; Rough-leg- ged Hawk; Northern Harrier, Buteo jamaicensis; Buteo la- gopus; Circus cyaneus; Falco sparverius; Pennsylvania; re- claimed surface mine. Although agricultural lands are decreasing in Penn- sylvania due to farm abandonment and subsequent re- forestation (Litvaitis 1993), over 300 surface mines av- eraging 157 ha in size occur in northcentral and north- western regions of the state (Energy Information Ad- ministration 1989). If managed properly, reclaimed surface mines have the potential to provide a signifi- cant amount of breeding habitat for grassland raptors such as American Kestrels {Falco sparverius) , Northern Harriers {Circus cyaneus), and Red-tailed Hawks {Buteo jamaicensis). Current procedures for reclaiming sur- face mines in Pennsylvania result in open grasslands dominated by herbaceous plant species interspersed with some woody plants. Reclaimed surface mines, however, are a relatively new habitat type in the east- ern U.S. and little is known about their value as breed- ing habitat for raptors (Yahner and Rohrbaugh 1996). Our objective was to compare spring abundance of di- urnal raptors associated with reclaimed surface mines and agricultural habitats in two geographic regions of northern Penn.sylvania. Study Area and Methods We selected two counties each in the northwestern (NW; Clarion and Butler Counties) and the northcen- tral (NC; Centre and Clearheld Counties) geographic regions of Pennsylvania. Clarion and Clearfield Coun- ties represented areas with abundant reclaimed sur- face mine habitat; conversely, Butler and Centre Coun- ties contained considerable agricultural habitat (En- ergy Information Administration 1989, Rohrbaugh and Yahner 1996, Yahner and Rohrbaugh 1996). Re- claimed surface mines were dominated by herbaceous plant species, such as fescue {Festuca spp.), orchard grass {Dactylis spp.), timothy {Phleum spp.), red top {Agrostis spp.), bird’s-foot trefoil {Lotus spp.), clover {Trifolium spp., Melilotus spp.) , and goldenrod {Solidago spp.) (see details in Yahner and Rohrbaugh 1996). De- ‘ Present address: Cornell Laboratory of Ornithology, 159 Sapsucker Woods Road, Ithaca, NY 14850 U.S. A. ciduous and coniferous trees, such as black locust {Ro- binia pseudoacacia) and Austrian pine {Pinus nigra), were planted on the mines in small plantations (<5 ha). In addition, reclaimed mines were associated with numerous human-created wetlands (e.g., cattail \ Typha spp.] marshes) that were designed to leach metals from water and soil. Agricultural habitat was characterized by crop spe- cies, such as corn (Maze spp.), grain {Triticum spp, and Hordeum spp.), soybeans {Glycine max), alfalfa {Medi- cago spp.), and mixed-herbaceous species used for hay production. Livestock pastures and confined feeding areas also were common throughout the agricultural survey routes. These agricultural areas were inter- spersed with fencerows, woodlots, and riparian forests We chose 10 road survey routes per geographic re- gion, giving five survey routes per habitat type (re- claimed surface mines or agricultural habitat) in each region. Routes were established by initially identifying unimproved and light-duty roads on 16 USGS topo- graphic quadrangles (1:24 000 scale) that traversed ei- ther reclaimed surface mine or agricultural habitats in the four counties. We then drove along all 1-km sec- tions of potential routes to determine if they were a suitable part of a survey route. A suitable section of a route was defined as one with relatively low rates of vehicular trafhc, characterized by at least 150 m of open habitat (surface mine or agricultural) perpen- dicular and adjacent to the road edge that extended at least 1 km either on one or both sides of the road, and devoid of developed areas (e.g., urbanization) or active surface mine operations. The mean length of survey routes was 13 km (range = 8-18 km); the min- imum distance separating survey routes was arbitrarily designated as 3 km (Yahner and Rohrbaugh 1996). We established survey stations along each route at 0.8-km intervals; the mean number of survey stations per route was 16 (range = 10-20). Surveys were con- ducted by driving a vehicle along the route at slow speeds (16-40 km/hr) while looking for perched or flying raptors. In addition, we stopped the vehicle for 5 min at each station to scan for raptors (Yahner and Rohrbaugh 1996). We recorded species, sex, age, lo- cation, distance (m) at initial detection and behavior at initial sighting (e.g., flying or perched) of each rap- tor observed during surveys. We also noted time of day and weather conditions for each raptor sighting. Spring surveys were conducted twice per month dur- ing April-June in both 1993 and 1994, yielding a total of 240 spring surveys. Two survey routes were driven June 1998 Short Communications 179 Table 1. Number of Red-tailed Hawks, American Kestrels and Northern Harriers observed in reclaimed surface mines (SM) and agricultural (AG) habitat types of northwestern and northcentral Pennsylvania during spring. Species Habitat Region Northwestern Northcentral Total Type No. No./Hr No. No./Hr No. No./Hr Red-tailed Hawk SM 61 0.94 19 0.32 80 0.64 AG 82 1.02 79 1.04 161 1.03 Total 143 0.99 98 0.72 241 0.86 American Kestrel SM 49 0.75 27 0.45 76 0.61 AG 33 0.41 39 0.51 72 0.46 Total 82 0.56 66 0.489 148 0.53 Northern Harrier SM 14 0.21 5 0.08 19 0.15 AG 1 0.01 3 0.04 4 0.03 Total 15 0.10 8 0.06 23 0.08 per day, one each in the morning and afternoon; morning and afternoon surveys were initiated between 0700-0900 H and 1300-1500 H, respectively. Because of logistical constraints, all 10 routes in a given geo- graphic region were surveyed during the same 5 days; furthermore, on a given day, the same habitat type was surveyed, but habitat type surveyed was alternated from one day to the next. Also, a given survey route was driven an equal number of times in morning and afternoon. We stratified the number of sightings of each raptor species into four habitat-region types, including those in surface mines and agricultural habitats. Data were pooled from both springs to give a better measure of habitat-use patterns (Rice et al. 1984) and because numbers of sightings per species did not vary between years (Yahner and Rohrbaugh 1996). Numbers of rap- tor sightings per type were converted to number per hour, based on the amount of stationary survey time spent within each habitat and region. This enabled us to compare observation rates (no. /hr) among habitat- region types. We compared observed vs. expected numbers of sightings of each raptor species among the four habi- tat-region types using G-tests for goodness-of-fit (Sokal and Rohlf 1995). If a significant difference occurred among the four types, then a G-test for goodness-of-fit was used in the habitat-region type of interest. Ex- pected numbers of sightings for a given species were calculated by multiplying the proportion of stationary survey time per type by the total observed numbers of sightings for that species. Results and Discussion We observed 412 raptors during our two spring sur- veys including 241 (58%) Red-tailed Hawks, 148 (36%) American Kestrels, and 23 (6%) Northern Harriers (Table 1). In addition, we noted 10 Cooper’s Hawks {Accipiter cooperii), three Ospreys (Pandion haliaetus), two Broad-winged Hawks {Buteo platypterus), two Red- shouldered Hawks {Buteo lineatus), one Sharp-shinned Hawk {Accipiter striatus), and one Northern Goshawk {Accipiter gentilis) . The observed number of Red-tailed hawks, Ameri- can Kestrels, and Northern Harriers each differed sig- nificantly from expected among the four habitat-re- gion types (G^ 11.1, df= 3, P < 0.001 ; Table 1 ) . Red- tails were seen less often than expected at reclaimed surface mines in the northcentral region {N =19, 0.32 hawks/hr; G = 36.6, df = 1, P< 0.001) but were more common than expected in each of the other types {N = 61-82, 0.94-1.02 hawks/hr; G > 5.4, df = 1, P < 0. 025). Kestrels occurred more often than expected at reclaimed surface mines in the northwestern region {N = 49, 0.75 hawks/hr; G = 15.3, df = 1, P > 0.005) and less than expected in agricultural habitats in the northwestern region {N = 33, 0.41 hawks/hr; G = 5.8, df = 1, P < 0.025). In the northwestern region, more harriers than expected were observed at reclaimed surface mines {N = 14, 0.21 hawks/hr; G = 16.1, df = 1, P < 0.001), but fewer harriers than expected were seen in agricultural habitats {N = 1, 0.01 hawks/hr; G = 9.4, df = 1, P< 0.005). Reclaimed surface mines seemed to be preferred habitat for spring populations of raptors, particularly those in the northwestern region of Pennsylvania. Per- haps this trend occurred because surface mines in the northwestern region tended to be larger and more abundant, thereby providing more breeding habitat (Yahner and Rohrbaugh 1996). Probable and con- firmed breeding attempts of Northern Harriers, for example, have been shown to be significantly higher in regions of Pennsylvania containing abundant re- claimed surface mines (Rohrbaugh and Yahner 1996). Previous studies conducted in other regions of Penn- sylvania, which are virtually devoid of reclaimed sur- face mines, have noted that breeding Red-tailed Hawks and American Kestrels prefer agricultural lands as 180 Short Communications VoL. 32, No. 2 breeding habitat (Bednarz 1992, Rohrbaugh 1994, Rohrbaugh and Yahner 1997). Thus, our study indi- cated that grasslands created by surface mine recla- mation in Pennsylvania may also serve as valuable breeding habitat for open-country raptors (see also Rohrbaugh and Yahner 1996). Small mammals are major prey of the raptors ob- served in our study area (see Weller et al. 1955, Bart 1977). Meadow voles {Microtus pennsylvanicus) are quite abundant on reclaimed surface mines in the southcentral region of Pennsylvania (Alberici et al. 1989; Yahner and Rohrbaugh unpubl. data). Dormant or fallow agricultural fields often contained residual plant material and seeds, which probably provided cov- er and food resources for small mammal populations during winter (Yahner and Rohrbaugh 1996). While they are important areas for grassland raptors, we must caution that larger reclaimed surface mines (>100 ha) may have a negative effect on forest-depen- dent raptors such as Broad-winged Hawks. They may also be important nesting habitats for the long-term conservation of a variety of grassland bird species such as the Bobolink {Dolichonyx oryzivorus) (Yahner and Rohrbaugh 1996). Resumen. — Comparamos la abundancia de aves rapa- ces en primavera en terrenos de mineria en recuper- acion y en zonas de agricultura en dos regiones geo- graficas del norte de Pennsylvania desde Abril-Junio en 1993 y 1994. De las 412 aves rapaces observadas, la mas comun fue Buteo jamaicensis (58%), Falco sparverius (36%) y Circus cyaneus (6%). Las tierras de mineria en recuperacion registraron mas aves rapaces que las de agricultura lo cual indica una posible preferencia por estas areas. Las tierras de mineria presentaron mayor abundancia de presas, tales como Microtus pennsylvan- icus. Concluimos que las tierras de minera en recla- macion de >100 ha proveen habitat de pastizales va- liosos para aves rapaces en reproduccion. Adicional- mente, estas superficies constituyen el habitat para la conservacion a largo plazo de otras aves tales como Dolichonyx oryzivorus. Acknowledgments Funding for this study was provided by the Pennsyl- vania Game Commission, the Pennsylvania Agricultur- al Experiment Station, and the Max McGraw Wildlife Foundation. Literature Cited Alberici, T, W.E. Sopper, G.L. Storm and R.H. Yah- ner. 1989. Trace metals in soil, vegetation, and voles from mine land treated with sewage sludge./ Environ. Qual. 18:115-120. Bart, J. 1977. Winter distribution of Red-tailed Hawk in central New York State. Wilson Bull. 89:623-625. Bednarz, J. 1992. Red-tailed Hawks. Pages 106-107 in D.W. Brauning [Ed.], Atlas of breeding birds in Pennsylvania. Univ. Pittsburgh Press, Pittsburgh, PA U.S.A. Energy Information Administration. 1989. Coal pro- duction 1989. DOE/EIA-O118(89) . Distribution Category UC-98. U.S. Dept. Energy, Washington, DC U.S.A. Litvaitis, J.A. 1993. Response of early successional ver- tebrates to historic changes in land use. Conserv. Biol 7:866-881. Rice, J., B.W. Anderson and R.D. Ohmart. 1984. Com- parison of the importance of different habitat at- tributes to avian community organization. J. Wildl Manage. 48:895—911. Rohrbaugh, R.W., Jr. 1994. Effects of macrohabitat, microhabitat, and microclimate on nest-box use and nesting success of American Kestrels in eastern Pennsylvania. M.S. thesis, Pennsylvania State Univ., University Park, PA U.S.A. AND R.H. Yahner. 1996. Reclaimed surface mines: an important nesting habitat for Northern Harriers in Pennsylvania. Pp. 307-314 mD.M. Bird, D.E. Varland and J.J. Negro [Eds.], Raptors in hu- man landscapes. Academic Press, London, U.K. AND . 1997. Effects of microhabitat and microclimate on nest-box use and nesting success of American Kestrels. Wilson Bull. 109:410-423. SoKAL, R.R. AND F.J. Rohlf. 1981. Biometry. W.H. Free- man and Co., San Francisco, CA U.S.A. Weller, M.W., I.C. Adams, Jr. and B.J. Rose. 1955. Winter roosts of Marsh Hawks and Short-eared Owls in central Missouri. Wilson Bull. 67:189-193. Yahner, R.H. and R.W. Rohrbaugh, Jr. 1996. Long- term status and management of Northern Harriers, Short-eared Owls, and associated wildlife species in grasslands of Pennsylvania. Final Rep., Pennsylva- nia Game Commission, Harrisburg, PA U.S.A. Received 1 July 1997; accepted 15 February 1998 June 1998 Short Communications 181 J. Raptor Res. 32(2):181-182 © 1998 The Raptor Research Foundation, Inc. A New Trap Design for Capturing Spotted Owls Charles L. Johnson and Richard T. Reynolds USDA Forest Service, Rocky Mountain Research Station, 240 West Prospect Street, Fort Collins, CO 80526 U.S.A. Key Words: Spotted Owls; Strix occidentalis; trap; recap- ture. Due to conditioned trap avoidance, Mexican Spotted Owls {Strix occidentalis ludda) are difficult to repeatedly capture. In studies requiring repeated captures (e.g., to replace radiotransmitters), it is necessary to either con- tinually modify existing traps or develop new trap de- signs. In this paper, we describe a new board leg-hold trap that we developed for capturing Spotted Owls. This trap may be useful for other raptor species. The trap consisted of a piece oiVi" (1.27 cm) plywood with a hole cut in the center. A live mouse was tethered in the center hole and a monofilament noose that sur- rounds the hole was pulled shut around the owl’s legs when it attempted to capture the mouse (Fig. 1). To construct the trap, a 5.72 cm radius hole was cut m the center of a 27 X 43 cm piece of plywood. Three 9.52- mm diameter holes were drilled 6.35 cm from the edge of the center hole and three 9.52-mm diameter wooden dowels, 4.5 cm long, were inserted and glued (Fig. 1). A 9.52-mm diameter hole was drilled 1.27 cm deep into the large end of three bottle corks. The three corks were glued onto the ends of the wooden dowels. A 7.62 X 3.81 43 cm 27 cm Plywood Eye screws on eye-screw plate Monofilament line Mouse attachment to binder clip and stake Figure 1. Diagram of the board leg-hold trap. 182 Short Communications VoL. 32, No. 2 X 2.54 cm piece of wood was glued lengthwise at the edge of the plywood base, centered on the bait hole. To this piece of wood, two small eye screws were attached about 2.54 cm apart, facing the bait hole. The eye screws were smaller than the eye ring that was used to make the monofilament noose, so that the eye ring would not pass through the eye screws. Two small holes were drilled through the plywood base about 3 cm from the outer ends of the plywood. Small metal stakes are inserted through these holes to anchor the trap to the ground. With a razor blade or sharp knife, a horizontal slit was cut into the side of each cork, facing the bait hole. The trap was sanded to remove any sharp edges and splinters and the entire trap was painted to match the environ- ment. When trapping Spotted Owls, we first located the owl and placed the trap so that the owl could see it. The trap was staked to the ground and covered with litter or soil. The noose was then inserted in the cork slits so that the noose was fairly taut, with the end of the noose line run- ning through the eye screws to the observer. A live mouse was tethered in the middle of the center hole by clipping a small black binder clip to its tail. The binder clip was attached to a small metal stake that was pushed complete- ly into the ground allowing the mouse to move in circles within the center hole. Once the owl’s attention was on the mouse, we backed slowly away (6 m was often sufficient) while feeding out monofilament line. When the owl landed on the mouse, we pulled rapidly and held pressure on the monofila- ment line. After being caught, we quickly secured the owl. This trap worked best with two people; one to hold the line and one to secure the owl. A small tape recorder, emitting mouse vocalizations and scurrying sounds, can also be placed near (or beneath) the trap to entice hes- itant owls. Although our trap is manually operated, we are aware that other more sophisticated noose traps have been de- veloped for owls that are automatically tripped when the owl lands on the spring-loaded trigger mechanism (Eric Forsman pers. comm.). Although both types of traps may be effective, we think that the manually-operated version has several advantages. In particular, it is less complicat- ed, cheap to construct, easily camouflaged, and allows the operator considerable flexibility regarding the best time to pull the noose shut. Although we have used our trap only to catch Spotted Owls, we believe it should also be effective for other owls and diurnal raptors. Resumen. — Una variedad de trampas (de nudos, redes de niebla, bal chatri, etc.), han sido empleadas para la cap- tura de buhos. Debido a su familiarizacion, usualmente es dificil atrapar en forma repetida a Strix occidentalis. Este articulo describe un tipo de trampa que permite en for- ma efectiva capturar a Strix occidentalis en Colorado, Acknowledgments We thank Eric Forsman, Joe Ganey, and Suzanne Joy for reviewing this manuscript. Funding was provided by Fort Carson Department of Defense, Air National Guard, Bureau of Land Management, and USDA Forest Service Received 6 August 1997; accepted 11 February 1998 Letters J. Raptor Res. 32(2): 183 © 1998 The Raptor Research Foundation, Inc. Unusual Nesting of the Rufous-legged Owl? The Rufous-legged Owl (Strix rufipes King) is a forest specialist that inhabits the woodlands of southern South America. Dietary studies have shown it to be a generalist predator (D.R. Martinez 1993,/, Raptor Res. 27:214-216). Recent research on habitat use indicates it is associated with old-growth forests (D.R. Martinez and F.M. Jaksic 1996, Ecosdence 3:259-263) , and cavity-bearing dead trees have been suggested as the key features and limiting factors in suitable habitat (R. Rozzi et al. 1996, pages 135-152 in}.]. Armesto et al. [Eds.], Ecologia de los bosques nativos de Chile, Editorial Universitaria, Santiago, Chile). Apparently, in late September in mature forests, it selects a nest tree with an upward-facing cavity where it lays a clutch of 2-3 eggs (A. W. Johnson 1967, The birds of Chile and adjacent regions of Argentina, Bolivia and Peru. Vol. 2. Platt Establecimientos Grahcos, Buenos Aires, Argentina; R. Housse 1945, Las aves de Chile: su vida y costumbres. Ediciones Universidad de Chile, Santiago, Chile). To our knowledge, no firsthand descriptions of an active nest have been published. On 11 October 1996, while conducting an assessment of the use of exotic pine {Firms radiata) plantations by diurnal birds in Constitucion, central Chile, we found a Rufous-legged Owl nest in an approximately 15-yr-old pine stand (35°27'32'S, 72°52'40"W). The nest, which contained two eggs, was a small depression on the ground covered with pine needles and surrounded by woody debris. The vegetation in the understory was sparse (37.5% cover) and was composed mainly of shrubby individuals of Cryptocaria alba and Aristotelia chilensis. The pine canopy had a mean height of 11 m and a coverage of 55%. The closest patch of native forest was 420 m away. We observed the nest every 4—5 d, and there was always an owl incubating the eggs. By 23 October, one egg had disappeared from the nest, but the other was still being incubated. On 6 November, we visited the site for the last time. The egg had not yet hatched, so we could not conhrm nesting success. During the same day, however, we discovered a second nest located in a pine plantation only 350 m north from the first one (35°27'24"S, 72°52'40"W). This nest was also on the ground at the base of a clump of grasses and contained only one egg. The composition of the understory (33.7% cover) was similar to that of the first site, but the pine plantation was older (approximately 25-yr-old) , with a mean canopy height of 16 m and a coverage of 62.5%. The nest was 100 m from the nearest native forest. We found these two nests by chance while surveying less than 0.5% of the 14073-ha of pine plantations in our study area. This suggests that the Rufous-legged Owl might be relatively common. Ground nesting has been reported for other Strix owls as a response to the lack of nesting sites due to forestry practices (S.J. Petty et al. 1994,/. Raptor Res. 28:134—142, S. Sulkava and K. Huhtala 1997,/. Raptor Res. 31:151-159). In our case, there were nearby stands of native forest with trees large enough to contain potential nesting cavities. Nonetheless, our survey in 51 ha (2.55% available area) of native forest found no owls. The clutches we observed were smaller than those reported for the species and certainly smaller than the average sizes of clutches of temperate Sthxowls in North America (2.4— 3.2; P.A. Johnsgard 1988, North American owls. Biology and natural history. Smithsonian Institution Press, Washington, DC U.S.A.) and mid-Europe (4.6; H. Mikkola 1973, pages 116-146 m J.A. Burton [Ed.], Owls of the world, their evolution, structure and ecology. E.P. Dutton & Co , New York, NYU.S.A.). Small clutches and the loss of one egg during the observation period indicate that success of ground nests in pine plantations may be low for the species. However, the apparent behavioral shift from cavity nesting in old-growth native forests to ground nesting in exotic pine plantations could be an indication that the Rufous-legged Owl is not as dependent on old-growth forests as previously assumed. We thank the Zoological Society of Milwaukee County and the Max MacGraw Wildlife Foundation for providing partial funding for this study and Forestal Copihue for kindly allowing us to conduct research in its property. We are also grateful to R.P. Gerhardt, F.R. Gehlbach and to an anonymous reviewer for their helpful comments on an earlier version of this note. — Cristi^ F. Estades, Department of Wildlife Ecology, University of Wisconsin, 1630 Linden Drive, Madison, WI 53706 U.S.A. and Departamento de Manejo de Recursos Forestales, Universidad de Chile, Casilla 9206, Santiago, Chile; Stanley A. Temple, Department of Wildlife Ecology, University of Wisconsin, 1630 Linden Drive, Madison, WI 53706 U.SA.; and Alvaro F. Gajardo, Departamento de Manejo de Recursos Forestales, Univer- sidad de Chile, Casilla 9206, Santiago, ChUe. 183 184 Letters VoL. 32, No. 2 J. Raptor Res. 32(2) :184 © 1998 The Raptor Research Foundation, Inc. Apparent Siblicide in Peregrine Falcons Siblicide occurs regularly in many species of raptors but has not been reported for falcons. Ian Newton (1979, Population ecology of raptors, Buteo Books, Vermillion, SD U.S.A.) found no record of serious sibling aggression among the falcons. We here report an instance of apparent siblicide in the Peregrine Falcon {Falco peregrinus) . On 17 June 1994, we banded a brood of four young peregrines at Palisade Head, a cliff on the north shore of Lake Superior, Lake County, Minnesota. The young were about 17-d old. One of the four was weak and very bloody on its back from its shoulders to its tail, apparently picked by its nest mates. We removed it from the nest but it died within 2 hr despite careful handling. None of the other three nestlings showed any injuries, although all appeared thin. The Palisade Head site seems not to be ideal location for peregrine nesting. Lake Superior is deep and cold, with rocky shores. Bird life on the lake and its mostly forested shores is sparse. The same female has nested at Palisade Head since 1988 with four different males. Brood size at fledging has been one young (3 yr), two young (once), three yoixng (4 yr) , and four (once) . Prey species in years with broods of three or fewer have included small passerines. Blue Jays ( Cyanocitta cristata) , Northern Flickers ( Colaptes auratus) , shorebirds, small ducks, an occasional Rock Dove (Columba livia), bats and chipmunks taken off the cliff face. In 1990, the peregrines hatched four young. When the brood was about 3-wk old, the female began killing Ring-billed Gulls (Larus delawarensis) . Over the next 6 wk, we identified about two dozen ring-bills that were taken. All were juveniles except for one adult. Herring Gulls {Larus argentatus) are abundant along the lakeshore, but ring-bills are scarce. The female seemed to hunt the ring-bills selectively. In years when the brood size is smaller than four, Ring-billed Gulls were taken infrequently and Herring Gulls were not taken at all. A further indication that prey is scarce at Palisade Head is our impression that single young at banding seem robust and well fed, while young in large broods appear malnourished. However, we have not quantified this observation. The apparent siblicide we observed may well have been due to hunger. Well-fed peregrine nestlings seem to get along with little aggression. About 12 pairs of peregrines currently nest close to Lake Superior. Most of these probably face foraging problems similar to those at Palisade Head, as do many other pairs nesting in mostly forested habitats in the north. Therefore, hunger and starvation in peregrine broods may be common. Dead nestlings, whether starved or killed by siblings, would be quickly eaten by adults or fed to young. Peregrine nestlings can feed themselves starting at about 25 d of age and could eat dead siblings. Until siblicide is directly observed in peregrines, however, the bloody chick at Palisade Head remains the most convincing evidence that it occurs in Peregrine Falcons. Adult Peregrine Falcons have been observed feeding dead young to their siblings (T. Cade pers. comm.). We have twice noted the disappearance of dead nestlings within a few hr of nest flooding by heavy rain; presumably they were eaten or removed by the parents. There has also been an instance where a small male peregrine nestling in a brood with two large females disappeared between nest visits in circumstances suggesting siblicide as a possible explanation (B. Mutch pers. comm.). We suggest that siblicide may occur fairly regularly in peregrine broods that are food- stressed. — Harrison B. Tordoff, Bell Museum of Natural History and Department of Ecology, Evolution, and Behavior, University of Minnesota, 1987 Upper Buford Circle, St. Paul MN 55108 U.SA, and Patrick T. Redig, The Raptor Center, University of Minnesota, 1920 Fitch Avenue, St. Paul MN 55108 U.S.A. BOOK REVIEWS Edited by Jeffrey S. Marks /. Raptor Res. 32(2) :185-186 © 1998 The Raptor Research Foundation, Inc. Buteo Books is pleased to sponsor the Book Re- view section of the Journal of Raptor Research. Buteo Books stocks a comprehensive selection of orni- thology books, both new and used. A Fascination with Falcons: A Biologist’s Adven- tures from Greenland to the Tropics. By Bill Burn- ham. 1997. Hancock House, Blaine, WA. 233 pp., 46 color photographs, numerous pen-and-ink drawings and vignettes by John Schmitt. ISBN 0- 88839-415-2. Cloth, $29.95, — ^For me, this was an engaging book, and I was pleased to be asked to review it. In part, because Bill took a master’s de- gree with me working on Greenland falcon biology and, in a sense, because I watched the author go through his professional development and embark on many of the adventures he describes. Burnham explains his plan for presenting the book’s layout and subject matter on page 8, ar- ranging the chapters geographically and topically, but not necessarily chronologically. Starting in Greenland, he travels southward to the mid-lati- tudes, primarily Colorado and Idaho, where his main adventures were with falconry. Then, he moves south to the tropics, with adventures mainly in Central America. There is a series of pages with photographs clustered together about a third of the way through the book. They were well-chosen to illustrate a variety of themes, from falcons on their prey, to colleagues and researchers in the field, to landscapes he visited. In all there are seven chapters and an appendix. Chapter 1 is entitled “Events, Great and Small.” It starts with an introduction to the general actions of pesticides on raptors, moves into Bill’s late teen- age years and his introduction to falcons and fal- conry, and ends with an introduction to, and ra- tional for, the breeding of falcons in captivity and his involvement. Chapter 2 describes “In Search of the Gyrfalcon.” Much of this chapter depicts the adventures and problems he and Steve Sherrod (another master’s student of mine) had in working on Disco, a large island on the west coast of Green- land about one-third of the way up from the south- ern tip. I well remember Bill and Steve telling me about learning how to use a native (Inuit) Green- lander’s kayak, especially in mastering the ability of righting oneself when upside down in freezing water. Chapter 3, “First Time North,” essentially is a description of working throughout western Greenland on falcons, from the time Bill started in 1972, and the findings of the research teams with whom he worked. Several important publications in the open literature have resulted from all the extended work in Greenland that is, by the way, still ongoing — so, in a sense. Bill’s book is not yet done. Two chapters then cover falconry, sharing a plethora of stories about birds, people, and fun times, either alone or in solitude or with others. The penultimate chapter takes us into the trop- ics with “Neotropical Falcons.” Bill and others at the Peregrine Fund have spearheaded an extensive program in the tropics cleverly called the Maya Project — most of the studies were in the Central American home of the Mayan people. An enor- mous amount of what we know about Neotropical raptors resulted from this project, which is really the foundation for Bill’s chapter. Finally, there is a well-thought-out chapter on “Conservation.” I par- ticularly liked his starting sentence in this chapter, when he was asked “Are you an environmentalist?” “No,” he responded “but I do care about the en- vironment as every man and woman in the world does.” This chapter has several subtitles that are self-explanatory: Nature Conservation; Conserving Falcons In The Americas; You Can Make A Differ- ence (with a set of sections that are self-evident from their tides, such as Support Conservation of Wildlife Habitat and Let Your Opinion Be Known) ; and lastly, Final Thoughts, in which he writes about the Endangered Species Act and our behavior in, and implied stewardship of, the natural world. A brief appendix on the “Biology and Ecology of Fal- cons” terminates the book. 185 186 Book Reviews VoL. 32, No. 2 The drawings and vignettes by Schmitt are well- executed and artfully done, but then he consis- tently does superb work with raptors. I counted 75 in all, some nothing more than a study of a falcon contour feather. The drawings were not necessarily original for the book, and 1 have seen many else- where. My sense is that many were overly dark, too much ink and some detail lost, but this clearly is a function of printing rather than the drawing. Some of the artwork seemed out of place and perhaps unnecessary. Although nonetheless nice, 1 missed the rationale for including the foot of a Great Horned Owl {Bubo virginianus) and the Osprey {Pandion haliaetus) in the chapter on “Falconry,” where the species are never mentioned, or the Red-tailed Hawk {Buteo jamaicensis) in the chapter on “First Time North,” which is a description of early experiences in Greenland, where the Red- tailed Hawk does not occur. Most of the other drawings have some relationship to the chapters in which they appear; e.g., an arctic fox {Alopex lago- pus) in the chapter on Greenland, hunting dogs in the chapter on falconry, and Swallow-tailed Kite {Elanoides forjicatus) in the chapter on Neotropical falcons. This is a fun book, light reading, informative, well-written and edited, and at the end of the day it has the most essential of all take home messages: conservation of landscapes and human behavior toward the earth and its biodiversity. Even if you are not interested in falcons, Greenland, or the tropics, this book is worth having and sharing with others. — Clayton M. White, Department of Zoolo- gy, Brigham Young University, Provo, UT 84602 U.S.A. J Raptor Res. 32(2): 186 © 1998 The Raptor Research Foundation, Inc. Rotmilan-Sonderheft. Edited by K. Richarz, B.- U. Meyburg, and M. Horsman. 1995. “Vogel und Umwelt” Vol. 8. World Working Group on Birds of Prey and Owls. 180 pp., photos, maps, charts. Drawings by Franz Muller. Paper, DM 17.50 (ap- proximately $20.00 U.S.). — This is the proceedings of a symposium on the Red Kite {Milvus milvus; “Rotmilan” in German) held in Saxony in 1994. Included are 18 papers, most of which are in Ger- man but with English summaries. All aspects of the biology of this, the only diurnal raptor with a range almost restricted to Europe, are thoroughly docu- mented. The Red Kite had been faring well, but it has decreased with the introduction of large-scale agriculture in the former East Germany. It is feared that the same may be occurring in much of Eu- rope, although in Great Britain efforts to expand the small nucleus of a population in Wales seem to be hopeful. This attractive and valuable little book and the symposium from which it originated were spon- sored by the World Working Group on Birds of Prey and Owls and by various German ornitholog- ical and conservation groups. It contains an insert by a power company with a striking color photo- graph of II “Rastende Rotmilan” perched on an electric transmission tower. — ^Dean Amadon, De- partment of Ornithology, American Museum of Natviral History, Central Park West at 79th Street, New York, NY 10024 U.S.A. /. Raptor Res. 32(2): 186-1 87 © 1998 The Raptor Research Foundation, Inc. A Field Guide to Birds of The Gambia and Sen- egal. By Clive Barlow and Tim Wacher. 1997. Yale University Press, New Haven, CT. 400 pp., 4 maps, 48 color plates and numerous pen-and-ink draw- ings by Tony Disley. ISBN 0-300-07454-9. Hard cov- er, $40.00. — ^At 11 000 km^, The Gambia is one of the smallest countries in the world. This narrow (25 to 30 km-wide) country is defined by the course of the Gambia River and virtually is sur- rounded by Senegal. This book focuses on The Gambia because the authors reside (Barlow) or have worked there (Wacher) , but Senegal is treat- ed completely as well. A 15-page introduction in- cludes information on geography, climate, vegeta- tion, major habitats and protected areas. The color plates by Tony Disley are simple but elegant, and, in my opinion, they are at least as good as those in June 1998 Book Reviews 187 The Collins Field Guide to the Birds of East Africa pub- lished in 1980. The plates depict 570 of the ap- proximately 660 bird species recorded for the “Se- negambian” region. Each species account includes brief information on identification (including flight characteristics), habits, voice, status and dis- tribution, and breeding biology. The region in- cludes an incredible diversity of raptors; the spe- cies accounts treat 58 species of falconiforms and 12 species of owls. Indeed, the falconiforms occupy 8 of the 48 color plates, 3 of which focus on flying raptors. The layout of the book is quite user-friend- ly. Having visited East Africa, I am now eager to expand my horizons to the western part of the con- tinent, especially because another great field guide is now available for that region. — Jeff Marks, Mon- tana Cooperative Wildlife Research Unit, Univer- sity of Montana, Missoula, MX 59812 U.SA. 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. Barn Owl (1). Carl D. Marti. 1992. 16 pp. Boreal Owl (63). G.D. Hayward and RH. 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, B.A. 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). Marc J. 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. 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, R James and I. Warkentin. 1993. 20 pp. Northern Saw-whet Owl (42). Wchard 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. 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). RW. 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, Marc J. Bechard and C. Stuart Houston. 1997. 28 pp. Swallow-tailed Kite (138). Kenneth D. Meyer. 1995. 24 pp. White-tailed Hawk (30). C. Craig Farquhar. 1992. 20 pp. White-tailed Kite (178). Jeffrey R. Dunk. 1995. 16 pp. THE RAPTOR RESEARCH FOUNDATION, INC. (Founded 1966 ) OFFICERS PRESIDENT: Michael N. Kochert VICE-PRESIDENT: David E. Andersen SECRETARY: Patricia A. Hall TREASURER: Jim Fitzpatrick BOARD OF DIRECTORS NORTH AMERICAN DIRECTOR #1: INTERNATIONAL DIRECTOR #3: Petra Bohall Wood NORTH AMERICAN DIRECTOR #3: Brian A. Millsap NORTH AMERICAN DIRECTOR #2: Massimo Pandolfi INTERNATIONAL DIRECTOR #2: Karen Steenhof INTERNATIONAL DIRECTOR #1: Michael McGrady DIRECTOR AT LARGE #1: Patricia L. Kennedy DIRECTOR AT LARGE #2: John A. Smallwood DIRECTOR AT LARGE #3; James C. Bednarz DIRECTOR AT LARGE #4: Cesar MArquez Reyes DIRECTOR AT LARGE #5: Lloyd Kiff DIRECTOR AT LARGE #6: Robert Kenward Reuven Yosef EDITORIAL STAFF EDITOR: Marc J. Bechard, Department of Biology, Boise State University, Boise, ID 83725 U.S.A. BOOK REVIEW EDITOR: Jeffreys. Marks, Montana Cooperative Research Unit, University of Montana, Missoula, MT 59812 U.S.A. SPECIAL PUBLICATIONS EDITOR: Daniel E. Varland, Rayonier, 3033 Ingram Street, Hoquiam, WA 98550 SPANISH EDITOR: Cesar MArquez Reyes, Instituto Humbolist Colombia, AA. 094766, Bogota 8, Colombia 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 (814 X 11 in.) or standard international, white, bond paper, with 25 mm (1 in.) margins. The cover page should contain a title, the author’s full name(s) and address(es). Name and address should be centered on the cover page. If the current address is different, indicate this via a footnote. A short version of the title, not exceeding 35 characters, should be provided for a running head. An abstract of about 250 words should accompany all research articles on a separate page. Tables, one to a page, should be double-spaced throughout and be assigned consecutive Arabic numer- als. Collect all figure legends on a separate page. Each illustration should be centered on a single page and be no smaller than final size and no larger than twice final size. The name of the author(s) and figure number, assigned consecutively using Arabic numerals, should be pencilled on the back of each figure. Names for birds should follow the A.O.U. Checklist of North American Birds (6th ed., 1983) or another authoritative source for other regions. Subspecific identification should be cited only when pertinent to the material presented. Metric units should be used for all measurements. Use the 24-hour clock (e.g., 0830 H and 2030 H) and “continental” dating (e.g., 1 January 1990). Refer to a recent issue of the journal for details in format. Explicit instructions and publication policy are outlined in “Information for contributors,” / Raptor Res., Vol. 27(4), and are available from the editor. ASSOCIATE EDITORS Allen M. Fish Gary R. Bortolotti Charles J. Henny Fabian Jaksic Daniel E. Varland