W\L • o HARVARD UNIVERSITY Ernst Mayr Library of the Museum of Comparative Zoology MCZ library JAN 1 8 20lt , harvard university OST Wilson Journal ofO rnithology Volume 120, Number 1, Mareh 2008 \ ' MC2 brarv UNlVERSiyv Published by the Wilson Ornithological Society THE WILSON ORNITHOLOGICAL SOCIETY FOUNDED 3 DECEMBER 1888 Named after ALEXANDER WILSON, the first American ornithologist. President— James D. Rising, Department of Zoology, University of Toronto, Toronto, ON M5S 3G5, Canada; e-mail: rising@zoo.utoronto.ca First Vice-President— E. Dale Kennedy, Biology Department, Albion College, Albion, MI 49224, USA; e-mail: dkennedy@albion.edu Second Vice-President— Robert C. Season, USDA, Wildlife Services, 6100 Columbus Ave, Sandusky, OH 44870, USA; e-mail: beason@netzero.com Editor Clait E. Braun, 5572 North Ventana Vista Road, Tucson, AZ 85750, USA; e-mail: TWILSONJO@ comcast.net Secretary— John A. Smallwood, Department of Biology and Molecular Biology, Montclair State University, Montclair, NJ 07043, USA; e-mail: smallwoodj@montclair.edu Treasurer MelindaM. Clark, 52684 Highland Drive, South Bend, IN 46635, USA; e-mail: MClark@tcservices.biz Elected Council Members— Robert L. Curry; Daniel Klem Jr., and Douglas W. White (terms expire 2008); Carla J. Dove, Greg H. Farley, and Mia R. Revels (terms expire 2009); Rebecca Holberton, Robert S. Mulvihill, and Timothy O’Connell (terms expire 2010). Membership dues per calendar year are: Active, $21.00; Student, $15.00; Family, $25.00; Sustaining, $30.00; Life memberships, $500 (payable in four installments). The Wilson Journal of Ornithology is sent to all members not in arrears for dues. THE WILSON JOURNAL OF ORNITHOLOGY (formerly The Wilson Bulletin) THE WILSON JOURNAL OF ORNITHOLOGY (ISSN 1559-4491) is published quarterly MarcN^ September, and December by the Wilson Ornithological Society, 810 East 10th Street, Lawrence, KS 66044-8897. The subscription pnce, both in the United States and elsewhere, is $40.00 per year. Penodicals postage paid at Lawrence, KS. POSTMASTER: Send address changes to OSNA, 5400 Bosque Boulevard, Suite 680, Waco, TX 76710. All articles and communications for publications should be addressed to the Editor. Exchanges should be addressed to The Josselyn Van Tyne Memorial Library, Museum of Zoology, Ann Arbor, Michigan 48109. Subscriptions, changes of address, and claims for undelivered copies should be sent to OSNA, 5400 Bosque Boulevard, Suite 680, Waco, TX 76710. Phone: (254) 399-9636; e-mail: business(gosnabirds.org. Back issues or single copies are available for $12.00 each. Most back issues of the journal are available and may be ordered from OSNA. Special prices will be quoted for quantity orders. All issues of the journal published before 2(^00 are accessible on a free Web site at the University of New Mexico library (http://elibrary.unm.edu/sora/). The site is fully searchable, and full-text reproductions of all papers (including illustrations) are available as either PDF or DjVu files. © Copyright 2008 by the Wilson Ornithological Society Printed by Allen Press, Inc., Lawrence, Kansas 66044, U.S.A. COVER: Wilson’s Snipe (Gallinago delicata). Illustration by Scott Rashid. 0 This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). MCZ LIBRARY MAY 19 2009 Harvard UNIVERSITY FRONTISPIECE. Zosterops somadikartai, sp. nov., Togian White-eye, a new species of white-eye endemic to mangroves and lowland secondary woodlands in the Togian Islands, Sulawesi, Indonesia. Adult male (holotype), original painting by Agus Prijono. VOL. 120, NO. 1 ‘TT’e Wilson Journal of Ornithology Published by the Wilson Ornithological Society March 2008 PAGES 1-238 The Wilson Journal of Ornithology 120(1): 1-9, 2008 A NEW WHITE-EYE (ZOSTEROPS) EROM THE TOGIAN ISLANDS, SULAWESI, INDONESIA MOCHAMAD INDRAWAN,' PAMELA C. RASMUSSEN,^'' AND SUNARTC'^ ABSTRACT. — We encountered white-eyes (Zosterops) that did not match any known species during ornitho- logical field observations in the Togian Islands, Gulf of Tomini, Sulawesi, Indonesia. Subsequently, we collected a specimen and made tape recordings. We consider the Zosterops of the Togian Islands to be a new species that differs most strikingly from the Black-crowned White-eye (Z. atrifrons) in lacking a white eye ring and in soft- part colors. The new species has a somewhat higher-pitched, less modulated song than Z. atrifrons. It seems uncommon and has been encountered only near sea level on three small islands, and it may be best considered Endangered. This brings the number of endemic species in the Togian Islands to two and, under BirdLife International criteria, this island group (which has recently been declared a National Park) now qualifies as an Endemic Bird Area. Received 17 April 2006. Accepted 25 May 2007. The white-eyes {Zosterops) of Indonesia comprise a bewildering and diverse array of taxa. No fewer than 10 taxa in six species (Ev- erett’s White-eye [Z. everetti^. Mountain White-eye [Z. montanus]. Lemon-bellied White-eye fZ. chloris]. Pale-bellied White-eye [Z. consohrinorum]. Lemon-throated White- eye IZ. anomcdus], and Black-crowned White- ' Department of Biology, Eaculty of Science and Mathematics, University of Indonesia, Depok Campus, Depok 16424, Indonesia. ^ Michigan State University Department of Zoology and Museum, We.st Circle Drive, East Lansing, MI 48824, USA. ^ Department of Fisheries and Wildlife Science, Col- lege of Natural Resources, 210 Cheatham Hall, Vir- ginia Tech University, VA 24060, USA. ■* Corresponding author; e-mail: rasmus39@msu.edu eye [Z. atrifrons]-, Mees 1961, White and Bruce 1986) have been conventionally rec- ognized from the Sulawesi subregion alone. Despite this, it recently has become clear that Mees’ (1961) classification substantially un- derestimates the diversity of Zosterops taxa of the Sulawesi subregion. Taxa formerly consid- ered subspecies of Z. atrifrons that are clearly better considered full biological species in- clude the Sangihe White-eye (Z. nelirkorni) and the extralimital Seram White-eye (Z. stalkeri) (Rasmussen et al. 2()0(), Dickin.son 2003). Lormerly synonymized subspecies of Z. atrifrons requiring recognition include Z. a. suhatrifrons and Z. a. siirdns (Rasmus.sen et al. 2000) and a probably undescribed form of Z. atrifrons occurs in south-central Sulawesi (Holmes and Holmes 1985). Most surprising- ly, at least two distinctive new taxa have re- 1 2 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 N FIG. 1. Relevant localities in the Togian Islands, eastern Sulawesi, and the Banggai Islands. cently been discovered; one, a member of the Z. consobrinorum group from Wangi Wangi in the Tukangbesi Islands (off Sulawesi’s southeast peninsula) is still formally unde- scribed (Owen 2004; David Kelly, pers. comm.), while the other is the subject of this paper. We now know there are at least 15 taxa and 9-10 full biological species of Zosterops in the Sulawesi subregion. Few of these forms overlap in geographic range and several are restricted to one or two peninsular “arms” of Sulawesi or one or a few of its satellite is- lands. Even more taxa of Zosterops may await discovery in this highly complex and poorly surveyed region, not only in remote islands, but also in the less explored mountain ranges of Sulawesi’s several peninsulas. The first documented ornithological work in the Togian Islands (now also spelled Togean Islands), Gulf of Tomini, Sulawesi, Indonesia (Fig. 1) was in August 1871 by A. B. Meyer (Indrawan et al. 2006). The only other bird collector to have worked there was J. J. Men- den in the fall of 1939 (Indrawan et al. 2006). Subsequent ornithological work in the Togian Islands has been of limited duration and strict- ly observational. However, longer-term field observations in the Togian Islands (1-28 Aug 1996, 3-27 Jun 1997, and 11-18 May 2001) and Bangkurung Island (Banggai Archipelago, 2-8 Nov 1997) resulted in discovery of a new species of owl (Indrawan and Somadikarta 2004) and observations of white-eyes (Fron- tispiece) in the field (by MI and Sunarto). White-eyes on the Togian Islands (but not those of Bangkurung) are similar in most characters to the Black-crowned White-eye (Z. atrifrons) of Sulawesi, except they entirely lack the white eye ring, which is prominent in adults of all forms of Z. atrifrons (Rasmussen et al. 2000, Dickinson 2003), and narrower but still conspicuous in juveniles (Rasmussen et al. 2000). We observed Z. atrifrons in the eastern and northern peninsulas of Sulawesi to compare with our sightings in the Togian Islands. The observations in the eastern peninsula were in Tanjung Api Nature Reserve (16-19 and 24- 27 Jan 2003), while those in the northern pen- insula were in Tomohon (4-5 Jan 2002, 31 Jul 2003) and Tangkoko Nature Reserve (30 Jul 2003) (MI, unpubl. data). MI revisited the Togian Islands in July 2003 and collected a single specimen (Frontis- piece). This is the first museum specimen of Zosterops known from either the Togian Is- lands or Bangkurung Island. The Togian Is- lands specimen was directly compared by PCR at AMNH (acronyms in Acknowledg- ments) with 29 Z. a. atrifrons, two Z. atrifrons (race uncertain) from Siuna (00° 45' S, 122 58' E), east-central Sulawesi, three Z. a. sur- dus from west-central Sulawesi, five Z. a. su- laensis from the Sula Islands, and four Z. a. subatrifrons from Peleng Island, Banggai Is- lands; at BMNH with six nominate atrifrons Indrawan et al. • NEW SPECIES OF WHITE-EYE 3 and three Z anomalus’ at USNM with series of nominate atrifrons and surdus\ and indi- rectly using photographs at RMNH with two anomalus, four nominate atrifrons, four su- laensis, and five Z. stalkeri. We compared songs and calls of the white- eye from the Togian Islands with those of Z a. atrifrons and surdus {n — 4), Z. a. sulaensis (1), Z. nehrkorni (1), Z h. hypoxanthus of New Britain (2), and Z h. admiralitatis of Manus (1); recordings of vocalizations of oth- er taxa were unavailable to us. The new white- eye differs strikingly in plumage from other known taxa, being closest to Z a. atrifrons, and it appears to differ consistently in song from Z atrifrons. Given that the magnitude of the several differences exceeds those between most other currently recognized species of Zosterops, and that lack of an eye ring, iris color, and song differences could well serve as isolating mechanisms, the new taxon should be considered specifically distinct un- der the Biological Species Concept. We pro- pose for it the name: Zosterops somadikartai, sp. nov. Togian White-eye Holotype.—MusQum Zoologicum Bogo- riensis 30366, male, from Pulau Malenge (00° 15' S, 122° 03' E, —50 m asl), Togian Islands, Gulf of Tomini, Sulawesi, collected 27 July 2003 by MI. Specimen in heavy molt of wing and body, and has lost feathers from right side of venter, lower undertail coverts, all uppertail coverts, and entire tail. Detached feathers saved at time of collection include one right and four left rectrices, one belly feather, two ventral flank feathers and three from more dorsally on the flanks, and five uppertail co- verts. Holotype an adult based on compact feather structure, especially on nape and man- tle; broad-tipped outer remiges in comparison with two known juvenile surdus; heavy molt; flaking tarsal scutes, and (as shown by a ra- diograph) evidently fully ossified cranium. Diagnosis. — Zosterops somadikartai differs from nominate atrifrons (of northern Sula- wesi) in its slightly less extensive black cap, which does not reach the upper rear eye and which appears to have a straight rather than oblique border at the rear edge; in its lack ol a white eye ring (vs. a medium-width white eye ring); its clearer yellow throat, less invad- ed by olive on the sides and the yellow not extending onto the upper breast; its whiter un- derparts; its slightly heavier bill than most atrifrons, with a distinctly pale base; and its reddish (vs. brown) iris. Measurements of the holotype suggest slightly different proportions from atrifrons (Table 1) but cannot be taken as definitive given that only a single individ- ual of somadikartai was available for mea- surement. From surdus (of west-central Sula- wesi), somadikartai differs in the same ways as from atrifrons, but is paler and brighter ol- ive above and much clearer yellow on the throat. Zosterops somadikartai differs from subatrifrons (of Peleng Island, Banggai Is- lands) by its lack of a white eye ring (vs. a fairly broad white eye ring); its grayer breast, and its less extensive black crown. From su- laensis (of the Sula Islands), somadikartai dif- fers by its lack of a white eye ring (vs. a very broad white eye ring); by its slightly narrower- based bill; its less extensive black crown; its grayish (vs. white) breast; and its duller yel- low-olive upperparts. From Z. m. minor (of western New Guinea, which also lacks a white eye ring but which also has tiny white specks around the orbital skin), somadikartai differs in having a narrower-based bill; blacker (vs. grayish) bare skin around the eye; in having a black forecrown (vs. forecrown olive, essen- tially concolorous with mantle); in being much duller olive above; in having a gray (vs. white) breast; and in having a much duller, less orange throat. From Z. anomalus (of southern Sulawesi, which lacks a white eye ring but also has tiny white specks around the orbital skin), somadikartai differs in its small- er size; distinctly thinner bill (on dorsal view); black (vs. bright yellow-olive) forecrown and lores (vs. yellow); duller olive upperparts; less extensive, less bright yellow throat; blacker (vs. dark brown) wingtips; and whiter central underparts {anomalus is incorrectly shown as having a white vent in Coates and Bishop [1997]; it actually has a yellow vent as with somadikartai). No other regional taxa ol Zos- terops are as similar to somadikartai as those listed above. Description of the Holotype. — Capitalized color and number descriptions follow Smithe (1975). Forehead from base of bill to above eye Jet Black (89); bare orbital ring blackish- gray with tiny white specks (latter visible only 4 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 at close range), surrounded by Jet Black feath- ered eye ring; rear crown and nape Citrine (51), grading on auriculars to Yellowish Ol- ive-Green (50), and on lower sides of face to Olive- Yellow (52), then to Sulfur Yellow (157) at center of throat. Back Citrine, rump Yellowish Olive-Green. Wing coverts Olive- Green (Basic [46]), marginal coverts paler than Sulfur Yellow, alula Blackish Neutral Gray (82), primary and secondary edgings Yellowish Olive-Green, tips of primaries and secondaries between Blackish Neutral Gray and Jet Black, tertials Citrine, underwing co- verts whitish. Upper tail coverts and rectrices detached and central rectrices missing and color cannot reliably be described. Rectrices dark brownish-black with weakly olive lateral edgings (as in many atrifrons and surdus) when arranged as in folded tail. Undertail co- verts missing. Upper breast paler than Pale Neutral Gray (86), lower breast to belly and flanks white, tibial feathers white on internal side, grayer externally, vent Sulfur Yellow. Iris dark red (from photographs taken of ho- lotype while living). In life, the basal third of the lower mandible was pale flesh, although when dried the bill of the same individual ap- pears completely black except for the rami of the lower mandible, which are a dull pinkish horn. The bare fleshy eye ring was dark gray- ish in life, but is blackish in the specimen. The feet are fairly pale in the photograph in which they are visible, but are dark metallic horn in the specimen. Measurements of the Holotype. — Measure- ments (mm) were taken by PCR. Head and body length of dried specimen (forecrown at base of bill to vent) ~65; culmen (from feath- ering) 1 1.0; wing arc 57; tail missing, but lon- gest feather (from tip to proximal end of pig- mentation on dorsal shaft) 40.2; tarsus 14.9 (Table 1). Etymology. — We are pleased to name this new species after Professor Soekarja Soma- dikarta, Indonesia’s leading avian taxonomist. Dr. Somadikarta encouraged the senior author to work on taxonomic aspects of birds of Su- lawesi (resulting in the discovery of two new bird species), and for many years contributed significantly to Indonesian ornithology, espe- cially in taxonomy. Specimens. — The holotype is the only known specimen of Z. somadikartai. REMARKS Voice Song. — The song of Z. somadikartai is a thin sweet warble, fairly similar overall to that of Z. atrifrons (Fig. 2), which is somewhat variable among available recordings. Succes- sive song strophes of what is evidently the same individual of atrifrons are typically sim- ilar to one another. Five successive song stro- phes for somadikartai, probably from the same individual, are available and, despite the complexity of the song, all these strophes are extremely similar, differing at most by the dropping of a single note near the end of the strophe. The song sample of somadikartai dif- fers consistently from that of atrifrons in sounding distinctly higher overall. There is broad overlap of frequencies between the two taxa, but the frequency range of atrifrons is broader (4.2 to 7.3 kHz for somadikartai vs. 3.0 to 7.7 kHz for atrifrons). The song of so- madikartai also differs from atrifrons in being notably less modulated, with nearly all ele- ments having a much narrower frequency range (maximum modulation per element 0.4 kHz for somadikartai vs. 2.4 kHz for atrif- rons). Two of the three available song samples of atrifrons end with a faint 2- to 4-note “trill” of short chip notes and, although this appears to be lacking in the third sample, it may possibly have been too faint to appear in this poor-quality recording. None of the avail- able song strophes of somadikartai has trills at the end or elsewhere in the song. The song of somadikartai is much higher- pitched overall (frequency range 2.1 to 5.2 kHz in sulaensis) compared to Z. a. sulaensis (Fig. 2), considered in Rasmussen et al. (2000) to probably be a good species pending further study. The song of sulaensis has less frequency modulation than atrifrons, but more than somadikartai. Calls.— M\ tape-recorded numerous call notes in the presence of Z. somadikartai, but none resemble known call notes of Z. a. su- laensis and no tapes were available for call notes of Z. a. atrifrons or Z. a. surdus for comparison. Small flocks of somadikartai gave twittering chirrups, evidently similar to those described by Watling (1983) for Z. a. surdus. The calls, uttered while the birds were moving, aided in their detection. Indrawan et al. • NEW SPECIES OF WHITE-EYE 5 C/5 c « 3 ^ C/5 s ° 3 N I T3 a i 3: £ (u ^ £ ^ c/3 S P Ni t£ ci O (u x: ^ 3 ^ £ -38 ^8 N- - o I ^ 3 S £ +1 1 li-; -■*- U i£ 3 S 5 P ~ c/3 3 3 ~ 1) ^ 3 CQ ^ < 5 H S in in vO — d d +1 +1 tn (N in d mmmmmininininin — H in m dodo cO (N (N rs d d d d d ^ +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 On in O in in — ■ d d \0 ^ ■3" O "d d d 3- in in ON m ^ O ON m 3; p O p 3; 00 < 1“ — Ol O', P 3 ,3 p ^ ^ UQ3UJ>£l,CuQu£udt-S3:ZCH 3 .3 C c/3 — J2 = p O >5 y. y. X X V 3 o 3 - £ O p oi O', p ,r3 .3 O 4J X X X 3 6 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 FIG 2 Sonagrams of (A) Z somadikartai, recording by MI; (B) Z. atrifrons surdus, NSA 32916, Lore Lindu, Sulawesi, recording by Alan Greensmith; (C) Z. atrifrons surdus, NSA 45860, Lore Lindu, Sulawesi, recording by Steven Smith; (D) Z. atrifrons (subspecies uncertain, either surdus or atrifrons), NSA 13189, Sulawesi, recording by R. J. Watling; (E) Z. atrifrons surdus, NSA 1 13121, Sulawesi, recording by Clide Carter; (F, G) two recordings of Z. a. sulaensis, NSA 73891, Taliabu, Sula Islands, by Robert Lucking. Rationale for Consideration as a Full Biological Species The alpha taxonomy of few groups of birds is as difficult and ambiguous as that of the Zosterops white-eyes, given the large number of mostly allopatric taxa and the small number of morphological characters that vary among taxa. Taxa from widely disparate parts of the world can appear extremely similar without necessarily indicating close relationships. Nu- merous similar-looking white-eye taxa that were long treated as conspecific have recently been found to have high levels of genetic di- vergence inconsistent with subspecies status, for example among Micronesian (Slikas et al. 2000) and African and Indian Ocean taxa (Warren et al. 2006). In at least one African case, taxa formerly united as conspecific are not even sister taxa, instead being widely sep- arated on the phylogenetic tree (Warren et al. 2006). The former unification of many dissim- ilar taxa under Black-crowned White-eye {Zosterops atrifrons) is the only case of which we are aware in which just one “subspecies” {minor) of a polytypic species lacks a white eye ring, but minor was subsumed under atrif- rons on the basis of a single putative hybrid specimen and is best treated as a separate spe- cies (Rasmussen et al. 2000). In the case of Z. somadikartai, it might be argued that it is just a strongly marked subspecies, but we con- sider that its level of distinctiveness is equiv- alent to that of Black-fronted White-eye (Z. minor) and numerous other taxa now usually treated as full species. In our view, the lack of an eye ring, the iris and bill color differ- ences, and the divergence in song would like- ly act as isolating mechanisms. Hence we choose to describe somadikartai at the level of full species under the Biological Species Concept. Distribution Zosterops somadikartai is endemic to the Togian Islands, Gulf of Tomini, Sulawesi. It Indrawan et al • NEW SPECIES OE WHITE-EYE 7 has been recorded from several sites on Mal- enge Island (00° 15' S, 122° 03' E), Binuang, Talatakoh Island (00°21'S, 122° 06' E), and two sites at Batudaka Island (00 28 S, 121 48' E). All records except for the holotype are sight records. All localities were below 100 m asl. Several observations of the new white-eye were made by MI and Sunarto in the Togian Islands prior to the collection of the type spec- imen. 23 and 27 August 1996, and 21 June 1997 (Malangkat, Malenge Island). — Two, three, and two birds, respectively were observed traveling and foraging on low bushes (proba- bly Lantana camara) near mangroves {Avi- cennia and Sonneratia spp. behind the mud- living Rhizophora sp. community). The ob- servations lasted between 30 sec and 1 min each from a distance of —20 to 40 m. 12 June 1997 (Malenge Village, Malenge Island).— Tsno birds were seen moving through a garden with coconuts (Cocos nuci- fera), other trees, and bushes. The observa- tions were for about 30 sec from a distance of -30 m. 23 and 28 June 1997 (Binuang, Talatakoh Island, 00°21'S, 122° 06' E).— Three and two birds, respectively, were seen foraging in re- generating scrub at a logged-over forest site. Each group was observed for about 2 min each at a distance of —10 and 20 m. 24 May 2001 (Tanempo, Batudaka Island, 00°28'S, 121° 48' E).— One white-eye was briefly detected as it gleaned on exposed Lan- tana branches at 2-m height for —5 sec, before being chased into the scrub by another white- eye. It perched and gleaned at a distance of — 15 to 20 m from us for about 3 sec. In this fleeting glimpse the bird appeared to have a narrow streak of white above the eye, al- though it clearly did not form an eye ring. This observation on Batudaka might differ from the others and confirmation is needed. 22 January' 2003 (Botnha, Batudaka Is- land).— A possible additional record on Ba- tudaka was made by Yunus Masala. Two white-eyes were observed calling and mating in a coconut grove on 22 January 2003. The birds were identifled as Zosterops anomalus, based on their lack of a white eye ring (Y. Masala, pers. comm.). However, given the lo- cality they were almost certainly Z. somadi- kartai. In each of the above sightings by MI and Sunarto, the birds had a black mask covering the eye and cheek, not reaching the back of head or crown. They lacked a distinct white eye ring. The upperparts were olive and the tail dark. The throat was bright yellow, con- trasting with the pale breast and underparts. The vent was yellow like the throat and con- trasted with the rest of the underparts. The ex- tent of yellow on the vent was moderate. The bill appeared black and the feet were grayish. These characters would have identified the birds as Z. atrifrons, except that no white eye ring was visible in all the birds observed, even from close range. The observations were of at least nine birds from four different localities on three islands within the Togian Islands. Obserx’ations of White-eyes in the Banggai Islands. — MI made four sightings of at least three different white-eyes in the scrubby gar- dens of Kalopapi (Bangkurung or Bangkulu Island, 01°50'S, 123° 06' E; BAKOSUR- TANAL [1993]) in the Banggai Islands. In these observations, the eye ring was strikingly large, resembling that of Z. a. sulaensis. This contrasted to MTs earlier observations (Indra- wan et al. 1997) of at least eight individual white-eyes in Peleng and Banggai (Banggai Island group, which is closer to the Sulawesi mainland) in which the eye ring appeared less extensive, resembling Z. a. atrifrons and Z. a. surdus of mainland Sulawesi. The four AMNH specimens from Peleng, Banggai Is- lands (race Z. a. subatrifrons) are somewhat variable in eye ring size but are intermediate between sulaensis and atrifrons. Birds appar- ently with large eye rings were also observed on Banggai Island itself in 1981, but were identified as Zosterops cf. chloris (Bishop 1992). It may well be that a distinct form oc- curs on the outlying island of Bangkurung, or this population may be indistinguishable from sulaensis, but further observations and prob- ably specimen collection will be necessary. The white-eye observations for Bangkurung were mostly in the hills, above 100 m, where the bird appeared to be moderately common in small flocks, in contrast to .somadikartai of the Togian Islands. Observations of White-eyes in the Eastern Peninsula of Sidawesi. — It is not clear trom 8 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 1, March 2008 the literature (e.g.. White and Bruce 1986) whether any form of athfrons is known from the eastern peninsula of Sulawesi. However, MI observed white-eyes in mixed garden shrubs in Tanjung Api Nature Reserve (00° 50' S. 121° 37' E) in the eastern peninsula of Sulawesi in 2003. Four birds from different flocks that MI observed fairly well had a clear white eye ring, and other characters (black face mask, olive mantle, and dull yellow throat) identifying them as Zosterops atrif- rons. Their bills appeared dark and their legs gray. Two AMNH specimens recently collect- ed from Siuna (00° 45' S, 122° 58' E), close to the base of the Balantak Mountains near the eastern tip of east-central Sulawesi, are similar in appearance to nominate atrifrons. Ecology and Behavior The ecology of birds and habitats in the To- gian Islands is described in Indrawan et al. (2006). Monsoon and evergreen forest, inter- spersed with mixed gardens of coconut, clove {Syzygiion aromaticiim), cacao {Theobroma cacao), and durian {Durio zibethinus) char- acterize the general vegetation of the island group. The habitats in which we found Z. so- madikartai range from mangrove to secondary vegetation and mixed gardens, all in the low- lands (<100 m asl). We did not detect the spe- cies at higher altitudes in our survey. Zosterops somadikarlai is gregarious, for- aging in twos and threes, and it roosts in groups of up to five birds. The birds often foraged actively throughout the day until about sunset before retiring to roost. They fed on insects, including caterpillars, frequently by gleaning them from branches and below leaves. They often foraged in dense shrubs, possibly because these microhabitats had more insects. The roosts were typically in shrubs 5 to 15 m high, such as Macaranga and larger bamboos, in which they perched at 4 to 10 m above ground. An observation was also made of two birds roosting in the mid- canopy of mangroves at 1130 UTC + 8. Conservation Status Our limited observations suggest that Zos- terops somadikartai may be localized and un- common. The bird definitely occurred in at least three of the six main islands of the To- gian Islands, namely Malenge, Talatakoh, and Batudaka (Fig. 1). It has not been found on Togian and the Walea islands, although its presence on these islands could have been overlooked. That all previous collectors and observers failed to detect this species supports the conclusion from our fieldwork of its scar- city. The present surveys allow us to suggest an lUCN threat category (lUCN 2001) for Z. so- madikartai. The extent of occurrence of the entire population is <5,000 km-. The popu- lation appears to be severely fragmented and the species is known to exist on just three is- lands. Our surveys also project that area, and extent and quality of habitat are likely to de- crease further due to conversion and overex- ploitation of resources. We believe the taxon should be assigned “Endangered” status based on these lUCN criteria (EN, B, 1, a, b, iii). However, it does occur in some anthro- pogenic habitats, and further fieldwork is clearly needed to establish its population size and ecological requirements. Endemic Bird Areas (EBAs) encompass the complete ranges of two or more restricted- range species, defined as species with an es- timated total global breeding range below 50,000 km- (ICBP 1992, Stattersfield et al. 1998). The Togian Islands merit EBA status, given the presence there of another recently discovered endemic, Togian Hawk Owl (M- nox burhani) (Indrawan and Somadikarta 2004). The recently created Togean Islands National Park should afford these new bird species some of the protection necessary for their survival. Note Added in Proof. — Zosterops somadi- kartai was observed in two different areas of Malenge Island on 14 and 15 December 2007, and a freshly dead individual was found just outside Malenge village (Raphael Jordan and Benoit Segerer, in litt.). This bird was not sal- vaged but a digital photograph shows it to be similar to the holotype. There is a large area of bare dusky circumorbital skin, bluish in front of and above the eye and below and be- hind the eye, with a few tiny white feathers forming barely visible white specks around the upper and lower front edges of the eye (but not on the cranial or caudal edges and not nearly forming a complete ring). Iris dark red; base of lower mandible pale pinkish horn. Indrawan et al. • NEW SPECIES OE WHITE-EYE 9 contrasting with the black distal two-thirds; tarsi, toes, and claws light bluish-silver, joints slightly darker; undertail coverts pale clear yellow; and rectrices dusky black with indis- tinct yellowish-green edgings. ACKNOWLEDGMENTS This study was undertaken at the behest of Dr. Soe- karja Somadikarta, who also kindly commented on the manuscript. We thank the Biology Research Center Indonesian Institute of Sciences for research permis- sion. We thank Dr. S. N. Prijono, Don Darjono, and Sudaryanti for sending the specimen from the Museum Zoologicum Bogoriense to the USA for study, and for their patience with the length of the loan period. Agus Prijono graciously painted the Plate of the type spec- imen and Yunus Masala provided data on a sighting. We thank the staff of the American Museum of Natural History (AMNH), The Natural History Museum (BMNH), Naturalis, Leiden (RMNH), the National Museum of Natural History, Smithsonian Institution (USNM), and Michigan State University Museum for access to and/or loan of specimens. Eieldwork in the eastern peninsula and offshore islands has been sup- ported by grants, including Nagao — Natural Environ- ment Foundation (NEE) and the Oriental Bird Club. An Earthwatch Expeditions grant (awarded to Dr. Jatna Supriatna) further augmented the funding support. We thank D. S. Rosenstein and J. G. Sikarskie for radio- graphs, and Cheryl Tipp and Richard Ranft for gen- erously providing recordings of other species. LITERATURE CITED BAKOSURTANAL [Indonesian Surveillance and Mapping Agency, in collaboration with DISH- IDROS TNI AL — Hydro-oceanographic Office OF THE Indonesian Navy]. 1993. Map of national marine environment Number 24. Central Sulawesi (scale 1:500,000). BAKOSURTANAL, Cibinong, Indonesia. Bishop, K. D. 1992. New and interesting records of birds in Wallacea. Kukila 6:8-34. Coates, B. J. and K. D. Bishop. 1997. A guide to the birds of Wallacea. Dove Publications, Alderley, Queensland, Australia. Dickinson, E. C. (Editor). 2003. The Howard and Moore complete checklist of the birds of the world. Revised and enlarged. Third Edition. Christopher Helm, London, United Kingdom. Holmes, P. R. and H. J. Holmes. 1985. Notes on Zos- terops spp. from the Lake Matano area of south- east Sulawesi, Indonesia. Bulletin of the British Ornithologi.sts’ Club 105:136-139. ICBP. 1992. Putting biodiversity on the map: priority areas for global conservation. International Coun- cil for Bird Preservation, Cambridge, United Kingdom. Indrawan, M. and S. Somadikarta. 2004. A new spe- cies of hawk-owl from the Togian Islands, Gulf of Tomini, central Sulawesi, Indonesia. Bulletin of the British Ornithologists’ Club 124:160-171. Indrawan, M., Y. Masala, and L. Pesik. 1997. Recent observations in the Banggai Islands, Sulawesi. Kukila 9:61-71. Indrawan, M., S. Somadikarta, J. Supriatna, M. D. Bruce, Sunarto, and G. Djanubudiman. 2006. The birds of the Togian Islands, Central Sulawesi, Indonesia. Forktail 22:7-22. lUCN. 2001. lUCN red list of threatened species. 2001. Categories and criteria (Version 3.1). lUCN, Gland, Switzerland, www.redlist.org/info/categories_ criteria2001.html (accessed 2 April 2006). Mees, G. E 1961. A systematic review of the Indo- Australian Zosteropidae (Part II). Zoologische Verhandlingen 50:1-168. Owen, J. 2004. Mystery bird discovered on Indonesian island. National Geographic News. news, nationalgeographic. com/news/2004/0 1 /0 1 26_ 040126_wangiwangi.html (accessed 20 January 2006). Rasmussen, P. C., J. C. Wardill, E R. Lambert, and J. Riley. 2000. On the specific status of the San- gihe White-eye Zosterops nehrkorni, and the tax- onomy of the Black-crowned White-eye Z. atrif- rons complex. Forktail 16:69-81. Slikas, B., I. B. Jones, S. R. Derrickson, and R. C. Fleischer. 2000. Phylogenetic relationships of Micronesian white-eyes based on mitochondrial sequence data. Auk 117:355—365. Smithe, E 1975. Naturalist’s color guide. American Museum of Natural History, New York, USA. Stattersfield, a. j., M. J. Crosby, A. J. Long, and D. G. Wege. 1998. Endemic bird areas of the world. BirdLife Conservation Series Number 7. BirdLife International, Cambridge, United King- dom. Warren, B. H., E. Bermingham, R. P. Prys-Jones, AND C. Thebaud. 2006. Immigration, speciation and extinction in a highly diverse songbird line- age: white-eyes on Indian Ocean islands. Molec- ular Ecology 15:3769-3786. Watling, D. 1983. Ornithological notes from Sula- wesi. Emu 83:247-261. White, C. M. N. and M. D. Bruce. 1986. The birds of Wallacea (Sulawesi, The Moluccas & Les.ser Sunda Islands, Indonesia). An annotated check- list. B.O.U. Check-list Number 7. British Orni- thologists’ Union, London, United Kingdt)m. The Wilson Journal of Ornithology 120( 1 ): 10-25, 2008 THE WHITE-EYED FOLIAGE-GLEANER (FURNARIIDAE: AUTOMOLUS) IS TWO SPECIES KEVIN J. ZIMMER'-^ ABSTRACT. New information on vocalizations of populations of the White-eyed Foliage-gleaner (Auto- molus leucophthalmus), and analysis of biometric and plumage characters, reveal that it consists of two biological species. One form is restricted to the Pernambuco Center of Endemism in coastal northeastern Brazil and the second occupies much of the remainder of humid Atlantic Forest from Bahia, Brazil south to northeastern Argentina and eastern Paraguay. The northeastern form, although cryptically similar morphologically to other subspecies of Automoliis leucophthalmus, is highly differentiated in several vocal characters. The vocal difference between the two groups exceeds that between other accepted species pairs within the genus. Reciprocal tape- playback experiments offer supporting evidence that vocal differences within the White-eyed Foliage-gleaner complex are sufficient to act as isolating mechanisms in the event of secondary contact between the two groups. The northeastern form is shown to have vocal and morphological similarities to the Para Foliage-gleaner {Au- tomolus paraensis) of southeastern Amazonian Brazil, supporting hypothesized Amazonian origins of the leu- cophthalmus complex, as well as a sister relationship with the Automolus infuscatus! paraensis complex. Received 6 December 2006. Accepted 12 May 2007. The genus Automolus is a group of fairly large, mostly drab plumaged foliage-gleaners (subfamily Philydorinae), which, as currently recognized (Zimmer 2002, Remsen 2003), comprises eight species {ochrolaemus, infus- catus, paraensis, leucophthalmus, melanope- zus, roraimae, rubiginosus, and rufipileatus). Members of the genus are distributed widely through tropical South America and Central America (Remsen 2003). The White-eyed Fo- liage-gleaner (A. leucophthalmus) has a dis- junct distribution relative to all other conge- ners (which do not extend south of Amazon- ia), occurring in the Atlantic Forest region of eastern Brazil, eastern Paraguay, and north- eastern Argentina (Remsen 2003). Originally described as Anabates leucophthalmus by Wied-Neuwied (1821) from specimens col- lected at Rio do Cachoeira, Bahia, Brazil, its range was subsequently considered to extend south through much of southeastern Brazil to Paraguay and Misiones, Argentina (Hellmayr 1925). Cory (1919) and Hellmayr (1925) not- ed that specimens of leucophthalmus from the southern part of the range were smaller and paler than those from Bahia. J. T. Zimmer (1947) reported the oldest name applicable to the smaller, paler birds from south of Bahia ' Los Angeles County Museum of Natural History, 900 Exposition Boulevard, Los Angeles, CA 90007, USA. 2 Current address; 1665 Garcia Road, Atascadero, CA 93422, USA; e-mail: kjzimmer@charter.net was sulphurascens (Sphenura sulphurascens Lichtenstein 1823). Concomitantly, he de- scribed a new subspecies, lammi, from a male collected at Recife, Pernambuco, Brazil by Donald Lamm. Subsequent authors (e.g., Pe- ters 1951, Remsen 2003) followed Zimmer (1947) in recognizing sulphurascens and lam- mi as subspecies of Automolus leucophthal- mus. Andrew Whittaker and I were conducting field work in January 1996 in Alagoas, Brazil and noticed the local population of White- eyed Foliage-gleaners (A. /. lammi) varied dra- matically in songs and calls from populations with which we were familiar in the Brazilian states of Espfrito Santo, Rio de Janeiro, Sao Paulo, and Parana. The songs of the Alagoas birds were more similar to those of birds then treated as Automolus infuscatus paraensis (Ol- ive-backed Foliage-gleaner) of southeast Amazonia than they were of other populations of A. leucophthalmus (Zimmer 2002). I was intrigued by the vocal distinctiveness of birds that appeared, under field conditions, to be “typical” White-eyed Foliage-gleaners in plumage characters, and began investigating vocal and morphological variation among all of the recognized subspecies in the complex. The objectives of this paper are to: (1) pre- sent new evidence of vocal differences among populations of White-eyed Foliage-gleaners, and (2) re-evaluate geographic variation in plumage and biometric characters within the 10 Zimmer • WHITE-EYED EOLIAGE-GLEANER SPECIES LIMITS 11 species. I demonstrate the subspecies A. 1. lammi is sufficiently differentiated vocally to be considered a biological species distinct from other members of the group. METHODS I observed and tape recorded “White-eyed” Foliage-gleaners at sites throughout the Bra- zilian states of Alagoas, Pernambuco, Sergipe, Bahia, Espirito Santo, Minas Gerais, Rio de Janeiro, Sao Paulo, Parana, and Santa Catarina on multiple occasions from 1996 to 2006. All measurements used in behavioral data (dis- tances, heights, etc.) are estimates. Mapped distributions are based on label data from specimens that I examined, and by more re- cent records documented with tape recordings. These localities, along with the type localities for each taxon, were entered into a geographic information system (Isler 1997) and mapped by M. L. Isler. I assume that vocalizations of foliage- gleaners, like those of other suboscines, are mostly or entirely inherited (Kroodsma 1984, 1989; Kroodsma and Konishi 1991; Remsen 2003) and provide potentially important char- acters for systematic study (Lanyon 1978, Is- ler et al. 1997, Krabbe and Schulenberg 1997, Zimmer 1997). I assembled recordings of Aw- tomolus leucophthalmus from across its geo- graphic range to analyze vocalizations (Ap- pendix 1). Recordings were assigned to sub- species by location following the reported dis- tribution of the three subspecies {lammi, leucophthalmus, sulphurascens) in current lit- erature (Remsen 2003). I reviewed each re- cording to identify the number of individuals vocalizing and to label every vocalization as to type. Vocalizations were categorized as loudsongs, calls, and long calls for compari- son. Loudsongs were consistently patterned multi-note vocalizations (Isler et al. 1997) giv- en seemingly in the context of territorial ad- vertisement. Vocalizations characterized as calls usually were structurally simple (typi- cally involving well-spaced repetition of iden- tical notes or pairs of notes), and most often were given in the context of contact notes be- tween mates, or as aggression calls during ter- ritorial conflicts with conspecifics, or in re- sponse to tape playback. Vocalizations cate- gorized as long calls were infrequently given series of mostly similar single note calls de- livered in rapid succession, most often seem- ingly as signals of extreme agitation. I made auditory comparisons of all record- ings and visually compared spectrograms of all individuals of A. /. lammi and A. /. leuco- phthalmus recorded, as well as of a large, geo- graphically balanced sample of recordings of A. /. sulphurascens. Songs selected for illus- tration as spectrograms were considered rep- resentative based on visual comparison of spectrograms of the large sample. I also ex- amined my own collection of recordings of songs and calls of A. rubiginosus, A. ochro- laemus, A. infuscatus, A. paraensis, A. rufi- pileatus, and A. melanopezus for qualitative comparison with the vocalizations of A. leu- cophthalmus. This involved both auditory comparison and superficial visual comparison of spectrograms. Visual (qualitative) characters were ob- tained through examination of printed copies of all spectrograms. I considered a character to be diagnostic visually when examination unambiguously distinguished one population from another. My tape recordings were made with a Sony TCM-5000 recorder with Senn- heiser ME-80, ME-66, and MKH-70 shotgun microphones. Spectrograms used in illustra- tions were made by R R. Isler, using Canary 1.2 and Raven 1.2 software (Bioacoustics Re- search Program, Cornell University Labora- tory of Ornithology, Ithaca, NY, USA), Can- vas graphics software (Version 5.0.3, Deneba Software, Miami, FL, USA), and a Hewlett Packard LaserJet 6MP printer. Playback experiments were conducted to examine reactions of the three taxa currently considered subspecies of A. leucophthalmus to vocalizations of the other subspecies. In each case, an individual or pair of foliage-gleaners was first presented with a prerecorded tape of another taxon. Each taxon was represented for all trials by a single playback tape consisting of 10-15 songs interspersed with calls from three different individuals from three different localities. The sulphurascens tape contained recordings of individuals Irom Parana, Sao Paulo, and Espirito Santo, Brazil. The lammi tape contained recordings ol individuals Irom two sites in Alagoas and one in Sergipe, Bra- zil. The leucophthalmus tape contained vocal- izations from two sites in Bahia. The tape ot 10-15 songs plus calls was played during 12 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 1, March 2008 each trial to conclusion. A buffer period of about 2 min was used to separate playback of vocalizations of different taxa to avoid con- founding effects. Foliage-gleaners that did not respond to playback of other taxa were pre- sented with a playback tape of their own vocal type. This was done to account for potential seasonal influences by resolving the question of whether birds that were unresponsive in the first trials were discriminating between the different vocalizations, or were generally non- territorial during the trial period and therefore unresponsive to playback of any kind. Re- sponses to playback were characterized as “strong,” “moderate,” “weak,” or “none.” A strong response involved immediate and re- peated vocalizations (from previously nonvo- calizing birds) as well as approach toward the sound source. A moderate response involved either a cautious approach without vocalizing, or sustained vocalizing without approach. A weak response involved single or unsustained vocalizations (from previously nonvocalizing birds) without approach. The none category includes instances in which birds remained si- lent (in the case of previously nonvocalizing birds) and did not approach the sound source, as well as instances in which already vocal- izing birds neither changed the delivery or rate of their vocalizations, nor approached the sound source. A potential fifth category would be a “negative response,” in which the sub- ject responded to the sound stimulus by mov- ing farther away. No negative responses to playback were observed from any taxa. I examined representative specimens of A. 1. leucophthalmus (n = 5), A. /. sulphurascens (n = 32), and A. /. lammi (n = 6) to identify morphological differences. These specimens are housed at the Academy of Natural Scienc- es of Philadelphia, Philadelphia (ANSP); Field Museum of Natural History, Chicago (FMNH); Los Angeles County Museum of Natural History, Los Angeles (LACM); and the National Museum of Natural History, Washington, D.C. (USNM) (Appendix 2). A wing ruler with a perpendicular stop at zero was used to measure wing chord (wing) and tail length (tail) while dial calipers were used to measure tarsus length (tarsus), culmen length from the anterior end of the nares to the tip (culmen), bill depth at the anterior end of the nares (bill depth), and bill width at the anterior end of the nares (bill width). All mea- surements with calipers were to the nearest 0.1 mm; those taken with the wing ruler were to the nearest 0.5 mm. Measurement terminology conforms to Pyle et al. (1987). Plumage was described from specimens and compared to a standard color reference (Smithe 1975). RESULTS Distribution. — The distributions of the three subspecies of Automolus leucophthalmus, as ascertained from examination of specimens and tape recordings, varied (Fig. 1). Habitat. — All forms of Automolus leuco- phthalmus occur in tropical lowland and foot- hill evergreen forest, and taller second-growth woodlands. All three subspecies occur near sea level. Nominate leucophthalmus was re- corded to 750 m, A. /. lammi to 550 m, and A. /. sulphurascens to 900 m elevation. Behavior. — All forms of Automolus leuco- phthalmus occupied the lower and mid levels of the forest strata. They were frequently, but not habitually, encountered as members of mixed-species flocks of insectivores com- prised primarily of woodcreepers, other fo- liage-gleaners, antshrikes, and antwrens. The Red-crowned Ant-tanager (Habia rubica) ap- peared to be a nuclear species in most flocks in which A. /. sulphurascens occurred in Sao Paulo and Parana. My observations of A. /. lammi, A. 1. sulphurascens, and A. /. leuco- phthalmus suggest they are dead-leaf search- ing specialists (>75% of all foraging maneu- vers involved searches of dead leaves for ar- thropods). Foliage-gleaners moved through vine tangles during these searches, going from one suspended dead leaf to another without scanning intervening live vegetation. Larger dead leaves were often vigorously pecked and pulled apart to get at arthropod prey concealed within the curls and folds. A. /. sulphurascens was observed on many occasions to rummage in leaves trapped in the bases of epiphytes, in the crowns of understory palms, and in the interiors of climbing bamboo thickets. Both A. /. sulphurascens and A. /. lammi were ob- served foraging to heights of 8-12 m, al- though most observations of all three forms were from much lower. Plumage. — The three subspecies of White- eyed Foliage-gleaner are so similar in plum- age characters that silent birds could easily Zimmer • WHITE-EYED EOLIAGE-GLEANER SPECIES LIMITS 13 FIG. 1. Distribution of “White-eyed” Foliage-gleaners {Automolus leucophthalmus) confirmed by exami- nation of specimens or tape recordings: stars = A. /. leucophthalmus', circles = A. /. sulphurascens', squares = A. /. lammi ; “T” next to a symbol indicates the type locality for that taxon; “S” next to a symbol indicates specimen confirmation for that taxon; “V” next to a symbol indicates vocal confirmation (tape recording) for that taxon; “B” next to a symbol indicates both specimen and vocal confirmation for that taxon; and “note 1” refers to a lack of a precise type locality for A. /. sulphurascens. defy identification in the field. Specimens of A. /. lammi in direct comparison usually were separable from both A. 1. leucophthalmus and A. 1. sulphurascens by their darker and brown- er (less rufescent) dorsal color (closest to col- or #121 A, Front’s Brown); darker and redder (less orange) tail (between color #34, Russet, and color #223A, Mars Brown); creamier (less white) throat; and dingier buff abdomen, which was brownest on the flanks, lower belly and vent (between color #239, Ground Cin- namon, and color #26, Clay Color, but closer to the latter). The two ANSP specimens of A. /. leucophthalmus were distinctly dirtier white on the throat (nearly grayish) with more red- dish-brown on the lower Hanks and abdomen than were any specimens of A. /. sulphuras- cens', these distinctions did not hold for the two USNM specimens of A. /. leucophthal- mus. Size. — Measurements for the three subspe- cies of White-eyed Foliage-gleaner varied (Ta- ble 1). Gender of four of the five specimens of nominate leucophthalmus was unknown and measurements for that taxon are not di- vided by male or female. Vocalizations. — 1 assembled recordings of 103 different White-eyed Foliage-gleaners ( 17 leucophthalmus, 68 sulphurascens, and 1 8 lammi), including 844 individual songs (75 leucophthalmus, 523 sulphurascens, and 246 lammi) and 2,058 individual calls and long calls (771 leucophthalmus, 992 sulphuras- cens, and 295 lammi). 14 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 S -3 5 ^ £; u D. rr~, +1 7 +1 7 4 ^• Tj- ^ ^ sC 3 'O +11+11+11 — oc n- P >C ^ >4 S' q T +1 +1 ir-, ^ ^ W3 +1 T +1 , ^ S c^ >4 3 P t- s :c s d s 4 s 4 Spectrographic analysis confirmed my field impressions that A. /. lammi differed strikingly and consistently in its loudsongs from both A. /. sulphiirascens and A. /. leucophthalmus . The loudsong of this subspecies (Fig. 2A-F) was a countable series of 2-1 1 closely spaced (3_4/sec), frequency-modulated notes or dou- blets, each of which had a particularly harsh, grating quality. Doublets were formed by the close pairing of harsh notes with sharp thin notes. The sharp notes were seldom discem- able in the field, except at close range when they were at times apparent as a sharp “hic- cup” prior to the first note of a song. How- ever, they are clearly visible in some spectro- grams (Fig. 2D) as chevron-shaped notes pre- ceding each frequency-modulated note. The frequency-modulated notes are readily identi- fied in the spectrograms because they cover a wide band width (the difference between the highest and lowest frequency within a note, either at a given point in time or over the en- tire course of the note), and produce a trace that resembles a “zig-zag” sewing machine stitch within the confines of the general shape of the note. The frequency-modulated notes in A. /. lammi songs appear as rectangular- shaped or as shallowly inverted-U-shaped notes in the spectrograms. They have a dis- tinctive strident quality to the human ear and it was this quality that allowed immediate identification of A. /. lammi songs in the field. The number of notes and length of songs varied within a song bout in the same indi- vidual. Typical songs contained 3-6 notes or doublets (92% of all songs) with 4 notes/dou- blets being most common (35% of all songs). Individuals within a song bout tended to sing 3-5 consecutive songs with an identical num- ber of notes before adding or subtracting notes. Twelve of 15 individuals for which loudsongs were recorded varied the number of notes/doublets per song within a recorded song bout. Responses to playback were within the same range of notes found in natural (un- solicited) songs. The peak frequency pattern within the songs was flat; that is, peak fre- quencies of successive notes in the song were roughly equal with little rise or fall. Occa- sional loudsongs contained a well-differenti- ated, longer first doublet (Fig. 2C), but these were the exception. Automolus species are sexually monochromatic and both males and Zimmer • WHITE-EYED EOLIAGE-GLEANER SPECIES LIMITS 15 6-j A A. 1. lammi loudsong 3 A. 1. lammi loudsong C A. 1. lanmti loudsong t, t l|. Ml' '■r 1 l|l)ll igiu iijiM 'iiiJiil 'l ... - t A A \ 3 ill !1 li :i rww V r MiW H H n 1 lilibllii. 1 K Ik lib! lik I|y iHiBj C 0.5 1.'o 1 .5 2.0 2.5 3.0 3.S i 4.0 4.5 5.( D A. L lammi loudsong E A. 1. latmti loudsong F A. 1. lammi loudsong 5 - .n fh 2 2 A A A 4 Miff ffifflHlI iaZaaa - kAiiJ.lbl# li 1 ,m PW A C a 1 o.'s 1.0 -1 1.5 r 2.0 1 2.5 3.0 3.' i 4.0 4.5 5.1 G A. l^ lammi long call N O - m ^ A ^ ^ I A. t w ^ - — TT i ^ / (i. f fj U m B — n ri ¥ y u I a 2- B* 4) -1 — ^ IT fj . a E * u " ^ Eli ' 7 J u - ( c 3 0.5 1 'o — 1 1.5 2.0 1 2.5 3.0 3.' 5 4.0 4.5 5.1 O - c H A. L lanptfi^ Idn^ call 5 - A 1 rcT r , fl f !_L V 4 - ,1 j i J 5 / fl 1 n tf 1 fl M 1 ^ o - m M d 2 - if U U -J . ri ji u a J 1 - f F 0 - ( a D 0.5 1.'o ^ — 1.5 2.( 3 2.5 3.0 3.5 4.0 4.5 5. I A. 1. lammi call J A. 1. lammi calls 1 1 i li 1 1 / i I 1 f \ Q f f i 1 G 1 — 1 1 \ “ f f] 1 1 ^ — I 1 ^ Time (see) FIG. 2. Types of vocalizations of Automolus leucophthalmus lammi: (A) loudsong from Sergipe, Brazil; (B) short loudsong from Sergipe, Brazil; (C) loudsong from Sergipe, Brazil; (D) loudsong from Alagoas, Brazil; (E) loudsong from Alagoas, Brazil; (F) loudsong from Alagoas, Brazil; (G) long call from Sergipe, Brazil; (H) long call from Sergipe, Brazil; (I) two-note call from Alagoas, Brazil; and (J) two-note and one-note calls from Sergipe, Brazil. All spectrograms from recordings by K. J. Zimmer. females sing; thus, the sexual identity of sing- ing birds could not be ascertained. However, when presumed mated pairs responded to playback, the songs of each member of the pair were similar in pattern and quality to one another, and differed only slightly in peak fre- quency. The loudsongs of A. /. siilphurascens (Fig. 3A-D) in contrast, consisted of a barely countable, faster-paced (5-7 doublets/sec) se- 16 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 6nr A A 1. suiphurascms [oudsoag B A. L sulphurascens Doudsong 5 - ■ ; 4 -j ! L !ll. 1 A ^ rt *-i LA./iA-H 3 - W 2 - u 1 - iV^ ly ‘ Y Y ^ 0 -I c ) 0.5 1.'o 1.5 2.' 0 2.5 3.0 3.5 4.0 4.5 5.( 6 - C A. 1. sulphurascens toudsong D A.L sulphurascens toudseng 5 - ■, '| \ “ \ \ J 1 4 - — yr’ i II il llll ''"i '1, 'li llii^ 3 - I — Mil MfI ><| i/l ^ 2 - N aA aA a a a a r iHi 1 ] ' ^ i 1^' ih *1^ i./^ — H 1 - n _ jjt — n — h4-i 0.5 LO 1-5 2 0 2.5 3.0 3.5 4.0 4.5 5.0 E A.L sulphurascats tong call F A. 1. sulphurascens long call ]■ i ( k 1 .1 i , J « fl : : ' fl f, 1 .*/ Ui\ .7^ - J m ff[ fi f I A k- — At / / , 1^/ f U t % ~ I 'A 1 1 J f P- 'J Ji J LUl r f i f f 1 f < ^ J >— f d 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 G A. 1. sulphurasexns calls II A, L sulphurascens call I A. L sulphurascens call s I 1 — r^i — I ! 1 — !-i 4 p— TD zz — t\ ^ n — - — — -j — ~ — —1 — — 1 — ^ — “ f ' 0 0.5 1.0 1.5 2.0 2.5 3. 0 3.5 4.0 4.5 5.0 FIG. 3. Types of vocalizations of Automolus leucophthalmus sulphurascens: (A) loudsong from Ubatuba, Sao Paulo, Brazil; (B) loudsongs of presumed mated pair from Ubatuba, Sao Paulo, Brazil; (C) loudsong from Iguacu, Parana, Brazil; (D) loudsong from Linhares, Espirito Santo, Brazil; (E) long call from Ubatuba, Sao Paulo Brazil; (F) long call from Linhares, Espirito Santo, Brazil; (G) two-note call from Ubatuba, Sao Pau o, Brazil; (H) one-note call from Ubatuba, Sao Paulo, Brazil; and (I) one-note and two-note calls from Iguacu, Parana, Brazil. All spectrograms from recordings by K. J. Zimmer. ries of 2-12 closely spaced doublets. The in- dividual notes which were clear in tone lacked the frequency modulation that defined the component notes of A. /. lammi loudsongs. The doublets of A. /. sulphurascens loudsongs consisted of two closely paired chevron- shaped notes, the second of which was typi- cally higher in peak frequency than the first and longer in length. Visually diagnostic dif- ferences in note shape and clarity (frequency- modulated vs. clear) between the two taxa were apparent in all spectrograms. Loudsongs of A. /. sulphurascens, as with A. /. lammi. exhibited a fairly flat peak frequency pattern and only occasionally (Fig. 3D) had well-dif- ferentiated, longer introductory doublets. The number of notes and the length of loud- songs in A. /. sulphurascens, as with A. /. lam- mi, varied within a song bout in the same in- dividual. Typical loudsongs contained 4-9 doublets (94% of all songs) with 5 doublets being most common (31% of all songs). In- dividuals within a song bout tended to sing 3-5 consecutive loudsongs with an identical number of notes before adding or subtracting notes. Forty-nine of 51 individuals for which Zimmer • WHITE-EYED FOLIAGE-GLEANER SPECIES LIMITS 17 FIG. 4. Types of vocalizations of Automolus leucophthalmus leucophthalmus: (A) loudsong from Pto. Se- guro, Bahia, Brazil; (B) loudsong from Fazenda Palmeiras, Bahia, Brazil; (C) loudsong from Fazenda Palmeiras, Bahia, Brazil; (D) one-note and two-note calls from Pto. Seguro, Bahia, Brazil; and (E) one-note and two-note calls from Fazenda Palmeiras, Bahia, Brazil. All spectrograms from recordings by K. J. Zimmer. loudsongs were recorded varied the number of doublets per song within a recorded song bout. Responses to playback were within the same range of notes in natural (unsolicited) loudsongs. Males and females of presumed mated pairs sang nearly identical loudsongs that differed only slightly in note shape and mean peak frequency (Fig. 3B). The smaller sample (10 individuals; 75 total songs) of A. /. leucophthalmus loudsongs (Fig. 4A-C) was visually indistinguishable in note shape, clarity, and peak frequency pattern from those of A. /. sulphurascens. Loudsongs of nominate leucophthalmus consisted of 2—8 closely spaced doublets with typical loud- songs containing 4-8 doublets (96%) and 5-doublet loudsongs being most common (33%). Seven of 10 individuals for which loudsongs were recorded varied the number of doublets per song within a recorded song bout. Long calls also differed noticeably between A. /. lammi and A. /. sulphurascens. Long calls were highly variable between and within in- dividuals of the same subspecies. Most of the variation involved the number of notes (8-20) which, in turn, often was seemingly inllu- enced by extent of agitation. Birds that were particularly upset typically gave longer calls. Long calls of A. /. lammi consisted of a series of loud, slightly liquid sounding week notes, the first several of which were closely and evenly spaced, followed by a slight gap and then another shorter cluster of notes before tailing off with two or more widely spaced individual notes (Fig. 2G-H). Notes in the ter- minal half of these long calls were longer and more drawn-out {week week week week week weekk weekk weekk weekkk weekkk). In spec- trograms, the shape of these terminal notes ap- pears as a letter “Z” turned on its side, where- as the introductory notes appear more as in- verted check marks with the longer tail at the beginning of the note (Fig. 2G-H). Long calls of A. /. sulphurascens typically began with a prelude of 1-3 differentiated notes that rose and fell in frequency and then settled into an uncountable series of 5-10 identical, closely- spaced notes, before decreasing in frequency and slowing in pace with two or more slightly differentiated terminal notes (Fig. 3E-F). The variation in frequency of the introductory notes imparted the effect of a sputtering gas motor just before it begins to burn on all cyl- inders. No prelude was apparent in long calls of A. /. lammi, whose component notes showed a nearly flat frequency pattern over the duration of the call. Note shapes varied visibly between the two subspecies. Both forms had elements that appeared on spectro- 18 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 grams as inverted check-marks or chevrons, but in A. /. lammi the tops of these note-trac- ings are more rounded and open, whereas in A. /. sulphurascens they appear as more point- ed and closed. The difference is apparent to the human ear with the notes of A. I. laimni sounding more liquid in quality, whereas those of A. /. sulphurascens sound sharper. Some long calls of A. /. sulphurascens also contained several strongly upslurred notes whose tracings appear as nearly vertical sweeps lacking any tail (Fig. 3E); no similar notes were apparent in long calls of A. /. lam- mi. The limited sample of A. /. leucophthal- nius recordings lacked long calls for compar- ison. Other calls of the three subspecies were more similar. All three subspecies frequently gave single-note kwek calls and double-noted k\vek-kwaah calls (Figs. 21— J, 3G— I, 4D— E), mostly as contact calls between separated pair members, but also as agonistic calls in re- sponse to tape playback. Both members of presumed mated pairs gave nearly identical calls which, if anything, differed only slightly in pitch. These vocalizations were reminiscent in quality to calls of species of Synallaxis spinetails (e.g., Synallaxis rutilans) as well as those of the Olive-backed Foliage-gleaner (Automolus infuscatus) (K. J. Zimmer, tape re- cordings). Tape Playback Experiments.— Playhsick ex- periments with recordings of Automolus offer further evidence of the significance of vocal differences among the taxa and their role as potential isolating mechanisms. I performed tape playback experiments on 32 individuals or pairs of sulphurascens in Parana (12 in Sep 2003, 6 in Oct 2004, and 3 in Sep 2005) and Espirito Santo (7 in Oct 2003 and 4 in Oct 2004), Brazil. All birds were first presented with recordings of lammi. None of the 32 in- dividuals or pairs of sulphurascens indicated any response to tape playback of vocalizations of lajnmi. Birds were next presented with re- cordings of nominate leucophthalmus. Thirty of 32 individuals or pairs of sulphurascens re- sponded strongly to tape playback of leuco- phthalmus vocalizations and, in each instance, these same individuals and pairs responded strongly to playback of sulphurascens vocal- izations as well. Birds responded in each case by vocalizing immediately followed by direct and rapid approach toward the sound source. Birds that were already vocalizing prior to playback invariably switched to a different vocalization, either from loudsongs to calls, or from calls to loudsongs, and frequently deliv- ered a long call. In two instances (both in Pa- rana), individuals or pairs failed to respond to tape playback of any Automolus vocalizations. The birds involved on both occasions were moving with mixed-species understory flocks, and may have moved beyond the broadcast range of my tape recorder by the time I had completed playback of lammi vocalizations and commenced playback of leucophthalmus vocalizations. I performed tape playback experiments on six individuals or pairs of lammi in Sergipe (2 in Jan 2004) and Alagoas (4 in Jan 2004), Brazil. All birds were first presented with re- cordings of nominate leucophthalmus fol- lowed by recordings of sulphurascens. None of the six pairs or individuals indicated any response to either leucophthalmus or sulphur- ascens vocalizations. Birds were then pre- sented with recordings of their own vocal type. Five of six individuals or pairs respond- ed to playback with an immediate vocal re- sponse followed by direct and aggressive ap- proach toward the sound source. Already vo- calizing birds, as was the case in the sulphur- ascens trials, changed their vocalization to a different type and frequently gave long calls when responding to playback. One individual failed to respond to playback of any kind, even when it was in full view <20 m distant. Tape playback experiments were conducted on six individuals or pairs of leucophthalmus in Bahia, Brazil in February 2006. All birds were first presented with recordings of latnmi followed by recordings of sulphurascejis and nominate leucophthalmus. None of the six pairs or individuals indicated any response to playback of lammi recordings. Birds respond- ed strongly in each instance to playback of sulphurascens and leucophthalmus recordings with immediate vocal response followed by direct and aggressive approach toward the sound source. DISCUSSION Spectrographic comparison of vocal char- acters reveals that Automolus leucophthalmus lammi differs dramatically in several aspects Zimmer • WHITE-EYED FOLIAGE-GLEANER SPECIES LIMITS 19 of its vocal repertoire from all other popula- tions in the complex. The loudsongs of A. /. lammi differed diagnosably from loudsongs of the other two taxa in two visual characters; the shape of all individual notes, and in being strongly frequency modulated. Both distinc- tions are clearly evident in all spectrograms. Loudsongs of A. /. lammi were also consis- tently slower paced than those of the other subspecies, as evidenced by the obviously longer between-note intervals visible in the spectrograms. Differences in the long calls of A. /. lammi from those of A. /. sulphurascens were less striking, but still substantial, and in- cluded differences in individual note shape throughout the calls, flatter peak frequency pattern, and absence of well-differentiated in- troductory and terminal notes, all characters that were visually diagnostic in all spectro- grams. Isler et al. (1998) compared vocal characters between eight pairs of syntopic ant- bird (Thamnophilidae) species and found that within each species-pair, the members were diagnosable in a minimum of three different characters of the loudsong and calls. This standard is exceeded in a comparison of vocal differences between A. /. lammi and A. /. sul- phurascensUeucophthalmus. The syntopic ant- bird study and those subsequent (e.g., Isler et al. 2001, 2002, 2005) incorporated use of vi- sually diagnostic characters (including note shape, presence or absence of frequency mod- ulation, note shape pattern, and peak frequen- cy pattern) obtained from spectrograms as di- agnosable differences in assessing species limits. The vocal distinctions between A. /. lammi and other White-eyed Foliage-gleaners are even more noteworthy when the lack of geo- graphic variation within the remainder of the complex is considered. Tape recordings of A. /. sulphurascens from Igua^u, Parana, (across the Rio Iguagu from Misiones Province, Ar- gentina) and from Itapoa, Santa Catarina, both near the southern boundaries for the species, were indistinguishable from those of sulphur- ascens from northern Espirito Santo (near the northern geographic limit of sulphurascens) and from those of nominate leucophthalmus from Bahia. Remsen (2005), in his overview of the use of vocal characters in assessing questions of species limits, made it clear that broad geographic sampling from throughout a species’ range (not just from the endpoints) is needed. My inventory of White-eyed Foliage- gleaner recordings nearly spanned the north- south and east-west boundaries of the known range with representative recordings from many points in between. The biological and taxonomic significance of these vocal distinctions is suggested by the results of the playback experiments. Sample sizes were small for two of the pairwise tests, but those performed were unequivocal. None of the A. /. sulphurascens or A. /. leucophthal- mus tested showed any response to repeated playbacks of A. /. lammi, nor did any of the A. /. lammi respond to playbacks of either A. /. sulphurascens or A. /. leucophthalmus . One of the frequent criticisms of the experimental design and subsequent interpretation of play- back experiments is failure to account for sea- sonal differences in responsiveness (Kroods- ma 1986, Remsen 2005). This possibility was accounted for in my playback trials when over 93% of all birds tested (30 of 32 A. /. sul- phurascens, 6 of 6 A. /. leucophthalmus, and 5 of 6 A. /. lammi) responded strongly to play- back of their own vocal types, after ignoring playbacks of one or more of the other taxa. The lack of response from any taxon to songs of one of the others was not due to a seasonal lack of interest in territorial defense. The extent of divergence in vocal charac- ters between lammi and the other two taxa greatly exceeds that of morphological diver- gence. Although lammi is larger than sulphur- ascens in several mensural characters (partic- ularly culmen length), there is overlap in the ranges of most characters, and nominate leu- cophthalmus is similar in most measurements to lammi. Similarly, although plumage char- acters can allow separation of specimens with direct comparison, their utility in the field is uncertain at best. The lack of plumage dis- tinctions between forms that are so dillerent vocally should not be unexpected as the Fur- nariidae are known for their remarkable ho- mogeneity in coloration (Remsen 2003). Re- cent studies of neotropical suboscine passer- ines have revealed numerous cases of cryptic biodiversity in which distinct species-level taxa with dramatically different vocalizations were long overlooked because ol their mor- phological similarity to more widespread forms (e.g., Pierpont and Fitzpatrick 1983. 20 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. I, March 2008 Willis 1992a, Bierregaard et al. 1997, Krabbe and Schulenberg 1997, Zimmer et al. 2001). The distribution of A. /. lammi as a species- level taxon distinct from other members of the complex is consistent with demonstrated bio- geographic patterns. The entirety of its range is included in the Sierra do Mar center of en- demism as defined by Cracraft (1985) and the southeastern Brazil center of endemism as de- fined by Haffer (1985), areas which essential- ly include the entire ranges of A. /. leiico- phthalmus and A. /. sidphurascens. However, Cracraft (1985:72) noted that “This center will surely be subdivided into smaller subcen- ters when bird distributions are examined in more detail.” Other authors (Muller 1973, Jackson 1978, Kinzey 1982) had already rec- ognized (based partly on the distribution of primate and reptile taxa) three to four sepa- rate, smaller centers of endemism from coastal Brazil, one of which is the narrow belt of coastal and montane forests lying north of the Rio Sao Francisco, usually referred to as the “Pernambuco Center”. Thirty-eight bird taxa (including A. /. lammi) have been recognized as endemic to the Pernambuco Center (Sil- veira et al. 2003). Stattersfield et al. (1998: 277), in their compilation of “endemic bird areas (EBAs)”, recognized the “Atlantic Slope of Alagoas and Pernambuco” as an EBA comprising the Atlantic coastal forest in northeast Brazil, primarily in Pernambuco and Alagoas, the primary habitat of which is trop- ical evergreen forest. A distinct group of en- demic, range-restricted birds (Philydor novae- si, Myrmotheriila snowi, Terenura sicki, and Phylloscartes ceciliae) that are confined to tropical evergreen forests in Alagoas and Per- nambuco, primarily above 500 m, occur with- in this EBA. Three of these four species, all of which occur syntopically with A. /. lammi in portions of its limited range, have a pattern of geographic replacement by their presumed closest relatives {Philydor atricapillus, Myr- motherida unicolor, and Terenura maculata, respectively) in the Atlantic coastal forests be- tween southern Bahia and northern Rio de Ja- neiro (Remsen 2003, Zimmer and Isler 2003). This essentially parallels the replacement of A. /. lammi by A. /. leucophthalmus and A. /. sul- phurascens. Spectrographic and auditory comparison of vocal characters throughout Automolus sug- gests a close relationship between all mem- bers of the A. leucophthalmus complex and both A. infuscatus, the Olive-backed Foliage- gleaner, and A. paraensis, the Para Foliage- gleaner (Zimmer 2002). Single-note calls, double-note calls, and long calls of all three species are more similar to one another than they are to homologous vocalizations of other Automolus species (K. J. Zimmer, recordings). Loudsongs of A. paraensis are also similar in pattern to those of all members of the A. leu- cophthalmus complex and share with A. /. lammi the distinctive, frequency-modulated strident notes that are otherwise unique within the genus (Zimmer 2002). Willis (1988a) sug- gested a possible superspecies relationship be- tween A. leucophthcdmus and A. infuscatus based on certain vocal similarities, but did not elaborate. In the type description of lammi, Zimmer (1947:101) stated that: “The more extensively shaded and darker bill of this form may possibly indicate a trend toward Auto- molus infuscatus, one form of which, paraen- sis, occurs in the Para district of eastern Bra- zil. The darker and duller hues of the general coloration may possibly point in the same di- rection, but there is still too great distinction between the two groups to justify specific union.” Vaurie (1980), on the basis of mor- phological characters, also treated A. leuco- phthalmus and A. infuscatus as being more closely related to one another than to other members of the genus. More recently, Remsen (2003) proposed that A. leucophthalmus (in- cluding lammi) and A. infuscatus (including paraensis) formed a superspecies as suggested by voice, plumage pattern, and biogeography. The Pernambuco Center of Endemism has been considered an area of historical inter- change between Amazonian and Atlantic For- est floral and faunal elements (Silveira et al. 2003). Willis (1992b:7), in discussing the pos- sible Amazonian origins of Atlantic Forest taxa, specifically suggested a “northern cross- ing” in the Pernambuco region as a point of origin for Automolus leucophthalmus based on the “Pernambucan subspecies [=A. /. lammi] showing some intermediate features to A. in- fuscatus.'' The vocal similarities between A. /. lammi and A. paraensis (formerly A. infus- catus paraensis), the geographically most proximate member of the A. infuscatus com- plex, lend further support to Willis’ hypothe- Zimmer • WHITE-EYED FOLIAGE-GLEANER SPECIES LIMITS 21 sis of an Amazonian origin for the White-eyed Foliage-gleaner complex and of the suggested sister relationship of the infuscatus group. TAXONOMIC CONCLUSIONS Automolus leucophthalmus lammi differs substantially in several characters of its loud- songs and long calls from the other subspecies in the complex. It also differs morphologically from the other members of the group, al- though these distinctions are subtle. The extent of the vocal differences is more than sufficient to support the recognition of A. /. lammi as a separate species under any of the widely accepted species concepts (McKitrick and Zink 1988). The most difficult species concept to satisfy is the Biological Species Concept (BSC) because A. /. lammi is allo- patrically or parapatrically distributed with re- spect to other members of the complex. A pri- mary challenge in applying the BSC to allo- patrically distributed taxa is the need to judge whether or not the taxa are sufficiently differ- entiated as to prevent extensive hybridization in the event of secondary contact. Johnson et al. (1999) advocated using the extent of dif- ferentiation between accepted biological spe- cies in the same genus as a measure for as- sessing whether a taxon has diverged suffi- ciently to be considered a separate species un- der the Biological Species Concept. Zimmer (2002) presented spectrographic comparisons of loudsongs of Chestnut-crowned {Automolus rufipileatus) and Olive-backed foliage-glean- ers, two species widely sympatric in Amazon- ia, as an appropriate measure of species-level vocal differentiation within the genus. The loudsongs of A. rufipileatus and A. infuscatus were far more similar to one another than are the loudsongs of A. /. lammi to those of A. /. leucophthalmus and A. /. sulphurascens in note shape, note shape pattern, presence or ab- sence of frequency modulation to notes, be- tween-note intervals, and pace. Results of playback experiments offer supporting evi- dence that vocal differences within the White- eyed Foliage-gleaner complex are sufficient to act as isolating mechanisms in the event of secondary contact between the two groups {lammi vs. leucophthalmus and sulphuras- cens). 1 could find no diagnosable differences be- tween A. /. leucophthalmus and A. /. sulphur- ascens. The smaller sample of nominate leu- cophthalmus vocalizations were indistinguish- able from the large sample of sulphurascens vocalizations. Biometrics and plumage data suggest that leucophthalmus is slightly larger and darker than sulphurascens, but clinal var- iation in the extent of color saturation within sulphurascens has been noted (Remsen 2003). The number of specimens of nominate leu- cophthalmus available was too small for either statistical analysis or meaningful qualitative generalizations. Until more specimens become available, I suggest that nominate leuco- phthalmus and sulphurascens continue to be treated as valid subspecies. Therefore, I sub- mit the complex consists of two biological species: Automolus lammi Zimmer-Pemambuco Fo- liage-gleaner Automolus leucophthalmus (Wied-Neu- wied)- White-eyed Foliage-gleaner A. /. leucophthalmus (Wied-Neuwied) A. /. sulphurascens (Lichtenstein) The English name chosen for A. lammi highlights the Pernambuco Center of Ende- mism where the species is endemic, as well as one of the four Brazilian states to which it is restricted. I have chosen to retain the estab- lished name of “White-eyed Foliage-gleaner” as the English name for the other two taxa in the complex. CONSERVATION IMPLICATIONS The White-eyed Foliage-gleaner is uncom- mon to common throughout its range and is considered “not globally threatened” (Rem- sen 2003). Its range is more extensive than most Atlantic Forest endemics and includes a number of protected parks and reserves (Rem- sen 2003). Furthermore, it persists in taller second-growth woodlands, which also bodes well for its survival (Remsen 2003). This as- sessment pertains only to the widespread sub- species sulphurascens. Nominate leucophthal- mus remains a poorly known bird from a handful of localities in Bahia (Remsen 2003). There are few specimens and, for most ol these, the locality is given simply as “Bahia". It is almost certainly confined to the humid coastal plain and adjacent mountain ranges, but more work is needed to assess its northern and southern distributional limits and popu- 22 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 lation size, as well as to confirm its status as a distinct taxon relative to sulphurascens. It would be of particular interest to learn if the range of nominate leucophthalmus contacts that of lammi in northern Bahia or Sergipe. Automolus lammi should be assigned the conservation status of “threatened” (Remsen 2003). Its known range is restricted to coastal Paraiba, Pernambuco, Alagoas, and Sergipe, all of which have suffered catastrophic loss of forest for agricultural purposes, primarily sug- ar-cane production. Brown and Brown (1992) estimated that only 2% of the original forest cover remained in Alagoas and Pernambuco states, and only 6% in Paraiba. Lowland for- ests in this region have been largely replaced by vast sugar-cane plantations with most re- maining remnant forests restricted to steep slopes and ridge-tops (Stattersfield et al. 1998; KJZ, pers. obs.). Virtually all remaining for- ests in the region are true fragments, lacking even the narrowest of wooded corridors to other such fragments. Most of these exist on sugar-cane mill properties or usinas, rather than on formally protected lands (Silveira et al. 2003). This presents a particularly dire conservation picture for certain groups of for- est-dependent birds with poor dispersal capa- bilities, many of which have been demonstrat- ed to disappear from forest fragments below a minimum critical size, or in which anthro- pogenic activities such as selective logging are conducted. Automolus species have been shown to be among the most adversely af- fected birds by selected logging activities in French Guianan forests (Thiollay 1992) and by forest fragmentation in Amazonas, Brazil (Stouffer and Bierregaard 1995, Stouffer and Borges 2001). Automolus lammi is known only from a few formally protected areas, notably from the pri- vately owned Murici Biological Reserve and Pedra Talhada Federal Biological Reserve in Alagoas State. The former site has been iden- tified as one of the most important sites for conservation in the Neotropics (Wege and Long 1995, BirdLife International 2000); de- spite its protected status, illegal selective log- ging of larger trees was an ongoing problem as recently as 2004 (pers. obs.). Recent sur- veys of privately owned forest fragments in the region found A. lammi in only a few sites and concluded that it appeared to be “genu- inely rare” (Silveira et al. 2003:41). The same surveys found that lammi and other understo- ry insectivores were inexplicably absent from some sites with seemingly suitable habitat. The authors suggested that widespread recent use of aerially dispersed pesticides in the sur- rounding usinas could be seriously impacting populations of insectivorous birds in the forest fragments. Automolus lammi is only one of several dis- tinctive avian taxa endemic to northeast Brazil in general, and the Pernambuco Center in par- ticular, that have languished in relative obscu- rity due to their current taxonomic treatment as subspecies of broader-ranging species. Many of these forms, as noted by Silveira et al. (2003), were originally described as dis- tinct species and later “lumped” into geo- graphically widespread, polytypic species without analysis or even comment, as part of a generalized lumping trend in the mid 1900s. Recent fieldwork and preliminary investiga- tions, particularly of vocal characters, sug- gests that many of these taxa merit recognition as distinct species under any of the widely recognized species concepts. The populations in question continue to suffer for our adher- ence to plumage-based taxonomies that great- ly underestimate the true species-level biodi- versity of tropical ecosystems. Other workers (e.g., Willis 1988b, 1991; Silveira et al. 2003) have reported the lack of attention given to conservation of endangered subspecies rela- tive to endangered species, but the signifi- cance of this distinction cannot be overstated. Lack of understanding of species limits in widespread, polytypic groups, along with in- complete knowledge of distributions may re- sult in serious underestimation of regional biodiversity and endemism, as well as ob- scuring historical patterns of biogeography and evolution. These miscalculations could have serious repercussions when decisions re- garding allocation of scarce conservation dol- lars are considered. ACKNOWLEDGMENTS Special thanks are due M. L. and R R. Isler for lending their time and talents in producing the map and spectrograms (respectively) for this paper. Andrew Whittaker was an enthusiastic partner in much of the fieldwork upon which this study was based. K. L. Gar- rett, K. C. Molina, and the staff at LACM were of inestimable help in coordinating specimen loans from Zimmer • WHITE-EYED EOLIAGE-GLEANER SPECIES LIMITS 23 different institutions, as well as logistical support for the author’s frequent museum visits. T S. Schulenberg was available with helpful advice, and also provided access to several key references. I also thank J. M. Bates, S. J. Hacked, T. S. Schulenberg, and D. E. Wil- lard (FMNH); S. L. Olson, G. R. Graves and Phillip Angle (USNM); and R. S. Ridgely and D. J. Agro (ANSP) for arranging specimen loans from their re- spective institutions. G. E Budney, Andrea Priori, and the rest of the staff at the Library of Natural Sounds, Cornell Laboratory of Ornithology were of great help in providing access to the LNS collection of record- ings. Andrew Whittaker and C. A. Marantz generously provided additional tape recordings to add to my in- ventory. I thank C. E. Braun, D. E Stotz, and an anon- ymous reviewer for their many helpful comments on the manuscript. Special thanks go to Victor Emanuel Nature Tours, Inc. for providing me with many of the travel opportunities that made this research possible. LITERATURE CITED Bierregaard Jr., R. O., M. Cohn-Haft, and D. E Stotz. 1997. Cryptic biodiversity: an overlooked species and new subspecies of antbird (Formica- riidae) with a revision of Cercomacra tyrannina in northwestern South America. Ornithological Monographs 48: 1 1 1-128. BirdLife International. 2000. Threatened birds of the world. BirdLife International, Cambridge, United Kingdom and Lynx Edicions, Barcelona, Spain. Brown Jr., K. S. and G. G. Brown. 1992. Habitat alteration and species loss in Brazilian forests. Pages 119-142 in Tropical forest and extinction (T. C. Whitmore and J. A. Sayer, Editors). Chap- man and Hall, London, United Kingdom. Cory, C. B. 1919. Descriptions of three new South American birds. Auk 36:540-541. Cracraft, j. 1985. Historical biogeography and pat- terns of differentiation within the South American avifauna: areas of endemism. Ornithological Monographs 36:49-84. Haffer, j. 1985. Avian zoogeography of the neotrop- ical lowlands. Ornithological Monographs 36: 113-146. Hellmayr, C. E. 1925. Catalogue of birds of the Americas. Part IV. Field Museum Natural History and Zoology Series 13:1-390. ISLER, M. L. 1997. A sector-based ornithological geo- graphic information system for the Neotropics. Ornithological Monographs 48:345-354. ISLER, M. L., P. R. ISLER, AND R. T. Brumfield. 2005. Clinal variation in vocalizations of an antbird (Thamnophilidae) and implications for defining species limits. Auk 122:433-444. ISLER, M. L., P. R. ISLER, AND B. M. Whitney. 1997. Biogeography and sy.stematics of the Thamnophi- lus punctatiis (Thamnophilidae) complex. Orni- thological Monographs 48:355-381. ISLER, M. L., P. R. ISLER, AND B. M. WniTNEY. 1998. Use of vocalizations to establish species limits in antbirds (Passeriformes: Thamnophilidae). Auk 115:577-590. IsLER, M. L., J. Alvarez A., P. R. Isler, and B. M. Whitney. 2001. A new species of Percnostola antbird (Passeriformes: Thamnophilidae) from Amazonian Peru, and an analysis of species limits within Percnostola rufifrons. Wilson Bulletin 113: 164-176. Isler, M. L., J. Alvarez A., P. R. Isler, T. Valqui, A. Begazo, and B. M. Whitney. 2002. Rediscov- ery of a cryptic species and description of a new subspecies in the Myrmeciza hemimelaena com- plex (Thamnophilidae) of the Neotropics. Auk 119:362-378. Jackson, J. E 1978. Differentiation in the genera En- yalius and Strobilurus (Iguanidae): implications for Pleistocene climatic changes in eastern Brazil. Arquivos de Zoologia (Sao Paulo) 30:1-79. Johnson, N. K., J. V. Remsen Jr., and C. Cicero. 1999. Resolution of the debate over species con- cepts in ornithology: a new comprehensive bio- logic species concept. Proceedings of the Inter- national Ornithological Congress 22:1470-1482. Kinzey, W. G. 1982. Distribution of primates and for- est refuges. Pages 455—482 in Biological diversi- fication in the Tropics (G. T. Prance, Editor). Co- lumbia University Press, New York, USA. Krabbe, N. and T. S. Schulenberg. 1997. Species limits and natural history of Scytalopus tapaculos (Rhinocryptidae), with descriptions of the Ecua- dorian taxa, including three new species. Ornitho- logical Monographs 48:47-88. Kroodsma, D. E. 1984. Songs of the Alder Flycatcher (Empidonax alnorum) and Willow Flycatcher (Empidonax traillii) are innate. Auk 101:13—24. Kroodsma, D. E. 1986. Design of song playback ex- periments. Auk 103:640-642. Kroodsma, D. E. 1989. Male Eastern Phoebes {Say- ornis phoebe\ Tyrannidae, Passeriformes) fail to imitate songs. Journal of Comparative Psychology 103:227-232. Kroodsma, D. E. and M. Konishi. 1991. A suboscine bird (Eastern Phoebe, Sayornis phoebe) develops normal song without auditory feedback. Animal Behavior 42:477-487. Lanyon, W. E. 1978. Revision of the Myiarchiis fly- catchers of South America. Bulletin of the Amer- ican Museum of Natural History 161:429-627. Lichtenstein, M. H. K. 1823. Verzeichniss der Doub- letten des Zoologischen Museums der Kbnigl. Universitiit zu Berlin nebst Beschreibung vieler bisher unbekannter Arten von Siiugethieren, Vii- geln, Amphibien und Fischen // herausgegbcn von H. Lichtenstein. Zoologisches Museum: 41 \An- tomolus leucophthabnus sidphnrascens], x + 1 18 pp. Berlin, Germany. McKitrick, M. C. and R. M. Zink. 1988. Species con- cepts in ornithology. Condor 90:1-14. Mullf:r. P. 1973. Dispersal centers of terrestrial ver- tebrates in the neotropical realm. Biogeographica Number 2. 24 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 Peters, J. L. 1951. Checklist of birds of the world. Volume VII. Harvard University Press, Cam- bridge, Massachusetts, USA. PiERPONT, N. AND F. W. FITZPATRICK. 1983. Specific status and behavior of Cymbilaimus sanctaemar- iae, the Bamboo Antshrike, from southwestern Amazonia. Auk 100:645—652. Pyle, P, S. N. G. Howell, R. P Yunick, and D. F. DeSante. 1987. Identification guide to North American passerines. Slate Creek Press, Bolinas, California, USA. Remsen Jr., J. V. 2003. Family Furnariidae (oven- birds). Pages 152-357 in Handbook of birds of the world. Volume 8. Broadbills to tapaculos (J. del Hoyo, A. Elliott, and D. Christie, Editors). Lynx Edicions, Barcelona, Spain. Remsen Jr., J. V. 2005. Pattern, process, and rigor meet classification. Auk 122:403—413. SiLVEiRA, L. F, F. Olmos, and a. j. Long. 2003. Birds in Atlantic Forest fragments in north-east Brazil. Cotinga 20:32-46. Smithe, F. B. 1975. Naturalist’s color guide. American Museum of Natural History, New York, USA. Stattersfield, a. j., M. J. Crosby, A. J. Long, and D. C. Wege. 1998. Endemic bird areas of the world. Priorities for biodiversity conservation. BirdLife Conservation Series 7. BirdLife Inter- national, Cambridge, United Kingdom. Stouffer, P. C. and R. O. Bierregaard. 1995. Use of Amazonian forest fragments by understory insec- tivorous birds. Ecology 76:2429-2445. Stouffer, P. C. and S. H. Borges. 2001. Conservation recommendations for understory birds in Ama- zonian forest fragments and second-growth areas. Pages 248-261 in Lessons from Amazonia: the ecology and conservation of a fragmented forest (R. O. Bierregaard, C. Gascon, T. E. Lovejoy, and R. Mesquita, Editors). Yale University Press, New Haven, Connecticut, USA and London, United Kingdom. Thiollay, j. M. 1992. Influence of selective logging on bird species diversity in a Guianan rain forest. Conservation Biology 6:47-63. Vaurie, C. 1980. Taxonomy and geographical distri- bution of the Furnariidae (Aves, Passeriformes). Bulletin of the American Museum of Natural His- tory 166:1-357. Wege, D. C. and A. J. Long. 1995. Key areas for threatened birds in the Neotropics. BirdLife Inter- national Conservation Series 5. Cambridge, Unit- ed Kingdom. Wied-Neuwied, M., Prinz zu. 1821. Reise nach Bras- ilien in den Jahren 1815 bis 1817 von Maximilian, Prinz zu Wied-Neuwied — mit zwei und zwanzig Kupfern, neunzehn Vignetten und drei Karten. Volume 2:141 [Automolus leucophthalmus]. Frankfurt am Main, Heinrich Ludwig Bronner, Germany. Willis, E. O. 1988a. Behavioral notes, breeding rec- ords, and range extensions for Colombian birds. Revista de la Academia Colombiana de Ciencias 16:137-150. Willis, E. O. 1988b. Drymophila rubricollis (Bertoni, 1901) is a valid species (Aves, Formicariidae). Revista Brasileira de Biologia 48:431-438. Willis, E. O. 1991. Sibling species of greenlets (Vi- reonidae) in southern Brazil. Wilson Bulletin 103. 559-567. Willis, E. O. 1992a. Three Chamaeza ant-thrushes in eastern Brazil (Formicariidae). Condor 94:110- 116. Willis, E. O. 1992b. Zoogeographical origins of east- ern Brazilian birds. Ornitologia Neotropical 3:1- 15. Zimmer, J. T. 1947. New birds from Pernambuco, Bra- zil. Proceedings of the Biological Society of Washington 60:99-106. Zimmer, K. J. 1997. Species limits in Cranioleuca vul- pina. Ornithological Monographs 48:849-864. Zimmer, K. J. 2002. Species limits in Olive-backed Foliage-gleaners (Automolus: Furnariidae). Wil- son Bulletin 114:20-37. Zimmer, K. J. and M. L. Isler. 2003. Family Tham- nophilidae (typical antbirds). Pages 448-681 in Handbook of birds of the world. Volume 8. Broad- bills to tapaculos (J. del Hoyo, A. Elliott, and D. Christie, Editors). Lynx Edicions, Barcelona, Spain. Zimmer, K. J., A. Whittaker, and D. C. Oren. 2001. A cryptic new species of flycatcher (Tyrannidae: Suiriri) from the cerrado region of central South America. Auk 118:56-78. APPENDIX 1 Recording locations and recordists. — Num- bers following each name represent the num- ber of individual birds recorded by the re- cordist at each site. All recordings are from Brazil. Capitalized names refer to Brazilian states. Only recordings of the taxa in the Au- tomolus leucophthalmus complex and used in the vocal analysis are included. Recordings of other Automolus species were compared su- perficially. Automolus I lammi. ALAGOAS: Itabeguara (Zimmer 1); Murici Reserve (Zimmer 5, C. Marantz 3). SERGIPE; Crasto Reserve (Zimmer 9). Automolus /. sulphurascens. ESPIRITO SAN- TO: Caetes (Zimmer 2); Linhares CVRD Reserve (Zimmer 3); Augusto Ruschi Re- serve (Zimmer 4); Santa Teresa (Zimmer 4); Sooretama Biological Reserve (Zimmer 1). MINAS GERAIS: Serra do Cara9a Nat- ural Reserve (Zimmer 2). PARANA: Igua- 9u Falls National Park (Zimmer 33). SAO PAULO: Ubatuba (Zimmer 17). SANTA Zimmer • WHITE-EYED FOLIAGE-GLEANER SPECIES LIMITS 25 CATARINA: Volta Velha Reserve, Itapoa (Zimmer 2). Automolus L leucophthalmus . BAHIA: Boa Nova (A. Whittaker 1); Fazenda Palmeiras (Zimmer 12); Porto Seguro (Zimmer 2); Una Ecological Park (Zimmer 2). APPENDIX 2 List of localities and lending institutions for specimens examined. — All specimens were from the following institutions: Academy of Natural Sciences of Philadelphia, Philadel- phia, Pennsylvania (ANSP); Field Museum of Natural History, Chicago, Illinois (FMNH); Los Angeles County Museum, Los Angeles, California (LACM); National Museum of Nat- ural History, Washington, D.C. (USNM). Automolus 1. lammi (4 males, 2 females). BRAZIL: Engenho Riachao, Quebrangulo, Alagoas (LACM, I male); Usina Sinambii, Alagoas (LACM, 2 males, 2 females); Re- cife, Pernambuco (USNM, 1 male). Automolus /. sulphurascens (18 males, 11 fe- males, 3 unknown). ARGENTINA: Arroyo Urugua, Misiones (LACM, 1 male, 2 fe- males). BRAZIL: Linhares, Espirito Santo (USNM, 1 male); Serra Dourada, Goias (LACM, 1 male, 1 female, 1 unknown); Carrego Boa Esperanca, Mato Grosso (LACM, 1 male); Raul Scares, Minas Ger- ais (LACM, 3 males); Rio das Velhas, Mi- nas Gerais (FMNH, 1 male); Parque Na- cional de Itatiaia, Rio de Janeiro (LACM, 3 males, 1 unknown); Federal District, Rio de Janeiro (USNM, 1 female); Pedra Bran- ca, Parati, Rio de Janeiro (LACM, 1 male, 2 females); Guapi, Rio de Janeiro (LACM, 1 male); Boa de Terezopolis, Rio de Janeiro (LACM, 2 males); Brusque, Santa Catarina (LACM, 1 female, 1 unknown); Joinville, Santa Catarina (FMNH, 1 female); Barra do Leapara, Sao Paulo (FMNH, 1 female); Costao dos Engenhos, Sao Paulo (FMNH, 1 male, 1 female); Paranapiacaba, Sao Pau- lo (FMNH, 1 male); Fazenda Bela Vista, Taquar, Sao Paulo (LACM, 1 male); Can- tareira, Sao Paulo (LACM, 1 female). Automolus 1. leucophthalmus (4 unknown). BRAZIL: Bahia (USNM, 2 unknown; ANSP, 2 unknown); Espirito Santo (USNM, I male). The Wilson Journal of Ornithology 120( 1 );26-37, 2008 RECENT ADVANCES IN THE BEHAVIORAL ECOLOGY OE TROPICAL BIRDS The 2005 Margaret Morse Nice Lecture BRIDGET J. M. STUTCHBURY'-^ AND EUGENE S. MORTON''^ ABSTRACT— Tropical birds offer unique opportunities to test ecological and evolutionary theory because their life history traits are so diverse and different from temperate zone models upon which most empirical studies are based. We review recent studies on the behavioral ecology of tropical birds, studies that explore new advances in this held. Life histories and their evolution remain the focus of research on tropical birds. Cue size manipulations in two species showed that food limitation does not explain small clutch size. In antbuds enlarged clutches decreased post-fledging survival whereas in thrushes, enlarged broods were costly due to high nest predation. Small clutches may be favored via different ultimate selective forces and shared underlying tradeoffs between the immune, metabolic, and endocrine systems m the body may account for the commonly observed ‘slow pace of life’ in tropical birds. The physiological tradeoff between testosterone and immunocom- petence may explain the evolution of low testosterone levels in tropical passerines where adult survival is paramount. In contrast to life history theory, few studies have explored temperateqropical differences m tem- toriality mating systems, and song function. The idea that low breeding synchrony m tropical birds is associated with low levels of extra-pair fertilizations was supported by several new paternity studies conducted on tropical passerines. Seasonally breeding tropical birds have higher testosterone levels than tropical birds with prolonged breeding seasons, although it is unclear if this pattern is driven by mating systems per se or selection from pathogens. Recent work on relations between pair members in permanently paired tropical passerines focuses on the question of mate defense versus territorial defense and the extent of cooperation versus selfish interests in inter-sexual relations. Received 29 January 2007. Accepted 15 July 2007. Theory and empirical tests in the behavioral ecology of birds are generally constrained by a distributional problem. Whereas 80% of pas- serine birds reside in tropical latitudes, studies of temperate latitude species dominate our world view of avian biology. For example, does breeding occur when food for raising young is most abundant? Is the mating system of most passerine birds dominated by extra- pair behavior? Is song driven by sexual selec- tion because it functions to attract mates for males? Does testosterone underlie the main- tenance of aggression in territoriality? Well, this is what we think. But this is based on temperate passerines, popular test subjects for much of behavioral ecology theory. Studies of temperate birds outnumber those of tropical birds by a margin >100:1 (Stutchbury and Morton 2001). What if an equivalent research effort had been spent on studying the same phenomena, but with birds in the tropics instead of birds ' Department of Biology, York University, 4700 Keele Street, Toronto, ON, M3J 1P3, Canada. 2 Hemlock Hill Field Station, 22318 Teepleville Flats Road, Cambridge Springs, PA 16403, USA. 2 Corresponding author; e-mail: bstutch@yorku.ca in the temperate zone? Alexander Skutch (1985) provided a preview when he stated that the question of latitudinal clutch size variation should not be “why do tropical birds lay so few eggs?” but, rather “why do temperate zone birds lay so many?” Studies of tropical birds test existing theory, but they also lead to entirely new questions that would other- wise not be asked. Several years ago, we re- viewed research on the behavioral ecology of tropical birds to show how run-of-the-mill tropical birds have fundamentally different ad- aptations for life history strategies, mating systems, territory defense, and communica- tion (Stutchbury and Morton 2001). Here we highlight recent studies to illustrate how trop- ical birds can be used to test and develop new theories for understanding the evolution of bird behavior. Temperate zone birds converge in their ad- aptations to the overwhelmingly huge and predictable temperature and resource swings of temperate latitudes, and the vast majority of temperate breeding passerines have highly seasonal territorial defense and breeding. This creates a high level of male-male competition for territories, an intense period of social mate choice, and a sudden rush of nesting and egg- 26 Stutchbury and Morton • BEHAVIORAL ECOLOGY OF TROPICAL BIRDS 27 laying activity among females. Thus, many temperate species face a common suite of se- lection pressures on breeding adaptations that are imposed by climate and short breeding season. Most tropical birds are free from the constraints of the spring rush to establish ter- ritories and find mates; they typieally have ac- cess to mates, territories, and food year round. Many tropical birds, often both males and fe- males, keep singing whether breeding or not, whereas temperate zone birds stop singing when breeding wanes and females sing in only a few species (Morton 1996). One might ask, in the Skutchian vein, not “why do so many female birds sing in the tropies?” but, rather, “why do so few females sing in the temperate zone?” Temperate-tropical comparisons are funda- mental to understanding the different selective pressures acting on birds, and have been ap- plied most frequently to the question of life history evolution (Wikelski et al. 2003a, Tie- leman et al. 2005). The great diversity in life histories among tropical birds should also be viewed as a resource, one that can be used to document the full evolutionary potential of avian adaptation. Within one study area it is straightforward to find tropieal passerines that are highly seasonal breeders and others that are territorial and paired year-round and have a prolonged 6 to 8 month breeding season. Seasonal effects on food availability and breeding can occur even in the same forest, depending upon whether the species forages on the ground or in the bushes and vines above (Ahumada 2001). Similarly, in a single study site one finds species that breed in the dry season as opposed to the wet season, spe- cies in which males build nests and incubate eggs as opposed to only feeding young, and species in which adults rely primarily on fruit as opposed to insects. This diversity is the largely untapped resource offered by tropical birds, a powerful tool for testing behavioral ecology theory and developing new ideas that have a broad evolutionary scope (Stutchbury and Morton 2001). Our review focuses on re- cent advances in life history traits of tropical birds, their slow pace of life, their mating sys- tems, and territoriality. Life History Traits. — Tropical birds have a much smaller clutch size than their temperate zone counterparts, a fundamental latitudinal difference that has long been recognized (Skutch 1949), and still generates a great deal of theoretical and empirical study (Ghalambor and Martin 2001, Martin et al. 2001, Ricklefs and Wikelski 2002). There is little agreement, however, on whether small clutch sizes of tropical birds are favored because of food lim- itation, high predation risk, limited recruit- ment opportunities for juveniles, pathogens, or some combination of these selective factors (Ricklefs and Wikelski 2002). Clutch size ma- nipulations have been a standard tool for test- ing optimal clutch size theory in temperate species, but few such experiments have been done with tropical birds (Stutchbury and Mor- ton 2001). Styrsky et al. (2005) experimentally re- duced and enlarged clutches of Spotted Ant- birds {Hylophylax naevioides) to created clutches of one, two, or three eggs. They found that parents of enlarged broods could successfully feed additional young by increas- ing their provisioning rate and the young fledged at normal weight. Furthermore, in- creased parental activity at the nest did not result in increased nest predation. Instead, the eost of having a large brood occurred after the young fledged by reducing juvenile survival. Broods of three had only 30% post-fledging survival compared with 70% for broods of one chick. Even natural broods of two eggs had lower juvenile survival (45%) than re- duced clutches of one egg. Brood division af- ter fledging normally results in each chick from a brood of two being eared for by a sin- gle parent. In enlarged clutches a single parent had to care for two chicks, but one of these chicks usually died in the first few days ot the post-fledging period (Styrsky et al. 2005). This cost of a large clutch may be a general phenomenon because tropical passerines have a longer period of fledgling care and a longer delay before juvenile dispersal than temperate zone species (Russell et al. 2004). Ferretti et al. (2005) manipulated clutch size in two populations of the Rulous-bellied Thrush (Turdus rufiventris) in subtropical Ar- gentina to test whether lower food availability or high nest predation best explained inter- population variation in average clutch size (3.2 vs. 2.7 eggs). Clutch size in both popu- lations was experimentally standardized at three eggs, and manipulated so the eggs 28 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 hatched synchronously. If low food availabil- ity results in selection for small clutch sizes, one would predict the population with the smaller natural clutch size would experience low provisioning rates, slow nestling growth, and high starvation of nestlings. However, the opposite was true; the population with the larger initial clutch size apparently had less food available because nestlings experienced more starvation and slower growth rate. No starvation was observed in the population where clutches were enlarged, suggesting that food supply was not constraining clutch size. Instead, differences in nest predation (50 vs. 75%) appear to shape life history traits in this species because the population with the high- est predation had small clutch sizes, longer incubation bouts, faster nestling growth, and lower parental visitation rates to the nest. These two examples illustrate how the di- rect selective influence of environmental fac- tors on clutch size can differ among species. We can also add to the mix an earlier clutch manipulation study on Clay-colored Thrushes {Turdus grayi) in Panama in which parents could not raise additional young during the dry season when there were abundant fruit re- sources for adults but arthropod food for nest- lings was scarce (Stutchbury and Morton 2001). Clay-colored Thrushes bred at a poor time for feeding nestlings because predation was lower in the dry season (Morton 1971). A comparative study of tropical passerines in Africa found that seasonality of food re- sources had a strong influence on life history traits (Peach et al. 2001). Nectarivores and in- sectivores had high survival (72%) and low clutch size whereas granivores had low annual survival (54%) and large clutch sizes. The un- predictable and highly seasonal rainfall, which directly affects seed availability, may reduce the survival of granivores and select for larger clutch sizes. The diversity of tropical avian adaptations to food availability is exemplified by a recent study of the endangered Hawaiian Akepa {Loxops coccineus) (Freed et al. 2007). This specialized bird, feeding only on insects and spiders in the foliage of ohia trees {Me- trosideros polymorpha), illustrates “nestling overgrowth” where, at peak weight, nestlings weigh more than their parents. To attain this overgrowth, Akepa nestlings are fed more than they need thus banking this food for use during the 4-month fledgling period when food resources are declining. Akepa fledglings actually lose weight as they grow! Clay-col- ored Thrushes illustrate the opposite strategy. They fledge at much lower weights than the adult weight (e.g., at 35 vs. 80 g for adult weight) and continue to grow for long periods after fledging, even replacing remiges and rec- trices in a complete molt so these will match their increasing size. Fledglings of the largest species of robin, the Great Thrush {Turdus fuscater) (33 cm long), are similar in size to fledgling robins of normal size for the genus (20-25 cm) but continue growing for an un- known length of time after fledging (E. S. Morton, unpubl. data). Slow Pace of L//e.— Tropical birds exhibit a lower basal metabolic rate than temperate birds (Wikelski et al. 2003a). Ricklefs and Wi- kelski (2002) argue that life history variation falls on a slow-fast continuum in all birds and that basic physiological tradeoffs can explain why tropical birds generally converge on one end of this continuum, despite a wide array of environments and selective forces. Under their model, different environmental factors act on separate physiological systems (immune, met- abolic, and endocrine systems). Internal trade- offs between the physiological systems result in a uniform life history (low clutch size, long life span) and ‘slow pace of life’ among trop- ical birds. More resources may be allocated to adult survival than reproductive effort given a slow pace of life (Ricklefs and Wikelski 2002, Tie- leman et al. 2005). Tropical birds appear to have a slower pace of life because they lay smaller clutches, hatchlings grow more slow- ly, and adults live longer than temperate zone birds, all of which suggest a lower resting metabolic rate (RMR) in tropical birds (re- viewed in Wikelski et al. 2003a). The basic question is whether latitudinal relations be- tween RMR are genetically based or whether local climate affects metabolic rates through acclimation. Weathers (1997) showed that RMR may vary by habitat in tropical birds with birds in open, sunny areas having higher metabolic rates than birds in shady forests. Eurasian Stone Chats {Saxicola torquatus) showed population-specific RMRs with birds from Kenya, Austria, Ireland, and Kazakhstan differing in RMR even though all were housed Stutchbury and Morton • BEHAVIORAL ECOLOGY OF TROPICAL BIRDS 29 under identical conditions in captivity. The tropical birds from Kenya had lower RMR (Wikelski et al. 2003b). They suggested that it was not migration or year-round territorial- ity but rather adaptation to temperature fluc- tuations that influences RMR. For example, birds from northern areas can experience cold temperatures during breeding and a higher RMR would be favored. The important point of the Eurasian Stone Chat study is that ad- aptation to local habitat may be under genetic control (Wikelski et al. 2003b). A comparison of temperate-breeding and tropical House Wrens {Troglodytes aedon) found that Panamanian wrens had a lower field metabolic rate during reproduction, a smaller number of total provisioning trips to the nest (but not fewer trips per nestling), and fewer chicks per brood (Tieleman et al. 2006). Overall field metabolic rates during breeding were reduced by about 35% in tropical wrens, suggesting a slower pace of life, although some of this reduction may be the result of the much shorter day length in the tropics. In contrast to the temperate-breeding population, the tropical wrens did not show temporal var- iation in field metabolic rates while breeding which is likely due to their long breeding sea- son. The physiological tradeoff between testos- terone and immune function is the best studied interaction among physiological systems that impacts life history evolution. This is often called the immunocompetence handicap hy- pothesis (Folstad and Karter 1992, Hillgarth and Wingfield 1997) and there is considerable variation among studies, and taxa, in the strength of evidence supporting this tradeoff (Roberts et al. 2004). Testosterone is impor- tant in the breeding biology of temperate male passerines, as plasma testosterone titers in- crease enormously during the breeding season and influence important reproductive process- es such as song, courtship, sperm maturation, and territoriality (Balthazart 1983, Wingfield et al. 1990, Wingfield and Hahn 1994). Trop- ical birds, however, break this temperate zone rule because even during the peak of the breeding season males have low levels of tes- tosterone (Levin and Wingfield 1992, Hau 2001). Testosterone plasma titers in temperate zone passerines are in the range of 2.1 to 5.5 ng ml"', whereas breeding tropical passerines have values such as 0.2 ng ml"' in Spotted Antbirds (Hau et al. 2000) and 0.3 ng ml"' in White-bellied Antbirds {Myrmeciza longipes) (Fedy and Stutchbury 2006). In Spotted Ant- birds, brain sex steroid receptor expression is increased in the non-breeding season when testosterone levels are reduced (Hau 2007), a different pattern from temperate zone birds where circulating testosterone levels and sex steroid receptors are temporally linked (e.g., both highest during the breeding season). Tra- ditional ideas about the role of testosterone in mediating aggression seem not to apply to tropical species, especially those with year- round territoriality (Wiley and Goldizen 2003). The pace of life model proposes the great importance of immune function in tropical birds has a cascading effect; it reduces basal metabolic rate and reproductive effort to save energy and reduces testosterone to maintain immunocompetence, all of which result in a longer adult life span. The idea is that high testosterone can be tolerated in birds with short breeding seasons (e.g., temperate zone breeders) because the short time intervals al- leviate the fitness-lowering effects of testos- terone such as interference in parental care and possibly impaired immune function (Wingfield et al. 2001). Tropical birds that live at high altitudes may be able to sustain higher concentrations of testosterone than lowland birds because parasite loads are low in the highlands (Goymann et al. 2004). There is support for higher prevalence of hematozoan parasites in lowland than in highland tropical birds, but it may also be true that infections occur at a higher rate in temperate than in tropical birds, arguing against the tradeoff be- tween testosterone and hematozoan preva- lence as a causal factor underlying differences in the levels of testosterone (Ricklefs 1992, Durrant et al. 2006) Even with almost undetectable levels of tes- tosterone (Hau et al. 2000, Fedy and Stutch- bury 2006), tropical male passerines vigor- ously defend territories, sing, attack intruders, and pair with females (Wikelski et al. 1999). Female tropical birds can be fiercely territorial yet have low testosterone levels (Levin 1996). It is not fully understood whether female ag- gression in general is mediated by testosterone (Ketterson et al. 2005) and studies of female 30 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. I, March 2008 tropical birds would provide good subjects for resolving these questions. The general pattern for tropical passerines that defend territories throughout the year is low levels of testoster- one during most of the year (Levin and Wing- held 1992, Hau et al. 2000, Wikelski et al. 2003b, Day et al. 2006, Fedy and Stutchbury 2006). Tropical birds may not be able to main- tain high androgen levels due to their negative effects on immunocompetence and, instead, elevate testosterone only when needed; when territorial conflicts and instability arise (Wi- kelski et al. 1999). Testosterone did not in- crease in W^hite-bellied Antbirds after simu- lated territorial intrusions or during temporary removal experiments that created territorial conflicts among males (Fedy and Stutchbury 2006). If tropical birds have a slower pace of life, is it caused by a heightened immune function resulting from low levels of testosterone? Par- tial support for this causal relationship was found in House Sparrows {Passer domesti- cus). House Sparrows living in Panama laid smaller clutches over a long breeding season relative to a population in temperate New Jer- sey. Martin et al. (2006) found that in tropical House Sparrows, secondary antibody response to an administered novel antigen was faster and “energetic investment in immune activi- ty” was greater than in temperate House Spar- rows, as predicted by differences in pace of life. However, cell-mediated immune response was greater in the fast-living temperate birds. They concluded the relation of testosterone to a reduction in immune function is equivocal because different components of the immune system differ in their relative costs and ben- efits and only some differences may comple- ment variation in life histories (Martin et al. 2006). Furthermore, a laboratory study of male House Sparrows from New Jersey found that testosterone did not suppress immune function during the nonbreeding season, so seasonal variation in immune function could not be explained by seasonal changes in tes- tosterone levels (Greenman et al. 2005). Comparisons of the role of testosterone in temperate and tropical populations of the Eur- asian Stone Chat have been influential be- cause of the depth of study in both field and captive situations. The Stone Chat has a large breeding range that includes migratory north- ern populations with a short breeding season to tropical populations that are resident and have year-long territoriality. Begun by Dittami and Gwinner (1985), the levels of testosterone and gonadal cycles have been studied recently in both the laboratory (Rodl et al. 2004) and in field situations (Goymann et al. 2006). Di- rect measurements of extra-pair mating behav- ior have not been published but the length of the breeding season and changes in testoster- one suggest that extra-pair mating is common to all populations (Goymann et al. 2006). Stone Chats illustrate a reversal in breeding synchrony with temperate populations breed- ing over 5 months and tropical birds for a shorter period. Equatorial Stone Chats are sin- gle-brooded with individual pairs raising a clutch of three within a period of less than 3 months. Northern birds, in contrast, raise up to three clutches of five young each within a breeding season of 5—6 months (Helm et al. 2005, Schwabl et al. 2005). Equatorial Stone Chats, like other tropical birds, showed lower levels of testosterone (0.5 ng mL^or less) than their temperate counterparts, except during the stage when females were fertile (nest-building and laying), when testosterone levels were higher than those of European populations. This surge in testosterone was related to the relatively short time of female fertility (Goy- mann et al. 2006). Another latitudinal comparison of testoster- one has been accomplished within the genus Zonotrichia; northern populations of the White-crowned Sparrow (Z. leucophrys) and the tropical, but high altitude-living, Rufous- collared Sparrow (Z. capensis) (Moore et al. 2002, 2004a, 2004b; Wada et al. 2006). The tropical species of Zonotrichia can have levels of testosterone similar to northern latitude species (Moore et al. 2002). Moore et al. (2004a) report that testosterone in Rufous-col- lared Sparrows was higher during the main breeding period than the early breeding period (before incubation had begun), the opposite of the situation in the White-crowned Sparrow, whose testosterone levels are highest during territory establishment and decrease when they are feeding young (Morton et al. 1990). Rufous-collared Sparrows showed an elevated plasma testosterone when challenged by play- backs early in the breeding season. Oddly, tes- tosterone was elevated much more in response Stutchbury and Morton • BEHAVIORAL ECOLOGY OE TROPICAL BIRDS 31 to playbacks of a heterospecific song (a sym- patric seedeater) than by playback of the local conspecific dialect (4.2 vs. 2.0 ng ml^O^ Tes- tosterone-implanted Rufous-collared Spar- rows exhibited enhanced testosterone concen- trations but were no more aggressive than blank-implanted controls (Moore et al. 2004a). The opposite experiment, blocking testosterone pharmacologically, also did not affect territorial aggression in Rufous-collared Sparrows (Moore et al. 2004b). Moore et al. (2004a) suggested that high altitude Rufous- collared Sparrows likely have an extra-pair mating system and the sustained levels of tes- tosterone may reflect a genetically polygynous population where males would be favored to have elevated concentrations of testosterone throughout the breeding season (Moore et al. 2004a). We predicted that high testosterone is due to competition for extra-pair fertilizations (EPFs) in birds (Stutchbury and Morton 2001), regardless of latitude, and this view is supported by recent reviews (Hirschenhauser et al. 2003, Goymann et al. 2004, Garamszegi et al. 2005). We emphasized that extra-pair mating systems result in increased male-male competition, but more than simply male/male aggression and territoriality. Part of the en- hanced performance provided by testosterone likely includes influencing female choice of extra-pair mates. Studies of testosterone in birds to date have not been accompanied by study of the mating system and breeding syn- chrony of the target species, making it diffi- cult to evaluate the relative roles of immune function and mating system. Breeding Synchrony and Extra-pair Mat- ing.— Breeding synchrony among fertile fe- males is generally much higher for temperate breeding than tropical birds due to the dra- matic differences in the length of the breeding season (Stutchbury and Morton 1995). For Nearctic-neotropical migrants, breeding syn- chrony (% of females simultaneously fertile) typically ranges from 20 to 40% among spe- cies (Stutchbury et al. 2005a). The breeding synchrony for tropical passerines studied to date ranges from 8 to 30% with most species <15% (Table 1). Low values are typical for species that are territorial and paired year- round with long breeding seasons, and the highest values (25—30%) occurred in species TABLE 1. Erequency of extra-pair fertilizations and extent of breeding synchrony in socially monog- amous tropical passerines. Values give % of extra-pair young and broods that contained at least one extra-pair young (sample size in parentheses), and the breeding synchrony index (Kempenaers 1993). Species EPF frequency, % Breeding Young Broods synchrony (%) Cercomacra tyrannina^ 0 (15) 0 (12) 8 Tachycinita albilinea'^ 15 (98) 26 (30) 8 Elaenia flavogaster^ 4 (24) 8 (13) 9-10 Thryothorus leucotis'^ 4 (53) 3 (31) 10 Zosterops lateralis^ 0 (122) 0 12 Loxiodes bailleui^ 0 (20) 0 (12) low Geospiza scandens^ 8 (159) 15 (66) low Elaenia chiriquensis^ 37 (14) 67 (15) 15-18 Turdus grayi' 38 (37) 53 (19) 25 Volatinia jacarina^ 50 (20) 64 (7) 30 Fleischer et al. 1997; Moore et al. 1999; Stutchbury et al. 2007; d Gill et al. 2005; ^ Robertson et al. 2001; f Fleischer et al. 1994; § Petren et al. 1999; ‘ Stutchbury et al. 1998; J Carvalho et al. 2006. that defend seasonal territories and have a rel- atively short breeding season (Table 1). Many variables can affect extra-pair effort in breeding birds, but we have long argued that breeding synchrony is an important pre- dictor of extra-pair mating systems at the spe- cies level (Stutchbury and Morton 1995, Mor- ton et al. 1998, Stutchbury 1998). The fre- quency of EPFs is closely correlated with breeding synchrony among New World pas- serines (Stutchbury 1998, Stutchbury et al. 2005b) and is an important predictor of extra- pair mating systems in more general compar- ative analyses (Mpller and Ninni 1998). We predicted that extra-pair mating systems would not be the norm because most tropical passerines are paired and territorial year round, and have long asynchronous breeding seasons. Due to the scarcity of paternity stud- ies, our first evidence for low EPFs in tropical passerines came from their low testes size (Stutchbury and Morton 1995) and then from a handful of tropical passerines that were found to have few or no EPFs (Stutchbury and Morton 2001). Given that mating systems may so thor- oughly affect a species’ biology, it is discon- certing that only three studies on tropical pas- serines have been published in the last 5 years, adding only four species to the tropical list (Gill et al. 2005, Carvalho et al. 2006, Stutch- 32 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 bury et al. 2007). To be fair, this is partly due to the logistical challenges of obtaining a de- cent sample size when faced with small brood sizes and high nest predation. There are now 10 species of tropical passerines, all from the Neotropics, where paternity data are available (Table 1). Half have fewer than 5% extra-pair young, and can be considered genetically mo- nogamous. Tropical species with relatively high syn- chrony, like the Lesser Elaenia (Elaenia chi- riquensis). Clay-colored Thrush, and Blue- black Grassquit (Volatinia jacarina), have abundant EPFs (Stutchbury et al. 1998, 2007; Carvalho et al. 2006). Blue-black Grassquits are sexually dimorphic granivores that have high breeding synchrony due to rapid and syn- chronous maturation of grass seed crops in the rainy season. Males have a distinctive jump display from the grass accompanied by a com- plicated, but short, buzzy song. Females seem to prefer to mate with males having the high- est jumps and this highly synchronized breed- ing is coupled with an extra-pair mating sys- tem (Carvalho et al. 2006). Studies of closely related species that differ in breeding synchrony offer a more powerful test of the tie between synchrony and extra- pair behavior. One study compared EPF fre- quency in two Elaenia flycatchers that differ in breeding synchrony (Stutchbury et al. 2007). The Lesser Elaenia and the Yellow-bel- lied Elaenia {Elaenia flavogaster) breed dur- ing the dry season but the former is migratory and more synchronous whereas the latter is permanently paired and defends territories throughout the year. Both species are monog- amously paired, share the same habitat, and have a fruit-influenced territorial system (Stutchbury and Morton 2001), but the Lesser Elaenia had a high level of extra-pair fertil- izations (67% of broods, 37% of nestlings) while we detected only a single instance of extra-pair paternity in the Yellow-bellied Elaenia (8% of broods, 4% of nestlings). We hope the common occurrence of near-genetic monogamy in tropical passerines, and the in- triguing differences among species document- ed thus far, will spur others to incorporate pa- ternity analyses into their studies. Testosterone and Mating System. — We sug- gested that testosterone appeared to relate more to achieving success in extra-pair mating than as a normal and necessary component of breeding behavior, which explains why tem- perate species have such high levels of testos- terone compared with tropical passerines (Stutchbury and Morton 2001). Testosterone functions to maintain the exhausting aggres- sion and displays that are associated with strong sexual selection and mate choice com- mon to extra-pair mating systems (Raouf et al. 1997). Tropical species that do breed syn- chronously and have extra-pair matings are expected to have high testosterone levels, which is the case for the Clay-colored Thrush (1.8 ng mL*) (Wikelski et al. 2003a). Testos- terone is also important in lekking species, like the Golden-collared Manakin (Manacus vitellinus), where testosterone increases dis- play behavior (Day et al. 2006). Testosterone has recently been tied specifi- cally to extra-pair mating systems in temper- ate zone passerines (reviewed in Garamszegi et al. 2005). The model that best explains lev- els of testosterone suggests a primary rela- tionship between relative testis size and extra- pair fertilizations with a secondary conse- quence on levels of testosterone (Garamszegi et al. 2005). An extra-pair mating system causes higher levels of testosterone, presum- ably because testosterone enhances behavior involved in the relative success of extra-pair fertilizations. Goymann et al. (2004) con- firmed that average levels of testosterone were higher in northern temperate species than tropical species (2.8 ± 0.4 vs. 1.3 ± 0.2 ng mL^ P < 0.0002). They also showed that birds with short breeding seasons, including high altitude breeders, had high testosterone relative to those with long breeding seasons. Seasonally-breeding tropical birds with breed- ing season territoriality like temperate birds, had >2.0 ng mL* testosterone whereas birds with year-round territories had <1.0 ng mL^ They concluded that high testosterone can evolve in the tropics under conditions of short breeding seasons and high altitudes (Goymann et al. 2004). Short breeding seasons are as- sociated with higher breeding synchrony and more intense pressure to establish a territory and find a mate. Short breeding seasons also appear to be associated with extra-pair mating systems in tropical birds (Table 1) which clearly would generate intense male-male competition. Analysis of testosterone relative Stutchbury and Morton • BEHAVIORAL ECOLOGY OF TROPICAL BIRDS 33 to type of territoriality and mating system was hampered by low statistical power — we don’t know enough about tropical birds! Whether the mating systems of lowland breeders or the high cost of lowered immu- nocompetence explains the evolution of low testosterone requires studies that uncouple the two variables. Many species in the lowland are seasonal breeders and have extra-pair mat- ing systems, and presumably there are also high altitude breeders with relatively long breeding seasons. Comparisons of testosterone function in congeneric tropical birds that share the same habitat but differ in mating system, for instance the Elaenia example (Stutchbury et al. 2007) would be a powerful way to con- trol for exposure to parasites and explore the role of testosterone in territory defense and immune function. Territoriality and Song. — High adult sur- vival is expected to result in relatively low turnover of territories, which could constrain opportunities for territory and mate switching by adults, as well as territory and mate choice by juveniles (Morton et al. 2000). In Dusky Antbirds {Cercomacra tyrannina), experimen- tally created vacancies are filled rapidly (<24 hrs), usually by neighboring territory holders (Morton et al. 2000). Both males and females of several species will quickly divorce their mates to fill vacant territories, likely to obtain a higher quality territory. Other species in similar habitat and the same study area in cen- tral Panama show a pattern of voluntary mate fidelity. In White-bellied Antbirds, for in- stance, experimentally-created vacancies were filled slowly (2-3 days) or not at all, and di- vorce was uncommon (Fedy and Stutchbury 2004). Annual survival of territory holders in Buff-breasted Wrens {Thryothorus leucotis) is 68% for females and 76% for males. Natural vacancies occur regularly and are filled rap- idly (<24 hrs), but in this species divorce did not occur between pairs which already bred together (Gill and Stutchbury 2006). A wid- owed bird is conspicuous to neighbors be- cause it sings only it’s half of the duet and one can presume that neighbors are aware of vacancies but choose not to move. Once a pair breeds together they remain paired until one member dies, so “til death do us part’’ actu- ally applies. In contrast, divorce was relatively common in young birds that had been paired <5 months and had not yet bred together. The interesting twist is the propensity for divorce only applied to young birds that immigrated into the population. Juveniles which lived on their parent’s territory before acquiring a ter- ritory nearby (1-2 territories away) generally did not divorce once paired, even when there were opportunities to switch territories. This suggests that local recruits can successfully assess territory and mate quality prior to ac- quiring a territory, whereas outsiders fill a va- cancy first and then begin to assess other op- portunities on neighboring territories. We do not yet know why mate fidelity is so advan- tageous for experienced pairs. Male and female Dusky Antbirds exhibit sex-specific territory defense (Morton and Derrickson 1996). Male song playbacks elicit a stronger response from the male than the female (Bard et al. 2002), whereas female res- idents are more aggressive than their mates in response to female song playback. A recent study of another antbird, the Warbling Antbird {Hypocnemis cantator) went much further by showing that females sang more and more quickly, thus initiating more duets, when re- sponding to female solo playbacks than to duet or male solo playbacks (Seddon and To- bias 2006). This supports the idea that duets are in defense of the mate rather than a form of cooperation in joint defense of the territory. Male White-bellied Antbirds responded more aggressively than females regardless of whether the stimulus was male song, female song, or duetting (Fedy and Stutchbury 2005). Both genders were more aggressive to play- backs during the non-breeding (dry) season than the breeding (wet) season, perhaps be- cause the low arthropod abundance during the dry season increases the benefits ot exclusive use of the territory. Duetting did not appear to function in territory defense since pairs did not duet more during playbacks, even when duet songs were used as a stimulus. A typical White-bellied Antbird pair sang only several (1—3) duets/hr, <10 male songs/ hr, and <3 female songs/hr (Fedy and Stutch- bury 2005). There was no dawn chorus (e.g., elevated song rate) during the dry or wet sea- son. In contrast, the Chestnut-backed Antbird (Myntieciza e.wsnl) has a dawn chorus (Stutch- bury et al. 2()()5b). Possibly, the acoustic dif- 34 THE WILSON JOURNAL OF ORNITHOLOGY • VoL 120, No. I, March 2008 ferences between scrubby edge (White-bellied habitat) and shaded forest (Chestnut-backed habitat), in combination with habitat differ- ences in food availability, have influenced the tendency to sing at dawn (Morton 1977, Sed- don 2005). Antbirds do not learn songs and their duets are simple relative to the elaborate repertoires of duets found in song-learning oscine pas- serines (Seddon and Tobias 2006). Perhaps the best example of how complex ducting can be- come is found in the Plain-tailed Wren (Thry- othorus euophrys) where groups of birds sing in synchronized choruses, coordinating their repertoires of 20 phrases of sex-specific songs for up to 2 min (Mann et al. 2007)! Buff- breasted Wrens have a conspicuous dawn cho- rus following daybreak, when pairs typically sing over 50 duets in 30 min, often switching to different duet combinations (Gill et al. 2005). The role of duet singing in territorial de- fense has taken on new significance with the discovery of a “duet code ’ in Black-bellied Wrens (Thryothorus fasciatoventris) (Logue and Gammon 2004, Logue 2006). As in most ducting bird species, males and females differ in their responses to intra- and inter-sexual playbacks with greater responses to same-sex songs. Female Black-bellied Wrens respond more to female songs but males do not dif- ferentiate to the same extent. However, the song contributions by pair members to each duet are not random. For each pair there exists a specific pairing of songs which is somewhat reciprocal. For example, if a male sings song A, his mate duets with her song W, and if her song W is played back, her mate responds with his song A. These couplings between songs in each duet constitute the duet code in this species and serve to identify specific in- dividuals, useful information in defending ter- ritories. Rufous-and-white Wrens {Thryotho- rus rufalbus) sing mainly solo songs with males singing much more than females and males only singing in a dawn chorus. Play- back experiments using stereo speakers imi- tating an intruding pair showed that Rufous- and-white Wrens also have a duet code that functions in territorial defense (Mennill and Vehrencamp 2005, Mennill 2006). There are exciting new aspects of neotrop- ical avian biology that have developed in the last 5 years with the traditional focus on pas- serine birds (Stutchbury and Morton 2001). The use of the diversity of tropical birds to evaluate and test ideas about general avian ad- aptations is still a largely untapped resource. We hope this mini-review encourages others to step outside the temperate zone bias and into the field to study tropical birds. ACKNOWLEDGMENTS We thank the Wilson Ornithological Society for awarding us the 2005 Margaret Morse Nice Medal. It is a great honor. We dedicate this review to the natu- ralists that Margaret Morse Nice, Alexander Skutch, and Eugene Eisenmann, among many others, mentored and hope the tradition of “knowing your organism” will increase as a way to enhance our scientific un- derstanding of birds. LITERATURE CITED Ahumada, J. a. 2001. Comparison of the reproductive biology of two neotropical wrens in an unpredict- able environment in northeastern Colombia. Auk 118:191-210. Balthazart, J. 1983. Hormonal correlates of behavior. Pages 221-365 in Avian biology. VII (D. S. Ear- ner, J. R. King, and K. C. Parkes, Editors). Aca- demic Press, New York, USA. Bard, S. C., M. Hau, M. Wikelski, and J. C. Wing- field. 2002. Vocal distinctiveness and response to conspecific playback in the Spotted Antbird, a neotropical suboscine. Condor 104:387-394. Carvalho, C. B. V, R. H. Macedo, and J. A. Graves. 2006. Breeding strategies of a socially monoga- mous neotropical passerine: extra-pair fertiliza- tions, behavior, and morphology. Condor 108: 579-590. Day, L. B., J. T. McBroom, and B. A. Schlinger. 2006. Testosterone increases display behaviors but does not stimulate growth of adult plumage in male Golden-collared Manakins (Manacus vitel- linus). Hormones and Behavior 49:223-232. Dittami, j. P. and E. Gwinner. 1985. Annual cycles in the African Stonechat Saxicola torquata axil- laris and their relationship to environmental fac- tors. Journal of Zoology 207:357-370. Durrant, K. L., j. S. Beadell, E Ishiaq, G. R. Graves, S. L. Olson, E. Gering, M. A. Peirce, C. M. Milensky, B. K. Schmidt, C. Gebhard, AND R. C. Fleischer. 2006. Avian hematozoa in South America: a comparison of temperate and tropical zones. Ornithological Monographs 60:98- 111. Fedy, B. C. and B. j. M. Stutchbury. 2004. Territory switching and floating in a resident tropical pas- serine, the White-bellied Antbird. Auk 121:486- 496. Fedy, B. C. and B. J. M. Stutchbury. 2005. Territory defence in tropical birds: are females as aggres- Stutchbury and Morton • BEHAVIORAL ECOLOGY OF TROPICAL BIRDS 35 sive as males? Behavioural Ecology and Socio- biology 58:414-422. Fedy, B. C. and B. J. M. Stutchbury. 2006. Testos- terone does not increase in response to conspecific challenges in the White-bellied Antbird {Myrme- ciza longipes), a resident tropical passerine. Auk 123:61-66. Ferreti, V., P. E. Llambias, and T. E. Martin. 2005. Life-history variation of a neotropical thrush chal- lenges food limitation theory. Proceedings of the Royal Society of London B 272:769-773. Fleischer, R. C., C. Tarr, and T. K. Pratt. 1994. Genetic structure and mating system in the Palila, an endangered Hawaiian honeycreeper, as as- sessed by DNA fingerprinting. Molecular Ecology 3:383-392. Fleischer, R. C., C. L. Tarr, E. S. Morton, A. Sang- MEISTER, AND K. C. Derrickson. 1997. Mating system of the Dusky Antbird, a tropical passerine, as assessed by DNA fingerprinting. Condor 99: 512-514. Folstad, I. AND A. J. Karter. 1992. Parasites, bright males, and the immunocompetence handicap. American Naturalist 139:603-622. Freed, L., J. S. Fretz, and M. C. Medeiros. 2007. Adaptation in the Hawaiian Akepa to breed and moult during a seasonal food decline. Evolution- ary Ecology 9:157-167. Garamszegi, L. Z., M. Eens, S. Hurtrez-Bousses, AND A. P. M0LLER. 2005. Testosterone, testes size, and mating success in birds: a comparative study. Hormones and Behavior 47:389-409. Ghalambor, C. K. and T. E. Martin. 2001. Fecun- dity-survival trade-offs and parental risk-taking in birds. Science 292:494-497. Gill, S. A. and B. J. M. Stutchbury. 2006. Long- term mate and territory fidelity in neotropical Buff-breasted Wrens (Thryothorus leucotis). Be- havioral Ecology and Sociobiology 61:245-253. Gill, S. A., M. J. Vonhof, B. J. M. Stutchbury, E. S. Morton, and J. S. Quinn. 2005. No evidence for acoustic mate-guarding in duetting Buff- breasted Wrens (Thryothorus leucotis). Behavioral Ecology and Sociobiology 57:557-565. Goymann, W., 1. T. Moore, A. Scheuerlein, K. Hir- SCHENHAUSER, A. GrAFEN, AND J. C. WiNGFIELD. 2004. Testosterone in tropical birds: effects of en- vironmental and social factors. American Natural- ist 164:327-334. Goymann, W., D. Geue, 1. Schwabl, H. Funks, D. ScHMiDL, H. Schwabl, and E. Gwinner. 2006. Testosterone and corticosterone during the breed- ing cycle of equatorial and European Stonechats (Saxicola torquata axillaris and S. t. ruhicola). Hormones and Behavior 50:779-785. Greenman, C. G., L. B. Martin II, and M. Had. 2005. Reproductive state, but not testosterone, reduces immune function in male House Sparrows (Passer domesticus). Physiological and Biochemical Zo- ology 78:60-68. Hau, M. 2001. Timing of breeding in variable envi- ronments: tropical birds as model systems. Hor- mones and Behavior 40:281-290. Hau, M. 2007. Regulation of male traits by testoster- one: implications for the evolution of vertebrate life histories. BioEssays 29:133-144. Hau, M., M. Wikelski, K. K. Soma, and J. C. Wing- field. 2000. Testosterone and year-round territo- rial aggression in a tropical bird. General and Comparative Endocrinology 117:20-33. Helm, B., E. Gwinner, and L. Trost. 2005. Flexible seasonal timing and migratory - behavior results from Stonechat breeding programs. Pages 2 lb- 227 in Bird hormones and bird migrations: ana- lyzing hormones in droppings and egg yolks and assessing adaptations in long-distance migration. (W. Goymann, S. Jenni-Eirmann, and U. Bauchin- ger. Editors). Annals of the New York Academy of Sciences, Blackwell Publishing, New York, USA. Hillgarth, N. and j. C. Wingfield. 1997. Parasite- mediated sexual selection: endocrine aspects. Pag- es 78-104 in Host-parasite evolution: general principles and avian models (D. H. Clayton and J. Moore, Editors). Oxford University Press, Oxford, United Kingdom. Hirschenhauser, K., H. Winkler, and R. F. Oliveira. 2003. Comparative analysis of male androgen re- sponsiveness to social environment in birds: the effects of mating system and paternal incubation. Hormones and Behavior 43:508-519. Kempenaers, B. 1993. The use of a breeding synchro- ny index. Ornis Scandinavica 24:84. Ketterson, E. D., V. Nolan Jr., and M. Sandell. 2005. Testosterone in females: mediator of adap- tive traits, constraint on sexual dimorphism, or both? American Naturalist 166:885-598. Levin, R. N. 1996. Song behaviour and reproductive strategies in a duetting wren, Thryothorus nigri- capillus. 1 . Removal experiments. Animal Behav- iour 52:1093-1106. Levin, R. N. and J. C. Wingfield. 1992. The hormonal control of territorial aggression in tropical birds. Ornis Scandinavica 23:284-291. Logue, D. M. 2006. The duet code of the female Black-bellied Wren. Condor 108:326-335. Logue, D. M. and D. E. Gammon. 2004. Duet song and sex roles during territory defence in a tropical bird, the Black-bellied Wren, Thryothorus fascia- toventris. Animal Behaviour 68:721-731. Mann, N. I., K. A. Dingess, and P. J. B. Slater. 2006. Antiphonal four-part synchronized chorusing in a neotropical wren. Biology Letters 2:1-4. Martin, T. E., A. P Mpller, S. Merino, and J. Ct.oB- ERT. 2001. Does clutch size evolve in response to parasites and immunocompetence? Proceedings ol the National Academy of Sciences 98:2071-2076. Martin II, L. B., D. Hasseloui.st, and M. Wikf:lski. 2006. Investment in immune defense is linked to pace of life in House Sparrows. Oecologia 147: 565-575. Mf:nnii.l, D. .1. 2006. Aggressive responses ot male 36 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 and female Rufous-and-white Wrens to stereo duet playback. Animal Behaviour 71:219-226. Mennill, D. J. and S. L. Vehrencamp. 2005. Sex dif- ferences in singing and duetting behavior of neo- tropical Rufous-and-white Wrens {Thryothorus ru- falbus). Auk 122:175—186. M0LLER, A. P. AND P. Ninni. 1998. Sperm competition and sexual selection: a meta-analysis of paternity studies in birds. Behavioral Ecology and Socio- biology 43:345-358. Moore, I. T, B. G. Walker, and J. C. Wingfield. 2004b. The effects of combined aromatase inhib- itor and anti-androgen on male territorial aggres- sion in a tropical population of Rufous-collared Sparrows, Zonotrichia capensis. General and Comparative Endocrinology 135:223-229. Moore, 1. T, N. Perfito, H. Wada, T. S. Sperry, and J. C. Wingfield. 2002. Latitudinal variation in plasma testosterone levels in birds of the genus Zonotrichia. General and Comparative Endocri- nology 129:13-19. Moore, I. T, H. Wada, N. Perfito, D. S. Buch, T. P. Hahn, and J. C. Wingfield. 2004a. Territoriality and testosterone in an equatorial population of Ru- fous-collared Sparrows, Zonotrichia capensis. An- imal Behaviour 67:41 1-420. Moore, O. R., B. J. M. Stutchbury, and J. S. Quinn. 1999. Extrapair mating system of an asynchro- nously breeding tropical songbird: the Mangrove Swallow. Auk 116:1039-1046. Morton, E. S. 1971. Nest predation affecting the breeding season of the Clay-colored Robin, a trop- ical song bird. Science 181:920—921. Morton, E. S. 1977. On the occurrence and signifi- cance of motivational-structural rules in some bird and mammal sounds. American Naturalist 111: 855-869. Morton, E. S. 1996. A comparison of vocal behavior among tropical and temperate passerine birds. Pages 258-268 in Ecology and evolution of acoustic communication in birds (D. E. Kroodsma and T. E. Miller, Editors). Cornell University Press, Ithaca, New York. Morton, E. S. and K. C. Derrickson. 1996. Song ranging by the Dusky Antbird, Cercomacra tyr- annina: song ranging without learning. Behavioral Ecology and Sociobiology 39:195—201. Morton, E. S., B. J. M. Stutchbury, and K. C. Der- rickson. 2000. Territory switching behavior in a sedentary tropical passerine, the Dusky Antbird Cercomacra tyrannina. Behavioral Ecology 11: 648-653. Morton, E. S., B. J. M. Stutchbury, and W. Piper. 1998. Genetic monogamy in Blue-headed Vireos, and a comparison with a sympatric vireo with ex- trapair paternity. Behavioral Ecology 9:515-524. Morton, M. L., L. E. Peterson, D. M. Burns, and N. Allan. 1990. Seasonal and age-related changes in plasma testosterone levels in mountain White- crowned Sparrows. Condor 92:166-173. Peach, W. J., D. B. Hanmer, and T. B. Oatley. 2001. Do southern African songbirds live longer than their European counterparts? Oikos 93:235-249. Petren, K., B. R. Grant, and P. R. Grant. 1999. Low extrapair paternity in the Cactus Finch (Geospiza scandens). Auk 116:252—256. Ricklefs, R. E. 1992. Embryonic development period and the prevalence of avian blood parasites. Pro- ceedings of the National Academy of Sciences of the United States of America 89:4722-4725. Ricklefs, R. E. and M. Wikelski. 2002. The physi- ology/life-history nexus. Trends in Ecology and Evolution 17:462-468. Raouf, S. a., P. G. Parker, E. D. Ketterson, V. No- lan Jr., and C. Ziegenfus. 1997. Testosterone in- fluences reproductive success by increasing extra- pair fertilizations in male Dark-eyed Juncos, J un- co hyemalis. Proceedings of the Royal Society of London, Series B, 264:1599—1603. Roberts, M. L., K. L. Buchanan, and M. R. Evans. 2004. Testing the immunocompetence handicap hypothesis: a review of the evidence. Animal Beahviour 68:227-239. Robertson, B. C., S. M. Degnan, J. Kikkawa, and C. C. Moritz. 2001. Genetic monogamy in the ab- sence of paternity guards: the Capricorn Silvereye, Zosterops lateralis chlorocephalus, on Heron Is- land. Behavioral Ecology 12:666-673. RODL, T, W. GOYMANN, I. SCHWABL, AND E. GWINNER. 2004. Excremental androgen metabolite concen- trations and gonad sizes in temperate zone vs. tropical Stonechats (Saxicola torquata ssp.). Gen- eral and Comparative Endocrinology 139: 124- BO. Russell, E. M., Y. Yom-Tov, and E. Geffen. 2004. Extended parental care and delayed dispersal: northern, tropical, and southern passerines com- pared. Behavioral Ecology 15:831-838. ScHWABL, H., H. Flinks, and E. Gwinner. 2005. Tes- tosterone, reproductive stage, and territorial be- havior of male and female European Stonechats Saxicola torquata. Hormones and Behavior 47: 503-512. Seddon, N. 2005. Ecological adaptation and species recognition drives vocal evolution in noetropical suboscine birds. Evolution 59:200—215. Seddon, N. and J. A. Tobias. 2006. Duets defend mates in a suboscine passerine, the Warbling Ant- bird (Hypocnemis cantator). Behavioral Ecology 17:73-83. Skutch, a. E 1949. Do tropical birds rear as many young as they can nourish? Ibis 91:430-455. Skutch, A. E 1985. Clutch size, nesting success, and predation on nests of neotropical birds, reviewed. Pages 575-594 in Neotropical ornithology (P. A. Buckley, M. S. Foster, E. S. Morton, R. S. Rid- gely, and E G. Buckley, Editors). Volume 36. American Ornithologists’ Union, Washington, D.C., USA. Stutchbury, B. J. M. 1998. Female mate choice of extra-pair males: breeding synchrony is important. Stutchbury and Morton • BEHAVIORAL ECOLOGY OF TROPICAL BIRDS 37 Behavioural Ecology and Sociobiology 43:213- 215. Stutchbury, B. J. and E. S. Morton. 1995. The effect of breeding synchrony on extra-pair mating sys- tems in songbirds. Behaviour 132:675-690. Stutchbury, B. J. M. and E. S. Morton. 2001. Be- havioral ecology of tropical birds. Academic Press, London, United Kingdom. Stutchbury, B. J. M., E. S. Morton, and W. H. Piper. 1998. Extra-pair mating system of a synchronous- ly breeding tropical songbird. Journal of Avian Biology 29:72-78. Stutchbury, B. J. M., E. S. Morton, and T. E. Pitch- er. 2005a. Sex roles in migrants: extra-pair mating systems and winter social systems. Pages 307-320 in Birds of two worlds: the ecology and evolution of migratory birds (R. Greenberg and P. Marra, Editors). Johns Hopkins University Press, Balti- more, Maryland, USA. Stutchbury, B. J. M., E. S. Morton, and B. Wool- FENDEN. 2007. Comparison of the mating systems and breeding behavior of a resident and a migra- tory tropical flycatcher. Journal of Field Ornithol- ogy 78:40-49. Stutchbury, B. J. M., B. E. Woolfenden, B. C. Fedy, AND E. S. Morton. 2005b. Nonbreeding territorial behaviour of two congeneric antbirds, Myrmeciza exsul and M. longipes. Ornithologia Neoptropical 16:397-404. Styrsky, j. N., j. D. Brawn, and S. K. Robinson. 2005. Juvenile mortality increases with clutch size in a neotropical bird. Ecology 86:3238-3244. Tieleman, B. I., J. B. Williams, R. E. Ricklefs, and K. C. Klasing. 2005. Constitutive innate immu- nity is a component of the pace-of-life syndrome in tropical birds. Proceedings of the Royal Society of London Series B-Biological Sciences 272: 1715-1720. Tieleman, B. I., T. H. Dijkstra, J. R. Lasky, R. A. Mauck, G. H. Visser, and J. B. Williams. 2006. Physiological and behavioural correlates of life- history variation: a comparison between tropical and temperate zone House Wrens. Functional Ecology 20:491-499. Wada, H., I. T. Moore, C. W. Breuner, and J. C. Wingfield. 2006. Stress responses in tropical sparrows: comparing tropical and temperate Zon- otrichia. Physiological and Biochemical Zoology 79:784-792. Weathers, W. W. 1997. Energetics and thermoregu- lation by small passerines of the humid, lowland tropics. Auk 114:341-353. WiKELSKi, M., M. Hau, and j. C. Wingfield. 1999. Social instability increases plasma testosterone in a year-round territorial neotropical bird. Proceed- ings of the Royal Society of London Series B, Biological Sciences 266:551-556. WiKELSKi, M., M. Hau, W. D. Robinson, and J. C. Wingfield. 2003a. Reproductive seasonality of seven neotropical passerine species. Condor 105: 683-695. WIKELSKI, M., L. Spinney, W. Schelsky, A. Scheuer- LEiN, AND E. Gwinner. 2003b. Slow pace of life in tropical sedentary birds: a common-garden ex- periment on four Stonechat populations from dif- ferent latitudes. Proceedings of the Royal Society of London Series B, Biological Sciences 270: 2383-2388. Wiley, C. J. and A. W. Goldizen. 2003. Testosterone is correlated with courtship but not aggression in the tropical Buff-banded Rail, Gallirallus philip- pensis. Hormones and Behavior 43:554-560. Wingfield, J. C. and T. P. Hahn. 1994. Testosterone and territorial behaviour in sedentary and migra- tory sparrows. Animal Behaviour 47:77-89. Wingfield, J. C., S. E. Lynn, and K. K. Soma. 2001. Avoiding the ‘costs’ of testosterone: ecological bases of hormone-behavior interactions. Brain Be- havior and Evolution 57:239-251. WiNGnELD, J. C., R. E. Hegner, j. a. Dufty, and G. E Ball. 1990. The “Challenge hypothesis”: the- oretical implications for pattern of testosterone se- cretion, mating systems, and breeding strategies. American Naturalist 136:829-846. The Wilson Journal of Ornithology 120(1):38^9, 2008 DISTRIBUTION, BEHAVIOR, AND CONSERVATION STATUS OF THE RUFOUS TWISTWING (CNIPODECTES SUPERRUEUS) JOSEPH A. TOBIAS, DANIEL J. LEBBIN,^ ALEXANDRE ALEIXO,^ MICHAEL J. ANDERSEN,-* EDSON GUILHERME,^ PETER A. HOSNER,® AND NATHALIE SEDDON' ABSTRACT The Rufous Twistwing (Cnipodectes superrufus), a newly described Amazonian tyrant-fly- catcher, is known from five specimens and five localities in Cuzco and western Madre de Dios departments, Peru. We report three additional specimens and eight new localities extending the known range of the species east across Dpto Madre de Dios, Peru, into Dpto. Pando, Bolivia, and Acre State, Brazil. The new localities increase the distribution from -3,400 to -89,000 km^. We collected biometric data from five individuals, made behavioral observations in the field, and recorded three separate types of vocalizations, two of which (including the song) were previously unknown. We provide quantitative description of these vocalizations, consider their function, and compare them with vocalizations of the only known congener, the Brownish Twistwmg {Cnipo- dectes subbrunneus). Unique vocal repertoires support the classification of these two forms as sister species. The Rufous Twistwing resembles the Brownish Twistwing in producing loud vocalizations from regular song posts and both species appear to have a polygamous mating system. We provide further evidence consistent with the hypothesis the Rufous Twistwing is a Guadua bamboo specialist and recommend that it be listed as Vulnerable on the lUCN Red List. Received 5 July 2006. Accepted 23 March 2007. The Rufous Twistwing {Cnipodectes super- rufus) is a large, distinctive tyrant-flycatcher from southwest Amazonia. It occurs in a re- gion relatively well studied by ornithologists, but was only recently described and little is known about its behavior, distribution, and population size (Lane et al. 2007). The only previous publication considered the Rufous Twistwing likely to be a Guadua bamboo specialist, but noted that it was al- most certainly heard in “river-edge thicket vegetation with no bamboo growth visible” at Manu Lodge (12° 06' S, 71° 06' W; elevation 340 m) in 1991 and 2000 (Lane et al. 2007: 768). The five confirmed localities (Lane et al. ' Edward Grey Institute, Department of Zoology, University of Oxford, South Parks Road, Oxford, 0X1 3PS, United Kingdom. 2 Department of Ecology and Evolutionary Biology, Cornell University, El 48 Corson Hall, Ithaca, NY 14853, USA. ^ Museu Paraense Emflio Goeldi, Coordena^ao de Zoologia, Caixa Postal 399, 66040-170, Belem PA, Brazil. "^Macaulay Library, 159 Sapsucker Woods Road, Ithaca, NY 14850, USA. 5 Universidade Federal do Acre, Departmento de Ciencias da Natureza, BR-364, km 04, Campus, 69915-900, Rio Branco AC, Brazil. 6 Cornell Museum of Vertebrates, 159 Sapsucker Woods Road, Ithaca, NY 14850, USA. Corresponding author; e-mail: joseph.tobias@zoo.ox.ac.uk 2007) encompass a small area (~3,400 km^) in western Dpto. Madre de Dios and Dpto. Cuzco, Peru. However, the existence of large blocks of Guadua bamboo-dominated terra firme forest in southwestern Amazonian Bra- zil, southeastern Peru, and northwestern Bo- livia (Silman et al. 2003) suggest the global distribution of the species may be extensive (Lane et al. 2007). The only published vocalization is an agi- tation call (Lane et al. 2007), whereas the full vocal repertoire is likely to be similar to that of the Brownish Twistwing (C. subbrunneus), a presumed sister species with at least three distinct vocalizations (Parker and Remsen 1987, Schulenberg et al. 2000, Ridgely and Greenfield 2001). Male Brownish Twistwings are notably disinterested in playbacks of their songs, which are given from small, stable, and often adjacent territories, suggesting a polyg- ynous mating system (Parker and Remsen 1987, Ridgely and Tudor 1994, Ridgely and Greenfield 2001, Fitzpatrick 2004). Lane et al. (2007) reported no information regarding the breeding behavior of Rufous Twistwing. In this paper we report; (1) new localities, (2) vocalizations, (3) habitat preferences, and (4) behavior of the Rufous Twistwing to add to the details presented by Lane et al. (2007). We also assess the conservation status of the species, based on our results, following lUCN Red List criteria (lUCN 2001). 38 Tobias et al. • RUFOUS TWISTWING BEHAVIOR AND DISTRIBUTION 39 boundaries. Crosses indicate the previously known range of Rufous Twistwing in Peru: (1) Rio Urubamba centered on Camisea, and (2) Rio Manu centered on Pakitza, Dpto. Madre de Dios (type locality). Stars indicate new localities for the species: (3) Oceania, Dpto. Madre de Dios, Peru; (4) Extrema, Dpto. Pando, Bolivia, (5) Esta^ao Ecologica do Rio Acre, Acre State, Brazil; (6) Ramal Jarinal, Acre State, Brazil; (7) Rio Branco, Acre State, Brazil; (8) Timpia, Dpto. Cuzco, Peru; (9) Montetoni, Dpto. Cuzco, Peru; and (10) Tambopata, Dpto. Madre de Dios, Peru. Open circles indicate study sites where C. superrufus was absent: (11) Camino Mucden, Dpto. Pando, Bolivia; (12) Palmera, Dpto. Pando, Bolivia; (13) San Sebastian, Dpto. Pando, Bolivia; and (14) CICRA, Dpto. Madre de Dios, Peru. METHODS Study Sites and Survey Methods. — We sur- veyed six Gwc/c/wa-dominated forest sites in southwestern Amazonia, including one site in Bolivia and three in Brazil. Avifaunal surveys were undertaken by three independent teams at held sites using different methods at each site except CICRA and Oceania where the same methods were used. Playback of pre-re- corded Rufous Twistwing calls was used op- portunistically at sites 1-3; in all cases we used the agitated calls recorded by Thomas Valqui H. at Kirigueti, Dpto. Cuzco, (i.e.. Fig. 3A in Lane et al. 2007). Additional records were sought from experienced observers. Lo- calities were mapped (Fig. 1) and range sizes measured using equal-area projections with ArcView 3.2. 40 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 1. Centro de Investigacion y Capacitacion Rio Los Amigos (CICRA), 270 m, Dpto. Ma- dre de Dios, Peru (12°35'S, 69° 12' W), 2001-2007. This research station is at the con- fluence of the Rio Los Amigos and the Rio Madre de Dios in the Los Amigos Conser- vation Concession, a 135,000-ha private con- cession managed by Amazon Conservation Association, DJL, MJA, PAH, JAT, and NS surveyed multiple patches of Guadua weber- baueri and G. sarcocarpa bamboo for multi- ple bird species over a 7-year period using random transects, field observations, tape-re- cording, and intensive mist netting. This site includes camp CMl on the south bank of the Rio Los Amigos. 2. Oceania, 250 m, Dpto. Madre de Dios, Peru (11°23'S, 69° 32' W), 11-18 October 2004. This locality is 6 km west of Iberia, a town -130 km north of Puerto Maldonado on the Inapari road. DJL, MJA, and PAH sur- veyed a large patch of Guadua weberbaueri and G. sarcocarpa bamboo, in places collaps- ing to form impenetrable thickets with a can- opy at ~4 m. The bamboo patch was on an upland terrace, bordered by land recently cleared and burned for agriculture, extending to the nearby Rio Tahuamanu (to the east and south), as well as successional forest and ma- ture forest (to the north, south, and west). The presence of burned and chained logs, and a logging road revealed a recent history of hu- man disturbance. We conducted daily census- es using random transects between 0445 and 1000 hrs EST (12-16 Oct), and operated mist nets daily from dawn until midday (11—17 Oct). A combination of 6- and 12-m nets were rotated over 30 net locations covering a total length of 700 m of trail in 7 days. Total effort was —400 net hrs (for 12-m nets). 3. Extrema, 250 m, Dpto. Pando, Bolivia (11° 28' S, 69° 15' W), 6-9 November 2004. This military outpost is -45 km east-south- east of the Oceania site on the banks of the Rio Tahuamanu where this river flows from Peru into Bolivia. The natural habitat appears to be forest with an open canopy and an un- derstory largely composed of Guadua bam- boo. JAT and NS surveyed habitat here and at the adjacent guard-post of Alto Peru, 1 km north in Dpto. Madre de Dios, Peru. We sur- veyed habitat with and without bamboo on random transects between 0445 and 1800 hrs daily (6-9 Nov), spending 18 hrs in bamboo at Extrema and 2 hrs in bamboo at Alto Peru, Madre de Dios, Peru (2 km from Extrema). 4. Esta9ao Ecologica do Rio Acre, 250- 350 m. Acre State, Brazil (11° 03' S, 70° 12' W) on the border with Peru, 13-24 August 2005 and 3-15 February 2006. Site A was surveyed in 2005 by AA, including riparian forest along Rio Acre and a small tributary (Igarape do Tombo); A A and EG surveyed site B, —7 km west of site A, in 2006. Exten- sive patches of Guadua bamboo were present at both sites. Field effort included —250 hrs of random transects through the main vege- tation types, observing, mist netting, tape-re- cording, and collecting birds. 5. Rio Branco, 250 m. Acre State, Brazil (09° 57' S, 67° 57' W), 1998-2006. Long-term fieldwork is being conducted by EG at Parque Zoobotanico, a field station managed by Univ- ersidade Federal do Acre. The site is within the suburban limits of Rio Branco and the habitat (-100 ha of humid forest with patches of Guadua sarcocarpa and G. weberbaueri bamboo) is becoming increasingly disturbed and isolated by peripheral clearance (Guilher- me 2001). The main survey technique was mist netting: 10 mist nets (12 m) were oper- ated for 6 hrs/day (usually dawn until midday) on -35 days/year (evenly split between dry and wet seasons). Annual mist netting effort was —2,000 net hrs. 6. Ramal Jarinal, 320 m, -100 km WNW of Rio Branco, Acre State, Brazil (09° 54' S, 68° 28.5' W), 12-22 November 2006. EG and Marcos Persio D. Santos (Universidade Fed- eral do PiauO surveyed this site and collected specimens. The habitat was terra firme forest dominated by bamboo {Guadua sarcocarpa and G. weberbaueri) and palms (e.g., Attalea sp.) with high human disturbance including recently cleared cattle pastures and selective logging. Sound Recording, Analysis, and Playback Experiments. — Recordings of vocalizations were made at Oceania and Extrema using a Sennheiser ME 67 shotgun microphone with a Sony PCM-Ml MiniDAT or a Sony TC-D5 Pro II Cassette Recorder. All recordings are deposited at the Macaulay Library (ML) at Cornell University. Recordings were automatically filtered and digitized at a sampling frequency of 44. 1 kHz Tobias et al. • RUFOUS TWISTWING BEHAVIOR AND DISTRIBUTION 41 using Avisoft SASLabPro Version 4.15 (© 2002 Raimund Specht, Berlin, Germany) with a 16-bit acquisition sound card (0 VIA [Wave] 5.10). Sonograms were generated using stan- dard broadband (216 Hz) filter settings (FFT = 1,024, Frame = 75%, Window - FlatTop, Overlap = 87.5%). Sonograms were inspected visually, assigned to three types, and de- scribed quantitatively using on-screen cursors to measure standard temporal (sec) and fre- quency (kHz) parameters: overall duration, duration of first note, duration of middle note, duration of final note, number of notes, pace of notes (notes/min within a vocalization), pace of calls (calls/min within a recording), maximum overall frequency, minimum over- all frequency, and bandwidth. Playback experiments were conducted at Extrema to investigate the function of vocal- izations. Recordings were played through a Sony SRS-58 speaker from a range of 10-40 m; the subject of most playbacks (7/10) was the individual from which the tape was re- corded. Samples were too small to permit sta- tistical analysis. RESULTS Field Observations. — Intensive field work at CICRA has not yet produced records of the Rufous Twistwing, despite extensive bamboo and a broad community of bamboo-specialist birds. The species was more readily found at other sites, including Oceania, where it was encountered 19 times (1-4 times daily) during an 8-day survey. Fifteen of these encounters were aural and four were visual; two obser- vations lasted <1 min, one lasted —20 min, and one lasted —45 min. At least three indi- viduals were involved (2 were color-banded and 1 unbanded bird was later seen). The first record for Bolivia involved at least four individuals at Extrema where the species was encountered on 1 1 occasions during a 4-day survey. Eight observations were aural and three were visual. Ten encounters were within 2 km of the Extrema guard-post and one encounter was within Peru near the Alto Peru guard-post. One bird was photographed and tape-recorded on 7 November 2004; at least three other individuals were heard or seen daily along a 2-km transect. The first record for Brazil involved a single individual mist netted and photographed at Parque Zoobotanico, Rio Branco on 22 May 1998, although it was only identified to spe- cies in February 2006. The species was en- countered repeatedly at Parque Zoobotanico from 1998 onwards: nine individuals were mist netted, one specimen was collected, and another window-killed individual was found and prepared as a specimen. Both specimens are currently housed at Museu Paraense Em- ilio Goeldi (MPEG), Belem, Brazil (Table 1). The Rufous Twistwing was encountered at two further Brazilian sites. It was found once at Esta^ao Ecologica do Rio Acre during —250 hrs of field work: AA heard and tape- recorded a single individual on 15 August 2005 in the dense understory of an extensive thicket of Guadua bamboo along a narrow creek (Igarape do Tombo). Another individual was found by EG and Marcos Persio D. San- tos on 20 November 2006 at Ramal Jarinal and a male was heard calling consistently throughout the morning from a Guadua bam- boo thicket in terra firme forest. It was attract- ed with playback of its own calls and was col- lected; the skin is stored in MPEG (Table 1). Three other localities have recently come to light (Fig. 1). The Rufous Twistwing was found three times near the Machiguenga com- munity of Timpia (12° 04' S, 72° 49' W; ele- vation 410 m) on the Rio Urubamba, Dpto. Cuzco, in 1998-2001 (N. G. Gerhart, pers. comm.). It was also recorded once nearby at Rio Shihuaniro, a tributary of the Rio Timpia, in October 1998 and once near Montetoni (1 1° 54' S, 72° 21 ' W; elevation 550 m) on the upper Rio Camisea, Dpto. Cuzco in April 1998 (N. G. Gerhart, pers. comm.). A retro- spective report involves a single individual in Guadua bamboo at Explorer’s Inn, Tambopata (13°08'S, 69° 36' W; elevation 250 m) in March 1984 (Michael Kessler, pers. comm.). The Rio Shihuaniro locality is documented with a sound recording (ML 125894), as are the Timpia and Montetoni localities (N. G. Gerhart, pers. comm.); the Tambopata record is supported by a convincing description by an experienced observer. Adding the eight new localities to the original five produces a global range of —89,000 km- (Fig. 1). Mist Nett in and Specimen Data. — Two birds were mist netted at Oceania on 12 Oc- tober 2004, a capture rate of one bird per —200 net hrs (for 12-m nets). Both birds had 42 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. I, March 2008 1 o O m B £ cd 3 loo-' U c o o o (NCNCnIiTj — i TtTf O 0 0 0 [L. z (N CM in 0 -H — 0 CM CM 3 3 ~o 3 ' « 3 c ^ § -j v2 v2 2 2 ~ S S CQ CQ (U 250 hrs of field work at Esta^ao Ecologica do Rio Acre. In addition, it is the only regional bamboo spe- cialist that has not been recorded during 7 years of intensive fieldwork at CICRA. This absence is particularly of interest as the site contains extensive bamboo patches supporting all 19 of Kratter’s (1997) bamboo specialist Tobias et al. • RUFOUS TWISTWING BEHAVIOR AND DISTRIBUTION 47 birds (obligate, near-obligate, and facultative), along with six other bird species strongly or seasonally associated with bamboo in this re- gion. These results suggest the species is rel- atively specialized, or confined to larger or older blocks of habitat, as suggested by Lane et al. (2007). Results from three sites indicate the Rufous Twistwing can be relatively common and eas- ily encountered, especially when vocalizing. We recorded it 19 times in an 8-day survey at Oceania and 11 times in a 4-day survey at Extrema suggesting that, at least in optimum habitat, encounter rates can be fairly high. Consistent captures at Rio Branco also indi- cate that some populations can be sedentary for periods of several years in suitable habitat. Overall, our data suggest that populations are patchy and unpredictable, and vary widely in density. The Rufous Twistwing is known to sally from perches to catch insects and to lift one or other of its wings repeatedly (Lane et al. 2007). Our results affirm that wing raising is a common behavior in this species and sug- gest that wing raises tend to be given in one- sided bouts. The function of wing raises in this species and several other genera of tyrant- flycatcher remains unclear. They may serve as signals in intersexual contexts, although they are often given by solitary birds. The function of modified primaries in both Cnipodectes species has not been fully ex- plained, but primaries presumably are used in sound production (Zimmer 1939, Ridgely and Greenfield 2001, Lane et al. 2007). The wings of the male Brownish Twistwing make an au- dible mechanical pr’r’r’r’r’ wing-rattle in flight (Hilty and Brown 1986) and a similar sound has been heard from the Rufous Twistwing (Lane et al. 2007). It is likely that sound production in both species is prominent during courtship display, as in other understo- ry birds with modified wing feathers, i.e., Pi- pra manakins and Smithornis broadbills, for example. The Rufous Twistwing is more often heard than seen. The only vocalization previously reported is an agitation call which closely re- sembles an aggressive call of the Brownish Twistwing recorded in Ecuador (Krabbe and Nilsson 2003). Our vocal data enlarge the known repertoire of the Rufous Twistwing to at least three call types. In addition to the ag- itation call (Vocalization #1) noted by Lane et al. (2007), we describe a 5-7 note song (Vo- calization #2), and a 1-3 note pyew call (Vo- calization #3). One Rufous Twistwing at Ex- trema, Bolivia responded strongly to playback of Vocalization #2 and gave this call-type from a small area over 3 consecutive days, suggesting that it functions in advertising and courtship. The song of the Brownish Twistwing con- tains 1-3 notes (Eig. 3C in Lane et al. 2007). It is perhaps analogous to the Rufous Twist- wing’s pyew call (Vocalization #3), which is also given 1-3 times in succession (Eig. 2C, double version and Eig. 3A, triple version). A relationship between the pyew note of the Ru- fous Twistwing (Eig. 2C) and the song of the Brownish Twistwing (Eig. 2D in Lane et al. 2007) is suggested by similarities in their structure; both have a 2 kHz fundamental with one major harmonic and a ridge-shaped intro- duction which leads into a level section. How- ever, we believe it is unlikely the pyew note is the main song of the Rufous Twistwing as it is given rarely and from random points. Vocalization #2 is given at shorter intervals from fixed points and may be analogous to the stereotyped song of the Brownish Twistwing, although it is a longer, faster series of notes. The Rufous Twistwing does not appear to give an upwardly inflected cueet or kuuuu-wit note, the unsolicited single-note call of the Brown- ish Twistwing (Eig. 3D in Lane et al. 2007). It is possible the pyew call (Vocalization #3) is homologous with the cueet call. This con- clusion is supported by our observation that both calls are given at fairly long and irregular intervals. The vocal repertoire of the Rufous Twist- wing is similar to the Brownish Twistwing. However, presumed homologous vocalizations are diagnosably different between the two forms supporting their treatment as separate species. They may produce different mechan- ical noise; songs of the Brownish Twistwing are often introduced by bill snapping sounds (Hilty and Brown 1986), which have not yet been reported for the Rufous Twistwing. The wing-rattles of the two species have not been compared. Male Brownish Twistwings sing from ha- bitual song posts and do appear to defend 48 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. I, March 2008 pair-territories (Parker and Remsen 1987). They therefore differ from males of most ty- rant-flycatchers and resemble those of certain manakin genera (e.g., Neopelma and Tyran- neutes) with polygamous breeding systems. Our observations suggest the behavior of male Rufous Twistwing is similar. The sexual size dimorphism in both Cnipodectes species is ex- treme for tyrant-flycatchers (Lane et al. 2007), again suggesting polygamy. Further research is required to clarify breeding biology and ranging behavior in Cnipodectes . CONSERVATION STATUS We estimate a global range of -89,000 km^ on the basis of all 1 3 known localities for the Rufous Twistwing (Fig. 1). However, our re- sults fit the prediction of Lane et al. (2007) that the global range coincides with the dis- continuous block of Guadua— dominated forest that extends over much of southeastern Peru, northern Bolivia, and western Amazonian Brazil. The total area of this habitat is un- known, but has been estimated at -121,000 km2 (Nelson 1994) to -180,000 km^ (Saatchi et al. 2000). The Rufous Twistwing is probably the least abundant and perhaps the most threatened of all bamboo specialists in Amazonia. The risk of extinction in the short term is low but re- cent development projects, including the paved Trans-Oceanica Highway, will increase human settlement and habitat destruction in the re- gion (Nepstad et al. 2001, Conover 2003, To- bias and Brightsmith 2007). This may not be detrimental to the Rufous Twistwing because Guadua bamboo at times proliferates on de- forested land and abandoned chacras (Gris- com and Ashton 2003; D. F. Lane, pers. comm.), thereby increasing suitable habitat for the species. We note, however, that it is usu- ally recorded in large, mature patches of Guadua bamboo and appears to be absent from stretches of more recent Guadua re- growth along the proposed route of the Trans- Oceanica Highway. We also note the socio- economic value of large bamboos and the in- creasing tendency to harvest them (Bystria- kova et al. 2004). These factors suggest the extent of suitable habitat for the species may decline in the future. The Rufous Twistwing is more or less re- stricted to patchy bamboo forests. It is gen- erally scarce, and we have shown it to be ab- sent from some apparently suitable sites. More field data are needed before population size can be calculated but, for the purpose of clas- sifying its conservation status according to current data and lUCN guidelines (lUCN 2001), we estimate this species may have a declining global population of <10,000 indi- viduals, and that all subpopulations may con- tain fewer than 1,000 individuals. We there- fore recommend that it be listed on the lUCN Red List as Vulnerable under criterion C2a (i) (lUCN 2001) with a caveat that data quality is poor and that further surveys may lead to future changes in status. Only a small part of the proposed distribution of the Rufous Twistwing has been surveyed and the threat from habitat disturbance remains unclear. Field surveys are required throughout the bamboo forests of southwest Amazonia and the extent of bamboo should be assessed at intervals, perhaps through remote sensing techniques. ACKNOWLEDGMENTS Permission for field work was granted in Peru by Institute Nacional de Recursos Naturales (INRENA) and in Brazil by Institute Brasileiro do Meio Ambiente e dos Recursos Naturais Renovaveis (IBAMA). Logis- tical support and/or funding was provided by SOS Amazonia and WWF-Brazil (to Alexandre Aleixo), and Centro Nacional de Pesquisa para Conserva^ao das Aves Silvestres (CEMAVE) (to Edson Guilherme). We are indebted to Macaulay Library for loan of sound recording equipment, and to N. C. A. Pitman and R. B. Huayllapuma for logistical support. We thank J. M. Barnett, S. H. M. Butchart, C. D. Cadena, N. G. Ger- hart (deceased), Michael Kessler, D. E Lane, P C. Pul- garm R., Marcos Persio D. Santos, G. P. Servat, and Thomas Valqui for useful data, comments, and correc- tions Fieldwork was funded in part by a Fulbright grant (to D. J. Lebbin), a CNPq/SECTAM joint Re- gional Development Research Program research grant #35.0415/2004-8 (to Alexandre Aleixo), and a Junior Research Fellowship from Newnham College, Univer- sity of Cambridge (to Nathalie Seddon). LITERATURE CITED Alverson, W. S., D. K. Moskovits, and J. M. Shop- land (Editors). 2000. Bolivia: Pando, Rio Tahua- manu. Rapid Biological Inventories Report 1. The Field Museum, Chicago, Illinois, USA. Bystriakova, N., V. Kapos, and I. Lysenko. 2004. Bamboo biodiversity. UNEP-WCMC/INBAR, Cambridge, United Kingdom. Conover, T. 2003. Peru’s long haul: highway to nches or ruin. National Geographic June: 80-111. Tobias et al. • RUFOUS TWISTWING BEHAVIOR AND DISTRIBUTION 49 Fitzpatrick, J. W. 2004. Family Tyrannidae (tyrant flycatchers). Pages 170-462 in Handbook of the birds of the world (J. del Hoyo, A. Elliott, and D. A. Christie, Editors). Volume 9. Cotingas to pipits and wagtails. Lynx Edicions, Barcelona, Spain. Griscom, B. W. and P. M. S. Ashton. 2003. Bamboo control of forest succession: Guadua sarcocarpa in southeastern Peru. Forest Ecology and Man- agement 175:445-454. Guilherme, E. 2001. Comunidade de aves do campus e Parque Zoobotanico da Universidade Federal do Acre, Brasil. Tangara 1:57-73. Hilty, S. L. and W. L. Brown. 1986. A guide to the birds of Colombia. Princeton University Press, Princeton, New Jersey, USA. lUCN. 2001. lUCN Red List categories and criteria. Version 3.1. lUCN Species Survival Commission, Gland, Switzerland and Cambridge, United King- dom. Krabbe, N. and j. Nilsson. 2003. Birds of Ecuador, DVD-ROM. Bird Songs International, Westernie- land. The Netherlands. Kratter, a. W. 1997. Bamboo specialization by Am- azonian birds. Biotropica 29:100-110. Lane, D. E, G. P. Servat, T. Valqui H., and F. R. Lambert. 2007. A distinctive new species of ty- rant flycatcher (Passeriformes: Tyrannidae: Cni- podectes) from southeastern Peru. Auk 124:762- 772. Meyer de Schauensee, R. 1966. The species of birds of South America with their distribution. Acade- my of Natural Sciences, Philadelphia, Pennsylva- nia, USA. Montambault, j. R. 2002. Informes de las evalua- ciones biologicas de Pampas del Heath, Alto Mad- idi, Bolivia, y Pando, Bolivia. Conservation Inter- national, Washington, D.C., USA. Moskovits, D. K., W. S. Alverson, and I. Halm (Ed- itors). 2003. Bolivia: Pando, Federico Roman, Rapid Biological Inventories 06. The Field Mu- seum, Chicago, Illinois, USA. Nelson, B. W. 1994. Natural forest disturbance and change in the Brazilian Amazon. Remote Sensing Reviews 10:105-125. Nepstad, D., G. Carvalho, A. Barros, A. Alencar, J. Capobianca, j. Bishop, P. Moutinho, P. Lefeb- VRE, S. U. Lopes, and E. Prins. 2001. Road pav- ing, fire regime feedbacks, and the future of Am- azon forests. Forest Ecology and Management 154:395-407. Parker, T. A. and J. V. Remsen. 1987. Fifty-two Am- azonian bird species new to Bolivia. Bulletin of the British Ornithologists’ Club 107:94-107. Ridgely, R. S. R. and P. J. Greenheld. 2001. The birds of Ecuador: a field guide. Volume 2. Prince- ton University Press, Princeton, New Jersey, USA. Ridgely, R. S. R. and G. Tudor. 1994. The birds of South America: the suboscine passerines. Volume 2. University of Texas Press, Austin, USA. Saatchi, S. S., B. Nelson, E. Podest, and J. Holt. 2000. Mapping land cover types in the Amazon Basin using 1 km JERS-1 mosaic. International Journal of Remote Sensing 21:1201-1234. SCHULENBERG, T. S., C. A. MaRANTZ, AND P. H. EN- GLISH. 2000. Voices of Amazonian birds: birds of the rainforest of southeastern Peru and northern Bolivia. Cornell Laboratory of Ornithology, Itha- ca, New York, USA. SiLMAN, M. R., E. J. Ancaya, and j. Brinson. 2003. Los bosques de bambii en la Amazonia occidental. Pages 63-73 in Alto Purus: biodiversidad, con- servacion y manejo (R. Leite Pitman, N. Pitman, and P. Alvarez, Editors). Center for Tropical Con- servation, Lima, Peru. Tobias, J. A. and D. J. Brightsmith. 2007. Distribu- tion, ecology and conservation status of the Blue- headed Macaw Primolius couloni. Biological Conservation 139:126-138. Zimmer, J. T. 1939. Studies of Peruvian birds: notes on the genera Myiotriccus, Pyrrhomyias, Myi- ophobus, Onychorhynchus, Platyrinchus, Cnipo- dectes, Sayornis, and Nuttallornis. American Mu- seum Novitates 1043:1-15. The Wilson Journal of Ornithology 1 2()( 1 ):50-61 , 2008 NATURAL HISTORY AND BEHAVIOR OF THE ALDABRA RAIL (DRYOLIMNAS [CUVIERl] ALDABRANUS) ROSS M. WANLESS* 2 AND PHILIP A. R. HOCKEY’ ABSTRACT. — The Aldabra Rail {Dryolimnas [cuvieri} aldahranus) is endemic to Aldabra Atoll, Seychelles and is the last remaining flightless bird in the tropical western Indian Ocean. We studied it over two breeding seasons from 1999 to 2001. Pairs formed strong bonds, defended territories year-round, and were mate and territory faithful across seasons. Reproductive duties are shared by males and females. Breeding was closely tied to the rainy season and pairs responded to favorable conditions by increasing clutch size and clutch fre- quency. Clutches contained 1-4 eggs and replacement clutches were laid if nests failed early in the season. The incubation period was 19-24 days and hatching was usually synchronous. Mayfield estimates of daily survival of eggs and nesting success were 98.8% and 77.0%, respectively. Hatching success was 60.9% and 57% of chicks that hatched were successfully reared to independence. Chicks are sub-precocial and fledged chicks were cared for in the natal territory for at least 2 months, after which they were forcibly evicted. A large repertoire of behaviors and ritualized displays are described including pseudo-copulation and reverse mounting. Received 25 August 2006. Accepted 30 March 2007. The Aldabra Rail {Dryolimnas [cuvieri] al- dabranus) is endemic to Aldabra Atoll, Sey- chelles. Historically, distinct representatives of the genus Dryolimnas occurred throughout the Aldabra Group (Aldabra and Cosmoledo atolls and Astove and Assumption islands), as well as on Madagascar and the Mascarenes (Rand 1936, Rountree et al. 1952, Benson 1967, Mourer-Chauvire et al. 1999). Only the White-throated Rail (D. c. cuvieri) of Mada- gascar and the Aldabra Rail remain — all other forms are extinct. The Aldabra Rail is truly flightless (Wanless 2003a) unlike its sister tax- on on Madagascar (Rand 1936, Taylor and van Perlo 1998). The range of the Aldabra Rail had contracted by the 1970s to two is- lands within Aldabra (Malabar and Polymnie) and the small lagoon islet of lie aux Cedres (Benson and Penny 1971, Collar 1993, Taylor and von Perlo 1998). A reintroduction of Aldabra Rails to Picard Island was successfully undertaken in 1999 (Wanless et al. 2002). This provided the op- portunity to study the natural history and breeding biology of the Aldabra Rail. The species received some research attention in the 1970s, but the autecological study went unpublished (Huxley 1982). We located the manuscript and received permission (C. R. ' DST-NRF Centre of Excellence at the Percy FitzPatrick Institute, University of Cape Town, Ron- debosch. South Africa 7701. ^ Corresponding author; e-mail; rwanless@botzoo.uct.ac.za Huxley, pers. comm.) to use the results to cor- roborate many of our findings, particularly those relating to pair behaviors. The objec- tives of our paper are to describe aspects of the natural history of the Aldabra Rail, in- cluding adult, juvenile, and chick plumages. We also describe use of a reliable genetic technique for ascertaining gender of Aldabra Rails, and a suite of territorial and reproduc- tive behaviors and displays. METHODS Study Area. — Aldabra Atoll (9° 24' S, 46° 20' E) is 1,100 km southwest of the granitic Seychelles and 400 km north of Madagascar. It is a large, slightly raised coral atoll (34 X 14.5 km, land area = 155 km^). Four large islands (Grande Terre, Malabar, Picard, and Polymnie) form a rim around a substantial, shallow lagoon that contains numerous islets (Fig. 1). The climate is dry and tropical with an average annual temperature of 27° C. An- nual rainfall is variable, but averages about 1 ,200 mm with a pronounced wet season, usu- ally November- April (Stoddart 1971). The vegetation is low and either dense or open mixed scrub. The former includes large, al- most monospecific stands of Pemphis acidula. Aldabra Rails were studied mainly on Mala- bar and Picard islands between 1999 and 2001. Identification of Gender. — The method for gender identification in the field using bill col- or (Penny and Diamond 1971) proved unre- 50 Wanless and Hockey • ALDABRA RAIL BIOLOGY 51 FIG. 1. Aldabra Atoll, Seychelles, showing the four rimming islands, larger lagoon islets, and the four channels to the open sea. Islands with Aldabra Rail populations are marked with asterisks. liable based on necropsy of birds. We used a universal genetic technique (Fridolfsson and Ellegren 1999) to assign gender to all individ- uals that were trapped and banded. PCR blanks and positive controls were included to ensure reliable reactions. We compared bill and tarsus length measurements from 22 males and 22 females to investigate sexual di- morphism. Differences were tested with a r-test at the 0.05 significance level. Gender of birds was also assigned when handled follow- ing Penny and Diamond (1971), and the re- sults were compared to with genetic identifi- cation. Breeding Biology. — The location and con- struction materials of 23 active Aldabra Rail nests and the dimensions of nine were record- ed. Elevation was measured from the ground to the base of the nest. Nest length (longest axis) and width were measured to the nearest 5 mm across the centre of the cup. Cup depth was measured vertically from the deepest part of the nest to the cup rim. Eggs were measured to the nearest 0. 1 mm using vernier callipers and weighed to the nearest 0.5 g using a 100 g Pesola spring bal- ance. Many eggs were of unknown age and weights should be treated with caution. Average clutch size was estimated from nests visited at least twice during incubation. In instances when nests were not found, but newly hatched chicks were located, a mini- mum estimate of clutch size was assumed to equal the number of chicks. The response of birds to favorable conditions was examined by comparing clutch size and number of repeat clutches between the reintroduced population on Picard (no density-dependent constraints on reproduction) and on Malabar (high den- sity-dependence) (Wanless et al. 2002). Where appropriate, means ± one standard deviation are presented. We performed a Kaplan-Meier survival function analysis (which calculates the cu- mulative proportion of nests surviving over time) for the incubation period with survival censored at 20 days (Nur et al. 2004). Hatch- ing success was calculated from complete clutches only. We estimated the daily survival rate for eggs and nest success for the incu- bation period following Mayfield (1975): dai- ly survival = 1 — losses/e and nest success = (daily nest survival)'/', where e = exposure, the total number of active days per egg/nesl, and ip is the incubation period. A nest was defined as successful if at least one chick fledged. We considered chicks in full juvenile plumage and capable of foraging for them- selves to have been reared to independence. Chick survival is the number of chicks reared to independence divided by the number that hatched and breeding success is the number of chicks reared to independence per nesting attempt. Behavior. — We took detailed notes in the field describing interactions between birds. Particular attention was given to posture and the role of the white throat patch and undertail 52 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 coverts in displays. Birds were also photo- graphed in various postures from which line drawings were produced. Interactions were ei- g ther common (e.g., pairs duetting and greet- ^ ing) or only seen opportunistically (e.g., mat- m ing). Consequently, no attempts were made to quantify the relative frequency or individual | variation of behaviors, but Huxley’s (1982) ^ findings were incorporated or used to corrob- orate our observations. 49 - 47 - aA^ a ^ 45 - A A A Aa a AA A Aa □ 43 - qA cP ^ o □ . □ ^ □ 41 - □ mn ni □□ o A Male 39 - 37 - □ □ □ Female RESULTS Structure and Plumage. — The Aldabra Rail is a medium-sized rail with somewhat reduced wings that do not extend beyond the tail at rest. The body is slender, emphasized by the noticeable reduction of breast musculature. The upper parts are a dark, olive green, and the white throat patch contrasts strongly with the dark body plumage. The undertail coverts are also white. The bill is black, long, slender, and sharply pointed with the base variably pink or red. The head, neck, and breast of adults from Malabar and Polymnie are rich, dark, and red-brown with the flanks barred with white. Adults from lie aux Cedres are noticeably duller, paler, and tinged with pink. Newly hatched chicks are typical of rallids, being thickly covered with black down. Pri- mary feathers begin to emerge from their sheaths at about 2 weeks. Juveniles resemble adults but are duller with reduced contrast be- tween upper and lower body coloration. Identification of Gender.— The gender of 47 (81%) of 58 adults ascertained in the field was confirmed to be correct by genetic analysis; of the remaining 1 1 birds, six males were incor- rectly identified as females and five females as males. Males were significantly larger than females in bill length (43.5 ± 2.7 vs. 39.4 ± 1.5 mm, df = 42, P < 0.001) and tarsus length (45.4 ± 5.7 vs. 41.3 ± 1.3 mm, df = 42, P = 0.007). However, there was overlap be- tween large females and small males in both measures (Fig. 2), and we conclude that gen- der of Aldabra Rails cannot be reliably iden- tified in the field or in the hand. General Eco/ogy.— Paired birds roosted on the ground, usually less than 10 m apart, and were diurnal becoming active at first light by immediately performing a duet. This was fre- quently echoed by other nearby pairs in a 35 37 39 41 43 45 47 49 Bill length (mm) FIG. 2. Sexual dimorphism in correlated morpho- logical characters of the Aldabra Rail from 22 adult females and 22 adult males. domino effect. Occasionally, birds called or pairs duetted in the middle of the night. Aldabra Rails use their bills to remove leaf litter, probe soil, and eatch food; the feet were not used to scratch in litter or to manipulate food. Size of prey items was, in most cases, too small to estimate and no identifiable re- mains were found in fecal samples. A single regurgitation contained finely crushed arthro- pod exoskeleton(s) and 18 tiny land-snail shells. The latter might have been ingested as grit to aid digestion. Aldabra Rails hunted geckos, skinks, and crabs opportunistically. These prey were held in the bill when cap- tured, while the bird alternately flicked its head rapidly from side to side and then beat the prey against the ground or a branch. Larg- er crabs were killed by rapid, powerful blows with the bill directed between the eyes. Crab exoskeletons and hermit crab shells were bro- ken open by placing them on the ground and repeatedly hammering them from above. Rails washed large food items if water was nearby. Breeding Biology. — Courtship behavior pri- or to pair formation has not been reported. New pair formation occurred within 5 days after reintroduction to Picard Island, but could have happened earlier. Aldabra Rails are mo- nogamous and both males and females de- fended the territory year-round. Territory fi- delity between years was high; of 14 pairs on Picard Island where both individuals were col- or-banded, only two pairs were suspected of having shifted territory. Males and females shared incubation and chick-provisioning du- Wanless and Hockey • ALDABRA RAIL BIOLOGY 53 TABLE 1. Clutch sizes and breeding success (defined as chicks reared to independence per nest) of Aldabra Rails. ND denotes no data. Total nests Mean clutch Total Chick Breeding Parameter (eggs) size hatched survival success All clutches (Picard and Malabar) 23 (65) 2.83 ND ND 1.11 All clutches (Picard only) 17 (51) 3.0 30 20 1.05 First clutches (Picard only) 9(28) 3.11 21 14 1.27 Repeat clutches (Picard only) 8 (23) 2.88 9 6 0.75 ties. There was no evidence of cooperative breeding in the species and young were in- variably evicted from the natal territory, usu- ally at 8-10 weeks of age, once they had at- tained full juvenile plumage. The most conspicuous pair behavior was the duet, which consists of a series of loud, high-pitched, ascending whistles that increase in pitch to a sustained crescendo. Duets were given after a territorial dispute, after the Greeting Display, or in response to other pairs duetting nearby. We interpret duetting as an integral part of pair-bonding behavior. Pair- bonding activities were increasingly notice- able as the breeding season approached. Du- etting and response to playback of duets were noticeably reduced during chick rearing, and other pair-bonding behaviors were seldom seen. Nesting Ecology. — Nest site prospecting be- gan in November or early December and males did most of the building. Two or more nests may be built and material was often placed in several potential sites. Once a site was selected, one or both birds stood on or near the site and gave a series of low, groan- ing crreaak calls. Eight of 23 active nests were on the ground; the other 15 were in low shrubs, Pandanus tectorius bushes, or natural cavities in trees. Elevated nests ranged from barely above ground to 1.6 m above ground (mean = 0.8 m, n = 14). Only 1 of 23 nests lacked some overhead shelter. A wide diversity of dry plant material (usu- ally the dominant litter of the area) was used in nest construction. Two active nests, how- ever, had almost no added material and were natural depressions (under a shrub and in a tree cavity, respectively), sparsely lined with dry grasses or Casuarina needles. Nests were usually constrained by the na- ture of the site, but most were quite large and roughly circular (260 ± 33 X 221 ± 22 mm, n = 23). The mean depth was 116 ± 61 mm (n = 9). The mean cup length was 149 ± 22 mm, mean width was 134 ± 27 mm and the mean depth was 48 ± 24 mm (n = 9 for all cup measurements). The cup was usually suf- ficiently deep that only the head of an incu- bating adult protruded; the adult’s black- streaked, dull olive-green back camouflaged it well on the nest. Three pairs used brooding nests to brood downy chicks; others may have gone undetected. These were built on the ground under shelter and consisted of a shal- low cup lined with a thin layer of dry material. Mean egg dimensions were 43.3 ± 1.2 X 30.6 ± 1.1 mm, and mean mass was 21.4 ± 2.1 g {n = 39). Eggs varied in shape, but most were elliptical or biconical. The moderately glossy, off-white shell is covered in maroon- brown speckling and has a finely granulated surface. Speckling is most dense at the obtuse end and many speckles are clouded or ad- umbrated, giving a multi-toned appearance. Twenty-three clutches were monitored from egg-laying or incubation to completion (Table 1). Small unequal sample sizes and the non- independence of samples (repeat clutches) prevented testing for differences in mean clutch size between the reintroduced popula- tion from Picard and the source population on Malabar. Average clutch sizes were slightly larger on Picard (3.0, range 2-4 eggs, n = 17) than on Malabar (2.33, range 1-3 eggs, n = 6). Excluding repeat clutches increased mean clutch size on Picard to 3.11 eggs {n = 9 nests). Six pairs on Picard attempted second clutches after successfully rearing chicks, tour of which were successful. Two pairs laid third clutches, both of which failed at the egg stage (Guy Esparon, pers. comm.). A pair on Mal- abar in 2()()0/2()() 1 built a new nest and laid immediately after their first clutch was dep- 54 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 0.95 - 0.9 - at s 0.85 B 0.8 0.75 o- 6- 5 10 15 20 Survival time (days) FIG. 3. Kaplan-Meier survival probabilities for Aldabra Rail eggs in the 2000/2001 breeding season with all data censored at 20 days. redated; we found no other repeat clutches on Malabar. Timing of rail reproduction was closely tied to onset of the wet monsoon and concomitant increase in invertebrate densities. Laying of first clutches in the 2000/2001 breeding sea- son on Picard was concentrated around the end of December and early January, although one successful clutch was laid in mid-Novem- ber, before the monsoon began. Eggs were laid at intervals of 1-4 days, typically every other day. The incubation period (from last egg to first-hatched chick) was 19-24 days (mean = 21.2 days, n = 5 nests). Individual eggs took 22-24 days from laying to hatching (mean = 22.6 days, mode = 22 days, n = U eggs). The hatching interval was 0-3 days, synchro- ny being the norm, implying that incubation typically starts after clutch completion. Chicks are semi-precocial (sensu Starck and Ricklefs 1998); they hatch downy and are able to fol- low parents from hatching onwards. They can leave the nest within a few hours of hatching and abandoned the nest within 3 days. Egg- shells were not found and one female was ob- served to eat the hatched egg shells. Infertile eggs were abandoned in the nest {n = 2). Newly hatched chicks left the nest for periods in the care of one parent when infertile or late- developing eggs were being incubated. Repeat clutches were timed so that juveniles from the first brood were evicted from the territory be- fore the next clutch hatched {n = 6). Parents adopted one of two caring and provisioning strategies once the nest was abandoned. They either divided the brood between themselves and provisioned young independently, or they left chicks well concealed, foraged alone, and returned to the chicks with food. The latter strategy was only used for downy chicks. Downy chicks often gave a continuous, high- pitched peep begging call while following for- aging adults. Chicks immediately ran to the nearest dark place when parents gave a danger call. Parents provisioned chicks until they were 6-8 weeks of age and became aggressive towards chicks when they attained full juve- nile plumage, at ~8 weeks of age. Aggressive interactions continued until offspring left the natal territory (generally at 8-10 weeks of age). Two of 12 nests found during incubation failed to produce any chicks. The Kaplan- Meier survival function for eggs (Fig. 3) showed relatively low failure rate due to pre- dation. The overall survival rate was inflated because eggs that were infertile but survived Wanless and Hockey • ALDABRA RAIL BIOLOGY 55 beyond 20 days were not counted as failures in the analysis. Overall, the daily survival rate of eggs was 97.5% and the Mayfield estimate of nest success during incubation was 77.0%. Hatching success was 60.9% (28 of 46 eggs from 15 complete clutches). The 18 eggs that failed did so because of presumed predation (7), breakage (3), and presumed infertility/oth- er causes (8). Only one nest (containing 2 eggs) was entirely depredated: all other pre- dation events involved the loss of single eggs. Chick survival was 57% with 16 of 28 hatch- lings surviving to independence. Breeding success on Picard was 1 .05 chicks per nesting attempt {n = \1 attempts), although breeding success of repeat clutches following success- ful first clutches was only 0.75 chicks per nesting attempt (n = ^ attempts. Table 1). Behavior. — Adult Aldabra Rails used a va- riety of vocalizations, postures, and displays, mostly in territorial or sexual contexts. Two prominent plumage features, the white throat patch and white undertail coverts, had partic- ular significance in postures and displays. These features contrast strongly with the dark- er plumage of the rest of the bird and also stand out against the surroundings of the un- derstory. The Relaxed Posture was used when rest- ing, foraging, and walking. When standing, the neck was bent or retracted with the head pointing horizontally or down, hiding the throat patch. The tail was flicked with each step, but the undertail coverts were not fanned. There were no associated vocaliza- tions except during the early breeding period when males mate-guard and both males and females gave the mpclick call continuously and partially displayed the undertail coverts. When sleeping, the head and neck were pulled in and the bill held horizontally, at times while the bird stood on one leg. Aldabra Rails bathed regularly if fresh wa- ter was available. Bathing was done in shal- low water and the head dipped repeatedly un- der the water in a bobbing fashion, spilling water over the upper parts on each rise and wetting the underparts on each dip. A wet bird may sun itself with the wings fanned horizon- tally. There are no records of alternative bath- ing activities (e.g., dust bathing or bathing in the sea). When preening, the eyes were kept shut. Allopreening was initiated by gently pecking around the head area. It was initiated by males or females and was also targeted at young birds. Allopreening often preceded oth- er pair-bonding behavior such as duetting and copulation. The Greeting Display (Fig. 4A) was usually performed when pairs reunited after foraging apart; they trotted towards each other with necks partially stretched and vertical, display- ing their throat patches. Both gave soft, purr- ing calls and occasionally touched bills. Once birds met, contour feathers were slightly raised and undertail coverts fanned, but the white throat patch was concealed by the low head position and retracted neck. The Singing Posture (Fig. 4B) was used by individuals when calling singly and when du- etting. Birds often ran towards each other when duetting and stood close. The bill was held slightly below horizontal and the neck stretched and angled forward and upwards. The body was angled slightly upwards and the legs held in a partial crouch. Calls were var- iable in length; after calling, birds usually adopted the Relaxed Posture. The Curious Posture (Fig. 4C) was given when a rail encountered a novel stimulus such as humans and/or their equipment. The con- text was similar to that of the Alert Posture, although the latter was used when a bird’s at- tention was drawn to something further away, such as when territorial intruders were detect- ed. The neck was stretched and held horizon- tal. The body was also horizontal and the head at times cocked at an angle. The neck was often retracted and extended at a new angle. The legs were at times in a half-crouch, as if ready to flee. The contour feathers were not raised. The Alert Posture was adopted when a bird was in a similar state of arousal as for the Curious Posture, but was not restricted to in- vestigating novel stimuli. It is similar to the Upright Posture, but the undertail coverts were not displayed. The body was upright with the neck fully or partially stretched and the throat patch partially obscured. The tail pointed down and coverts and contour feath- ers were not fanned. The Territorial Defense Display (Fig. 4D) was given when a conspecific intruded on a territory or in a boundary dispute. The body was horizontal and the neck pulled in, par- 56 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 FIG 4 Postures and displays of the Aldabra Rail based on photographs and field observations. (A) Greeting Display (B) Singing Posture, (C) Curious Posture, (D) Territorial Defense Display, Upright Posture, Invhation Display, (G) Mounting, (H) Coitus, (I) Post-copulatory Display, (J) Full Nest Defense Display, ( ) Partial Nest Defense Display, and (L) Aggressive and Submissive displays. Wanless and Hockey • ALDABRA RAIL BIOLOGY 57 tially obscuring the throat patch. The tail was raised, coverts fully fanned, and the contour feathers were all raised. Wings were occasion- ally held slightly away from the body with primaries partially splayed. Birds approached each other and moved from side to side, flick- ing the tail with each step. If neither protag- onist retreated, they adopted the Upright Pos- ture and typically a fight ensued. The Upright Posture (Fig. 4E) was also used in the context of territorial defense. The body was upright and the neck fully extended and held vertically, displaying the throat patch. The tail was held horizontal or pointing slightly downwards with coverts fully fanned. Contour feathers were slightly raised on oc- casions. It was used during Aggression Dis- plays by the aggressor. After the Upright Pos- ture was adopted by one or both individuals, if one bird backed down it was usually chased, if not, they fought. Chasing birds held their necks up and out and displayed their throat patches and undertail coverts, whereas fleeing birds kept the tail down, concealing the un- dertail coverts. Chasing birds flapped their wings to assist in maneuvering. The Chasing Posture was also used by adults when evicting their offspring from the natal territory and, in a trespassing situation, against neighbors or their offspring. The pursuing bird pecked the fleeing bird whenever it could reach it. Birds circled each other as a prelude to fighting or approached head-on in the Terri- torial Defense Stance. They then suddenly adopted the Upright Posture (fully displaying their throat patch) and attacked, leaping and kicking each other simultaneously, while flap- ping their wings vigorously and directing blows with their bills. They either landed on their feet and repeatedly circled and jumped at each other, alternating between the Upright and Territorial Defense Postures, or they fell to the ground, facing each other while balanc- ing on their tails and wings and kicking vig- orously. Fights were violent and occasionally resulted in serious injury or death. A series of postures was used before, dur- ing, and after copulation. The sequence of events appeared to be fixed with the exception of the initiation. Incomplete events may have been due to the presence of observers, but we believe mounting was a pair-bonding activity; when cloacal contact was absent, the behavior is referred to as Pseudo-copulation. Pseudo- copulation was frequent in the months before reproduction began and after young had left the natal territory. Copulation was usually preceded by duet- ting. It was initiated in one of two ways. Birds adopted the Approach Posture, moved to- wards their mates with the neck stretched ver- tically, body angled up (but not as much as in the Upright Posture), displaying the throat patch and undertail coverts. Alternatively, ini- tiation of the sequence began when a bird faced away from its mate and performed the Invitation Display (Fig. 4F). This was present in all copulation sequences observed. A bird in the Invitation Display arched its neck and pointed the bill downwards. The body was an- gled upwards and the bird waddles forward a few paces in a half crouch, slowing quickly to a standstill. The tail was pointed down, the undertail coverts fanned and the wings folded. A responsive mate adopted the Approach Pos- ture, walked or trotted towards its mate, and then mounted the inviting bird (Fig. 4G). It then crouched on the latter’s back, the body almost horizontal, tarsi touching the other bird’s back, the head pointing down at 45°, with wings folded and the undertail coverts fully fanned. The mounted bird then raised its tail while displaying the undertail coverts. The head and neck positions changed repeatedly but were probably related to maintaining bal- ance. The mounting bird then tread on the oth- er bird’s back a few times, leaned back, flut- tered its wings for stability, and moved its tail under that of the inviting bird’s to make clo- acal contact, usually for no more than 1-2 sec (Fig. 4H). Normally, the entire copulation se- quence was repeated with roles reversed, in- cluding reverse mounting. After dismounting, the bird performed the Post-copulatory Dis- play (Fig. 41), walking in a semi-circle around the front of its mate, which remained in the Invitation Display. The wings were fanned and vertical, with the leading edge towards the ground. The contour feathers were raised, with the head held low and the neck held the same way as in the Nest Defense Display. Nest Defense Displays (Fig. 3J-K) were di- rected at any intruder that approached too close to the nest or chicks, except for crabs, which were vigorously pecked. Birds rushed towards the intruder in the Full Nest Defense 58 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 Display (Fig. 4J) with wings held fully open and vertical, and the leading edge pointing to- wards the ground. The body was held low and horizontal and the head kept close to the body, facing the intruder while obscuring the throat patch. The undertail coverts were fully fanned. The mantle and back feathers were slightly raised. The displaying bird stopped and walked back and forth ~1 m in front of the intruder. Birds also made rapid attacks with a single, jabbing peck at the intruder. If the in- truder didn’t advance further, the displaying bird would at times circle the intruder. The intensity of the display decreased quite rapidly and birds adopted the Partial Nest Defense Display (Fig. 4K) with both wings partially folded or, more usually, the wing away from the intruder completely folded. This was fol- lowed by a Staring Display identical to the Territorial Defense Display except the con- tour feathers were completely raised and the displaying bird usually stood still and stared directly at the intruder; the contour feathers were gradually lowered as the display ended. An incubating or brooding bird stayed on the nest until the intruder was within 1 m if a mate was present and displaying. The bird left the nest in the Full Nest Defense Display if the intruder came closer, and either charged at the intruder or moved to the side (possibly in an attempt to divert attention from the nest). If an intruder attempted to raid the nest, the incubating bird stood next to or over the nest and attacked the intruder, making powerful, jabbing pecks and giving a loud, piercing squeal that was only given in this context. Birds often followed a retreating intruder, at times in the Partial Nest Defense Display. At this ostensibly critical juncture, birds occa- sionally stopped mid-display to feed; this may have been displacement behavior. When ‘courtship feeding’ or provisioning young, a bird with food in its bill approached the intended recipient and adopted the Provi- sioning Posture. The head was pointed verti- cally downwards so the bill almost touched the ground. The rest of its body was in the Relaxed Posture. This posture was maintained until the food item had been taken. If a chick dropped food or struggled to handle it, the adult would pick it up repeatedly, at times breaking the item into smaller pieces. The chick usually approached the adult in this pos- ture from the side and underneath, small chicks even standing under the chest of the adult. Aggressive and Submissive Displays (Fig. 4L) were distinct from the Territorial Defense Display and differed from the positions for fighting in that only one bird was aggressive. These displays formed part of intra-pair be- havior and were also used when adults evicted young from the natal territory. When a bird performed the Aggressive Display, it adopted the Upright Posture while approaching anoth- er bird. It then stood over the other bird and leaned forward when attacking. The aggressor would also grab the other bird’s head or neck in its bill or push the other bird with its chest. The bird being attacked either adopted a Sub- missive Posture or fled. In the Submissive Posture, the bird crouched on the ground with its neck retracted and the head pointing down and away from the aggressor, hiding the throat patch. No undertail coverts were visible, the wings were kept folded, and the contour feath- ers were flattened. We only observed intra- pair aggression arising from direct competi- tion for food outside the breeding season. During the breeding season it appeared to be a pair-bonding behavior, but was also used in competition for food. Where gender of birds in intra-pair aggressive interactions was known, the aggressor was usually male. DISCUSSION Aldabra Rails form monogamous pairs that vigorously defend their territories against in- truders; both pair bonds and territory bound- aries were stable across breeding seasons. Males and females shared duties associated with breeding (e.g., incubation and chick pro- visioning) and territoriality (e.g., territorial de- fense). They foraged in all vegetation types on Aldabra but preferred dense mixed scrub (Penny and Diamond 1971, Wanless 2002). They were seldom encountered in monospe- cific stands of Casuarina equisetifolia, possi- bly reflecting a paucity of understory cover or low invertebrate prey densities (Spaull 1979). They regularly foraged in mangrove {Rhizo- phora sp.) swamps, but vacated these at high tide (Penny and Diamond 1971, this study). In addition to terrestrial arthropods, which dom- inate the diet, they also eat berries, flower pet- als, intertidal invertebrates, organic jetsam on Wanless and Hockey • ALDABRA RAIL BIOLOGY 59 beaches, turtle and tortoise eggs and hatch- lings, skinks, geckos, crabs, carrion, and kitchen scraps (Penny and Diamond 1971, Frith 1977, Huxley 1982, this study). Neither this study nor Huxley’s (1982) quantified the diet of Aldabra Rails. This was due to the dif- ficulties of observing birds in dense vegeta- tion without disrupting their feeding. Slow- motion replay of video recordings is probably the only way to gather reliable data on feeding ecology. A significant finding of this study is that gender of Aldabra Rails cannot be reliably as- certained in the field. We incorrectly identified gender of 1 1 birds that were in the hand at the time and could be closely examined. This casts doubt over interpretations of Aldabra Rail biology and behavior from earlier studies (e.g.. Penny and Diamond 1971, Huxley 1982). It also led to the discovery of reverse mounting in the species. A second finding that refutes published in- formation about Aldabra Rails relates to their territoriality. Penny and Diamond (1971) ar- gued that territories are not necessarily con- tiguous, but conform to a ‘neighborhood’ sys- tem, where core territories are defended but communal foraging areas exist. Neither Hux- ley (1982) nor we found any evidence to sup- port this assertion. Huxley (1979) described a mutualistic re- lationship between Aldabra Rails and Aldabra giant tortoises {Dipsochelys dussumieri). The rails reportedly glean flies, dead skin or ec- toparasites from tortoises, which in turn assist by adopting a distinctive posture, exposing the full surface of skin to be gleaned. Neither we nor anyone else with whom we have worked or consulted has witnessed this interaction. On the contrary, Aldabra Rails regularly foraged less than a meter from giant tortoises and nei- ther species showed the slightest interest in the other. A recent text on the Aldabra Rail (e.g., Sin- clair and Langrand 1998) treated it as a dis- tinct species from the White-throated Rail of Madagascar. A taxonomic review based on genetics and vocalizations is presently under- way, but a greater understanding of the Al- dabra Rail's biology is necessary in under- standing its evolution, distinctiveness, conser- vation status, and threats to its persistence on Aldabra. In addition, there are consistent dif- ferences in plumage coloration between birds from lie aux Cedres versus Malabar and Po- lymnie islands that merit closer investigation. The reintroduced population (on Picard) produced clutches of up to four eggs, in con- trast to the limited evidence of three-egg clutches being the maximum on Malabar. This suggests the species’ maximum clutch size is four. We failed to find second clutches follow- ing successful first clutches on Malabar, but at least six pairs on Picard attempted double- brooding. The Aldabra Rail’s response to good breeding conditions appears to be to in- crease reproductive effort by increasing clutch size and clutch frequency. Huxley (1982) re- ported two repeat clutches from Malabar pairs (no total sample size given). In general, it is likely that replacement clutches will be laid but repeat clutches are expected to be quite rare. There are three points of particular interest regarding reproductive displays of the rails. One, termed pseudo-copulation, involves mounting but excludes cloacal contact. We in- terpret this as a pair-bonding activity. The sec- ond is reverse mounting (where the female mounts the male). This behavior has been re- ported from a diversity of bird species (James 1983), including rallids (e.g., Anderson 1975, Brooke 1992). Reverse mounting is known to be a regular component of pair bonding/sexual behavior in relatively few bird species (James 1983, Nuechterlein and Storer 1989). Last, the Post-copulatory Display was described by Frith (1977) who conjectured that it was re- stricted to extra-pair copulations — this is not the case. Aldabra has five species that are potential nest predators: the terrestrial Robber Crab (Birgus latro), Malagasy Kestrel (Falco new- toni), Malagasy Coucal (Centropus toulou). Pied Crow (Corvus alhiis), and black rat {Rat- tus rattus), the latter being an introduced spe- cies. The loss of seven eggs (5 of which were individual losses from successful clutches) and 12 of 28 chicks suggests that nest pred- ators are successful in raiding Aldabra Rail nests and in preying on nestlings. However, Aldabra Rails have strong nest and chick de- fense instincts. First, they do not Hush from the nest in the same way as species that have high adult mortality rates at the nest (cf. Con- way and Martin 2()()(), Lloyd 2004). Second, 60 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. I, March 2008 the hiding behavior of chicks in response to the danger call from parents is strong. Fur- thermore, no adult Aldabra Rail has ever been reported to be attacked by predators. These data lend support to the hypothesis that eggs and chicks are at risk of predation, but adults are not, or at least much less so, given the absence of other (introduced) predators. We posit that Aldabra Rails have evolved on Al- dabra in the absence of predators of adults and lack appropriate behaviors. The Nest Defense displays described would be entirely inappro- priate for a predator capable of taking adult Aldabra Rails. The present-day absence of significant predators of adult Aldabra Rails is fortuitous. They are not at risk from black rats (Penny and Diamond 1971; Frith 1977; Huxley 1982; Wanless 2002, 2003b). Rats typically retreat in the face of aggression from Aldabra Rails, which are capable of killing them (Wanless 2003b). The ranges of Aldabra Rails and feral domestic cats {Fells catus) do not overlap on Aldabra and cats are thought to be largely re- sponsible for the local extinction of Aldabra Rails on Grande Terre and Picard islands (Wanless 2002). However, local extinctions happened before many interactions were re- corded and there are only two records of Al- dabra Rails interacting with cats (Huxley 1982, Hambler et al. 1993). In both instances, Aldabra Rails adopted a Nest or Territory De- fense Posture rather than flee. This implies that Aldabra Rails do not have an appropriate (i.e., fleeing) response to cats, and do not treat them as potential predators. The presence of the Norway rat {Rattus norvegiciis) elsewhere in the Seychelles is a source of concern, as it is bigger and more aggressive than the black rat, and its impacts on insular avifauna are much greater (Towns et al. 2006). We predict that inappropriate responses of Aldabra Rails to cats (and to humans close to their nests) are likely to be similarly maladaptive for other novel predators. Dispersal of cats, Norway rats or other mammalian predators to any of the four islands where the Aldabra Rail now occurs will pose a significant threat to the sur- vival of the species. Quarantine/bio-security procedures that minimize risk of further col- onizations should be a priority and put in place for all visits to Aldabra. ACKNOWLEDGMENTS Rachel Wiseman and Richard White provided in- valuable help in the field. Gabby Raaff produced il- lustrations of the postures and displays. Jessica Cun- ningham performed the genetic identification of gen- der. Some of the descriptive terminology we used to classify and describe postures and displays was origi- nally used by C. R. Huxley. Barry Taylor, John Cooper, J. R. Young and an anonymous referee greatly im- proved earlier drafts. We received financial and/or lo- gistical support from the Seychelles Islands Lounda- tion, Dutch Trust Lund, Conservation International, and the Center for Applied Biodiversity Science, the (South African) National Research Loundation, and the University of Cape Town. LITERATURE CITED Anderson, A. 1975. A method of sexing Moorhens. Wildfowl 26:77-82. Benson, C. W. 1967. The birds of Aldabra and their status. Atoll Research Bulletin 118:63-111. Benson, C. W. and M. J. Penny. 1971. The land birds of Aldabra. Philosophical Transactions of the Royal Society of London, B 260:417—527. Brooke, R. K. 1992. Apparent reverse mounting in the African Rail Rallus caerulescens. Ostrich 63:185. Collar, N. J. 1993. The conservation status in 1982 of the Aldabra White-throated Rail Dryolimnas cuvieri aldabranus. Bird Conservation Interna- tional 3:299-305. Conway, C. J. and T. E. Martin. 2000. Evolution of avian incubation behavior: influence of food, cli- mate and nest predation. Evolution 54:670-685. Eridolfsson, a. K. and H. Ellegren. 1999. A simple and universal method for molecular sexing of non- ratite birds. Journal of Avian Biology 30:1 16-121. Erith, C. B. 1977. Life history notes on some Aldabra land birds. Atoll Research Bulletin 201:1-15. Hambler, C., J. M. Newing, and K. Hambler. 1993. Population monitoring for the flightless rail Dry- olimnas cuvieri aldabranus. Bird Conservation International 3:307-318. Huxley, C. R. 1979. The tortoise and the rail. Philo- sophical Transactions of the Royal Society of London, B 286:225-230. Huxley, C. R. 1982. Unpublished typescript on the Aldabra Rail. Archived in Royal Society of Lon- don Library, London, United Kingdom. James, P. C. 1983. Reverse mounting in the Northwest Crow. Journal of Eield Ornithology 54:418-419. Lloyd, P. 2004. Variation in nest predation among arid-zone birds. Ostrich 75:228-235. Mayfield, H. P. 1975. Suggestions for calculating nest success. Wilson Bulletin 87:456—466. Mourer-Chauvire, C., R. Bour, S. Ribes, and E Moutou. 1999. The avifauna of Reunion Island (Mascarene Islands) at the time of the arrival of the first Europeans. Pages 1—38 in Avian paleon- tology at the close of the 20th Century: Proceed- ings of the 4th International Meeting of the So- Wanless and Hockey • ALDABRA RAIL BIOLOGY 61 ciety of Avian Paleontology and Evolution (S. L. Olson, Editor). Washington, D.C., USA. Nuechterlein, G. L. and R. W. Storer. 1989. Reverse mounting in grebes. Condor 91:341-346. Nur, N., a. L. Holmes, and G. R. Geupel. 2004. Use of survival time analysis to analyze nesting suc- cess in birds: an example using Loggerhead Shrikes. Condor 106:457-471. Penny, M. J. and A. W. Diamond. 1971. The White- throated rail, Dryolimnas cuvieri on Aldabra. Phil- osophical Transactions of the Royal Society of London, B 260:529-548. Rand, A. L. 1936. The distribution and habits of Mad- agascar birds. Bulletin of the American Museum of Natural History 72:143-499. Rountree, R. R. G., R. Guerin, S. Pelte, and J. Vin- son. 1952. Catalogue of the birds of Mauritius. Mauritius Institute Bulletin 3:155-217. Sinclair, I. and O. Langrand. 1998. Birds of the In- dian Ocean islands. Struik, Cape Town, South Af- rica. Spaull, V. W. 1979. Distribution of soil and litter ar- thropods on Aldabra Atoll. Philosophical Trans- actions of the Royal Society of London, B 286: 109-117. Starck, j. M. and R. E. Ricklefs. 1998. Patterns of development: the altricial-precocial spectrum. Pages 3-30 in Avian growth and development (J. M. Starck and R. E. Ricklefs, Editors). Oxford University Press, New York, USA. Stoddart, D. R. 1971. Rainfall on Indian Ocean coral islands. Atoll Research Bulletin 147:1-21. Taylor, B. and B. van Perlo. 1998. Rails: a guide to the rails, crakes, gallinules and coots of the world. Pica Press, Sussex, United Kingdom. Towns, D. R., I. A. E. Atkinson, and C. H. Daugh- erty. 2006. Have the harmful effects of intro- duced rats on islands been exaggerated? Biologi- cal Invasions 8:863-891. Wanless, R. M. 2002. The reintroduction of the Al- dabra Rail Dryolimnas cuvieri aldabranus to Pi- card Island, Aldabra Atoll. Thesis. University of Cape Town, South Africa. Wanless, R. M. 2003a. Can the Aldabra White- Throated Rail Dryolimnas cuvieri aldabranus fly? Atoll Research Bulletin 508:1-7. Wanless, R. M. 2003b. Flightless Aldabra Rail Dry- olimnas cuvieri aldabranus kills black rat Rattus rattus. Ostrich 74:134. Wanless, R. M., J. Cunningham, P. A. R. Hockey, J. Wanless, R. W. White, and R. Wiseman. 2002. The success of a soft-release reintroduction of the flightless Aldabra Rail {Dryolimnas [cuvieri] al- dabranus) on Aldabra Atoll, Seychelles. Biologi- cal Conservation 107:203-210. The Wilson Journal of Ornithology 120(l):62-73, 2008 POST-FLEDGING MOVEMENT AND SPATIAL HABITAT-USE PATTERNS OF JUVENILE SWAINSON’S THRUSHES JENNIFER D. WHITE' AND JOHN FAABORG' ABSTRACT We used radio telemetry to study post-fledging dispersal patterns of juvenile Swainson s Thrushes {Catharus ustulatus) in central coastal California from 2000 to 2002. We followed 30 different broods during the dependent period and 35 juveniles after independence. Adults with dependent juveniles had two types of movements, drifting and stationary, similar to those described for Wood Thrush {Hylocichla mustelina) A juveniles established at least one post-fledging dispersal area (n = 29) after independence and most (62%) established a second; therefore, most home ranges were comprised of disjunct dispersal areas. Median home range (95% fixed-kernel) and core (50% fixed-kernel) areas were 2.0 ha (range = 0.2-9.0) and 0.41 ha (range = 0.05-1.6), respectively. Juveniles spent from 3 to 39 days in different post-fledging dispersal areas; they generally had a primary dispersal area where they spent the majority of their time. Juveniles spent an average of 22 7 ± 15 days in primary and 9.6 ± 0.9 days in secondary dispersal areas. The median post-fledgmg dispersal distance was 147 m (range = 28-1,040 m) for initial dispersal from a brood-rearing area and 418 m (range = 154-2,624 m) for subsequent dispersal movements. Initial dispersal was not directed; however, final dispersal (brood-rearing area to final location) was directed to the northwest for all plots combined and west for one plot. We suggest that post-fledging movements and spatial habitat-use patterns are affected by adult b^ee mg strategies, and by spatially and temporally clumped resources. Received 2 October 2006. Accepted 17 July 2007. Post-nesting movements and spatial habitat- use patterns may reveal important aspects of post-fledging ecology for juveniles and post- breeding ecology for adults; however, little in- formation exists on movements and spatial patterns of juvenile birds. This is largely be- cause juveniles are non-territorial and cryptic and, therefore, difficult to follow as they dis- perse from the natal area. Post-fledging move- ment patterns have only been described for a few passerine species including Wood Thrush {Hylocichla mustelina) (Anders et al. 1998, Vega Rivera et al. 1998, Lang et al. 2002, Fink 2003), White-throated Thrush {Turdus assimilis) (Cohen and Lindell 2004), Lark Bunting {Calamospiza melanocorys) (Yackel Adams et al. 2001), Ovenbird {Seiurus auro- capillus) (Bayne and Hobson 2001), Hooded Warbler {Wilsonia citrina) (Rush 2003), Dick- cissel {Spiza americana) (Berkeley et al. 2007, Suedkamp Wells et al. 2008), and East- ern Meadowlark (Sturnella magna) (Kershner et al. 2004, Suedkamp Wells et al. 2008). Studies of movements during the post- * Division of Biological Sciences, 110 Tucker Hall, University of Missouri, Columbia, MO 65211, USA. 2 Current address: Puerto Rico Conservation Foun- dation, c/o International Institute of Tropical Forestry, USDA, Forest Service, Sabana Field Research Station, HC 02 Box 6205, Luquillo, Puerto Rico 00773, USA. 2 Corresponding author; e-mail: jend.white@gmail.com fledging dependent period, when juveniles are still with their parents, have shown that adults may have two types of movement patterns, one where adults stay on the breeding territory during the post-fledging period (stationary movements) and another where adults move from the breeding territory (drifting move- ments) (Anders et al. 1998, Vega Rivera et al. 2000). Vega Rivera et al. (2000) considered these movements (remaining on or moving from the breeding territory) to be different adult strategies, and the strategy used depend- ed on whether the adult had completed breed- ing activities and on where it wanted to molt. Adults also may split broods during the de- pendent period with males and females each caring for different fledglings (Green and Cockburn 2001, Yackel Adams et al. 2001), and the distances that dependent fledglings move may vary depending on whether the at- tending parent is male or female (Rush 2003). Movements during the dependent period also may vary depending on whether the adult is in a fragmented or contiguous forest (Rush 2003), and males with young may not move as far or be as likely to cross gaps as males without young (Bayne and Hobson 2001). There is little information on post-fledging movements of juvenile passerines after inde- pendence from their parents. However, studies of independent Wood Thrush juveniles dem- onstrated that juveniles used vegetation types 62 White and Faaborg • POST-FLEDGING MOVEMENTS OF SWAINSON’S THRUSH 63 different from those used by breeding adults (Anders et al. 1998, Vega Rivera et al. 1998, Fink 2003), made exploratory movements (Vega Rivera et al. 1998), moved in a south- erly direction (Anders et al. 1998), congre- gated in conspecific and mixed-species flocks (Vega Rivera et al. 1998, Fink 2003), re- mained solitary (Anders et al. 1998, Lang et al. 2002), and had from one to as many as five distinct post-fledging dispersal areas (Anders et al. 1998, Vega Rivera et al. 1998, Lang et al. 2002, Fink 2003). However, movement and spatial use patterns described for Wood Thrush may not be relevant to other species, vegetation types, landscapes, or regions. We describe post-fledging movements and dispersal for Swainson’s Thrushes (Catharus ustulatus) in a western habitat mosaic of ri- parian forest, coastal scrub, mixed-hardwood forest, and grazed and ungrazed annual grass- lands. The study subspecies (C. u. oedicus) is endemic to California; this subspecies nests in shrubs and is unique among Swainson’s Thrushes in that it breeds primarily in riparian forests rather than coniferous forests. Swain- son’s Thrushes are insectivore/frugivores (Evans Mack and Wang 2000) and foraging patches with abundant fruit were shown to be important to juvenile resource selection (White et al. 2005). The objectives of our study were to: (1) describe movements of adults with dependent broods, (2) describe dispersal of juveniles after they become in- dependent from their parents, and (3) estimate juvenile home range size during the post- fledging period. METHODS Study Area. — We studied Swainson’s Thrushes in riparian forests along three creeks in Marin County, California: Redwood Creek (37° 51 ' N, 122° 34' W, Muir Woods National Monument), Lagunitas Creek (38° 02' N, 122° 45' W, Golden Gate National Recreation Area), and Muddy Hollow (38°02'N, 122° 52' W, Point Reyes National Seashore). Ri- parian vegetation was typical of central coast- al California and was primarily red alder {Al- tius rubra) and arroyo willow {Salix lasiole- pis). The adjacent vegetation types at each creek were coastal scrub (Muddy Hollow and Redwood Creek), mixed-hardwood forest (La- gunitas and Redwood creeks). Bishop pine {Pinus muricata) forest (Muddy Hollow), and grazed (Lagunitas Creek) and ungrazed (Mud- dy Hollow and Redwood Creek) annual grass- es. Field Methods. — We used radio telemetry to study post-fledging movements of juvenile Swainson’s Thrush from 2000 to 2002. We searched for and monitored nests following standardized protocols (Martin and Geupel 1993) on pre-existing nest plots established by PRBO Conservation Science (PRBO). There were two nest plots along each creek. Nest plots at Redwood Creek were 1,400 m apart and plots A and B were 4.9 and 4.2 ha, re- spectively. Nest plots at Lagunitas Creek were 640 m apart and plots A and B were —7.4 and 3.9 ha, respectively. Nest plots at Muddy Hol- low were 500 m apart and plots A and B were 8.3 and 3.1 ha, respectively. We did not sur- vey Muddy Hollow for nesting activity due to time constraints after 2000. Each nest plot at Redwood and Lagunitas creeks had an aver- age of 21 Swainson’s Thrush territories each year from 2000 to 2002 (J. D. White, unpubl. data). We banded nestlings on days 9 and 10 of the nestling period; each nestling received a USGS aluminum band and a unique com- bination of three color-bands. At least one (rarely up to 3) nestling per brood was fitted with a radio transmitter using the leg-harness method (Rappole and Tipton 1991). We used a harness made of cotton cov- ered elastic string with adjustable leg loops that were attached to the transmitter prior to placement on nestlings in the field. This al- lowed for quick adjustment of harness length and attachment in the field. We used 1.4-g transmitters during 2000 with an average life of 60 days (Advanced Telemetry Systems [ATS], Itasca, MN, USA). We used refur- bished ATS transmitters and new 1.1 -g trans- mitters during 2001 with an average life of 50 days (Holohil Systems Ltd. Ottawa, ON, Can- ada). Transmitters weighed 4% (2000) and 3% (200 1-2002) of average adult thrush body mass (30 g). We used hand-held receivers (model R-IOOO) and Yagi three-element direc- tional antennas (model RA-150, Communi- cations Specialists, Inc. Orange, CA, USA) to locate radio-marked individuals. We located juveniles once every 2 to 4 days after fledging until independence. We checked juveniles after independence every other day. 64 THE WILSON JOURNAL OF ORNITHOLOGY • Vol 120, No. 1, March 2008 homing in on the animal to estimate its loca- tion (White and Garrott 1990). We attempted to visually relocate juveniles; when we were unable to observe a juvenile we circled the individual within a 10-20 m radius. We were able to correctly identify the vegetation type the juvenile was using by standing at the edge of two vegetation types and plotting juvenile location. We marked two GPS locations per day, for a total of 20-30 locations per bird, using the Universal Transverse Mercator (UTM) grid system. We used a handheld Gar- min GPS unit (GPS II Plus) and only recorded points with a low estimate of error (Figure of Merit [FOM] <10). We alternated checks of individuals between morning and afternoon on subsequent visits to ensure that each bird was monitored during different time periods throughout the day. Twice-daily locations were separated by 2 hrs to ensure the juvenile had sufficient time to move if it desired (Pas- inelli et al. 2001) and to guard against auto- correlation of points (Otis and White 1999). We monitored juveniles for 15 to 30 min and recorded behavior, food items taken, and pres- ence of conspecifics or other avian species. We used this information to ascertain if ju- veniles were with parents and siblings, and whether they were foraging independently. Juveniles that left the natal area without re- turning and were no longer observed with par- ents or siblings were considered biologically independent. Swainson’s Thrush juveniles af- ter biological independence from siblings and parents, are expected to be statistically and spatially independent and to disperse indepen- dently from each other as was reported for Wood Thrush (Anders et al. 1998, Vega Ri- vera et al. 1998). Siblings were considered spatially independent as long as their locations did not track each other in space or time (Er- ickson et al. 2001). We plotted locations of biologically independent siblings in ArcView’ GIS (ESRI 2000) to visually examine if sib- ling locations tracked each other; in no in- stance did siblings or non-related individuals track each other in space and time. Analyses. — We projected juvenile locations onto aerial photographs in ArcView to study movements of adults with dependent broods (movements from when juveniles leave the nest until they become biologically indepen- dent from their parents). We considered any location where a juvenile was attended by an adult to be a “dependent-period” location and. therefore, a brood-rearing location. We calculated and plotted the center of all brood- rearing locations, and measured the distance from the nest to the center of the brood-rear- ing area, which is defined as the area where adults raise their young; these areas can be close to or far from the nest. We measured nest to brood-rearing area in one case sepa- rately for two juveniles from the same brood because they w^ere 17 m apart. Broods reared within the natal area (<50 m of the nest) were considered stationary and broods reared out- side the natal area (>50 m from the nest) were considered drifting. We chose 50 m to sepa- rate brood types because stationary brood types were within 50 m of the nest. We want- ed to quantify differences between the two brood rearing strategies (stationary and drift- ing) and examine whether differences in de- pendent period movements resulted in differ- ences in independent period movements. We used non-parametric one-way ANOVA (PROG NPARIWAY) to examine whether mean distance moved from nest to brood-rear- ing area, and mean initial and subsequent dis- persal distances differed by stationary or drift- ing natal movements (SAS Institute 1999). We separated nests into early versus late nests and used Fisher’s exact test to examine if there were differences by brood type (early vs. late fledging) (SAS Institute 1999). We considered nests initiated on or before 15 June early nests and those initiated after this date late nests. This date (15 Jun) was the median date of egg laying for Swainson’s Thrushes during the study and first fledglings of the year fledge during early June (J. D. White, unpubl. data). We measured mobility (m/day) for each in- dividual by estimating daily movements from calculations of the distance between subse- quent locations (Batschelet 1981) and divided that distance by the number of days between relocations (Forsman et al. 2002). We calcu- lated median mobility by age group, and by drifting and stationary brood movements. Post-fledging dispersal is the movement of an independent juvenile from the natal area, or from parents and siblings in a brood-rearing area (Anders et al. 1998). Post-fledging dis- persal differs from natal dispersal, which is the movement of a bird from where it hatched White and Faaborg • POST-FLEDGING MOVEMENTS OF SWAINSON’S THRUSH 65 to its first breeding site (Greenwood 1980). We considered post-fledging dispersal to have started either when juveniles left the natal area on their own without returning, or when birds dispersed from a brood-rearing area and were no longer associating with their parents or sib- lings. We did not use a minimum distance moved as a criterion to define the onset of post-fledging dispersal because a juvenile could move a short distance and be outside the natal area. We defined initial dispersal as the distance from the center of the brood-rearing area to the center of the first post-dispersal area. We defined subsequent dispersal as the distance between the centers of different post-dispersal areas (e.g., first, second, third, etc.). We con- sidered a post-dispersal area distinct if a bird stayed in the area for >3 days. Juveniles may have more than one area of concentrated use within one post-dispersal area; we refer to these as multiple centers of activity. We mea- sured initial and subsequent dispersal distanc- es and the distance between multiple centers of activity using the distance measurement tool in the Animal Movements extension in Arc View (Hooge and Eichenlaub 1997). We calculated direction traveled from the center of the brood-rearing area to the first post-fledging dispersal area (initial dispersal direction) and to the final location (final dis- persal direction) for each bird (Batschelet 1981, Kernohan et al. 2001). We used the Rayleigh test to examine if the initial or final dispersal directions were random or oriented (Batschelet 1981). We tested for orientation at Redwood Creek (by plot), at Lagunitas Creek (both plots combined, due to smaller sample size), and at both creeks combined. We quantified two other types of move- ments, exploratory and wandering move- ments. An exploratory movement was defined as a movement greater than 300 m from an established dispersal area and subsequent re- turn to that area. Wandering movements were defined as locations where a bird stayed for <3 days in between established dispersal ar- eas. We calculated home ranges for individual Swainson’s Thrush juveniles with al least 25 locations (Seaman et al. 1999). A home range is defined as the extent of area with a defined probability of occurrence of an animal during a specified time period (Kernohan et al. 2001). We considered all post-fledging juvenile lo- cations (dependent and independent) to be a part of their post-fledging home range. Kernel methods are recommended over oth- er methods for home range estimation with small sample sizes (Kernohan et al. 2001). We used program KERNELHR (Seaman et al. 1998) to estimate home range areas; this pro- gram calculates volume percentages which es- timate the true use distribution (UD) rather than point percentages which only give esti- mates of UD. KERNELHR also calculates smoothing parameters separately for the x and y dimensions, allowing the kernel to better fit linear geographic features (e.g., rivers and mountain canyons). We used the fixed-kernel with the Least Squares Cross Validation smoothing parame- ter (h-LSCV) to estimate home range (95% fixed-kernel) and areas of concentrated use or core areas (50% fixed-kernel). While there are limitations to h-LSCV smoothing (Home and Garton 2006) it is generally recommended over reference smoothing options available in software programs (Kernohan et al. 2001). The smoothing parameter failed (h-LSCV went to zero) for eight juveniles due to tightly clustered locations. We moved clustered or duplicate locations slightly (1 or 2 m) for five of these individuals which corrected smooth- ing parameter failures. We deleted all but one or two of the clustered locations to correct smoothing parameter failures for another three individuals, and used KERNELHR with the reduced data set to obtain smoothing param- eter values (h,x) and h,y)), and then used KER- NELHR with the full data set using these val- ues. Results are reported as means ± SE un- less otherwise indicated. RESULTS We color banded 178 Swainson's Thrush nestlings from 62 nests. Sixty-seven nestlings were radiomarked from 61 nests (18 nestlings in 2()0(), 32 in 2001, and 17 in 2002) from which 35 juveniles reached independence from 30 different nests. We followed 10. 12, and 13 juveniles to independence in 2000, 2001, and 2002, respectively. Immediately after fledging, juveniles used thicket areas near the nest. Thickets could be tall shrubs (e.g., arroyo willow or red elder- 66 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 TABLE 1 Distances (m) moved for Swainson’s Thrush broods with stationary and drifting dependent period movements and for independent juveniles, Marin County, California, 2000—2002. Movement type n Mean SE Median Range Stationary, nest to brood-rearing area 22 28 2 30 9-49 Drifting, nest to brood-rearing area 9 189 102 70 25-994 Initial dispersal distanee 34 242 41 147 28-1,040 Subsequent dispersal distance Exploratory movement 28 11 598 439 101 31 418 422 154-2,624 289-651 berry [Sambucus racemosa]) or trees with low lying branches (e.g., red alder or boxelder [Acer negundo]) that reached into herbaceous cover, where fledglings remained close to the ground. Brood-rearing areas were in riparian, mixed-hardwood forest, and coastal scrub vegetation types. Five juveniles (14%) used only riparian vegetation after post-fledging dispersal, three (9%) used only upland vegetation types (coastal scrub or mixed-hardwood forest), and 27 (77%) used both riparian and adjacent up- land vegetation types. Brood Movements. — Swainson’s Thrush adults had both stationary (n = 21 broods) and drifting movements {n — 9 broods) during the post-fledging dependent period (Table 1, Fig. 1). The proportion of drifting broods was the same among early (30%) and late (27%) nests {P = 1.0, two-tailed Fisher’s exact test). Mobility was low for the first few days post- fledging for both movement types (Table 2). Mobility increased for drifting broods during days 3 to 7 post-fledging and again during days 8 to 12 post-fledging. Drifting adults led fledglings to brood-rearing areas that were on average 150 m farther from the nest than sta- tionary broods (F = 6.31, df = 1, P < 0.02; Table 1). Although drifting adults led young birds hundreds of meters from the natal area, the maximum daily distance moved was about 40 m/day (Table 2, Fig. 1). Median mobility was 3 to 12 m/day greater for drifting broods than stationary broods during the first 12 days after fledging (Table 2). Mobility increased for stationary juveniles during days 8 to 12 and 13 to 19 after fledging when these juveniles began dispersing independently. Brood-rearing areas used by drifting adults in 8 of 9 cases, consisted of fruiting shrub thickets with abundant fruit (twinberry [Lon- icera involucrata], Himalayan blackberry [Rubus discolor], and dogwood [Cornus ser- cia\y, in the other case, there were abundant oak moths {Phryganidia californica) in the brood-rearing area. Adults led juveniles on three occasions to areas identified as “popu- lation-level core areas”. Population-level core areas were areas that contained abundant fruit and were used by more than one radio-marked thrush each year, and were used during each year of the study. Adults on one occasion also led a juvenile across an old field gap of 78 m between the riparian natal area to coastal scrub foraging area; the fledgling had been out of the nest for 10 days. Juveniles from drifting broods spent an av- erage of 8.6 fewer days in the natal area than stationary broods (Fig. 2A); however, they stayed with parents and siblings 5 days longer (Fig. 2B, C) on average. Two siblings in one case dispersed from the natal area at different ages, one at 15 and the other at 22 days post- fledging. Spatial Use Patterns and Home Range Es- timates.—AW Swainson’s Thrushes after in- dependence established at least one post- fledging dispersal area {n = 29), 18 individ- uals (62%) established a second, three indi- viduals (10%) established a third, and one established a fourth. Therefore, the home range for most juveniles was comprised of disjunct dispersal areas. The median home range area (95% fixed-kernel) was 2.0 ha (range = 0. 2-9.0, mean [± SE] = 2.58 ± 0.52 ha) and the median core area (50% fixed-ker- nel) was 0.41 ha (range = 0.05-1.6, mean = 0.49 ± 0.09 ha). The mean number of loca- tions used to estimate individual home ranges was 33 (range = 25-47). Time spent in different post-fledging dis- persal areas ranged from 3 to 39 days. Juve- niles generally established a primary dispersal area where they spent most of their time; this White and Faaborg • POST-FLEDGING MOVEMENTS OF SWAINSON’S THRUSH 67 □ Nest o Brood locations • Independent locations A Transmitter location Movements juvenile 1 — Movements juvenile 2 ■■ Red AlderAA/illow riparian Mixed-hardwood forest Coastal scrub [ I Grazed annual grassland [771 Ranch stmctures \ \ \ • > \ \ 7 Sep 200 Meters A FIG. 1. Nest site, brood, and independent juvenile locations of two juvenile Swainson’s Thrushes at Lagun- itas Creek, Marin County, California, 2000. The juvenile on the left had stationary brood movements; the juvenile on the right had drifting brood movements. could be either the first or subsequent dis- persal areas. The mean (± SE) number of days spent in the primary dispersal area was 22.7 ± 1 .5 days {n = 29, median = 24, range = 6-39) and the mean number of days in the secondary dispersal area was 9.6 ± 0.9 days {n = 18, median = 10, range = 3-15). Two juveniles had two centers of activity within one dispersal area, and both juveniles traveled between these centers daily (Fig. 3). The dis- tance between centers was 231 m and 131 m. Mobility. — Mobility ranged from a low near zero for all age groups to highs in the hun- dreds of meters for age groups past 8 days post-lledging (Table 2). Mobility varied great- ly because, within a dispersal area, birds gen- erally did not move far, often remaining in the same shrub thicket; however, they made long- distance movements between dispersal areas. Distance and Direction. — Dispersal dis- tances of independent juveniles did not differ between stationary or drifting natal move- 68 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 TABLE 2. Median mobility (m/day) for Swainson’s Thrush juveniles with stationary (n = 25) and drifting {n = 12) natal movements, and for all independent juveniles pooled after post-fledging day 20, Marin County, California, 2000-2002. Sample size (n) refers to the number of movements by juveniles in each age group. Days post-fledging Stationary Drifting Mobility (n) Range Range Mobility (n) 0-2 4.5 (30) 0.0-24.2 7.9 (20) 0.5-40.0 3-7 5.9 (38) 0.5-101.5 14.9 (22) 0.0-38.9 8-12 14.1 (59) 1.0-756.7 25.7 (23) 0.5-206.9 13-19 28.9 (114) 0.0-211.8 22.6 (48) 1.0-218.9 20-29 25.5 (236) 0.5-880.3 30-39 37.0 (178) 0.3-724.6 40-49 39.6 (124) 0.0-475.9 >50 103.9 (38) 2.1-393.1 merits (initial dispersal, F = 0.03, df = \, P = 0.86; subsequent dispersal, F = 0.08, df = P = 0.78) so these data were pooled. Most initial dispersal distances were <300 m from the natal area and most subsequent dispersal distances were >200 m (Fig. 4). The average initial dispersal distance was 242 m and av- erage subsequent dispersal distances were 598 m (Table 1). The greatest distance a Swain- son’s Thrush juvenile traveled between dis- persal areas was 2,624 m. Initial post-hedging dispersal direction was not oriented signihcantly at Redwood Creek (plots A and B), Lagunitas Creek, or both creeks combined (Table 3). However, initial dispersal at Redwood Creek plot B tended to- ward a southwest direction {P = 0.05, Table 3); the angular deviation (a measure of dis- persion) was relatively small {s = 54°) and the vector length (a measure of concentration around the mean direction, where 0 is not di- rected and 1 is perfectly directed) was rela- tively large (r = 0.55, Table 3). Final dispersal was oriented to the west for Redwood Creek plot B and northwest for both creeks com- bined (Table 3). However, hnal dispersal di- rection for both creeks combined may have been an artifact of the strong orientation at Redwood Creek plot B or due to a large sam- ple size. The relatively large angular deviation (s = 64°) and small vector length (r = 0.37) for hnal dispersal for both creeks indicate that with a smaller sample, this result would not be signihcant (Table 3). Wandering and Exploratory Movements. — Eleven individuals had wandering locations of 1-2 days duration. The majority of juveniles wandered after dispersal from the natal area and prior to settling into a hrst post-hedging dispersal area (64%, n = 7). The remaining juveniles wandered before settling into a sec- ond post-hedging dispersal area (36%, n = 4). Nine juveniles made exploratory move- ments (e.g.. Fig. 4) and one individual made three different exploratory movements. The average distance of exploratory movements was 439 ± 31 m (Table 1). The juvenile in two cases (18%) subsequently returned to the exploratory site and remained there for a lon- ger period of time. Two additional juveniles made movements that were functionally ex- ploratory but were less than 300 m; one of these returned to the exploratory site for its second dispersal area. DISCUSSION Juvenile Swainson’s Thrushes were gener- ally near the nest and took cover in thicket areas within a day or two post-hedging. Mo- bility was low for the hrst few days after hedging, as has been found for recent hedg- lings in other studies (Vega Rivera et al. 2000, Yackel Adams et al. 2001). Swainson’s Thrush adults with dependent broods had both stationary and drifting modes of movement similar to those described for Wood Thrush (Anders et al. 1998, Vega Ri- vera et al. 2000). The type of movement cho- sen may have depended on whether the adults attempted a second brood or whether they had completed breeding activities and moved to a foraging or molting site (Vega Rivera et al. 2000). Swainson’s Thrush adults that made drifting movements led hedglings to areas with abundant fruit; these adults were not lim- ited to riparian vegetation and were able to Number of days post-fledging White and Faaborg • POST-FLEDGING MOVEMENTS OF SWAINSON’S THRUSH 69 16 -1 A. Days in natal area 14 - 12 - 10 - 1 T 8 - 6 - 4 - 2 - 0 -- 25 1 B. Age (in days) last observed with parent 20 - 15 - 5 - 0 C. Age (in days) first observed alone 25 20 15 10 - 5 - 0 -I ^ ' Stationary Drifting FIG. 2. (A) Mean number of days juvenile Swain- son’s Thrushes spent on the natal territory or with par- ents and siblings, (B) age (days post-fledging) last ob- served with parents, and (C) age at first observation of independence by dependent movement type, stationary (n = 22) or drifting (n = 9), Marin County, California, 2()()() to 2002. Mean ± SE and 95% confidence inter- vals are shown for each box plot. cross old field gaps to foraging sites. Broods did not appear to be as restricted to forest sites as has been described for some birds in forest fragments (Bayne and Hobson 2001); how- ever, this may not hold true in isolated riparian areas. Adults were site faithful not only to terri- tories and nest sites but also to brood-rearing sites, leading fledglings to the same areas in subsequent years. Therefore, brood-rearing ar- eas may have an important role in territory or nest-site selection. Swainson’s Thrush fledglings from drifting broods became independent at 20 days post- fledging on average. This age was similar to the average age at dispersal and presumably independence for Wood Thrush (Anders et al. 1998, Vega Rivera et al. 1998, Lang et al. 2002, Fink 2003). However, the average age at independence for Swainson’s Thrush fledg- lings from stationary broods was 15 days post-fledging. The average initial (242 m) and subsequent (598 m) post-fledging dispersal distances for Swainson’s Thrushes were less than average dispersal distances reported for Wood Thrush (2.1 km, 1.5 km, 2.2 km, and 825 m, respec- tively) (Anders et al. 1998, Vega Rivera et al. 1998, Lang et al. 2002, Fink 2003). Juvenile Swainson’s Thrush may not have had to dis- perse great distances because of abundant nearby resources (White et al. 2005). The ri- parian areas in which Swainson’s Thrushes bred were narrow and bordered by coastal scrub, which had areas of abundant fruit. Ri- parian forests also were bordered by mixed- hardwood forest that had oak moth caterpillar outbreaks during late summer and could also have patches of abundant fruit (White et al. 2005). Juvenile Wood Thrushes from frag- mented forests in northern Missouri (Fink 2003) also did not disperse as far as Juvenile Wood Thrushes from more contiguous forests (Anders et al. 1998, Vega Rivera et al. 1998, Lang et al. 2002). Fink (2003) attributed this to nearby second growth or edge in forest fragments, which offered either abundant cov- er or food; another explanation was that ju- veniles might be constrained to forest frag- ments by the surrounding agricultural matrix. However, similar to Swainson’s Thrush broods, independent Juvenile Swainson’s Thrushes were not restricted to riparian forests and readily used adjacent coastal scrub and mixed-hardwood forest. Southerly movements after fledging may in- dicate migratory movements (Anders et al. 70 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 □ Nest o Brood locations • Independent locations Movement Red Alder/Willow riparian MW Mixed-hard\A/ood forest Coastal scrub I I Annual grassland 24Jui 100 Meters FIG. 3. Nest site, brood, and independent juvenile locations for a juvenile Swainson’s Thrush at Redwood Creek, Marin County, California, 2002. This juvenile had two dispersal sites, one of which had two centers of activity. An exploratory movement is indicated by the date, 24 July. 1998); however, three studies of post-fledging Wood Thrushes did not find evidence for southerly movements (Vega Rivera et al. 1998, Lang et al. 2002, Fink 2003). Swain- son’s Thrush post-fledging dispersal was not generally oriented; however, there was clear evidence for westerly orientation at one plot along Redwood Creek. Therefore, migration likely was not driving movements of Swain- son’s Thrush juveniles during the post-fledg- ing period (10 Jun-10 Sep). Rather, optimal foraging likely was a more important factor affecting post-fledging independent move- ments, including the westerly movements at Redwood Creek where birds moved into coastal scrub and areas with abundant ripe fruits (White et al. 2005). Some Swainson’s Thrush juveniles wan- dered prior to establishing a post-fledging dis- persal area, while others also made wandering movements between different dispersal areas. Wandering movements prior to settling and White and Faaborg • POST-FLEDGING MOVEMENTS OE SWAINSON’S THRUSH 71 (D O C CD -4— < -O o CD CD _c c g ■■c o Q. O i_ a. Distance (km) EIG. 4. Proportion of Swainson’s Thrush juveniles dispersing by distance category from natal to first post- fledging dispersal area (n = 34), and between subsequent post-fledging dispersal areas (n = 26), Marin County, California, 2000-2002. between dispersal sites have also been de- scribed for Wood Thmsh (Anders et al. 1998, Vega Rivera et al. 1998, Lang et al. 2002, Fink 2003). Juvenile Swainson’s Thrushes also made exploratory movements from their dispersal areas; however, few exploratory movements resulted in resettlement at the ex- ploratory site. Exploratory movements may be important in gaining a sense of location and in searching for new foraging patches (Baker 1982). Swainson’s Thrush home-range estimates (2 ha) were similar to those of juvenile Wood Thrush (average = 2.8 ± 0.5 ha for one dis- persal site [Vega Rivera et al. 1998] and 1.5 ± 1.6 ha for first dispersal sites [Anders et al. TABLE 3. Initial and final post-fledging dispersal direction of Swainson’s Thrush juveniles at Lagunitas Creek and at two sites at Redwood Creek, Marin County, California, 2000-2002. Number of juveniles followed at each site (n), mean bearing from true north (/?), mean mean vector length (r), and Rayleigh test P-values. angle (a). and mean angular deviation (s) in degrees. n h a S p Initial dispersal Lagunitas Creek 1 1 75 15 74 0.16 0.763 Redwood Creek, A 10 330 120 63 0.39 0.241 Redwood Creek, B 10 230 220 54 0.55 0.050 Creeks combined 33 273 177 74 0.17 0.468 Final dispersal Lagunitas Creek 1 I 307 143 6S 0.30 0.380 Redwood Creek, A 10 324 126 61 0.43 0.173 Redwood Creek, B 10 275 175 51 0.61 0.023 Creeks combined 33 297 153 64 0.37 0.019 72 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 1998]). Juvenile home ranges were comprised of disjunct dispersal areas and multiple cen- ters of activity, likely due to spatially and tem- porally clumped resources. The availability of fruit varied in both space and time, depending on fruiting phenology and ecology of different fruiting-shrub species (J. D. White, unpubl. data). Oak moth caterpillars upon which thrushes fed also were patchily distributed in space and time. These caterpillars reached outbreak proportions in mid August and Sep- tember, and specialized on coast live oak (Quercus agrifolia) and California bay {Um- bellularia californica) trees primarily in mixed-hardwood forest. Home-range esti- mates were relatively small and used sites of- ten were far apart. For conservation purposes managers may need to preserve large areas that encompass widely dispersed patches of appropriate foraging sites to ensure that re- sources are available for the post-fledging pe- riod. Swainson’s Thrush spatial use patterns were similar to those described for Wood Thrush, despite these species occupying different veg- etation types. This is likely due to the similar ecology of these closely related species. Both are insectivore/frugivores that use fruiting shrub thickets during the post-fledging period. Wood Thrushes breed in mature eastern de- ciduous forest but juveniles sought areas of second growth during the post-fledging peri- od, likely due to abundant fruit resources and cover in these areas (Anders et al. 1998, Vega Rivera et al. 1998, Fink 2003). Home ranges comprised of disjunct areas have also been de- scribed for Common Ravens {Corvus corax\ Roth et al. 2004) and Eurasian Jay {Garrulus glandarius; Rolando 1998) that use concen- trated and patchily distributed resources. Only one Swainson’s Thrush had nest-centered movements as described for Dickcissels (Suedkamp Wells et al. 2008) and Burrowing Owls {Athene cunicularia; Davies and Restani 2006). Research is needed on post-fledging move- ments and spatial habitat-use patterns in other species, particularly species that presumably rely on a more evenly distributed resource base, to learn if patterns described for Swain- son’s Thrush and Wood Thrush are general. It will also be important to study these patterns in other vegetation and landscape types to learn how juveniles use space, and the ease with which they can disperse from natal areas and find food and cover during this critical period in the annual cycle. ACKNOWLEDGMENTS The Avian Ecology Group at the University of Mis- souri and two anonymous reviewers made helpful comments on this manuscript. We thank the Golden Gate National Recreation Area (especially Daphne Hatch) and the Point Reyes National Seashore (espe- cially Sarah Allen) for ongoing support of bird moni- toring and research. We are grateful for the general support of PRBO, especially Geoffrey Geupel, Nils Warnock, and Thomas Gardali. We also thank PRBO’s Terrestrial Ecology Division field crew, especially Roy Churchwell, Jill Harley, Chris Rintoul, and Viola Ton- iolo for help in the field. This project was funded by a GAANN fellowship. University of Missouri-Colum- bia. Canon National Parks Science Scholars Program, the National Park Service, National Park Service Fee Demonstration Project, and the Marin Audubon Soci- ety. LITERATURE CITED Anders, A. D., J. Faaborg, and E R. Thompson III. 1998. Postfledging dispersal, habitat use, and home-range size of juvenile Wood Thrushes. Auk 115:349-358. Baker, R. R. 1982. Migration: paths through time and space. Holmes and Meier Publishers, Inc., New York, USA. Batschelet, E. 1981. Circular statistics in biology. Academic Press, New York, USA. Bayne, E. M. and K. A. Hobson. 2001. Movement patterns of adult male Ovenbirds during the post- hedging period in fragmented and forested boreal landscapes. Condor 103:343-351. Berkeley, L. L, J. P. McCarty, and L. L. Wolfen- BARGER. 2007. Postfledging survival and move- ment in Dickcisssels {Spiza americana): implica- tions for habitat management and conservation. Auk 124:396-409. Cohen, E. B. and C. A. Lindell. 2004. Survival, hab- itat use and movements of fledgling White-throat- ed Robins (Turdus assimilis) in a Costa Rican ag- ricultural landscape. Auk 121:404-414. Davies, J. M. and M. Restani. 2006. Survival and movements of juvenile Burrowing Owls during the postfledging period. Condor 108:282—291. Erickson, W. P, T. L. McDonald, K. G. Gerow, S. Howlin, and j. W. Kern. 2001. Statistical issues in resource selection studies with radio-marked animals. Pages 209-242 in Radio tracking and an- imal populations (J. J. Millspaugh and J. M. Mar- zluff. Editors). Academic Press, New York, USA. ESRI. 2000. ArcView GIS. Version 3.2a. Environmen- tal Systems Research Institute, Inc., Redlands, California, USA. White and Faaborg • POST-FLEDGING MOVEMENTS OF SWAINSON’S THRUSH 73 Evans Mack, D. and Y. Wang. 2000. Swainson’s Thrush (Catharus ustulatus). The birds of North America. Number 540. Fink, M. 2003. Post-fledging ecology of juvenile Wood Thrush in fragmented and contiguous land- scapes. Dissertation. University of Missouri, Co- lumbia, USA. Forsman, E. D., R. G. Anthony, J. A. Reid, P. J. Loschl, S. G. Sovern, M. Taylor, B. L. Biswell, A. Ellingson, E. C. Meslow, G. S. Miller, K. A. Swindle, J. A. Thrailkill, F. F. Wagner, and D. E. Seaman. 2002. Natal and breeding dispersal of Northern Spotted Owls. Wildlife Monographs 149. Green, D. J. and A. Cockburn. 2001. Post-fledging care, philopatry and recruitment in Brown Thorn- bills. Journal of Animal Ecology 70:505-514. Greenwood, P. J. 1980. Mating systems, philopatry and dispersal in birds and mammals. Journal of Animal Behavior 28:1140-1162. Hooge, P. N. and B. Eichenlaub. 1997. The animal movement extension to arcview. Version 1.1. USGS, Alaska Science Center — Biological Sci- ence Office, Anchorage, USA. Horne, J. S. and E. O. Garton. 2006. Liklihood cross- validation versus least squares cross-validation for choosing the smoothing parameter in kernel home-range analysis. Journal of Wildlife Manage- ment 70:641-648. Kernohan, B. j., R. a. Gitzen, and J. J. Millspaugh. 2001. Analysis of animal space use and move- ments. Pages 125-166 in Radio tracking and an- imal populations (J. J. Millspaugh and J. M. Mar- zluff. Editors). Academic Press, New York, USA. Kershner, E. L., j. W. Walk, and R. E. Warner. 2004. Postfledging movements and survival of ju- venile Eastern Meadowlarks (Sturnella magna) in Illinois. Auk 121:1146-1154. Lang, J. D., L. A. Powell, D. G. Krementz, and M. J. Conroy. 2002. Wood Thrush movements and habitat use: effects of forest management for Red- cockaded Woodpeckers. Auk 119:109-124. Martin, T. E. and G. R. Geupel. 1993. Nest-monitor- ing plots: methods for locating nests and moni- toring success. Journal of Field Ornithology 64: 507-519. Otis, D. L. and G. C. White. 1999. Autocorrelation of location estimates and the analysis of radi- otracking data. Journal of Wildlife Management 63:1039-1044. Pasinelli, G., j. Heglebach, and H.-U. Reyer. 2001. Spacing behavior of the Middle Spotted Wood- pecker in central Europe. Journal of Wildlife Man- agement 65:432-441. Rappole, j. H. and a. R. Tipton. 1991. New harness design for attachment of radio transmitters to small passerines. Journal of Field Ornithology 62: 335-337. Rolando, A. 1998. Factors affecting movements and home ranges in the Jay {Garrulus glandarius). Journal of Zoology 246:249-257. Roth, J. E., J. P. Kelly, W. J. Sydeman, and M. A. Colwell. 2004. Sex differences in space use of breeding Common Ravens in western Marin County, California. Condor 106:529-539. Rush, S. 2003. The effects of forest fragmentation on post-fledging survival and dispersal of a forest songbird. Thesis. York University, Toronto, On- tario, Canada. SAS Institute Inc. 1999. SAS/STAT user’s guide. Version 8. SAS Institute Inc., Cary, North Caro- lina, USA. Seaman, D. E., B. Griefith, and R. A. Powell. 1998. KERNELHR: a program for estimating animal home ranges. Wildlife Society Bulletin 26:95- 100. Seaman, D. E., J. J. Millspaugh, B. J. Kernohan, G. C. Brundige, K. j. Raedeke, and R. A. Gitzen. 1999. Effects of sample size on kernel home range estimates. Journal of Wildlife Management 63: 739-747. SuEDKAMP Wells, K. M., J. J. Millspaugh, M. R. Ryan, and M. W. Hubbard. 2008. Eactors affect- ing home range size and movements of post-fledg- ing grassland birds. Wilson Journal of Ornithology 120:120-130. Vega Rivera, J. H., C. A. Haas, J. H. Rappole, and W. J. McShea. 2000. Parental care of fledgling Wood Thrushes. Wilson Bulletin 1 12:233-237. Vega Rivera, J. H., J. H. Rappole, W. J. McShea, AND C. A. Haas. 1998. Wood Thrush postfledging movements and habitat use in northern Virginia. Condor 100:69-78. White, G. C. and R. A. Garrott. 1990. Analysis of wildlife radio-tracking data. Academic Press, New York, USA. White, J. D., T. Gardali, F. R. Thompson 111, and J. Faaborg. 2005. Resource selection by juvenile Swainson’s Thrushes during the post-fledging pe- riod. Condor 107:388-401. Yackel Adams, A. A., S. K. Skagen, and R. D. Ad- ams. 2001. Movements and survival of Lark Bun- ting fledglings. Condor 103:643-647. The Wilson Journal of Ornithology 1 20( 1 ):74-84, 2008 SPRING MIGRATORY STOPOVER OF SWAINSON’S THRUSH ALONG THE PACIFIC COAST OF SOUTHWEST COSTA RICA SCOTT WILSON, KEITH A. HOBSON,^ DOUGLAS M. COLLISTER,^ AND AMY G. WILSON' ABSTRACT. — Stopover behavior and migratory pathways of neotropical migrant birds in Central and South America have received little study. We examined stopover ecology of Swainson’s Thrush (Catharus ustulatus)^ on the Osa Peninsula, Costa Rica, during spring migration, 2002-2005. Capture rates per net hour were high (x = 21.8 individuals/ 100 net hrs) suggesting large numbers pass through lowland coastal areas in spring. Mean passage date of males was ~6 days earlier than females. Timing of passage by age class was variable; after- second year (ASY) birds preceded second-year (SY) birds in 2 years, SYs preceded ASYs in 1 year, and both had the same mean passage date in 1 year. We also observed annual variation in relative abundance of the two age groups. Only 1.7% of marked Swainson’s Thrushes were recaptured on subsequent days suggesting most individuals left the immediate area soon after initial capture. Stopover lengths ranged from 1 to 9 days, although most were <4 days. Many individuals had some energy stores on arrival (fat score ^ 1) but reserves varied between years and tended to be lower in SYs compared to ASYs. Regression of body mass against time of day indicated that individuals tended to gain mass throughout the morning (0.67% of lean body mass/hr on average). Our findings for stopover lengths, rates of mass gain, and recapture rates are within the range observed at North American stopover sites in spring. However, our capture rates were relatively higher, perhaps because the small land mass of southern Central America concentrates individuals at stopover sites in these regions. Received 26 September 2006. Accepted 31 March 2007. Annual movements between breeding and wintering areas allow neotropical migrant songbirds to benefit from the seasonal flush of food in temperate breeding areas, while en- hancing survival by wintering in the tropics (Greenberg 1980). However, long-distance migration is energetically expensive and re- serves required to traverse these distances may exceed the amount individuals can store (Alerstam and Lindstrom 1990, Schaub and Jenni 2000a). Thus, migrants need to stopover en route where they must forage quickly in unfamiliar habitats (Moore and Aborn 2000, Petit 2000) while coping with weather (Rap- pole and Warner 1976, Richardson 1978), predators (Lindstrom 1990), and other indi- viduals competing for the same resources (Moore and Wang 1991). Failure to meet these demands may be costly. Most annual mortal- ity for migrant songbirds may occur during migration even though these periods only rep- ' Center for Applied Conservation Research, 2424 Main Mall, University of British Columbia, Vancou- ver, BC, V6T 1Z4, Canada. ^ Environment Canada, 1 1 Innovation Boulevard, Saskatoon, SK, S7N 3H5, Canada. ^ Calgary Bird Banding Society, 247 Parkside Cres- cent SE, Calgary, AB, T2J 4J3, Canada. Corresponding author; e-mail: wilsonwa@interchange.ubc.ca resent —25% of the annual cycle (Sillett and Holmes 2002). An individual’s inability to re- fuel quickly at stopover sites can also lower fitness through delayed arrival in breeding ar- eas and a subsequent reduction in breeding success (Marra et al. 1998, Smith and Moore 2005). Nearly all studies of spring stopover of neo- tropical migrants have been conducted in North America, representing the latter half of the migration for most species. Little is known about migratory behavior and stopover ecol- ogy during the early stages of spring migra- tion in tropical areas of Central and South America. Central America has a small land- mass relative to North America and, for some species that breed across the boreal forest and winter in South America, the entire breeding population may pass through a narrow geo- graphic corridor during migration. Even minor disturbances to stopover sites in these areas could have implications for population dy- namics (Newton 2006). Deforestation and habitat disturbance have been extensive throughout much of Central America and are expected to continue (Achard et al. 2002). Thus, it is important to further our understand- ing of migratory routes and stopover areas used by neotropical migrants throughout sites in the tropics. 74 Wilson et al. • STOPOVER ECOLOGY OE SWAINSON’S THRUSH 75 We examined migratory timing and stop- over behavior of Swainson’s Thrush (Catha- rus ustulatus) at a site on the Osa Peninsula, Costa Rica. The Sw^ainson’s Thrush is a neo- tropical migrant that breeds across the boreal forest, western mountain regions, and Pacific coast of North America. Coastal breeding populations winter in southern Mexico and northern Central America, while continental breeding populations primarily winter in west- ern South America (Mack and Wang 2000, Ruegg and Smith 2002). Swainson’s Thrush is a common migrant in Pacific coast regions of southwestern Costa Rica in spring (Stiles and Skutch 1995), and we previously showed that individuals passing through these regions like- ly breed in areas of the west-central boreal forest and interior western mountains of North America (Wilson et al. 2008). Our specific ob- jectives in the present study were to examine: (1) capture rates over time, (2) differences in migratory timing between age and gender classes, and (3) stopover length, recapture rates, and rates of mass gain. METHODS Study Area. — The study was conducted on the northwest tip of the Osa Peninsula, Costa Rica, during spring migration 2002-2005. Our study area (08° 41' N, 83° 38' W) was 3 ha in size —10 km north of Corcovado National Park. Annual mean temperatures on the Osa Peninsula are 27° C and precipitation is 550 cm with dry and wet seasons occurring in De- cember-April and May-November, respec- tively (Holdridge 1967). Habitat in the region consists of tropical wet and premontane forest (Sanchez-Azofeifa et al. 2002). The immedi- ate study area was mostly secondary forest, 15-25 years of age, with larger stands of pri- mary forest along the eastern and southern edges. There were two small open areas (~30 by 30 m) within the 3-ha plot. This is typical of habitat structure along coastal areas outside Corcovado National Park, where settlement has led to fragmentation and patches of youn- ger-age forests. Much of the habitat adjacent to our site was secondary forest, particularly along the coast in both directions. The habitat beyond our site towards the interior of the peninsula was a mixture of primary and sec- ondary forest with some cleared areas for ag- riculture (wSanchez-Azofeifa et al. 2002). Com- mon tree species included Spondias monbin, Hyeronima alchorneoides, Elaeis oleifera, So- cratera exorrhiza, and Ochroma pyramidale; common understory shrubs, vines, and small trees included Apebia tiboubou, Heliconia spp., Vosychia spp.. Philodendron spp., and Piper spp. Capture Methods. — The site was operated from late March through April annually. We captured individuals using 12-15 mist nets (3-tier, 12-m length, 30-mm mesh) that were opened at dawn and operated for —6 hrs. Nets were evenly spaced throughout the study area and mostly within or on the edge of secondary forest habitats. A few were just inside primary forest at the edge of the study area. We made minor adjustments to the location of some net lanes between years because of slight changes in land use. These changes did not influence our survey effort and we do not believe they influenced our results. Individuals captured were banded with a numbered USGS alumi- num band and classified to age where possible as second year (SY) or after-second year (ASY) based on plumage, feather shape and wear, and wing morphology (Pyle 1997). We also recorded date and time of capture, wing chord, body mass, and fat score (Helms and Drury 1960). We measured all individuals re- captured on subsequent days but only record- ed time of capture for those recaptured on the same day. Plumage examination indicated that Swainson’s Thrushes captured were from con- tinental populations (olive-backed group) that breed across the boreal forest and interior western mountains (Mack and Wang 2000, Ruegg and Smith 2002). The Swainson’s Thrush is sexually monomorpic and individ- uals were genetically assigned to gender fol- lowing Griffiths et al. (1998) and Wilson et al. (2008). The Swainson’s Thrush is a rare win- ter resident in the area (Stiles and Skutch 1995) and a separate study failed to capture any individuals during three 8-day periods from early December to early March 2005- 2006 (D. M. Collister, unpubl. data). Thus, nearly all individuals captured were likely passage migrants. Statistical Analyses. — SwainsoiTs Thrushes primarily migrated through the site between late March and late April (Wilson et al. 2008). We used only 2004 and 2005 data for some analyses because data from 2002 and 2003 76 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 TABLE 1. Rica. Timing of study, effort, and capture data for Swainson’s Thrushes on the Osa Peninsula, Costa Year Dates Net hrs Individuals captured Individuals recaptured (subsequent days) 2002 17 Mar-12 Apr^ 2,041 439 2003 15 Apr- 10 May 2,123 404 10 (2.5%) 2004 28 Mar-29 Apr 2,468 546 9 (1.6%) 2005 30 Mar-29 Apr 2,213 540 7 (1.3%) ^ We did not band all Swainson’s Thrushes in 2002 and did not estimate recapture rates for that year. A small number of the 439 captures in 2002 were likely recaptures. were outside the late March-late April period. We used general linear models (GLM) to compare timing of migration by age and gen- der, and we present means with standard error for most measures. Energetic condition was compared using fat scores (Helms and Drury 1960) and residuals from a regression of mass against wing length. We used mass-wing re- siduals rather than deviation from the average mass because an individual’s size will affect mass, and wing length is a useful way to stan- dardize for size (Schulte-Hostedde et al. 2005). Examination of mass-wing length re- siduals indicated a linear relationship and no tendency for body size to influence energetic condition (Schulte-Hostedde et al. 2005). We also report fat scores for comparison with oth- er studies but did not conduct analyses using these scores because they are a subjective measure. We estimated stopover length by calculat- ing minimum stopover duration where the in- dividual’s date of first capture was subtracted from the last date recaptured (Cherry 1982). This method is conservative because it does not incorporate time the bird spent at the site before first capture or after last capture. We also used open population capture-recapture models (Schaub et al. 2001). This approach uses Cormack-Jolly-Seber models to estimate the likelihood of capture for an individual present at the site and recruitment analyses to estimate probability of arrival in the study area prior to capture. Parameter estimation us- ing this method is limited when recapture sample sizes are small, as was the case in our study. We used the multiple day constancy method of Morris et al. (2005) which holds parameters fixed across days reducing the number of parameters that need to be esti- mated without pooling data. We calculated change in mass of recaptured individuals by subtracting the mass at first capture from mass at final capture. This value was divided by the stopover duration to cal- culate change in mass per day. We also esti- mated rates of mass gain by regressing mass against time of capture for all individuals be- cause recaptured individuals may not be a ran- dom sample of the entire population (Winker et al. 1992, Dunn 2001). We included wing length in the model to control for effects of size on mass gain. The regression coefficient for time of capture represents the average change in mass per minute. We used the mass versus time of capture regression equation to estimate the expected rate of mass gain for an individual of average wing length over the 6-hr period since sunrise. Statistical analyses were conducted using the R Statistical Lan- guage and Environment package (R Devel- opment Core Team 2006). RESULTS Capture Rates and Migration Timing. — We captured 1,931 Swainson’s Thrushes over the 4-year study. These numbers equated to 19.1, 22.1, and 24.4 Swainson’s Thrushes/ 100 net hrs for each year, 2003-2005, respectively (Table 1). We did not band all individuals in 2002 and could not estimate recaptures. How- ever, based on recapture rates from the other 3 years, we likely captured ~21 individuals/ 100 net hrs in 2002. The lower numbers in 2002 and 2003 are likely because banding did not encompass the full migratory period of Swainson’s Thrushes. Data from 2002 and 2003 indicated most individuals passed through the site between late March and late April. In 2002, banding extended from 17 March to 12 April, but only 7 and 17% of individuals were caught prior to 28 March and Wilson et al. • STOPOVER ECOLOGY OF SWAINSON’S THRUSH 77 Date of passage FIG. 1. Timing of spring migration of male and female Swainson’s Thrushes through the Osa Peninsula, Costa Rica, in 2004 (males = 51, females = 48). Date of passage refers to Julian day with 1 Apr = 92. 30 March (start dates for 2004 and 2005), re- spectively. In 2003, banding was conducted from 15 April to 10 May but only 4% of in- dividuals were caught after 29 April (end date for 2004 and 2005). We may have missed some birds, especially early in migration but our data for 2004 (28 Mar-29 Apr) and 2005 (30 Mar-29 Apr) likely represent —80-90% of the individuals passing through the site (Fig. 1). Recapture rates during 2003-2005 were low with 10 (2.5%), 9 (1.65%), and 7 (1.3%) individuals caught on subsequent days in each year, respectively. There were no year to year recaptures. The mean passage date of males in 2004 was —6 days earlier than for females (males: 98 ± 1.0 (mean Julian day ± SE), n = 51; females: 104 ± 1.2, /i = 48). We also exam- ined passage date through the site with all known-age birds in 2004 and 2005. In 2004, 46% of birds captured were SY, 27% were ASY, and 27% were of unknown age. The mean passage date of ASYs (103 ± 0.6, n = 149) was —3 days earlier than SYs (106 ± 0.5, n = 249) in 2004. Passage dates were earliest for ASY males, intermediate for SY males and ASY females, and latest for SY fe- males for birds of known age and gender (Fig. 1). More individuals captured in 2005 were SY (84%) versus ASY (11%) but there was little difference in timing of passage between the two groups (ASY: 103 ± 0.9, n = 59; SY: 103 ± 0.4, n = 459). We only attempted to classify age of 167 of 439 individuals cap- tured in 2002. Of these, 49% were SYs, 31% were ASYs, and 20% could not be classified to age. SY’s within the shortened 2002 period passed through on average 1 day earlier than ASYs (SY: 92 ± 0.6, n = 82; ASY: 93 ± 0.9, n = 51). Of 404 individuals in 2003, 73% were SYs, 13% were ASYs, and 14% could not be classified to age. ASY’s preceded SYs by 2 days on average in 2003 (SY: 1 1 1 ± 0.3, n = 295; ASY: 109 ± 0.6, n = 54). Estimates should only be compared within years because the data do not cover the same periods each year. Stopover Length and Mass Gain. — Mean minimum stopover lengths ranged from 1 to 9 days with a mean of 3.2 days {n = 9) in 2004 and 2.3 days {n = 1) in 2005. The dis- tribution tended to be skewed as 14 of 16 in- dividuals had stopover lengths of 1-4 days while two stayed for 6 and 9 days. The me- dian minimum stopover length was 2 days in both years. Stopover lengths estimated by the multiple day constancy (MDC) method were 4.0 ± 3.9 days (mean ± SD) in 2004 and 1.2 ± 0.7 days in 2005. The minimum stopover duration and MDC capture-recapture methods produced similar results with the capture-re- capture method estimating slightly longer 78 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 TABLE 2. Regression coefficients for equation of Swainson’s Thrushes on the Osa Peninsula, Costa Rica, average change in mass (grams) per minute. mass against wing chord and time of capture for Standard errors are in parentheses. Time represents Class Intercept Wing chord Time All birds 1,042 1.096 (2.10) 0.278 (0.022) 0.003 (0.0007) ASY 201 0.924 (4.72) 0.283 (0.048) 0.0018 (0.0017) SY 675 -1.153 (2.73) 0.302 (0.029) 0.0036 (0.0009) stopovers in 2004 and shorter stopovers in 2005. We did not compare stopover length by age and gender because of the low number of recaptures. Regression equations for body mass against wing chord and time of capture indicated a significant influence of both wing and time since sunrise on body mass = 0.15, F21039 = 92.65, = 12.89, < 0.001, = 4.79, Rt,me < 0.001, Table 2). We predicted from the regression equation that an average- sized individual (mean wing = 95.7 mm) would add about 1 .08 grams over the 6-hr pe- riod from sunrise equating to 0.18 g/hr. This estimate would be —0.67% of lean body mass/ hr using the average mass for individuals with a fat score of 0. We examined age effects on rates of body mass gain by regressing mass against time of capture for SYs and ASYs separately. The equation for SYs revealed a significant effect of wing length and time since sunrise on mass {R^ = 0.16, = 0.30 (SE = 0.03), P_g < 0.001, = 0.004 (SE = 0.001), Ptime < 0.001). We predict an av- erage-sized SY individual with a mean wing length of 95.3 mm would add 1.3 g over the 6-hr morning period. Wing chord for ASYs had a significant effect but time since sunrise did not although the coefficient was positive (R2 = 0.15, p,.,g = 0.28 (SE = 0.05), P^.„g < 0.001, = 0.002 (SE - 0.002), = 0.284). Analysis of fat reserves at time of banding revealed that most individuals had some fat (72.5% of individuals had a fat score >1, Fig. 2). There was a tendency for ASYs to have higher fat scores than SYs and higher fat scores in 2004 than 2005 for both groups. Measures of energetic condition from residu- als of a mass-wing regression did not reveal distinct differences by age or gender and there was variability within classes. Residual esti- mates for each group were [mean ± SE (n)] ASY male: 0.43 ± 0.51 g (14), SY male: 0.44 ± 0.44 g (21), ASY female: -0.06 ± 0.49 g (14), SY: female -0.10 ± 0.32 g (25). Recap- tured and non-recaptured individuals had sim- ilar fat scores (mean ± SE) (recapture =1.4 ± 0.35, non-recapture = 1.5 ± 0.05). Analy- ses of mass change from recaptured birds re- vealed variability among individuals. Seven of 15 individuals in 2004 and 2005 lost mass during the stopover (average loss = 0.6 g/day) while 8 gained mass (average gain = 0.3 g/day). The average mass change based on these individuals was lower than predicted by regressing mass against time of capture using all birds. DISCUSSION Captures. — Swainson’s Thrushes were common spring migrants at a stopover site along the Osa Peninsula, Costa Rica. This spe- cies is a rare winter resident in the region and all individuals captured were likely passage migrants. Individuals from South American wintering areas likely reached the study site by following the Panama isthmus or via an ocean crossing from Colombia making land- fall between the Azuero Peninsula of Panama and the Osa Peninsula of Costa Rica (Wilson et al. 2008). Swainson’s Thrushes are also common spring migrants along the Caribbean coast of Central America (Galindo et al. 1963, Ridgely and Gwynne 1989, Ralph et al. 2005). Most continental birds enter North America along the Gulf coast (Rappole and Warner 1976; Wang and Moore 1993, 1997) before moving to breeding locations across the boreal forest and mountainous areas from Alaska to eastern Canada and the northeastern United States (Mack and Wang 2000, Ruegg and Smith 2002). Swainson’s Thrush capture rates averaged 20-25 individuals/ 100 net hrs and are among the highest reported for this species in spring. Wilson et al. • STOPOVER ECOLOGY OF SWAINSON’S THRUSH 79 Fat score FIG. 2. Fat scores for second year (SY) and after-second year (ASY) Swainson’s Thrushes on the Osa Peninsula, Costa Rica, in (A) 2004 (SY = 249, ASY = 149) and (B) 2005 (SY = 450, ASY = 58). Spring capture rates along the northern Gulf coast of Mexico for four thrush species com- bined were —8-10 individuals/ 100 net hrs; Swainson’s Thrushes accounted for about 36% of those captures (Wang and Moore 1997). Elsewhere, spring capture rates of Swainson’s Thrushes (individuals/ 100 net hrs) were 0.49 in Minnesota (Winker et al. 1989), 1.5 in southern Alberta (Collister et al. 2005), and 2.35 in South Dakota (vSwanson et al. 2003). At Almirante on the Caribbean coast of Panama, 365 Swainson’s Thrushes were captured during netting in April 1963 (net hrs not available, Galindo et al. 1963), slightly lower than numbers caught in April at our site. Caution is required when inferring abundance from capture rates of mist-net studies because of differences in methodology, effort, dates of study, and habitat (Remsen and Good 1996). Mist-netting studies may also be biased against larger species and those that primarily inhabit the canopy and subcanopy layers 80 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 (Blake and Loiselle 2001). Swainson’s Thrushes primarily forage in lower forest lev- els (Mack and Wang 2000) and were usually observed in the understory at our site. Thus, mist-net captures likely provide an accurate index of abundance for this species. The high number of captures in our study relative to North American sites may have been due to the small landmass of southern Central Amer- ica, which likely concentrates individuals dur- ing migration. Migration Timing. — Males passed through the site ~6 days earlier than females in 2004 with mean passage dates of 6 April and 12 April, respectively. Few studies have exam- ined gender differences in migratory timing of Catharus thrushes because they are sexually monomorphic and require genetic techniques to accurately ascertain gender. Wing lengths of migrating thrushes along the Gulf coast in spring declined over time, suggesting earlier passage of larger males (Wang and Moore 1997). A similar pattern was noted among Swainson’s Thrushes at a spring stopover site in Minnesota (Winker et al. 1989). Gender- specific spring migration patterns are fre- quently reported in sexually dimorphic species (Francis and Cooke 1986, Wang et al. 1998, Swanson et al. 1999). Mean passage date of male Northern Wheatears (Oenanthe oenanthe leucorhoa) in Scandinavia was 5 days earlier than females (Dierschke et al. 2005) while on Appledore Island, Maine, male American Redstarts (Setophaga ruticilla) and Common Yellowthroats (Geothlypis trichas) preceded females by 2 and 5 days, respectively (Morris and Glasgow 2001, Morris et al. 2003). Earlier spring passage of males is likely re- lated to male-male competition in breeding ar- eas and the need for males to establish terri- tories prior to arrival of females (Lack 1954). Earlier passage of males through stopover sites could be due to: (1) earlier departure from wintering areas (Marra et al. 1998), (2) faster migration rates (Swanson et al. 1999), or (3) more northerly wintering areas and a shorter distance to stopover sites (Chandler and Mulvihill 1990). Although our site was relatively close to wintering areas, differences between males and females were similar to patterns from other species at North American stopover sites. This might suggest that earlier passage of males was due to earlier departure and/or more northern wintering areas, al- though it is still possible that males travel more quickly than females in spring. We found mixed results for age-specific timing of migration. Using all known-age in- dividuals, ASYs preceded SYs by about 3 days in 2004 but the two groups had similar mean passage dates in 2005. Within the short- ened study periods in the first 2 years, SYs preceded ASYs by 1 day in 2002 and ASYs preceded SYs by 2 days in 2003. However, these data were based on all known-age birds and, given gender-specific differences in pas- sage date, it is difficult to interpret timing by age group when males and females are pooled. Furthermore, there tended to be more unknown age individuals at the start of the season, which may not have been equally rep- resentative of both age groups. Although the sample size was smaller, our data for individ- uals of known age and gender in 2004 may be most informative and indicated that ASY males passed through earliest, SY males and ASY females were intermediate, and SY fe- males were the latest to arrive. This is consis- tent with earlier migration by males and a ten- dency for older birds to precede first-year birds (Francis and Cooke 1986, Stewart et al. 2002). We are uncertain why there was annual variation in age ratios. There was a relatively higher proportion of ASYs in 2002 and 2004 than in 2003 and 2005. Annual variation in the abundance of age groups at stopover sites may be due to greater productivity the previ- ous spring, age-specific differences in surviv- al, or variability in the routes or stopovers sites used by different age groups (Ralph 1981, Wang et al. 1998, Dean et al. 2004). Changes in study design or survey effort may also influence age ratios at stopover sites (Kel- ly and Finch 2000). Our data from 2002 and 2004 included a higher number of individuals from the early migratory period which, if there was a tendency for older birds to move earlier, may explain the higher proportion of ASYs captured in those years. Further re- search on variation in the abundance and tim- ing of passage by age class at other tropical stopover sites would be useful. Stopover Ecology. — Only 26 of 1,490 in- dividuals (1.7%) captured in 2003 to 2005 were recaptured on subsequent days suggest- ing the majority of individuals moved from Wilson et al. • STOPOVER ECOLOGY OE SWAINSON’S THRUSH 81 the study area soon after capture. Spring re- capture rates were similar to those from the St. Croix River Valley, Minnesota (1.2%, Winker et al. 1989) but were lower than rates observed elsewhere. Recapture rates were 9 and 7% from two separate studies along the northern Gulf coast (Kuenzi et al. 1991, Wang and Moore 1997), 5.7% along coastal regions of south Texas (Rappole and Warner 1976), and 5.2% along the Caribbean coast of Pan- ama (Galindo et al. 1963). One reason for the low recapture rate at our study site may have been that it was small (3 ha) with extensive areas of similar nearby habitat. Thus, individuals may continue to move through these habitats without returning to the study site. Recapture rates may be high- er at sites that are spread over a wider region increasing the likelihood of capturing individ- uals that wander but remain within the general area. Galindo et al. (1963), on the Caribbean coast of Panama, surveyed four plots over a 25.9 km^ area and recorded higher recapture rates. Cochran and Wikelski (2005), using ra- dio-telemetry, found that Swainson’s Thrushes at stopover sites in Illinois established small foraging areas of ~ 1 00 m in diameter and re- mained in that area until the next flight. How- ever, the large number of individuals at our site may have influenced how many remained in the area (Moore and Wang 1991, Wang et al. 1998). Differences in food abundance be- tween locations may also lead to variation in stopover behavior and recapture rates (e.g.. Petit 2000). Individuals may also be less likely to re- main at stopover sites if they have sufficient reserves to continue on migration. Stopover length is often negatively related to energetic condition (Moore and Kerlinger 1987, Coch- ran and Wikelski 2005). Experimental studies have found that, following a trans-Gulf cross- ing, individuals with greater energetic re- serves have higher migratory restlessness than those in poor condition (Wang and Moore 1993, but see Smith and Norment 2005). Mul- ti-day stopovers may be necessary if all fat reserves are depleted, as is often the case along the gulf coast (e.g., Kuenzi et al. 1991). Many individuals in our study had fat scores ^1 and gained mass through the morning, which combined may provide sufficient re- serves to continue on that night. Migrants at stopover sites in Costa Rica would also not cross a major geographic barrier until they reached the Gulf of Mexico and it may not be necessary to build reserves beyond what is re- quired for a few hours of flight (e.g., Schaub and Jenni 2000b, Rubolini et al. 2002). Indi- viduals that did stopover typically remained for ~1 to 4 days, similar to lengths observed for this species in temperate areas further north (Winker et al. 1989, Wang and Moore 1997). Rates of mass gain through the morning av- eraged —0.18 g/hr, which equates to 0.67% of the lean body mass (average mass of individ- uals with no visible fat) (Dunn (2002). This is similar to the pooled estimates for spring mi- grants across Canada (0.40% of lean body mass), although rates for Swainson’s Thrushes in that study were low (average = —0.06% of lean body mass/hr, Dunn 2001, 2002). Daily mass gain for Swainson’s Thrushes in Min- nesota (Winker et al. 1989) was —4.08 g/day in spring, slightly higher than our data. Vari- ation in rates of mass gain among sites could be due to differences in food availability (Dunn 2001) or whether migrants are using a site for other purposes such as to rest or re- hydrate (Kuenzi et al. 1991). Entire populations of some neotropical mi- grant species may pass through a narrow geo- graphic corridor during early stages of spring migration because of the small landmass of southern Central America. Habitat disturbance in these regions could negatively impact pop- ulations of these species (Sillett and Holmes 2002, Newton 2006). We know little about migratory routes, habitat requirements, and stopover behavior of neotropical migrants in tropical regions. Our study revealed that Swainson’s Thrush use lowland habitats along Costa Rica’s Pacific coast in spring; high numbers relative to North American sites in- dicate thrushes are particularly concentrated during this period. Recapture rates, stopover lengths, and rates of mass gain were similar to those in temperate areas suggesting these behaviors may be consistent across the spring migratory period. Studies that use a combi- nation of mist netting and counts in high and low elevation habitats of southern Central America would further our knowledge of stopover ecology of neotropical migrant song- birds within the tropics. 82 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 ACKNOWLEDGMENTS We thank S. R. Morris for conducting analyses on stopover length using capture-recapture methods, C. E. Ritland, A. E. Miscampbell, and H. H. Yeh for help with genetic analyses, and Rex Kenner for assisting with gender classification of museum specimens. We also thank Jolanda Hess, David Argello, and Jay Tress for access to their property. Yong Wang and an anon- ymous reviewer provided valuable comments on an earlier draft of this manuscript. We thank members of the Calgary Bird Banding Society (A. M. Cole, Jen Lane, Steve Lane, P. E. Mitchell, Mike Mulligan, E. A. Peterson, Carl Savignac, G. J. Smiley, W. P. Taylor, B. R. Trakalo, and C. M. Watson) and Alberto and Elizabeth Quesada-Calvo for assistance in the field. Funding for this project was provided by members of the Calgary Bird Banding Society, Alberta Gaming and Liquor Commission, Canadian Wildlife Service, NSERC (doctoral grants to SW, AGW) and a Killam doctoral award to SW. LITERATURE CITED Achard, E, H. D. Eva, H. J. Stibig, P. Mayaux, J. Gallego, T. Richards, and J. P. Malingreau. 2002. Determination of deforestation rates of the world’s humid tropical forests. Science 297:999- 1002. Alerstam, T. and a. Lindstrom. 1990. Optimal bird migration: the relative importance of time, energy and safety. Pages 331-351 in Bird migration: physiology and ecophysiology (E. Gwinner, Edi- tor). Springer- Verlag, Berlin, Germany. Blake, J. G. and B. A. Loiselle. 2001. Bird assem- blages in second growth and old-growth forests, Costa Rica: perspectives from mist nets and point counts. Auk 118:304-326. Chandler, C. R. and R. S. Mulvihill. 1990. Inter- preting differential timing of capture of sex clas- ses during spring migration. Journal of Field Or- nithology 61:85-89. Cherry, J. D. 1982. Fat deposition and length of stop- over of migrant White-crowned Sparrows. Auk 99:725-732. Cochran, W. W. and M. Wikelski. 2005. Individual migratory tactics of New World Catharus thrush- es: current knowledge and future tracking options from space. Pages 274-289 in Birds of two worlds (R. Greenberg and P. P. Marra, Editors) Johns Hopkins University Press, Baltimore, Maryland, USA. CoLLisTER, D. M., R. D. Dickson, and G. Smiley. 2005. 2004 Calgary Bird Banding Society Annual Technical Report. Calgary, Alberta, Canada. Dean, K. L., H. A. Carlisle, and D. L. Swanson. 2004. Age structure of neotropical migrants dur- ing fall migration in South Dakota: is the northern Great Plains region an inland “coast”? Wilson Bulletin 116:295-303. Dierschke, V., B. Mendel, and H. Schmaljohann. 2005. Differential timing of spring migration in northern wheatears Oenanthe oenanthe: hurried males or weak females? Behavioral Ecology and Sociobiology 57:470-480 Dunn, E. H. 2001. Mass change during migration stop- over: a comparison of species groups and sites. Journal of Field Ornithology 72:419-432. Dunn, E. H. 2002. A cross-Canada comparison of mass change in birds during migration stopover. Wilson Bulletin 1 14:368-379. Francis, C. M. and E Cooke. 1986. Differential tim- ing of spring migration in wood warblers (Paru- linae). Auk 103:548-556. Galindo, P, E. Mendez, and A. J. Adames. 1963. Banding of migrant thrushes in Almirante, Pana- ma. Bird Banding 34:202-209. Greenberg, R. 1980. Demographic aspects of long- distance migration. Pages 493-516 in Migrants in the Neotropics: ecology, behavior, distribution and conservation (A. Keast and E. S. Morton, Editors). Smithsonian Institution Press, Washington, D.C., USA. Grieeiths, R., M. C. Double, K. Orr, and R. J. G. Dawson. 1998. A DNA test to sex most birds. Molecular Ecology 7:1071-1075. Helms, C. W. and W. H. Drury. 1960. Winter and migratory weight and fat: field studies on some North American buntings. Bird-Banding 31:1-40. Holdridge, L. R. 1967. Life zone ecology. Tropical Science Center, San Jose, Costa Rica. Kelly, J. F. and D. M. Finch. 2000. Effects of sam- pling design on age ratios of migrants captured at stopover sites. Condor 102:699-702. Kuenzi, a. j., E R. Moore, and T. R. Simons. 1991. Stopover of neotropical landbird migrants on East Ship Island following Trans-Gulf migration. Con- dor 93:869-883. Lack, D. 1954. The natural regulation of animal num- bers. Oxford University Press, London, United Kingdom. Lindstrom, A. 1990. The role of predation risk in stop- over habitat selection in migrating Bramblings Fringilla montif ring ilia. Behavioral Ecology 1: 102-106. Mack, D. E. and Y. Wang. 2000. Swainson’s Thrush {Catharus ustulatus). The birds of North America. Number 540. Marra, P. P, K. A. Hobson, and R. T. Holmes. 1998. Linking winter and summer events in a migratory bird by using stable-carbon isotopes. Science 282: 1884-1886. Moore, E R. and D. A. Aborn. 2000. Mechanism of en route habitat selection: how do migrants make habitat decisions during stopover? Studies in Avi- an Biology 20:34-42. Moore, E R. and P. Kerlinger. 1987. Stopover and fat deposition by North American wood warblers (Parulinae) following spring migration over the Gulf of Mexico. Oecologia 74:47-54. Moore, E R. and Y. Wang. 1991. Evidence of food- based competition during migratory stopover. Be- havioral Ecology and Sociobiology 28:85-90. Wilson et al. • STOPOVER ECOLOGY OF SWAINSON’S THRUSH 83 Morris, S. R. and J. L. Glasgow. 2001. Comparison of spring and fall migration of American Redstarts on Appledore Island, Maine. Wilson Bulletin 113: 202-210. Morris, S. R., C. R. Pusateri, and K. A. Battaglia. 2003. Spring migration and stopover ecology of Common Yellowthroats on Appledore Island, Maine. Wilson Bulletin 115:64-72. Morris, S. R., D. A. Liebner, A. M. Larracuente, E. M. Escamilla, and H. D. Sheets. 2005. Multiple day constancy as an alternative to pooling for esti- mating mark-recapture stopover length in Nearctic- neotropical migrant landbirds. Auk 122:319-328. Newton, I. 2006. Can conditions experienced during migration limit the population levels of birds? Journal of Ornithology 147:146-166. Petit, D. R. 2000. Habitat use by landbirds along Ne- arctic-neotropical migration routes: implications for conservation of stopover habitats. Studies in Avian Biology 20:15-33. Pyle, P. 1997. Identification guide to North American birds. Part I. Slate Creek Press, Bolinas, Califor- nia, USA. R Development Core Team. 2006. R: a language and environment for statistical computing, reference in- dex version 2.3.1. R Foundation for Statistical Com- puting, Vienna, Austria. http://www.R-project.org (accessed 24 February 2007). Ralph, C. J. 1981. Age ratios and their possible use in determining autumn routes of migrant passer- ines. Wilson Bulletin 93:164-188. Ralph, C. J., M. J. Widdowson, R. I. Frey, P. A. Her- rera, AND B. P. O’Donnell. 2005. An overview of a landbird monitoring program at Tortuguero, on the Caribbean coast of Costa Rica. USDA, For- est Service, General Technical Report PSW-GTR- 191. Albany, California, USA. Rappole, j. H. and D. W. Warner. 1976. Relation- ships between behavior, physiology and weather in avian transients at a migration stopover site. Oecologia 26:193-212. Remsen, j. V. AND D. A. Good. 1996. Misuse of data from mistnet captures to assess relative abundance in bird populations. Auk 113:381-398. Richardson, W. J. 1978. Timing and amount of bird migration in relation to weather: a review. Oikos 30:224-272. Ridgely, R. S. and j. a. Gwynne. 1989. A guide to the birds of Panama. Princeton University Press, Princeton, New Jersey, USA. Rubolini, D., a. G. Pastor, A. Pilastro, and F. Spi- na. 2002. Ecological barriers shaping fuel stores in Barn Swallows Hirundo rustica following the central and western Mediterranean flyways. Jour- nal of Avian Biology 33:15-22. Ruegg, K. C. and T. B. Smith. 2002. Not as the crow flies: a historical explanation for circuitous migra- tion in Swainson's Thrush {Catluinis ustnlatus). Proceedings of the Royal Society of London Se- ries B 269:1375-1381. Sanchez-Azofeifa, G. A., B. Rivard, J. Calvo, and I. Moorthy. 2002. Dynamics of tropical defor- estation around national parks: remote sensing of forest change on the Osa Peninsula of Costa Rica. Mountain Research and Development 22:352- 358. ScHAUB, M. AND L. Jenni. 2000a. Body mass of six long-distance migrant passerine species along the autumn migration route. Journal fiir Ornithologie 141:441-460. ScHAUB, M. AND L. Jenni. 2000b. Fuel deposition of three passerine bird species along the migration route. Oecologia 122:306-317. ScHAUB, M., R. Pradel, L. Jenni, and J. D. Lebreton. 2001. Migrating birds stopover longer than usu- ally thought: an improved capture-recapture anal- ysis. Ecology 82:852-859. Schulte-Hostedde, a. I., B. Zinner, J. S. Millar, AND G. J. Hickling. 2005. Restitution of mass-size residuals: validating body condition indices. Ecol- ogy 86:155-163. SiLLETT, T. S. AND R. T. HoLMES. 2002. Variation in survivorship of a migratory songbird throughout its annual cycle. Journal of Animal Ecology 71: 296-308. Smith, R. J. and F. R. Moore. 2005. Arrival timing and seasonal reproductive performance in a long- distance migratory landbird. Behavioral Ecology and Sociobiology 57:231-239. Smith, S. B. and C. J. Norment. 2005. Nocturnal ac- tivity and energetic condition of spring landbird migrants at Braddock Bay, Lake Ontario. Journal of Field Ornithology 76:303-310. Stewart, R. L. M., C. M. Francis, and C. Massey. 2002. Age related timing of spring migration within sexes in passerines. Wilson Bulletin 114: 264-271. Stiles, F. G. and A. F. Skutch. 1995. A guide to the birds of Costa Rica. Cornell University Press, Ith- aca, New York, USA. Swanson, D. L., H. A. Carlisle, and E. T. Liknes. 2003. Abundance and richness of neotropical mi- grants during stopover at farmstead woodlots and associated habitats in southeastern South Dakota. American Midland Naturalist 149:176-191. Swanson, D. L., E. T. Liknes, and K. L. Dean. 1999. Differences in migratory timing and energetic- condition among sex/age classes in migrant Ruby- crowned Kinglets. Wilson Bulletin 1 1 1:61-69. Wang, Y. and F. R. Moore. 1993. Relation between migratory activity and energetic condition among thrushes (Turdinae) following passage across the Gulf of Mexico. Condor 95:934-943. Wang, Y. and F. R. Moore. 1997. Spring stopover of intercontinental migratory thrushes along the northern coast of the Gulf of Mexico. Auk 1 14: 263-278. Wang, Y, D. M. Finch, F R. Moorf;, and J. F Kelly. 1998. Stopover ecology and habitat use of migra- tory Wilson’s Warblers. Auk 1 15:829-842. WiL.soN, S., K. A. Hobson, D. M. Colli.stf;r, and A. G. WiL.soN. 2008. Breeding destinations and 84 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 spring migration patterns of Swainson’s Thrush at a Costa Rican stopover site. Auk 125: In Press. Winker, K., D. W. Warner, and A. R. Wiesbrod. 1989. The Northern Waterthrush and Swainson’s Thrush as transients at a temperate inland stopover site. Pages 384-404 in Ecology and conservation of neotropical migrant landbirds (J. M. Hagan III and D. W. Johnston, Editors). Smithsonian Insti- tution Press, Washington D.C., USA. Winker, K., D. W. Warner, and A. R. Wiesbrod. 1992. Daily mass gains among woodland migrants at an inland stopover site. Auk 109:853-862. The Wilson Journal of Ornithology 120( 1):85-91, 2008 DEMOGRAPHY OF EASTERN YELLOW WAGTAILS AT CAPE ROMANZOF, ALASKA HEATHER M. RENNERS AND BRIAN J. McCAEEERY^ ABSTRACT. — We quantified demographic parameters of Eastern Yellow Wagtails (Motacilla tschutschensis) breeding at Cape Romanzof, Yukon Delta National Wildlife Refuge, Alaska. We monitored 79 nests in an 837- ha area during 1997-1999 and banded 160 individuals. Mayfield nest success differed among years and ranged from 0.14 to 0.63/year. Most nest failures were attributed to predation. Annual fecundity (mean number of fledglings/female) ranged from 0.7 to 3.7. At least 8.8% of nests had polygynous males; females paired with polygynous males had the same fecundity as monogamous females. Forty-two to 100% of the breeding males returned the following year, usually to the same territory while no adult females returned. Four nestlings banded in the study area returned to nest the following year. The best model for annual survival accounted for differences between both age groups and years. The demography of Eastern Yellow Wagtails at Cape Romanzof varied and was characterized by relatively high adult male survival and site fidelity, female-biased dispersal, and weak natal philopatry. The absence of returning females is significant and possible differences in migration stopovers and wintering locations should be investigated. Moderate levels of male immigration may be necessary in periodic pulses to maintain a local population, but female immigration would need to be massive and sustained. Received 8 September 2006. Accepted 17 June 2007. Four major demographic factors affect life history patterns in bird populations: age-spe- cific fecundity, recruitment, survival from fledging to breeding, and adult survival (Nur and Sydeman 1999). These factors can be dif- ficult to evaluate because individuals are rare- ly seen after hatching or fledging. Some spe- cies of migratory birds demonstrate fidelity to breeding sites over a period of years allowing mark-recapture techniques to be used to esti- mate demographic parameters. Demographic parameters have been esti- mated for few long-distance migrant passer- ines in North America. Specifically, survival estimates based on recaptures of banded birds are rare in the published literature, although recent publication of Monitoring Avian Pro- ductivity and Survivorship (MAPS) data on- line (Michel et al. 2006) provided estimates for many species. Other than MAPS data for Arctic Warbler {Phylloscopus borealis), de- mographic parameters have not been reported for any of the other five Paleotropic passerine ' Department of Natural Re.sources, Fernow Hall, Cornell University, Ithaca, NY 14853, USA. 2 Yukon Delta National Wildlife Refuge, U.S. Fish and Wildlife Service, P. O. Box 346, Bethel, AK 99559, USA. Current address: Alaska Maritime National Wild- life Refuge, U.S. Fish and Wildlife Service, 95 Sterling Highway, Suite 1, Homer, AK 99603, USA. ^ Corresponding author; e-mail: heather_renner@fws.gov species nesting in western Alaska: Bluethroat {Luscinia svecica). Northern Wheatear {Oen- anthe oenanthe). White Wagtail {Motacilla alba). Eastern Yellow Wagtail (M. tschut- schensis), and Red-throated Pipit {Anthus cer- vinus). Our objective in this 3-year study was to obtain the first demographic information for one of these species. Specifically, we quanti- fied the following demographic parameters of Eastern Yellow Wagtails breeding at Cape Ro- manzof, Alaska: reproductive success, annual adult and first-year survival, site fidelity and natal philopatry, and earliest age at first re- production. METHODS Study Area. — We studied Eastern Yellow Wagtails at Cape Romanzof Long Range Ra- dar Site, Alaska (61° 49' N, 166° 5' W) during May to August 1997-1999. Cape Romanzof projects into the Bering Sea at the western end of the Askinuk Mountains between Scammon and Kokechik bays. The Askinuk Mountains rise from the delta lowlands at sea level to 714 m and extend inland about 50 km. Dwarf shrub meadows and block-fields dominate the study area which includes two drainages, Nil- umat and ISouth creeks. Nest Searching, Capturing Birds, and Banding Protocols. — We searched for nests soon after Eastern Yellow Wagtails arrived in breeding areas in late May by watching adults fly to the nest, (a) with nesting materials or 85 86 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 with food, or (b) after mobbing potential pred- ators (including human observers). Nests were marked with flags >5 m distant to minimize the likelihood of attracting predators to the nest. We used mist nets to capture adults at nest sites in 1997-1999 and we marked indi- viduals with standard USGS aluminum leg bands as well as unique combinations of three color-bands. We captured adults only after eggs had hatched to minimize the risk of nest abandonment. We marked chicks with aluminum leg bands prior to fledging in 1997 and 1998. We recorded natural fledging dates in 1997 and banded young only after at least one individ- ual in the nest had departed to avoid prema- ture fledging due to our activities. We attempt- ed to mark every nestling that fledged on the study area in 1998 by banding them in the nest as soon as they were sufficiently large to re- tain a band, generally 10 days after hatching. We did not color-band nestlings because por- tions of their tarsi were too large to accom- modate more than one band. We also captured Eastern Yellow Wagtails distant from nest sites in mist nets after young had fledged in both 1997 and 1998 to document post-fledg- ing movements of previously banded hatch- ing-year birds, and to mark adults and fledg- lings not previously banded. We operated an array of 6-13 mist nets around our camp and at the beach near the mouth of Nilumat Creek daily during late July and August from 0600 to 1300 hrs (weather permitting). Reproductive Success.— Wq checked nests at least twice each week during the egg-laying and incubation periods, and every second day after hatching. We recorded the contents of the nest as well as presence/absence of adults and of mobbing behavior. We defined fledging as successful nest departure, although the chicks could not necessarily fly at this age. Nests from which at least one young fledged were considered successful. We confirmed fledging by observing fledglings or adults carrying food in the immediate nest area. We attributed nest failure to abandonment if eggs or dead chicks remained in the nest bowl and, other- wise to predation. We estimated daily survival rate (5), and Mayfield nest success (P2) (Mayfield 1960, 1975; Klett et al. 1986) for all nests with >2 visits while the nest was active. The estimated number of exposure days for a successful nest varied slightly each year (Renner and Mc- Caffery 2006k We calculated confidence in- tervals for daily survival using Johnson’s ( 1 979) formula to compare Mayfield nest suc- cess and daily survival between years. Hatch- ing (chicks/eggs) and fledging success were calculated for each year and for all 3 years combined. All nests were not included in this analysis; for example, if we found a nest after hatching, we were unable to ascertain the original clutch size. Return Rate and Dispersal. — We defined a returning bird as a banded bird that nested in the study area in 1 year and was observed in the study area during the following year. A recaptured bird was one that was previously captured in the study area during the same year. A local bird was a young-of-the-year that was definitely raised in the study area. A hatching-year (HY) bird was a young-of-the- year that may have been raised in the study area. This latter definition refers to unbanded birds captured in the post-fledging mist net ar- ray that were sufficiently developed to be ca- pable of sustained flight. Breeding dispersal distance is the distance between nests of an individual in consecutive years (Haas 1990). Distances between nests in consecutive years were measured with a 50-m tape or, when the distance exceeded 500 m, on a 1:25,000 to- pographic map. We used a 20-60 X spotting scope to iden- tify banding status whenever adults were ob- served. In some cases, returning adults had not been color banded and were captured to identify the individual from its metal band. We calculated return rates by dividing the number of banded birds that returned to breed on the study site in the current year by the total number of banded birds breeding on the study site in the previous year (Haas 1998). We considered all marked adults that were re- observed or recaptured, regardless of whether or not their nests were found, to have re- turned. We evaluated the effect of nest success in year n on the probability of returning in year n + I using 2X2 contingency tables and Fisher’s exact test. Local Annual Survival Estimates. — Annual apparent survival (actual survival minus dis- persal) was estimated using program MARK (White and Burnham 1999) following Lebre- Renner and McCaffery • EASTERN YELLOW WAGTAILS IN ALASKA 87 TABLE 1. Nest success of Eastern Yellow Wagtails at Cape Romanzof, Alaska, 1997-99. The Mayfield exponent is the estimated number of exposure days for a successful nest (from Renner and McCaffery 2006). Year # nests monitored (# successful) Mayfield success Mayfield exponent Daily survival rate 95% CI for s 1997 28 (20) 0.632 28.15 0.984 0.973-0.995 1998 34 (14) 0.332 27.08 0.960 0.942-0.978 1999 17 (2) 0.136 28.00 0.931 0.897-0.965 ton et al. (1992) and Nichols et al. (1993). No marked birds were seen that had not been marked or seen in the previous year and the probability of reobservation could not be dis- tinguished from 1 . The probability of reobser- vation was assumed to be high and “known fate” survival estimates were calculated (Pol- lock et al. 1989a, b). Known fate parameter estimates assume that if birds were not ob- served, they did not survive (return to the population). Known fate models typically es- timate actual survival but, in this case, they estimate only apparent survival. Model nota- tion followed Lebreton et al. (1992). The fac- torial structure of the model is represented by subscripting the primary survival parameter (S) using “t” for annual differences and “g” for group (age, adult or juvenile, at banding) differences. Constancy over time is indicated with a dot (.). Relationships among factors were indicated using standard linear model notation. We first defined potential models for each year separately and all years considered to- gether (Burnham and Anderson 1998). The candidate model set included a fully parame- terized global model and a series of reduced parameter models involving the survival rate. Model selection was based on comparison of the Akaike Information Criterion (AIC^). The model with the lowest AIC^ was accepted as being most parsimonious for the data. Com- parisons among models in the candidate mod- el set were based on indices of relative model plausibility using Akaike weights (Burnham and Anderson 1998). The ratio of Akaike weights between any two models indicates the relative extent to which a particular model is better supported by the data than the other model. We report parameter estimates and as- sociated standard errors derived by averaging over all models in the candidate model set weighted by Akaike model weights to account for uncertainty in model selection (Buckland et al. 1997, Burnham and Anderson 1998). RESULTS Nest Searching and Banding. — We moni- tored 79 nests during 1997-1999 in the Nil- umat and South Creek drainages. We banded 9 adults, 9 local nestlings, and 43 HYs in 1997. All adults banded were successful breeders. In 1998, we banded 35 adults, 61 local nestlings, and 3 HYs. We captured at least one adult at all 14 suceessful nests in- cluding the males at 13 (93%) and the females at 11 (79%). We captured males at only 6 of 20 (30%) unsuccessful nests and females at 7 of 20 (35%) unsuccessful nests. We captured 49 HY birds during post-fledging mist-netting operations at the mouth of Nilumat Creek of which three were recaptures from elsewhere on our study site. Reproductive Success. — Nest success was highest in 1997 and lowest in 1999 (Table 1). These 2 years differed more than four fold (i.e., non-overlapping 95% confidence inter- vals for .S' and P2); 1998 did not differ signif- icantly from either year. Mean number of fledglings per female ranged from 0.67 to 3.71 across the 3 years {n = 12 nests with known fates; Table 2) and all years differed (ANO- VA, df = 2, R < 0.05; Tukey’s pairwise com- parisons, P < 0.05). Mean fecundity was higher for polygynous than for monogamous males {P < 0.05). Breeding success of fe- males did not differ between those mated to monogamous and polygynous males. Polygy- nous males averaged 4.67 young lledged/male {n = 1 [nests of 3 males in 3 yearsf SE = 2.9 1 ), whereas 7 females averaged 2 fledg- lings per female (wSE = 0.78). Causes of Nest Failure. — We did not ob- serve nest predation, but assumed it was the major cause of nest failure at Cape Romanzof based on frequent observations of Long-tailed 88 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120. No. I. March 2008 TABLE 2. Measures of fecundity for 72 Eastern Yellow Wagtail nests at Cape Romanzof, Alaska, 1997- 99. Parameter Number eggs laid Number eggs hatched (%) Number eggs fledged (%) Proportion hatched eggs that fledged young Mean ± SE fledglings/female Mean ± SE fledglings/successful female Mean ± SE fledglings/monogamous male Mean ± SE fledglings/polygynous male Measure 396 (mean clutch size = 5.5) 301 (76) 148 (37) 49% 2.1 ± 0.30 (range 0-6, n = 72) 4.4 ± 0.20 (range 2-6, n = 38) 2.2 ± 0.32 (range 0-6, n = 65) 4.7 ± 2.91 (range 0-10, n = 3) Jaegers {Stercorarius longicaudus) hunting near Eastern Yellow Wagtail nest sites, as well as red fox {Vulpes vulpes) and weasels {Mus- tela spp.) being mobbed by wagtails and other ground-nesting birds. We observed no obvi- ously weather-related mortality. Predation of young probably accounted for all nest failures in 1997 (8) and 1999 (15), and most in 1998 (17). Three nests were abandoned by the par- ents in 1998; one after part of the clutch was lost to predators during incubation and two during early brood-rearing. Neither of the abandonments during brood-rearing was relat- ed to banding, because we had not attempted to capture birds at those nests. Return Rate —OvQmW return rates were 52% (males), 0% (females), and 26% (genders combined). None of 23 banded adult females returned to the study site in any year after banding. All four adult males banded in 1997 returned to the study area in 1998. At least 8 (42%) of the 19 adult banded males present in 1998 (including 1997 returns), returned in 1999. Four nests were lost to predators before we could learn if the males were banded; the return rate of males in 1999 may have been as high as 63%. At least two of the four males banded in 1997 returned in 1999. One of these did not nest in 1999, but sang and displayed persis- tently throughout the season. Return rates for males, females, and both genders combined did not differ between those that nested suc- cessfully and unsuccessfully, respectively, in the previous year (Fisher’s exact test, P — 0.26). Ten of 17 (59%) successful males and 2 of 6 (33%) unsuccessful males returned the following year; no females returned. Return rates were 30% (10/33) for successful and 15% (2/13) for unsuccessful birds when gen- ders were combined. We were less likely to capture birds at nests that eventually failed, because some nests did not survive until our scheduled trapping day. We had a small marked sample of “unsuccessful” breeders, particularly in 1998, and the power to detect differences in return rates of successful breed- ers was low. Dispersal/Site Eleven of 12 re- turning males in 1998 and 1999 nested. Mean breeding dispersal distance for these 1 1 birds was 64.3 ± 32.3 m (range = 0-300 m). Two males used the same nest bowl as in the pre- vious year. Ten of the 12 returns had nested successfully in the previous year. All returning adults nested in the same drainage as in the previous year and exhibited fidelity to their previous year’s breeding territory. The two that did not eventually nest on their previous territory were often observed displaying there early in the spring and, in both cases, appar- ently moved because snow covered most of the territory until long after nest-building had begun in the rest of the population. Natal Philopatry. — One of 70 nestlings (1.4%) banded in the study area returned to nest the following year. This female fledged from a nest in 1998 at the mouth of South Creek. In 1999, she nested (unsuccessfully) along Nilumat Creek, 3,400 m distant. We have circumstantial evidence for natal philo- patry for three birds (2 females, 1 male) band- ed as HYs during July 1997 which nested in the study area in 1998. These birds were not banded at the nest (which would have con- firmed they were raised locally), but records during 1998 showed that at least 80% of birds captured during July (12 of 15) were recap- tures known to have been raised within 1 km of the net array. A more plausible return rate Renner and McCaffery • EASTERN YELLOW WAGTAILS IN ALASKA 89 TABLE 3. Initial model testing for survival (S) of individually marked Eastern Yellow Wagtails nesting at Cape Romanzof, Alaska. Models are sorted by in- creasing AICc value. Subscripts reflect different fac- tors in the model (t = year marked), g = group (ju- veniles or adult males), “t/.” indicates year-depen- dence in that parameter for the first group and con- stancy in that parameter for the second group. Model pa AICc*’ AAlCc AIC weighC S(g-t/.) 3 69.088 0.00 0.526 S(g*t) 4 69.532 0.44 0.421 S(g) 2 73.709 4.62 0.052 S(g-a2 t/.) 3 87.199 18.11 0.000 S(t) 2 104.372 35.28 0.000 ^ Number of estimated parameters. ^ Akaike Information Criterion adjusted for small sample size. AIC weights calculated following Burnham and Anderson (1998). for all 119 HYs banded was 3.6%. Thus, the minimum age of first breeding is 1 year. One other locally-banded bird returned in 1998. An adult female captured in a mist net 100 m north of the mouth of Nilumat Creek on 8 August 1997 nested on lower Nilumat Creek in 1998. She may have nested nearby in 1997 but, when captured, she had no traces of a brood patch and had completed prebasic molt. We could not ascertain if the condition of this female’s brood patch or molt was typ- ical for a local breeder because no females known to have nested locally were captured in early August. Annual Survival Estimates. — The most par- simonious models accounted for differences between both groups (adult male vs. juvenile) and years (Table 3). The model best supported by the data indicates variation among years for adult males but not juveniles. Survival es- timates differed between years for both groups and between young and adult males. Apparent survival estimates (actual survival minus dis- persal) were zero for adult females, because none returned to the study area. Apparent sur- vival of adult males was 0.973 ± 0.0052 in 1998 and 0.386 ± 0.1057 in 1999. Apparent survival of juveniles was 0.044 ± 0.0232 in 1998 and 0.026 ± 0.0158 in 1999. DISCUSSION The demography of Eastern Yellow Wag- tails at Cape Romanzof varied in our 3-year study and was characterized by relatively high adult male survival and site fidelity, female- biased dispersal, and weak natal philopatry. Apparent survival of adult males from 1997 to 1998 was more than double that in 1998- 1999, and corresponded with a nearly propor- tional reduction in nesting attempts between those years (34 to 17 nests). We are confident that we had sufficient observations to detect any color-banded birds had they arrived in our study area. We could not assess the relative contributions of dispersal (perhaps resulting from habitat covered by snow) and mortality to the local population decline. However, Ash- ford (1970) and Wood (1976, 1979) reported high mortality (—50%) for Western Yellow Wagtails (Motacilla flava) in wintering areas in west Africa, as did McClure (1998) for Yel- low Wagtails (M. flava flavissima) in Taiwan, due in large part to harvest by humans for food. If these rates apply to M. tschutschensis, and given the observed return rates (42-63%) of males banded at Cape Romanzof coupled with low natal philopatry, it is plausible that most males not re-observed at Cape Roman- zof probably died rather than dispersed. Eastern Yellow Wagtails at Cape Roman- zof, relative to studies of wagtails in Europe, exhibited high male return/survival rates. Drost’s (1948) 5-year banding study on Hel- goland, Germany (54° 1 1 ' N, 7° 55' E, a l-km^ island, 45 km off the coast) documented re- turn rates for Yellow Wagtails: 1 1 to 43% of banded birds returned after 1 year (5-year mean == 21%), and 7 to 22% of the original cohort (mean = 8%) returned after 2 years. Drost’s data indicated that females returned in nearly equal proportions to males. No other published return rates are available for Yellow Wagtails although Martin and Clobert (1996) suggested European passerines (especially open-nesting species) have lower survival than their North American counterparts, bal- anced by higher fecundity. Indeed, Western Yellow Wagtails averaged >1 successful brood attempt each year (leading to higher overall fecundity despite a lower clutch size) while no wagtails at Cape Romanzof double- brooded (Renner and McCaffery 2006). No nest-site-based data exist regarding fi- delity of female Eastern or Western Yellow Wagtails at other locations (but see Paradis et al.l998 for calculated breeding and natal dis- persal distances based on banding recoveries in Britain). The Paradis ct al. (1998) study. 90 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 unlike our study, was designed to document large dispersal distances rather than precisely examine short distances between nests in sub- sequent years. However, the data from Paradis et al. (1998) fit with our results in suggesting breeding dispersal distances <1 km and sub- stantially larger mean natal dispersal distanc- es. Kovshar (1979) reported high philopatry among a small sample of female White Wag- tails, where four of five females color-banded nested within 100 m of their previous year’s nest and the other nested 1 km distant. One of 100 marked White Wagtail nestlings was recovered 2 years later, 300 m from its natal site, and seven more were recovered the fol- lowing year, >1 km away (Kovshar 1979). The demographic patterns we observed at Cape Romanzof (higher survival of males than females, and of adults than juveniles) generally fit with expectations for passerines. Greenwood (1980) showed that most bird spe- cies demonstrate female-biased breeding dis- persal (i.e., females dispersed further than males after breeding), including 25 of 28 spe- cies and 14 of 15 families analyzed. The com- plete absence of returning females at Cape Romanzof (in contrast to the relatively high return rate of males) is extreme and warrants investigation into possible differences in mi- gration stopovers and wintering locations. Greenwood (1980) demonstrated that natal dispersal in birds is typically greater than breeding dispersal and is, as adult dispersal, also female-biased (i.e., females disperse fur- ther) in 21 of 24 species (11 of 12 families) examined. The small number of returning ju- veniles we observed matches the pattern re- ported by Greenwood (1980), but we did not find female-biased natal dispersal (3 of 4 re- turning fledglings were female). True juvenile survival is probably much higher than appar- ent survival. Ricklefs (1973) and Greenberg (1980) suggested that juvenile survival of most small land birds from fledging until re- turn to breeding areas is 15-30%. Local sur- vival estimates using mark-recapture data, as reported here, may significantly underestimate actual survival because juvenile dispersal is high in many species (Greenwood 1980). Female passerines, particularly long-dis- tance migrants, generally have higher mortal- ity than males (Ricklefs 1973). Our data are consistent with this pattern but we consider it more plausible that our study population of Eastern Yellow Wagtails demonstrates a strong female bias in dispersal as well as mor- tality. If our estimate of adult female survival is representative, then observed fecundity rates are certainly too low to sustain this pop- ulation without substantial and sustained im- migration of females. ACKNOWLEDGMENTS We thank C. R. Smith, M. E. Richmond, and C. C. Krueger of the Department of Natural Resources at Cornell University for support throughout this project. Funding was provided by the U.S. Fish and Wildlife Service, Cornell University, Western Bird Banding As- sociation, and a Walter Benning research grant. Logis- tical support at Cape Romanzof was provided by the PMC/Frontec staff We thank the U.S. Air Force and particularly Gene Augustine (61 1th Air Support Group Conservation Resources Program, Elmendorf Air Force Base). The field work at Cape Romanzof would not have happened without their support and encour- agement. Field data were collected by the authors, C. M. Harwood, M. E. Hopey, M. B. Krosby, Christine McCaffery, C. K. Melin, and A. M. Moreland. E. G. Cooch conducted the MARK survival analyses and provided substantial advice. Valuable comments on the manuscript were also provided by I. L. Jones, S. M. Matsuoka, Nadav Nur, Martin Renner, and B. K. San- dercock. Translation of German articles was provided by Harmony Hall. LITERATURE CITED Ashford, R. W. 1970. Yellow wagtails at a Nigerian winter roost: analysis of ringing data. Bulletin of Nigerian Ornithology 7:24-26. Buckland, S. T, K. P. Burnham, and N. H. Augustin. 1997. Model selection: an integral part of infer- ence. Biometrics 53:603-618. Burnham, K. P. and D. R. Anderson. 1998. Model selection and inference — a practical information- theoretic approach. Springer- Verlag, New York, USA. Drost, V. R. 1948. Populationsstudien an der En- glischen Schafstelze, Motacilla flava flavissima Blyth, auf Helgoland. Vogelwarte 1:18-28. Greenberg, R. 1980. Demographic aspects of long- distance migration. Pages 493-504 in Migrant birds in the Neotropics: ecology, behavior, distri- bution, and conservation (A. Keast and E. S. Mor- ton, Editors). Smithsonian Institution Press, Wash- ington, D.C., USA. Greenwood, P. J. 1980. Mating systems, philopatry, and dispersal in birds and mammals. Animal Be- havior 28:1140—1162. Haas, C. 1990. Breeding ecology and site fidelity of American Robins, Brown Thrashers, and Logger- head Shrikes in shelterbelts in North Dakota. Dis- Renner and McCajfery • EASTERN YELLOW WAGTAILS IN ALASKA 91 sertation. Cornell University, Ithaca, New York, USA. Haas, C. 1998. Effects of prior nesting success on site fidelity and breeding dispersal: an experimental approach. Auk 115:929-936. Johnson, D. H. 1979. Estimating nest success: the Mayfield method and an alternative. Auk 96:651- 661. Klett, a. T, H. E Duebbert, C. A. Eaanes, and K. E Higgins. 1986. Techniques for studying nest success of ducks in upland habitats in the prairie pothole region. USDI, Fish and Wildlife Service, Resource Publication 158. Washington, D.C., USA. Kovshar, a. E 1979. The singing birds of the subal- pine belt of Tien Shan. Nauka, Alma-Ata, Ka- zakhstan. Lebreton, J. D., K. P. Burnham, J. Clobert, and D. R. Anderson. 1992. Modeling survival and test- ing biological hypotheses using marked animals: a unified approach with case studies. Ecological Monographs 62:67-118. Martin, T. E. and J. Clobert. 1996. Nest predation and avian life history evolution in Europe versus North America: a possible role of humans? Amer- ican Naturalist 147:1028-1046. Mayheld, H. E 1960. Nest success calculated from exposure. Wilson Bulletin 73:255-261. Mayfield, H. E 1975. Suggestions for calculating nest success. Wilson Bulletin 87:456-466. McClure, H. E. 1998. Migration and survival of the birds of Asia. White Lotus Co., Ltd, Bangkok, Thailand. Michel, N., D. E DeSante, D. R. Kaschube, and M. P. Nott. 2006. The Monitoring Avian Productivity and Survivorship (MAPS) Program annual re- ports, 1989-2003. NBII/MAPS Avian Demo- graphics Query Interface. http://www.birdpop. org/nbii/NBIIHome.asp (accessed 18 December 2006). Nichols, J. D., C. Brownie, J. E. Hines, K. H. Pol- lock, AND J. B. Hestback. 1993. The estimation of exchanges among populations or subpopula- tions. Pages 264-280 in Marked individuals in the study of bird populations (J. D. Lebreton and P. M. North, Editors). Birkhauser-Verlag, Basel, Switzerland. Nur, N. and W. j. Sydeman. 1999. Demographic pro- cesses and population dynamic models of sea- birds. Current Ornithology 15:149-188. Paradis, E., S. R. Baillie, W. J. Sutherland, and R. D. Gregory. 1998. Patterns in natal and breeding dispersal in birds. Journal of Animal Ecology 67: 518-536. Pollock, K. H., S. R. Winterstein, and M. J. Con- roy. 1989a. Estimation and analysis of survival distributions for radio-tagged animals. Biometrics 45:99-109. Pollock, K. H., S. R. Winterstein, C. M. Bunck, and P. D. Curtis. 1989b. Survival analysis in teleme- try studies: the staggered entry design. Journal of Wildlife Management 53:7-15. Renner, H. M. and B. J. McCaffery. 2006. Nesting biology of Eastern Yellow Wagtails at Cape Ro- manzof, Alaska. Journal of Field Ornithology 77: 1-9. Ricklefs, R. L. 1973. Fecundity, mortality, and avian demography. Pages 366-435 in Breeding biology of birds (D. S. Earner, Editor). National Academy of Science, Philadelphia, Pennsylvania, USA. White, G. C. and K. P. Burnham. 1999. Program MARK: survival estimation from populations of marked animals. Bird Study 46(Supplement):120- 138. Wood, B. 1976. The biology of Yellow Wagtails Mo- tacilla flava L. overwintering in Nigeria. Disser- tation. University of Aberdeen, Aberdeen, United Kingdom. Wood, B. 1979. Changes in numbers of over-wintering Yellow Wagtail Motacilla flava and their food supplies in a West African savanna. Ibis 121:228- 231. The Wilson Journal of Ornithology 120(l):92-98, 2008 BREEDING ECOLOGY OF THE NARCISSUS FLYCATCHER IN NORTH CHINA NING WANG,' YANYUN ZHANG,' AND GUANGMEI ZHENG'-^ ABSTRACT. We studied the breeding ecology of the Narcissus Flycatcher (Ficedula narcissina elisae) in subalpine secondary broad-leaf forest near Beijing, China during 2003-2006. The Narcissus Flycatcher arrived in the breeding area at the beginning of May with first-spring males arriving later than older males. Nests of first-spring male Narcissus Flycatchers were at lower altitudes (1,271.0 ± 20.7 m) than those of older males (1,363.8 ± 22.5 m). Narcissus Flycatchers nested in diverse sites on different plant species with Betula dahurica being most commonly used. Nests of Narcissus Flycatchers were exposed and 2.5 ± 0.3 m above the ground with mean clutch size of 4.3 ± 0.2. Mean egg length was 17.2 ± 0.1 mm, mean egg width was 13.2 ± 0.1 mm. and mean egg mass was 1.5 it 0.0 g. The mean incubation period was 13.4 i 0.2 days and mean nestling period was 13.2 ± 0.2 days. Fledging success was 4.0 ± 0.3 fledglings per successful nest. The Narcissus Flycatcher had low breeding success (51.2%), mainly because of predation. Received 28 November 2006. Ac- cepted 16 July 2007. The Narcissus Flycatcher {Ficedula narcis- sina) is a gender-dimorphic small passerine that breeds in northeast Asia and winters in southeast Asia. One subspecies (F. n. elisae) breeds only in north China (Cheng 1987, Dickinson 2003, Zheng 2005) and was con- sidered to be a distinct species based on mor- phological, acoustic, and molecular evidence (Weigold 1922, Li 2004, Zhang et al. 2006). This subspecies has delayed plumage matu- ration in first-spring males (Zhang et al. 2006). The plumage of first-spring males is similar to females instead of older males; these female-looking males sing like adult males and can breed successfully. The first- spring male was even mistakenly identified as a new species: Ficedula beijingnica (Zheng et al. 2000). Little information exists on the breeding ecology of Ficedula narcissina eli- sae and our objectives are to present data on the (1) reproductive phenology; (2) character- istics of nests, eggs, and nestlings; (3) repro- ductive behavior; and (4) breeding success of this subspecies. METHODS Study Area. — We studied Narcissus Fly- catchers in the subalpine Xiao Longmen For- est (40°00'N, 115°26'E), 114 km west of ’ College of Life Science, Ministry of Education Key Laboratory for Biodiversity Science and Ecolog- ical Engineering, Beijing Normal University, Beijing, 100875, China. 2 Corresponding author; e-mail: zhenggm@bnu.edu.cn Beijing, China. This forest is 705.4 ha in size at 1,000-1,763 m altitude. The terrain is rug- ged with steep slopes (30-60°) and the width of drainages ranges from 10 to 60 m. The mean annual temperature of the study area is —4.8 °C and the annual precipitation is 500- 700 mm. The vegetation is dominated by sec- ondary temperate deciduous broad-leaf forest with patches of coniferous plantations. The broad-leaf forest is dominated by Populus davidiana, P. cathayana, Salix viminalis, S. caprea, Juglans mandshurica, Betula platy- phylla, B. dahurica, Quercus mongolica, Acer truncatum, Fraxinus rhynchophylla, and Sy- ringa pekinensis. Conifer plantations are dom- inated by Larix principis-rupprechtii, L. kaempferi, and Pinus tabulaeformis . Local people graze goats, cut firewood, and collect herbs, edible plants, and mushrooms in the forest. Our study site was 196.5 ha of largely broad-leaf trees in the southern part of the for- est. Nest-box Arrangement. — We erected 200 nest boxes (8.5 X 8.5 X 17.5 cm) in the study area from 2004 to 2006 before flycatchers ar- rived in spring. Half of the nest boxes had a large square entrance (9 X 8.5 cm) and the other half had a small round entrance (3. 5 -cm diam). Nest boxes were placed 3-4 m above ground on live tree trunks on both slopes of a drainage with entrances facing down slope. We collected data from May to July in 2004, 2005, and 2006 after a preliminary study from June to July 2003. Breeding Surveys. — We searched for and 92 Wang et al. • BREEDING ECOLOGY OF NARCISSUS FLYCATCHER 93 TABLE 1 . Arrival dates of Narcissus and Yellow-rumped flycatchers in Xiao Longmen Forest, China. Narcissus Flycatcher Male Yellow-rumped Flycatcher Year Older First-spring Female Male Female 2004 5 May 13 May 8 May 7 May 8 May 2005 3 May 15 May 6 May 6 May 9 May 2006 3 May 17 May 10 May 5 May 1 1 May listened for songs of flycatchers every day each year starting on I May. After locating a bird, we tried to ascertain its species, gender, and age if it was a male. Older male Narcissus Flycatchers (>2 years) have yellow supercil- iums, yellow rumps, and white patches on the greater coverts. First-spring males are without yellow superciliums and yellow rumps, and instead of white patches on greater coverts, they have two white wing bars. The date we found the first individual each year was used as the arrival date for that year. Nests Searching and Monitoring. — We lo- cated nests by following adults displaying nesting behavior such as entering a hole or crevice, carrying plant fibers for nest-building, and taking invertebrates to feed nestlings, or we searched the vicinity of where adults were giving alarm calls. We monitored nests by re- cording date of the first egg, clutch size, date of hatching, date of fledging, and calculated reproductive success. Length, width, and mass of eggs were measured on day of laying. A breeding attempt was considered successful if at least one young fledged. We backdated egg- laying dates for nests found during incubation or after nestlings had hatched based on clutch- es with complete information. We calculated the duration between the arrival date and lay- ing date of each clutch for statistical analyses. Nest Site Characteristics. — We measured characteristics of nest sites after breeding ceased, including altitude, aspect, nesting tree species, height of the nest above ground, and dimension of the cavity where the nest was located. Characteristics of each nest site were included only once in statistical analyses even if the nest site was used several times in dif- ferent years. Statistical Analyses. — We used ,SPSS for Windows 1 1.0 for all analyses. The normality of data was examined using the one sample Kolmogorov-Smirov test. We compared means by t-tests for data with normal distri- bution and calculated Pearson’s correlation coefficients. Equality of variances was tested by Levene’s test. We used Mann-Whitney U-tQsts and Spearman’s correlation coeffi- cients for data that did not show normal dis- tribution. Data are presented as T ± SE, with maximum and minimum values. Significance was set at P = 0.05. RESULTS Arrival Dates. — Both Narcissus Flycatchers and Yellow-rumped Flycatchers {Ficedula zanthopygia) occurred on our study area. Nar- cissus Flycatchers arrived in the beginning of May with males arriving 3-7 days earlier than females, while first-spring males arrived 8-12 days later than older males. Yellow-rumped Flycatchers arrived 2-3 days later than Nar- cissus Flycatchers and males arrived 1-6 days earlier than females (Table 1). Nest Site Characteristics. — We found 43 Narcissus Flycatcher nests, 2 in 2003, 17 in 2004, 7 in 2005, and 17 in 2006. Fourteen nests were found during nest-building, 5 dur- ing egg-laying, 12 during incubation, 5 during the nestling stage, and 7 just after fledging when parents were still feeding their young nearby. Seven nests had first-spring males. Narcissus Flycatchers nested in tree forks formed by upward reaching branches (10 nests), hollows at the top of stakes ( 1 1 nests), cavities on tree trunks (13 nests), and in nest boxes with larger square entrances (9 nests). Both first-spring and older males used the four types of nest-sites with no difference detected between age groups (Chi-square test, X“ = 2.525, df = 3, P = 0.471). None of the nest boxes with small round entrances was used by Narcissus Flycatchers. Some nest sites were used repeatedly: one cavity in a tree trunk was used in 2003, 2004, and 2006; another cavity in a tree trunk was used in 2004 and 2006; 94 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 one hollow at the top of a stake was used in 2004 and 2006. We found 27 Yellow-rumped Flycatcher nests, 4 in 2003, 7 in 2004, 7 in 2005, and 9 in 2006. Nine nests were found during nest- building, 2 during egg-laying, 10 during in- cubation, and 6 during nestling stages. Yel- low-rumped Flycatchers nested in cavities in tree trunks (11 nests), nest boxes with small round entrances (15 nests), and one nest was in a cavity on a brick wall. None of the nest boxes with larger square entrances was used by Yellow-rumped Flycatchers. One cavity in a tree trunk was used in 2003 and 2004. Cav- ities in tree trunks used as nest sites by Yel- low-rumped Flycatchers significantly differed from those used by Narcissus Flycatchers both in the smallest diameter (Yellow-rumped Fly- catcher: 4.6 ± 0.3 cm, range = 4.0-6. 0 cm, = 10; Narcissus Flycatcher: 6.7 ± 0.5, range = 3. 5-8.0 cm, n = 9: Mann- Whitney f/-test, Z = -2.583, P = 0.010) and in the largest diameter (Yellow-rumped Flycatcher: 5.5 ± 0.6 cm, range = 4.0— 9.0 cm, n = 10; Narcis- sus Flycatcher: 18.5 ± 6.9, range = 5.0-72.0 cm, n = 9; Mann- Whitney f/-test, Z = -3.052, P = 0.002). Narcissus Flycatchers nested at higher ele- vation (1,347.1 ± 19.6 m, range = 1,150- 1,577 m, n = 39) than Yellow-rumped Fly- catchers (1,237.0 ± 21.8 m, range = 1,070- 1,473 m, ^ = 26; Independent-samples Mest: t = 3.688, df = 63, P < 0.001). First- spring male Narcissus Flycatchers nested at lower elevation than older males (first-year males: 1,271.0 ± 20.7 m, range = 1,173- l, 350 m, n = 1- older males: 1,363.8 ± 22.5 m, range = 1,150—1,577 m, n = 32; Indepen- dent-samples r-test: t = —3.032, df = 31, P = 0.006). Twenty-five of 39 Narcissus Flycatcher nests were on north-facing slopes, 5 were on south-facing slopes, and 9 were in drainages (Chi-square test, ~ 17.231, df = 2, P < 0.001). First-spring males nested on north-fac- ing slopes {n = 4) and in drainages {n = 3). Twelve of 26 Yellow-rumped Flycatcher nests were on north-facing slopes, 3 were on south- facing slopes, and 1 1 were in drainages (Chi- square test, “ 5.615, df = 2, P = 0.060). No difference (Chi-square test, x^ = 5.517, df = \ P = 0.063) was detected between Yel- low-rumped and Narcissus flycatchers in slopes chosen for nesting. Narcissus Flycatchers nested lower above ground (2.5 ± 0.3 m, range = 0.4-8.0 m, n = 39) than Yellow-rumped Flycatchers (4.2 ± 0.3 m, range = 2.2-7.0 m, ^ = 26; Indepen- dent-samples Ltest, t = —4.299, df = 63, P < 0.001). Nests of first-spring male Narcissus Flycatchers were at the same height as older males (Independent-samples r-test, t = -0.604, df = 37, P = 0.55). Narcissus Flycatchers nested on nine plant species (Table 2), with 20 of 39 nests on Bet- ula dahurica (Chi-square test, x^ — 67.846, df = 8, P < 0.001). First-spring males nested on five plant species with these species used also by older males. Sixteen of 39 Narcissus Fly- catcher nests were on dead trees or stakes (Chi-square test, x^ = 1.256, df = 1, P = 0.26). One of seven nests of first-spring males was on a stake with no difference (Chi-square test, x^ = 0.2.982, df = 1, P = 0.084) detected between age groups. Yellow-rumped Flycatchers also nested on nine plant species (Table 2), but only three of these species (Juglans mandshurica, Betula platyphylla, and B. dahurica) were used by Narcissus Flycatchers. Nine of 25 nests of Yellow-rumped Flycatcher (excluding the nest on brick wall) were on dead trees (Chi-square test, x^ = 1-960, df = 1, P = 0.16), not dif- ferent (Chi-square test, x^ ^ 0.262, df = 1, P = 0.608) from that recorded for the Narcissus Flycatcher. Egg-laying, Clutch Size, and Egg Charac- teristics.—T\\q earliest Narcissus Flycatcher clutch was laid in the second half of May and the latest was laid in the beginning of July. We could not confirm whether second clutches were laid after completion or failure of the first nesting attempt in a breeding season. The mean duration between the arrival date of Narcissus Flycatchers and date of the first egg of every clutch that year was 34.3 ± 2.5 days (range = 16-54 days, n = 21). The mean clutch size of Narcissus Flycatcher was 4.3 ± 0.2 (range = 3-6, n ^ 19). Two clutches had 3 eggs, 10 had 4 eggs, 6 had 5 eggs, and 1 had 6 eggs. Clutch size of first-spring males did not differ from that of older males (Inde- pendent-samples r-test, t = —0.192, df = 17, P = 0.85). There was a negative correlation between clutch size and date of the first egg Wang et al. • BREEDING ECOLOGY OF NARCISSUS FLYCATCHER 95 TABLE 2. Plants used by Narcissus and Yellow-rumped flycatchers for nesting in Xiao Longmen Forest, China. Narcissus Flycatcher Yellow-rumped Flycatcher Plant species Live trees Dead trees Totals Live trees Dead trees Totals Pinus tabulaeformis Populus davidiana 2 2 2 1 3 P. cathayana 2 2 Salix viminalis S. caprea 1 5 6 3 1 4 Juglans mandshurica 2 2 7 7 Betula platyphylla 1 2 3 2 2 B. dahurica 15 or 5 20 (3) 1 1 2 Corylus mandshurica Ulmus laciniata 1 (1) 1 (1) 1 1 U. propinqua Malus baccata 2 (1) 2 (1) 1 1 Acer truncatum Fraxinus rhynchophylla 2 (1) 2 (1) 1 1 Syringa pekinensis Abelia biflora 1 (1) 1 (1) 1 1 2 ^ Numbers of nests of first-spring male Narcissus Flycatchers. (Pearson’s correlation coefficient r = —0.748, P < 0.001). The color of Narcissus Flycatcher eggs ranged from whitish pink to light blue with rusty speckles. Mean egg length was 17.2 ±0.1 mm (range = 14.8-18.6 mm, n — 38) with mean width of 13.2 ± 0.1 mm (range - 12.0-13.8 mm, n = 38). The mean mass of Narcissus Flycatcher eggs was 1.5 ± 0.0 g (range -- 1.2-1. 7 g, n = 23). The mean clutch size of Yellow-rumped Flycatchers was 4.8 ± 0.2 (range -- 4-6, n 1 0). Three of the clutches had four eggs, six had five eggs, and one had six eggs. The clutch size of Yellow-rumped Flycatchers did not differ (Independent-samples r-test, t = -1.740, df = 21, P = 0.093) from that of Narcissus Flycatchers. The color of Yellow- rumped Flycatcher eggs was similar to that of Narcissus Flycatcher eggs. Mean egg length was 17.1 ± 0.2 mm (range == 15.0-17.9 mm, n = 16) and mean width was 13.1 ±0.1 mm (range = 12.8-13.5 mm, n = 16). Neither measurement differed (Independent-samples r-test, t = 0.495, df = 52, P = 0.62 for length; r = 0.661, df = 52, P = 0.51 for width) from those of Narcissus Flycatchers. The mean mass of Yellow-rumped Flycatcher eggs was 1.4 ± 0.0 g (range = 1.1-1. 6 g, n = 15), slightly less than those of Narcissus Flycatch- ers (Independent-samples /-test, / = 2.824, df = 36, P = 0.008). Incubation and Parental Care. — Incubation of Narcissus Flycatcher nests was by the fe- male alone and males were observed feeding their incubating mate. Some females initiated incubation before their clutches were com- plete and this most likely led to asynchronous hatching (within 2 days) observed in several clutches. The mean incubation period was 13.4 ± 0.2 days (range = 12-14 days, n = 17). Length of incubation was negatively cor- related to laying date (Pearson’s correlation coefficient, r — —0.499, P = 0.049) and pos- itively with clutch size (Pearson’s correlation coefficient, r = 0.699, P = 0.002). Mean sur- vival of nests to hatching was 0.91 ± 0.04 (range = 0.60-1.00, n = 17). Mean hatching success was 3.9 ± 0.2 nestlings (range = 2-6 nestlings, n = 17) with brood size posi- tively correlated to clutch size (Pearson’s cor- relation coefficient, r = 0.637, P = 0.008). There was no difference between nests of first-spring versus older males in incubation periods, survival of nests to hatching, and hatching success (Independent-samples r-test, all P >0.05). Both male and female Narcissus Flycatcher fed their young. The mean nestling period was 13.2 ± 0.2 days (range = 12-14 days, n = 13). Length of nestling period correlated neg- atively to laying date (Spearman's correlation coefficient, r = —0.657, P = 0.015) and pos- 96 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 itively to brood size (Spearman’s correlation coefficient, r = 0.853, P < 0.001). All nest- lings from the same clutch fledged in the same morning. Fledging success was 4.0 ± 0.3 fledglings (range = 3-6 fledglings, « = 13) per successful nest. There was no difference (Mann-Whitney U-test, all P >0.05) between first-spring and older males in nestling periods and numbers of fledglings per successful nest. Breeding Success. — Twenty-two of 43 Nar- cissus Flycatcher nests in Xiao Longmen For- est successfully fledged young, and the pro- portion of successful and failing nests was nearly the same (Chi-square test, = 0.023, df = I, P = 0.88). Three of seven nests of first-spring males succeeded and the propor- tion of successful nests did not differ from that of older males (Chi-square test, ^ 0.281, df = 1, F = 0.596). Five nests failed during nest-building, 4 during egg-laying, 6 during incubation, and 6 during the nestling stage. Predation accounted for failure of 17 nests. Likely predators included Swinhoe’s striped squirrel (Tamiops swinheoi), rock squirrel {Sciurotamias davidianus), Siberian weasel (Mustela sibirica), Japanese Sparrow- hawk (Accipiter gularis). Great Spotted Woodpecker (Picoides major), Eurasian Jay {Garrulus glandarius). Red-billed Blue Mag- pie (Urocissa erythrorhyncha), and Chifeng beauty snake {Elaphe anomala). Other causes for nest failure were storms (2 nests), human disturbance (1 nest), and ants taking over the cavity (1 nest) after an egg was laid. Twenty of 27 nests of Yellow-rumped Fly- catcher in Xiao Longmen Forest successfully fledged young and the proportion of success- ful nests was higher than that of failing nests (Chi-square test, x^ ~ 6.259, df = 1, P = 0.012). Yellow-rumped Flycatchers had higher breeding success than Narcissus Flycatcher (Chi-square test, x^ ^ 5.679, df = \, P = 0.017). Four nests failed during nest-building, 1 during egg-laying, 1 during incubation, and 1 during the nestling stage. Two nests were depredated, 2 were destroyed by humans, 1 was taken over by wasps during the nest- building stage, 1 was destroyed by a storm, and 1 was abandoned after an egg was laid. Molt. — Fledgling Narcissus flycatchers were olive green with light yellow speckles on the dorsal side and light yellow with olive green streaks on the ventral side. Remiges were olive brown, and wing coverts were ol- ive brown with light yellow tips, which formed horizontal bars on greater and middle coverts. Rectrices were olive brown. Fledg- lings began their post-natal molt about a month after fledging. Remiges, wing coverts, and rectrices were not involved in this molt. The post-natal molt ended at the beginning of September. Juvenile males and females were alike in plumage, both olive green above and light yellow below. Males which returned to the breeding area in their second calendar year (first-spring) still looked like females. First-spring males began their first postnuptial molt at the end of June when their contour feathers, remiges, wing co- verts, and rectrices went through a complete molt. Primaries molted descendantly and sec- ondaries molted ascendantly. Rectrices molted centrifugally. Males obtained their typical adult plumage in early September of their second cal- endar year with head to back olive green and with yellow superciliums and mmp. The under- parts were yellow, remiges and wing coverts were blackish brown with white patches on greater coverts, and rectrices were blackish brown. Female plumage color did not change noticeably after their first postnuptial molt. DISCUSSION Arrival Dates. — Older Narcissus Flycatcher males arrived earlier than young males similar to European Pied Flycatchers {Ficedula hy- poleuca) (Potti and Montalvo 1991, Potti 1998a), Collared Flycatchers (F. albicollis) (Mitrus et al. 1996, Mitrus 2004), and Red- breasted Flycatchers (F. parva) (Mitrus 2006, Mitrus et al. 2006). Delayed settlement of young males may result from delayed depar- ture from wintering areas or delayed matura- tion. Males which bred in previous seasons return to their known breeding places (Part and Gustaffsson 1989) while young males may have to spend time searching for suitable breeding sites. Male Narcissus and Yellow-rumped fly- catchers both arrived earlier than females sim- ilar to Pied, Collared, and Red-breasted fly- catchers (Dement’ ev and Gladkov 1968). This suggests pair bonding in both species occurred after arrival in breeding areas. We observed males of both Narcissus and Yellow-rumped flycatchers displaying to females by going in Wang et al. • BREEDING ECOLOGY OE NARCISSUS ELYCATCHER 97 and out of their nest holes followed by the females entering the holes to inspect the nest sites similar to that reported for European Pied Flycatcher (Rinden et al. 1995). This type of nest-site display suggests that quality of the nest site occupied by a male is an important criterion in mate selection. Nest Sites Characteristics. — Narcissus Fly- catchers nested at higher altitudes than Yel- low-rumped Flycatchers in Xiao Fongmen Forest indicating separation in habitat use by elevation. Narcissus Flycatchers near Beijing bred in subalpine forest while Yellow-rumped Flycatchers bred in low elevation woodlands (Zheng 1984, Cai 1987). Replacement of con- geners by elevation was observed in other passerines near Beijing (Zheng 1984) includ- ing Marsh (Poecile palustris) and Willow tits (P. montana). Meadow (Emberiza cioides) and Godlewski’s buntings {E. godlewskii). Black {Dicrurus macrocercus) and Hair-crest- ed drongos {D. hottentottus), and White- throated (Monticola gularis) and Blue Rock thrushes (M. solitarius). Breeding habitat of Narcissus Flycatchers in our study, where the two flycatchers were sympatric, was dominat- ed by Betula dahurica while the breeding hab- itat of Yellow-rumped Flycatcher was domi- nated by Juglans mandshurica (Wang et al. 2006). This implies partitioning in microhab- itat use by the two species when they were sympatric. Narcissus Flycatchers used more diverse and exposed nest sites closer to the ground than those used by Yellow-rumped Flycatch- ers and other typical secondary hole-nesting passerines such as tits {Poecile spp.) and nut- hatches {Sitta spp.) that bred sympatricly in Xiao Fongmen Forest. This is similar to sym- patricly-breeding Red-breasted and European Pied flycatchers, and other secondary cavity nesters in Poland (Mitrus and Socko 2004) where Red-breasted Flycatchers used more open, lower, and more diverse-shaped nest cavities than other species. In Europe, Euro- pean Pied Flycatchers nested in higher holes with smaller entrances in trees than sympa- tricly breeding Collared Flycatchers (Czesz- czewik and Walankiewicz 2003). The separation between Narcissus and Yel- low-rumped flycatchers in nest sites could be affected by inter-specific conflicts. We ob- served male Yellow-rumped Flycatchers expel male Narcissus Flycatchers near nest sites of the former, suggesting Yellow-rumped Fly- catchers are dominant in inter-specific con- flicts. Accepting diverse nest sites could max- imize the possibility that subdominant Narcis- sus Flycatchers may breed. Fower breeding success of Narcissus Flycatchers suggested that subdominant species may use suboptimal nest sites. Breeding Behaviors and Reproductive Pa- rameters.— Fate breeding Narcissus Flycatch- er pairs had smaller clutches, shorter hatching periods, fewer nestlings, and shorter nestling periods than pairs that bred early. This is con- sistent with patterns reported for European Pied Flycatchers breeding in temperate areas (Jarvinen and Vaisanen 1983, Alatalo and Fundberg 1984, Moreno and Carlson 1989, Sanz 1996, Eeva et al. 2000). Short breeding seasons in temperate subalpine areas such as Xiao Fongmen Forest compel breeding birds to complete their reproductive cycle within a short time, particularly for late breeders. Most passerines that breed in Xiao Fongmen Forest fledge their young before late July. Male Narcissus Flycatcher fed their mates during incubation and brooding. This was also reported for European Pied Flycatchers (Sanz 1996). Male provisioning could decrease the frequency and duration when a female was ab- sent from the nest (Moreno and Carlson 1989), so that continuous warming of the eggs was possible. In subalpine areas with low tem- peratures near dawn, such as in Xiao Fong- men Forest, it may be important that females remain on their nests. Asynchronous hatching within a Narcissus Flycatcher clutch is similar to European Pied Flycatchers (Potti 1998b). The proportion of asynchronous hatching for European Pied Fly- catchers was higher in better habitat and milder weather than in poor habitat or severe weather (Slagsvold 1986, Slagsvold and Litjeld 1989). It also varied with apparent quality of the female (Potti 1998b). ACKNOWLEDGMENTS We thank the managers of Xiao I.ongmen Forest for accommodation, J. Zhang, Y. Sun, Q. L. Ma, and L. Dt)iig for field assistance. We also thank L. L. Sever- inghaus for improving the English of the draft, editor C. Fi. Braun for proofreading the Faiglish of the man- uscript, and two anonymous reviewers for commenting on the manuscript. Financial support was provided by 98 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 National Sciences Foundation of China (Grant 30170115). LITERATURE CITED Alatalo, R. V. AND A. Lundberg. 1984. Density-de- pendence in breeding success of the Pied Fly- catcher {Ficedula hypoleuca). Journal of Animal Ecology 53:969-977. Cai, Q. K. 1987. Birds of Beijing. Beijing Press, Bei- jing, China. Cheng, T. H. 1987. A synopsis of avifauna of China. Science Press, Beijing, China. CzESZCZEWiK, D. AND W. Walankiewicz. 2003. Natu- ral nest sites of the Pied Flycatcher Ficedula hy- poleuca in a primeval forest. Ardea 91:221-229. Dement’ev, G. P. and N. A. Gladkov. 1968. Birds of the Soviet Union. Volume 4. Israel Program for Scientific Translations Ltd., Jerusalem, Israel. Dickinson, E. 2003. The Howard and Moore complete checklist of the birds of the world. Third Edition. Christopher Helm, London, United Kingdom. Leva, T, S. Veistola, and E. Lehikoinen. 2000. Tim- ing of breeding in subarctic passerines in relation to food availability. Canadian Journal of Zoology 78:67-78. Jarvinen, a. and R. a. Vaisanen. 1983. Egg size and related reproductive traits in a southern passerine Ficedula hypoleuca breeding in an extreme north- ern environment. Ornis Scandinavica 14:253-262. Li, W. 2004. Studies on the molecular phylogeny and taxonomic status of part of the genus Ficedula. Thesis. Beijing Normal University, Beijing, Chi- na. Mitrus, C. 2004. No effect of age of males on repro- ductive success of the Collared Flycatcher Fice- dula albicollis. Folia Zoologica 53:319-328. Mitrus, C. 2006. The influence of male age and phe- nology on reproductive success of the Red-breast- ed Flycatcher {Ficedula parva Bechst.). Annales Zoologici Fennici 43:358—365. Mitrus, C. and B. Socko. 2004. Natural nest sites of the Red-breasted Flycatcher Ficedula parva in a primeval forest. Acta Ornithologica 39:53-57. Mitrus, C., N. Kleszko, and B. Socko. 2006. Habitat characteristics, age, and arrival date of male Red- breasted Flycatchers Ficedula parva. Ethology, Ecology and Evolution 18:33-41. Mitrus, C., W. Walankiewicz, D. Czeszczewik, and P M. JabIonski. 1996. Age and arrival date of Collared Flycatcher Ficedula albicollis males do not influence quality of natural cavities used. Acta Ornithologica 31:101-106. Moreno, J. and A. Carlson. 1989. Clutch size and the costs of incubation in the Pied Flycatcher Fi- cedula hypoleuca. Ornis Scandinavica 20:123- 128. Part, T. and L. Gustafsson. 1989. Breeding dispersal in the Collared Flycatcher {Ficedula albicollis): possible causes and reproductive consequences. Journal of Animal Ecology 58:305-320. PoTTi, J. 1998a. Arrival time from spring migration in male Pied Flycatchers: individual consistency and familial resemblance. Condor 100:702-708. PoTTi, J. 1998b. Variation in the onset of incubation in the Pied Flycatcher {Ficedula hypoleuca): fitness consequences and constraints. Journal of Zoology 245:335-344. PoTTi, J. AND S. Montalvo. 1991. Male arrival and female mate choice in Pied Flycatchers Ficedula hypoleuca in central Spain. Ornis Scandinavica 22:45-54. Rinden, H., S. Dale, and T. Slagsvold. 1995. Male display behaviour and the assessment of male quality during visits by female Pied Flycatchers Ficedula hypoleuca. Fauna Norvegica, Series C: Cinclus 18:107-117. Sandz, j. j. 1996. Effect of food availability on incu- bation period in the Pied Flycatcher {Ficedula hy- poleuca). Auk 113:249-253. Slagsvold, T. 1986. Asynchronous versus synchro- nous hatching in birds: experiments with the Pied Flycatcher. Journal of Animal Ecology 55:1115— 1134. Slagsvold, T. and J. T. Lifjeld. 1989. Constraints on hatching asynchrony and egg size in Pied Fly- catchers. Journal of Animal Ecology 58:837-849. Wang, N., Y. Y. Zhang, and G. M. Zheng. 2006. Home ranges and habitat vegetation characters in breeding season of Narcissus Flycatcher and Yel- low-rumped Flycatcher. Journal of Beijing Normal University (Natural Science) 42:295-299. Weigold, H. 1922. Muscicapa elisae n. sp. Falco 18: 1-2. Zhang, Y. Y, N. Wang, J. Zhang, and G. M. Zheng. 2006. Acoustic difference of Narcissus Flycatcher complex. Acta Zoological Sinica 52:648-654. Zheng, G. M. 1984. Ecological distribution of birds in Beijing and its vicinity during summer. Chinese Zoological Research 5:29-40. Zheng, G. M. 2005. A checklist on the classification and distribution of the birds of China. Science Press, Beijing, China. Zheng, G. M., J. Song, Z. W Zhang, Y. Y. Zhang, AND D. S. Guo. 2000. A new species of flycatcher Ficedula from China (Aves: Passeriformes-Mus- cicapidae). Journal of Beijing Normal University (Natural Science) 36:405-409. The Wilson Journal of Ornithology 120(1):99-104, 2008 REPRODUCTIVE SUCCESS OE HOUSE WRENS IN SUBURBAN AND RURAL LANDSCAPES MICHAEL J. NEWHOUSE,' PETER P. MARRA.' ^ AND L. SCOTT JOHNSON^ ABSTRACT. — We investigated the impacts of urbanization on reproductive success of House Wrens {Trog- lodytes aedon). We compared reproductive effort and success for 33 nesting attempts in suburban sites (2.5-10 buildings/ha) and 43 nesting attempts in rural sites (<2.5 buildings/ha) in and around the Washington, D.C.- Baltimore, Maryland, metropolitan area. There were no differences in clutch initiation dates or clutch sizes between suburban and rural nests. However, nestlings at suburban nests weighed less and had smaller body size prior to fledging compared to nestlings at rural nests. Parental feeding rates differed between suburban and rural nests during the “early nestling stage” (day 3 to day 6), but not in the “late nestling stage” (day 8 to day 12) suggesting average quality of prey for nestlings may be lower at suburban sites. Overall, suburban nests fledged more young than rural nests largely because of higher rates of nest predation on rural nests. Further research on how food availability and predation affects nesting success of House Wrens and other birds along urbanization gradients may provide important insights into impacts of urbanization on birds. Received 1 November 2006. Accepted 1 April 2007. Land development can directly or indirectly impact native bird populations. For example, development alters habitat thereby reducing available areas for breeding and associated ac- tivities. Developed areas can have higher lev- els of human disturbance (Geise 1996, Beale and Monaghan 2004) and more chemical con- taminants than rural areas (Obendorf et al. 2006; K. E. Roux and R P. Marra, pers. comm.), which can indirectly impact birds through ingestion of contaminated food (e.g., fruit, phytophagous insects). Native bird spe- cies often rely on structurally complex habi- tats with a high proportion of native vegeta- tion for nest sites and food making them par- ticularly vulnerable in developed areas where invasive species often dominate and vegeta- tion structure has been changed drastically (Goldstein et al. 1986, Mills et al. 1991, Case 1996, Pavlik and Pavlik 2000, Hennings and Edge 2003). Finally, as natural areas become fragmented by urbanization the percentage of edge versus interior habitat also increases re- sulting in birds nesting in edge habitats ex- periencing higher rates of nest predation ' Smithsonian Environmental Research Center, Edgewater, MD 21037, USA. 2 Department of Biological Sciences, Towson Uni- versity, Towson, MD 21252, USA. ^ Current address: Smithsonian Migratory Bird Cen- ter, National Zoological Park, Washington, D.C. 20008, USA. ■’ Corresponding author; New Jersey Meadowlands Commission, Lyndhurst, NJ 07071, USA; e-mail: michael.newhouse@njmeadowlands.gov (Robinson et al. 1995, Donovan et al. 1997, Donnelly 2002, Donnelly and Marzluff 2004). Despite the challenges that affect birds in developed environments, habitats in urban/ suburban areas may attract birds by presenting them with historically reliable cues to suitable nesting habitat. Patches of apparently suitable breeding habitat that remain in developed en- vironments may become “ecological traps” (i.e., habitats that lure birds to settle, but result in lower overall fitness) (Gates and Gysel 1978, Purcell and Vemer 1998, Donovan and Thompson 2001, Kokko and Sutherland 2001, Schlaepfer et al. 2002, Weldon and Haddad 2005). Developed areas may be population sinks — areas occupied by breeding birds but with individuals producing too few surviving young to sustain the local population (e.g., Misenhelter and Rotenberry 2000, Resumes 2000). Few studies have examined the impacts of urbanization on bird reproductive success and on individual fitness. Most research has ex- amined how rates of nest depredation vary along urban to rural land-use gradients or with increasing forest fragmentation (e.g., Robin- son et al. 1995, Patten and Bolger 2003). The general conclusion from these studies is that depredation increases with proximity to forest edge. Leston and Rodewald (2006) examined reproductive success of Northern Cardinals (Carclinali.s carclinalis) in forests within urban and suburban land-use types: they found these sites provided suitable breeding habitat and 99 100 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 were not ecological traps. However, response to the effects of urbanization may vary be- tween species. We examined the reproductive success of House Wrens {Troglodytes aedon) nesting in the rural and suburban habitat matrix. Our pri- mary objective was to ascertain whether sub- urban development negatively impacted re- productive success of this cavity-nesting spe- cies. House Wrens are well-suited for study of the impacts of development on reproductive success because they prefer open woodlands and readily nest in suburban areas. METHODS Study Species.— WousQ Wrens are small (10-12 g), migratory, insectivorous, second- ary cavity-nesting songbirds that breed across much of North America (Johnson 1998). This species nests naturally in pre-formed tree cav- ities but readily uses nest boxes. Females nor- mally produce two broods each season, laying 6-8 eggs for first breeding attempts and 4-7 eggs for second attempts. Only females brood young but both adults feed nestlings. Study Sites and Nest Box Examination.— We conducted this study during spring and summer 2004 in and around the Washington, D.C. -Baltimore, Maryland, metropolitan area. The Washington, D.C. area has experienced a five-fold increase in human population density during the last century from 1.1 million resi- dents in 1900 to approximately 5.3 million in 2000 (Negative Population Growth 2005). In this study, 202 of 210 nest boxes avail- able to wrens were on properties (i.e., yards) owned by participants in the Neighborhood Nestwatch Program. This project is operated by the Smithsonian Migratory Bird Center (SMBC) of the National Zoological Park (http://www.fonz.org/nestwatch.htm). We classified properties as suburban or rural based on the density of human dwellings in the area surrounding a nest site (Cam et al. 2000, Marzluff 2001, Thorington and Bow- man 2003). We followed Marzluff (2001) and classified the habitat surrounding a nest as suburban if it had 2.5—10 dwellings/ha and ru- ral if there were <2.5 dwellings/ha. Several landowners had existing nest boxes and we added 65 additional nest boxes to properties prior to start of our study, distrib- uting them equally between suburban and ru- ral habitats. All added nest boxes were sus- pended from trees approximately 2. 5-3.0 m above ground and were not equipped with predator deterrent devices. Ultimately, each site contained an average of three nest boxes. Of the eight nests that were not part of the Nestwatch program, seven were at environ- mental education centers in rural Maryland while one was in a suburban neighborhood. Three of 91 boxes (3%) in suburban habitats and 83 of 119 boxes (69%) in rural habitats had predator deterrent devices. We checked nest boxes every 4-5 days be- ginning in mid-April until males present had paired. Thereafter, we visited each nest every other day to identify dates on which the first eggs were laid (all dates are reported as Julian dates). Female House Wrens usually lay one egg/day. We captured adults with mist nets and marked them with a USGS numbered alu- minum leg band and two colored bands. Males were captured after pairing. Females were not captured until their eggs had hatched to reduce risk of nest abandonment. No nests were abandoned immediately after female capture. We visited nest boxes 7-10 days after the first egg was laid to measure clutch size. We did not check nests daily after laying began and could not compare rates of partial clutch loss in suburban and rural habitats. We checked nests daily starting 1 1 days after lay- ing ended to identify the day that hatching be- gan. We termed the date of hatching Nestling Stage Day (NSD) 0. We banded, weighed nestlings, and measured both their wing chord and tarsus to the nearest 0.01 mm on NSD 1 1. We recorded the number of young fledged from a nest as the number of young present on NSD 1 1 minus the number of young found dead in the nest after fledging. We considered a nesting attempt successful if >1 nestling fledged from the nest. All nests that failed ap- peared to do so as a result of abandonment, depredation, conspecific nest destruction, or hypothermia. There were no nests in which nestlings disappeared gradually or as a result of apparent starvation. We assumed parental abandonment when eggs were present but un- attended. We assumed abandonment or hy- pothermia when all nestlings were found dead in the nest. We assumed predation when eggs or nestlings were missing and nest contents Newhouse et al • HOUSE WREN NESTING SUCCESS 101 were substantially disturbed. We also assumed predation when nest contents were not dis- turbed, eggs or young had disappeared, and the nest had not been occupied by a new male. Finally, we assumed conspecific nest destruc- tion had occurred when the nest was largely intact, all eggs or nestlings were missing from the nest, and a new male had occupied the nest. Parental Provisioning Behavior. — We re- corded the number of trips made by parents to deliver food to nestlings for one continuous hour between the 0700 and 1100 hrs EST. We monitored nests with binoculars from a dis- tance of 15-20 m. Data collection began when both parents appeared to be accustomed to our presence (typically about 5 min). We docu- mented provisioning rates at two points during the nestling stage, “early” (NSD 2 to NSD 6, n = 16) and “late” (NSD 8 to NSD 12, n = 24). Some nests {n = 11) were observed dur- ing both early and late nestling stages. Adults typically deliver a single prey item per trip; thus, the number of provisioning trips gener- ally equals the number of prey items trans- ported to the nest (Morrison and Johnson 2002). We also documented the total time that females spent brooding young. We assumed that time in the nest box >1 min involved brooding. All observations were at a similar time of season. We observed two parents provisioning nest- lings at most nests during the early nestling stage including 9 of 10 suburban and 7 of 1 1 rural nests. Only females were provisioning at the remaining nests. In contrast, we observed only one parent provisioning at 10 of 15 sub- urban nests (eight females, two males) and 9 of 10 rural nests (seven females, two males) during the latter part of the nestling stage. De- sertion of the nest by one parent, usually the male, is not uncommon in House Wrens (Czapka and Johnson 2000). Our intent was to use provisioning rates as an index of prey availability as the number of parents provi- sioning at a nest will influence feeding rates. Thus, we included only those early stage ob- servations in which two parents were provi- sioning and only those late-stage observations with only one parent provisioning in our anal- ysis. Statistical Analyses. — We compared the mean date the first egg was laid in the first 15 nests of the season in the two habitat types using a r-test to examine whether onset of breeding differed in suburban and rural habi- tats. We used analysis of covariance (AN- COVA) with date of first egg as a covariate to compare clutch size. We report least- squares means adjusted for effects of laying date. We used the same analysis to compare the number of young fledged from successful nests. We could not use ANCOVA to compare number of young fledged from all nests, suc- cessful and unsuccessful, because data were not normally distributed. Therefore, we com- pared numbers of fledglings from all nests us- ing a Wilcoxon Rank Sum test. We used the same test to compare number of fledglings per egg laid. We compared measures of nestling mass and size using a /-test, using mean brood values to avoid pseudoreplication. Broods tended to be smaller in rural nests and we also compared nestling measures using ANCOVA with brood size as a covariate and report least- squares means adjusted for brood size. We re- port interactions only if significant. We report means ± SE unless otherwise noted. All sta- tistical analyses were performed using SAS Version 8.1 (SAS Institute 1999). RESULTS We monitored reproduction at 76 nests (Ta- ble 1), 33 in suburban areas (17 first and 16 second broods) and 43 in rural habitats (24 first and 19 second broods). Birds in suburban and rural areas commenced breeding at the same time and laid similarly sized clutches. However, House Wrens in rural habitats were less successful {P < 0.001) at fledging at least one nestling than those in suburban habitats (60 vs. 88% of nests, respectively). Seven of the 17 (41%) unsuccessful nests in rural hab- itats were most likely destroyed by snakes and the remaining 10 were either destroyed by predators or by conspecifics. We could not clearly assign cause of failure in two of the four cases where nests failed in suburban hab- itats. In the remaining two cases, the entire brood was found dead suggesting either aban- donment or death from hypothermia. A great- er proportion of nests in rural compared to suburban habitats were destroyed by predators (0 vs. 41%; = 3.69, P = 0.05). This ex- plains, in part, why rural nests fledged signif- icantly fewer young per nest, on average, than 102 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 00 +1 C/D u o e ^ D OOCOOrOONCN^"^ r- o q — ; On q q (N ooooooooo V Os o r- (N ^ d d 'O cN o (N (N 1 in On (N (N I I N I ^ I 1 ^ 'vO ^ in m m c> cm (N (N ^ El ^ ^ ^ ' — ^ I — ^ ' ' (N ^ (M -H m ^ — ‘ d d — < (N d + 1 +1 O +1 +1 +1 +1 +1 +1 in d d in cl '^. Os O (N 00 m o d d Tt Os cc d d rc -H d d [1 d I — ^ I — I Os Os Os m Os , OX) o "O c 0 0-0 TD -O > 2 OX) OX) CL c c Oi) OX) T3 "O (U (U 3 u U Z fc tX) c/2 t/) W) o o w z z e - s ^ %o ^ c ^ O w C X (U suburban nests. Successful rural nests did not produce fewer fledglings than suburban nests after adjusting for seasonal effects on fledg- ling production. The mean number of fledg- lings per egg laid was also nearly identical for successful rural and suburban nests. The mean mass of nestlings on NSD 1 1 was greater in rural nests than in suburban nests (P < 0.02). The difference in nestling mass was significant even after controlling for brood size which were, on average, smaller in rural nests (Fi 52 = 4.67, P < 0.04). Similarly, nestlings had longer wing chords, on average, on NSD 1 1 in rural compared to suburban nests, even after controlling for brood size (Fi 52 = 5.97, P < 0.02). Tarsus length of nest- lings in rural and suburban nests did not differ. House Wren pairs early in the nestling stage delivered prey at a greater (P < 0.01) rate at suburban nests than at nests in rural areas (Ta- ble 2). In contrast, we found no difference in the rate at which parents provisioned older nestlings in suburban and rural habitats. Fe- males at rural nests spent more time brooding nestlings than females at suburban nests (P < 0.02; Table 2). Brooding time declined as pro- visioning rates increased in both habitats (r = -0.56, n ^ U, P = 0.03). DISCUSSION Reproductive success of House Wrens nest- ing in rural and suburban habitats varied with lower fledging success in rural habitats be- cause of greater rates of nest predation. Our results, however, were mixed. Nestlings reared in suburban areas weighed significantly less and had shorter wings compared to nestlings reared in rural areas. In an Illinois population of House Wrens, a nestling’s probability of be- ing recruited into the breeding population was positively associated with its body mass late in the nestling stage (C. F. Thompson, unpubl. data). Similar results have been reported for several species (Newton 1998). Our result that nests in rural areas were dep- redated at significantly higher rates compared to nests in suburban areas was striking given that far more rural than suburban nest boxes had predator-deterrent devices (69 vs. 3% of nest boxes, respectively). Rural nests may have been depredated significantly more often because most nest boxes were on the edge of large forest fragments facing more open land- Newhouse et al • HOUSE WREN NESTING SUCCESS 103 TABLE 2. Parental behavior (x ± SE, [«]) of House Wrens nesting in Washington, D.C.-Baltimore, Maryland metropolitan area in 2004. suburban and rural habitats in the Suburban Rural Statistic P-value Feeding rate, early^ 3.2 ± 0.3 [9] 2.0 ± 0.2 [7] 64 = -3.31 <0.01 Feeding rate, late*’ 4.0 ± 0.57 [10] 3.40 ± 0.6 [9] 6v = -0.79 0.44 Minutes brooding/hr 29.0 ± 4 [8] 43.0 ± 6 [6] 62 = 2.81 <0.02 Provisionings/nestling/hr, 1-4 days post-hatching. '’Provisionings/nestling/hr, 9-13 days post-hatching. scapes. Edge habitats are known to have high- er numbers of predators (e.g., Patten and Bol- ger 2003). Most rural nest boxes were older and had been in place longer than the subur- ban nest boxes, giving predators the chance to learn nest locations and how to elude predator deterrent devices. Differences in nestling body mass (Table 1) between habitats may be caused by differenc- es in prey quantity or quality. We have no data to address this hypothesis directly and are un- aware of existing data directly relevant to birds. Parents in suburban habitats appeared to deliver as many or more prey to nests com- pared to parents in rural habitats and yet pro- duced lighter, smaller offspring; this suggests that prey quality may be lower in suburban areas. Consistent with this observation are the findings of Ishitani et al. (2003) who reported that insects in suburban habitats tend to be smaller than species in more natural habitats. Females in suburban habitats also spent less time brooding nestlings during our watches. We did not observe any human disturbance while watching nests in both habitats. Nest- lings in suburban areas may grow more slowly because lower body temperatures cause them to process food less efficiently and invest more energy in thermoregulation at the ex- pense of growth (Johnson and Kermott 1993, Dawson et al. 2005). Highly disturbed environments, such as res- idential suburban habitats, may represent an ecological trap for some species. Leston and Rodewald (2006) failed to find any differenc- es in reproductive success or apparent survival of Northern Cardinals breeding in forest patches within urban versus rural landscapes. Our finding that House Wrens in suburban habitats produce lighter, smaller offspring sug- gests these habitats may act as ecological traps. However, counter to this suggestion is the finding that suburban nests also experi- enced significantly lower rates of nest preda- tion. Further work is needed to learn whether this result is an artifact of our use of nest box- es. House Wrens are common in suburban habitats which suggest sufficient young are produced in these habitats to sustain popula- tions. Alternatively, suburban habitats may be sinks and populations rely on dispersal of in- dividuals from other locations to maintain cur- rent densities. Researeh quantifying factors di- rectly associated with habitat quality (e.g., prey availability and nest predation rates), measures of annual survival, and habitat se- lection in this system should provide insight into the extent urban and suburban areas act as population sinks and ecological traps for House Wrens and other native song birds. ACKNOWFEDGMENTS We thank the Neighborhood Nestwatch participants for permission to work on their property, J. Smith, R. J. Peters, and K. E. Roux for help in the field, and R. S. Ebersole for valuable comments on the manuscript and moral support. We also thank the Mills Corpora- tion for financial support. FITERATURE CITED Beale, C. M. and P. Monaghan. 2004. Human dis- turbance; people as predation free predators? Jour- nal of Applied Ecology 41:335-343. Cam, E., j. D. Nichols, J. R. Sauer, J. E. Hines, and C. H. Feather. 2000. Relative species richness and community completeness: birds and urbani- zation in the Mid-Atlantic States. Ecological Ap- plications 10:1 1 96- 1210. Case, T. J. 1996. Global patterns in the establishment and distribution of exotic birds. Biological Con- servation 78:69-96. CzAPKA, S. J. AND L. S. Johnson. 2000. Consequences of mate sharing for first-mated females in a po- lygynous songbird, the House Wren. Wifson Bul- letin 1 12:72-81. Dawson, R. D., C. C. Lawrie, and E. L. O’Brien. 2005. The importance of microclimate variation in determining size, growth and survival of avian 104 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 offspring: experimental evidence from a cavity nesting passerine. Oecologia 144:499-507. Donnelly, R. 2002. Design of habitat reserves and settlement for bird conservation in the Seattle metropolitan area. Dissertation. University of Washington, Seattle, USA. Donnelly, R. and J. M. Marzluff. 2004. Importance of reserve size and landscape context to urban bird conservation. Conservation Biology 18:733-745. Donovan, T. M. and E R. Thompson. 2001. Modeling the ecological trap hypothesis: a habitat and de- mographic analysis for migrant songbirds. Eco- logical Applications 11:871-882. Donovan, T. M., P. W. Jones, E. M. Annand, and E R. Thompson. 1997. Variation in local-scale edge effects: mechanisms and landscape context. Ecol- ogy 78:2064-2075. Gates, J. E. and L. W. Gysel. 1978. Avian nest dis- persion and fledging success in field-forest eco- tones. Ecology 59:871—883. Giese, M. 1996. Effects of human activity on Adelie Penguin Pygoscelis adeliae breeding success. Bi- ological Conservation 75:157-164. Goldstein, E. L., M. Gross, and R. M. DeGraff. 1986. Breeding birds and vegetation: a quantita- tive assessment. Urban Ecology 9:377-385. Hennings, L. A. and W. D. Edge. 2003. Riparian bird community structure in Portland, Oregon: habitat, urbanization, and spatial scale patterns. Condor 105:288-302. ISHITANI, M., D. J. KOTZE, and J. NiEMELA. 2003. Changes in carabid beetle assemblages across an urban-rural gradient in Japan. Ecography 26:481- 489. Johnson, L. S. 1998. The House Wren {Troglodytes aedon). The birds of North America. Number 380. Johnson, L. S. and L. H. Kermott. 1993. Why is reduced male parental assistance detrimental to the reproductive success of secondary female House Wren? Animal Behaviour 46:1111-1120. Kokko, H. and W. j. Sutherland. 2001. Ecological traps in changing environments: ecological and evolutionary consequences of a behaviourally me- diated Allee effect. Evolutionary Ecology Re- search 3:537-551. Leston, L. and a. D. Rodewald. 2006. Are urban forests ecological traps for understory birds? An examination using Northern Cardinals. Biological Conservation 131:566-574. Marzluff, J. M. 2001. Worldwide urbanization and its effects on birds. Pages 19-47 in Avian ecology and conservation in an urbanizing world (J. M. Marzluff, R. Bowman, and R. Donnelly, Editors). Kluwer Academic, Norwell, Massachusetts, USA. Mills, G. S., J. B. Dunning Jr., and J. M. Bates. 1991. The relationship between breeding bird den- sity and vegetation volume. Wilson Bulletin 103: 468-479. Misenhelter, M. D. and J. T. Rotenberry. 2000. Choices and consequences of habitat occupancy and nest site selection in Sage Sparrows. Ecology 81:2892-2901. Morrison, B. L. and L. S. Johnson. 2002. Feeding of House Wren nestlings afflicted by hematophagous ectoparasites: a test of the parental compensation hypothesis. Condor 104:183—187. Negative Population Growth. 2005. Maryland’s population in 2050: is smart growth enough? http://www.npg.org/polls091200/md (accessed 22 June 2005). Newton, I. 1998. Population limitation in birds. Aca- demic Press. London, United Kingdom. Obendorf, S. K., a. T. Lemley, A. Hedge, A. A. Kline, K. Tan, and T. Dokuchayeva. 2006. Dis- tribution of pesticide residues within homes in central NY State. Archives of Environmental Con- tamination and Toxicology 50:31-44. Patten, M. A. and D. T. Bolger. 2003. Variation in top-down control of avian reproductive success across a fragmentation gradient. Oikos 101:479- 488. Pavlik, J. and S. Pavlik. 2000. Some relationships between human impact, vegetation, and birds in urban environments. Ekologia-Bratislava 19:392- 408. Purcell, K. L. and J. Verner. 1998. Density and re- productive success of California Towhees. Con- servation Biology 12:442-450. Resumes, V. 2000. How can maladaptive habitat choice generate source-sink population dynamics? Oikos 91:579-582. Robinson, S. K., F. R. Thompson III, and T. M. Don- ovan. 1995. Regional forest fragmentation and the nesting success of migratory birds. Science 267. 1987-1990. SAS Institute. 1999. SAS/STAT user’s guide. Release 8.1. SAS Institute Inc., Cary, North Carolina, USA. SCHLAEPFER, M. A., M. C. RUNGE, AND P. W. ShERMAN. 2002. Ecological and evolutionary traps. Trends in Ecology and Evolution 17:474—480. Thorington, K. K. and R. Bowman. 2003. Predation rate on artificial nests increases with human hous- ing density in suburban habitats. Ecography 26: 188-196. Weldon, A. J. and N. M. Haddad. 2005. The effects of patch shape on Indigo Buntings: evidence for an ecological trap. Ecology 86:1422—1431. The Wilson Journal of Ornithology 120( 1 ): 105-1 10, 2008 HABITAT SELECTION AND REPRODUCTIVE SUCCESS OE CERULEAN WARBLERS IN INDIANA KIRK L. ROTH' AND KAMAL ISLAM' ^ ABSTRACT. — We studied the breeding habitat of Cerulean Warblers (Dendroica cerulea) in southern Indiana in the Ohio River Valley in 2002-2003 to identify similarities and differences in habitat characteristics compared to breeding habitats reported for this species in other geographical regions. Ten 259-ha study plots were surveyed for Cerulean Warblers and territories were mapped using locations of perched, singing males. We measured slope and vegetation characteristics including canopy height and cover, ground cover, number of shrubs and shrub species, number of trees, DBH, and number of snags at Cerulean Warbler territories and non-use sites. Habitat characteristics associated with Cerulean Warbler territories compared to non-use sites were higher canopy height (28 m) and cover (84%), larger trees (>38 cm DBH), higher slope (11°), fewer number of trees (30), and fewer trees between 3 and 23 cm DBH. Calculated Mayfield estimate of nest productivity (0.165) was lower compared to Mississippi Alluvial Valley sites (0.242). Received 23 December 2006. Accepted 17 July 2007. The Cerulean Warbler {Dendroica cerulea) is one of the fastest-declining neotropical mi- grants in North America (Robbins et al. 1992) and populations declined at an average annual rate of 3.04% between 1966 and 2000 (Rob- bins et al. 1992, Link and Sauer 2002). The breeding range of the Cerulean Warbler in- cludes most of the Midwest, from the Missis- sippi Alluvial Valley north into southern On- tario, Canada (Hamel 1992) and much of the northeastern United States. Large population losses have been reported in the Midwest, which represents the center of the species’ dis- tribution (Dunn and Garrett 1997). Despite an increase in forest cover during the last 50 years in the northeastern United States, loss of suitable breeding habitat likely is a major cause of Cerulean Warbler population de- clines. Fragmentation of suitable forest habitat is prominent in the agricultural areas of the Midwest and Mississippi Alluvial Valley within the range of this species (Robbins et al. 1992). The breeding habitat of Cerulean Warbler has been described as deciduous forest with large, tall trees (Lynch 1981, Hamel 2000a, Jones and Robertson 2001). However, differ- ences in habitat characteristics have been not- ed across the range of the species (Hamel 2000a) and Hamel (2000b) suggested that Ce- rulean Warblers do not exhibit a rangewide nesting tree preference. Specific tree prefer- ' Department of Biology, Ball State University, Muncie, IN 47306, USA. ^Corresponding author; e-mail: kislam@bsu.edu ences may occur in different localities within the breeding range (Oliarnyk and Robertson 1996; Dunn and Garrett 1997; Hamel 2000a, b; Barg et al. 2006). Other aspects of Cerulean Warbler breeding habitat, such as nest height (Hamel 2000a) and forest tract size (Peterjohn and Rice 1991, Robbins et al. 1992) also vary rangewide. Differences in specific breeding habitat characteristics among geographical re- gions suggest that habitats suitable for Ceru- lean Warblers in one area may be unsuitable for populations in other locations. Variability in habitat preferences, which could reflect tree species availability or specific site-dependent parameters, has implications for site-specific habitat management to benefit Cerulean War- bler populations. Identification of preferred vegetative struc- ture is considered a high research priority for conservation of Cerulean Warblers (Hamel 2000a). Studies of the breeding biology of Ce- rulean Warblers are needed for many geo- graphical locations throughout its distribution (Hamel 2000a, Hamel et al. 2004). Research focusing on breeding biology of Cerulean Warblers is difficult because the species typi- cally nests in the canopy 9-15 m above ground level (Hamel 2000a, Jones and Rob- ertson 2001). Investigation of nest productiv- ity requires intensive study and has been rare- ly undertaken (Oliarnyk and Robertson 1996, Hamel 2000a). The only published studies on Cerulean Warbler nest characteristics and pro- ductivity arc from research sites in Ontario, Canada (Oliarnyk and Robertson 1996, Jones and Robertson 2001, Hamel et al. 2004, Jones et al. 2004), and Michigan (Rogers 2006). 105 106 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 1, March 2008 Our objectives were to: (1) identify breed- ing habitat characteristics of Cerulean Warbler territories in southern Indiana, (2) compare characteristics of Cerulean Warbler territories with randomly selected non-use sites, (3) compute nest survival, and (4) ascertain if there were any demonstrated site-specific breeding habitat preferences. METHODS Study Area.— We conducted our study in the 20,234-ha Big Oaks National Wildlife Refuge (39°03'N, 85°25'W) in Jefferson, Jennings, and Ripley counties, Indiana. Ten 259-ha (1 miO study plots with deciduous for- est were selected in the northern third of the refuge based on the presence of Cerulean Warblers. The forest in the study plots was dominated by oak (Quercus spp.), hickory {Cary a spp.), and maple {Acer spp.). Each (259-ha) plot was grid-marked at 200-m in- tervals resulting in 49 interval points to facil- itate mapping of territories and vegetation sampling. Bird Surveys.— Wq conducted surveys of male Cerulean Warblers within the study plots from early May through late June in 2002 and 2003. Each plot was surveyed once during each breeding season. We used a 100-m fixed- radius point count for 3-min at each grid in- terval (Hutto et al. 1986). Cerulean Warblers in our study area prominently and strongly sang from selected song perches and playback was often unnecessary. We attempted to elicit a response using playbacks of Cerulean War- bler songs to simulate territory occupancy if Cerulean Warblers were not detected (Ealls 1981). We played the recording for 2 min at a broadcast covering an area with a 100-m radius using a Sony WM-FX281 Walkman with a Radio Shack Model 540-1441 speaker. Bird surveys were conducted during peak ac- tivity times between 0530 and 1000 hrs EST and not during rainy or windy weather. We mapped Cerulean Warbler territories from mid-May through early July. Territories were defined by locations of perched, singing males and were demarcated using locations of ob- servations and bird responses to playbacks. Playbacks of another territorial male outside the territory boundary have been demonstrat- ed to cause male songbirds to sing at the edge of their own territories (Falls 1981). All birds drawn by playback moved distances ^50 m (K. L. Roth, pers. obs.). No territories in the study plots overlapped or were adjacent to one another. A territory was considered to be com- pletely mapped when the bird indicated no new song perches, even in response to play- backs. Complete territory mapping involved a time period of 1—4 days. Universal Transverse Mercator (UTM) coordinates were recorded along territorial boundaries using a Global Po- sitioning System (GPS) unit with NAD 83 da- tum and later mapped using the Arcinfo com- puter program (ESRI 1996). Vegetation Sampling.— sampled vege- tation within each of the 7 1 territories detected and at an equal number of randomly selected sites in the study plots from early July to mid- August each year (James and Shugart 1970). The approximate center of each territory, lo- cated by pacing and halving the length and width of the territory, was used as the center of the vegetation sampling plot. Random sites were selected by applying a random numbers table to the grid setup of the study plots. One random site was selected for each mapped ter- ritory. All vegetation sampling plots com- prised a 11.3-m radius circle (0.04 ha). We sampled canopy cover and ground cover at 2-m intervals in each cardinal direction within the plot, using a densiometer, which resulted in a total of five readings in each direction. Slope was measured using a clinometer at the edges of the plot in each cardinal direction. Habitat characteristics measured within the 11.3-m radius plots included diameter at breast height (DBH) for trees and snags with a DBH >5 cm, mean tree height, percent can- opy cover, and density of trees, snags, and shrubs. We measured DBH for woody species <5 cm (defined as shrubs) within a 5.0-m ra- dius of the center of the plot. We estimated canopy height using a clinometer in each quarter of the plot. Trees and shrubs were sep- arated into five size classes to categorize tree, snag, and shrub densities: (3-7.9, 8-14.9, 15- 22.9, 23-37.9, and >38 cm). We measured vertical stratification using methods modified from Hubbel and Foster (1986). Vertical stratification was estimated from 22.8 m toward the opposite border at plot borders in each cardinal direction. Verti- cal stratification was recorded using cover es- timates at five height intervals (<2, <5, <10, Roth and Islam • CERULEAN WARBLER BREEDING IN INDIANA 107 <20, and >20 m) on a stratification pole. We scored vertical stratification at each height in- terval on a scale of 1 to 6: 1 (1-7% cover), 2 (8-25% cover), 3 (26-50% cover), 4 (51-75% cover), 5 (76-93% cover), or 6 (94-100% cover). Vegetation Analysis. — Differences between vegetation components at territorial and ran- domly selected non-use sites were analyzed with r-tests for equality of means using SPSS (SPSS Inc. 2002). Logistic regressions were used for variable comparisons between ran- dom and territorial sites for: number of shrub species, number of shrubs, vertical stratifica- tion classes, overall mean DBH, number of trees including snags, all DBH classes, canopy height, canopy cover, ground cover, and slope. Canopy cover, ground cover, canopy height, and slope readings were combined for analy- sis in each plot. Vegetative characteristics are reported as means ± standard deviation. Nest Site Characterization and Nest Surviv- al.— We located Cerulean Warbler nests using behavioral cues from adults. Vegetative pa- rameters of nest trees were sampled using the Bbird field protocol (Martin et al. 1997). We measured DBH, horizontal distance to the fo- liage edge of the nest tree, and horizontal dis- tance to the nest tree trunk for each nest tree. We recorded nest tree species and vegetative species that provided concealment above the nest. Nest height was estimated using a cli- nometer. Tree species and DBH of Cerulean Warbler song perch trees from territory map- ping were used as comparisons for nest trees (Barg et al. 2006). Nest and perch tree differ- ences were analyzed using Utests of equality of means using SPSS (SPSS Inc. 2002). Dif- ferences between nest tree species and ex- pected tree species based upon occurrence in the territory were considered, but nest tree sample size was too small for an associative test. Nest tree vegetative characteristics are re- ported as means ± standard deviation. Each nest was monitored from the ground using binoculars and spotting scopes. We observed and recorded the number of days the nest was in the building, incubation, and brooding stag- es using parental activity as an indicator. Nest- ing stages were designated conservatively giv- en the difficulty of observing Cerulean War- bler nests. Thus, a bird was required to show definitive behavior to indicate the nesting stage. Timing of these stages was confirmed using average nest data (Hamel 2000a). Due to frequent nest observation, no nests had un- certain fail dates. These data were used for Mayfield estimates of nest productivity (May- field 1975). RESULTS Territorial Characteristics. — There were differences {P < 0.05) in five characteristics (canopy cover and height, slope, number of trees, and DBH) between Cerulean Warbler territories {n = 71) and random (n = 71) sites (Table 1). Overall canopy height and cover were significantly greater in territories than at random sites. Cerulean Warbler territories had significantly higher slope but fewer number of trees than random sites (Table 1). Number of shrubs, number of shrub species, number of snags, and vertical stratification at all height intervals did not differ between territories and random sites. The number of trees for three of five DBH classes (3-7.9, 8-14.9, and 15-22.9 cm) was significantly greater in random sites than ter- ritories but there were significantly more trees >38 cm DBH present in territories than ran- dom sites (Table 1). The number of trees be- tween 23 and 37.9 cm DBH did not differ between territories and random sites. Important predictive components, based on logistic regression analyses, included canopy height (p = 0.089, SE = 0.030, P = 0.003), number of trees ((3 = —0.025, SE = 0.012, P = 0.036), and slope ((3 = 0.093, SE = 0.039, P = 0.016) with a 67.2% predictive value for Cerulean Warbler territories. The probability that a site is used as a Cerulean Warbler ter- ritory increased with increasing canopy height and slope, and decreased with increasing num- ber of trees. Nest Tree Characteristics and Nest Surviv- al.— Tulip trees {Liriodendron tulipifera) were widely available, but use as nest trees by Ce- rulean Warblers was low. Fifty-three percent of nest trees were either black walnut {Juglans nigra) or white oak {Quercus alba). Fourteen nest trees included 9 black walnuts, 2 sweet- gums {Licpiidatnhar styraciflaa), 2 black lo- custs (Rohinia pseudoacacia), and 1 eastern sycamore (Platanus occidental is). Thirty- three percent of nests were concealed by Vir- ginia creeper (Parthenocissus cptinquefolia). 108 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 1, March 2008 TABLE 1 Vegetative characteristics at Cerulean Warbler territories and random sites at Big Oaks National Wildlife Refuge, Indiana during 2002 and 2003. Canopy and ground cover measurements were on a scale of 0- 5, with 5 being complete cover. Stratification values were coded on a scale of 1-6 following Hubbel and Loster (1986). Territory Random Results of f-test Vegetation parameter x SD n x SD n t df ^ Canopy cover 4.19 0.59 58 Ground cover 2.77 1.11 58 Slope (degrees) 10.87 6.9 58 Canopy height (m) 28.19 6.49 56 Number shrubs 1.92 2.51 56 Number shrub species 2.69 1.7 39 Number trees 29.7 13.13 56 DBH (cm) 22.8 2.98 56 Number snags 5.05 4.98 56 Number trees 3-7.9 DBH 11.8 10.02 56 8-14.9 DBH 5.8 5.78 56 15-22.9 DBH 3.6 3.61 56 23-37.9 DBH 4.4 4.37 56 >38 DBH 3.6 3.59 56 Vertical cover 0-2.5 m 3.25 1.52 58 2.5-5 m 2.88 1.24 58 5-10 m 3.29 1.07 58 10-15 m 3.83 1.04 58 >15 m 3.95 1.02 58 3.91 0.78 58 -2.181 114 0.031 2.97 1.14 58 0.972 114 0.33 7.13 5.31 58 -3.275 114 0.001 24.07 7.35 58 -3.18 112 0.002 1.61 1.54 54 -1.135 235 0.25 2.51 1.79 41 -0.46 78 0.64 42.9 29.27 54 3.047 72.9 0.003 20.7 5.19 54 -2.595 83.6 0.011 7.01 7.51 54 1.611 91.6 0.11 17.6 18.66 54 2.055 80.5 0.043 10.6 10.59 54 3.49 73.7 0.001 6.6 6.61 54 3.683 81 <0.001 4.9 4.91 54 0.798 105.2 0.42 2.6 2.56 54 -2.482 96 0.015 3.34 1.3 57 0.316 113 0.75 3.08 1.24 57 0.902 113 0.36 3.3 1.24 57 0.064 113 0.94 3.7 1.23 57 0.614 113 0.54 4.06 1.16 57 -0.553 113 0.58 All nests not concealed by Virginia creeper were concealed by the foliage of the nest tree. DBH was greater for nest trees {x = 50.4 ±25.8 cm) compared to trees at random sites, trees in territories (x = 22.81), and song perch trees. Nest tree height was variable and aver- aged 18.38 ± 5.08 m{n = 43). Nest distance from trunk averaged 4.8 ± 2.4 m. Nest distance from foliage edge averaged 2.9 ± 1.5 m. Forty of 43 nests located contained eggs or nestlings. Thirty-three of 40 active nests failed and 7 fledged young (Table 2). Mayfield es- timates of nest productivity gave a seasonal TABLE 2. Number of Cerulean Warbler territories surveyed and nest productivity data, including May- field estimates (Mayfield 1975) at Big Oaks National Wildlife Refuge, Indiana during 2002 and 2003. Parameter 2002 2003 Totals Territories 42 29 71 Lledged nests 4 3 7 Tailed nests 10 23 33 Mayfield estimate 0.335 0.0899 0.165 survival probability of 0.165 for both years. The Mayfield estimate was higher during 2002 (0.335) than during 2003 (0.0899) (Table 2). DISCUSSION Important characteristics associated with Cerulean Warbler territories were tall tree can- opy, higher percent of canopy cover, higher slope, more trees >38 cm DBH, fewer num- ber of trees, and fewer trees between 3 and 22.9 cm DBH than random sites. Factors that did not appear to be important in Cerulean Warbler territory selection included vertical stratification, number of snags, shrub cover, and ground cover. The vegetative characteristics of Cerulean Warbler territories in our study area were sim- ilar to those reported in other studies of Ce- rulean Warblers (Lynch 1981, Kahl et al. 1985, Jones and Robertson 2001, Weakland and Wood 2005). Territorial canopy cover in our study averaged 83.7%, similar to the 85% in Missouri (Kahl et al. 1985). There was a Roth and Islam • CERULEAN WARBLER BREEDING IN INDIANA 109 significant difference in tree density between territorial and random sites in our study but not in Ontario (Jones and Robertson 2001). Our study supports others (Robbins et al. 1992, Hamel 2000a, Jones and Robertson 2001, Barg et al. 2006, Jones and Islam 2006) that Cerulean Warblers select large trees for nesting and perching. Oliamyk and Robertson (1996) reported that large nest tree size was an important factor for nest success. DBH was larger for nest trees compared to territorial site and song perch trees in our study, indicating that Cerulean Warblers select for the largest trees within their territories for nest sites. Some form of nest concealment was found at all nests of Cerulean Warblers at Big Oaks National Wildlife Refuge. Live Virginia creeper was an important component for nest concealment, overhanging a third of all nests located. No Cerulean Warbler nest was con- cealed by any other vine, despite an abun- dance of grapes {Vitis sp.) and poison ivy {Toxicodendron radicans) (K. L. Roth, pers. obs.). The denser foliage of Virginia creeper may present a more attractive cover for Ce- rulean Warblers at Big Oaks National Wildlife Refuge, which otherwise build nests only un- der extensive cover provided by the leaves of the nest tree. Cerulean Warblers appear to be consistent in selecting the largest trees in territories, but there is no definitive optimal tree size. The average nest height of 18.4 m in our study is higher than the average height of 11.8 m for nests in Ontario (Oliamyk and Robertson 1996) and the 1 1.4 m (n = 80) rangewide re- ported by Hamel (2000b). However, nest height at sites reported in Michigan was slightly higher (19-20 m; Rogers 2006). Thus, Cerulean Warblers select the largest trees for territorial and nesting activities. Most territo- ries in our study were in topographically di- verse areas and particularly near drainages. This is evident from higher slopes associated with Cerulean Warbler territories compared to random sites. However, a small number of ter- ritories were also noted in several floodplain areas of our study area. We recorded low nest success with our Mayheld estimate of 0.165 compared with a 0.242 success for 66 nests at three Mississippi Alluvial Valley sites (Hamel 2000a). Our Mayheld estimate assumed that all eggs pre- sent at hatch time will produce living young and actual survival of young may be lower than estimated. We did not observe any brood parasitism. We suspected predation was a major cause of nest loss. Potential nest predators including American Crows {Corvus brachyrhynchos). Blue Jays (Cyanocitta cristata), and rat snakes (Elaphe obsoleta) were commonly found in Cemlean Warbler nesting areas. We observed one nest being depredated by an American Crow and another failed nest had remains of a tom wing with pinfeathers directly under- neath (K. L. Roth, pers. obs.). Another pos- sible cause of nest failure is weather-related events (Jones et al. 2001). Southern Indiana experienced two unseasonably cold periods of below 5° C temperatures in late May-early June 2003. Four of 14 (29%) and 4 of 7 (57%) active nests failed during the hrst and second cold spells, respectively. Some habitat characteristics appear to be common to Cerulean Warblers throughout the breeding range of the species. However, ter- ritorial characteristics may be unique to a spe- cihc population within a region. Tree size and forest moisture regime may not be useful as rangewide dehning characteristics of Cemlean Warbler habitat. These factors, combined with the variability in size of forest tract selected by this species (Peterjohn and Rice 1991, Robbins et al. 1992, Hamel 2000a) make hab- itat characterization difficult. Our data may only be useful for Cerulean Warbler habitat management in the Ohio River Valley. A sin- gle rangewide management approach may not be an appropriate strategy for this species. We recommend greater emphasis on formulating and implementing site-specific management recommendations to conserve important hab- itat features. These studies should be under- taken to identify Cerulean Warbler habitat in all major areas of this species’ breeding range including sites within Appalachia, Wisconsin, and the Ozark Mountains. Cerulean Warblers have an affinity for the habitat in our study area, but further research is required to iden- tify critical habitat characteristics associated with nest success/productivity. Research at- tention should be given to habitat issues such as canopy gaps, horizontal vegetative com- plexity, and canopy foliage configuration (Ha- mel 2()()()a), no THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 ACKNOWLEDGMENTS This research was supported, in part, by the U.S. Fish and Wildlife Service, Garden Club of America, Indiana Academy of Science, Sigma Xi, and Ball State University Office of Academic Research and Spon- sored Programs. We thank Joseph Robb for permission to work at Big Oaks National Wildlife Refuge and his staff for field support. We thank refuge staff, Teresa Vanosdol, Jason Lewis, and Laura Lake, and our field crew, K. C. Jones, Joseph Allen, Holly Adams, Dustin Varble, Elizabeth Shaefer, Kendra Darnell, Kate Had- dan, Amy Saylor, Eric Suttles, and C. R. Ward for assistance in data collection. We are grateful to Jason Jones and Randy Dettmers for contributing their ex- pertise and time in the field. Analytical assistance was provided by James Jones. This manuscript benefited from the critical comments and suggestions of two anonymous reviewers and the editor. LITERATURE CITED Barg, J. J., D. M. Aiama, J. Jones, and R. J. Robert- son. 2006. Within-territory habitat use and micro- habitat selection by male Cerulean Warblers (Den- droica cerulea). Auk 123:795—806. Dunn, J. L. and K. L. Garrett. 1997. A field guide to warblers of North America. Houghton-Mifflin, Boston, Massachusetts, USA. ESRI. 1996. ArcView GIS. Environmental Systems Research Institute Inc., Redlands, California, USA. Falls, J. B. 1981. Mapping territories with playback: an accurate census method for songbirds. Studies in Avian Biology 6:86-91. Hamel, P. B. 1992. Cerulean warbler, Dendroica cer- ulea. Pages 385-400 in Migratory nongame birds of management concern in the Northeast (K. J. Schneider and D. M. Pence, Editors). USDI, Fish and Wildlife Service, Newton Corner, Massachu- setts, USA. Hamel, P. B. 2000a. Cerulean Warbler {Dendroica cer- ulea). The birds of North America. Number 511. Hamel, P. B. 2000b. Cerulean Warbler status assess- ment. USDI, Fish and Wildlife Service, Minne- apolis, Minnesota, USA. Hamel, P. B., D. K. Dawson, and P. D. Keyser. 2004. How we can learn more about the Cerulean War- bler {Dendroica cerulea). Auk 121:7—14. Hubbel, S. P. and R. B. Foster. 1986. Canopy gaps and the dynamics of a neotropical forest. Pages 77-95 in Plant ecology (M. J. Crawley, Editor). Blackwell Scientific Publications, Oxford, United Kingdom. Hutto, R. L., S. M. Pletschet, and T. P. Hendricks. 1986. A fixed-radius point-count method for non- breeding and breeding season use. Auk 103:593- 602. James, F C. and H. H. Shugart Jr. 1970. A quanti- tative method of habitat description. Audubon Field Notes 24:727-736. Jones, J. and R. J. Robertson. 2001. Territory and nest-site selection of Cerulean Warblers in eastern Ontario. Auk 118:727-735. Jones, J., R. D. DeBruyn, J. J. Barg, and R. J. Rob- ertson. 2001. Assessing the effects of natural dis- turbance on a neotropical migrant songbird. Ecol- ogy 82:2628-2635. Jones, J., J. J. Barg, T. S. Sillett, M. L. Veit, and R. J. Robertson. 2004. Minimum estimates of survival and population growth for Cerulean War- blers {Dendroica cerulea) breeding in Ontario, Canada. Auk 121:15-22. Jones, K. and K. Islam. 2006. Selection of song perches by Cerulean Warblers. Proceedings of the Indiana Academy of Science 115:37-43. Kahl, R. B., T. S. Baskett, J. A. Ellis, and J. N. Burroughs. 1985. Characteristics of summer hab- itats of selected nongame birds in Missouri. Ag- riculture Experiment Station Research Bulletin 1056. University of Missouri, Columbia, USA. Link, W. A. and J. R. Sauer. 2002. A hierarchical analysis of population change with application to Cerulean Warblers. Ecology 83:2832-2840. Lynch, J. M. 1981. Status of the Cerulean Warbler in the Roanoke River Basin of North Carolina. Chat 45:29-35. Martin, T. E., C. R. Paine, C. J. Conway, W. M. Ho- CHACHKA, P. Allen, and W. Jenkins. 1997. BBIRD Field Protocol. Montana Cooperative Wildlife Research Unit, University of Montana, Missoula, USA. Mayheld, H. 1975. Suggestions for calculating nest success. Wilson Bulletin 87:456—466. Oliarnyk, C. j. and R. J. Robertson. 1996. Breeding behavior and reproductive success of Cerulean Warblers in southeastern Ontario. Wilson Bulletin 108:673-684. Peterjohn, B. G. and D. L. Rice. 1991. The Ohio breeding bird atlas. Division of Natural Areas and Preserves, Ohio Department of Natural Resources, Columbus, USA. Robbins, C. S., J. W. Fitzpatrick, and P. B. Hamel. 1992. A warbler in trouble: Dendroica cerulea. Pages 549-562 in Ecology and conservation of neotropical migrant landbirds (J. M. Hagan and D. W. Johnston, Editors). Smithsonian Institute Press, Washington D.C., USA. Rogers, C. M. 2006. Nesting success and breeding biology of Cerulean Warblers in Michigan. Wilson Journal of Ornithology 118:145—151. SPSS Inc. 2002. SPSS. Version 1 1.5. Chicago, Illinois, USA. Weakland, C. a. and P. B. Wood. 2005. Cerulean Warbler {Dendroica cerulea) microhabitat and landscape-level habitat characteristics in southern West Virginia. Auk 122:497—508. The Wilson Journal of Ornithology 120(1):1 1 1-1 19, 2008 NESTING BIOLOGY OF GRASSLAND BIRDS AT FORT CAMPBELL, KENTUCKY AND TENNESSEE JAMES J. GIOCOMO,! 3 E. DANIEL MOSS,^ DAVID A. BUEHLER,i AND WILLIAM G. MINSERi ABSTRACT. — Grassland birds have experienced greater population declines than any other group of birds monitored by the North American Breeding Bird Survey. Our goal was to compare demographic rates among years within species and among species of grassland birds. Eight-hundred and eleven nests of Henslow’s Spar- rows (Ammodramus henslowii). Grasshopper Sparrows (A. savannarum). Field Sparrows {Spizella pusilla), Dick- cissels {Spiza americana), and Eastern Meadowlarks (Sturnella magna) were monitored between 1999 and 2003. Mayfield nest success including the egg-laying stage, as well as the incubation and nestling periods, was 20, 34, 15, 20, and 18%, respectively. Most nest failures were attributed to predation. Nest parasitism by Brown- headed Cowbirds (Molothrus ater) was infrequent (<2% of all nests parasitized). Clutch size decreased during the nesting season for Dickcissels, Grasshopper Sparrows, and Field Sparrows. Nesting phenology suggests the possibility of multiple-brooding for all five species in this study. Received 21 February 2006. Accepted 14 April 2007. Grassland bird population declines have been attributed to the dramatic decrease of na- tive grasslands during the 20th century be- cause of clearing of non-forested land for ag- riculture or development, and discontinued use of prescribed fire (Askins 1993, Herkert et al. 1996, Peterjohn and Sauer 1999). Mili- tary lands in the eastern United States are an exception to the trend in loss of native grass- lands. Some installations have maintained large areas of grasslands through use of pre- scribed burning and mowing to facilitate mil- itary training. For example. Fort Campbell Military Reservation in Kentucky and Ten- nessee, a 42,000-ha U.S. Army Base, has — 10,000 ha of grassland habitat including remnant patches of native grassland (Moss 2001). Several other military installations in the eastern United States including Fort Knox, Kentucky, Fort Bragg, North Carolina, and Fort Drum, New York, also have large areas of early-successional habitats (Eberly 2002). It is important to understand not only the distribution of grassland bird species in the eastern United States, but also their productiv- ity. Many studies report densities and diver- ' Department of Forestry, Wildlife and Fisheries, University of Tennessee, 274 Ellington PSB, Knox- ville, TN 37996, USA. 2 Directorate of Public Works, IMSE-CAM-PWE, 865 16th Street, Fort Campbell, KY 42223, USA. ^Corresponding author; e-mail: jgiocomo@utk.edu sity of bird species, but these measures may be misleading indicators of habitat quality or breeding success (Van Home 1983, Vickery et al. 1992). Few studies have collected de- tailed demographic information needed to un- derstand productivity (i.e., nesting success, clutch size) (Marzluff and Sallabanks 1998). Many grassland bird nests are difficult to find and monitor, and relatively few studies have attempted to monitor more than one to two species for more than a few years (Winter 1998). Managers of military installations need demographic information to understand if and how their management actions and military training may impact bird populations. We monitored Henslow’s Sparrow {Ammo- dramus henslowii). Grasshopper Sparrow {A. savannarum). Field Sparrow {Spizella pusil- la), Dickcissel {Spiza americana) and Eastern Meadowlark {Sturnella magna) nests at Fort Campbell from 1999 through 2003. Our ob- jectives were to obtain annual, species-specif- ic demographic information including nest success, clutch size, young produced per suc- cessful nest, causes for nest failure, nest par- asitism rates, timing of nest initiation, and sea- sonal clutch size variation, and to compare these demographic rates among years within species and among species. METHODS Study Area. — The study was conducted on Fort Campbell on the Kentucky-Tennessee THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 1 12 border. Fort Campbell contains some of the largest remaining blocks of native prairie “barrens” east of the Mississippi River. Bar- rens are grass-dominated, treeless areas oc- curring on the hilly, karst topography of west- central Kentucky and northwestern Tennessee (Chester et al. 1997). Historically, these grass- lands were maintained primarily through burning by native Americans (Delcourt et al. 1993). Many grasslands on Fort Campbell contain at least some native warm-season grasses including little bluestem {Schizachyr- ium scoparium), big bluestem (Andropogon gerardii), switchgrass {Panicum virgatum), Indian grass {Sorghastrum nutans), and broomsedge {Andropogon virginicus) . Ap- proximately 60% of the area is in oak {Quer- cus sp.)-hickory (Carya sp.) forests with many leased agricultural fields (cool-season grasses, corn, millet, and soybeans) interspersed among the grasslands. Most of the grassland areas monitored in this study were managed mainly for military training and not specifi- cally for hay production. Nest Searching.— concentrated primar- ily on finding an adequate sample (—20) of nests each year for our five target species in- cluding Henslow’s Sparrow, Grasshopper Sparrow, Dickcissel, Eastern Meadowlark, and Field Sparrow. Fields with appropriate grassland habitat, such as dead standing clumps of grass, were systematically searched by one crew leader and four field assistants for males of target species defending territo- ries or exhibiting nesting behavior between 1 May and 31 July, 1999-2003. Fields were se- lected to provide the maximum opportunity to find a representative sample of nests for the more difficult species to monitor, especially Henslow’s Sparrows. Most fields on Fort Campbell are burned every 1-3 years; thus, a different set of fields was searched each year allowing us to search in areas with dead stand- ing vegetation that is not present the first year after a field is burned. Up to 20 fields were searched each year, but most nest searching was concentrated in a few larger fields (>150 ha) that provided habitat for all five species in all years. Field sizes ranged from 4 to 600 ha. Behavioral cues, such as birds flushing close to an observer, chipping, carrying nesting ma- terial, or carrying food or fecal sacs, were used to locate nest sites. Once nests were located, a flag was placed at least 5 m from the nest and detailed maps of the nest locations were drawn. Nests were monitored every 2-4 days to ascertain nest fate. We calculated apparent yearly nest suc- cess (successful nests/total nests), Mayfield (1961, 1975) nest success, and standard error for individual species where sample sizes were sufficient (-9 nests, Johnson 1979). Demographic Rate Estimates.— clutch size was calculated using the highest number of eggs or young for nests with a sta- ble clutch size for two consecutive visits. Suc- cessful nests were defined as any nest fledging at least one host young. Nests with no expo- sure time (e.g., induced fledging when the nest was found) were not included in nest success calculations. We calculated daily survival rates and the probability of nesting success for five periods: egg laying, incubation, nestling, incubation and nestling combined, and all pe- riods. The combined probability of nesting success during the incubation and nestling stages was calculated to facilitate comparison with studies that did not explicitly include the egg laying stage. We used mean period lengths as exponents to calculate the probability of nest success from daily survival probabilities for each spe- cies. We rounded mean clutch size to the near- est half-egg for the mean number of days dur- ing the egg laying stage for each species. Gen- erally, one egg is laid per day for the five tar- get species until the clutch is completed with incubation starting with laying of the last egg. We used published values for mean number of days for the incubation and nestling stage (Ehrlich et al. 1988). The number of days in the incubation and nestling stages combined, and all stages combined was the sum of the appropriate number of days in the respective component stages. Mean and standard errors for daily survival probabilities, and mean nest success were calculated by nesting period and year for each species (Johnson 1979). Statistical Analysis. — Yearly means of young per successful nest were compared within species using one-way ANOVA. Nest initiation dates were estimated to the day lay- ing started by back dating from the day the nest was found. Average start and end dates of nest initiation were calculated by averaging the first and last 10% of all nests combined Giacomo et al. • NESTING GRASSLAND BIRDS 113 TABLE 1. Demographic data for Henslow’s Sparrow, Grasshopper Sparrow, Dickcissel, Eastern Meadow- lark, and Field Sparrow at Fort Campbell, Kentucky, 1999-2003. Henslow’s Sparrow (n = 113)^ Grasshopper Sparrow (n = 131)2 Dickcissel (n = 204)2 Eastern Meadowlark (n = 87)2 Field Sparrow (n = 276)2 Successful nests 65 85 87 36 126 Unknown fate 1 0 0 1 0 Failed nests 47 46 117 50 150 Depredated 44 38 97 45 139 Abandoned 3 3 9 2 7 Abandoned-parasitized 0 0 0 0 1 Mowing for hay 0 4 4 1 2 Military activity 0 1 3 0 0 Weather 0 0 4 2 1 Apparent nest success, % 58 65 43 41 46 Clutch size average (n) 4.1(108) 4.4(131) 4.3(191) 4.6(87) 3.6(264) Hatching success, % (n) 90.4(80) 93.2(104) 90.3(116) 94.1(53) 95.9(171) Young fledged/nest 2.2 2.6 1.7 1.7 1.6 Young fledged/successful nest 3.9 4.1 3.9 4.0 3.6 Parasitized nests 1 0 0 0 3 ^ n = number of nests. for each species. We used the difference be- tween start and end dates as a measure of the minimum window of time available for each species to initiate nesting (nest initiation win- dow). We used linear regression to examine the relationship between clutch size and nest initiation dates. We used program CON- TRAST (Sauer and Hines 1989) to test for differences among daily survival rates within species, comparing different nesting intervals, and among species. The level of significance was set at a = 0.05 and we used a Bonferroni correction to adjust our level of significance for multiple tests (Zar 1998). RESULTS Eight-hundred and eleven nests were mon- itored between 1999 and 2003; apparent nest success ranged between 41 and 65% for each species (Table 1 ). Most nest failures were at- tributed to predation. The primary predators of nests appeared to be snakes based on nu- merous observations of snakes in nests. Other causes of nest failures included abandonment, hay mowing and harvesting, military training activities, and Brown-headed Cowbird (Mol- othnis ater) parasitism (Henslow’s Sparrow = 1, Field Sparrow = 3; Table 1). Average clutch size ranged from 3.6 for Field .Sparrows to 4.6 eggs for Eastern Meadowlarks. Hatch- ing success ranged from 90% for Dickcissels and Henslow’s Sparrows to 96% for Field Sparrows (Table 1). Average young fledged per nest ranged from 1.6 to 2.6, and the av- erage number of young per successful nest ranged from 3.6 for Field Sparrows to 4.1 for Grasshopper Sparrows (Table 1). Eastern Meadowlarks initiated nests earliest with average nest incubation starting on 16 April. Field Sparrow nest initiation started on 25 April followed by Henslow’s Sparrows (27 Apr) and Grasshopper Sparrows (1 May). Dickcissels consistently were the last species to arrive and began incubation on 10 May. The average end of the nest initiation window was between 28 June and 4 July for all five species, but active nests were found through August for all species. The length of the nest initiation window was longest for Eastern Meadowlarks (75 days), intermediate for Hen- slow’s Sparrows (63 days). Grasshopper Spar- rows (64 days), and Field Sparrows (64 days) and shortest for Dickcissels (51 days). Clutch size decreased during the nesting season for Dickcissels (/ = -6.19, df = 190, P < 0.001), Grasshopper Sparrows (/ = -2.23, df = 130, P = 0.03), and Field Spar- rows it = -5.52, df = 259, P < 0.001). On average, Dickcissel clutch size was reduced by one egg every 50 days, and Grasshopper Sparrow and Field Sparrow clutch sizes were reduced by 0.5 eggs every 62 and 52 days, respectively. We did not detect a linear de- crease in clutch size during the nesting season 114 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 for Henslow’s Sparrow {t = -0.37, df - 104, p = 0.71) and Eastern Meadowlark {t = -0.94, df = 85, P = 0.35). The clutch size for both Henslow’s Sparrows and Eastern Meadowlarks initially increased, peaked in the middle of the season, and then gradually de- creased. Nesting success was generally greatest for nests found during the laying stage and least during the incubation stage. Combining nests found in all years, nesting success for Eield Sparrows was lower than for Grasshopper Sparrows because of the difference in nest success during the incubation stage (Table 2). Nesting success varied among years, but we were unable to detect differences after Bon- ferroni correction (Table 3). DISCUSSION Overall mean nesting success rates (May- field 1975) were within the range of values previously reported for Henslow’s Sparrows (27% [incubation and nestling stages only]; Table 4), and Dickcissels (26%). Eastern Meadowlark (22%) and Grasshopper Sparrow (41%) nesting success was near the high end of previously reported values (Table 4). Al- most all nests of these four species were found in the largest fields (>150 ha). A few Hen- slow’s Sparrow nests were found in fields as small as 4 ha, but these small fields were with- in 10 km of larger fields. Henslow’s Sparrow densities were too low in small fields to allow for an adequate sample of nests each year. Field Sparrow nesting success was lower at Fort Campbell (20%) than previously reported (Table 4). Low nesting success may be related to the ubiquitous distribution of Field Sparrow nests in grassland fields, including some fields as small as 2 ha. Smaller fields had more hab- itat features that might attract potential pred- ators (e.g., small trees for perch sites), possi- bly accounting for reduced nesting success rates (Herkert 1994). Published nesting success rates usually do not include the egg laying stage. Between 6 and 36% of our nests, depending upon species and year, were found during the egg laying stage. This study is one of only a few that explicitly reports a daily survival rate of nests during the laying stage. We detected nest fail- ures for three of the five species monitored during this short but critical part of the nesting cycle, resulting in estimates of nest success during the egg-laying stage of 65-79%. The egg laying stage should be treated separately from the incubation stage because incubation usually starts between laying the penultimate egg and up to a few days thereafter. Eggs usu- ally are less conspicuous when the female is on the nest during incubation, reducing the probability predators will find the nest through visual cues. Exposed eggs during the laying stage may be more vulnerable to predators such as common raccoons {Procyon lotor) or Blue Jays (Cyanocitta cristata). The laying stage tended to have the greatest nest success rates followed by the nestling stage and the incubation stage (Table 2). Brown-headed Cowbird parasitism rates were low at Fort Campbell for these grassland species, but were within the range of reported parasitism rates for each species (Table 4). The lack of Dickcissel nest parasitism was particularly noteworthy when compared with other areas, but was consistent with records from Tennessee (Nicholson 1997). Our nest parasitism rates probably were low because most nests were in large grassland fields (>100 ha) far from forest edges or other tall woody perch sites (Hauber and Russo 2000, Jensen and Cully 2005), except those of Field Sparrows, which were found in the full range of field sizes (4-600 ha). Nest parasitism rates may be related to the proximity of songbird populations to the greatest densities of Brown- headed Cowbirds (Basili 1997, Winter et al. 2004). Fort Campbell is outside the greatest density areas for Brown-headed Cowbird pop- ulations (Sauer et al. 2004). Morris and Thompson (1998) found Brown-headed Cow- birds were most associated with grazed pas- tures, regardless of grass height. There is no livestock grazing activity at Fort Campbell. Clutch sizes of Henslow’s Sparrows, Grass- hopper Sparrows, Dickcissels, and Eastern Meadowlarks were slightly greater at Fort Campbell than average but were within the range of published rates (Table 4). Field Spar- row clutch size was less than average, but within the range of published rates. Average clutch sizes included two nests with more than twice the average number of eggs, one Dick- cissel with nine eggs and one Eastern Mead- owlark with 10 eggs. These individual nests may represent egg laying efforts of more than Giocomo et al. • NESTING GRASSLAND BIRDS 115 TABLE 2. Mayfield nesting success of grassland birds at Fort Campbell, Kentucky, 1999-2003, by nest cycle interval. Species Nest cycle interval^ Mean days in period^ Failures (n)‘^ Exposure days^ Daily survival SE Success (%)g HESP'^ Laying 4.0 8 2 19.5 0.897 0.069 64.9 GRSP Laying 4.5 8 0 17.0 1.000 0.000 100.0 DICK Laying 4.0 75 13 229.0 0.943 0.015 79.2 EAME Laying 4.0 13 0 26.0 1.000 0.000 100.0 FISP Laying 3.5 47 11 124.0 0.911 0.026 71.6 HESP Incubating 11.0 54 21 297.5 0.929 0.015 44.7 GRSP Incubating 11.5 73 23 445.0 0.948 0.010 54.3 DICK Incubating 12.5 145 67 997.0 0.933 0.008 41.9 EAME Incubating 12.5 54 25 354.5 0.929 0.014 40.1 FISP Incubating 12.0 176 83 912.5 0.909 0.010 31.8 HESP Nestling 9.5 88 23 412.5 0.944 0.011 58.0 GRSP Nestling 9.0 108 23 621.0 0.963 0.008 71.2 DICK Nestling 9.0 124 37 694.5 0.947 0.009 61.1 EAME Nestling 9.0 61 25 387.0 0.935 0.012 54.8 FISP Nestling 7.5 183 56 849.5 0.934 0.009 60.0 HESP Inc. and nestling 20.5 111 44 710.0 0.938 0.009 26.9 GRSP Inc. and nestling 20.5 129 46 1,066.0 0.957 0.006 40.5 DICK Inc. and nestling 21.5 191 104 1,691.5 0.939 0.006 25.6 EAME Inc. and nestling 21.5 86 50 741.5 0.933 0.009 22.3 FISP Inc. and nestling 19.5 266 139 1,762.0 0.921 0.006 20.1 HESP All stages 24.5 113 46 729.5 0.937 0.009 20.3 GRSP All stages 25.0 129 46 1,083.0 0.958 0.006 33.8 DICK All stages 25.5 204 117 1,920.5 0.939 0.005 20.1 EAME All stages 25.5 86 50 767.5 0.935 0.009 17.9 FISP All stages 23.0 276 150 1,886.0 0.920 0.006 14.7 ^ Nesting cycle intervals include laying, incubating, nestling, incubation and nestling combined, and all stages combined. '’Expected length (days) of each stage from Ehrlich et al. (1988). Number of nests monitored in each nest cycle interval. Total number of failed nests. ^ Total number of exposure days (Mayfield 1975). Probability of daily nest success (Mayfield 1975). 8 Probability of nest success (Mayfield) through the nesting cycle interval. HESP = Henslow’s Sparrow, GRSP = Grasshopper Sparrow, DICK = Dickcissel, EAME = Eastern Meadowlark, FISP = Field Sparrow. TABLE 3. Annual daily survival and nest success rates for grassland birds at Fort Campbell, Kentucky, 1999-2003. Year Rate Henslow’s Sparrow Grasshopper Sparrow Dickcissel Eastern Meadowlark Field Sparrow 1999 Days(n)“ 25.5(6) 96.0(19) 1 1 1.0(14) 125.5(12) 150.5(23) DSR(SE)'’ 0.948(0.023) 0.910(0.027) 0.920(0.024) ().9()7(0.()24) Success 26.3 9.0 12.0 10.5 2000 Days(n) 276.5(40) 257.5(30) 324.0(40) 149.0(17) 522.5(84) DSR(SE) 0.920(0.016) 0.973(0.010) 0.935(0.014) 0.913(0.023) 0.9 12(0.0 12) Success 13.1 50.2 18.1 9.7 1 1.9 2001 Days(n) 235.5(26) 215.0(26) 755.5(74) 201.5(23) 536.0(71) DSR(SE) 0.958(0.013) ().944(0.()16) ().943(0.()08) ().945(0.()16) ().92()(().012) Success 34.5 23.8 22.4 23.9 14.5 2002 Days(n) 1 16.0(20) 23 1 .0(24) 3 I 1 .0(30) 142.5(15) 298.0(47) DSR(SE) ().94()(().022) ().952(0.()14) ().949(0.013) 0.95 1(0.0 18) ().9()6(().017) Success 21.8 29.5 26.0 27.7 10.2 2003 Days(n) 77.5(20) 283.5(30) 419.0(46) 149.0(19) 379.0(51 ) DSR(SE) ().948(0.025) 0.96 1(0.01 1) 0.936(0.012) ().94()(().()20) ().95()(().01 1 ) Success 27.3 37.2 18.3 20.4 30.5 “ Days = total exposure days, n = number of nests. Nest exposure includes laying, incubation, and nestling stages. DSR = Daily survival rate (Mayfield 1975). 116 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 1, March 2008 TABLE 4. Weighted average and range of published demographic rates for Henslow’s Sparrow, Grasshopper Sparrow, Lield Sparrow, Eastern Meadowlark, and Dickcissel. Species Clutch size Parasitized Apparent success Mayfield success^ Henslow’s Sparrow^ Grasshopper Sparrow^’ Dickcissel' Eastern Meadowlark® Lield Sparrow'’ Average’’ Range’' Average Range Average Range Average Range Average Range 3.8 3.3- 4.4 4.2 3.7- 4.6 4.0 3.7- 4.7 4.4 4. 1-5.2 3.7 3.4- 4.0 0.03 0.00-0.08 0.09 0.00-0.50 0.40 0.03-0.95 0.08 0.00-0.70 0.12 0.00-0.80 0.51 0.19-0.74 0.44 0.15-0.80 0.48 0.46-0.67 0.32 0.30-0.70 0.36 0.10-0.72 0.29 0.07-0.46 0.32 0.07-0.52 0.26 0.12-0.50 0.10 0.10-0.25 0.27 0.21-0.47 “ Probability of nest success corrected for exposure time (Mayfield 1975). Robins 1971, Peck and James 1987, Winter 1998, Reinking et al. 2000, Monroe 2001, Burhans 2002. Average for published values weighted by sample size. ' p“f%l4'!tltu9T8" 1982. McNai, 1987, Vick.ry e, al, 1992, Winter 1998, ReInking e, al. 2000. B.len, and Normen. 2003. V„a 2003, ?963- G^s r9™Zimn,erm.n 1982, 1983: Basil. 1997; Winter 1998: Vos 2003: Jensen and Finck 2004, Fletcher e, al. 2006_ Blunders 1932, Johnston 1964, Roseberry and Klimstra 1970, Elliot 1978, Peck and James 1987, Knapton 1988, Lanyon 1995, Granfors et al. 1996, ^'"14^1^1934, Walkinshaw 1939, Crooks 1948, Best 1978, Wray et al. 1982, Carey et al. 1994, Barber et al. 2001, Vos 2003. one female each. Both nests were lost to pre- dation before the eggs hatched, and we were not able to confirm the number of parents vis- iting the nest. Grassland birds have relatively low nesting success, compensated by several nesting at- tempts within a single season (Wiens 1969, Martin 1995, Winter 1999). The length of the nest initiation window (51-75 days) suggests the possibility of multiple breeding attempts or multiple successful broods within a single breeding season for all five species. Henslow’s Sparrows and Grasshopper Sparrows are gen- erally considered at least double brooded; two pairs from a color-banded Henslow’s Sparrow population in Kentucky had three successful broods in one season (Monroe 2001). Some nests initiated in July could represent third at- tempts or successful broods for some nesting pairs. The amount of time from the start of the nesting season (late Apr) and the last nests (early Aug) allows for the possibility of three broods if the time to finish a complete nesting cycle is less than 32 days including nest build- ing (Ehrlich et al. 1988). Dickcissels are con- sidered single brooded, or may move to a dif- ferent location to re-nest (Winter 1998), but the nest initiation window for Dickcissels at Fort Campbell was sufficiently long to allow for the possibility of double brooding (5 1 days). Field Sparrows are considered double brooded, but their nesting success was so low at Fort Campbell that few pairs could produce two successful nests within the nesting season without many unsuccessful nesting attempts. Field Sparrows had sufficient time to com- plete at least two successful nests with several unsuccessful attempts (64 days). The length of the nest initiation window for Eastern Meadowlarks at Fort Campbell was sufficiently long for three successful nesting attempts, but Eastern Meadowlarks may delay for a longer period between successive nests than expected. Kershner et al. (2004) radio- tracked female Eastern Meadowlarks in Illi- nois and reported although they had time in the season to nest more than once, many birds did not re-nest in the same territory. This be- havior would spread the distribution of nest- ing attempts across the season, and could ac- count for the long nesting season for Eastern Meadowlark (75 days) in our study. Clutch size decreased during the nesting season for Dickcissels, Grasshopper Spar- rows, and Field Sparrows. The second brood would be reduced by about one for Dickcis- sels, and about 0.5 for Grasshopper Sparrows and Field Sparrows if these species were dou- ble brooded. Clutch size did not show a linear relationship with time of egg laying during the nesting season for Henslow’s Sparrows and Eastern Meadowlarks. Winter (1998) also re- Giocomo et al. • NESTING GRASSLAND BIRDS 117 ported a lack of relationship between clutch size and time in nesting season for Henslow’s Sparrow. Eastern Meadowlarks and Henslow’s Sparrows tended to have smaller clutch sizes at the beginning and end of the nesting season. The fields monitored at Fort Campbell were used extensively for army training exercises, such as airborne-troop parachute drops and as- sociated vehicle activity, throughout the breeding season for grassland birds. However, most (88%) recorded nest losses were attri- buted to predation and few (<1%) nests were affected directly by military activities. Vehicle or troop movements crushed a few nests. Mowing for hay and weather accounted for more recorded nest losses (3 and 1.7%, re- spectively) than military activities. Nest searching activities were concentrated in grasslands not managed specifically for hay production, and land management effects ob- served were not representative of all grass- lands in the area. Undoubtedly, a much larger proportion of nests failed because of mowing in grassland fields managed for hay produc- tion. However, nest searching was concentrat- ed in areas used extensively for military train- ing, and nest failure rates may be considered representative of direct military training im- pacts at Fort Campbell. ACKNOWFEDGMENTS The Department of Defense Legacy Resource Man- agement Program, Fort Campbell, and the University of Tennessee Agricultural Experiment Station spon- sored this project. We thank the staff in the Conser- vation Branch at Fort Campbell for assistance with project logistics and the many field assistants who helped search for nests. LITERATURE CITED Askins, R. a. 1993. Population trends in grassland, shrubland, and forest birds in eastern North Amer- ica. Current Ornithology 1 1:1-34. Balent, K. L. and C. J. Norment. 2003. Demographic characteristics of a Grasshopper Sparrow popula- tion in a highly fragmented landscape of western New York State. Journal of Field Ornithology 74: 341-348. Barber, D. R., T. E. Martin, M. A. Melchiors, R. E. Thill, and T. B. Wigley. 2001. Nesting success of birds in different silvicultural treatments in .southeastern U.S. pine forests. Con.servation Bi- ology 15:196-207. Basili, G. D. 1997. Continental scale ecology and con- servation of the Dickcissel. Dissertation. Univer- sity of Wisconsin, Madison, USA. Best, L. B. 1978. Field Sparrow reproductive success and nesting ecology. Auk 95:9-22. Burhans, D. E. 2002. Conservation assessment for Henslow’s Sparrow Ammodramus henslowii. USDA, Forest Service, General Technical Report NC-226. North Central Research Station, St. Paul, Minnesota, USA. Carey, M., D. E. Burhans, and D. A. Nelson. 1994. Field Sparrow {Spizella pusilla). The birds of North America. Number 103. Chester, E. W., B. E. Wofford, J. M. Baskin, and C. C. Baskin. 1997. A floristic study of barrens on the southwestern Pennyroyal Plain, Kentucky and Tennessee. Castanea 62:161-172. Crooks, M. P. 1948. Field Sparrow life history in cen- tral Iowa. Iowa Bird Life 23:10-13. Delcourt, P. a., H. R. Delcourt, D. F. Morse, and P. A. Morse. 1993. History, evolution, and orga- nization of vegetation and human culture. Pages 47-80 in Biodiversity of the southeastern United States (W. H. Martin, S. G. Boyce, and A. C. Echternacht, Editors). John Wiley and Sons, New York, USA. Eberly, C. 2002. Defending the stepping stones of mi- gration. Birding 34:450-458. Ehrlich, P. R., D. S. Dobkin, and D. Wheye. 1988. The birder’s handbook. Simon and Schuster Inc., New York, USA. Elliot, P. F. 1978. Cowbird parasitism in the Kansas tallgrass prairie. Auk 95:161-167. Fletcher Jr., R. J., R. R. Koford, and D. A. Seaman. 2006. Critical demographic parameters for declin- ing songbirds breeding in restored grasslands. Journal of Wildlife Management 70:145-157. Giocomo, J. 2005. Grasshopper Sparrows nesting in eastern Tennessee. The Migrant 75:48-49. Grantors, D. A., K. E. Church, and L. M. Smith. 1996. Eastern Meadowlarks nesting in rangelands and Conservation Reserve Program fields in Kan- sas. Journal of Field Ornithology 67:222-235. Gross, A. O. 1968. Dickcissel. Pages 158-191 in Life histories of North American cardinals, grosbeaks, buntings, towhees, finches, sparrows and their al- lies (O. L. Austin, Editor). Part 1. U.S. National Museum Bulletin 237. Hauber, M. E. and S. a. Russo. 2000. Perch prox- imity correlates with higher rates of cowbird par- asitism of ground nesting Song Sparrows. Wilson Bulletin 112:150-153. Herkert, j. R. 1994. The effects of habitat fragmen- tation on midwestern grassland bird communities. Ecological Applications 4:461-471. Herkeri, j. R., D. W. Sample, and R. E. Warner. 1996. Management of Midwestern grassland land- scapes for the conservation of migratory grassland birds. Pages 89-1 16 in Management of Midwest- ern landscapes for the conservation t)f neotropical migratory birds (E R. 'fhompson HI. Editor). USDA, Forest Service, General fechnical Report NC-187. North Central Research Station. St. Paul. Minnesota. USA. 118 THE WILSON JOURNAL OF ORNITHOLOGY • VoL 120, No. 1, March 2008 Hicks, L. E. 1934. A summary of cowbird host species in Ohio. Auk 51:385-386. Jensen, W. E. and J. E Cully Jr. 2005. Density-de- pendent habitat selection by Brown-headed Cow- birds {Molothrus ater) in tallgrass prairie. Oecol- ogia 142:136—149. Jensen, W. E. and E. J. Finck. 2004. Edge effects on nesting Dickcissels {Spiza americana) in relation to edge type of remnant tallgrass prairie in Kan- sas. American Midland Naturalist 151:192-199. Johnson, R. G. 1979. Estimating nest success: the Mayfield method and an alternative. Auk 96:651- 661. Johnston, R. E 1964. The breeding birds of Kansas. Museum of Natural History. University of Kansas Publication 12:575-655. Kershner, E. L., j. W. Walk, and R. E. Warner. 2004. Breeding-season decisions, renesting, and annual fecundity of female Eastern Meadowlarks {Sturnella magna) in southeastern Illinois. Auk 121:796-805. Knapton, R. W. 1988. Nesting success is higher for polygynously mated females than for monoga- mously mated females in the Eastern Meadowlark. Auk 104:325-338. Lanyon, W. E. 1995. Eastern Meadowlark {Sturnella magna). The birds of North America. Number 160. Long, C. A. 1963. Production of sterile eggs in the Dickcissel. Wilson Bulletin 75:456. Martin, T. E. 1995. Avian life history evolution in relation to nest sites, nest predation, and food. Ecological Monographs 65:101-127. Marzluff, j. M. and R. Sallabanks. 1998. Past ap- proaches and future directions for avian conser- vation biology. Pages 5-14 in Avian conservation: research and management (J. M. Marzluff and R. Salabanks, Editors). Island Press, Washington, D.C., USA. Mayfield, H. 1961. Nest success calculated from ex- posure. Wilson Bulletin 73:255—261. Mayfield, H. E 1975. Suggestions for calculating nest success. Wilson Bulletin 87:456—466. McNair, D. B. 1987. Egg data slips — are they useful for information on egg-laying dates and clutch size? Condor 89:369-376. Monroe, M. S. 2001. Reproductive success of Hen- slow’s Sparrows in Kentucky: effects of season, nest characteristics, and vegetation structure. The- sis. Eastern Kentucky State University, Richmond, USA. Morris, D. L. and E R. Thompson III. 1998. Effects of habitat and invertebrate density on abundance and foraging behavior of the Brown-headed Cow- bird. Auk 115:376-385. Moss, E. D. 2001. Distribution and reproductive suc- cess of native grassland birds in response to burn- ing and field size at Fort Campbell Military Res- ervation: special focus on Henslow’s and Grass- hopper sparrows. Thesis. University of Tennessee, Knoxville, USA. Nicholson, C. P. 1997. Atlas of the breeding birds of Tennessee. University of Tennessee Press, Knox- ville, USA. Peck, G. and R. James. 1987. Breeding birds of On- tario: nidiology and distribution. Volume 2. Pas- serines. Miscellaneous Publications. Royal Ontar- io Museum of Life Sciences, Toronto, Ontario, Canada. Peterjohn, B. G. and J. R. Sauer. 1999. Population status of North American grassland birds from the North American Breeding Bird Survey 1966- 1996. Studies in Avian Biology 19:27-44. Price, H. E 1934. A few cowbird notes. Oologist 51: 107-108. Reinking, D. A., D. A. Wiedenfeld, D. H. Wolfe, and R. Rohrbaugh. 2000. Distribution, habitat use, and nesting success of Henslow’s Sparrows in Oklahoma. Prairie Naturalist 32:219-232. Robins, J. D. 1971. A study of Henslow’s Sparrow in Michigan. Wilson Bulletin 83:39—48. Roseberry, j. L. and W. D. Klimstra. 1970. The nest- ing ecology and reproductive performance of the Eastern Meadowlark. Wilson Bulletin 82:243- 267. Sauer, J. R. and J. E. Hines. 1989. Testing for differ- ences among survival or recovery rates using pro- gram CONTRAST. Wildlife Society Bulletin 17: 549-550. Sauer, J. R., J. E. Hines, and J. Fallon. 2004. The North American Breeding Bird Survey, results and analysis 1966—2003. Version 2004.1. USGS, Patuxent Wildlife Research Center, Laurel, Mary- land, USA. Online at (http://www.mbr-pwrc.usgs. gov/bbs/bbs.html) (accessed 1 February 2005). Saunders, G. B. 1932. A taxonomic revision of the meadowlarks of the genus Sturnella (Vieillot) and natural history of the Eastern Meadowlark, Stur- nella magna magna (Linnaeus). Dissertation. Cor- nell University, Ithaca, New York, USA. Sutter, B. and G. Ritchison. 2005. Effects of grazing on vegetation structure, prey availability, and re- productive success of Grasshopper Sparrows. Journal of Field Ornithology 76:345—351. Van Horne, B. 1983. Density as a misleading indi- cator of habitat quality. Journal of Wildlife Man- agement 47:893—901. Vickery, P. D. 1996. Grasshopper Sparrow (Ammo- dramus savannarum). The birds of North Ameri- ca. Number 239. Vickery, P. D., M. L. Hunter, and J. V. Wells. 1992. Is density an indicator of breeding success? Auk 109:706-710. Vos, S. M. 2003. Distribution, abundance, and pro- ductivity of grassland birds on Fort McCoy mili- tary base. Thesis. University of Wisconsin, Mad- ison, USA. Walkinshaw, L. H. 1939. Nesting of Field Sparrow and survival of young. Bird-banding 10:107-114, 149-157. Wiens, J. A. 1969. An approach to the study of eco- Giocomo et al. • NESTING GRASSLAND BIRDS 119 logical relationships among grassland birds. Or- nithological Monographs 8. Winter, M. 1998. Effect of habitat fragmentation on grassland-nesting birds in southwestern Missouri. Dissertation. University of Missouri, Columbia, USA. Winter, M. 1999. Nesting biology of Dickcissels and Henslow’s Sparrows in southwestern Missouri prairie fragments. Wilson Bulletin 111:515-527. Winter, M., J. A. Shaffer, D. H. Johnson, T. M. Don- ovan, W. D. SVEDARSKY, R W. JONES, AND B. R. Euliss. 2004. Habitat and nesting of Le Conte’s Sparrows in the northern tallgrass prairie. Journal of Eield Ornithology 76:61-71. Wray II, T. K. A. Strait, and R. C. Whitmore. 1982. Reproductive success of grassland sparrows on a reclaimed surface mine in West Virginia. Auk 99: 157-164. Zar, J. H. 1998. Biostatistical analysis. Eourth Edition. Prentice-Hall, Upper Saddle River, New Jersey, USA. Zimmerman, J. L. 1982. Nesting success of Dickcissels in preferred and less preferred habitats. Auk 99: 292-298. Zimmerman, J. L. 1983. Cowbird parasitism of Dick- cissels in different habitats at different nest den- sities. Wilson Bulletin 95:7-25. The Wilson Journal of Ornithology 1 20( 1 ): 1 20-1 30, 2008 FACTORS AFFECTING HOME RANGE SIZE AND MOVEMENTS OF POST-FLEDGING GRASSLAND BIRDS KIMBERLY M. SUEDKAMP WELLS, JOSHUA J. MILLSPAUGH,' MARK R. RYAN,' AND MICHAEL W. HUBBARD^ ABSTRACT.^We describe post-fledging movements and evaluate the effects of local vegetation, temporal, and biological factors on home range size for two species of declining grassland birds m southwestern Missouri from 2002 to 2004. We obtained >30 detections for 74 individual juvenile Dickcissels (Spizn amencana)and wTuvenile Eastern Meadowlarks (Sturnella magna) during the post-fledging period. Juvenile Eastern Mead- owlarks had a greater total number of days (6.7 ± 0.6) with large 0300 m) movements J;'™"''" D.ek^^^ (30 + 03 days). Average Dickcissel home range size was larger and three times as variable in 2002 (77^ 22 ha“) compared to 2003 (31.4 ± 7.5 ha) and 2004 (34.9 ± 7.5 ha). There were year-specific effects of the variabmtv in vegetation height on home range size of juvenile Dickcissels. Home range size was similar among years for juvenile Eastern Meadowlarks (2002, 109.4 ± 39.2 ha; 2003, 82.7 ± 29.4 ^ but there was a year-specific effect of variability in grass cover on home range size of juvenile Eastern Mead owlaAs Local vegetaton conditions are important factors affecting home range size and movements during post-fledging period. Received 28 August 2006. Accepted 28 March 2007. The post-fledging period is widely recog- nized as the least understood portion of the avian life cycle (Baker 1993), primarily due to constraints associated with monitoring ju- veniles during natal dispersal (Greenwood and Harvey 1982). Natal dispersal typically in- cludes the period between the time a juvenile fledges and returns to breed the following year as an adult (Greenwood and Harvey 1982). During the post-fledging period (a subset of the natal dispersal period), defined as the pe- riod between fledging and departure for mi- gration or settling in wintering areas (Anders et al. 1998), juveniles become independent from parents and develop flight and foraging skills. Simultaneously, juveniles must also learn to avoid predators and cope with adverse weather conditions that shape components of post-fledging dispersal such as movement pat- terns. Movement patterns may subsequently influence population dynamics by affecting the probability of juvenile survival (Magrath 1991, Cohen and Lindell 2004, Kershner et al. ’ Department of Fisheries and Wildlife Sciences, University of Missouri-Columbia, Columbia, MO 65211, USA. 2 Resource Science Division, Missouri Department of Conservation, Jefferson City, MO 65102, USA. 3 Current address; H.T. Harvey & Associates, 983 University Avenue, Building D, Los Gatos, CA 95032, USA. Corresponding author; e-mail; kwells@harveyecology.com 2004, Wiens et al. 2006, Yackel Adams et al. 2006), resource selection (Anders et al. 1998, Vega Rivera et al. 1998, Kennedy and Ward 2003, Berkeley 2005), recruitment (Green and Cockburn 2001), genetic structure (Dinge- manse et al. 2003), and breeding dispersal as an adult (Kenward et al. 2001). Thus, describ- ing post-fledging dispersal and potential fac- tors affecting movement patterns is important for developing bird conservation strategies. Vegetation, temporal, and biological factors are the dominant hypotheses that have been used to explain movement patterns during the post-fledging period. Local vegetation condi- tions may be important for providing site-se- lection cues associated with food availability or protection from predators. For example, ju- venile Northern Goshawks (Accipiter gentilis) that received supplemental food on their natal territories had earlier onset of dispersal move- ments from the natal area and no permanent dispersal from the study area compared to control birds (Kennedy and Ward 2003). Veg- etation structure has also been shown to be an important local-scale factor affecting move- ments of juvenile Dickcissels {Spiza ameri- cana) that seek favorable vegetation condi- tions for maximizing concealment from pred- ators (Berkeley 2005). Temporal factors, such as year-specific trends in weather conditions, have also been associated with movement pat- terns of Lark Buntings {Calamospiza melan- ocorys) during the post-fledging period (Yack- 120 Suedkamp Wells et al. • POST-FLEDGING MOVEMENTS OE GRASSLAND BIRDS 121 el Adams et al. 2001). Individual or cohort- level biological factors are the third major hy- pothesis associated with movements during the post-fledging period. Juvenile Greater Fla- mingos (Phoenicopterus roseus) in good body condition had a higher probability of move- ment during the post-fledging period than ju- veniles in poor body condition (Barbraud et al. 2003). Low nestling rank or body mass has also been associated with higher dispersal probabilities in juvenile House Sparrows {Passer domesticus) that were in poor condi- tion compared to nestlings of higher rank or body mass (Altwegg et al. 2000). We describe home range size and daily movement patterns during the post-fledging period for two species of declining, grassland songbirds: Dickcissel and Eastern Meadow- lark (Sturnella magna). Both are grassland- breeding songbirds that have shown regional and national declines over the last several de- cades (Sauer et al. 2005). Dickcissels are also on the Partners in Flight Continental Watchlist as a species of conservation concern (Rich et al. 2004). We were interested in home range size and daily movement patterns of these spe- cies because both aspects of the post-fledging period are largely undescribed for grassland birds. Our specific objectives were to: (1) compare daily movements of both species, (2) estimate home range size, and (3) model the effects of vegetation, temporal, and biological factors on home range size. METHODS Study Area. — We conducted the study at Ta- berville Conservation Area (38° N, 93° W) and Wah’Kon-Tah Prairie (37° N, 94° W) in Cedar and St. Clair counties in southwestern Missouri. Taberville Conservation Area is a 680-ha prairie owned and managed by the Missouri Department of Conservation (MDC). Wah’Kon-Tah Prairie is a 1,930-ha prairie at the northern periphery of El Dorado Springs, Missouri. It is owned by the Missouri Chapter of The Nature Conservancy (TNC), and joint- ly managed by MDC and TNC. Both sites are similar and embedded in an agricultural ma- trix of crops (wheat, soybeans, and corn) and grazing lands (pasture, hay fields) on sur- rounding private property. Dominant land management practices on both sites include livestock grazing, prescribed burning, seed harvesting for native plants, and hay produc- tion. The study sites are divided into units that receive management (primarily prescribed burning or mowing) applied by the same MDC personnel at least once every 3 years. Dominant vegetation is native grasses includ- ing big bluestem (Andropogon gerardi), little bluestem (Schizachyrium scoparium), and In- dian grass {Sorghastrum nutans). Eorb species include coneflowers {Echinacea spp.), white wild indigo {Baptisa alba), blazing star {Lia- trus spp.), compass plant {Silphium lacinia- tum), milkweeds {Aesclepias spp.), and sun- flowers {Helianthus spp.). Woody species in- clude smooth sumac {Rhus glabra), persim- mon {Diospyros virginiana), blackberry {Rubus spp.), and roses {Rosa spp.). The av- erage annual precipitation is 102 cm with most occurring during the growing season be- tween April and August (Missouri Climate Center 2005). Eall 2002 marked the start of a severe drought that lasted into summer 2003 and was the most severe drought on record in the last two decades. Bird Handling and Marking. — We located nests of both species using systematic search- es and focused surveys in areas with Dickcis- sels and Eastern Meadowlarks, and appropri- ate nesting vegetation. We conducted nest searches every day from 0600 to 1400 hrs CDT during the nesting season between the third week of April and the second week of August from 2002 to 2004. We used behav- ioral cues of parents to indicate the presence of a nest. We recorded GPS coordinates and marked the location of each nest by placing colored flagging tape at least 5 m distant. We recorded the species, content, and parental ac- tivity at each nest. If the nest contained nest- lings, we attempted to classify their age based on presence of down feathers, whether eyes were open or not, extent of pin feather devel- opment, or presence of a full complement of feathers (KSW, unpubl. data) as well as pub- lished information (Eastern Meadowlarks only; Lanyon 1995). We monitored each nest every 3-4 days until 2-3 days prior to the expected fledging date and then switched to daily nest checks. We attached a metal UvSGS band to the left leg and a unique combination of plastic, UV-rcsistant Dar\ ic bands (Avinct, Dryden, NY, USA) to the right leg 2—3 days prior to fledging (6-7 days of age for Dick- 122 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 cissels; 9-1 1 days of age for Eastern Mead- owlarks), and weighed each individual to the nearest gram using a spring scale (Avinet, Dryden, NY, USA). We modified the Rappole and Tipton meth- od (1991) (Suedkamp Wells et al. 2003) and attached 0.7-g transmitters with a 10-cm whip antenna (Biotrack, Dorset, UK) to the back of each bird using a leg harness. We constructed the leg harness from cotton, elastic beading cord to allow room for growth. Battery life for each transmitter was expected to range be- tween 55 and 60 days. We secured the bottom of the transmitter to the back of the bird using super glue (Duro, Avon, OH, USA) and placed nestlings back in the nest. Forced fledging was occasionally observed during the beginning of the study in 2002, but modifi- cation of our protocol by attaching transmit- ters to younger nestlings minimized subse- quent forced fledging events. We generally at- tached transmitters to every member of a brood because prior experience during a pilot effort indicated we would have insufficient sample sizes if we only selected one individ- ual per nest (KSW, pers. obs.). Inclusion of more than one sibling per brood may artifi- cially appear to reduce variability, but empir- ical evaluations have demonstrated that ran- dom subsampling of multiple individuals within a cohort does not improve the precision or accuracy of home range estimates or their associated variability with kernel estimators (De Sofia et al. 1999). We also demonstrated elsewhere that juvenile survival estimates for multiple individuals from the same brood are independent (KSW, unpubl. manuscript) Thus, we assumed home range sizes from individ- uals in the same brood were also independent. Handling and processing time usually was be- tween 2-5 min per bird. Radiotracking. — We began radiotracking birds using homing techniques (Mech 1983) the day after attaching transmitters. We were successful at obtaining visual locations of each bird during a tracking attempt at least 90% of the time (Suedkamp Wells 2005). We returned each subsequent morning if the brood remained in the nest and began tracking when at least one brood member fledged. Our goal was to track at least 25 individuals of both species for a minimum of 50 detections each nesting season (Carton et al. 2001). Initially we located each juvenile twice each day. We reduced tracking to one attempt per day after obtaining at least 50 detections of an individ- ual within the post-fledging season to accom- modate tracking more individuals during peak nesting periods. We systematically located ju- veniles during four daily time blocks corre- sponding to key periods of biological activity. The four tracking blocks were early morning (0600-0930 hrs), mid-morning (0930-1230 hrs), afternoon (1230-1700 hrs), and evening (1700-2130 hrs). We alternated among the four time periods during each tracking day to obtain two detections per day, one each in non-consecutive detection time periods. We avoided tracking before 0600 hrs and after 2130 hrs to reduce risk of stepping on juve- niles when they could not be visually located. We tracked each individual until we recov- ered the transmitter (with or without a carcass) through the end of August each post-fledging season. We performed extensive searches of the immediate area on foot immediately after being unable to locate an individual. If unsuc- cessful, we broadened the search to include all roads within a 3-km radius of the last known location and study site boundaries. We at- tempted to locate missing birds from aerial searches of 5 km strips over the study area in a helicopter twice monthly between 1 June and 30 August each year. Statistical Analyses. — We restricted our sample of marked juveniles to those individ- uals with >30 detections for subsequent anal- yses because simulation research has indicated 30 is the minimum sample size required for stable home range estimates using kernel es- timators (Seaman et al. 1999). We calculated the distance moved between locations using GPS coordinates from tracking sessions for each individual in the Animal Movements extension (Hooge and Eichenlaub 2000) of Arc View 3.3 (Environmental Systems Re- search Institute, Redlands, CA, USA). We cal- culated average daily movements by classi- fying each tracking event by age (days since fledging) of the juvenile and then averaging within age classes across individuals. We ex- tracted the first age that each individual had daily movements >300 m and the total num- ber of days with movements >300 m. We used 300 m as a threshold because visual in- spection of average daily distances moved for Suedkamp Wells et al. • POST-FLEDGING MOVEMENTS OF GRASSLAND BIRDS 123 < Time since fledging (days) FIG. 1. Average (± SE) daily movements (m) for juvenile Dickcissels (n = 74) and Eastern Meadowlarks (n = 64) for individuals with >30 detections in southwestern Missouri, 2002-2004. both species indicated relatively consistent trends in point estimates and their standard er- rors until movements exceeded this threshold (Fig. 1). We used two-sample r-tests in SYS- TAT (SPSS 1999) to compare the first age that juveniles of both species showed daily move- ments exceeding 300 m and the total number of days where daily movements exceeded 300 m. We also included reference to large move- ments (>600 m) because these movements were associated with biologically meaningful periods such as flocking with other juveniles of the same species. We describe average dai- ly movements for comparison with other pub- lished studies that commonly include move- ment parameters and because we were inter- ested in fine-scale movements that comprise the home range estimate. We elected not to evaluate year effects on average daily move- ments because they represent a non-indepen- dent subset of all movements used to estimate home range size for an individual. We used an alpha level of 0.05 to infer significant re- sults. We calculated 95% home range contours using a fixed-kernel estimator (Worton 1987, 1989) from the ‘KDE folder’ (Beardah and Baxter 1995) in MATLAB (Mathworks 1999). Kernel estimators are favored for home range estimates because they are superior to several other methods (Kernohan et al. 2001). We used the “plug in” method and smoothed the jc and j coordinates independently for the fixed-kernel estimator (Wand and Jones 1995, Jones et al. 1996). We tested for year effects on home range size using 95% confidence in- tervals. We reduced mu Iticol linearity and tested whether assumptions were violated with each explanatory factor prior to analysis. We used probability plots in wSYSTAT to assess the dis- tribution of each variable. We applied a log transformation for percent data (grass or forb cover) and a square root transformation for all 124 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 other variables if we detected evidence of non-normality (Steel et al. 1997). We used a data reduction approach in SYSTAT to detect potential problems with multicollinearity by visually inspecting scree plots from the Prin- cipal Components Analysis (PCA). We re- moved all variables except the most highly correlated variable for subsequent analyses of the reduced data set if multiple variables were highly correlated with the first PCA axis and did not contribute unique information based on visual inspection. We used the General Linear Model module in SYSTAT to evaluate the effects of local vegetation, temporal, and biological factors on home range size with the reduced data set. Lo- cal vegetation factors included vegetation height (cm), grass cover (%), and forb cover (%) measured in a 20 X 50 cm Daubenmire frame (Daubenmire 1959) at each tracking lo- cation for each individual. We selected these variables a priori because prior analyses in- dicated they were important predictors of ju- venile resource selection (Suedkamp Wells 2005). Other habitat features, such as woody cover, were also important for micro-scale re- source selection (Suedkamp Wells 2005), but less appropriate for this analysis because pat- terns of variability were negligible when av- eraged across all movement locations used to generate home range size for an individual. We used the CV instead of average values of local vegetation variables because prior habi- tat selection analyses showed variability in vegetation structure was more important and most appropriately reflected the patchy envi- ronment of grassland vegetation (Suedkamp Wells 2005). We included temporal effects of year (2002-2004) because prior analyses dem- onstrated year-specific variability in home range size (Suedkamp Wells 2005). We did not include site as a temporal factor because prior analyses indicated large-scale differenc- es in resource selection were weak and that within-site variability was more important than variability across sites (Suedkamp Wells 2005). Biological factors included juvenile body mass (g) at time of transmitter attach- ment, maximum observed clutch size (range 1-5 for both species), maximum number of siblings observed (range 1-5), and order of fledging (coded 1 for first fledging event and 2 for subsequent fledging events). Tracking re- cently fledged juveniles twice daily allowed us to record first and second fledging bouts, but our resolution was coarse because we were unable to differentiate order of fledging among juveniles that fledged after the second bout. We fit single parameter models and bi- ologically meaningful interactions to the re- duced data set. We used 95% confidence in- tervals to elucidate the simple effects driving the interaction if significant interactions were present. RESULTS Sample Description. — We located 258 Dickcissel nests and 113 Eastern Meadowlark nests during the 3-year study period. Average (± SE) clutch and brood sizes were 4.1 ± 0.9 eggs and 2.9 ± 1.3 nestlings for Dickcissels, and 4.4 ± 1.0 eggs and 3.3 ± 1.1 nestlings for Eastern Meadowlarks. We attached radio transmitters to 248 Dickcissel nestlings from 94 broods and 164 Eastern Meadowlark nest- lings from 46 broods. Juveniles of both spe- cies generally transitioned to independence at 2-3 weeks of age after beginning gradual movements from the natal site. The subset of individuals with at least 30 detections used for subsequent home range and movement anal- yses included 74 juvenile Dickcissels (2002, n = 25; 2003, n = 19; 2004, n = 30) and 64 juvenile Eastern Meadowlarks (2002, n = 8; 2003, n = 26; 2004, n = 30). Individuals with >30 detections reflected the full range of brood sizes (1-5 nestlings) for both species in 2002 and 2004, but did not include the larger brood sizes for Dickcissels (4-5 nestlings) in 2003. Average nestling body mass at time of marking for Dickcissels with >30 detections (15.3 ± 0.4 g; 95% Cl 14.6-16.0 g) was com- parable to individuals with transmitters but with <30 detections (14.7 ± 0.4 g; 95% Cl 14.0-15.4 g). Average nestling body mass for juvenile Eastern Meadowlarks was also simi- lar for individuals with >30 detections (44.9 ± 0.8 g; 95% Cl 43.3-46.5 g) and <30 de- tections (44.8 ± 1.3 g; 95% Cl 42.2-47.5 g). Mortality during the first 2 weeks out of the nest was high with only 56% of juvenile Dick- cissels and 57% of juvenile Eastern Mead- owlarks surviving. The mean date of fledging (day zero out of the nest) was 25 June ± 1 day (range 5 Jun-24 Jul) for Dickcissels and 5 June ± 2 days (range 7 May-23 Jul) for Siiedkanip Wells et al. • POST-FLEDGING MOVEMENTS OF GRASSLAND BIRDS 125 Eastern Meadowlarks with >30 detections per individual. The average number of detections per individual with at least 30 detections was 58.0 ± 1.2 (range 34-75) and 53.8 ± 1.1 (range 30—88) respectively, for Dickcissel and Eastern Meadowlark juveniles. Movements. — Average daily movements by juvenile Dickcissels were relatively consistent although increasing at a slow rate until 28 days post fledging, which corresponds to the first 2 weeks of independence. Dickcissels had large daily movements (>600 m) starting at 28 days of age (Fig. 1 ). The period of largest movements occurred between days 32 and 38. Juvenile Eastern Meadowlarks also had slow but consistently increasing rates of daily movement until day 28. Juvenile Eastern Meadowlarks after day 28 had slightly larger movements and had three isolated peaks of large movement (>600 m) around days 40, 44, and 54 (Fig. 1 ). Peaks of large movements corresponded with altered periods of activity for juvenile meadowlarks when large daily movements were more characteristic as juve- niles began congregating with other juvenile meadowlarks in roving flocks. Juveniles of both species frequently foraged or moved into surrounding private land beginning at 3-4 weeks of age. The mean age that juveniles had daily movements >300 m was similar for Dickcis- sels (17.7 ± 1.4 days) and meadowlarks (20.1 ± 1.4 days; P = 0.20). The total number of days with daily movements >300 m differed between Dickcissel (5.0 ± 0.5 days) and East- ern Meadowlark (6.7 ± 0.6 days) juveniles {P = 0.020). The majority of juvenile Dickcissels (62%, n = 46) had 8 or fewer days with daily movements >300 m (Fig. 2). The peak of dai- ly movements >300 m occurred between days 4 and 8 for most juvenile Eastern Meadow- larks (41%, n = 26) and the distribution of individuals in each category decreased more slowly compared to Dickcissel juveniles (Fig. 2). Home Ran^e 57ze.— Preliminary data reduc- tion for Dickcissels with PCA showed that or- der of fledging = 0.59), clutch size = 0.77), and number of siblings (P = 0.85) were highly correlated in order of increasing mag- nitude. Therefore, we removed order of fledg- ing and clutch size from subsequent analyses to reduce multicollinearity. fhe reduced data 30 0 4 8 12 16 Total days with movements exceeding 300 m FIG. 2. Total number of days with large move- ments (>300 m) for Juvenile Dickcissels {n = 74; open bars) and Eastern Meadowlarks (n = 64; filled bars) with >30 detections in southwestern Missouri, 2002- 2004. set was tested for several single parameter ef- fects (juvenile body mass, number of siblings, year, forb cover CV, and vegetation height CV) and two interactions (year and forb cover CV, year and vegetation height CV). Average Dickcissel home range size was larger and at least three times as variable in 2002 (77.0 ± 22.1 ha; 95% Cl 31.3-122.6 ha) compared to 2003 (31.4 ± 7.5 ha; 95% Cl 15.7-47.0 ha) and 2004 (34.9 ± 7.5 ha; 95% Cl 19.6-50.2 ha). However, 95% confidence intervals over- lapped across years indicating home range siz- es were not statistically different. Average home range size pooled across years was 48.8 ± 8.8 ha. There was a significant (P = 0.018) interaction between year and the CV of veg- etation height, which explained a large amount of variation in Dickcissel home range size (multiple = 0.54). The average CV of vegetation height for Dickcissels was higher and twice as variable in 2003 (0.59, 95% Cl 0.49-0.69) compared to 2002 (0.40, 95% Cl 0.36-0.45) and 2004 (0.47, 95% Cl 0.43- 0.51) based on non-overlapping 95% confi- dence intervals. The relationship between Dickcissel home range size and vegetation height CV varied more by home range size than vegetation height CV in 2002. This in- dicated the range of vegetation heights select- ed by juvenile Dickcissels during post-lledg- ing dispersal was relatively consistent but as- sociated with larger home ranges (Fig. 3). However, Dickcissel home ranges in 2003 and 2004 were characterizetl by a wider range of vegetation heights, indicating juveniles were selecting more variable vegetation heights that Home range size (ha) 126 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 3001 250 200 ■ 150- X. ■ ■ ■* 100 50- 0- 0 1 1 0.2 0.4 0.6 0.8 Vegetation height CV CM CO o o o o o o CM CM CM ■ <1 X 500 -| 'cO £ 400 - 30 detections in southwestern Missouri. FIG. 4. Home range size (95% fixed kernel) for juvenile Eastern Meadowlarks (n = 64) as a function of grass cover CV and year (2002, filled squares; 2003, open triangles; 2004, asterisks) for individuals with >30 detections in southwestern Missouri. may have facilitated smaller home ranges (Fig. 3). Preliminary PCA results for Eastern Mead- owlarks indicated that order of fledging (r^ = 0.22), number of siblings (r^ = 0.59), and clutch size (r^ = 0.71) were correlated in in- creasing order of magnitude and we removed the former two variables from subsequent analyses to reduce multicollinearity. The cor- relation between order of fledging and the oth- er two variables was lower than the other var- iables removed, but visual inspection of the scree plot indicated order of fledging did not contribute additional information. The re- duced data set was tested for effects of several single parameters (juvenile body mass, clutch size, year, grass cover CV, and vegetation height CV) and several interactions (year and grass cover CV, year and vegetation height CV). Patterns of home range size for Eastern Meadowlarks were similar across years (2002, 109.4 ± 39.2 ha, 95% Cl 19.0-199.8 ha; 2003, 82.7 ± 29.4 ha, 95% Cl 22.2-143.2 ha; 2004, 70.7 ± 11.6 ha, 95% Cl 47.0-94.5 ha) indicated by overlapping 95% confidence in- tervals, but variability decreased with succes- sive years and increasing sample sizes in the study. Average home range size pooled across years was 80.9 ± 13.9 ha. There was a sig- nificant (P = 0.030) interaction between year and the CV of grass cover, which explained a large proportion of the variation in Eastern Meadowlark home range size (multiple = 0.56). The mean CV of grass cover was higher and twice as variable in 2002 (1.21, 95% Cl 1.00-1.42) compared to 2003 (0.74, 95% Cl 0.62-0.85) and 2004 (0.79, 95% Cl 0.71- 0.86). The relationship between home range size and the CV of grass cover in 2002 varied more by the CV of grass cover, indicating ju- venile Eastern Meadowlarks were selecting a wider range of grass cover than was associ- ated with larger home range sizes (Fig. 4). In contrast, juvenile Eastern Meadowlarks in 2003 and 2004 selected a narrower range of grass cover than was associated with smaller home range sizes. However, limited sample sizes in the first year (2002, n = 9) likely in- fluenced the variability in home range size. DISCUSSION Movement Patterns. — Daily movement pat- terns for Dickcissels were not consistent with other studies, but daily movement patterns for Eastern Meadowlarks were consistent with other studies. Daily movement patterns for ju- venile Dickcissels were consistently smaller than estimates for juvenile Lark Buntings, a similar-sized grassland bird in shortgrass prai- ries (Yackel Adams et al. 2001). Three weeks after fledging, average daily movements by Lark Buntings (238.5 ± 56.1 m) were still ~ 70% greater than those of juvenile Dickcissels (172.6 ± 21.8 m) despite comparable average body masses (Dickcissels, 15.3 ± 0.4 g; Lark Buntings, 21.4 ± 0.5 g). We would not predict large differences in daily movements between these two similarly sized species with similar ecological niches but drought conditions in the Colorado study or differences in landscape structure may have been responsible. Daily movements by Eastern Meadowlarks in our study were similar to patterns for the same species in highly fragmented grasslands in Il- linois reported by Kershner et al. (2004). Juvenile Meadowlarks at 2 weeks of age, in Suedkamp Wells et al • POST-FLEDGING MOVEMENTS OE GRASSLAND BIRDS 127 our study had movements ranging between 100 and 150 m on average, which is similar to the range of movements between 100 and 120 m on average in Kershner et al.’s (2004) study. Juvenile meadowlarks in both studies were also reported foraging and using private land, particularly agricultural crops. Although fine-scale movement distances for this species were similar between studies in fragmented landscapes in the Midwest, additional infor- mation on this species and other grassland birds that reflects a range of landscape frag- mentation is needed. Patterns of average daily movement be- tween species were somewhat surprising giv- en their differences in body size (Dickcissels, 15.3 ± 0.4 g; Eastern Meadowlarks, 44.9 ± 0.8 g) and territory size (Dickcissel range = 0.16 to 1.10 ha, Zimmerman 1966, Harmeson 1974; Eastern Meadowlark range = 1.2 to 6.1 ha, Lanyon 1995). Greenwood and Harvey (1982) recommended framing natal dispersal distances as territory size to reflect potential biological constraints on dispersal distances. We would have expected juvenile Eastern Meadowlarks to have larger average daily movements at an earlier age and a higher total number of days with large movements com- pared to Dickcissel juveniles. Our results were only consistent with our prediction that larger, juvenile Eastern Meadowlarks would show a greater total number of days with large move- ments (>300 m). Our observations suggest the greater number of days with large move- ments for Eastern Meadowlarks may be as- sociated with land use patterns and associated availability of food resources. For example, while nest searching we frequently observed female Eastern Meadowlarks repeatedly trav- eling >500 m from their nests to forage on private land outside our study areas where sweet clover {Melilotus spp.) was abundant. Later, the same females were frequently ob- served leading their broods to these food sources. Soon after juveniles left our study ar- eas for private lands, we observed the initia- tion of larger daily movements and the for- mation of loose, conspecific flocks that made tracking more difficult. Several authors have reported similar observations for juvenile meadowlarks (Kershner et al. 2004) and ju- venile Wood Thrush (Hylocichla inusteliiui) (Vega Rivera et al. 1998). Flock formation may be an adaptive strategy by juveniles to collectively reduce the risk of predation and share information about potential foraging sites in fragmented agricultural landscapes. Future research describing flock formation would be helpful for understanding selective pressures that influence dispersal of older ju- veniles as they begin navigating novel land- scapes. Home Range Size. — This is the first study, to our knowledge, to describe home range size for post-fledging grassland birds. Differences in average home range size across species (Dickcissels, 48.8 ± 8.9 ha; Eastern Mead- owlarks, 80.9 ± 13.9 ha) reflect differences in body size and reported territory sizes. Previ- ous studies on area sensitivity have suggested 50 ha is an appropriate minimum for many grassland birds (Herkert 1994, Helzer and Je- linksi 1999, Winter and Faaborg 1999), which encompasses the smaller of the two average home range sizes in our study. Both species in our study have been reported to be area sensitive (Herkert 1994, Helzer and Jelinski 1999, Winter and Faaborg 1999), indicating they are less likely to occur and potentially reproduce successfully in small patches com- pared to large patches. Future studies inves- tigating the relationship between minimum area sizes required by adults and post-fledging juveniles would be helpful for incorporating all phases of the avian life history cycle into guidelines for grassland conservation and planning efforts. Our finding that local vegetation and year- specific weather factors influence post-fledg- ing dispersal is consistent with other studies of grassland birds (Kershner et al. 2004, Berkeley 2005, Yackel Adams et al. 2006). Home range sizes of juvenile Dickcissels in our study were larger when average vegeta- tion height was shorter and less variable (as indicated by the CV), emphasizing the impor- tance of taller grass and more variability in structural heterogeneity in grassland environ- ments for this species. Dickcissels are consid- ered a tallgrass prairie species and are fre- quently associated with taller plant heights characteristic of tallgrass prairie communities (Zimmerman 1966, Winter 1999, Suedkamp Wells 2005). We believe the interaction between year and CV of vegetation height on home range size 128 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 FIG. 5. Average precipitation by month of the growing season (Apr— Aug) and year (2002, filled di- amonds; 2003, filled squares; 2004, filled triangles) for study sites in southwestern Missouri. The long-term average over 30 years (asterisks) is depicted for com- parison (Missouri Climate Center 2005). for juvenile Dickcissels was driven by abnor- mal patterns of precipitation early in the breeding season in 2002 (Fig. 5) and drought conditions in 2003 (Missouri Climate Center 2005). The 2002 growing season was unusual because twice the normal precipitation oc- curred in May, which triggered flooding that resulted in standing water and colder than nor- mal conditions. May is a critical time for breeding pairs because females begin arriving in Missouri during the first week of May and initiate their first nesting attempts during the second week of May (KSW, pers. obs.). Dick- cissels are normally assumed to be single brooded and females which lost or abandoned their nests in the storm may have renested in poor condition. This may have influenced sub- sequent brood sizes and their respective con- dition, which may influence their ability to disperse. However, we were unable to detect potential effects of clutch size or nestling body mass on home range size. Potential dif- ferences may have been more pronounced if we were able to compare seasonal reproduc- tive effort between females which were suc- cessful with their first attempt and those that renested. The variability of vegetation height (as in- dicated by the CV) within Dickcissel home ranges in 2003 was nearly twice the 2002 and 2004 averages. This pattern indicates the per- sistent drought that started in 2002 and con- tinued through summer 2003 affected local vegetation structure by altering the types of vegetation present (forb vs. grass) and the re- sulting vegetation height. The absence of large broods (4-5 nestlings) of Dickcissels from our sample in 2003 may reflect the energetic strain on breeding females during prolonged drought periods when food is scarce due to limited availability of preferred substrates for domi- nant insect foods (e.g., forbs) (KSW, unpubl. data). One explanation for the change in veg- etation height is that the early pulse of spring rain in April 2003 may have favored growth of forbs that typically compete with grasses for light and moisture, and provide important substrates for dominant insect foods. If grass- es were physiologically stressed from the drought that started at the end of the 2002 growing season (Missouri Climate Center 2005), a competitive release in 2003 may be a viable explanation. Results from microhab- itat selection by juvenile Dickcissels in com- panion parts of this study indicate the avail- ability of grasses and forbs increased in 2003 compared to 2002 and the forb increase was proportionately larger (Suedkamp Wells 2005). This shift in dominant plant groups supports our explanation for the effects of drought on average vegetation height. Local vegetation and year-specific factors also affected home range size for juvenile Eastern Meadowlarks. Interactions between year and the variability in grass cover (as in- dicated by the CV) were driven by patterns of grass composition that were twice as variable in 2002 versus 2003 and 2004. Results from a companion part of this study indicated the transition from 2002 to 2003 was associated with a two-fold increase in grass cover se- lected by juvenile Eastern Meadowlarks, de- spite similar levels of availability (Suedkamp Wells 2005). Grass cover was an important predictor of the presence of juvenile Eastern Meadowlarks in 2002 but not 2003 (Sued- kamp Wells 2005), indicating grass cover was a more limiting habitat feature in 2002 versus 2003. Grass cover is important for breeding habitat because females typically build nests at the base of grass clumps and use litter to construct a roof and two side walls (KSW, pers. obs.). CONSERVATION IMPLICATIONS Our results provide support for the hypoth- esis that local vegetation factors primarily af- fect post-fledging dispersal of grassland birds. Conservation planners need realistic guide- lines for grassland acquisition and manage- ment that incorporate adult and juvenile life history needs on public and private land if de- clining grassland bird populations are to be Suedkamp Wells et al. • POST-FLEDGING MOVEMENTS OF GRASSLAND BIRDS 129 stabilized. Our results coupled with previous recommendations demonstrate the importance of using minimum area requirements as the threshold for grassland acquisition and man- agement plans that address all life history stages. Grassland managers should be aware that larger areas are required for species with larger body and territory sizes, such as mead- owlarks, Upland Sandpipers (Bartramia lon- gicauda), and Greater Prairie-chickens {Tym- panuchus cupido) that are also present and species of conservation concern (Missouri De- partment of Conservation 2007). Home range sizes may be manipulated by managing for structural heterogeneity, which should reduce the minimum area required for both species in our study. Conservation planners should select and use management practices (prescribed burning, herbicides, grazing, mowing) to ob- tain the minimum required amounts of critical habitat components (vegetation height and grass cover in this case). Minimum area re- quirements may be affected by the relative proportion of public and private land within an individual or population home range. The majority of birds in this study left public lands and spent large periods of time on surrounding private lands, which is managed primarily for row crops and livestock grazing. The land- scape context around public land may be equally important and must be considered when evaluating land acquisition and manage- ment plans for grassland birds to insure the necessary resources exist within appropriate distances for dispersing juveniles. ACKNOWLEDGMENTS We thank Frank Thompson, John Faaborg, Eric Bol- linger, Maiken Winter, and David Sample for com- ments on earlier drafts that improved the manuscript. The Missouri Department of Conservation, and the University of Missouri USGS Cooperative Wildlife Research Unit, Department of Fisheries and Wildlife Sciences, and Conservation Biology Fellowship Pro- gram provided funding. Jason Wells provided assis- tance with data manipulation in Access. Len Gilmore, Kristin Austin, Sharon Gough, Jay Bowmaster, and Josh Cussimanio provided housing and logistical sup- port. Matt Bahm, Kellie Alsup, Lisa Fitzgerald, Micah Zucarelli, John Quinn, Maren Gimple, Shelby Sturgis, Tim Bull, Debbie Morton, Krista Adamek, Jody Bart/. Courtney McCusker, Melissa Hough, Steve Fullington, Audrey DeRose-Wilson, Angie Merritt, and Craig Re- koske provided field assistance. We also thank the landowners in St. Clair and Cedar counties, especially the Bartz, Morton, Seigismund, and Morton families. LITERATURE CITED Altwegg, R., T. H. Ringsby, and B. Saether. 2000. Phenotypic correlates and consequences of dis- persal in a metapopulation of House Sparrows Passer domesticus. Journal of Animal Ecology 69: 762-770. Anders, A. D., J. Faaborg, and F. R. Thompson III. 1998. Postfledging dispersal, habitat use, and home-range size of juvenile Wood Thrushes. Auk 115:349-358. Baker, R. R. 1993. The function of post-fledging ex- ploration: a pilot study of three species of passer- ines ringed in Britain. Ornis Scandinavica 24:71- 79. Barbraud, C., a. R. Johnson, and G. Bertault. 2003. Phenotypic correlates of post-fledging dis- persal in a population of Greater Flamingos: the importance of body condition. Journal of Animal Ecology 72:246-257. Beardah, C. C. and M. J. Baxter. 1995. MATLAB routines for kernel density estimation and the graphical representation of archaeological data. Anelecta Prehistorica Leidensia 28. Leiden Uni- versity, Rapenburg, The Netherlands. Berkeley, L. 2005. The postfledging ecology of Dick- cissels (Spiza americana). Thesis. University of Nebraska, Omaha, USA. Cohen, E. B. and C. A. Lindell. 2004. Survival, hab- itat use, and movements of fledgling White-throat- ed Robins {Turdiis assimilis) in a Costa Rican ag- ricultural landscape. Auk 121:404-414. Daubenmire, R. 1959. A canopy-coverage method of vegetation analysis. Northwest Science 33:43-63. De Solla, S. R., R. Bonduriansky, and R. J. Brooks. 1999. Eliminating autocorrelation reduces biolog- ical relevance of home range estimates. Journal of Animal Ecology 68:221-234. Dingemanse, N. j., C. Both, A. J. van Noordwijk, A. L. Rutten, and P. j. Drent. 2003. Natal dispersal and personalities in Great Tits (Parus major). Pro- ceedings of the Royal Society of London B 270: 741-747. Garton, E. O., M. j. Wisdom, F. A. Leban, and B. K. Johnson. 2001. Experimental design for radiote- lemetry studies. Pages 16 — 14 in Radio tracking and animal populations (J. J. Millspaugh and J. M. Marzluff, Editors). Academic Press, San Die- go, California, USA. Green, D. J. and A. Cockbern. 2001. Post-fledging care, philopatry, and recruitment in Brown Thorn- bills. Journal of Animal Ectilogy 70:505-514. Greenwood, P. J. and P. H. Harvey. 1982. The natal and breeding dispersal of birds. Annual Review of Ecology and Systematics 13:1—21. Harmf-son, j. P. 1974. Breeding ecology of the Dick- cissel. Auk 91:348-359. Hf;i.zf:r, C. J. and D. E. Jf;i. inski. 1999. The relative importance of patch area and perimeter-area ratio 130 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. I, March 2008 to grassland breeding birds. Ecological Applica- tions 9:1448-1458. Herkert, J. R. 1994. The effects of habitat fragmen- tation on midwestern grassland bird communities. Ecological Applications 4:461-471. Hooge, P. N. and B. Eichenlaub. 2000. Animal move- ments extension to ArcView. Version 2.0. USGS, Alaska Science Center, Biological Science Office, Anchorage, USA. Jones, M. C., J. S. Marron, and S. J. Sheather. 1996. A brief survey of bandwidth selection for density estimation. Journal of the American Statistical As- sociation 91:401—407. Kennedy, P. L. and J. M. Ward. 2003. Effects of ex- perimental food supplementation on movements of juvenile Northern Goshawks {Accipiter gentilis atricapillus). Oecologia 134:284-291. Kenward, R. E., S. S. Walls, and K. H. Hodder. 2001. Life path analysis: scaling indicates priming effects of social and habitat factors on dispersal distances. Journal of Animal Ecology 70:1-13. Kernohan, B. j., R. a. Gitzen, and J. J. Millspaugh. 2001. Analysis of animal space use and move- ments. Pages 126-168 in Radio tracking and an- imal populations (J. J. Millspaugh and J. M. Mar- zluff. Editors). Academic Press, San Diego, Cali- fornia, USA. Kershner, E. L., j. W. Walk, and R. E. Warner. 2004. Postfledging movements and survival of ju- venile Eastern Meadowlarks {Sturnella magna) in Illinois. Auk 121:1146—1154. Lanyon, W. E. 1995. Eastern Meadowlark {Sturnella magna). The birds of North America. Number 160. Magrath, R. D. 1991. Nestling weight and juvenile survival in the Blackbird, Turdus merula. Journal of Animal Ecology 60:335-351. Mathworks. 1999. MATLAB Version 5.3. Natick, Massachusetts, USA. Mech, L. D. 1983. Handbook of animal radio-tracking. University of Minnesota Press, Minneapolis, USA. Missouri Climate Center. 2005. Climate bulletin. University of Missouri, Columbia, USA. Missouri Department of Conservation. 2007. Mis- souri species and communities of conservation concern. January 2007 checklist. Department of Conservation, Jefferson City, Missouri, USA. Rappole, j. H. and a. R. Tipton. 1991. New harness design for attachment of radio transmitters to small passerines. Journal of Field Ornithology 62: 335-337. Rich, T. D., C. J. Beardmore, H. Berlanga, P. J. Blancher, M. S. W. Bradstreet, G. S. Butcher, D. W. Demarest, E. H. Dunn, W. C. Hunter, E. E. Inigo-Elias, j. a. Kennedy, A. M. Martell, A. O. Panjabi, D. N. Pashley, K. V. Rosenberg, C. W. Rustay, j. S. Wendt, and T. C. Will. 2004. Partners in Flight North American Landbird Con- servation Plan. Cornell Laboratory of Ornitholo- gy, Ithaca, New York, USA. Sauer, J. R., J. E. Hines, and J. Fallon. 2005. The North American breeding bird survey, results and analysis 1966-2005. Version 2005.2. USGS, Pa- tuxent Wildlife Research Center, Laurel, Mary- land, USA. www.mbr-pwrc.usgs.gov/bbs/bbs. html (accessed 19 February 2007). Seaman, D. E., J. J. Millspaugh, B. J. Kernohan, G. C. Brundige, K. j. Raedeke, and R. A. Gitzen. 1999. Effects of sample size on kernel home range estimates. Journal of Wildlife Management 63: 739-747. SPSS Institute Inc. 1999. SYSTAT. Version 9. SPSS Institute Inc., Chicago, Illinois, USA. Steel, R. G., J. H. Torrie, and D. A. Dickey. 1997. Principles and procedures of statistics. McGraw- Hill, Boston, Massachusetts, USA. SUEDKAMP Wells, K. M. 2005. Resource selection, movement patterns, and survival of post-fledging grassland birds in Missouri. Dissertation. Univer- sity of Missouri, Columbia, USA. SuEDKAMP Wells, K. M., B. E. Washburn, J. J. Millspaugh, M. R. Ryan, and M. W. Hubbard. 2003. Effects of radio-transmitters on fecal glu- cocorticoid levels in captive Dickcissels. Condor 105:805-810. Vega Rivera, J. H., J. H. Rappole, W. J. McShea, AND C. A. Haas. 1998. Wood Thrush postfledging movements and habitat use in northern Virginia. Condor 100:69-78. Wand, M. P. and M. C. Jones. 1995. Kernel smooth- ing. Chapman and Hall, London, United King- dom. Wiens, J. D., B. R. Noon, and R. T. Reynolds. 2006. Post-fledging survival of Northern Goshawks: the importance of prey abundance, weather, and dis- persal. Ecological Applications 16:406-418. Winter, M. 1999. Nesting biology of Dickcissels and Henslow’s Sparrows in southwestern Missouri prairie fragments. Wilson Bulletin 111:515—527. Winter, M. and J. Faaborg. 1999. Patterns of area sensitivity in grassland-nesting birds. Conserva- tion Biology 13:1424-1436. WORTON, B. J. 1987. A review of models of home range for animal movement. Ecological Modelling 38:277-298. WoRTON, B. J. 1989. Kernel methods for estimating the utilization distribution in home-range studies. Ecology 70:164—168. Yackel Adams, A., S. K. Skagen, and R. D. Adams. 2001. Movements and survival of Lark Bunting fledglings. Condor 103:643-647. Yackel Adams, A., S. K. Skagen, and J. A. Savidge. 2006. Modeling post-fledging survival of Lark Buntings in response to ecological and biological factors. Ecology 87:178—188. Zimmerman, J. L. 1966. Polygyny in the Dickcissel. Auk 83:534-546. The Wilson Journal of Ornithology 120(1):131-138, 2008 ECOLOGICAL FACTORS AFFECTING RESPONSE OF DARK-EYED JUNCOS TO PRESCRIBED BURNING JINELLE H. SPERRY,>-2'' T. LUKE GEORGE," AND STEVE ZACK^ ABSTRACT. — We compared abundance, daily survival rate, nest site characteristics, food availability, nest activity, and nestling size of Dark-eyed Juncos (Junco hymenalis) between burned and unburned mechanically- thinned ponderosa pine (Pinus ponderosa) forest units. Dark-eyed Junco territory density, number of detections in point counts, and daily nest survival were similar between treatments. Average bare ground was 4.8 times higher and litter cover was 2.6 times lower at nest sites in burned units compared to unburned nest sites. However, there was 28% less burned area around nests compared to random points in burned units, indicating that juncos placed nests in unburned portions of burned units. They also selected non-traditional nesting sites in burned units such as root holes and in trees. Arthropod abundance was higher in burned units 1-year post burn although numbers were similar in the second-year post burn. Nest attentiveness and feeding rates were three times higher in burned units, possibly in response to increased food availability. The potentially negative effect of prescribed burning through reduction of litter and increase in bare ground was offset by novel nesting strategies and increased food availability. Received 21 November 2006. Accepted 26 May 2007. Traditional management practices, such as logging and fire suppression, have dramati- cally altered the structure and composition of ponderosa pine {Pinus ponderosa) forests in the western United States (Covington and Moore 1994, Veblen et al. 2000). These for- ests were frequently swept by low-intensity fires, prior to European settlement, that main- tained open stands (Covington and Moore 1994, Skinner and Change 1996, Fry and Ste- phens 2006). Decades of fire suppression have resulted in an accumulation of understory fu- els, increased stand density, and encroachment of fire- intolerant species, such as white fir {Abies concolor), which can lead to high in- tensity, stand replacing fires (Covington and Moore 1994). Land managers often use pre- scribed burning in conjunction with silvicul- tural thinning to restore ponderosa pine forests to historical conditions and minimize fire risk (Weatherspoon 1996, Covington et al. 1997). Prescribed burns result in changes in forest ' Department of Wildlife, Humboldt State Univer- sity, 1 Harpst Street, Areata, CA 95521, USA. 2 Current address: Program in Ecology, Evolution and Conservation Biology, University of Illinois Ur- bana-Champaign, 606 E. Healey, Champaign, IL 61820, USA. ^ Current address: Wildlife Conservation Society, 718 SW Alder Street, Suite 210, Portland, OR 97205, USA. '‘Corresponding author; e-mail: jhutch@uiuc.edu Structure, soil properties, plant species com- position, understory vegetation, ground cover, and arthropod biomass (Rogers 1996, DeLuca and Zouhar 2000, Griffis et al. 2001, Waltz et al. 2003). These changes may positively or negatively affect forest nesting birds by alter- ing habitat structure or food availability. The effects of prescribed burning on birds gener- ally vary with ground and shrub-foraging spe- cies typically increasing in abundance follow- ing burning (Bock and Lynch 1970, Finch et al. 1997, Bock and Bock 1983, Saab and Pow- ell 2005). Increased abundance of ground- nesting birds following a burn seems unex- pected given that burning alters the habitat in ways that should be detrimental. Our objective was to examine the response of Dark-eyed Juncos {Junco hyemalis) to pre- scribed burning in a mechanically-thinned for- est. Specifically, we examined if predicted in- crease in abundance following prescribed burning results from habitats becoming eco- logical traps through decreases in ground nest cover or whether behavioral flexibility allows juncos to exploit these habitats successfully. METHODS Study Area. — We conducted the study with- in the Goosenest Adaptive Management Area (GAMA) on the Klamath National Forest in northern California (41°30N, 121° 52 W) at an elevation of 1,500 to 1,700 m. The site is 131 132 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 dominated by ponderosa pine, white fir, in- cense-cedar {Lihocedrus decurrens), and sug- ar pine (Pinus kimhertiana). The U.S. Forest Service and other cooper- ators initiated a long-term study in 1996 at GAMA to examine alternative approaches to accelerating late-succession forest character- istics. Five replicates of three silvicultural treatments and a non-treatment “control” were randomly applied to 20 units within GAMA. Each unit was 40 ha and included a 100-m buffer where the silvicultural treatment extended beyond the unit boundary to reduce possible edge effects within the unit. The pine emphasis treatment involved removal of small-diameter trees while retaining large-di- ameter pines. The pine-emphasis-with-fire treatment involved the same thinning treat- ment as the pine emphasis units followed by a prescribed bum. Another thinning treatment was applied to five of the units but was not monitored in this study (Ritchie and Harcksen 1999). The most intensive aspects of our study were conducted on the five pine emphasis and five pine-emphasis-with-fire units. We refer to these treatments as unburned and burned treat- ments, respectively. We also monitored nests on two additional control (unthinned) units al- though too few juncos were found to include these units in analyses. Thinning treatments started in 1998 and were completed in 2000. The five burned units were prescribed burned in fall 2001. We conducted this study during May-August 2002 and 2003. Bird Abundance and Nest Survival — We estimated junco density in each unit using spot mapping and point counts. A 9-ha area (300 X 300 m) in the center of each treatment unit was marked with wire flags every 50 m to assist with spot mapping. Each unit was vis- ited 7-8 times in both 2002 and 2003, and the location, movements, and behavior of all Dark-eyed Juncos were plotted on maps. At the conclusion of the nesting season, territo- ries were delineated and counted following methods described in Bibby et al. (1992). Par- tial territories were counted if the majority of the territory was within the spot-mapping area. Point counts were conducted during May and June for both years. Nine points, each 200 m apart, were established within the 9-ha grid area on each unit and each point was surveyed twice by different observers in each year. Point count surveys were started within 15 min of local sunrise and continued until all points on a unit were completed. Count du- ration was 8 min and started upon arrival at the point. Point counts were not conducted during steady rain, snow, or strong winds (>20 km/hr). All junco detections <100 m from the point were used in analysis. Nest searches were conducted following Martin and Geupel (1993). Nests were monitored ev- ery i_4 days until their fate was ascertained. Vegetation. — Vegetative characteristics were measured around nest sites and at ran- dom grid points in both burned and unbumed units. Overstory canopy cover was estimated as the average of four densiometer readings taken 5 m from the site in the four cardinal directions and at the site for a total of 20 read- ings per site. Shrub cover and woody debris were estimated along two perpendicular 30.8-m transects centered on the nest or ran- dom site. Only woody debris with a diameter greater than 7.5 cm and shrubs greater than 0.25 m in height were included. We also measured vegetative characteristics in a 1 -m circle centered on the nest or random site. Litter depth was measured 1 m from the nest or random site in the four cardinal direc- tions. Herbaceous, grass, bare ground, log, and rock cover were visually estimated in a 1-m circle around each nest or random loca- tion (Martin et al. 1997). We also estimated the proportion of the area within 1 m of the nest or random site that was burned. Nest con- cealment directly over the nest cup was vi- sually estimated at every nest where the nest remained intact. All nests, including those in which eggs were not observed, were included in habitat measurements. Arthropod Sampling. — Arthropod abun- dance was estimated using pitfall and sticky board traps. Dark-eyed Juncos typically for- age on or near the ground (Holmes and Rob- inson 1988) and our methods targeted ground dwelling or low-flying arthropods. Sticky board traps were made of 30.5 X 30.5 cm cor- rugated plastic. Each sticky board was cov- ered with a thin layer of Tanglefoot® and placed vertically 4 cm above the ground on metal wires. Eollowing collection, sticky boards were covered in plastic wrap to pre- serve the samples. Pitfall traps consisted of Sperry et al. • JUNCOS AND PRESCRIBED BURNING 133 TABLE 1. Mean (± SE) number of Dark-eyed Junco territories on 9-ha spot-mapping plots, point count detections, and nesting success of Dark-eyed Juncos on five burned and five unburned units at Goosenest Adaptive Management Area, Klamath National Forest, California, 2002-2003. Year Variable Territories Detections Nests («) Obs. days Nest success Daily survival (95% Cl) 2002 Burned 5.8 ± 0.6 16.6 ± 3.1 18 207.5 0.29 0.95 (0.92-0.97) Unburned 5.0 ± 0.3 17.8 ± 3.8 14 99.5 0.23 0.95 (0.93-0.95) 2003 Burned 2.8 ± 0.7 8.0 ± 0.6 12 191.5 0.41 0.97 (0.96-0.97) Unburned 2.8 ± 0.5 7.8 ± 1.1 16 189.5 0.34 0.96 (0.95-0.97) 450-ml containers with holes cut in a circle around the top. Containers were placed with the holes flush to the ground. A lid was placed on the trap to ensure that precipitation and non-target species were excluded. Pitfall traps contained nontoxic propylene glycol as a pre- servative. Both trap types were set for two 7-day sam- pling periods. The first sampling period co- incided with Dark-eyed Junco nest building stage (late May 2002, early Jun 2003) and the second coincided with the nestling phase (late Jun 2002, early Jul 2003). Five sticky board and 10 pitfall traps were placed at randomly chosen grid points in each unit. Arthropods were identified to taxonomic order and mea- sured to nearest millimeter. Length- weight ra- tios were calculated using the general equa- tion in Rogers et al. (1976). Only arthropods 3-20 mm in length were used in the analysis as smaller prey items (<3 mm) and extremely large prey have been shown to be under-rep- resented in the diet of many bird species (Quinney and Ankney 1985, Raley and An- derson 1990, Van Home and Bader 1990). Nest Activity. — Digital video cameras were placed at nests for 3 hrs, starting at sunrise, to measure food delivery rates. Nests were vid- eotaped when young were 6 days old and only nests that had four nestlings were used to con- trol for variation in food demands as a func- tion of nestling age and number. We recorded the times when an adult delivered food to the nestlings and the start and end time of each brooding event. Nestling Measurements. — Wing chord, tar- sus, and weight of each nestling in 2003 were measured on the sixth day after hatching. Wing chord and tarsus lengths were recorded with calipers to the nearest 0.5 mm, and weight was measured with an electric balance to the nearest 0. 1 g. Statistical Analysis. — Dark-eyed Junco ter- ritory numbers and point count detections were compared between burned and unbumed units using two-sample r-tests. We used PROC LOGISTIC following Hazier (2004) to ana- lyze daily survival rates (DSR) of nests (SAS Institute 2000). Models were developed using three variables: treatment, mean arthropod biomass by unit, and year. These models were likely to explain variation in daily survival rate of junco nests on our study sites. The can- didate set of models was evaluated using Akaike’s Information Criteria (Burnham and Anderson 2002) corrected for small sample size (AICc). All models were ranked based on relative AIC^ weight (w,). The model with the lowest AAICc and highest w, is considered the best approximating model from the set of can- didate models tested (Burnham and Anderson 2002). Actual DSR values were calculated by back transforming logit-scale regression equa- tions. Only nests in which an egg was ob- served were included in nest survival analysis. Nest/random site characteristics, arthropod biomass, nest activity rates, and nestling mea- surements were compared between treatments with ANOVAs with unit nested within treat- ment. Nestling length and weight measure- ments were averaged for each nest. All AN- OVAs and r-tests were completed using NCSS (Hintze 2001) statistical program. RESULTS Abundance and Nest Survival. — Dark-eyed Junco abundance did not differ significantly between burned and unburned units for either year (Table 1 ). However, both average number of junco territories and point count detections per unit were higher in 2002 than 2003 (/ = 2.23, df = 18, P = 0.04 and / = 3.99, df = 18, P = <0.001, respectively) (Table 1). Thirty-two nests were monitored in 2002 134 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. I, March 2008 TABLE 2. Model selection results from Mayfield logistic regression explaining survival of Dark-eyed Junco nests on burned and unburned units at Goosenest Adaptive Management Area, Klamath National Forest, Cali- fornia, 2002-2003. K = number of parameters in each model including the intercept, AAIC^ = difference between each model and the model with the lowest AIC, score, and AIC, weight = the relative support for each model. Models are ranked based on AIC^ weights. Model K AAlCc AICc (wt) U'/ Rank Constant survival 1 0 244.24 0.601 1 Arthropod biomass 2 2.45 246.68 0.177 2 Year 2 3.55 247.79 0.101 3 Treatment 2 4.02 248.26 0.081 4 Year + Arthropod 3 6.47 250.71 0.024 5 Year + Treatment 3 7.54 251.78 0.014 6 Year * Treatment 4 11.68 255.91 0.002 7 with 18 nests in burned units and 14 in un- four nests videotaped in burned and two in burned units. We monitored 28 nests in 2003 unbumed units. No differences were detected with 12 in burned units and 16 on unbumed. between burned and unburned units for time Daily nest survival was similar between spent at nests or number of provisioning trips/ burned and unbumed units in both 2002 and min. 2003 (Table 1). The constant daily survival model received the lowest AAIQ and the greatest AIC^ weight (Table 2). Nest Activity. — Sample sizes in 2002 were small as few nests survived to day 6 of the nestling period. Four nests in burned and three in unbumed units were videotaped. Females spent an average of 35% of their time on the nest in the burned units compared to 12% in the unbumed {F = 8.15, df = \, P = 0.036). Females spent more time on nests in burned units and we observed more provisioning trips to nests in burned than unburned units (mean ± SE; 0.33 ± 0.06 and 0.11 ± 0.08 trips/min, respectively, F = 4.27, df = 1, P = 0.09). Sample sizes were also small in 2003 with Nestling A/ze.— Mean (±SE) nestling wing chord length on day 6 did not differ between burned and unbumed units (21.9 ± 2.02 and 19.5 ± 1.47 mm, respectively; F = 0.98, df = \ P ^ 0.25) nor did mean nestling mass (10.7 ± 0.52 and 9.6 ± 0.67 g, respectively; F = 1.52, df = 1, P = 0.25). Nest Site.—Limx depth in 2002 was lower and percent bare ground cover was higher at random sites and around Dark-eyed Junco nests (Table 3) in burned compared to un- burned units. Nest concealment did not differ between burned and unburned units (Table 3). There was less burned area around nests than at random locations in burned units (23.8 and 51.7%, respectively; F = 4.09, df = 1, P = TABLE 3. Mean (± SE) values of habitat variables at random sites on burned and unburned units at Goosenest Adaptive Management Area, Klamath National Forest, California, 2002-2003. Year Variable Random Random burned unburned P Nest burned Nest unburned P 2002 2003 Woody debris, % 0.86 ± 0.45 2.58 ± 0.44 Shrub cover, % 1.15 ± 2.33 6.12 ± 2.29 Litter depth, cm 0.52 ± 0.32 2.75 ± 0.31 Bare ground, % 49.27 ± 4.72 24.01 ± 4.62 Nest concealment, % n 25 25 Woody debris, % 1.27 ± 0.85 4.88 ± 0.85 Shrub cover, % 2.82 ± 1.28 5.35 ± 1.28 Litter depth, cm 1.29 ± 0.46 2.51 ± 0.41 Bare ground, % 36.8 ± 4.10 24.00 ±4.10 Nest cover, % n 25 25 0.14 0.82 -H 0.56 2.53 ± 0.56 0.068 0.21 3.01 -H 2.83 8.10 -+- 2.83 0.36 0.001 0.89 -+- 0.40 2.35 0.40 0.038 0.005 35.59 -+- 6.31 7.35 -H 3.27 0.002 80.0 6.23 87.79 ± 3.63 0.33 17 17 0.052 2.29 -H 1.06 6.24 -h 1.03 0.090 0.37 3.32 -H 1.66 6.35 -F 1.51 0.35 0.037 1.96 0.46 3.26 ± 0.41 0.066 0.058 20.62 -t- 4.66 20.00 + 7.61 0.95 86.64 ± 5.44 82.87 -h 4.66 0.61 15 18 Sperry et al. • JUNCOS AND PRESCRIBED BURNING 135 TABLE 4. Mean (± SE) values of arthropod biomass (g) on burned and unburned units at early and late sampling dates at Goosenest Adaptive Management Area, Klamath National Eorest, California, 2002-2003. Arthropod biomass was estimated using general length- weight ratios (Rogers et al. 1976). Year Trap type Season Burned Unbumed 2002 Sticky board Early 40.39 ±2.16 22.27 ± 2.68 Sticky board Late 28.33 ± 2.31 20.97 ± 2.68 Pitfall Early 24.84 ± 3.62 35.80 ± 4.02 Pitfall Late 22.64 ± 3.44 21.14 ± 3.90 2003 Sticky board Early 16.22 ± 1.05 15.54 ± 1.14 Sticky board Late 9.53 ± 1.40 13.04 ± 1.51 Pitfall Early 19.74 ± 2.67 20.66 ± 3.00 Pitfall Late 20.65 ± 3.34 27.02 ± 2.75 0. 048). Percent bare ground in 2003 was high- er, litter depth was lower, and there was less woody debris at random sites in burned com- pared to unbumed units (Table 3). None of the nest site variables differed between burned and unbumed units and none differed between nest and random sites on either treatment in 2003 (Table 3). Two severe wind events during winter 2002-2003 caused extensive windfall on the study units. There were more fallen trees in 2003 around nest sites than at random sites in both the burned (mean ± SE; 3.07 ± 0.65 and 1.52 ± 0.50, respectively; F = 5.51, df = 1, P = 0.026) and unburned units (3.3 ± 0.59 and 1.6 ± 0.50, respectively; F = 3.41, df = 1, P = 0.074). Arthropod Abundance. — Coleopterans and dipterans accounted for most arthropods caught on sticky board traps. Both treatment (burned vs. unbumed) and sampling periods differed in 2002 (Tables 4, 5) while only sam- pling period differed in 2003 (Table 5). Mean biomass was higher in burned units compared to unbumed and higher in May compared to June (Table 4). Most arthropods caught in pit- fall traps in both burned and unbumed units were Hymenoptera (ants) and Diptera. Mean arthropod biomass in pitfall traps did not dif- fer between treatments, sampling period or year (Tables 4, 5). DISCUSSION Several studies have documented increases in abundance of ground-nesting birds follow- ing prescribed burning (Bock and Lynch 1970, Bock and Bock 1983, Saab and Powell 2005). We found Dark-eyed Juncos were equally abundant in burned and unbumed units suggesting that response of ground-nest- ing birds to prescribed fire is variable. Re- sponses may depend on a variety of factors including the nesting and foraging ecology of the bird species, size of the bum, bum sever- ity, and time since burn (Finch et al. 1997). Daily nest survival did not differ between TABLE 5. Effects of treatment (burned and unburned, df = 1), unit (df = 8), sampling date (df = 1) and interactions from nested ANOVA on biomass of arthropods in sticky board and pitfall traps at Goosenest Adaptive Management Area, Klamath National Eorest, California, 2002-2003. Arthropod biomass was estimated using general length-weight ratios (Rogers et al. 1976). Treatment Date Unit Treatment X Date Unit X Date F P FPFPFPFP 2002 Sticky board 13.61 0.006 Pitfall 0.49 0.50 2003 Sticky board 0.36 0.57 Pitfall 2.60 0. 1 5 7.34 0.045 4.75 0.061 5.04 0.055 2.75 0.14 12.73 0.007 2.66 0.14 1.52 0.25 0.85 0.38 1.98 0.045 1.01 0.43 1.88 0.058 0.58 0.80 2.77 0.005 0.81 0.59 1.28 0.25 2.18 0.02( 136 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 treatments (burned vs. unburned), year, or as a function of arthropod biomass. Juncos achieved equal nesting success in burned and unbumed units indicating they were able to overcome the effects of reduced nesting cover. Burned habitats did not function as ecological traps and our results were broadly consistent with the behavioral flexibility hypothesis. Nest concealment is a major factor affecting nest success of passerines and an important aspect of nest-site selection (Martin and Roper 1988, Howlett and Stutchbury 1996, Flaspoh- ler et al. 2000, Weidinger 2002). We expected that nest concealment would be substantially lower on burned than unbumed units but we found no difference between treatments. Sim- ilarity in nest concealment, despite a decline in suitable cover in burned units, can be ex- plained by two nesting strategies used by jun- cos. First, juncos often placed their nests in unbumed patches in burned units as ground cover on the burned units did not bum uni- formly, which is common for both prescribed bums and wild fires in this forest type (Weath- erspoon 1996). Second, juncos used non-tra- ditional nesting sites. Two nests were on mats of pine needle >10 m high in trees and many nests were found under overhanging rocks or in burned root holes. Juncos were not ob- served nesting in these types of sites in un- bumed units or in burned units prior to burn- ing. These results suggest that nest conceal- ment was an important nest site criterion for juncos on our study area. A large wind throw event during winter 2002-2003 provided abundant down wood on all units in 2003, further reducing the differ- ence between burned and unbumed units. Dark-eyed Juncos used this increased ground structure by placing their nests in areas of greater wind throw compared to random sites, regardless of treatment. We documented an overall increase in low- flying arthropod biomass, particularly dipter- ans and coleopterans, in burned units in the spring following the prescribed bum. Biomass of ground arthropods in pitfall traps was sim- ilar between burned and unburned units. These results are broadly consistent with pre- vious studies that demonstrated that wood and bark boring beetles increased following fire to exploit weakened trees (McCullough et al. 1998, Santoro et al. 2000) while ground ar- thropods generally declined due to decreased litter, altered soil properties, and direct fire mortality (McCullough et al. 1998, Wikars and Schimmel 2001, Dress and Boerner 2004). Many arthropods associated with ground or shrub foliage, such as homopterans and Lepidoptera larvae, were not present in large numbers at our study site due to reduc- tion in vegetation caused by thinning treat- ments and prescribed burns. All arthropod Or- ders known to be used as prey by juncos in- creased in burned units 1 year after burning suggesting there may have been more food available. Analysis of junco nest activity demonstrat- ed that adults spent more time on the nest in burned than in unbumed units. Provisioning rates were also greater in burned units, pos- sibly due to higher nest visitation rates by the non-brooding adult. Higher arthropod abun- dance in burned units may have influenced nest attentiveness and provisioning rates. Dark-eyed Juncos demonstrated flexibility in nest placement in areas that were prescribed burned. By using novel nesting strategies, jun- cos were able to find suitable nesting cover. The behavioral plasticity displayed by Dark- eyed Juncos seems likely to be an evolved re- sponse that has allowed them to exploit re- curring changes in habitat. The positive re- sponse to prescribed burning documented in other ground-nesting bird species may be a product of similar behavioral plasticity. ACKNOWLEDGMENTS We are grateful to W. W. Oliver, M. W. Ritchie, and Bill Reynolds of the Pacific Southwest Research Sta- tion, USDA Forest Service in Redding, California and the Goosenest Ranger District for funding and support. We thank Michael Verbit, Jarred Wolfe, Donna Ger- mann, Katisha Jasper, Alex Powell, and Emily Camp- bell for assistance in the field and with insect identi- fication. We also thank M. D. Johnson, M. A. Camann, D. M. Sperry, J. D. Brawn, P. J. Weatherhead, M. J. Willis, and an anonymous reviewer for helpful com- ments and suggestions on earlier versions of this man- uscript. LITERATURE CITED Bibby, C. j., N. D. Burgess, and D. A. Hill. 1992. Bird census techniques. Academic Press, London, United Kingdom. Bock, C. E. and J. H. Bock. 1983. Responses of birds and deer mice to prescribed burning in ponderosa Sperry et al. • JUNCOS AND PRESCRIBED BURNING 137 pine. Journal of Wildlife Management 47:836- 840. Bock, C. E. and J. E Lynch. 1970. Breeding bird pop- ulations of burned and unburned conifer forest in the Sierra Nevada. Condor 72:182-189. Burnham, K. P. and D. R. Anderson. 2002. Model selection and inference: a practical information- theoretic approach. Second Edition. Springer- Ver- lag. New York, USA. Covington, W. W. and M. M. Moore. 1994. South- western ponderosa forest structure: changes since Euro- American settlement. Journal of Forestry 92: 39-47. Covington, W. W., P. Z. Fule, M. M. Moore, S. C. Hart, T. E. Kolb, J. N. Mast, S. S. Sackett, and M. R. Wagner. 1997. Restoring ecosystem health in ponderosa pine forests of the Southwest. Jour- nal of Forestry 95:39-47. DeLuca, T. H. and K. L. Zouhar. 2000. Effects of selection harvest and prescribed fire on the soil nitrogen status of ponderosa pine forests. Forest Ecology and Management 138:263-271. Dress, W. J. and R. E. Boerner. 2004. Patterns of microarthropod abundance in oak-hickory forest ecosystems in relation to prescribed fire and land- scape position. Pedobiologia 48:1-8. Finch, D. M., J. L. Ganey, Y. Wang, R. T. Kimball, AND R. Sallabanks. 1997. Effects and interac- tions of fire, logging, and grazing. Pages 103-136 in Songbird ecology in southwestern ponderosa pine forests: a literature review (W. M. Block and D. M. Finch, Technical Editors). General Techni- cal Report RM-292. USDA, Forest Service, Rocky Mountain Forest and Range Experiment Station, Fort Collins, Colorado, USA. Flaspohler, D. j., S. a. Temple, and R. Roseneield. 2000. The relationship between nest success and concealment in two ground-nesting passerines. Journal of Field Ornithology 71:736-747. Fry, D. L. and S. L. Stephens. 2006. Influence of humans and climate on the fire history of a pon- derosa pine-mixed conifer forest in the southeast- ern Klamath Mountains, California. Forest Ecol- ogy and Management 223:428-438. Griffis, K. L., J. A. Crawford, M. R. Wagner, and W. M. Moir. 2001. Understory response to man- agement treatments in northern Arizona ponderosa pine forests. Forest Ecology and Management 146:239-245. Hazler, K. R. 2004. Mayfield logistic regression: a practical approach for analysis of nest survival. Auk 121:707-716. Hintze, j. 2001. NCSS and PASS. Number Cruncher Statistical Systems, Kaysville, Utah, USA. Holmes, R. T. and S. K. Robinson. 1988. Spatial pat- terns, foraging tactics, and diets of ground-forag- ing birds in a northern hardwoods forest. Wilson Bulletin 100:377-394. Howleti, j. S. and B. j. Stutchbuky. 1996. Nest con- cealment and predation in Hooded Warblers: ex- perimental renn)val of nest cover. Auk 1 13:1-9. Martin, T. E. and G. R. Geupel. 1993. Nest-monitor- ing units: methods for locating nests and moni- toring success. Journal of Field Ornithology 64: 507-519. Martin, T. E. and J. J. Roper. 1988. Nest predation and nest-site selection of a western population of the Hermit Thrush. Condor 90:51-57. Martin, T. E., C. R. Paine, C. J. Conway, W. M. Ho- CHACHKA, P. Allen, and W. Jenkins. 1997. BBIRD Field Protocol. Montana Cooperative Wildlife Research Unit, University of Montana, Missoula, USA. McCullough, D. G., R. A. Werner, and D. Neu- mann. 1998. Fire and arthropods in northern and boreal forest ecosystems of North America. An- nual Review of Entomology 43:107-127. Quinney, T. E. and C. D. Ankney. 1985. Prey size selection by Tree Swallows. Auk 102:245-250. Raley, C. M. and S. H. Anderson. 1990. Availability and use of arthropod food resources by Wilson’s Warblers and Lincoln’s Sparrows in southeastern Wyoming. Condor 92:141-150. Ritchie, M. W. and K. A. Harcksen. 1999. Long-term interdisciplinary research on the Goosenest Adap- tive Management Area, Klamath National Forest, California. Forestry Chronicle 75:453-456. Rogers, L. E., W. T. Hinds, and R. L. Buschbom. 1976. A general weight vs. length relationship for arthropods. Annals of the Entomological Society of America 69:387-389. Rogers, P 1996. Disturbance ecology and forest man- agement: a review of the literature. General Tech- nical Report INT-GTR-336. USDA, Forest Ser- vice, Intermountain Research Station, Ogden, Utah, USA. Saab, V. A. and H. D. Powell. 2005. Fire and avian ecology in North America: process influencing pattern. Studies in Avian Biology 30:1-13. Santoro, A. E., M. J. Lombardero, M. P. Ayers, and J. J. Ruel. 2000. Interactions between fire and bark beetles in an old growth pine forest. Forest Ecology and Management 144:245-254. SAS Institute. 2000. SAS/STAT software: changes and enhancements. Release 8.1. SAS Institute, Cary, North Carolina, USA. Skinner, C. N. and C. R. Change. 1996. Fire regimes, past and present. Sierra Nevada Ecosystem Pro- ject: final report to Congress. Volume 2. Assess- ments and scientific basis for management op- tions. University of California, Davis. USA. Van Horne, B. and A. Bader. 1990. Diet of nestling Winter Wrens in relationship to food availability. Condor 92:4 1 3—420. VliBLEN, T. T, T. KrrZBERGIiR, and j. DoNNIiGAN. 2000. Climatic and human influences on fire regimes in ponderosa pine forests in the Colorado Front Range. Ecological Applications 10:1178-1195. Waltz. A. E., P. Z. Full, W. W. Covingion, and M. M. Moore. 2003. Diversity in ponderosa pine for- est structure following ecological restoration treat- ments. Forest .Science 49:885-900. 138 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 Weatherspoon, C. P. 1996. Fire-silviculture relation- ships in Sierra forests. Sierra Nevada Ecosystem Project; final report to Congress. Volume 2. Chap- ter 44. Centers for Water and Wildland Resources, University of California, Davis, USA. Weidinger, K. 2002. Interactive effects of conceal- ment, parental behaviour and predators on the sur- vival of open passerine nests. Journal of Animal Ecology 71:424—437. WiKARS, L. AND J. SCHIMMEL. 2001. Immediate effects of fire-severity on soil invertebrates in cut and un- cut pine forests. Forest Ecology and Management 141:189-200. The Wilson Journal of Ornithology 120(1):139-145, 2008 WINTER HABITAT USE BY BOREAL CHICKADEE FLOCKS IN A MANAGED FOREST ADAM HADLEY' AND ANDRE DESROCHERS'^ ABSTRACT. — There are increasing conservation concerns associated with boreal regions, but little is known about winter habitat requirements of bird species inhabiting them. We examined flock size, winter habitat pref- erence, and home range size of Boreal Chickadees (Poecile hudsonica) in a boreal forest harvested for timber near Quebec City, Quebec, Canada. We investigated whether and to what extent home range size was affected by clearcuts and regeneration forest stands. Flocks included an average of four individuals and occupied a mean winter home range of 14.7 ha. Flock membership and size were stable during the winter. Boreal Chickadees strongly preferred mature stands of commercial value (>7 m in height) and used regenerating stands (4-7 m in height) to a lesser extent. Younger stands (<4 m in height) and open areas were avoided. Home range size was not associated with landscape composition, but flocks with larger home ranges used them less evenly than those with smaller home ranges. This resident species prefers stands of commercial value and logging may contribute to apparent population declines of Boreal Chickadees. Received 23 September 2006. Accepted 31 March 2007. Logging is considered the most important threat to birds inhabiting boreal forest regions (Imbeau et al. 2001) and is rapidly modifying North America’s boreal forests. Approximate- ly 300,000 ha of boreal forest in Quebec, Can- ada have been clearcut annually in recent years (Ministere des Ressources naturelles et de la Faune 2006). This logging practice is leading to a dramatic regional reduction in large tracts of mature forest and a subsequent increase in younger successional stages. The remaining forest patches are typically restrict- ed to riparian buffer strips and buffers be- tween adjacent clearcuts (Ministere des Res- sources naturelles 1996). Among birds, resident species are hypoth- esized to be most exposed to loss and frag- mentation of boreal forests (Imbeau et al. 2001, Schmiegelow and Monkkonen 2002). Population dynamics of resident species in- habiting northern latitudes appear to be strongly affected by events occurring during the non-breeding season (Matthysen 1990, Lahti et al. 1998, Doherty and Grubb 2002). Many resident species inhabiting boreal for- ests in Europe have undergone marked pop- ulation declines hypothesized to result from forest harvesting (Imbeau et al. 2001). Despite increasing conservation concerns associated with boreal regions, little is known about win- ' Centre d’etude de la foret, Faculte de foresterie et de geomatique, Universite Laval, Quebec, Quebec, GIK 7P4, Canada. ^Corresponding author; e-mail: andre.desrochers@sbflulaval.ca ter habitat requirements of bird species inhab- iting them. The Boreal Chickadee {Poecile hudsonica) is a boreal forest resident which exemplifies the lack of knowledge of wintering birds with- in these regions. The Breeding Bird Survey, despite its weak sampling effort for boreal species, reports an alarming decline of Boreal Chickadees (mean annual change 1966-2004 = -3.59%, P = 0.0035) in eastern North America (Sauer et al. 2005). This species is listed as highly vulnerable to changes induced by modem forestry (Imbeau et al. 2001). The Boreal Chickadee is often considered the North American ecological equivalent of the Grey-headed Chickadee {P. cinctaf, a species that has also undergone dramatic declines due to the effects of forestry practices (Imbeau et al. 2001). Flocks of Boreal Chickadees form as soon as young fledge and persist throughout the non-breeding season (Sep-late Apr) (Ficken et al. 1996). Winter survival is thought to set population limits for Boreal Chickadees (Er- skine 1977, 1992) and concern has been ex- pressed about suitability of the remaining hab- itat in winter (Erskine 1992, Foss 1994, Cyr and Larivee 1995). Increasing the proportion of non-habitat and sub-optimal habitat within a landscape can increase the size ol an ani- mal’s home range (Gjerde and Wegge 1989, Storch 1993, Siffczyk et al. 2003). However, winter home range characteristics and habitat use, despite their obvious implications for management, are virtually unknown for this species (Ficken et al. 1996). 139 140 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 1, March 2008 Our objective was to measure the relative occurrence of Boreal Chickadees in mature and regenerating forest stands in a region managed for timber and recreational use. We tested whether flock size and interspecific composition were stable during winter months. We also tested whether size of Boreal Chickadee home ranges was associated with the area of clearcuts and regenerating stands. We predicted that flocks with large home ranges would occur predominantly in certain parts of their ranges (Carr and Macdonald 1986) and flocks with smaller home ranges would distribute their activities more evenly throughout their home range. METHODS Study Area.—^Q collected data during win- ters 2004-2005 and 2005-2006 at the Foret Montmorency, Quebec, Canada (47 20 N, 71° 10' W). The study area is a 66 km^ boreal forest mosaic managed for timber exploitation and recreational use. Mature coniferous stands (>7 m in height) cover —56% of the study area. Balsam fir {Abies balsamea) and occa- sionally black spruce {Picea mariana) domi- nate mature stands, interspersed with white birch {Betula papyrifera). Younger serai stag- es (4-7 m in height) were characterized by sapling balsam fir, black spruce and, to a less- er extent, mixed regeneration. These cover 24% of the study area. Open areas consisting of clearcuts, lakes, rivers, roads >7 m in width, and sapling stands <4 m in height cov- er 20% of the study area. An extensive road network (2.6 km/km^) crosses the research forest. Use of Forest Stands.— We investigated habitat use from 85 unmarked Boreal Chick- adee flocks. Seventy-two flocks were studied in the first winter (6 Jan-10 Mar 2004) and 13 flocks during the following winter (14 Feb-24 Mar 2005). We combined data from both years since no major differences in weather or food availability were observed. We located chickadee flocks each day using randomly selected points on a systematic 1-km spaced grid covering the study area. Grid points were visited only once during the study and we located flocks by snowshoeing a systematic search pattern. We started from the selected grid point and moved 500 m north, 500 m east, 1 ,000 m south, 500 m west. and 500 m north to return to the point of or- igin. All flock detections were passive (no use of playback or “pishing”). Flocks members were unmarked and we assumed that a new flock was monitored each day, as flock com- position differed in all but three cases of ad- jacent flocks (known to be separate due to si- multaneous observations). We followed flocks on snowshoes and plotted their locations in real time at 1-min intervals using a hand held Trimble® GPS receiver (PDOP < 8). One- minute sampling intervals were used because serial correlation is irrelevant when using the proportion of an animal’s trajectory contained within each habitat type for compositional analysis (Aebischer et al. 1993). Frequent sampling more closely approximates the un- derlying trajectory and provides a more pre- cise estimate of proportional habitat use (Ae- bischer et al. 1993, Barg et al. 2005). No po- sitions were recorded during the first 2.5 min following discovery of a flock. We followed each flock for as long as possible not exceed- ing 3 hrs. Positions were recorded only when we were at the approximate center of the flock; data recording ceased immediately if observer position no longer represented that of the flock. The following period for a flock was also terminated if the observer lost con- tact. We then moved to another grid sampling point before recommencing the search for dif- ferent flocks. The time we followed flocks ranged from 5 to 152 min (mean = 53 min) and path lengths ranged from 90 to 2,125 m (mean = 871 m). The total distance traveled following 85 flocks was 74 km in 75 hrs. We observed no discovery bias since there was no relationship (R^ = 0.001, F = 3.4, P = 0.06) between stand type and time elapsed since ini- tial discovery of a flock. We divided the habitat within the study area into three categories: (1) mature forest (stand >7 m in height), (2) regeneration (4-7 m in height), and (3) open areas (stands <4 m in height or areas devoid of vegetation above snow). Stand serai stages were characterized using existing GIS coverage for the study area (validated in situ, delimited by GPS, and mapped with ArcView 3.3 [ESRI 2002]). We established a measure of available habitat rep- resentative of home range size (Jones 2001) by delineating a 200-m buffer surrounding the movement path for each flock. We generated Hadley and Desrochers * BOREAL CHICKADEE WINTER HABITAT USE 141 a 10-m spaced grid of points within each buff- er. The resulting grid points were assumed to represent unbiased samples of the habitat due to the apparent lack of spatial periodicity in the stand types of the study area. We used compositional analysis (Aebischer et al. 1993) to examine if used habitat (observed loca- tions) differed from available habitat within 200 m (grid points). We replaced missing val- ues in log ratios (available but not used) with 0.001, at least one order of magnitude less than the smallest non-zero value for that hab- itat type, as suggested by Aebischer et al. (1993). We used SAS (SAS Institute Inc. 2004) and BYCOMP.SAS macro (Ott and Hovey 1997) to compute the randomization procedure recommended by Aebischer et al. (1993). Only flocks having all three habitat types represented within the “available habi- tat” were used in the analysis. This constraint reduced the sample size to 79 flocks. Home Range Estimation. — We calculated home range sizes using data collected from color-marked flocks of Boreal Chickadees during winter 2004-2005. Twenty-three mem- bers of seven flocks were captured from 6 De- cember 2004 to 3 February 2005 using mist nets and playbacks of chickadees mobbing a stuffed owl. We marked birds with USGS numbered aluminum bands and unique color combinations. We fitted color bands with ~1 cm long flags of colored electrician’s tape to enhance visibility for up to 10 m without use of binoculars (Desrochers et al. 1988). Each marked flock received seven visits be- tween 7 February and 24 March 2005. We randomized the order and time of visits among marked flocks, and successive visits were sep- arated by 2-7 days. We located marked flocks using short (5 sec) bursts of Boreal Chickadee calls on a portable speaker audible to 70 m and by searching on snowshoes in concentric circles around the initial capture locations. Bursts of playback were short and restricted in frequency to limit effects on flock move- ments. We ceased playbacks if answering calls were heard and proceeded directly to where the birds were located. We identified flocks by band combinations and flock composition was recorded at the beginning of each visit. We followed flocks during each visit using the same procedure as with unmarked Hocks. The time that we followed marked flocks during each visit ranged from 5 to 165 min (mean = 43 min). We recorded successive locations when they were separated by at least 5 -min intervals. Points were considered biologically independent as chickadees could easily move to any point within their home range during this time period (Barg et al. 2005). The num- ber of locations per flock ranged from 48 to 83 (mean ± SE = 69 ± 5 points). We calculated home range size using both the minimum convex polygon (MCP) and 95% kernel methods. The MCP was used to allow comparison with other studies of parids (Harris et al. 1990). However, the MCP allows little insight into internal configuration of used spaces, is highly affected by peripheral loca- tions, and can contain larger areas not used by the organism (Harris et al. 1990, Barg et al. 2005). We addressed these shortcomings by using a fixed kernel density estimator to study the density distribution of observations and to construct each flock’s distribution of use (UD). We used the 95% fixed kernel to define the kernel home range size and the 50% fixed kernel area (containing 50% of the observa- tions) to define the core area for each flock. Fixed kernel density estimations were per- formed using the Animal Movement exten- sion in Arc View 3.3 (Hooge and Eichenlaub 1997). We used least squares cross validation (LSCV) for each flock to calculate the optimal smoothing parameter. Least squares cross val- idation provides the least biased estimates for smoothing parameters (Worton 1995, Seaman and Powell 1996). Correlates of Home Range Size. — We com- piled the proportion of different stand serai stages within MCP and kernel home ranges, and examined their association with home range sizes. We also calculated a variable (USE) representing the distribution of loca- tions within the home range. USE was ob- tained by calculating the difference between the 95% and 50% (CORE) kernel estimations (Siffczyk et al. 2003). The USE variable pro- vided information on whether flock locations were concentrated in distinct areas of the home range (high USE values) or evenly dis- tributed throughout the home range (low USE values). Flocks with home ranges containing patchy, widely distributed, resources would be expected to have large USE values (concen- trates use in select locations) and larger home 142 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 TABLE 1. Relationships among home range size, landscape composition, flock size, and use patterns within home ranges of seven marked Boreal Chickadee flocks. Correlation coefficients are shown with /^-values in parentheses (Spearman rank correlation). Home range size = 100% minimum convex polygon (ha); USE - difference between the 95% and 50% kernels (ha). Habitat proportions represent proportions within MCP. Variable Home range size Proportion of mature stands in home range Proportion of regeneration stands in home range Proportion of open area in home range Flock size USE -0.57 (0.2) 0.14 (0.8) 0.39 (0.4) 0.26 (0.6) 0.89 (0.007) range sizes. Flocks with evenly distributed re- sources would be expected to have lower USE values (more even use of home range) and smaller home ranges. The association between landscape com- ponents and home range size was examined using Spearman rank correlations (SAS Insti- tute Inc. 2004). Effects were considered sig- nificant at a = 0.05. RESULTS Use of Forest Stands.— Boreal Chickadee flocks did not use mature forest, regenerating forest, and open areas at random (Wilks’ Lambda = 0.22, F = 134, P < 0.001, n = 79). Mature forest was used more frequently than regenerating forest (mean log-ratio = 7.16, P < 0.001) or open areas (mean log- ratio = 16.44, P < 0.001), and regenerating forest was used more frequently than open ar- eas (mean log-ratio = 3.60, P < 0.001). Flock and Home Range Size.— We used data from all 85 flocks of Boreal Chickadees and found that mean (±SE) flock size was 4 ± 0.2 with a range of three to eight Boreal Chickadees. Sixteen of 85 flocks contained at least one Black-capped Chickadee (Poecile atricapillus) (4 ± 0.4, range = 1—10 within these 16 flocks) and Red-breasted Nuthatches (Sitta canadensis) were present in 27 of 85 flocks (4 ± 0.3, range = 2-8 within these 27 flocks). Nine flocks contained all three spe- cies. Black-capped Chickadees appeared to form cohesive flocks with Boreal Chickadees and remained in close contact with Boreal Chickadees throughout the entire following period. Red-breasted Nuthatches were associ- ated with Boreal Chickadees only as loose for- aging groups (they followed at a distance, of- ten leaving and rejoining flocks). We used seven flocks of marked individuals to learn that flock membership and size remained sta- ble throughout winter months (Jan-Mar). We observed no apparent immigration, emigra- tion, or mortality over a 6-week period in the seven marked flocks. No banded individuals disappeared or moved between flocks and the total number of individuals within flocks re- mained unchanged. Boreal Chickadee flocks occupied a mean (± SE) home range size of 14.7 ± 3.2 ha (MCP, n = 1) with a range of 7.9 to 30.4 ha. Mean home range size using the 95% fixed kernel was actually larger, 16.9 ± 3.4 ha (95% kernel, n = 1) with a range of 7.6 to 33.9 ha. Flocks had a mean (± SE) core area of 2.3 ± 0.7 ha (50% kernel, n = 1) with a range of 1.0 to 5.8 ha. Landscape Structure and Home Range Size.— Boreal Chickadee MCP home ranges contained 61 ± 9% mature forest (range = 21-94%), 23 ± 10% regeneration forest (range = 0-79%), and 15 ± 6% open area (range = 0—43%). Vv^e found no relationship between landscape components and MCP home range size (Table 1). However, we be- lieve the trend of increasing MCP home range size with increasing proportion of open area (Fig. lA) and decreasing home range size with increasing proportion of mature forest (Fig. IB) has biological significance. Flocks with larger home ranges concentrated use in several distinct areas (large differences be- tween 95% and 50% core areas) while flocks with smaller home ranges used their home range more evenly (Table 1). Home range size was not associated with flock size (Table 1). DISCUSSION Boreal Chickadee flocks are similar to win- tering flocks of other northern parid species in size and member stability. Mean flock size (4 Hadley and Desrochers • BOREAL CHICKADEE WINTER HABITAT USE 143 0.4 0.6 Proportion of mature stands FIG. 1. Association between MCP home range size and (A) proportion of open area within home ranges, and (B) proportion of mature stands within home ranges of Boreal Chickadees, Quebec, Canada, 2004-2005. Trend lines are not significant. ± 0.2 individuals) was slightly larger than Grey-headed Chickadee winter flocks, which average between two and three individuals (Virkkala 1990). Willow tits (Poecile mon- tana) form, on average, four-bird flocks (Ek- man 1989, wSiffczyk et al. 2003), while Black- capped Chickadees usually form slightly larg- er flocks with six to eight individuals (Smith 1991). Our marked hocks showed member stability throughout the winter, as is thought to be the case with most parids (Desrochers and Hannon 1989, Ekman 1989). Boreal 144 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 Chickadees formed mixed species flocks sim- ilar to many wintering parids (Hogstad 1987, Smith 1991), but most often maintained flocks solely composed of conspecifics. Boreal Chickadee flocks preferred mature forest stands and avoided young successional stages or open areas (Fig. 1, Table 1). Mature forest stands were strongly preferred to regen- eration stands, but flocks frequently spent time in regenerating forest habitat. Boreal Chicka- dee flocks rarely spent time in open areas (when lone trees were available or when crossing gaps). These results are consistent with information on breeding habitat use by Boreal Chickadees. Whitaker and Montevec- chi (1997) considered the Boreal Chickadee to be a forest generalist within coniferous wood- ed areas. Erskine (1977) also showed Boreal Chickadees used both mature and young for- est during the breeding season. Our flocks had winter home range sizes of 14.7 ha, comparable to those of Black-capped Chickadees (9.5-14.6 ha [Smith 1991]; 22.4 ha [Desrochers and Fortin 2000]) and Willow Tits (12.6 ha [Siffczyk et al. 2003]) during the non-breeding season. Sizes of winter home ranges were not significantly associated with forest composition. However, our flocks had similar tendencies of increasing home range size with inclusion of a larger proportion of non-habitat to those demonstrated previously (Gjerde and Wegge 1989, Storch 1993, Siff- czyk et al. 2003). Differences in space use patterns depending on the size of winter home range likely reflect the patchiness of resources within large home ranges. Flocks with large home ranges fo- cused their activity in distinct locations within their home range while flocks with smaller home ranges distributed their activities more evenly across space. These results agree with the resource dispersion hypothesis (RDH), which predicts that home ranges will be large when patches of resources are widely spaced (Carr and Macdonald 1986). Similar results have been shown for Willow Tits (Siffczyk et al. 2003). Boreal forests in eastern North America are being subjected to intensive forest exploitation (Ministere des Ressources naturelles et de la Faune 2006). Extensive logging has resulted in a reduction in the proportion of mature for- est with a subsequent increase in proportion of young forest or open areas (Imbeau et al. 1999, Ministere des Ressources naturelles et de la Faune 2006). Consequently, forestry practices will result in substantial reduction of optimal Boreal Chickadee wintering habitat, at least over several decades. Our findings are consistent with the hypothesis that apparent population declines in this species result from loss of high-quality wintering habitat. ACKNOWLEDGMENTS We thank David Duchesne, Marc Lamarre, and Ge- nevieve d’ Anjou for help in the field. We are grateful to the staff at Foret Montmorency (Universite Laval) for logistical support. M. G. Betts, Louis Imbeau, S. G. Gumming, and M. S. Hadley provided useful com- ments on the manuscript. This study was funded by a NSERC Discovery grant to Andre Desrochers and a NSERC Postgraduate scholarship (PGS A) to A. S. Hadley. LITERATURE CITED Aebischer, N. J., P. a. Robertson, and R. E. Ken- ward. 1993. Compositional analysis of habitat use from animal radio-tracking data. Ecology 74: 1313-1325. Barg, J. J., J. Jones, and R. J. Robertson. 2005. De- scribing breeding territories of migratory passer- ines: suggestions for sampling, choice of estima- tor, and delineation of core areas. Journal of An- imal Ecology 74:139-149. Carr, G. M. and D. W. Macdonald. 1986. The so- ciality of solitary foragers - a model based on re- source dispersion. Animal Behaviour 34:1540- 1549. Cyr, a. and j. Larivee. 1995. Atlas saisonnier des oiseaux du Quebec. Presses de 1’ Universite de Sherbrooke et Societe de loisir ornithologique de I’Estrie, Sherbrooke, Quebec, Canada. Desrochers, A. and M.-J. Fortin. 2000. Understand- ing avian responses to forest boundaries: a case study with chickadee winter flocks. Oikos 91 .376— 384. Desrochers, A. and S. J. Hannon. 1989. Site-related dominance and spacing among winter flocks of Black-capped Chickadees. Condor 91:317-323. Desrochers, A., S. J. Hannon, and K. E. Nordin. 1988. Winter survival and territory acquisition in a northern population of Black-capped Chicka- dees. Auk 105:727-736. Doherty Jr., P. E and T. C. Grubb Jr. 2002. Survi- vorship of permanent-resident birds in a frag- mented forested landscape. Ecology 83:844-857. Ekman, j. 1989. Ecology of non-breeding social sys- tems of Parus. Wilson Bulletin 101:263-288. Erskine, A. J. 1977. Birds in boreal Canada: com- munications, densities, and adaptations. Canadian Wildlife Service Report Series 41. Canadian Wild- life Service, Ottawa, Ontario, Canada. Hadley and Desrochers • BOREAL CHICKADEE WINTER HABITAT USE 145 Erskine, a. J. 1992. Atlas of breeding birds of the Maritime Provinces. Nimbus Publications and Nova Scotia Museum, Halifax, Nova Scotia, Can- ada. ESRI. 2002. ArcView® 3.3 GIS. Environmental Sys- tems Research Institute Inc., Redlands, California, USA. Eicken, M. S., M. a. McLaren, and J. P. Hailman. 1996. Boreal Chickadee. The birds of North America. Number 254. Foss, C. R. (Editor). 1994. Atlas of breeding birds in New Hampshire. Audubon Society, Dover, New Hampshire, USA. Gjerde, I. AND P. Wegge. 1989. Spacing pattern, hab- itat use and survival of Capercaillie in a frag- mented winter habitat. Ornis Scandinavica 20: 219-225. Harris, S., W. J. Cresswell, P. G. Forde, W. J. Tre- WHELLA, 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 Re- view 20:97-123. Hogstad, O. 1987. Social rank in winter flocks of Wil- low Tits Parus montanus. Ibis 129:1-9. Hooge, P. N. and B. Eichenlaub. 1997. Animal move- ment extension to ArcView. Version 2.0. USGS, Alaska Science Center, Anchorage, USA. Imbeau, L., M. Monkkonen, and A. Desrochers. 200 1 . Long-term effects of forestry on birds of the eastern Canadian boreal forests: a comparison with Fennoscandia. Conservation Biology 15: 1151-1162. Imbeau, L., J.-P. L. Savard, and R. Gagnon. 1999. Comparing bird assemblages in successional black spruce stands originating from fire and logging. Canadian Journal of Zoology 77:1850-1860. Jones, J. 2001. Habitat selection studies in avian ecol- ogy: a critical review. Auk 118:557-562. Lahti, K., M. Orell, S. Rytkonen, and K. Koivula. 1998. Time and food dependence in Willow Tit winter survival. Ecology 79:2904-2916. Matthysen, E. 1990. Nonbreeding social organization in Parus. Current Ornithology 7:209-249. Ministere des Ressources naturelles. 1996. Regle- ment sur les normes d’ intervention dans les forets du domaine public. Gazette Officielle du Quebec 128:2750-2786. Ministere des Ressources naturelles et de la Fau- NE. 2006. Ressources et industries forestieres : portrait statistique. Gouvernement du Quebec, Quebec City, Canada. Ott, P. and F. Hovey. 1997. BYCOMPSAS. Clearing- house for ecology software. Available at: http:// nhsbig.inhs.uiuc.edu/wes/habitat.html. SAS Institute Inc. 2004. SAS software. Version 9.1. SAS Institute Inc., Cary, North Carolina, USA. Sauer, J. R., J. E. Hines, and J. Fallon. 2005. The North American breeding bird survey, results and analysis 1966-2004. USGS, Patuxent Wildlife Re- search Center, Laurel, Maryland, USA. SCHMIEGELOW, F. K. A. AND M. MONKKONEN. 2002. Habitat loss and fragmentation in dynamic land- scapes: avian perspectives from the boreal forest. Ecological Applications 12:375-389. Seaman, D. E. and R. A. Powell. 1996. An evaluation of the accuracy of kernel density estimators for home range analysis. Ecology 77:2075-2085. SiFECZYK, C., L. Brotons, K. Kangas, and M. Orell. 2003. Home range size of Willow Tits: a response to winter habitat loss. Oecologia 136:635-642. Smith, S. M. 1991. The Black-capped Chickadee: be- havioral ecology and natural history. Cornell Uni- versity Press, Ithaca, New York, USA. Storch, I. 1993. Patterns and strategies of winter hab- itat selection in alpine Capercaillie. Ecography 16: 351-359. ViRKKALA, R. 1990. Flock size of the Siberian Tit Pa- rus cinctus during the non-breeding season. Ornis Fennica 67:98-99. Whitaker, D. M. and W. A. Monte vecchi. 1997. Breeding bird assemblages associated with ripar- ian, interior forest, and nonriparian edge habitats in a balsam fir ecosystem. Canadian Journal of Forest Research 27:1 159-1 167. WoRTON, B. J. 1995. Using Monte-Carlo simulation to evaluate kernel-based home range estimators. Journal of Wildlife Management 59:794-800. The Wilson Journal of Ornithology 120( 1 ): 146-152, 2008 LONG-TERM EFFECTS OF WASTEWATER IRRIGATION ON HABITAT AND A BIRD COMMUNITY IN CENTRAL PENNSYLVANIA ADAM T. ROHNKEi ^ ^ND RICHARD H. YAHNER^^ ABSTRACT-We studied the possible long-term (1987 vs. 2003-2004) effects of wastewater irrigation of forests to habitat and a songbird community in central Pennsylvania. The study site, Toftrees, is a wastewater- irrigated deciduous forest interspersed with agriculture grasslands and crop fields that receives ^ 60 cm of wastewater annually. Habitat variables were quantified within 15 random 0.04-ha circular random plots in 1987. We conducted bird surveys twice weekly during 2003 and mist-netted birds for 1,250 hrs durtg 2004 Surveys and mist-netting efforts paralleled those conducted in 1987. A major decline in density o^understory trees and short shrubs was noted between 1987 and 2003; m contrast, cover of herbaceous vegetation and density of tall shrubs dramatically increased. Ground-shrub foraging bird species ^nd two canopy foragers. Red-eyed Vireo {Vireo olivaceus) and Eastern Wood-Pewee {Contopus virens), declined frorn 1987 to 2003-2004. Conversely, Gray Catbirds (Dumetella carolinensis) increased dramatically over this time period. Received 18 November 2005. Accepted 27 April 2007. Long-term trends in a bird community may be affected by many environmental factors, including habitat alteration, pollutants, inva- sive species, and exotic pathogens (Gill 1995, Rohnke 2005). Habitat alteration can have dramatic effects on food supply, nesting sites, cover, and other habitat features for native birds (Johnston 1970); bird density and diver- sity are related to vegetative structure of for- ests (Mac Arthur and Mac Arthur 1961, John- ston 1970, Dickson and Noble 1978). Sewage treatment and wastewater disposal impact to watersheds and associated habitats has been a major concern of municipalities across the United States (Kelso and Bowersox 2004). A growing number of municipalities (22 in 1985) have initiated land application of wastewater since the Federal Water Pollution Act of 1970 as an alternative to discharging treated wastewater directly into rivers and lakes (Rollfinke 1988). Several studies of waste water irrigated forests have documented effects including changes in structure and spe- cies composition of the overstory, understory. ’ School of Forest Resources, 1 1 1 Ferguson Build- ing, The Pennsylvania State University, University Park, PA 16802, USA. 2 School of Forest Resources, 1 19 Forestry Resourc- es Building, The Pennsylvania State University, Uni- versity Park, PA 16802, USA. 2 Current address; Central Mississippi Research and Extension Center, 1320 Seven Springs Road, Ray- mond, MS 39154, USA. Corresponding author; e-mail: rhy@psu.edu mid-canopy, and the forest floor (Cole et al. 1986, Kelso and Bowersox 2004). These changes are attributed to modification of soil moisture chemistry, and tree and shrub loss as a result of significant ice buildup during below freezing periods in winter (Cole et al. 1986, Rollfinke 1988, Kelso and Bowersox 2004). The Pennsylvania State University began ir- rigating wastewater experimentally onto for- ests, fields, and farmland at the Toftrees site in 1963 and expanded to the current facility in 1983 (Kelso and Bowersox 2004). Five years post expansion, Rollfinke (1988) docu- mented limited change in species composition and density of the overstory, understory, and herbaceous vegetation in the irrigated forest. Rollfinke (1988) concluded that small modi- fications to the winter irrigation schedule should alleviate much of the stress placed upon the forests. For example, limiting irri- gation to days with temperature above freez- ing or irrigating crop fields and open areas that are exposed to sun instead of shaded for- ested areas would reduce breakage of under- story tree limbs from ice. A comprehensive vegetation study by Kelso and Bowersox (2004) noted moderate change in species com- position of the overstory trees and substantial change in the species composition of sapling and seedling layers in the irrigated forest. In addition, herbaceous plants and non-native shrubs dominated the forest floor and under- story of the irrigated community. Kelso and Bowersox (2004) concluded the current man- 146 Rohnke and Yahner • WASTEWATER IRRIGATION EFEECTS 147 agement regime would not be able to sustain a forest community because of reduced regen- eration and selective pressure towards non-na- tive shrub species in the irrigated forest. Pre-expansion studies by Savidge and Da- vis (1971) and Greenwald (1981) documented limited changes to the local bird communities at Toftrees prior to facility expansion. Rollfin- ke and Yahner (1990) reported an increase in species richness in 1988 (5 years post-expan- sion), but diversity declined because the bird community became dominated by several for- est edge species (i.e.. Common Yellowthroat [Geothlypis trichas]. Indigo Bunting [Passer- ina cyanea], and Song Sparrow [Melospiza melodia]). Rollfinke and Yahner (1990) spe- cifically noted that wastewater irrigation may be detrimental to Wood Thrush (Hylocichla mustelina) and other ground-nesting and ground-dwelling bird species. They concluded wastewater irrigation had limited effect gen- erally on the bird community at Toftrees. However, Rollfinke and Yahner (1990) cau- tioned that effects of wastewater irrigation would continue to affect the structure of the forest (i.e., ice damage, limited regeneration of saplings, and dense herbaceous plant cover) and encouraged long-term studies to docu- ment gradual effects of wastewater irrigation on the forest bird communities at Toftrees. The objective of our study was to examine long-term (1987 vs. 2003-2004) effects of wastewater irrigation on habitats and a song- bird community at a wastewater irrigation site in central Pennsylvania. METHODS Study Area. — The Toftrees area is a 209-ha wastewater irrigated site in Centre County, Pennsylvania (40° 50' 15" N, 77° 53' 50" W). Overstory trees (>1.5 m tall, >7.5 cm dbh) are primarily oak (Quercus spp.) and red ma- ple {Acer ruhrum), interspersed with aspen (Popiilus spp.) and pine (Pinus spp.) (Rollfin- ke 1988, Rohnke 2005). Major understory trees (>1.5 m tall, 2. 5-7. 5 cm dbh) and shrubs (0.5-1.00 m tall, <2.5 cm diameter) include red maple, white ash {Fraxinus cimericana), tartarian honeysuckle {Lonicera tatarica), rose {Rosa spp.) (Rhoads and Klein 1993), dog- wood {Cornus spp.), and wild grape {Vitis spp.). Additionally, old fields and limited ac- tive agriculture fields are present. Agriculture areas were not included in either of the sur- veys. The wastewater irrigation system com- prised 3,100 rotating sprinkler heads connect- ed by a network of surface and buried lateral pipelines (Rollfinke and Yahner 1990). Ap- proximately 260 cm of wastewater are applied annually (Rollfinke and Yahner 1990, Kelso and Bowersox 2004) with wastewater more than doubling the average annual rainfall of 103 cm (National Climatic Data Center, The Pennsylvania State University). Several areas (each 0.01 ha or smaller) of standing water and permanent ponds have formed since in- stallation of the spray system in April 1983 (Rollfinke 1988). Habitat Sampling. — We quantified the same habitat variables within 15 random 0.04-ha circular sampling plots during summer 2003 that Rollfinke (1988) quantified in 20 plots in 1987. Habitat variables quantified within plots included densities (number/ha) of overstory live trees and snags (stems >1.5 m tall, >7.5 cm dbh), stumps (<1.5 m tall, >7.5 cm di- ameter), and logs (>1 m length, >7.5 cm di- ameter) (modified from James and Shugart 1970). Habitat variables measured within two perpendicular transects (1 X 10 m) inside the sampling plots included densities (number/ha) of understory trees (woody stems >1.5 m tall, 2. 5-7. 5 cm dbh), tall shrubs (woody stems >1.5 m tall, <2.5 cm diameter), and short shrubs (stems >1.5 m tall, <2.5 cm diameter); only stems rooted within the area of the per- pendicular transects were counted. Multi- stemmed species were counted as individual stems if a stem was unbranched and singularly emerged from the soil. Percent canopy cover, herbaceous ground cover, and leaf litter were recorded at 20 ocular tube readings at 2-m in- tervals along the perpendicular transects, giv- ing 10 readings/transect. Average vegetation height (cm) was measured along each perpen- dicular transect and once in the center of the 0.04-ha plot in field habitats only. We calculated means (± SE) of habitat var- iables from all plots combined. Means for each variable were compared between 1987 (Rollfinke 1988) and 2003 by percent change over this time period. Transect Surveys of Birds. — We established four transects, each 200 to 400 m in length, (at least 104 m apart) with a combined length 148 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. I, March 2008 of 1 ,000 m per site. A proportional sample of available habitats (i.e., field [30%], early suc- cessional [35%], and mature forests [35%]) totaling 10 ha (1,000 X 100 m) was surveyed. We counted all birds heard or seen within a 50-m lateral distance of transects while walk- ing '-1.6 km/hr (Conner and Dickson 1980). Birds flying over or through the forest canopy were not recorded. We did not conduct sur- veys during heavy precipitation or when winds exceeded 20 km/hr. Surveys were con- ducted twice weekly, June— August 2003 {n — 18), at each site from sunrise to 1100 hrs EDT to replicate sampling effort by Rollfinke (1988). Mean species richness (± SE), total number of contacts (number/ 10 ha) for all species combined, and number of contacts for indi- vidual species (number/ 10 ha) were calculated for 1987 (Rollfinke 1988) and at Toftrees in 2003. Raw data from 1987 at Toftrees were not available for statistical comparison and we calculated percent change in means between 1987 and 2003. Mist-netting of Birds. — We mist-netted birds from June through October 2004. Mist nets {n = 6-8) were placed between 10 and 50 m apart along the same transects used for bird surveys. We sampled birds at mist nets along each transect for 2 consecutive days; mist nets were generally in operation from 0.5 hr before sunrise to 1400 hrs each sampling day. Nets were closed because of precipita- tion, high wind velocity (>20 km/hr), or heat from direct sunlight to minimize levels of avi- an stress and possible mortality (Rohnke 2005). We opened nets during -250 hrs per transect, giving 1,250 total net hrs/year, as in 1987 (Rollfinke 1988). Nets were rotated and moved no more than 10 m from the center of the original net lanes (2 X 12 m plot cleared of all above-ground vegetation for placement of mist nets) during the season to increase capture success and reduce net avoidance (Rollfinke 1988). We fitted all captured birds with a single numbered federal aluminum band for identification purposes (Pyle et al. 1997) and recorded age, gender, mass (g), and wing and tarsus length (mm) (Rollfinke 1988). Birds previously captured during the season were recorded as recaptures and measured again. We used breeding ‘safe’ dates to ascertain whether captured individuals were breeding residents or migrants in central Pennsylvania. Safe dates calculated by the Pennsylvania Breeding Bird Atlas are conservative (1 Jun- 31 Jul; Robert Mulvihll and Mike Lanzone, pers. comm.). We used a more liberal range of dates because we were interested in not only recording breeding resident birds but also hatch-year and nonbreeding birds during sum- mer. Our ‘adjusted safe dates’ were 15 May-6 August and were based on Pennsylvania Breeding Bird Atlas safe dates and banding records at both sites. All birds captured within the adjusted safe dates were considered resi- dent and those captured outside of the adjust- ed safe dates (not previously banded) were migrants. Chi-square goodness-of-fit tests (So- kal and Rohlf 1995) were used to compare total individuals captured of each species be- tween 1987 and 2004. RESULTS Habitat Sampling.— OvQxsiory and under- story tree density decreased by 69 and 49%, respectively, in 2003 versus 1987 (Table 1). Density of tall shrubs increased (447%) in 2003 compared to 1987, whereas densities of short shrubs decreased by 80% in 2003 versus 1987. Densities of logs decreased 58% and densities of stumps decreased 96% in 2003 compared with 1987. Canopy cover decreased (28%) but herbaceous ground cover and leaf litter increased (21 and 92%, respectively) from 1987 compared with 2003. Transect Sampling of Birds.— Mean total contacts of all species combined and mean species richness increased by 89 and 16%, re- spectively, from 1987 to 2003 (Table 2). We noted 43 species in 2003 compared with 34 in 1987. Major increases (change >50%) in number of contacts of 15 species were noted in 2003 compared with 1987. Of these 15 spe- cies, we noted dramatic increases (change >1,000%) in three species including Ameri- can Robin {Turdus migratorius). Gray Catbird {Dumetella carolinensis), and Song Sparrow. Major decreases (change >-50%) were ob- served in number of contacts of nine species, including Hairy Woodpecker (Picoides villo- sus), Pileated Woodpecker {Dryocopus pilea- tus). Eastern Wood-Pewee {Contopus virens). Great Crested Elycatcher {Myiarchus crini- Rohnke and Yahner • WASTEWATER IRRIGATION EFFECTS 149 TABLE 1. Habitat variables (means ± SE) measured at 0.04-ha circular plots in 1987 {n = (« = 15) at Toftrees, Centre County, Pennsylvania. Data from 1987 are from Rollfinke (1998). 20) and 2003 Variable 1987 Year 2003 % change Overstory tree, all species, (number/ha) 708 ± 42 217 ± 49 -69 Understory tree, all species, (number/ha) 669 ± 101 336 ± 157 -49 Tall shrub, all species, (number/ha) 1,447 ± 304 7,909 ± 2,126 +447 Short shrub, all species, (number/ha) 6,645 ± 902 1,330 ± 4,725 -80 Physical features, (number/ha) Logs 305 ± 34 127 ± 30 -58 Stumps 68 ± 17 3 ± 2.3 -96 Percent cover Canopy 72 ± 3.8 52 ± 7.7 -28 Herbaceous ground cover 57 ± 3.8 69 ± 6.7 +21 Leaf litter 14 ± 3.0 27 ± 7.0 + 92 TABLE 2. Mean number of contacts (± SE) per species (number/ 10 ha), mean total contacts (all species combined), mean species richness, and total species observed along 1,000 m of transects (10 ha) at Toftrees, Centre County, Pennsylvania, during 1987 (Rollfinke 1988) and 2003. Only those 29 species observed in both years and only species with changes > ±50% are shown. Percent change was calculated between years. Toftrees Species 1987 2003 % change Mourning Dove“ {Zenaida macroura) 0.06 -H 0.06 0.56 H- 0.18 826 Red-bellied Woodpecker"* (Melanerpes carolinus) 0.06 -H 0.06 0.22 -+- 0.13 270 Hairy Woodpecker^ 1.00 -H 0.24 0.33 -+- 0.16 -67 Pileated Woodpecker^ 0.28 0.14 0.11 -+- 0.08 -61 Eastern Wood-Pewee^ 1.06 0.22 0.11 -+- 0.08 -90 Great Crested Flycatcher^ 0.89 -H 0.29 0.39 -+- 0.18 -56 Red-Eyed Vireo^ 8.00 ± 0.59 2.22 -+- 0.39 -72 Blue Jay"* (Cyanocitta cristata) 0.61 -H 0.33 1.94 -+- 0.36 219 American Crow"* 0.22 -H 0.10 1.83 -+- 0.51 733 Wood Thrush^ 3.28 -H 0.64 1.56 ± 0.28 -53 American Robin'* 0.44 0.18 8.1 1 -+- 1.43 1 ,743 Gray Catbird** 0.1 1 -t- 0.08 12.33 1.30 1 1,112 Cedar Waxwing"* 0.28 0.16 2.89 ± 0.54 932 American Redstart"* 1.39 0.23 4.61 -h 0.77 232 Ovenbird^ 0.33 0.12 0.06 ± 0.06 -83 Common Yellowthroat"* 4.50 0.35 1 0.50 ± 1 .00 133 Northern Cardinal"* 1.22 ± 0.3 1 3.00 ± 0.50 146 Rose- Breasted Grosbeak"* 0.17 0.09 1.78 ± 0.44 946 Indigo Bunting^ 3.89 ± 0.28 1.89 ± 0.45 -51 Eastern Towhee^ 3.89 ± 0.43 1.1 1 ± 0.24 -71 Song Sparrow** 0.17 + 0.09 12.89 ± 0.99 7,482 Brown-Headed Cowbird"* (Molothrus ater) 1.17 ± 0.35 2.50 -1- 0.71 1 14 Common Crackle"* {Qiiiscaliis quiscula) 0.44 ± 0.17 0.94 -t- 0.46 I 15 American Goldfinch"* 0.33 ± 0.14 2.44 + 0.50 641 Mean total contacts"* 44.60 ± 2.12 84.17 + 4.28 89 Mean species richness 17.50 0.57 20.33 0.65 16 “ Percent change > 50. Percent change > -50. Percent change > 1 .(XK). 150 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 TABLE 3 Total individuals mist-netted (recaptures not included) per species at Toftrees, Centre County, Pennsylvania during 1987 (Rollfinke 1988) and 2004. Only the 23 species with >10 contacts and captured in toth years are shown. Data were based on ~ 1,250 hrs of mist net operation during each year. Goodness-of-flt tests were performed on common (observed ^20 captures) species. Species Downy Woodpecker {Picoides pubescens) Black-Capped Chickadee (Poecile attricapillus) American Robin Gray Catbird"* American Redstart Ovenbird Common Yellowthroat Eastern Towhee Song Sparrow"* Northern Cardinal Indigo Bunting"* American Goldfinch"* Total individuals"* 1987 9 9 5 60 12 7 23 11 116 13 34 21 320 2004 1 5 10 115 8 3 15 7 54 9 7 3 237 > 10.83, df = \,P < 0.001. tus). Wood Thrush, Red-eyed Vireo {Vireo olivQC€us\ Ovenbird (^Sciurus ciuroccipillus^. Eastern Towhee {Pipilo erythrophthalmus), and Indigo Bunting. Mist-netting of Birds. — We mist-netted 280 individual birds representing 31 species in 2004 (Table 3). The six most common species included American Robin, Gray Catbird, American Redstart (Setophaga ruticilla). Common Yellowthroat, Song Sparrow, and Northern Cardinal {Cardinalis cardinalis). Rollfinke (1988) mist-netted 384 individuals representing 36 species at Toftrees in 1987 with the five most common being Gray Cat- bird, Common Yellowthroat, Song Sparrow, Indigo Bunting, and American Goldfinch (Carduelis tristis). Only the Gray Catbird increased (P < 0.001) from 1987 to 2004 (Table 3). We doc- umented declines (P < 0.001) for Indigo Bun- ting, Song Sparrow, and American Goldfinch. Song Sparrows were the second most abun- dant bird species in 2004. Total number of in- dividuals (P < 0.001) mist-netted declined in 2004 versus 1987. DISCUSSION Habitat Variables.— wastewater irrigation affected the vegetation at Toftrees as predicted by Greenwald (1981) and docu- mented by Rollfinke et al. (1990). Overstory trees were affected by winter ice damage and routine selective logging to remove safety hazards (Cole et al. 1986, Kelso and Bower- sox 2004). Tall and short shrubs densities also were impacted by ice buildup from year-round irrigation on the site (Cole et al. 1986, Roll- finke et al. 1990). The primary causes for the increase in herbaceous vegetation height and density were likely increases in soil moisture and nutrient concentrations (e.g., nitrogen) (Mastrota et al. 1989). Changes in species and density at the Toftrees site were likely the re- sult of secondary succession. Our observation of declining densities of overstory and understory trees at Toftrees since 1987 may be problematic because of minimal regeneration in the understory of tree saplings established after 1983, which was the year of installation of the wastewater irriga- tion system (Kelso and Bowersox 2004). The overstory canopy at Toftrees may be reduced more over time if this lack of regeneration continues, which will further alter vegetation structure and corresponding bird communities. Tall shrubs increased between 1987 and 2003 at Toftrees; consistent with findings of Kelso and Bowersox (2004), but their tall shrub den- sity was not as high as in our study. The Kelso and Bowersox (2004) study used similar methods and their sampling included several of our study plots but also several plots in different areas of the Toftrees site. We found that most tall shrubs were tartar- Rohnke and Yahner • WASTEWATER IRRIGATION EFEECTS 151 ian honeysuckle and rose, which are invasive, exotic shrub species that thrive in disturbed ecosystems (Rhoads and Klein 1993). Seed bank and sapling studies by Kelso and Bow- ersox (2004) documented that wastewater ir- rigation had strong negative effects on pres- ence of native tree and shrub species. They concluded a variety of factors could be lim- iting the establishment of native shrub and tree seedlings including overstory and under- story competition, browse pressure, and seed sources being mostly comprised of upland species. We noted that percent cover of leaf litter nearly doubled at Toftrees between 1987 and 2003 although Rollfinke (1988) concluded that high soil moisture accelerated decompo- sition of leaf litter. The increase of tall shrub density was likely attributed partially to an in- crease in foliage, resulting in greater amounts of leaf litter. Avian Community. — The Indigo Bunting, compared with 1987 (Rollfinke 1988), dra- matically decreased in 2003-2004; this spe- cies has declined slightly on a regional scale (Sauer et al. 2005). We have no biological ex- planation for the large decline in numbers of this species, except perhaps perching and singing sites have been reduced as a result of fewer overstory and understory trees. Total contacts combined and mean species richness increased in 2003-2004; consistent with findings by Rollfinke and Yahner (1990). Major declines in understory vegetation pre- sumably impacted some ground-shrub and mid-canopy foragers, as indicated by a decline in species richness and total individuals dur- ing mist-netting in 2004. Rollfinke and Yahner (1990) also reported population declines in Wood Thrush and Ovenbird since the onset of wastewater irrigation in 1983. We detected a decline in Eastern Towhee in 2004, which par- allels a steady decline of this species over the past two decades in the northeastern United States (Hagan 1993, Sauer et al. 2005). The Song Sparrow, another declining species on our study site, decreased on a regional scale but is considered common in Pennsylvania (McWilliams and Brauning 2000, Sauer et al. 2005). Greenwald (1981) predicted that canopy species, including Eastern Wood-Pewee and Red-eyed Vireo, should be found in similar abundance in both pre- and post-irrigation eras at Toftrees. Our study documented a de- crease in populations of both species based on transect surveys in 2003 compared to 1987. Decreases of overstory trees on the study area may partially account for the lower density of these species on a local scale. Regionally, the Eastern Wood-Pewee has steadily declined since 1966, whereas the Red-eyed Vireo has increased steadily since 1966 (Sauer et al. 2005). The “park-like understory,” as described by Kelso and Bowersox (2004) in wooded stands at Toftrees provides habitat for several edge species to expand into the forest interior (Rollfinke and Yahner 1990). For example, American Crow (Corvus brachyrhynchos), American Robin, American Redstart, Com- mon Yellowthroat, and American Goldfinch increased in the forest interior between 1987 and 2004. Frugivores (e.g.. Cedar Waxwing [Bomby cilia cedrorum]. Gray Catbird, and Rose-Breasted Grosbeak [Pheucticus ludovi- cianus}) may have benefited from the dramat- ic increase in tartarian honeysuckle on the site; these species are primary dispersal agents of its seeds (Nickell 1965). Long-term trends in the local bird com- munity indicated that local habitat alteration in response to wastewater irrigation have af- fected some species. Kelso and Bowersox (2004) concluded that natural processes can- not sustain a forest community under the cur- rent management plan at Toftrees. The avian community will lose forest dependent bird species (canopy and mid-canopy) over time as a result of the change in vegetation structure and will be mostly comprised of secondary successional bird species (i.e.. Gray Catbird, Song Sparrow, Common Yellowthroat, etc.). We predict continuing declines in ground nest- ing species, including Wood Thrush and Ov- enbird. These changes will result in a decline in mean bird species richness over time. ACKNOWLEGMENTS We thank K. C. Kim and G. .1. San Julian, and ref- erees for comments on earlier drafts. We thank M. C. Brittingham, R. R. Charnick, D. M. Cramer, J. T. De- marco, M. R. Marshall, M. E McDermott, P. M. McElhon, and B. D. Ross for logistical and field as- sistance. E. H. Hill assisted with preparation of the manii.script. This study was funded by the Pennsylva- nia Agricultural PAperiment Station and the Office of 152 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 Physical Plant and Wastewater Operations at The Pennsylvania State University. LITERATURE CITED Cole, D. W„ C. H. Henry, and W. L. Nutter (Edi- tors). 1986. The forest alternative for treatment and utilization of municipal and industrial wastes. University of Washington Press, Seattle, USA. Conner, R. N. and J. G. Dickson. 1980. Strip transect sampling and analysis for avian habitat studies. Wildlife Society Bulletin 8:4-10. Dickson, J. G. and R. E. Noble. 1978. Vertical dis- tribution of birds in a Louisiana bottomland hard- wood forest. Wilson Bulletin 90:19-30. Gill, F. B. 1995. Ornithology. Second Edition. W. H. Freeman and Company, New York, USA. Greenwald, C. M. 1981. Prediction of songbird re- sponses to habitat alteration resulting from waste- water irrigation. Thesis. Pennsylvania State Uni- versity, University Park, USA. Hagan, J. M. 1993. Decline of the Rufous-sided To- whee in the eastern United States. Auk 1 10:863- 874. James, E and H. H. Shugart Jr. 1970. A quantitative method of habitat description. Audubon Field Notes 24:727-736. Johnston, D. W. 1970. High density of birds breeding in a modified deciduous forest. Wilson Bulletin 79-82. Kelso, L. M. and T W. Bowersox. 2004. Long-term effects of wastewater irrigation on forested eco- systems in central Pennsylvania. Journal of the Pennsylvania Academy of Science 78:43—52. MacArthur, R. H. and J. W. MacArthur. 1961. On species diversity. Ecology 42:494-498. Mastrota, F. N., R. H. Yahner, and G. L. Storm. 1989. Small mammal communities in a mixed-oak forest irrigated with wastewater. American Mid- land Naturalist 122:388-393. McWilliams, G. M. and D. W. Brauning. 2000. The birds of Pennsylvania. Cornell University Press, Ithaca, New York, USA. Nickell, W. P. 1965. Habitats, territory, and nesting of the Catbird. American Midland Naturalist 73:433- 478. Pyle, P, S. N. G. Howell, D. F. DeSante, R. P. Yu- NiCK, AND M. Gustafson. 1997. Identification guide to North American birds. Volume 1. Slate Creek Press, Bolinas, California, USA. Rhoads, A. F. and W. M. Klein. 1993. The vascular flora of Pennsylvania: annnotated check-list and atlas. Proceedings of the American Philosophical Society, Philadelphia, Pennsylvania, USA. Rohnke, a. T. 2005. Possible ecological effects of hab- itat alteration and West Nile virus on songbird populations in central Pennsylvania. Thesis. Penn- sylvania State University, University Park, USA. Rollfinke, B. F. 1988. Avian communities in waste- water-irrigated deciduous forests in central Penn- sylvania. Thesis. Pennsylvania State University, University Park, USA. Rollfinke, B. F. and R. H. Yahner. 1990. Community structure and composition of breeding and winter- ing birds in a wastewater-irrigated oak forest. Journal of Wildlife Management 54:493-500. Rollfinke, B. F, R. H. Yahner, and J. S. Wakeley. 1990. Effects of forest irrigation on long-term population trends in breeding-bird communities. Wilson Bulletin 102:264—278. Sauer, J. R., J. E. Hines, and J. Fallon. 2005. The North American Breeding Bird Survey, results and analysis 1996—2004. Version 2005.2. USGS, Patuxent Wildlife Research Center, Laurel, Mary- land, USA. www.mbr-pwrc.gov/bbsd3bs.html (ac- cessed 28 May 2005). Savidge, I. R. AND D. E. Davis. 1971. Bird populations in an irrigated woodlot 1963-1967. Bird Banding 42:249-263. SOKAL, R. R. AND F J. Rohlf. 1995. Biometry. Third Edition. W. H. Freeman and Company, New York, USA. The Wilson Journal of Ornithology 120(1): 153- 158, 2008 LONG-TERM TRENDS IN BREEDING BIRDS IN AN OLD-GROWTH ADIRONDACK FOREST AND THE SURROUNDING REGION STACY A. MCNULTY, SAM DROEGE,^ AND RAYMOND D. MASTERS' ABSTRACT. — Breeding bird populations were sampled between 1954 and 1963, and 1990 and 2000 in an old-growth forest, the Natural Area of Huntington Wildlife Forest (HWF), in the Adirondack Mountains of New York. Trends were compared with data from regional North American Breeding Bird Surveys (BBS) and from a forest plot at Hubbard Brook Experimental Forest, New Hampshire. Trends for 22 species in the HWF Natural Area were negative, eight were positive, and one was zero; 20 were significant. Fifteen of 17 long-distance migrants declined, whereas 7 of 14 short-distance migrants and permanent residents declined. Most (74%) HWF Natural Area species, despite differences in sampling periods and local habitat features, matched in sign of trend when compared to Adirondack BBS routes, 61% matched northeastern BBS routes, and 71% matched eastern United States BBS routes, while 66% matched Hubbard Brook species. The agreement in population trends suggests that forest interior birds, especially long-distance migrants, are affected more by regional than local factors. The analysis indicated that bird trends generated from BBS routes may not be as biased toward roads as previously suggested. Received 14 February 2007. Accepted 22 June 2007. Avian trend data at both local and regional scales are important for setting conservation pri- orities, and documenting patterns of decline, in- crease, and stability (Robbins et al. 1989, Faa- borg 2002). Long-term monitoring of land-bird populations is uncommon because sustaining monitoring programs longer than a few years is difficult. As an example, forest birds in North America were not widely surveyed until —40 years ago, and data from only a handful of sites surveyed prior to the 1960s (or sites comparable to them) have continued to be available (cf. Holmes and Sherry 2001). Studies of longer du- ration are of immeasurable value in document- ing change and potential conservation needs for breeding bird species, especially because vari- ances in estimates tend to be high (Thompson and Schwalbach 1995). The primary source of long-term information on trends in populations of breeding landbirds in North America is the Breeding Bird Survey (BBS) (Peterjohn 1994). This survey, initiated in 1966, is both temporally and spatially exten- sive, and is based on permanent survey points along randomly selected secondary roads (blocked by latitude and longitude) throughout the United States and southern Canada. Publi- cation of trends resulting from BBS data led to ‘ Adirondack Ecological Center, SUNY Environ- mental Science and Fore.stry, 6312 Route 28N, New- comb, NY 12852, USA. ^ uses, Patuxent Wildlife Research Center, BARC- EAST, Building 308, Room 124, 10300 Baltimore Av- enue, Beltsville, MD 20705, USA. ^Corresponding author; e-mail: smcnulty@esf.edu programs in the late 1980s to survey forest in- teriors in selected parts of the continent (Cad- man et al. 1998, Faccio et al. 1998) to corrob- orate or complement large scale BBS trends from roadside counts. The BBS is affected by a number of known biases that influence estimates of population trends (Link and Sauer 1998). One area of po- tential bias in BBS data is placement of sur- vey points along roadways. The results may be biased toward trends that reflect human- disturbed habitat which may be disproportion- ately represented along roads (O’Connor 1989). Hagan et al. (1997) found that esti- mated population trends of landbirds in Maine did not correlate well with BBS results, likely because local habitat (industrial forest) dif- fered from that around survey routes. The objectives of this paper are to: ( 1 ) com- pare bird population trends from data collect- ed in 1954-1963 and 1990-2000 in an old growth tract in the central Adirondacks to trends based on BBS data from the region and to trends in a forest plot at Hubbard Brook Experimental Forest, New Hampshire, and (2) evaluate whether local trends match those at range-wide scales, a finding that would indi- cate the need to shift bird conservation pri- orities toward regional rather than local scales. METHODS Study Area. — Our sampling points were in the Natural Area on Huntington Wildlife Forest (HWF), a 6,(K)0-ha research station in the cen- tral Adirondack Mountains. HWF is operated by 153 154 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 the Adirondack Ecological Center of the State University of New York, College of Environ- mental Science and Forestry (44° 0' 26.3" N, 74° 16' 12.4" W, 488-579 m elevation) in the towns of Newcomb, Essex County, and Long Lake, Hamilton County, New York, USA. The HWF Natural Area is 365 ha of predominantly northern hardwood forest dominated by Amer- ican beech (Fagus grandifolia), yellow birch (Betula alleghaniensis), and sugar maple {Acer saccharum), interspersed with eastern hemlock (Tsuga canadensis), red spruce {Picea rubens), and balsam fir {Abies balsamea). The age of many trees in the HWF Natural Area is 200- 300 years: tree age was measured by increment boring. The shrub layer consists mainly of witch hobble {Viburnum lantanoides) and beech, with ground cover consisting of ferns (mostly Dryop- teris spinulosa) and forbs. The study portion of the HWF Natural Area is >2.75 km from the nearest road and is most likely primary or old- growth forest (Leopold et al. 1988; Adirondack Ecological Center, unpubl. data). No catastrophic events drastically altered the habitat during the years of the surveys. However, old-growth forests are subject to processes that alter the vegetative composition over time. For example, a wind storm in 1950 created several canopy openings in the stand that were largely closed by 1954, the first year of the original study (Webb et al. 1977). Other long term changes in vegetation structure and floristics have been documented. Timber in- ventories of the study area in 1948 and 1993 indicated that number of trees >9.14 cm di- ameter at breast height (dbh) decreased from 74.2 to 65.4/ha (Adirondack Ecological Cen- ter, unpubl. data). The number of sugar maple stems in this same size class declined from 20.2 to 10.5% of the stand, red spruce de- clined from 18.7 to 3%, and beech increased from 39.7 to 68.8%. Overall, conifers declined from 28.2 to 5.2% with most declines in the 9.14-29.46 cm diameter class. Basal area of all trees decreased from 32.62 to 26.56 mYha with the greatest changes occurring in trees >29.46 cm dbh where basal area declined from 23.35 to 18.07 mVha. The understory component of the stand (trees 1.5—9.14 cm, dbh) increased in stem density from 122.2 to 166.3/ha. This increase was driven by mortal- ity of beech >30 cm dbh due primarily to beech bark disease which resulted in extensive root sprouting (Jones and Raynal 1987). Sampling Methodology. — Six bird survey points spaced 0.27 km apart in a 1 X 6 grid were established in 1954 in the HWF Natural Area to serve as a control for a study of the effects of logging on songbirds (Webb et al. 1977). Surveys were conducted in June from 1954 to 1963 by a combination of five ob- servers and all points were visited an average 2.4 times per year (range 1-4). RDM sampled these points from 1990 to 2000 in the same manner for an average of 3.6 surveys (range 2-5) each year. Survey observers recorded all birds heard or seen without regard for distance for 10 min at each point. Surveys were conducted on calm, non-rainy mornings from dawn until noon. Data recorded after 0800 hrs EST were not included in this analysis to minimize the impact of decreased species detection in the later morning hours (Webb et al. 1977). Chim- ney Swifts {Chaetura pelagica) were excluded as it was unclear whether this aerial feeder was consistently counted across years. Statistical Analyses. — Trends in counts were calculated for the HWF Natural Area us- ing Poisson regression (Hedeker and Gibbons 1994) with date of survey as a co-variable (the BBS website calculator was not used). Signif- icance level was set at alpha = 0.05. English and scientific bird names follow the most re- cent AOU checklist (AOU 1998). We defined long-distance (or neotropical) migrants as spe- cies that winter largely outside of North America, short-distance migrants as species that winter largely in the southern United States, and permanent residents as species that remain in breeding areas year-round. We obtained estimates of bird population trends for 1966-2000 from the BBS (Sauer et al. 2001) for the following regions: Adiron- dack Mountains {n = 25 routes); northeastern United States (ME, NH, VT, RI, MA, CT, NJ, NY, PA, MD, WV, and VA) {n = 623 routes); and all eastern North American physiographic regions {n = 1,710 routes) (Sauer et al. 2001). Each BBS survey includes 50, 3-min counts taken at points spaced 0.8 km apart, conducted once each spring. Trends were calculated us- ing the estimating equation route regression method (Link and Sauer 1998). Estimates of population trends for north- McNulty et al • COMPARISON OF ADIRONDACK AND REGIONAL BIRD TRENDS 155 eastern forest birds for 1969-1998 were ob- tained from one 10-ha breeding bird census plot established in the Hubbard Brook Exper- imental Forest, New Hampshire (Holmes and Sherry 2001). This census plot is within a large contiguous tract of secondary northern hardwood forest resulting from selective log- ging in the early 1900s. The forest is domi- nated by sugar maple, American beech, and yellow birch with scattered red spruce. Pop- ulation trends were calculated using linear re- gression models for abundance over time. RESULTS Trends at the HWF Natural Area. — We had sufficient data from the HWF Natural Area for trend analysis of 31 species: the trends were negative for 22, positive for eight, and one showed no change (Table 1). Thirteen of 22 negative trends (59%) were significant, whereas seven (88%) positive trends were sig- nificant. Fifteen of 17 long-distance migrants declined, one increased, and one did not change; 12 of 15 negative estimates were sig- nificant. Seven of 12 short-distance migrants were declining (3 significantly) and five were increasing (all significantly). One of two res- ident species. Black-capped Chickadee {Poe- cile atricapillus), declined nonsignificantly while Hairy Woodpecker {Picoides villosus) increased significantly. Only American Red- starts {Setophaga ruticilla) and Winter Wrens {Troglodytes troglodytes) had a significantly declining trend with sampling date. The num- ber of birds not identified to species was tracked across all years and no significant trend was detected. The most precipitously-declining species at the HWF Natural Area were long-distance mi- grants. Chestnut-sided Warbler (Dendroica pensylvanica) declined most markedly, fol- lowed by Eastern Wood-Pewee {Contopus vi- rens), Canada Warbler (Wilsonia canadensis). Northern Parula {Parula americana), and Wood Thrush {Hylocichla mustelina) (Table 1). The top five increasing species were either short-distance migrants or residents and in- cluded Brown Creeper (Certhia americana). Golden-crowned Kinglet (Regulus satrapa). Blue-headed Vireo (Vireo solitarius). Hairy Woodpecker, and Yellow-rumped Warbler {Dendroica coronata). Comparisons of the HWF Natural Area TABLE 1. Breeding bird relative abundance, co- efficient of variation (CV), and trends in the Huntington Wildlife Forest Natural Area, Newcomb, New York 1954-63 and 1990-2000. AOU code Migration status^ X SD CV p YBSA s 0.46 0.77 158 0.010 HAWO p 0.43 0.61 131 0.001 EAPE L 0.93 1.35 106 <0.001 LEFL L 2.91 1.55 46 <0.001 REVI L 10.96 2.71 25 0.931 SOVI S 0.85 1.17 120 <0.001 BLJA S 0.55 0.94 160 0.003 BCCH P 0.85 1.29 153 0.690 RBNU s 0.28 0.69 240 0.102 WBNU s 0.70 0.89 127 0.326 BRCR s 0.81 1.12 104 <0.001 WIWR s 2.75 1.58 56 0.050 GCKI s 2.91 1.55 27 0.001 SWTH L 4.45 2.43 54 0.233 WOTH L 0.81 1.32 137 <0.001 HETH s 0.75 1.11 149 0.542 NPWA L 0.18 0.70 382 0.077 CSWA L 0.40 0.89 187 <0.001 MAWA L 0.22 0.55 245 0.517 BTBW L 5.73 2.38 41 0.060 YRWA S 1.63 1.09 67 0.656 BTGW L 5.34 1.94 36 0.603 BLWA L 4.18 2.58 62 0.450 BWWA L 1.94 0.47 23 0.016 AMRE L 1.48 1.74 100 <0.001 OVEN L 7.51 4.10 38 <0.001 CAWA L 0.45 0.97 194 <0.001 SCTA L 1.66 1.75 96 <0.001 WTSP S 0.40 0.72 166 0.001 DEJU s 1.25 1.37 108 0.072 RBGR L 0.57 0.84 133 <0.001 ** L = long-distance migrant, S = short-distance migrant, P = permanent resident. with Other Data Sets. — Twenty-three of the 31 species (74%) surveyed in both the HWF Nat- ural Area and the BBS had matching regression slope signs (e.g., both trends were increasing or both were decreasing) for the HWF Natural Area and the Adirondack BBS routes, 61% (/? = 19) for the Northeastern BBS routes, and 71% (az = 22) for the Eastern BBS routes (Table 2). Fourteen of the 21 species (66%) that oc- curred both at the HWF Natural Area and Hub- bard Brook matched in sign of slope. Surveys at Hubbard Brook and the BBS both began in the 1960s while HWF started in the mid-1950s with a hiatus in counts from 1963 to 1989; not all species surveyed at HWF were surveyed at Hubbard Brook. 156 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 8 8 S §8 CJ w I ^ S « iJ !Z O t£ (U -C Oh C £ X £ I B .5 ^ £ PQ ii 8 OX) P •S K T3 OJ 1) CQ W hJ CO CQ oa < CQ Match d z Z Z >- >- z >. >H Z Z ^ Z o o in o in o o o o o o o o o n o o o n- d V d V d V d V d V d d V V d V d d V V d V c ,o 00 ^ (N ^ o o 00 00 o o (N O o ri- o m (N O m CM o o 00 o 00 r- o (N o m O m (N o o o o in o o cn in ^ o o ro (N p JJ o 00 d d d d d d d d d 1 d 1 d 1 d 1 d d d d 1 d d 1 d 1 1 1 1 1 1 1 1 1 1 1 1 o o o o d d V V \D (N 00 IT) o o d d 1 I Z^h>h;h>hZ^Z>^Z>^ “ tOOOOOOOOm^O'-OO-a ^opppoooor^_por~^oop odddddddddddddddd V VVVVVVVV V VV o — ' o o o m o o 00 ^ o O ^ O p P d d d d d d V V oooooooo^ ^oorO'O'^'^pppppp I 1 I oc^oo— 'Or--^ OCNOppppp dddddddd V V V V o o o o oooooooooooooooopp _ — = ^ o— ^d^rOOxNOO Tt O (N rn (N O ro I I I I I z>h^h^>h>hzz>^^:^^^z >hZZ^>">^>^Z>^>^>^Z>^>^>^>^>^ 0 0 0 0 0-; o m CN O O O O ro O P d d> d> d> d> d> V o-^oooooo^'^o— ^ooooop 8o88oo^^oo>nOpopcN^^ ^o-^coooropppppppppppp ddddddddddddddoooo V V V V -H o o o ^ o o p p p p p O (N O p P p P d> d> d d> d d) d> V V V V oooooooop ooror-'-^cNcoppp (Nrnd--corod— '(N I I o o o in p p p p p d d nd— icsiro III I I I I O nO00-H(NaN00(NpOnpppp (NrOOO— hOP— 'OlNOOOppppp-ON OOOOOOOOOO — PPPPPPP ddddddddddddddooop I II II I I I I I I Oinro-HcxcNP'-' -HomoO'^'np oooppppp dddddddd I I I I I I r- m o P o r- (N P p o o o p p d d d d d I I I I I coOHi_lhJhJoo'>oQHCOoocncoco pjC/OjJJJCOi— IJi— Iddi— ^ ^ > H d ^ CQ CQ CQ ^ W ^ > < u d ^ 5 ^ H S < O U ^ ^ Q L = long-distance migrant, S = short-distance migrant, P = permanent resident. McNulty et al. • COMPARISON OF ADIRONDACK AND REGIONAL BIRD TRENDS 157 DISCUSSION HWF Natural Area Trends. — A clear pattern of decline among long-distance migrants that breed in the HWF Natural Area emerged when we compared population trends by migration strategy. Only two long-distance migrants, Magnolia Warbler {Dendroica magnolia) and Red-eyed Vireo (Vireo olivaceus) exhibited non-negative trends. Magnolia Warbler had a positive but non-significant trend while Red- eyed Vireo showed no change. In contrast, the ratio of positive to negative trends among the 14 species of short-distance migrants and per- manent residents was approximately 50:50; only three of seven declines were significant, whereas all seven positive trends were signifi- cant. These results support those of others that indicate declines in forest birds over the past several decades have involved primarily long- distance migrants (O’Connor 1989, Holmes and Sherry 2001). HWF Natural Area-Adirondack BBS. — The strength of the association between the two independent surveys, one along road corridors and the other several kilometers from the nearest road, supports the hypotheses that neo- tropical migrants breeding in the Adirondacks are declining and that those declines have their roots in factors that affect entire regions rather than those factors only affecting indi- vidual plots. The pattern of decline for long- distance migrants in the HWF Natural Area is mirrored in the BBS data for the Adirondacks (12 of 15 negative estimates were significant; Table 2). Other studies using BBS trend esti- mates show that declines in neotropical mi- grants are profound in the Adirondacks (Part- ners in Flight 2000). The directions of trends for 13 of the 17 long-distance migrants match when we compare the HWF Natural Area data and BBS data for the Adirondacks. Trends for Swainson’s Thrush (Catharus ustulatus), Ov- enbird (Seiurus aurocapilla). Northern Parula, and Black-throated Green Warbler {Dendroica virens) do not match, but only two of eight trend estimates for these four species are sig- nificant (Ovenbird and Northern Parula, both from the HWF Natural Area). HWF Natural Area— Northeastern and East- ern BBS. — We found a similar pattern when we compared trends from the HWF Natural Area with regional changes in neotropical mi- grants, but with proportionally fewer matches (11 of 17 for both Northeastern BBS and East- ern BBS). We expected the HWF Natural Area trends to agree more closely with the Adirondack BBS routes than with those of the larger BBS study regions, which was the case, but the HWF Natural Area trends agreed more closely with those of the Eastern BBS (11 of 14 species) than with the Northeastern BBS (6 of 14). Six of 14 short-distance migrants or permanent residents in the HWF Natural Area matched sign with BBS surveys while trends for 11 of 14 species in the HWF Natural Area matched the Eastern BBS. HWF Natural Area-Hubbard Brook. — Eleven species of neotropical migrants oc- curred in both the HWF Natural Area and Hubbard Brook; trends for seven species had matching signs, proportionally fewer matches in sign than with BBS routes. Trends were mismatched for Black-throated Blue Warbler {Dendroica caerulescens), Ovenbird, Black- throated Green Warbler, and Red-eyed Vireo. Twelve of the 14 remaining possible trends were significant, and it would appear there is less synchronicity in trends between these two montane, forested sites than between the BBS trends and the HWF Natural Area study site. Holmes and Sherry (2001) demonstrated good correspondence in trends in neotropical mi- grants on their plot with trends on BBS routes within New Hampshire with nine of 1 1 spe- cies matching in sign. CONSERVATION IMPLICATIONS The high correspondence we found in our study between trends from regional BBS routes, and both the HWF Natural Area and Hubbard Brook, indicates that regional rather than local factors may exert the greatest influ- ence on bird populations. Researchers and land managers often conclude that changes in bird populations are driven by local-scale changes in forest age, habitat quality, vegeta- tive composition, dispersion, and location, and these local factors are also responsible for much of the change in regional bird trends. The presence or absence of bird species is in large part based on presence or absence of the core habitat and associated environmental needs for those species. Our research indicates that regional or landscape-scale factors should receive consideration at least equal to that af- 158 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 forded local factors when assessing conser- vation needs for breeding birds, especially long-distance migrants. The mismatches in sign of trends among the species we investigated could be related to the variability of point counts, differences in de- tection of bird species, and/or observer bias among sites rather than indicators of a true difference between the bird populations stud- ied. We cannot ascribe the changes in these bird counts to any given factor. Many, mostly unmeasured, potential factors (e.g., changes in foraging substrates, insect outbreaks, increas- es or decreases in biomass, forest productivity, predator numbers, weather events, changes in vegetative composition and quality) may be driving these changes, and they may be doing so in breeding and wintering areas and/or mi- gration routes. The data we present from an old growth area distant from roads suggests that BBS sur- veys may not be as strongly biased by prox- imity to roads as previously suggested despite potential differences in habitats along roads and those in the HWF Natural Area. The long- term data set on 3 1 species of birds from the HWF Natural Area allowed for the detection of trends otherwise difficult to identify in bird populations whose sizes typically exhibit high annual variation (Collins 2001). ACKNOWLEDGMENTS We thank Deanna K. Dawson, David DeSante, Mer- cedes Foster, Russell Greenberg, Richard T. Holmes, Annie M. Woods, Benjamin Zuckerberg, and an anon- ymous reviewer for their helpful comments on earlier drafts of this manuscript. LITERATURE CITED American Ornithologists’ Union (AOU). 1998. Check-list of North American birds. Seventh Edi- tion. American Ornithologists’ Union. Washing- ton, D.C., USA. Cadman, M. D., H. J. Dewar, and D. A. Welsh. 1998. The Ontario Forest Bird Monitoring Program (1987_1997): goals, methods and species trends observed. Technical Report 325. Canadian Wild- life Service, Ottawa, Ontario, Canada. Collins, S. L. 2001. Long-term research and the dy- namics of bird populations and communities. Auk 118:583-588. Faaborg, J. 2002. Saving migrant birds: developing strategies for the future. The Corrie Herring Hooks Series. Volume 55. University of Texas Press, Austin, USA. Faccio, S. D., C. C. Rimmer, and K. P. McFarland. 1998. Results of the Vermont Forest Bird Moni- toring Program, 1989-1996. Northeastern Natu- ralist 5:293-312. Hagan, J. M., P. S. McKinley, A. L. Meehan, and S. L. Grove. 1997. Diversity and abundance of land- birds in a northeastern industrial forest. Journal of Wildlife Management 61:718-735. Hedeker, D. and R. D. Gibbons. 1994. A random- effects ordinal regression model for multilevel analysis. Biometrics 50:933-944. Holmes, R. T. and T. W. Sherry. 2001. Thirty-year bird population trends in an unfragmented tem- perate deciduous forest: importance of habitat change. Auk 118:589—609. Jones, R. H. and D. J. Raynal. 1987. Root sprouting in American beech: production, survival, and the effect of parent tree vigor. Canadian Journal of Forest Research 17:539-544. Leopold, D. L., C. Reschke, and D. S. Smith. 1988. Old-growth forests of Adirondack Park, New York. Natural Areas Journal 8:166-189. Link, W. A. and J. R. Sauer. 1998. Estimating pop- ulation change from count data: application to the North American Breeding Bird Survey. Ecologi- cal Applications 8:258-268. O’Connor, R. J. 1989. Population variation in relation to migrancy status in some North American birds. Pages 64-74 in Ecology and conservation of neo- tropical migrant landbirds (J. M. Hagan III and D. W. Johnston, Editors). Smithsonian Institution Press, Washington D.C., USA. Partners in Flight. 2000. Partners In Flight Landbird Conservation Plan: Physiographic Area 26: Adi- rondack Mountains. Draft. American Bird Con- servancy, Alexandria, Virginia, USA. Peterjohn, B. G. 1994. The North American Breeding Bird Survey. Birding 26:386-398. Robbins, C. S., J. R. Sauer, R. S. Greenberg, and S. Droege. 1989. Population declines in North American birds that migrate to the Neotropics. Proceedings of the National Academy of Sciences 86:7658-7662. Sauer, J. R., J. E. Hines, and J. Fallon. 2001. The North American Breeding Bird Survey, results and analysis 1966—2001. Version 2002.1. USGS, Patuxent Wildlife Research Center, Laurel, Mary- land, USA. Thompson III, F. R. and M. J. Schwalbach. 1995. Analysis of sample size, counting time, and plot size from an avian point count survey on Hoosier National Forest, Indiana. Pages 45-48 in Moni- toring bird populations by point counts (C. J. Ralph, J. R. Sauer, and S. Droege, Editors). USDA, Forest Service, General Technical Report PSW-GTR-149. Pacific Southwest Research Sta- tion, Albany, California, USA. Webb, W. L., D. F. Behrend, and B. Saisorn. 1977. Effect of logging on songbird populations in a northern hardwood forest. Wildlife Monographs 55. The Wilson Journal of Ornithology 120(1):159— 166, 2008 TIMING AND LOCATION OF MORTALITY OF FLEDGLING, SUBADULT, AND ADULT CALIFORNIA GULLS BRUCE H. PUGESEK' 3 AND KENNETH L. DIEM^ ABSTRACT. — We investigated patterns of mortality during post-breeding migrations of California Gulls {La- rus californicus) nesting near Laramie, Wyoming, USA. We used 151 recoveries and 647 sightings of banded and patagially-marked gulls to compare ratios of mortalities to observations of live birds (1) during four time periods (early and late fall migration, winter, and spring migration), (2) at two locations (Pacific coast and inland), and (3) among three age-classes of gulls (fledglings, 1- and 2-year-olds, and breeding-age adults). Mortality rates were higher in inland areas (35%) than in coastal areas (15%) and were dependent on season within inland areas, but not in coastal areas. Mortality in inland areas during early fall (21%) was comparable with that in coastal areas (13%) but was higher during late fall (68 vs. 13%) and spring migration (46 vs. 17%). Both fledgling (71%) and adult (64%) gulls experienced high mortality rates during late fall migration, possibly because some gulls were too weak to make their way to the Pacific coast and became trapped by poor weather conditions. Adult gulls also experienced high mortality inland during spring migration; few subadults made the costly migration to and from the breeding area. Some adults also skipped breeding and remained in coastal areas during the breeding season. Received 26 February 2004. Accepted 2 April 2007. California Gulls {Lams californicus) were first observed and described in California (Baird et al. 1884, Behle 1958). Except for a San Francisco Bay colony, the species breeds inland and spends much of the year away from the Pacific coast. All other breeding col- onies are inland and occur from the Salton Sea and Mono Lake, east to the Great Salt Lake, Utah Lake, and Colorado, into the Great Plains of Manitoba and the Dakotas, and north to the Northwest Territories (Cooke 1915, Jehl 1987, Winkler 1996, Molina 2000). The Bam- forth Lake, Wyoming population we studied has one of the longer distances to traverse be- tween breeding and wintering areas. This pop- ulation nests —130 km east of the continental divide in southeastern Wyoming, more than 1,500 km from the nearest Pacific coastline. California Gulls at Bamforth Lake follow a triangular migration pattern from breeding ar- eas in Wyoming to wintering areas on the Pa- cific coast (Pugesek et al. 1999), concentrating between San Francisco and British Columbia. During winter, they move south along the coastline as far as Baja California and return ' Northern Rocky Mountain Science Center, U.S. Geological Survey-Biological Resources Division, Forest Science Laboratory, 1648 South 7th Street, Bozeman, MT 59717, USA. ‘ Department of Zoology and Physiology, University of Wyoming, Laramie, WY 82071, USA. ’ Corresponding author; e-mail: Bruce.Pugesek @ usgs.gov to the Bamforth Lake area to breed, following routes primarily across the southwestern Unit- ed States. Previous research on California Gulls nest- ing at Bamforth Lake demonstrated that sur- vival declines with age (Pugesek and Diem 1990, Pugesek et al. 1995) and is related to increasing reproductive effort with age (Pu- gesek 1981, 1983, 1987; Pugesek and Diem 1990). We analyzed observation data of band- ed and wing-marked gulls during migration and wintering periods to further understand costs of migration and its relationship to re- productive effort. We compared ratios of ob- servations of live and dead gulls across sea- sonal time periods corresponding to post- breeding migration, the wintering period, and spring migration to the breeding colony. We categorized ratios by location of observation (inland and coastal), and among three age- classes of gulls (fledglings, 1- and 2-year-old subadults, and breeding-age adults 3 or more years of age). We hypothesized that: (1) mor- tality rates in inland areas are higher than in coastal areas, (2) mortality rates in coastal ar- eas are similar among age groups, (3) breed- ing-age adults experience higher mortality in inland compared with coastal areas during spring and fall migrations, (4) fledglings ex- perience higher mortality in inland compared with coastal areas during their first migration to the coast and do not migrate inland during spring, and (5) subadults remain in coastal ar- eas year-round. 159 160 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 METHODS Data Collection.— Wq studied a population of California Gulls banded in Albany County, near Laramie, Wyoming, USA. We banded 9,693 chicks and 1,650 adults between 1958 and 1994 with federal aluminum and monel steel leg bands. We also placed 7.6-cm cir- cular orange patagial markers on 764 chicks and 679 adults from 1966 to 1982. Each mark- er had a unique symbol and single-digit num- ber painted on both sides. Most gulls were banded and marked at the breeding colony at Bamforth Lake (41° 23' 39.67" N, 105 44 13.56" W). Some (1,203 chicks and 69 adults) were banded at locations within 12 km of this colony. An additional 473 chicks were banded at the Twin Buttes colony (41° 14' 33.73" N, 105° 51' 26.21" W) on the Laramie plains in 1962 and 1964. The Twin Buttes colony was abandoned in 1964 following 2 years of se- vere hail storms; the majority of gulls from this colony moved to the Bamforth Lake col- ony in 1964. We obtained gull locations from reports sent to us by amateur and professional orni- thologists, and from band recovery and return reports from the USGS Bird Banding Labo- ratory. The information obtained included the patagial marker number and symbol, band number, and the day, month, year, and location of the observation. We converted locations to latitude and longitude coordinates where nec- essary using NOAA Sectional Aeronautical Charts (1:500,000 for Canada and the United States, and 1:1,000,000 for locations in Mex- ico). Latitudes and longitudes were entered to the nearest 10 min. Gulls were classified by age by comparing observation dates to our banding records. Henceforth, we use “obser- vation” to refer generically to both reports of live birds and recoveries of bands from dead birds. Observations were obtained from 1958 to 1994. We removed repeated observations at the same location, 96 observations where age or date of observation was unknown, and ob- servations at the Bamforth Lake colony. We further limited the data set to observations during the migration and winter periods (i.e., Aug through Apr). The final data set consisted of 798 observations of 606 individuals. Full information on age group, season, and loca- tion was available for 790 observations. Pa- tagially-marked birds accounted for 630 ob- servations of which 586 were observed alive and 44 were found dead. The remaining 168 observations of banded gulls without wing markers consisted of 61 live birds and 107 found dead. Data Mapping and Analysis. — We parti- tioned data into three age groups: fledgling, subadult, and adult. Age was based on the number of years since gulls fledged until the time of the observation. The fledgling age group included all gulls that were < 1 year of age. Subadults included all gulls that were > 1 but <3 years of age. Adults were >3 years of age and included all gulls of reproductive age. We used ARC/INFO software to create lo- cation maps (Fig. 1). Observation coordinates of latitude and longitude were rounded to the nearest degree. Polygons were constructed around these 1 -degree points and appear rect- angular due to the UTM (Universal Transverse Mercator) projection typically used in map- ping. Fledglings and subadults were combined into one category, while breeding-age adults (>3 years of age) formed a second category. One observation of a dead fledgling gull in April was not plotted because this individual was observed ~560 km north of Vancouver, British Columbia on the Pacific coast at Greenville Channel (53° 10' N, 109° 10' W). We also partitioned data into four phases of the nonbreeding season: early fall migration (Aug and Sep), late fall migration (Oct through Dec), winter (Jan and Feb), and spring migration (Mar and Apr). We desig- nated locations of observations as being either coastal or inland. The coastal designation re- fers to lands west of the major mountain chains bounding the Pacific Ocean (Cascades in Oregon and Washington, Sierra Nevada in California, Mohawks in Arizona, and Sierra Madre Occidental in Mexico). Inland loca- tions were east of these mountain ranges. Given the number of categories, our data were too sparse to apply mark-recapture anal- ysis (Pollock et al. 1990; J. D. Nichols, pers. comm.). Log linear analysis (Jobson 1992) was performed on a matrix of the number of live versus dead bird observations partitioned by age group, season, and location. In log lin- ear analysis, a variety of models composed of main effects and interactions can be tested. Pugesek and Diem • MORTALITY OF CALIFORNIA GULLS 161 A B FIG. 1. California Gulls migrate in a triangular pattern with a high proportion of mortalities observed east of the coastal mountain ranges (dashed line) during August through September (A), October through December (B), January through February (C), and March through April (D). 162 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 TABLE 1. The only log linear model that fit the California Gull data required a main effect of location. a two-way interaction between age and season , and a three-way interaction between age. season, and loca- tion. Source df p Location ^ 29.98 0.001 Age group X season 6 20.73 0.002 Age group X season X location 6 13.55 0.035 Likelihood ratio 9 14.60 0.100 Model selection was based on an acceptable fit to the data as indicated by a nonsignificant {P > 0.05) model likelihood-ratio Chi-square. We began by fitting main effects models and progressed in complexity to models with in- teractions until a parsimonious model was ob- tained. Additional Chi-square and Fisher’s ex- act tests were performed on portions of the matrix using, whenever possible, data that were excluded from the log linear analysis be- cause full information was not available for age group, season, or location. Statistical sig- nificance was set at F < 0.05. RESULTS Seasonal Observations.— OhsQX\dX\ons dur- ing early fall (Fig. lA) included gulls in tran- sit between southeastern Wyoming and the Pacific coast. Large numbers of gulls were al- ready at the coast, including those that did not migrate inland during the breeding season. Most coastal observations were north of San Francisco. Numerous fledglings and adults were also observed in Colorado, south of the breeding island at Bamforth Lake. Locations of observations shifted during late fall toward greater numbers along the coast south of San Francisco, and there were fewer observations inland (Fig. IB). Gulls were observed as far south as Baja California, Mexico. Most inland polygons contained ob- servations of dead gulls. During winter, most observations occurred south of San Francisco along the coast (Fig. 1C), and only a few gulls were observed inland, primarily at extreme southerly locations in Nevada and Mexico. One dead adult was found east of the Sierra Nevada Mountains in California. During spring, gulls were observed along migration routes from southerly coastal regions to Bam- EF LF W S EF LF W S EF LF W S Fledgling Subadult Adult FIG. 2. Fledgling and subadult California Gulls displayed no trend in the percentage of live birds ob- served through time (EF = early fall, LF = late fall, W = winter, S = spring), but a declining trend was observed among adults. Sample sizes are shown above bars. forth Lake (Lig. ID). Large numbers of gulls remained along the southern Pacific coast. Mortality Analysis. — The only log linear model that fit the data included a main effect of location, a two-way interaction between age group and season, and a three-way inter- action between age group, season, and loca- tion (Table 1). All other combinations of main effects and interactions did not fit the data (maximum likelihood ratio Chi-square, P < 0.05). The significant main effect of location resulted from a lower percentage of live birds in inland areas (90 of 139; 65%) compared with coastal areas (557 of 659; 85%); data pooled across time period and age groups (x^ = 29.26, df = 1, P < 0.001). The significant age group X season inter- action was largely a consequence of seasonal variation observed among adults. The per- centage of live observations did not differ across time periods among fledglings (x^ = 3.31, df = 3, P = 0.35), or subadults (x^ = 2.80, df = 3, P = 0.42) but declined among adult gulls with each time period (x^ = 13.30, df = 3, P = 0.004; Pig. 2). The lowest per- centage of observations of live adult gulls oc- curred during spring migration. The significant three-way interaction among season, age group, and location (Table 1) resulted from (1) a consistently high per- centage of live birds in coastal areas across all seasons and age groups, and (2) a lower per- centage of live birds in inland areas compared Pugesek and Diem • MORTALITY OF CALIFORNIA GULLS 163 Aug-Sep Oct-Dec Jan-Feb Mar-Apr FIG. 3. Mortality among California Gulls in inland areas is high during migrations. Sample sizes are shown above bar. Coastal Inland Coastal Inland Fledglings Adults FIG. 4. Mortality among fledgling and adult Cal- ifornia Gulls during late fall is higher in inland areas (solid bars) compared with coastal areas (open bars). Sample sizes are shown above bar. with coastal areas, primarily in late fall and spring due to a low percentage of live fledg- lings during fall migration, and a low per- centage of live adults observed during spring and fall migrations. Analysis of the coastal location with all age groups pooled indicated high percentages of live birds in all seasons (x^ = 4.13, df = 3, P = 0.25; early fall: 87%, n = 153; late fall: 87%, n = 284; winter 80%, n = 125; spring: 83%, n = 93). There were no differences across seasons in percent live birds in the coastal area for fledglings (x^ = 2.07, df = 3, P = 0.56), subadults (x" = 2.48, df = 3, P = 0.48), or adults (x^ = 6.66, df = 3, P == 0.084). The lowest percentage of live adults (75%) occurred in winter. The percentage of live birds observed in in- land areas was lower compared with coastal areas in late fall (x^ — 44.03, df = 1, P < 0.001) and spring (x^ = 8.11, df = 1, P = 0.004). There were 247 live and 37 dead ob- servations (87%) in coastal areas, and 7 live and 15 dead observations (32%) during late fall. During spring, there were 77 live and 16 dead observations (83%) in coastal areas, and 12 live and 10 dead observations (54%) at in- land areas. During early fall, there was a slightly higher percentage of gulls alive in coastal areas (87%, n = 153) compared with inland areas (79%, n = 89), but the difference was not significant (x^ = 2.85, df = 1, P = 0.091). During winter, too few birds were ob- served inland for statistical analysis. There were 100 live and 25 dead bird ob.servations in coastal areas (80%), and 1 live and 1 dead bird observations (50%) at inland areas. Percentages of gulls observed alive at in- land locations differed with season (x^ = 19.48, df = 3, P < 0.001) (Fig. 3). The high- est percentage alive occurred during early fall and the lowest percentage occurred during late fall. Too few subadults were observed inland in all time periods to analyze statistically. The only times sufficient fledglings were present inland was during early and late fall. In both early and late fall, the percentage of live fledg- lings was lower inland (63%, n = 19) than on the coast (83%, n = 57), but the difference was only significant in late fall (early fall: x^ = 3.06, df = 1, P = 0.093; late fall: Fisher’s exact test, two-tailed probability, P < 0.001; Fig. 4). A similar pattern was observed among adult gulls (early fall: coastal: 89%, /? = 63; inland: 86%, n = 63; x^ = 0. 19, df = 1, P = 0.66) (late fall: x" = 18.25, df= 1,P< 0.001; Fig. 4). Too few adults were observed inland in winter for statistical analysis. In spring, the percentage of live adults inland, 5 of 15 (33%), was significantly lower (x^ = 9.75, df = 1, P = 0.002) than in coastal areas, 27 of 34 (79%). We combined data for early and late fall, and found that a significantly lower (x^ = 5.14, df = 1, P = 0.023) percentage of fledg- lings (56%; // = 27) was observed alive than adults (78%; n = 78) (fledglings: 15 of 27; adults 61 of 78). 164 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 DISCUSSION Comparing percentages of observations of live and dead birds is not equivalent to mark- recapture and other survival estimation meth- ods (Pollock et al. 1990). These comparisons do not provide precise survival estimates, but can be used to compare categories to illustrate when and where gulls experience greater or lesser amounts of mortality. Mark-recapture estimates of survival of breeding-age adult California Gulls in the Bamforth Lake popu- lation ranged from 88% among 3-year-old gulls to 76% among the oldest gulls (>20 years of age; Pugesek et al. 1995). The bulk of our population was <12 years of age (BHP, pers. obs.) and the overall survival rate of breeding-age adults is likely in the range of 82-87%. This survival rate is similar to our percent alive values for adult gulls, but we have no theoretical basis for claiming equiv- alency between survival rate estimates and our percent alive estimates. Data on the location of both live and dead gulls, and dates of observation of live gulls are precise because they are based on direct observations by qualified observers. Data on time of observation of dead gulls are less ac- curate because we have no way of knowing how long the gull had been dead before it was located. Thus, tabulations of seasonal catego- ries are biased toward shifting the time of death into a later time category. This inherent bias is probably unimportant because gulls are absent from areas critical for analysis during most periods other than the time to which they were assigned. For example, gulls found dead inland in spring likely died then because they are largely absent from the area in all other time periods. The same is true for gulls ob- served inland during the post-breeding fall migration to the Pacific coast. Some bias may occur, however, as a result of incorrectly as- signing observations to late fall when death occurred in the early fall. The sample revealed no differences in the percentage of live gulls observed across age groups. When the sample was additionally categorized by season, we found that adults incurred relatively high mortality during spring migration. The percentage of live gulls observed in coastal areas remained high during all sea- sons, 80-87% among all age groups com- bined. No significant variation within age groups was detected in percent live gulls in coastal areas. Fledglings of other gull species experience high mortality and subadult mortality is great- er than adult mortality but is more similar to adult than fledgling mortality (Coulson and Wooller 1976, Bradley et al. 1989). Our re- sults of high fledgling mortality are congruent with these findings. Fledgling mortality occurs during their first migration, exceeding that of adult gulls; thereafter, in coastal areas, fledg- ling mortality is similar to that of older gulls. The offspring of older parents grow more rap- idly and over a longer duration of parental care compared with offspring of younger par- ents (Pugesek 1995). This may translate into a lower mortality rate during fall migration and an overall higher fledgling survival rate among the offspring of older parents. Patterns of observations in inland areas in- dicated that migrations through these areas re- sulted in high mortality during spring and fall. The majority of Juveniles left the Bamforth Lake area after fledging and remained in coastal areas until they were of breeding age. Subadults also remained in coastal areas year- round. One-year-old gulls are rarely observed and 2-year-olds are infrequently observed m the Bamforth Lake and surrounding area dur- ing summer months (BHP, pers. obs.). Our results are consistent with earlier find- ings of cost of reproduction for California Gulls at Bamforth Lake measured in terms of reduced survival (Pugesek 1987, Pugesek and Diem 1990). As gulls age, they (1) skip breed- ing less frequently (Pugesek and Wood 1992), (2) exert higher levels of reproductive effort when breeding (Pugesek 1981, 1983), and (3) produce more offspring (Pugesek and Diem 1983). Those gulls that skip breeding increase their prospects for survival by remaining in coastal areas, thereby avoiding inland migra- tions to and from the breeding colony. Our observations of mortality among adults during the breeding season are insufficient to explain the declines in survival rate associated with high numbers of offspring fledged (Pugesek and Diem 1990, Pugesek et al. 1995) and lon- ger durations of parental care (Pugesek 1987). Few adult deaths were observed during the breeding season (Pugesek 1983). We suggest Pugesek and Diem • MORTALITY OF CALIFORNIA GULLS 165 the reduced survival rates among gulls that fledged greater numbers of offspring over lon- ger durations of parental care were due to higher mortality during the fall migration. This is true for older gulls, in particular, those that lose greater amounts of body mass during the breeding season compared with younger gulls (Pugesek and Diem 1990). We hypoth- esize that gulls that lose high amounts of body mass during the breeding season may be re- quired to remain in inland areas longer after the breeding season to regain their physical condition but may be trapped in late fall be- fore they can return to the Pacific coast. There is potential bias in categorizing time of death, but we note that large numbers of gulls re- mained in southeastern Wyoming and in Col- orado, while other gulls were on migration routes to the Pacific coast. If physical exhaustion prior to migration is a key source of reproductively-induced mor- tality, California Gulls at Bamforth Lake are probably on the extreme end of a severity scale for this phenomenon. Gulls typically ar- rived in the Bamforth Lake area for the breed- ing season before snow cover dissipated and lakes thawed. The Bamforth Lake area also provided severe thermal stress. During the ter- ritorial and egg-laying phases of reproduction, gulls were frequently exposed to average daily wind chills in excess of —20° C (BHP, unpubl. data) and June snowstorms were a common event. By late summer, gulls captured using cannon nets were frequently in poor physical shape as evidenced by sunken, boney keels and low body mass (Pugesek and Diem 1990). Finally, the post-breeding migration to the Pa- cific coast is longer for gulls east of the Con- tinental Divide. Populations with a less stress- ful nesting habitat and a shorter migration may experience lower reproductively-induced mortality compared with the Bamforth Lake population. Under the circumstance of high reproduc- tively-induced mortality, we expect that selec- tion pressures for increasing reproductive ef- fort with age are stronger in the Bamforth Lake population compared with populations nesting near the Pacific coast or in less stress- ful habitats. It would be of interest to compare life history characteristics among populations of California Gulls, for example, those nesting west of the Continental Divide with those nesting east of the divide where migrations are longer and environmental stress is likely harsher. Those nesting east of the Continental Divide may exhibit greater age-specific vari- ation in egg mass, clutch size, fledging suc- cess, age at first breeding, frequency of skipped breeding attempts, and age-specific variation in survival rates. Also, behavioral measures which are indicative of reproductive effort such as nest defense rates (Pugesek 1983) can be easily duplicated and compared among populations. ACKNOWLEDGMENTS Our research was supported by grants from NSF (DEB 791997), Sigma Xi, Wilson Ornithological So- ciety, Chapman Foundation, and NIA (T32-AG001 10- 03). We thank the U.S. Fish and Wildlife Service, Arapaho National Wildlife Refuge, and Wyoming Game and Fish Department for cooperation. We es- pecially thank C. R. Come for developing and testing the data base for the analysis. J. R. Jehl Jr., R. M. Erwin, and one anonymous reviewer provided helpful comments. LITERATURE CITED Baird, S. E, T M. Brewer, and R. Ridgway. 1884. The waterbirds of North America. Volume 2. Memoirs of the Museum of Comparative Zoology at Harvard College 13:1-549. Behle, W. H. 1958. The bird life of the Great Salt Lake. University of Utah Press, Salt Lake City, USA. Bradley, J. S., R. D. Wooller, I. J. Skira, and D. L. Serventy. 1989. Age-dependent survival of breeding Short-tailed Shearwaters Puffinus tenui- rostris. Journal of Animal Ecology 58:175-188. Cooke, W. W. 1915. Distribution and migration of North American gulls and their allies. U.S. De- partment of Agriculture Bulletin 292. CouLSON, J. C. AND R. D. WooLLER. 1976. Differential survival rates among breeding Kittiwake Gulls Rissa tridactyla (L.). Journal of Animal Ecology 45:205-214.' Jehl Jr., J. R. 1987. Geographic variation and evolu- tion in the California Gull {Larus californicus). Auk 104:421-428. JoBSON, J. D. 1992. Applied multivariate data analysis. Volume II. Categorical and multivariate methods. Springer- Verlag, New York, USA. Molina, K. C. 2000. The recent nesting of California and Laughing gulls at the Salton Sea. California. We.stern Birds 31:106-1 1 1. Pollock, K. IL, J. D. Nichols, C. Brownie, and J. E. Hine:s. 1990. Statistical inference for capture-re- capture experiments. Wildlife Monographs 107. PcGESEK, B. H. 1981. Increased reproductive effort 166 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 with age in the California Gull {Larus californi- cus). Science 212:822—823. PUGESEK, B. H. 1983. The relationship between paren- tal age and reproductive effort in the California Gull {Larus californicus). Behavioral Ecology and Sociobiology 13:161-171. PuGESEK, B. H. 1987. Age-specific survivorship in re- lation to clutch size and fledgling success in Cal- ifornia Gulls. Behavioral Ecology and Sociobiol- ogy 21:217-221. PUGESEK, B. H. 1995. Chick growth in the California Gull: reproductive effort versus foraging efficien- cy hypotheses. Animal Behaviour 49:641-647. PUGESEK, B. H. AND K. L. DiEM. 1983. A multivariate study of the relationship of parental age to repro- ductive success in California Gulls. Ecology 64: 829-839. PUGESEK, B. H. AND K. L. DiEM. 1990. The relationship between reproduction and survival in known-aged California Gulls. Ecology 71:811-817. PUGESEK, B. H. AND P. K. WOOD. 1992. Alternate re- productive strategies in the California Gull. Evo- lutionary Ecology 6:279-295. PUGESEK, B. H., K. L. Diem, and C. L. Cordes. 1999. Seasonal movement, migration, and home ranges of immature and adult California Gulls nesting at Bamforth Lake, Wyoming. Colonial Waterbirds 22:29-36. PUGESEK, B. H., C. Nations, K. L. Diem, and R. Pra- DEL. 1995. Mark-resighting analysis of a Califor- nia gull population. Journal of Applied Statistics 22:625-639. Winkler, D. W. 1996. California Gull {Larus califor- nicus). The birds of North America. Number 259. The Wilson Journal of Ornithology 120(1); 167-175, 2008 EFFECTS OF PREDATION AND FOOD PROVISIONING ON BFACK TERN CHICK SURVIVAE SHANE R. HEATH* 2 3 AND EREDERICK A. SERVELLO* ABSTRACT. — We placed predator exclosures around 31 Black Tern {Chlidonias niger) nests in Maine in 2001-2002 to measure growth and survival of chicks. Fifty-four percent of exclosed nests that hatched young were depredated in 2001 and four exclosed nests were abandoned prior to hatch. We modified our exclosure design in 2002 and only one nest (7%) was depredated and no nests were abandoned prior to hatch. Kaplan- Meier estimates of chick survival in the absence of predation were 0.87 to 18 days in 2001 and 0.90 to 15 days in 2002. Mass ratios among first, second, and third-hatched chicks indicated that size hierarchies were present in broods near time of brood completion, but linear growth rates and asymptotic mass were not affected by hatch order in 2- or 3 -chick broods. Predation was the primary determinant of chick survival in Black Tern colonies studied and food availability was not limiting chick growth. Predator exclosures did not prevent all depredation, but our exclosure design was effective at protecting and retaining chicks until fledgling age at 70% of nests; the majority of adults readily accepted predator exclosures. Received 23 December 2006. Accepted 19 April 2007. Black Terns {Chlidonias niger) are listed as endangered, threatened, or a species of con- cern in multiple states (USA) and Canadian provinces (Shuford 1999), and populations de- clined 3.1% annually during 1966-1996 in North America (Peterjohn and Sauer 1997). Low breeding productivity may be a key fac- tor limiting population growth (Servello 2000), especially among small, isolated pop- ulations along the southern portion of this spe- cies’ range. Nest success in this species is well studied (reviewed in Servello 2000), but chick survival estimates are lacking. Monitoring chick survival for wetland-nesting terns is problematic because chicks use dense vege- tation for cover and leave nests soon after hatching (Hall 1988, Cuthbert and Louis 1993). Thus, few studies have reported chick survival rates or investigated factors contrib- uting to chick mortality (Servello 2000). Few- er than 100 pairs of Black Terns nest annually in Maine and chick survival appears to be the primary cause of low breeding success (E A. Servello, unpubl. data). Identifying factors contributing to chick mortality is critical for understanding Black Tern ecology and devel- ' Department of Wildlife Ecology, 5755 Nutting Hall, University of Maine, Orono, ME 04469, USA. ^ Current address: Institute for Wildlife Studies, 2515 Camino del Rio S, Suite 334, San Diego, CA, 92108, USA. ^Corresponding author; e-mail; shaneheathl @gmail.com oping management strategies for increasing recruitment. Black Terns nest in wetlands where a wide variety of predators commonly occur. Reports of predation on Black Tern chicks have been limited to chicks held in small enclosures (Dunn 1979, Chapman and Forbes 1984) or anecdotal observations (Shealer and Haver- land 2000). Low food availability or inade- quate provisioning by adults may be a com- mon cause of chick mortality in Black Terns (Mosher 1986, Beintema 1997) as in other tern species (Safina et al. 1988, Monaghan et al. 1989, Nisbet et al. 1998, Eyler et al. 1999). However, food limitation in Maine eolonies has not previously been observed (Gilbert and Servello 2005). Black Terns exhibit laying and hatching asynchrony (Dunn and Agro 1995), which produces a size hierarchy among siblings and can result in lower growth and survival rates for the youngest chicks in broods when food resources are scarce (Lang- ham 1972, Skagen 1988, Bollinger et al. 1990, Brown and Morris 1996). We hypothesize that predation is the primary factor contributing to chick mortality of Black Terns in Maine, but annual or local variation in food availability may also have a contributing role. Predator exclosures have been successfully used with shorebirds to prevent egg depreda- tion (Rimmer and Deblinger 1990, Estelle et al. 1996, Mabee and Estelle 2000, Larson et al. 2002), but have not been designed to pro- tect chicks. With terns, smaller retention fenc- 167 168 THE WILSON JOURNAL OF ORNITHOLOGY • VoL 120, No. 1, March 2008 es have frequently been used to confine chicks, including those of Black Terns, for re- search purposes (e.g., monitoring growth, sur- vival, feeding behavior) but not for predator deterrence (Morris et al. 1976, Burger et al. 1996, Robinson and Hamer 2000, Shealer and Haverland 2000, Gilbert and Servello 2005). We sought to develop a predator exclosure that would hold chicks until fledging, deter mammalian and avian predators, and be suit- able for wetland substrates while simulta- neously allowing unimpeded feeding and brooding by adults. A useful exclosure would also provide opportunities for research on chick predation and growth. We report on the efficacy of our predator exclosures and test two hypotheses: (1) predator exclusion would result in chick survival near 100% if predation was the principal mortality factor and (2) growth rates of Black Tern chicks would de- crease with hatch order and increasing brood size if food provisioning was inadequate. METHODS Study Area.— This study was conducted at three Black Tern breeding colonies (Carlton Pond, Douglas Pond, and Messalonskee Lake) in Maine, USA, in May-July 2001-2002. Carlton Pond (44° 40' N, 69° 15' W) had a to- tal area of 431 ha and included 75 ha of semi- permanent emergent wetland bordering 1 1 3 ha of open water with dense mats of floating-leaf vegetation. Nesting areas were dominated by sedges (Carex spp.). Sphagnum spp., and pickerelweed {Pantedaria cordata). Douglas Pond (44° 50' N, 69° 21' W) was an impound- ed wetland on the Sebasticook River and had a total area of 227 ha, which included 44 ha of semi-permanent emergent vegetation bor- dering 85 ha of open water. Nesting areas at Douglas Pond were dominated by river bul- rush {Scirpus spp.) or sedges. Messalonskee Lake (44° 26' N, 69° 49' W) was a large lake (1,786 ha) with 55 ha of semi-permanent emergent wetland at one end; nesting areas were dominated by sedges and Sphagnum spp. Field Procedures. — We constructed 17 predator exclosures around individual nests in 2001 and 14 exclosures around nests in 2002. Black Terns differ from other species for which nest exclosures have been constructed in that adults fly rather than walk into exclo- sures, and our design had to accommodate this behavioral difference. Predator exclosures consisted of two primary parts: (1) a circular fence 1 m in diameter and 0.3 m in height placed around the nest (chick retention fence), and (2) a circular fence 4.6 m in diameter and 1 .4 m in height, placed concentrically around the retention fence (predator exclusion fence; Fig. 1). Landscaping cloth was attached to the inside of the chick retention fences to a height of 0.15 m and to the outside of the predator exclusion fences to a height of 0.9 m prior to deployment (Fig. 1). In addition, the retention fences had a 10-15 cm wide flap of chicken wire wrapped with landscaping cloth (“con- cealment flap”) attached inside the retention fence at a height of 0.15 m to provide over- head concealment. The purposes of the reten- tion fence were to reduce predation risk by limiting movements of chicks, keeping chicks from exclosure fences, and providing a loca- tion for horizontal and overhead concealment of chicks. The landscaping cloth in retention fences also prevented chicks from climbing out of the enclosure. The purposes of the ex- clusion fence were to reduce predator access and reduce visibility of chicks to ground pred- ators. We deployed retention fences around nests during the second half of incubation, and installation took less than 3 min. Retention fences were held tightly to the substrate with three wooden stakes. The “concealment flaps” were initially pressed flat against the retention fence until clutches hatched. We avoided hot sunny days when deploying exclosures to avoid egg heating while adults were acclimating to the structure. We removed eggs and placed them in an insulated container during erection of the exclusion fence. Exclu- sion fencing was typically deployed 3—4 days after the retention fencing, but before eggs hatched to allow terns to acclimate to these structures. However, retention and exclosure fences in 2001 were installed simultaneously at four nests and exclusion fencing was in- stalled after hatch at two nests in 2002. Two to four people were required to erect predator exclusion fencing and construction time was typically 15-25 min depending on water depth. Water depths were typically 0.2 to 1.0 m. Eight to 10 wooden stakes (2 m in length) were used to support exclusion fences and hold them flush to the substrate, and four rope guidelines attached to smaller stakes were Heath and Servello • BLACK TERN CHICK SURVIVAL 169 (A) 4.6 m Predator exclusion fence FIG. 1. (A) Predator exclusion fence including landscaping cloth, rope and stakes, and researcher access door, and (B) diagram showing predator exclosure around Black Tern nest, including predator exclusion and chick retention fences. 170 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 used to add tension and reduce movement by strong winds (Fig. 1). The skirt of landscaping cloth was partially submerged under the ex- tant water along with the lower portion of the fence to serve as an additional ground-level barrier. The two ends of the exclusion fencing, overlapped and fastened together with remov- able wires, served as the entrance for re- searchers. We selected entire clusters of nests or in- dividual nests on colony edges for predator exclosures to allow repetitive access to nests without major disturbance to colonies. We vis- ited exclosures approximately every 1-2 days to document hatching and to inspect exclo- sures. Chicks at each nest were assigned an alpha code based on hatch order: A-chick = first hatched, B-chick = second hatched, C-chick = third hatched. We visited exclo- sures to weigh chicks every 2 days following hatch, on average. Chicks missing from ex- closures prior to the anticipated fledge dates were considered depredated if no means of es- cape were evident. Dead chicks were removed from the exclosures to avoid attracting pred- ators. We retained broods in 2001 until A-chicks reached an approximate fledging age of 18 days (Dunn and Agro 1995) except when they flew from exclosures. We released broods in 2002 when A-chicks reached 15 days of age to reduce potential losses of older chicks to predators. Chicks became more ac- tive (increased wing flapping, vocalizations) at this age (Dunn 1979) and appeared to be more vulnerable to predators at older ages in exclosures during 2001. Statistical Analyses. — We calculated Kap- lan-Meier survival estimates for non-depre- dated chicks by censoring depredated, es- caped, or released chicks in analyses (Pollock et al. 1989). We compared survival rates be- tween years using log-rank tests (Pollock et al. 1989). We calculated mass ratios of chicks for each nest with a predator exclosure to ex- amine if size hierarchies were established by hatch order. Mass ratio is an index of the com- petitive ability of the latter-hatched chicks rel- ative to the A-chick during the early post- hatch period (Bollinger 1994). We defined mass ratio as a chick’s mass divided by the mass of the A-chick following brood comple- tion (modified from Bollinger 1994). Brood completion occurred when the final chick of each brood hatched. Mass ratios of B- and C-chicks were compared using ANOVA. B- or C-chicks were assumed to differ from A-chicks if 95% CIs for their mass ratios did not include 1.0. We used linear regression analysis to calculate linear growth rate (LGR; Emms and Verbeek 1991) for each chick for the age period 2-11 days when Black Tern chick growth is approximately linear (Bein- tema 1997). Only chicks with a minimum of three mass measurements during this interval were included. We iteratively fit the logistic equation. Chick mass = AM/{1 + exp[-Ar-(age - to)]}, to growth curves of individual chicks to esti- mate asymptotic mass (AM), where K is a. growth coefficient and tQ is time of inflection (Starck and Ricklefs 1998). We only calculat- ed AM for chicks that reached 13 days of age because weight measurements approaching the asymptote are required (Ricklefs 1967). Chick weights from broods that were even- tually depredated were included if meeting the above criteria for all growth analyses (mass ratio, LGR, AM). We analyzed data for 2- and 3-chick broods separately. We only included 2-chick broods if they resulted from a 2-egg clutch or the fail- ure of the third egg to hatch; broods that lost a chick from a 3-chick brood were excluded. We used linear regression to examine relation- ships of LGR and AM with hatch date. We first examined effects of colony and year on LGR using ANOVA, and then pooled data across colonies and years to separately ex- amine the dependence of LGR and AM on hatch order and brood size using nested AN- OVA (with brood as the nested term). Year and colony effects were analyzed separately because small sample sizes precluded tests of year by colony interactions. RESULTS Predator Exclosiires. — Four nests did not hatch in 2001 and were abandoned (Table 1). Three of these nests were abandoned imme- diately after exclusion fencing was erected, while the fourth nest was incubated past the expected hatch date and eventually aban- doned. Hatching success of nests incubated to Heath and Servello • BLACK TERN CHICK SURVIVAL 171 TABLE 1. Success of Black Tern nests with predator exclosures in Maine wetlands, 2001-2002. Hatching success is defined as the proportion of eggs hatched. Year Colony Nests exclosed Nests hatched (eggs) Eggs hatched Hatching success (%) 2001 Carlton 6 5 (15) 15 100 Douglas 6 4 (12) 11 92 Messalonskee 5 4 (10) 10 100 Totals 17 13 (37) 36 97 2002 Carlton 9 9 (26) 18 69 Douglas 5 5 (15) 15 100 Totals 14 14 (41) 33 81 term was 97% in 2001 and 36 eggs hatched (Table 1). Seven (54%) of 13 nests with young were depredated with 17 chicks lost (Table 2). The two depredated exclosures at Carlton con- tained three 16-day-old chicks and one 14- day-old chick, respectively, while the depre- dated exclosure at Douglas contained three chicks less than a week old. All four nests that hatched at Messalonskee were depredated within a 3 -day period. These four depredated exclosures contained three 15-day-old chicks, three 16-day-old chicks, two 10-day-old chicks, and two 6-day-old chicks, respective- ly. Overall, three chicks from three different broods died of causes unrelated to predation (Table 2). Sixteen chicks were released from exclosures in 2001 (Table 2). No nests were abandoned prior to hatch in 2002. Hatching success of nests incubated to term was 81% in 2002 and 33 eggs hatched (Table 1). Hatching success of exclosed nests at Carlton was 69%, but 100% at Douglas (Ta- ble 1). At Carlton, the third egg of a 3-egg clutch failed to hatch at three nests, and an- other nest containing one chick and two intact eggs was abandoned. One nest (7%) was dep- redated at Carlton in 2002 resulting in loss of three chicks <1 week of age. Three chicks from three different broods died of causes un- related to predation, while one brood of three chicks escaped their exclosure prior to fledg- ing age (Table 2). Twenty-four chicks were released from exclosures in 2002 (Table 2). Predators entering exclosures were not iden- tified in either year. Survival of N on-depredated Chicks. — Six chicks from six different broods died within exclosures in a manner unrelated to predation. Two of these chicks were <5 days of age and were found dead outside of the retention fenc- ing. These chicks may have escaped from the retention fence and perished after becoming separated from their respective broods, or may have died and been removed from the nest by the adults. Two other chicks died after appar- ent abandonment by adults, most likely as a result of investigator disturbance associated with predator exclosures. The remaining two chicks, a 16-day-old B-chick and an 8-day-old C-chick, were both in 3-chick broods and ex- hibited normal growth for 12 and 6 days, re- spectively, before declining in mass prior to TABLE 2. Fate of Black Tern chicks in predator exclosures, including cause of loss, 2001-2002. Year Colony Chicks depredated (nests) Chicks died of other causes (nests) Chicks escaped from exclosure (ne.sts) Chicks released from exclosures (ne.sts) 2001 Carlton 4 (2) 2=* (2) 0 9 (3) Douglas 3 (1) C (1) 0 7 (3) Messalonskee 10 (4) 0 0 0 Totals 17 (7) 3 (3) 0 16 (6) 2002 Carlton 0 3‘-- (3) 0 15 (8) Douglas 3 (1) 0 3(1) 9 (3) Totals 3(1) 3 (3) 3(1) 24 ( 1 1 ) ® One chick was found dead outside of retention fencing and one chick died of apparent starvation. Chick found dead outside of retention fencing. Two chicks from two separate broods died after apparent abandonment by adults, while the third chick died of apparent starvation. Survival rate 172 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 Chick age (days) FIG. 2. Kaplan-Meier survival rates for Black Tern chicks not depredated in predator exclosures in 2001- 2002 at Douglas Pond and Carlton Pond, Maine. Es- timates were based on 17 chicks from six broods in 2001 and 27 chicks from 13 broods in 2002. death. Including the above instances of mor- tality, chick survival (±SE) to 18 days in 2001 in the absence of predation was 0.87 ± 0.08 (95% Cl: 0.71-1.03) and 0.90 ± 0.05 (95% Cl: 0.80-1.00) to 15 days in 2002 (Fig. 2) and did not differ among years {P = 0.46). Growth Analyses. — Nineteen broods (3- chick broods, n = 13; 2-chick broods, n = 6) were used in growth analyses. Mass ratios dif- fered (Fj 24 = 1 1.38, P = 0.003) between B- and C-chicks in 3-chick broods. B-chicks were on average 91% of the mass of A-chicks (n = 13, 95% Cl = 0.83—0.98) near time of brood completion and C-chicks were 73% of the mass of A-chicks (n = 13, 95% Cl = 0.65-0.82). In 2-chick broods, B-chicks were on average 72% of the mass of A-chicks (n = 6, 95% Cl = 0.59-0.86) near brood completion. Neither LGR (P = 0.10, n = 51, = 0.034) nor AM (P = 0.48, n = 43, = 0.001) were affected by hatch date. Estimates of LGR for individual chicks ranged from 3.22 to 5.87 g/day, with a mean of 4.47 g/day. Colony (^2,10 ^ 0-24, P = 0.79) and year (Fj = 1.31, P = 0.28) did not affect LGR; therefore, data were pooled across years and colonies to test hatch order effects. Linear growth rates were not affected by hatch order for either 3 -chick broods (F224 = 2.13, P — 0.14) or 2-chick broods (Fi 5 = 1.06, P = 0.35). Growth rates did not differ (F, 32 = 1.33, P = 0.26) with brood size. Estimates of AM for individual chicks ranged from 39.0 to 73.5 g, with a mean of 61.6 g. Asymptotic mass did not dif- fer among colonies (F28 1.21, R = 0.35) or years (F, 9 = 1.00, P = 0.34) and data were pooled across years and colonies to test for hatch order effects. Asymptotic mass did not differ by hatch order for 3 -chick broods (F220 = 0.71, R = 0.50). We did not test effects of hatch order on AM for 2-chick broods because of small sample sizes. DISCUSSION Predator Exclosures. — Our predator exclo- sure design was not effective in excluding predators in 200 1 , but after our modifications we observed lower predation in 2002. How- ever, it is unknown if this lower predation (7%) resulted from our modifications or from changes in predator abundance between years. Adult terns generally accepted the small re- tention fences relatively quickly (3-15 min), but were more hesitant to initially enter the larger exclosures. Adults frequently circled the structure for >20 min. Once acclimated, adults appeared to travel normally to and from the nest during incubation and chick feeding. Four nests in 2001 were abandoned prior to hatch following deployment of predator ex- clusion fencing. Retention and exclusion fenc- ing were erected simultaneously at two of these nests, resulting in immediate abandon- ment. The predator exclusion fences at all four abandoned nests were erected >6 days prior to hatch, which may have been too early. The willingness of adult Black Terns to leave their nests for extended periods decreases as the in- cubation period lengthens (Cuthbert 1954) suggesting that adult acceptance of exclosures may increase closer to hatching. Nests where exclusion fences were constructed just before or immediately following hatch were not abandoned and we conclude this is the optimal time for deployment. We do not recommend deploying retention fencing and predator ex- clusion fencing on the same day. Predators successfully entered more than half of all exclosures where eggs hatched in 2001. A high proportion of depredated chicks were relatively old (58% at 14—17 days of age) suggesting that older chicks in exclosures may be susceptible to loss as they approach fledging. Chicks may become more conspic- uous as they approach fledging as a result of wing posturing behaviors and increased vo- Heath and Servello • BLACK TERN CHICK SURVIVAL 173 calization rates that could attract predators. We observed that chicks were often well hid- den around the inside perimeter of retention fences or remained immobile under vegetation when we visited exclosures. However, in 2001 when we used less physical or vegetative cov- er, chicks moved continuously in the retention fence seeking escape cover, which increased their visibility to potential predators. We mod- ified our techniques in 2002 by adding addi- tional vegetation and/or artificial camouflage netting in and around retention fences, as well as releasing broods at earlier ages. Only one exclosed nest was depredated in 2002 sug- gesting our modifications were highly effec- tive. However, placement of additional vege- tation and camouflage in 2002 likely resulted in lower hatching success of exclosed nests relative to 2001. The third egg of a 3 -egg clutch failed to hatch in three nests in 2002. We hypothesize the addition of dense vege- tation within retention fences following hatch of the first two chicks impeded the view of the nesting adults and resulted in their failure to properly incubate the remaining egg. Adults may have only landed in exclosures to feed chicks or brood chicks for short periods. We recommend waiting several days after all chicks have hatched before gradually adding vegetation/camouflage to the retention fence over the course of several visits. Chick Survival. — Nearly 90% of chicks sur- vived in the absence of predation to 13-18 days in 2001-2002. We observed only six cas- es of mortality due to other causes and four of these were most likely influenced by in- vestigator disturbance. These results suggest that predators were a primary cause of chick mortality in these wetlands and that weather, food resources, or disease were not signifi- cantly influencing chick survival at these sites in the 2 years studied. We cannot assess the magnitude of natural mortality due to preda- tion without data on chick survival for unex- closed broods; however, the difficulty we had in preventing predation of broods attests to a significant level of predator activity in these wetlands. We could not identify the primary predators of chicks in this study. Many species occurring in Maine have been directly ob- served preying on Black Tern chicks and nests, or else have been implicated by indirect evidence in other studies: Great Blue Heron (Ardea herodias) (Chapman and Forbes 1984, Shealer and Haverland 2000), mink (Mustela vison) (Dunn 1979; Hickey 1997; F. A. Ser- vello, pers. obs.). Northern Harrier {Circus cy- aneus) (this study). Great Homed Owl {Bubo virginianus) (Bailey 1977, Einsweiller 1988), Black-crowned Night-heron {Nycticorax nyc- ticorax) (Bailey 1977), common raccoon {Procyon lotor) (F. A. Servello, pers. obs.), Common Raven {Corvus corax) (this study), and fish (Don McDougal, pers. comm.). Inadequate food provisioning was not a sig- nificant cause of chick mortality during the 2 years of study. Despite the presence of a size hierarchy within broods during the early post- hatch period, differential growth (LGR and AM) with hatch order was not observed for either 2- or 3-chick broods. Furthermore, nei- ther LGR nor AM decreased with increasing brood size, as would be expected if food pro- visioning was inadequate. Gilbert and Servel- lo (2005) similarly found little evidence of food limitation during the 2 years prior to this study. Mean LGR and AM values for chicks were within ranges reported in other studies (Bailey 1977, Dunn 1979, Mosher 1986, Beintema 1997). Six chicks from three broods in 2001 had LGRs below 4.0 g/day and two of these chicks exhibited LGRs below the 3.32 g/day rate reported for starved Black Tern chicks in Europe (Beintema 1997). Four of these six chicks were depredated prior to release, one chick was found dead, apparently of starvation, and one chick was released. Two of these broods occurred in the same col- ony as broods exhibiting average to high LGRs suggesting factors other than food lim- itation may have been responsible for de- pressed growth. Starvation of chicks can be a direct result of insufficient food resources (poor foraging environment) or an indirect re- sult of low “parental quality,” whereby there is ample food but the parents do not provide sufficient food for their young (Nisbet et al. 1995). Predator exclosures may be a useful con- servation tool for increasing breeding produc- tivity of small, high-risk populations of Black Terns. Our predator exclosures allowed adults to successfully raise broods of fledglings with- in a fenced structure and, therefore, we rec- ommend the design as a research tool to ob- tain chick growth and survival data. Before 174 THE WILSON JOURNAL OF ORNITHOLOGY • VoL 120, No. 1, March 2008 exclosures are implemented as a management tool, we recommend that Black Tern chick predators be identified regionally to assess if our exclosure design would be effective. Pred- ator control (most commonly mammal trap- ping) has increased nest success for ducks (Sovada et al. 2001, Drever et al. 2004) and duckling survival (Pearse and Ratti 2004) in prairie pothole wetlands, but it is unclear if such control would be effective in Maine wet- lands because the primary predators of Black Tern nests and chicks are not known. Preda- tion should be considered a primary factor contributing to mortality of Black Tern chicks in Maine, and predation effects should be ex- amined throughout the breeding range of the species to increase our understanding of fac- tors limiting population growth. ACKNOWLEDGMENTS Funding or other support for this research was pro- vided by the Maine Agricultural and Forest Experi- ment Station, U.S. Fish and Wildlife Service, Maine Outdoor Heritage Fund, Maine Department of Inland Fisheries and Wildlife, and Florida Power and Light — Maine Hydro. We thank Dan Noble, David Pert, S. K. Lemin, Sarah Case, and Nathan Poore for field assis- tance. Shannon Kearney and T A. Gilbert were critical to the development of retaining fence and exclosure designs. We thank Mark McCollough and Don Mc- Dougal for valuable advice on our research and logis- tical support for fieldwork. We thank William Halte- man and Raymond O’Connor for assistance with sta- tistical analyses. We thank L. A. Hierl, T. J. Mabee, C. T. Collins, and C. E. Braun for useful reviews of this manuscript. This is Article Number 2952 of the Maine Agricultural and Forest Experiment Station. LITERATURE CITED Bailey, P. F. 1977. The breeding biology of the Black Tern {Chlidonias niger surinamensis). Thesis. University of Wisconsin, Oshkosh, USA. Beintema, a. J. 1997. European Black Terns {Chli- donias niger) in trouble; examples of dietary problems. Colonial Waterbirds 20:558-565. Bollinger, P. B. 1994. Relative effects of hatching order, egg-size variation, and parental quality on chick survival in Common Terns. Auk 111:263— 273. Bollinger, P. B., E. K. Bollinger, and R. A. Ma- LECKi. 1990. Tests of three hypotheses of hatching asynchrony in the Common Tern. Auk 107:696- 706. Brown, K. M. and R. D. Morris. 1996. From tragedy to triumph: renesting in Ring-billed Gulls. Auk 113:23-31. Burger, J., I. C. T. Nisbet, C. Safina, and M. Goch- FELD. 1996. Temporal patterns in reproductive suc- cess in the endangered Roseate Tern {Sterna dou- gallii) nesting on Long Island, New York, and Bird Island, Massachusetts. Auk 113:131-142. Chapman, B. A. and L. S. Forbes. 1984. Observations on detrimental effects of Great Blue Herons on breeding Black Terns. Journal of Field Ornithol- ogy 55:251-252. CUTHBERT, F. J. AND M. Y. Louis. 1993. The Forster’s Tern in Minnesota: status, distribution, and repro- ductive success. Wilson Bulletin 105:184—187. CuTHBERT, N. L. 1954. A nesting study of the Black Tern in Michigan. Auk 71:36-63. Drever, M. C., A. Wins-Purdy, T. D. Nudds, R. G. Clark, and D. A. Haukos. 2004. Decline in duck nest success revisited: relationships with predators and wetlands in dynamic prairie environments. Auk 121:497-508. Dunn, E. H. 1979. Nesting biology and development of young in Ontario Black Terns. Canadian Field- Naturalist 93:276-281. Dunn, E. H. and D. J. Agro. 1995. Black Tern {Chli- donias niger). The birds of North America. Num- ber 147. Einsweiller, S. S. 1988. Black Tern nesting biology in Cheboygan County, Michigan. Thesis. Central Michigan University, Mount Pleasant, USA. Emms, S. K. and N. A. M. Verbeek. 1991. Brood size, food provisioning and chick growth in the Pigeon Guillemot Cepphus columba. Condor 93:943-951. Estelle, V. B., T. J. Mabee, and A. H. Farmer. 1996. Effectiveness of predator exclosures for Pectoral Sandpiper nests in Alaska. Journal of Field Orni- thology 67:447-452. Eyler, T. B., R. M. Erwin, D. B. Stotts, and J. S. Hatfield. 1999. Aspects of hatching success and chick survival in Gull-billed Terns in coastal Vir- ginia. Waterbirds 22:54-59. Gilbert, T. A. and F. A. Servello. 2005. Insectivory versus piscivory in Black Terns: implications for food provisioning and growth of chicks. Water- birds 28:436-444. Hall, J. A. 1988. Early chick mobility and brood movements in the Forster’s Tern {Sterna forsteri). Journal of Field Ornithology 59:247—251. Hickey, J. M. 1997. Breeding biology and population dynamics of the Black Tern in western New York. Thesis. Cornell University, Ithaca, New York, USA. Langham, N. P. E. 1972. Chick survival in terns {Ster- na spp.) with particular reference to the Common Tern. Journal of Animal Ecology 41:385-396. Larson, M. A., M. R. Ryan, and R. K. Murphy. 2002. Population viability of Piping Plovers: effects of predator exclusion. Journal of Wildlife Manage- ment 66:361-371. Mabee, T. J. and V. B. Estelle. 2000. Assessing the effectiveness of predator exclosures for plovers. Wilson Bulletin 112:14-20. Monaghan, P, J. D. Uttley, M. D. Burns, C. Thaine, AND J. Blackwood. 1989. The relationship be- Heath and Servello • BLACK TERN CHICK SURVIVAL 175 tween food supply, reproductive effort and breed- ing success in the Arctic Tern Sterna paradisaea. Journal of Animal Ecology 58:261-274. Morris, R. D., R. A. Hunter, and J. E McElman. 1976. Factors affecting the reproductive success of Common Tern {Sterna hirundo) colonies on the lower Great Lakes during the summer of 1972. Canadian Journal of Zoology 54:1850-1862. Mosher, B. 1986. Factors influencing reproductive success and nesting strategies in Black Terns. Dis- sertation. Simon Fraser University, Burnaby, Brit- ish Columbia, Canada. Nisbet, I. C. T, J. A. Spendelow, and J. S. Hatfield. 1995. Variations in growth of Roseate Tern chicks. Condor 97:335-344. Nisbet, I. C. T, J. A. Spendelow, J. S. Hatfield, J. M. ZiNGO, AND G. A. Gough. 1998. Variations in growth of Roseate Tern chicks. II. Early growth as an index of parental quality. Condor 100:305- 315. Pearse, a. T. and j. T. Ratti. 2004. Effects of predator removal on Mallard duckling survival. Journal of Wildlife Management 68:342-350. Peterjohn, B. G. and J. R. Sauer. 1997. Population trends of Black Terns from the North American Breeding Bird Survey, 1996-1996. Colonial Wa- terbirds 20:566-573. Pollock, K. H., S. R. Winterstein, C. M. Bunck, and P. D. Curtis. 1989. Survival analysis in telemetry studies: the staggered entry design. Journal of Wildlife Management 53:7-15. Ricklefs, R. E. 1967. A graphical method of fitting equations to growth curves. Ecology 48:978-983. Rimmer, D. W and R. D. Deblinger. 1990. Use of predator exclosures to protect Piping Plover nests. Journal of Field Ornithology 61:217-223. Robinson, J. A. and K. C. Hamer. 2000. Brood size and food provisioning in Common Terns Sterna hirundo and Arctic Terns S. paradisaea: conse- quences for chick growth. Ardea 88:51-60. Safina, C., J. Burger, M. Gochfeld, and R. H. Wag- ner. 1988. Evidence for prey limitation of Com- mon and Roseate tern reproduction. Condor 90: 852-859. Servello, E A. 2000. Population and research priori- ties for Black Terns developed from modeling analyses. Waterbirds 23:440-448. Shealer, D. a. and j. a. Haverland. 2000. Effects of investigator disturbance on the reproductive be- havior and success of Black Terns. Waterbirds 23: 15-23. Shuford, W. D. 1999. Status assessment and conser- vation plan for the Black Tern (Chlidonias niger surinamensis) in North America. USDI, Fish and Wildlife Service, Denver, Colorado, USA. Skagen, S. K. 1988. Asynchronous hatching and food limitation: a test of Lack’s hypothesis. Auk 105: 78-88. SovADA, M. A., R. M. Anthony, and B. D. J. Batt. 2001. Predation on waterfowl in arctic tundra and prairie breeding areas: a review. Wildlife Society Bulletin 29:6-15. Starck, j. M. and R. E. Ricklefs. 1998. Avian growth and development: evolution within the altricial- precocial spectrum. Oxford University Press, Ox- ford, United Kingdom. The Wilson Journal of Ornithology 120( 1): 176-180, 2008 USE OF CLAY LICKS BY MAROON-FRONTED PARROTS (RHYNCHOPSITTA TERRIS!) IN NORTHERN MEXICO RENE A VALDES-PENA,'-2 sONlA GABRIELA ORTIZ-MACIEL,‘ SIMON O. VALDEZ JUAREZ,' ERNESTO C. ENKERLIN HOEFLICH,' AND NOEL E R. SNYDER' ABSTRACT. Geophagy has been documented in many species of birds, including many parrots, and its proposed functions include detoxification of dietary poisons, mineral supplementation, and acid buffering. Most geophagy reports involve tropical South American species; we present the first published report of clay-lick use by the Maroon-fronted Parrot (Rhynchopsitta terrisi), a species inhabiting high elevation temperate pine-oak (Pinus sw.-Quercus spp.) forest of the Sierra Madre Oriental. Thirty-six observation sessions were made at he four dispersed licks known from the restricted breeding range of this species. All known licks were near val ey bottoms far below most nesting cliffs. Parrot visitations to ingest clay were characteristically m groups and the average number of parrots per group was nine individuals. Group visits averaged 18.3 mm m duration and peaked between 0900 and 1 100 hrs. The total number of parrots visiting licks during any day represented only a fraction of the known population of the species, suggesting that unless additional licks have yet to be discov- ered, visits of individuals to licks are relatively infrequent. Received 18 September 2006. Accepted 22 May 2007. Geophagy is known primarily in herbivo- rous vertebrates including many ungulates, primates, elephants, reptiles, and a great di- versity of birds (Emmons and Stark 1979, Jones and Hanson 1985, Izawa 1993). In birds, the trait has been described in cracids, corvids, pigeons, grouse, cassowaries, and hornbills (Prendergast and Boag 1970, Em- mons and Stark 1979, Diamond et al. 1999), but is perhaps best known in parrots, where colorful aggregations of many species assem- ble at some of the earthen cliffs overlooking tributaries of the Amazon River. For example, hundreds of individuals of more than 20 parrot species are known to use licks overlooking the Tambopata River in the Tambopata-Candamo Reserve in Peru (Munn 1994, Brightsmith and Aramburu 2004). Other well known licks are in Manu National Park and along the Madre de Dios River of Peru (Gilardi et al. 1999); in Bolivia, >1,000 individuals of at least six spe- cies have been observed at a lick in the Valle de la Luna (Mee et al. 2005), while other licks occur in Amazonian regions of Colombia, Ec- uador, and Brazil (Izawa 1993, Kyle 2001). Geophagy is also known for Grey Parrots (Psittacus erithacus) in the Congo Basin for- est (May 2001) and a variety of Asian species ‘ Centro de Calidad Ambiental, ITESM, CEDES 5°piso, Avenida, Eugenio Garza Sada 2501 sur CP 64849, Monterrey, N.L. Mexico. 2 Corresponding author; e-mail; ravp@itesm.mx in Papua, New Guinea (Diamond et al. 1999). Other observations of parrots eating soil have involved Lear’s Macaws {Anodorhynchus lean) on their nest cliffs in eastern Brazil, Red-fronted Macaws (Ara rubrogenys) on their nest cliffs in Bolivia, and cockatoos in- gesting soil in a city park in Australia (J. D. Gilardi, pers. comm.). A variety of functions has been suggested for clay ingestion including gaining of mineral supplements, binding of toxic materials in di- ets, and regulation of pH in gut contents (Di- amond et al. 1999, Gilardi et al. 1999). Stud- ies by Gilardi et al. (1999) and Brightsmith and Aramburu (2004) in Peru suggest that both mineral supplementation and cytoprotec- tion from secondary poisons in plant foods may be especially important functions. Clay ingestion has not been well docu- mented for parrots in Mexico. There have been observations suggesting ingestion of soil by Military Macaws {Ara militaris) in north- eastern and southwestern Mexico (Juan Var- gas and Carlos Bonilla, pers. comm.), and Thick-billed Parrots {Rhynchopsitta pachy- rhyncha) and Lilac-crowned Parrots {Amazo- nafinschi) in northwestern Mexico (Juan Var- gas, pers. comm.). The trait was first clearly recognized in the Maroon-fronted Parrot {Rhynchopsitta terrisi) in 1995 at two mud- cliff sites still in use by the species today (NFRS). The objective of our study was to document basic features of clay-lick use by 176 Valdes-Pena et al. • MAROON-FRONTED PARROTS 177 the Maroon-fronted Parrot during the breeding season, quantifying such features as group size, daily peaks in activity, and associated be- haviors during clay ingestion. METHODS We observed four clay licks from June to September 2005 used by Maroon-fronted Par- rots in 12, 3-day visits. All licks were in the Sierra Madre Oriental west and southwest of Monterrey, Nuevo Leon and were well dis- tributed in the breeding range of the species in Nuevo Leon and Coahuila (Macias-Cabal- lero 1998). Nesting of Maroon-fronted Parrots is mainly colonial and is only known to occur in natural cavities in vertical limestone cliffs. The parrots occur in this area only during the summer breeding season and migrate south in the Sierra Madre Oriental for the remainder of the year. We do not know if other licks occur in breeding and wintering areas of the species; however, coverage of the breeding area has been sufficiently intensive, especially through interviews of local residents, that we believe most licks in this region have been located. Two of the licks (Santa Cruz and Santa Rosa) are in Nuevo Leon. The other two (El Tem- poral and Ultimagua) are in Coahuila. Topo- graphically, all four licks are close to valley bottoms in mud cliffs overlooking seasonal streams. None of the four licks is higher than 40 m above stream level and all are between 1,600 and 1,800 m above sea level. The most important nest cliffs of the species are at high- er elevations (2,500-3,000 m above sea level). The climate of the study area in the north- ern Sierra Madre Oriental corresponds to tem- perate subhumid forest with summer rains predominant. The annual temperature average is 14° C, with a range of 0 to 34° C. The av- erage annual precipitation is 600 mm of which 370 mm occurs between June and September. The vegetation is predominantly pine-oak {Pi- nas spp.-Quercus spp.) forest. The Maroon- fronted Parrot primarily subsists on pine seeds of several species (mainly Pinas cemhroides) with some use of acorns and agave seeds (Ma- cias-Caballero et al. 1996). Much of this diet is considered to be chemically well defended with tannins (Wong 2007). Clay licks were observed with binoculars and/or telescopes between 0700 (sunrise) and 1700 hrs (before sunset) from 50 to 100 m. depending on topography and availability of suitable viewing sites. We recorded the num- ber of parrots arriving at the clay licks, time of arrival, time spent eating clay per flock, weather, and notes about their behavior and/ or predator activity. We estimated the average number of parrots per flock that consumed clay and the average time these flocks were at clay licks. RESULTS Behavior. — Parrots usually arrived at clay licks in small groups (1-5 birds) giving typi- cal flight vocalizations. The birds perched in nearby trees prior to landing at the clay cliffs, possibly assessing their surroundings for nat- ural enemies. The first bird to land at the cliff was generally nervous in demeanor and nor- mally did not begin feeding until joined close- ly by other parrots (within the next 2 min). During feeding, parrots flew to and from the lick regularly to change location on the cliff or to increase spacing with other parrots. We noted some parrots remained alert while oth- ers fed. The alert posture switched from one individual to another over time and, occasion- ally, the entire group on the cliff took flight simultaneously when one individual uttered a sharp alarm call. Feeding on soil typically oc- curred just below the top of the low cliffs, allowing the birds to quickly gain speed in leaving the cliffs by losing altitude. Sites used by parrots for ingesting clay showed the typ- ical bird-created horizontal grooves, ledges, and hollows known for parrot clay licks in other regions and following particular soil ho- rizons, which can be an indicator of preferred soils. Periods of Clay Consamption. — Use of licks occurred throughout the day with the ex- ception of the last 2 hrs in late afternoon (1500-1700 hrs). Use was greatest during morning and declined during the afternoon (Fig. 1). Peak use occurred in late morning (0900-1 100 hrs), although at one lick the largest number of birds was observed between 0700 and 0900 hrs. The number of birds at another lick was equal between 0700-0900 and 0900-1 100 hrs (Table 1). Groap Size and Naniher of Parrots at Licks. — Thirty-six groups of parrots were re- corded in 12 days of observation with an av- erage of three group visits per day for each of THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. I, March 2008 FIG. 1. Time of feeding (all observations) by Maroon-fronted Parrots at clay licks. Sierra Madre Oriental, Mexico (approx, sunrise at 0700 hrs and sunset at 1800 hrs). the four clay licks (Table 1). The maximum simultaneous count of parrots at a lick was 28 and the maximum daily total for any lick was 63 individuals (in 5 groups). We recorded 326 birds eating clay and the average daily total was nine parrots per lick. These totals com- prised only a small fraction of the estimated total parrot population (based on roost counts to be ~3,500 individuals at the time of the study). Duration of Time at Licks.— Groups of par- rots stayed on the cliffs from 1 to 52 min (Ta- ble 2) with an average of 18.3 min per group visitation. The averages are for groups and not for individuals, many of which spent signifi- cant amounts of time during feeding sessions in alert postures. DISCUSSION The Maroon-fronted Parrot exhibited a strong preference for use of clay licks in late morning (0900—1100 hrs) in contrast to the temporal use of licks by parrots in tropical America. The main use period at tropical American licks is during the first hour after dawn (Brightsmith 2004; Mee et al. 2005; J. D. Gilardi, pers. comm.). Large macaws and some parakeets have their main activity in late morning and afternoons (Brightsmith 2004). Maroon-fronted Parrots also visit licks in ear- ly morning, but the emphasis on late morning visitations suggests clay ingestion followed first feedings. We are unsure if additional clay licks exist in the Maroon-fronted Parrot range. However, interviews of local residents and knowledge that all known licks were in low-elevation sites of relatively easy access, suggest the four known licks could represent all licks in use by the parrot population or at least those they most used and frequented. This possibility is also suggested by the dispersed distribution of the licks relative to known nesting colonies of the species. Twenty-one nesting colonies are presently known for the species (Macias- Caballero 1998) and the greatest distance from known colonies to the nearest lick is 17.5 km. This distance compares with dis- tances of 16 km estimated for parrots flying to and from licks in Peru (Brightsmith 2004). Ortiz-Maciel (2000) estimated individual Ma- roon-fronted Parrots flew as far as 23 km from breeding colonies and the known distribution TABLE 1 . Numbers of Maroon-fronted Parrots at clay licks in the Sierra Madre Oriental, Mexico. Numbers of parrots Time (hrs) Sta. Cruz Sta. Rosa Temporal Ultimagua Totals Range X ± SD Groups («) 0700-0900 6 44 8 25 83 2-16 8.3 ±4.4 10 0900-1100 42 34 70 25 171 2-28 11.4 ±7.8 15 1100-1300 21 22 0 12 55 1-8 7.8 ±5.8 7 1300-1500 4 10 3 0 17 3-6 4.2 ±1.2 4 1500-1700 0 0 0 0 0 0 0 0 0 Totals 73 110 81 62 326 1-28 9 ±6.4 36 Valdes-Pena et al. • MAROON-FRONTED PARROTS 179 TABLE 2. Time (min) of Maroon-fronted Parrots at clay licks. Sierra Madre Oriental, Mexico. Clay lick Time at clay licks (min) Groups (n) Range ;c ± SD Sta. Cruz 8 1-49 21.1 ± 15.89 Sta. Rosa 15 1-52 15.6 ± 12.05 Temporal 7 9-31 19.1 ± 8.15 Ultimagua 6 5-48 20.3 ± 17.85 Totals 36 1-52 18.3 ± 13.27 of clay licks could serve all known breeding colonies without need for additional licks. The two largest known Maroon-fronted Parrot colonies, comprising about 83% of the breeding population (in 2005), are 15.5 and 9.25 km from the nearest licks, and separated from the licks by 1,200 m elevation. The cost of flying to and from licks may cause parrots to delay visits to licks until late morning after their first feedings. Ortiz-Maciel (2000) indi- cated that peak morning movements of Ma- roon-fronted Parrots occurred between 0600 and 0900 hrs. The relatively low numbers of parrots vis- iting licks on any day strongly suggests that individual parrots may visit licks infrequently. Visits of the species to favored water sites for drinking, at times attracts up to 1,000 individ- uals per day (NFRS). Observations of the closely-related Thick-billed Parrot (Snyder et al. 1999) suggest daily drinking periods for individuals — a dramatic contrast to the appar- ent low frequency of clay ingestion in Ma- roon-fronted Parrots. Maroon-fronted Parrot groups at licks av- eraged nine individuals and spent an average of 18.3 min at the licks. In South America, where many parrot species use the same licks and group sizes are often much larger, the av- erage time individuals spent at licks ranged from 28.7 to 47.4 min for diverse species (Burger and Gochfeld 2003). The behavior of individual Maroon-fronted Parrots at licks was generally similar to that described for other species. Given the average of three groups per day with nine birds at each clay lick (108 birds), we hypothesize parrots use the lick once every 2.7 days based on a breeding population of ~300 birds in 2005. This hypothesis should be tested with uniquely marked parrots to ob- tain a more accurate average and the extent of individual variation in visitation rates. Maroon-fronted Parrots have a diet of pine seeds and acorns which is rich in protein, car- bohydrates, and fiber (Fonseca 2003). Wong (2007) found that pinion seeds (mainly Pinus cembroides) eaten by Maroon-fronted Parrots had low levels of Na, Ca, Mn, and S; these minerals occurred in high levels in consumed clay. Wong (2007) also found that clay had minerals such as kaolin, smectite, and micas that can absorb secondary compounds such as polyphenols and alkaloids suggesting the clays could also buffer acids. These findings suggest that parrots consume clay to help in absorption of dietary toxins, mineral supple- ments, and acid buffering capacity as also demonstrated by Gilardi et al. (1999). We recommend that areas used for geoph- agy be included as important sites for conser- vation of Maroon-fronted Parrots and that fu- ture studies of their ecology continue to doc- ument geophagy activity by including more visits to the clay licks. Future activities should involve local people to implement ecotourism in these areas. ACKNOWLEDGMENTS We thank the team work of the Centro de Calidad Ambiental at ITESM Campus Monterrey, without them this work could not have happened. Their friend- ship and desire to work for conservation of psitacids contributed to our success. We also thank Jamie Gi- lardi, Eduardo Inigo, and Donald Brightsmith for com- ments and revision of this paper, and CONACYT, and the National Commission of Natural Protected Areas of Mexico (CONANP) for financial support. We also thank all the volunteers that collaborated in field work and the local people whose hospitality made our field stays especially pleasant. LITERATURE CITED Brightsmith, D. J. 2004. Effects of weather on avian geophagy in Tambopata, Peru. Wilson Bulletin 1 16:134-145. Brightsmith, D. J. and R. Aramiu'RU. 2004. Avian geophagy and soil characteristics in southeastern Peru. Biotropica 36:534-543. Burger, J. and M. Gochfeld. 2003. Parrot behavior at a Rio Manu (Peru) clay lick: temporal patterns, associations, and antipredator responses. Acta Ethologica 6:23-34. Diamond, J., K. D. Bishop, and J. D. Gilardi. 1999. Geophagy in New Guinea birds. Ibis 14L1S1- 193. Emmons, L. and N. M. Stark. 1979. Elemental com- 180 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 position of a natural mineral lick in Amazonia. Biotropica 11:311—313. Fonseca, J. R. M. 2003. De pifias y pinones. Ciencias 69:64-65. Gilardi, J. D., S. S. Duffey, C. A. Munn, and L. A. Tell. 1999. Biochemical functions of geophagy in parrots: detoxification of dietary toxins and cy- toprotective effects. Journal of Chemical Ecology 25:897-922. IZAWA, K. 1993. Soil-eating by alouatta and ateles. In- ternational Journal of Primatology 14:229-242. Jones, R. L. and H. C. Hanson. 1985. Mineral licks, geophagy and biogeochemistry of North Ameri- can ungulates. Iowa State University Press, Ames, USA. Kyle, T. 2001. Geophagy in the Amazon: mammalian and avian utilization of clay licks in Amazonian Peru. Thesis. Frostburg State University, Frost- burg, Maryland, USA. Macias-Caballero, C. M. 1998. Comportamiento de anidacion y monitoreo de la productividad de las colonias de la cotorra serrana oriental (Rhynchop- sitta terrisi) en el Norte de Mexico. Thesis. Insti- tuto Tecnologico y de Estudios Superiores de Monterrey, Monterrey, Mexico. Macias-Caballero, C. M., E. C. Enkerlin-Hoeflich, N. F. R. Snyder, and A. Garza L. 1996. iVamos por buen camino! Estudio y conservacion de las cotorras serranas orientales. DUMAC 18:28-34. May, D. L. 2001. Grey Parrots of the Congo Basin forest. Psittascene 47:8-10. Mee, a., R. Denny, K. Fairclough, D. M. Pullan, AND W. Boyd-W ALLIS. 2005. Observations of par- rots at a geophagy site in Bolivia. Biota Neotro- pica 5:321-324. Munn, C. A. 1994. Macaws: winged rainbows. Na- tional Geographic 185:118-140. Ortiz-Maciel, S. G. 2000. Uso del paisaje por la Co- torra Serrana Oriental (Rhynchopsitta terrisi). Thesis. Instituto Tecnologico y de Estudios Su- periores de Monterrey, Monterrey, Mexico. Prendergast, B. a. and D. A. Boag. 1970. Seasonal changes in the diet of Spruce Grouse in central Alberta. Journal of Wildlife Management 34:605- 611. Snyder, N. F. R., E. C. Enkerlin-Hoeflich, and M. A. Cruz-Nieto. 1999. Thick-billed Parrot (Rhyn- chopsitta pachyrhyncha). The birds of North America. Number 406. Wong, K. 2007. Funciones bioquimicas de la geofagia en la Cotorra Serrana Oriental {Rhynchopsitta ter- risi). Thesis. Instituto Tecnologico y de Estudios Superiores de Monterrey, Monterrey, Mexico. The Wilson Journal of Ornithology 120(1): 18 1-1 89, 2008 CAVITY NUMBER AND USE BY OTHER SPECIES AS CORRELATES OF GROUP SIZE IN RED-COCKADED WOODPECKERS JOHN J. KAPPES JR.‘ 2 ABSTRACT. — I examined cavity number and use by other species as correlates of Red-cockaded Woodpecker (Picoides borealis) group size at Camp Blanding Training Site (CBTS) and Goethe State Forest (GSF), Florida. Group size was positively correlated with cavity number at CBTS but not at GSF. Other species occupied 1.75 (34%) and 1.19 (27%) cavities/territory at CBTS and GSF, respectively. Statistically controlling for use by other species yielded significant positive correlations between group size and cavity number in 4 of 5 years at CBTS and 2 of 4 years at GSF. Correlations between group size and total use by other species, controlling cavity number, were significantly negative at both sites. Separating total use by species, group size was negatively correlated only with southern flying squirrels (Glaucomys volans) at CBTS and only with Red-bellied Wood- peckers (Melanerpes carolinus) at GSF, as flying squirrels occupied >3 times more cavities/territory at CBTS than at GSF (0.87 vs. 0.28) while Red-bellied Woodpeckers occupied —0.75 cavities/territory at both sites. Flying squirrels may obscure the relationship between the two woodpecker species by limiting their cavity use. Received 10 January 2006. Accepted 28 June 2006. Woodpeckers (Picidae) and other excava- tors are generally less cavity-limited than non- excavating cavity nesters (Wesolowski 1983, Martin 1993). The quality of an excavator ter- ritory (as defined by survival and reproductive success) is a function of the number of suit- able cavities and depletion of this resource by other species (Kilham 1961, Short 1979, Ker- pez and Smith 1990, Li and Martin 1991, In- gold 1994, Sedgwick 1997, Kappes 2004). Cavity availability may limit the size and fit- ness of family groups in the case of cooper- atively breeding cavity nesters (Ligon and Li- gon 1990). The Red-cockaded Woodpecker {Picoides borealis) is a federally endangered coopera- tive breeder endemic to fire-maintained pine (Pinus spp.) forests of the southeastern United States. Family groups consist of one breeding pair and 0-3 helpers, and reproductive success and individual survival increases with group size (Walters 1990, Khan and Walters 2002, USDI 2003). The availability of cavities, which are excavated in living pines and used for nesting and year-round for roosting, is the primary factor affecting territory quality (Li- gon 1970, Copeyon et al. 1991, Walters et al. 1992, Davenport et al. 2000). Cavities require ' Department of Wildlife Ecology and Con.servation, University of Florida, Gainesville, FL 3261 1, USA. ^ Current address: Department of Biological Scienc- es, Virginia Tech, Blacksburg, VA 24061, USA; e-mail: kappes@vt.edu one to many years to excavate but often re- main suitable for a decade or longer, even af- ter periods of abandonment or use by other species (Conner and Rudolph 1995, Harding and Walters 2002). Sets of cavity trees, termed clusters (Walters 1990), are typically aggre- gated within a group’s territory. The most re- cently completed cavity usually becomes the breeding male’s roost (Conner et al. 1998) and the group’s nest cavity (Ligon 1970). Breed- ing females, helpers, and juveniles roost in- dividually in older cavities, including former nest holes, and may be more prone to mortal- ity or dispersal in cavity-limited territories (Carter et al. 1989, Doerr et al. 1989, Carrie et al. 1998, Daniels and Walters 2000, Kappes 2004). Thus, group stability and size can be cavity limited (Copeyon et al. 1991, Carrie et al. 1998). Cavities of Red-cockaded Woodpeckers are commonly usurped and/or pre-emptively oc- cupied by other cavity-nesting species (Kap- pes 1997, 2004). Research on the impact of these interactions on Red-cockaded Wood- peckers has focused on the effects of one spe- cies, southern Hying squirrels {Glaucomys vo- lans) (hereafter Hying squirrels), on the repro- ductive success of groups of Red-cockaded Woodpeckers containing at least a breeding pair (Conner et al. 1996, Laves and Loeb 1996, Mitchell et al. 1999). 1 monitored Red- cockaded Woodpecker group size, cavity number, and rate of cavity use by other spe- 181 182 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 cies at two forests from 1997 to 2001. I used correlative analyses to evaluate the hypotheses that; (1) cavity number may limit group size, and (2) cavity occupation by other species can cause or exacerbate this limitation by reducing cavity availability. Group size is defined as the number of ter- ritory occupants, including “groups” of 0 and 1 . This definition allows evaluation of the re- sponse of group size across a broader range of territory quality (as measured by cavity availability), including territories where cavity availability is too low to support a breeding pair. I predicted that clusters with low cavity availability (few suitable cavities and/or high use by other species) are more likely to be unoccupied (group size = 0), or occupied by single males (group size = 1). Clusters with intermediate levels of cavity availability are more likely to support an unassisted pair (group size = 2). Territories with relatively high cavity availability (high number of cav- ities and/or low use by other species) are more likely to have pairs with helpers (group size ^3). METHODS Study Area.—T\\Q study was conducted at two forests in northern peninsular Florida: Camp Blanding Training Site (CBTS) (81° 58' N, 29° 59' W) in Clay County and Goethe State Forest (GSF) (82° 38' N, 29° 16' W) in Levy County, where —16 (Kappes et al. 2004) and 30 groups of Red-cockaded Woodpeckers occurred, respectively. All cavity-tree clusters were in open, mature longleaf pine (Pinus pal- ustris) forest with an understory dominated by saw palmetto (Serenoa repens) and wiregrass (Aristida stricta). Prescribed fires were con- ducted at 3-5 year intervals. Fieldwork con- sisted of monitoring Red-cockaded Wood- pecker group size, cavity number, and cavity occupancy. I monitored 14 clusters at CBTS in 1997, 13 clusters in 1998 and 1999, and 16 clusters in 2000 and 2001 (total = 72 cluster years). I monitored 15 clusters at GSF in 1998 and 17 clusters from 1999 to 2001 (total = 66 cluster years). I only monitored suitable cav- ities within these clusters, as defined by en- trance diameter <60 mm, an intact floor, and occurrence in a living tree (Rudolph et al. 1990b, Kappes and Harris 1995). All study clusters included at least three suitable cavi- ties and were surrounded by sufficient forag- ing habitat to support a group of Red-cock- aded Woodpeckers (USDI 2003). Cavity Monitoring.— C3.V\iy clusters were censused 7-9 times per breeding season (Apr- Jul) at 10-14 day intervals using diurnal and nocturnal inspections, and roost checks. I ini- tially inspected each cavity in each census in- terval during the day using either ladders, a droplight and a mirror, or a video camera mounted on a telescoping pole (TreeTop Peep- er, Sandpiper Technologies Inc., Manteca, CA, USA). I inspected the bole of each cavity tree to ascertain the status of the resin barrier, which is maintained by Red-cockaded Wood- peckers on currently used cavity trees to deter rat snakes (Elaphe spp.) (Jackson 1974, Ru- dolph et al. 1990a). If a cavity was empty, but exhibited abundant evidence of fresh resin- barrier maintenance (clear sap exuding from resin-wells excavated by the woodpecker), I assumed it was occupied by Red-cockaded Woodpeckers in that census interval (Jackson 1977). This assumption was confirmed with an evening (pre-sunset) roost check in which the cavity was watched from an unobtrusive distance to ascertain if any birds entered it to roost. In most cases, multiple cavities were observed during a single roost check. If, dur- ing the daytime census, the cavity contained flying squirrels, a nocturnal species that dens in cavities during the day, I recorded that spe- cies as the occupant (flying squirrels enter den cavities before woodpeckers exit their roost cavities, and vice versa, precluding any tem- poral partitioning of cavity occupancy). If avi- an eggs or nestlings, or some other taxon were found, I recorded the appropriate species as the occupant at that census. If a cavity was empty during the daytime inspection and ex- hibited marginal evidence of resin-barrier maintenance, or if its resin-barrier status had changed since the previous census, it was re- visited for an evening roost check. Empty cav- ities exhibiting no evidence of current resin- barrier maintenance were included in an even- ing roost check, or re-inspected after dark for roosting birds using an infrared option on a video camera. I calculated the mean number of cavities occupied by each species for each cluster as the average across the 7-9 censuses. A species’ occupancy or cavity use was the mean number of Red-cockaded Woodpecker Kappes • WOODPECKERS: CAVITY NUMBER AND AVAILABILITY 183 cavities in a cluster year that were occupied by that species. Cavities refer specifically to being suitable for Red-cockaded Woodpeck- ers. Data Analysis. — I only used data from 1998 to 2001 to compare the composition of the animal communities occupying cavities at the two study sites as data from 1997 were lack- ing for GSF. Correlation analysis was used to evaluate the relationships between group size and (1) cavity number, (2) use by Red-bellied Woodpeckers (Melanerpes carolinus), (3) use by flying squirrels, and (4) total use by other species. The variables were rarely normally distributed because of small annual sample sizes or inherent skewness of the data and transformations failed to remedy this problem. I used Spearman’s rank methods, with few ex- ceptions, to calculate simple (bivariate) and partial correlations (Shipley 2000). One-tailed tests were used based on a priori predictions that group size would increase with increasing cavity number and decrease with increasing levels of occupancy by other species. P values <0.03 were considered strongly significant, P values >0.03 but <0.05 were considered sig- nificant, and P values >0.05 but <0.10 were considered marginally significant. RESULTS Community Compositions. — The annual means (a mean of means) of group size of Red-cockaded Woodpeckers at CRTS and GSF were 2.26 (range = 2.13-2.46) and 2.31 (range — 2.18-2.41), respectively (Fig. lA). The annual mean number of suitable cavities per cluster was 5.13 (range = 4.88-5.46) at CRTS and 4.35 (range = 4.20-4.47) at GSF. Mean group size was similar at the two for- ests, but the mean number of cavities was con- sistently higher at CRTS (Fig. lA). Annual mean occupancy by Red-bellied Woodpeckers was 0.73 (range = 0.63-0.82) cavities per cluster at CRTS and 0.77 (range = 0.66-0.84) at GSF (Fig. IR). Annual mean occupancy by flying squirrels at CRTS and GSF were 0.87 (range = 0.6-1.03) and 0.28 (range = 0.15-0.38) cavities per cluster, re- spectively. The annual mean number of cavi- ties per cluster occupied by passerines (Great Crested Flycatcher \Myiarchus (V7///7//.vl, Tuft- ed Titmouse [Baeolophus hicolor], and East- ern Rluebirds \Sialia .S7c///.vl) was similar at the CBTS o Cavities GSF □ Group size 1 j o Red-bellied Woodpeckers A Flying squirrels “r n I I I-t-4 (D I I I O j n I I -ro_j I (!) I ■ I u a 1 n 1 2 0 C 18 I 6- 14. 12' IQ- S' T I I I n I I J -r “■ I [] <> I I (in I -r I <> I I □ Other species 0 Vacant n I I $ 1998 1999 2000 2001 FIG. 1. (A) Red-cockaded Woodpecker group size and number of suitable cavities per cluster, (B) mean number of suitable Red-cockaded Woodpecker cavities per cluster occupied by Red-bellied Woodpeckers and southern flying squirrels, and (C) mean number of suit- able Red-cockaded Woodpecker cavities per cluster occupied by other species and mean number of vacant cavities per cluster. Camp Blanding Training Site (CBTS; dashed error bars) and Goethe State Forest (GSF; solid error bars). 1998-2()()1. 184 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. I, March 2008 TABLE 1. Coefficients of variation for annual mean use of Red-cockaded Woodpecker cavities by Red-bellied Woodpeckers, southern flying squirrels, and all other species (i.e., including Red-bellied Wood- peckers and southern flying squirrels) combined. Camp Blanding Training Site (CBTS; n = 5 years) and Goe- the State Forest (GSF; n = 4 years), Florida. Species CBTS GSF Red-bellied Woodpecker 0.116 0.099 Southern Flying Squirrel 0.223 0.401 All other species 0.054 0.070 two sites, averaging 0.10 (range — 0.05-0.13) cavities per cluster at CBTS and 0.12 (range = 0.07-0.17) at GSF. Total cavity use by other species (i.e., other than Red-cockaded Wood- peckers) averaged 1.75 (range = 1.62—1.84) cavities per cluster at CBTS and 1.19 (range = 1.12-1.31) at GSF (Fig. 1C). The mean number of vacant cavities per cluster was sim- ilar at the two sites (Fig. 1C). Cavity use by Red-bellied Woodpeckers was similar at the two sites, as was cavity use by passerines, but use by flying squirrels was 3.11 times greater (range = 2. 3-6.9 across years) at CBTS than at GSF. Total cavity use by other species was 1 .47 times greater (range = 1.40-1.53) at CBTS than at GSF due to greater use by flying squirrels (Fig. 1C). Total use by other species at both sites exhibited less annual variation than either Red-bellied Woodpeckers or flying squirrels (Table 1). Correlations Between Group Size and Mea- sures of Cavity Availability.— CorroXdiXions be- tween group size and cavity number at CBTS were positive but non-significant from 1997 to 1999, and strongly significant in 2000 and 2001 (Table 2). No relationship was detected at GSF (Table 2). Raw cavity numbers under- estimated availability because of high rates of use by other species. I calculated partial cor- relations between group size and cavity num- ber, controlling for use by other species. This substantially increased the correlation at CBTS; the relationship became strongly sig- nificant in all years except 1999 (Table 2). The strength of the relationship increased in all years at GSF, becoming strongly significant in 1998 and marginally significant in 2000 (Ta- ble 2). Correlations Between Group Size and Cav- ity Use by Other Species. — Variation in cavity number affected the relationship between group size and cavity use by other species. Thus, I calculated the partial correlations, con- trolling for cavity number. The partial corre- lations between group size and cavity use by Red-bellied Woodpeckers indicated no consis- tent association between these variables at CBTS (Table 3). The partial correlations be- tween cavity use by flying squirrels and group size were consistently negative, being margin- ally significant in 1997 and 2000, and strongly significant in 1999 and 2001. Partial correla- tions between group size and total cavity use by other species at CBTS revealed a strongly significant negative relationship in all years except 1999 (Table 3). The partial correlations between group size and cavity use by Red-bellied Woodpeckers were negative at GSF, being significant in 1998 and strongly significant in 1999 (Table 3). No relationship between group size and cavity use by flying squirrels was evident at GSF. The partial correlations between group size and total cavity use by other species were consistently negative, being strongly signifi- TABLE 2. Spearman’s rank correlations and partial Spearman’s rank correlations (controlling for use by other species) between Red-cockaded Woodpecker group size and cavity number. Camp Blanding Training Site (CBTS) and Goethe State Forest (GSF), Florida. Forest 1997 1998 1999 2000 2001 Simple correlations CBTS GSF 0.279 No data 0.179 0.335 0.363 0.000 0.514***" 0.121 0.558*** 0.000 Partial correlations CBTS GSF 0.655***" No data 0.624*** 0.627*** 0.027 0.253 0.681*** 0.411* 0.786*** 0.143 ^One-tailed test, *0.10 > P >0.05, ***P < 0.03. Kappes • WOODPECKERS: CAVITY NUMBER AND AVAILABILITY 185 TABLE 3. Spearman’s rank partial correlations, controlling for cavity number, between Red-cockaded Wood- pecker group size and mean number of cavities used by Red-bellied Woodpeckers, southern flying squirrels, and all other species (i.e., including Red-bellied Woodpeckers and southern flying squirrels) combined. Camp Blan- ding Training Site (CBTS) and Goethe State Forest (GSF), Florida. Species 1997 1998 1999 2000 2001 CBTS Red-bellied Woodpecker -0.391*^ -0.317 0.504** 0.034 0.042 Southern Flying Squirrel -0.421* -0.270 -0.695*** -0.414* -0.676*** All other species -0.619*** -0.632*** 0.231 -0.529*** -0.673*** GSF Red-bellied Woodpecker No data -0.472**^> —0 722*** -0.138 -0.310^’ Southern Flying Squirrel No data -0.299 0.016 -0.308 0.222 All other species No data -0.615*** -0.437** -0.564*** -0.231 a One-tailed test, *0.10 > P >0.05, **0.05 > P > 0.03, ***P < 0.03. ^ Pearson’s partial correlation. cant in 1998 and 2000, and significant in 1999 (Table 3). Potential Higher Order Interactions. — The partial correlations between group size and to- tal cavity use by other species were consis- tently and significantly negative at both for- ests, whereas those between group size and individual species differed between forests (Table 3). I hypothesized the relationship be- tween group size and cavity use by Red-bel- lied Woodpeckers changed between forests because of the inter-forest difference in cavity use by flying squirrels. I evaluated this hy- pothesis by calculating partial correlations be- tween group size and cavity use by Red-bel- lied Woodpeckers, controlling for both cavity number and use by flying squirrels. I predicted these correlations would be more negative than those controlling only for cavity number. Further, the negative relationship should in- crease more at CBTS where cavity use by fly- ing squirrels was greater. The results support- ed these predictions (Table 4). The correla- tions at CBTS became strongly significant in 1997 and marginally significant in 2000 and 2001. The correlations at GSF remained rel- atively unchanged, increasing only in 2000 (Table 4). DISCUSSION My primary hypotheses were: (1) Red- cockaded Woodpecker group size is cavity- limited, and (2) use of cavities by other spe- cies can cause or exacerbate this limitation. The data support these hypotheses indicating that group size is limited by cavity availabil- ity, as estimated by accounting for resource use by other species. The primary species causing the negative relationship between group size and total cav- ity use by other species changed between for- ests with level of use by flying squirrels (Table 3, Fig. IB), suggesting complex relationships among the three species studied. Where cavity use by flying squirrels was relatively high (CBTS), it limited both group size and cavity use by Red-bellied Woodpeckers, obscuring the relationship between the two species of woodpeckers. The negative effect of Red-bel- lied Woodpeckers on group size was mani- fested only where cavity use by flying squir- rels was low (GSF) or statistically controlled TABLE 4. Spearman’.s rank partial correlations between Red-cockaded Woodpecker group size and number of cavities used by Red-bellied Woodpeckers, controlling for both .southern flying .squirrel u.se and cavity number. Camp Blanding Training Site (CBTS) and Goethe State Forest (GSF), Florida. Forest 1997 1998 1999 2(KX) 2(K)1 CBTS GSF -0.602***“ No data -0.338 -0.409* -0.006 -0.741*** -0.391* -0.547*** -0.378* -0.214 => One-tailed te.st, *0.10 > P >0.05, ***P < 0.03. 186 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 1, March 2008 Flying Squirrel Red-bellied Woodpecker LIG. 2. Community interaction web for Red-cockaded Woodpeckers, southern flying squirrels, and Red- bellied Woodpeckers. Arrows point from the dominant to subdominant species for each direct, negative ( ) interaction illustrating a transitive dominance hierarchy. (GETS). Flying squirrels may impart a posi- tive indirect effect on group size by limiting cavity use by Red-bellied Woodpeckers (an interaction chain; Wootton 1994) in addition to its direct negative effect on group size (Lawlor 1979; Fig. 2). The management prac- tice of removing flying squirrels from cavities (Mitchell et al. 1999, USDI 2003) may com- petitively release Red-bellied Woodpeckers, which may preempt a response by Red-cock- aded Woodpeckers. Further evidence of a consequential inter- action between flying squirrels and Red-bel- lied Woodpeckers is that annual variation in mean occupancy by each of these species was greater than annual variation in total cavity use by other species (Table 1, Fig. IB, C). This pattern might result from compensation in cavity use by Red-bellied Woodpeckers in response to variation in use by flying squir- rels. I evaluated this explanation by calculat- ing the correlation between the annual mean cavity use of Red-bellied Woodpeckers and that of flying squirrels using data from GETS, where both species were well represented, and found a strongly negative relationship (r, = 0.900, P = 0.037, n = 5 years). The correlations between group-size and other variables were weaker at GSF than at GETS (Tables 2-4). Lower group size varia- tion (within years) at GSF compared to GETS may have contributed to this pattern (Fig. 1 A). Over 98% of the cluster years at GSF in- volved groups composed of 2—3 individuals. Seventy-six percent of the cluster years at GETS involved groups of 2-3 individuals, 15% involved 0-1 individual, and 8% in- volved 4-5 individuals. Another possible ex- planation for the weaker correlations at GSF was a greater rate of cavity tree mortality in that forest compared to GETS (0.29 cavity trees/cluster/year vs. 0.04 cavity trees/cluster/ year; J. J. Kappes, unpubl. data). This rela- tively high rate of cavity tree mortality at GSF may have precluded group size from tracking cavity availability because Red-cockaded Woodpeckers typically abandon cavity trees that die (Ligon 1970) as a resin barrier can no longer be maintained. Unmeasured components of territory qual- ity likely influenced group size at both forests. The most important of these factors is recur- rent fire, which improves the quality of for- aging habitat, thereby enhancing group for- mation, stability, reproductive success, and size (James et al. 1997, Walters et al. 2002, Kappes • WOODPECKERS: CAVITY NUMBER AND AVAILABILITY 187 USDI 2003, Kappes et al. 2004). Members of groups in high-quality foraging habitat may be in better physical condition enabling them to better survive bouts of roosting outside (Wil- liams et al. 1991), thereby bolstering group size even in cavity-limited clusters. Certain aspects of the ecology of Red-cock- aded Woodpeckers may predispose this spe- cies to cavity limitation and exacerbation of this limitation by other species. The slow sup- ply rate of cavities is often exceeded by rates of loss to tree mortality, cavity enlargement, and usurpation by other species, resulting in cavity shortages (Conner et al. 2001, Harding and Walters 2002). Moreover, because recur- rent fire in this ecosystem often maintains low densities of snags, which are used by other excavators as cavity sites (Ligon 1970, Kap- pes and Harris 1995), cavities used by Red- cockaded Woodpeckers constitute a dispro- portionate number of sites available to other cavity nesters, intensifying interactions at these sites. Caution is warranted in inferring a causal relationship between cavity use by other spe- cies and group size. Support for alternative explanations for the correlations, however, is lacking. That larger groups might be better at defending roost cavities from potential usurp- ers is doubtful because each group member defends only its own roost (Davis et al. 2005). Also, other species would be less likely to limit group size if they primarily occupied surplus cavities. However, cavity usurpation by other species was common, demonstrating these species regularly used cavities that would otherwise be used by Red-cockaded Woodpeckers. A Markov model analysis of cavity use data indicates the probabilities of transition in cavity use (over the 10-14 day census intervals), from Red-cockaded Wood- peckers to Red-bellied Woodpeckers and from Red-cockaded Woodpeckers to flying squirrels were 0.014 and 0.015, respectively at CBTS and 0.013 and 0.002, respectively at GSF (JJK, unpubl. data). Most of these transitions were due to direct displacement. The methods used in this study did not quantify outside roosting by Red-cockaded Woodpeckers, but displaced birds were observed roosting out- side or in lower quality cavities while other species occupied apparently suitable back-up sites. Cavity usurpation by other species was the sole cause of frequent outside roosting by adult Red-cockaded Woodpeckers in another north Florida study in which outside roosting could be quantified (Kappes 1993, 1997). This and previous work suggests that both Red-bellied Woodpeckers and flying squirrels are dominant over Red-cockaded Woodpeck- ers in cavity-associated interactions, support- ing the conclusion the negative correlation be- tween group size and other species was caused by the latter. The Red-bellied Woodpecker has been commonly observed physically displac- ing Red-cockaded Woodpeckers from roost cavities (Ligon 1970, Baker 1971, Jackson 1978, Kappes 1997, this study), whereas ob- servations of the reverse are rare (J. H. Carter, unpubl. data; JJK, unpubl. data). Interactions between flying squirrels and Red-cockaded Woodpeckers are rarely observed because they apparently occur at night. Repeated ob- servations of flying squirrels occupying active cavities of Red-cockaded Woodpeckers, while evicted birds roost outside or in lower quality cavities, indicate that flying squirrels often displace this woodpecker (Loeb 1993; JJK, unpubl. data). If Red-cockaded Woodpeckers displaced flying squirrels with any regularity, this should have been observed during the hundreds of pre-sunset roost checks conducted during this study, but no such displacement was observed. Apparent transitions from fly- ing squirrels to Red-cockaded Woodpeckers were often recorded, but most of these likely involved undetected intermediate transitions from flying squirrels to vacancy, which result from the habit of this species for moving fre- quently among multiple den cavities within their territories (Bendel and Gates 1987) or from predation by rat snakes (Kappes 2004). Removal experiments will be required to fur- ther evaluate the hypothesis that cavity occu- pation by other species can limit group size of Red-cockaded Woodpeckers. ACKNOWLEDGMENTS J. M. Davis, John Mahan, Michael Meisenberg, D. I. Stanish, J. A. Surclick, Diana Swan, Melissa Warren, and M. I. Williams assisted with fieldwork. The Flor- ida Army National Guard, Camp Blanding Training Site; Goethe State Forest; and the U.S. Fish and Wild- life Service, Clemson Field Office provided funding, study sites, and logistical support. K. E. Sieving, Ralph Costa, R. r. Engstrom, M. D. Moulton, C. W. Osen- berg, R. K. Schaefer, C. F^ Braun, four anonymous 188 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 reviewers, and the J. R. Walters Laboratory at Virginia Tech (J. R. Walters, L. A. Blanc, B. J. Olsen, J. M. Johnson, and J. L. Perkins) provided useful comments on earlier drafts. LITERATURE CITED Baker, W. W. 1971. Progress report on life history studies of the Red-cockaded Woodpecker at Tall Timbers Research Station. Pages 44-59 in Ecol- ogy and management of the Red-cockaded Wood- pecker (R. L. Thompson, Editor). USDI, Fish and Wildlife Service and Tall Timbers Research Sta- tion, Tallahassee, Florida, USA. Bendel, P. R. and j. E. Gates. 1987. Home range and microhabitat partitioning of the southern flying squirrel {Glaucomys volans). Journal of Mammal- ogy 68:243—255. Carrie, N. R., K. R. Moore, S. A. Stephens, and E. L. Keith. 1998. Influence of cavity availability on Red-cockaded Woodpecker group size. Wilson Bulletin 110:93-99. Carter III, J. H., J. R. Walters, S. H. Everhart, and P. D. Doerr. 1989. Restrictors for Red-cockaded Woodpecker cavities. Wildlife Society Bulletin 17:68-72. Conner, R. N. and D. C. Rudolph. 1995. Excavation dynamics and use patterns of Red-cockaded Woodpecker cavities: relationships with coopera- tive breeding. Pages 343—352 in Red-cockaded Woodpecker: recovery, ecology, and management (D. L. Kulhavy, R. G. Hooper, and R. Costa, Ed- itors). College of Forestry, Stephen F. Austin State University, Nacogdoches, Texas, USA. Conner, R. N., D. C. Rudolph, and J. R. Walters. 2001. The Red-cockaded Woodpecker: surviving in a fire-maintained ecosystem. University of Tex- as Press, Austin, USA. Conner, R. N., D. C. Rudolph, D. Saenz, and R. R. Schaefer. 1996. Red-cockaded Woodpecker nest- ing success, forest structure, and southern flying squirrels in Texas. Wilson Bulletin 108:697-711. Conner, R. N., D. Saenz, D. C. Rudolph, W. G. Ross, AND D. L. Kulhavy. 1998. Red-cockaded Wood- pecker nest-cavity selection: relationships with cavity age and resin production. Auk 115:447- 454. CoPEYON, C. K., J. R. Walters, and J. H. Carter III. 1991. Induction of Red-cockaded Woodpecker group formation by artificial cavity construction. Journal of Wildlife Management 55:549-556. Daniels, S. J. and J. R. Walters. 2000. Between-year breeding dispersal in Red-cockaded Woodpeckers, multiple causes and estimated cost. Ecology 81: 2473-2484. Davenport, D. E., R. A. Lancia, J. R. Walters, and P. D. Doerr. 2000. Red-cockaded Woodpeckers: a relationship between reproductive fitness and habitat in the North Carolina Sandhills. Wildlife Society Bulletin 28:426-434. Davis, J. M., K. E. Sieving, and J. J. Kappes Jr. 2005. Red-cockaded Woodpecker roost cavity defense during the non-breeding season. Florida Field Nat- uralist 33:81-92. Doerr, P. D., J. R. Walters, and J. H. Carter III. 1989. Reoccupation of abandoned clusters of cav- ity trees (colonies) by Red-cockaded Woodpeck- ers. Proceedings of the Annual Conference of the Southeastern Association of Fish and Wildlife Agencies 43:326-336. Harding, S. R. and J. R. Walters. 2002. Processes regulating the population dynamics of Red-cock- aded Woodpecker cavities. Journal of Wildlife Management 66:1083-1095. iNGOLD, D. J. 1994. Influence of nest-site competition between European Starlings and woodpeckers. Wilson Bulletin 106:227-241. Jackson, J. A. 1974. Gray rat snakes versus Red-cock- aded Woodpeckers: predator-prey adaptations. Auk 91:342-347. Jackson, J. A. 1977. Determination of the status of Red-cockaded Woodpecker colonies. Journal of Wildlife Management 41:448-452. Jackson, J. A. 1978. Competition for cavities and Red- cockaded Woodpecker management. Pages 103- 112 m Endangered birds: management techniques for the preservation of threatened species (S. A. Temple, Editor). University of Wisconsin Press, Madison, USA. James, E C., C. H. Hess, and D. Kuprin. 1997. Spe- cies-centered environmental analysis: indirect ef- fects of fire history on Red-cockaded Woodpeck- ers. Ecological Applications 7:118-129. Kappes, J. J. 1993. Interspecific interactions associated with Red-cockaded Woodpecker cavities at a north Florida site. Thesis. University of Florida, Gainesville, USA. Kappes, J. J. 1997. Defining cavity-associated inter- actions between Red-cockaded Woodpeckers and other cavity-dependent species: interspecific com- petition or cavity kleptoparasitism? Auk 1 14:778- 780. Kappes, J. J. 2004. Community interactions associated with Red-cockaded Woodpecker cavities. Pages 458-467 in Red-cockaded Woodpecker: road to recovery (R. Costa and S. J. Daniels, Editors). Hancock House Publishers, Blaine, Washington, USA. Kappes, J. J. and L. D. Harris. 1995. Interspecific competition for Red-cockaded Woodpecker cavi- ties in Apalachicola National Forest. Pages 458- 467 in Red-cockaded Woodpecker: recovery, ecology, and management (D. L. Kulhavy, R. G. Hooper, and R. Costa, Editors). College of For- estry, Stephen E Austin State University, Nacog- doches, Texas, USA. Kappes, J. J., K. E. Sieving, J. Davis, M. D. Adams, J. Garrison, P. Catlett, M. Corby, and R. Cos- ta. 2004. Status and management of Red-cock- aded Woodpeckers at Camp Blanding Training Site, Florida. Pages 198-202 in Red-cockaded Woodpecker: road to recovery (R. Costa and S. J. Kappes • WOODPECKERS: CAVITY NUMBER AND AVAILABILITY 189 Daniels, Editors). Hancock House Publishers, Blaine, Washington, USA. Kerpez, T. a. and N. S. Smith. 1990. Competition between European Starlings and native woodpeck- ers for nest cavities in saguaros. Auk 107:367- 375. Khan, M. Z. and J. R. Walters. 2002. Effects of help- ers on breeder survival in the Red-cockaded woodpecker (Picoides borealis). Behavioral Ecol- ogy and Sociobiology 51:336-344. Kilham, L. 1961. Reproductive behavior of Red-bel- lied Woodpeckers. Wilson Bulletin 73:237-255. Laves, K. S. and S. C. Loeb. 1996. Effects of southern flying squirrels (Glaucomys volans) on Red-cock- aded Woodpecker {Picoides borealis) reproduc- tive success. Animal Conservation 2:295-303. Lawlor, L. R. 1979. Direct and indirect effects of n- species competition. Oecologia 43:355-364. Li, P. and T. E. Martin. 1991. Nest-site selection and nesting success of cavity-nesting birds in high el- evation forest drainages. Auk 108:405-418. Ligon, J. D. 1970. Behavior and breeding biology of the Red-cockaded Woodpecker. Auk 87:255-278. Ligon, J. D. and S. H. Ligon. 1990. Green Woodhoo- poes: life history traits and sociality. Pages 33-65 in Cooperative breeding in birds: long term stud- ies of ecology and behavior (P. B. Stacey and W. D. Koenig, Editors). Cambridge University Press, Cambridge, United Kingdom. Loeb, S. C. 1993. Use and selection of Red-cockaded Woodpecker cavities by southern flying squirrels. Journal of Wildlife Management 57:329-335. Martin, T. E. 1993. Evolutionary determinants of clutch size in cavity-nesting birds: nest predation or limited breeding opportunities? American Nat- uralist 142:937-946. Mitchell, L. R., L. D. Carlile, and C. R. Chandler. 1999. Effects of southern flying squirrels on nest success of Red-cockaded Woodpeckers. Journal of Wildlife Management 63:538-545. Rudolph, D. C., R. N. Conner, and J. Turner. 1990a. Competition for Red-cockaded Woodpecker roost and nest cavities: effects of resin age and entrance diameter. Wilson Bulletin 102:23-36. Rudolph, D. C., H. Kyle, and R. N. Conner. 1990b. Red-cockaded Woodpeckers versus rat snakes: the effectiveness of the resin barrier. Wilson Bulletin 102:14-22. Sedgwick, J. A. 1997. Sequential cavity use in a cot- tonwood bottomland. Condor 99:880-887. Shipley, B. 2000. Cause and correlation in biology: a user’s guide to path analysis, structural equations and causal inference. Cambridge University Press, Cambridge, United Kingdom. Short, L. L. 1979. Burdens of the picid hole-excavat- ing habit. Wilson Bulletin 91:16-28. U.S. Department of Interior (USDI). 2003. Recov- ery plan for the Red-cockaded Woodpecker {Pi- coides borealis): second revision. USDI, Fish and Wildlife Service, Atlanta, Georgia, USA. Walters, J. R. 1990. The Red-cockaded Woodpecker: a “primitive” cooperative breeder. Pages 67-101 in Cooperative breeding in birds: long term stud- ies of ecology and behavior (P. B. Stacey and W. D. Koenig, Editors). Cambridge University Press, Cambridge, United Kingdom. Walters, J. R., C. K. Copeyon, and J. H. Carter III. 1992. Test of the ecological basis of cooperative breeding in Red-cockaded Woodpeckers. Auk 109:90-97. Walters, J. R., S. J. Daniels, J. H. Carter, and P. D. Doerr. 2002. Defining quality of Red-cock- aded Woodpecker foraging habitat based on hab- itat use and fitness. Journal of Wildlife Manage- ment 66:1064-1082. Wesolowski, T. 1983. Conjectures, venturous sugges- tions or ’evolution of hole-nesting in birds’? Ornis Scandinavica 14:63-65. Williams, J. B., M. A. Du Plessis, and W. R. Sieg- fried. 1991. Green Woodhoopoes {Phoeniculus purpureas) and obligate cavity roosting provide a test of the thermoregulatory insufficiency hypoth- esis. Auk 108:285-293. WooTTON, J. T. 1994. The nature and consequences of indirect effects in ecological communities. Annual Review of Ecology and Systematics 25:443-466. Short Communications The Wilson Journal of Ornithology 120(1 ): 190-195, 2008 First Description of Nests and Eggs of Two Hispaniolan Endemic Species; Western Chat-tanager {Calyptophilus tertius) and Hispaniolan Highland-tanager (Xenoligea montana) Christopher C. Rimrner,' Lance G. Woolaver,^ Rina K. Nichois,^ Eiadio M. Fernandez, ^ Steven C. Latta,-* and Esteban Garrido^ ABSTRACT. — We present the first nest descriptions for two Hispaniolan endemic songbirds, the Western Chat-tanager {Calyptophilus tertius) and Hispaniolan Highland-tanager {Xenoligea montana) from a mon- tane broadleaf forest site in the Sierra de Bahoruco of the Dominican Republic. Single Western Chat-tanager nests were found on 17 May 2002 and 9 June 2004. Both were coarsely-built, partially-domed, bulky struc- tures 1.0— 1.5 m above ground. One nest was freshly- depredated when found, whereas the second contained two eggs which hatched on 19 and 20 June. The nest- lings were depredated on 25 June. A Hispaniolan Highland-tanager nest found on 27 June 2004 fledged a single chick the following day. This nest, in a vine tangle 2.5 m above ground, was an open cup structure composed of moss, small herbaceous stems, leaf frag- ments, lichens, and other plant fibers. We describe the eggs of both species, the nestlings of Western Chat- tanager, and the juvenal plumage of Hispaniolan High- land-tanager. We believe that depredation by intro- duced feral cats {Felis domesticus) and rats {Rattus spp.) is a serious problem in these montane forests. Received 26 December 2006. Accepted 16 April 2007. Hispaniola supports more endemic bird spe- cies than any other Caribbean island, but the breeding biology of its avifauna remains poor- ly known. Breeding bird communities in the ' Vermont Institute of Natural Science, 6565 Wood- stock Road, Quechee, VT 05059, USA. 2 Wildlife Preservation Canada, 5420 Highway 6 North, Guelph, ON NIH 6J2, Canada. 3 Sociedad Ornitologica de la Hispaniola, Avenue, Maximo Gomez, esq. San Martin, Edificio Fundacion Progressio, Tercero Piso, Santo Domingo, Republica Dominicana. 4 National Aviary, Allegheny Commons West, Pitts- burgh, PA 15212, USA. 5 Grupo Jaragua, Inc., Calle El Vergel 33, El Vergel, Santo Domingo, Republica Dominicana. ^ Current address; Vermont Center for Ecostudies, P. O. Box 420, Norwich, VT 05055, USA. 2 Corresponding author; e-mail: crimmer@vinsweb.org island’s high elevation broadleaf and mixed pine-broadleaf forests have received little study because of the remoteness and difficulty of access of these forests. We documented previously undescribed nests of two species during investigations at a montane forest site in the Dominican Republic during 2002-2004. The Western Chat-tanager {Calyptophilus ter- tius) and Hispaniolan Highland-tanager {Xen- oligea montana) are among the most endan- gered species on Hispaniola^ understanding their breeding biology is fundamental to con- serving their populations. The Western Chat-tanager, recently classi- fied as a species distinct from the Eastern Chat-tanager {Calyptophilus frugivorus) (AOU 1998), occupies a disjunct and frag- mented range at elevations from 750 to 2,300 m in Haiti’s Massif de la Hotte and Massif de la Selle, and in the western Sierra de Baho- ruco of the Dominican Republic (Latta et al. 2006). The species inhabits dense understory of moist broadleaf forests where it is secretive and difficult to observe. The International Union for the Conservation of Nature (lUCN) has not recognized the recent split of C. tertius from C. frugivorus, but still considers the spe- cies complex as globally Vulnerable to ex- tinction because of its small range and overall population size (lUCN 2006). Data on abun- dance and population trends are lacking, but there is little doubt that declines have occurred because of habitat loss throughout Hispaniola (Keith et al. 2003, Latta et al. 2006). Two for- merly recognized subspecies are believed to have been extirpated (lUCN 2006). The Hispaniolan Highland-tanager, also known as the White-winged Warbler (AOU 1998), is restricted in Haiti to mesic montane forest' above 1,150 m elevation in the Massif de la Hotte of Haiti. This species was renamed 190 SHORT COMMUNICATIONS 191 by Latta et al. (2006) based on mtDNA anal- yses showing this speeies’ clear association with the tanagers of Hispaniola (Lovette and Bermingham 2002). It has apparently been ex- tirpated from the Massif de la Selle and other areas of former occurrence. The species’ Do- minican Republic distribution is limited to el- evations above 1,300 m in Sierra de Bahoru- co, parts of the Cordillera Central, and on up- per slopes of the southern Sierra de Neiba (Latta et al. 2006). The Hispaniolan Highland- tanager is classified as globally Vulnerable by the lUCN (2006) and is considered by some to be one of Hispaniola’s most endangered en- demic birds (Woods et al. 1992, Latta et al. 2006). No data exist on population size or trends, but local extirpations and ongoing se- vere deforestation suggest the species is in de- cline. The objectives of this paper are to describe: ( 1 ) the nest and eggs of Western Chat-tanager and Hispaniolan Highland-tanager, (2) the na- tal plumage of Western Chat-Tanager, and (3) the Juvenal plumage of Hispaniolan Highland- tanager. METHODS We conducted studies of the breeding bird community at a montane broadleaf forest site at Pueblo Viejo (18° 12' N, 71°32'W) at 1,775 m elevation in the Sierra de Bahoruco from May through July 2002-2004. General floristic features of this forest type have been described by Fisher-Meerow and Judd (1989) and Latta et al. (2003). Nest searching was conducted on a daily basis each year, primar- ily through systematic searching of suitable habitat. We fitted five Western Chat-tanagers with 1 .0-g radio transmitters (BD-2 model, Holohil Systems Ltd., Carp, ON, Canada) in 2002 and 2003 and tracked each individual daily to attempt to locate nests. Once located, nests of all species were monitored at 2-3-day intervals through termination (failure or fledg- ing); some nest visits occurred at 1-day inter- vals. We measured a standardized series of habitat and nest site characteristics following each nesting attempt. NEST DESCRIPTION Western C/uit-tcina^er. — CCR found a nest on 17 May 2002 by systematically searching an area in which a radio-marked female and male had been present during the previous several days. The nest was 1.3 m above ground in an extremely dense vine-shrub thicket in a tree-fall gap and could be accessed only by crawling on the ground. It was on top of the horizontal “roof” of the vine tangle, under the small overhanging branch of a 2.5-m tall broadleaf shrub that projected above the undergrowth. Estimated percent nest concealment from 1 m was 20% above, 90% below, 70% north, 80% east, 10% south, and 0% west. Fresh eggshell fragments, as well as the continued nearby presence of both female and male Chat-tanagers suggested the nest had been recently depredated and it was collected at this time. The nest was a bulky, coarsely-constructed, partially-domed structure with a west-facing side entrance that covered two-thirds of the nest cup. The exterior consisted of small woody stems 2-4.5 mm in diameter, robust herbaceous stems, peeled coarse exterior sheathing of herbaceous stems, vine tendrils, moss, and foliose lichens. About 40% of the outer layer was covered with the entire leaves of at least three broadleaf tree or shrub spe- cies. The inner cup was lined primarily with fine herbaceous stems and leaf fragments. The exterior diameter from back to the bottom rim of the entrance was 21.5 cm, from back to the entrance’s top rim 14.5 cm, and from side to side 23.5 cm. The nest cup measured 13.1 cm from front to back, 11.1 cm from side to side, and 5.1 cm from back to the top edge of the entrance. The overall nest height was 13.1 cm. A second nest of this species was discovered on 9 June 2004 by RKN and LGW when an adult flushed from close range. The nest was in a small clearing (~5 m diameter) within den.se, mature broadleaf forest, 3 m from an actively- u.sed foot trail. The nest was 1.1 m above ground on the eastern side of the lower trunk of a hardwood “ozua” (Pimenta oz.uci) tree. The nest tree was 13 m tall with its lower canopy 7.5 m above ground and a diameter at breast height of 19.7 cm. The nest was against the trunk and suppoiled beneath by a collection of vines with two main stems (4.0 and 3.0 mm diameter) interwoven with the top of the nest, two main stems on the bottom of the nest ( 10.0 and 5.0 mm diameter), and .several smaller stems 0.3— 1.0 mm in diameter touching the side of the nest. Estimated vegetation concealment 192 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 1, March 2008 LIG. 1. Western Chat-tanager nest with eggs, June 2004. Photograph by L. G. Woolaver. for the nest from a distance of 1 m was 0% overhead, 90% below, 95% north, 10% east, 20% south, and 85% west. This cover was pro- vided by the tree trunk and surrounding tangle of vine stems. This nest also was a large, bulky structure, roughly oval in shape (Fig. 1). The nest cup was partially covered by a roof, creating a south-facing entrance 8.7 cm high and 8.2 cm wide. The exterior was composed of moss and lichens interwoven with twigs and whole her- baceous stalks from 1 to 5.5 mm diameter, stripped sheathing of herbaceous stems, dried leaf fragments of broadleaf species, and frag- ments of dried bromeliad leaves. The inner cup was lined with fine twigs and herbaceous stems, dried leaf fragments, lichen, moss, and a few adult body feathers. The exterior nest dimensions were 28.5 cm from front to back, 27.0 cm from side to side, and 26.0 cm from top to bottom. The nest cup measured 9.1 cm from front to back, 7.7 cm from side to side. and 5.5 cm in depth. The nest cup height from the roof to the bottom was 14.3 cm. The nest contained two eggs when discov- ered on 9 June. The eggs had pale blue back- ground coloration and irregular light brown to dark brown speckling and mottling (Fig. 1). The eggs measured 27.8 X 18.7 and 28.9 X 18.8 mm and hatched on 19 and 20 June, re- spectively. No eggshells were found in the nest, suggesting the adults had removed them. The nestlings were dark pink with whitish bills and gapes, and covered with long, fine, black down. The nestlings were last confirmed alive at 1515 hrs EST on 25 June and both were missing from the nest at 0558 hrs on 26 June. The nest was damaged, suggesting dep- redation and it was collected on this date. Observations on adult behavior at this nest were recorded by EMF from a photographic blind during 5 hrs on 20 June, shortly after the eggs hatched. The smaller of the two adults was banded and presumed to be the female, based SHORT COMMUNICATIONS 193 FIG. 2. Hispaniolan Highland-tanager nest with egg, June 2004. Photograph by E. M. Fernandez. on brooding behavior. The nestlings were brooded for 10-16 min periods each hour dur- ing which the female faced the nest entrance. The male fed the female a small white grub during one brooding session and provided her with small arthropods on several occasions while she was foraging near the nest. Both adults fed the chicks, but the male’s feeding vis- its were infrequent and sporadic. The female fed the nestlings at 15-20-min intervals when not brooding. Each feeding bout consisted of 4-5 consecutive flights to the nest during which the female approached using a consistent pattern of perches. Her vocalizations changed from a short, two note '"chip-chip" while foraging to an even "tick, tick, tick, tick ...” as she ap- proached the nest. The female was twice ob- served removing fecal pellets from the nest, al- though it was unclear whether she ate the.se or carried them away. The male sang for 10-15 min bouts each hour from different perches, but within 10 m of the nest. Hispaniolan Highland-tanager. — Pablo Diaz found a nest of this species on 27 June 2004 containing one egg and one nestling that appeared close to fledging. The nest was in a dense vine tangle 2.5 m above ground. The vine thicket reached a height of 4.5 m and was in the understory of a closed-canopy broadleaf forest. Estimated percent nest cover from 1 m was 70% overhead, 70% north, 20% east, 60% .south, and 70% west. The nest was an open, cup-shaped structure (Fig. 2). The nest’s exterior was composed of moss, small herbaceous stems, broadleaf leaf fragments, lichens, and other plant libers. Its exterior cup was lined with fine, hairlike plant fibers and small numbers of herbaceous stems <1 mm in diameter. The exterior nest diam- eter was 10.5 X 7.9 cm, the nest cup diameter 194 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 was 5.6 X 5.2 cm, and the interior depth was 3.1 cm. The nest height was 5.5 cm. The sin- gle egg was oval in shape, pale greenish-white in background coloration, and faintly marked with reddish brown blotches and scrawls (Fig. 2). No measurements were obtained and the egg subsequently disappeared. The nestling fledged on 28 June. The juvenal plumage, previously unde- scribed, was documented through photographs taken by EMF. The head and nape were brownish-gray, whereas the remaining upper- parts were grayish tinged with olive-brown. The emerging flight feathers appeared dark grayish, the secondaries were edged greenish orange-brown proximally, and the greater co- verts were predominantly greenish orange- brown. The underparts were a smudgy off- white with some brownish tones. The bill was grayish-flesh, the legs pale grayish, and the eye dark brownish. DISCUSSION Confirmed descriptions of the nests of the Western Chat-tanager or Hispaniolan High- land-tanager have not been published prior to this paper. Bond (1943:122) reported a pos- sible Chat-tanager nest from the Massif de la Selle range in Haiti: “A single nest, contain- ing one addled egg (23.6 X 18.3 mm) . . . , probably pertained to this species. This nest (found on June 14) was situated in a fern about two feet above the ground, bordering a blackberry patch. There was a protesting pair of Chat Tanagers a few yards from the nest . . . The timing of this discovery and the egg dimensions are similar to our data, but more substantive information is lacking. Two descriptions of apparent Hispaniolan Highland-tanager nests from the early 1900s are less convincing. One was shown to Bond (1928) on 1 1 June 1928 from Mome La Selle in Haiti. Wetmore and Swales (1931:396) de- scribed the nest as “globular in shape, com- posed of moss and grasses, lined with grass stems and feathers, and placed in a bush five feet from the ground.” It contained “two fresh eggs, which are plain, creamy white in color without markings”. Bond (1928) reported the respective egg measurements as 21.6 X 15.5 and 21.7 X 15.5 mm. The second reputed nest was reported to Wetmore and Swales (1931: 397) by local residents at an unspecified lo- cation, and described as “oval with the en- trance from beneath”. It is unlikely that either of these two nests was that of a Hispaniolan Highland-tanager, based on our documenta- tion of the Pueblo Viejo nest. Our limited observations suggest the peak incubation period for Western Chat-tanagers occurs from mid-May to mid- June. Further evidence of this species’ nesting phenology is provided by two mist-netted females with ful- ly-developed incubation or brood patches. One was captured on 16 May 2002 and the second on 18 May 2003 (CCR, unpubl. data). Both encounters indicate that incubation was underway by mid-May. The Hispaniolan Highland-tanager fledging date of 27 June is later than the mean (± SD) fledging date of 1 June ± 25.5 days for the ecologically similar Green-tailed Ground-tanager (Microligea pal- ustris) at Pueblo Viejo (n = 14 nests; CCR and SCL, unpubl. data), but well within the range of expected variation. The depredation of both Western Chat-tan- ager nests highlights the generally high rates (-50%) of nest predation in montane broad- leaf forests of the Sierra de Bahoruco. Intro- duced predators, particularly feral cats {Felis domesticus), black (Rattus rattus) and Norway rats {R. norvegicus), appear to limit nest suc- cess of several endemic montane species, es- pecially those which forage and nest near the ground. The remains of the 2004 Western Chat-tanager nest do not allow us to identify the predator(s), as the nest had been pulled from below and the bottom ripped apart, lead- ing us to believe a mammal was responsible. High elevation broadleaf forests are consid- ered one of Hispaniola’s most endangered habitats (Latta and Lorenzo 2000). Ten of the 15 endemic bird species considered endan- gered or threatened on the island are concen- trated in montane forests (Latta et al. 2006). Understanding factors that limit populations of these species is crucial to implementing successful management and conservation practices. Our documentation of the first nests of Western Chat-tanager and Hispaniolan Highland-tanager contributes to an emerging base of information on the ecology of these two species, both of which are regarded at high risk of extinction. SHORT COMMUNICATIONS 195 ACKNOWLEDGMENTS We thank Jesus Almonte, Nicolas Corona, Pablo Diaz, J. D. Lambert, Danilo Mejia, Vinicio Mejia, Marisabel Paulino, Andrea Townsend, and Jason Townsend for assistance with very challenging field work. Funding was provided by the Bay Foundation, National Geographic Society, Scott Neotropical Fund of the Lincoln Park Zoo, Wildlife Preservation Cana- da, and friends of the Vermont Institute of Natural Sci- ence. Authorization to conduct our work was provided by the Subsecretaria de Areas Protegidas and Depar- tamento de Vida Silvestre, Dominican Republic. Con- structive reviews of the manuscript were provided by W. J. Arendt and J. W. Wiley. LITERATURE CITED American Ornithologists’ Union (AOU). 1998. Checklist of North American birds. Seventh Edi- tion. American Ornithologists’ Union. Washing- ton, D.C., USA. Bond, J. 1928. The distribution and habits of the birds of the Republic of Haiti. Proceedings of the Acad- emy of Natural Sciences of Philadelphia 80:483- 521. Bond, J. 1943. Nidification of the passerine birds of Hispaniola. Wilson Bulletin 55:115-125. Fisher-Meerow, L. L. and W. S. Judd. 1989. A flo- ristic study of five sites along an elevational tran- sect in the Sierra de Bahoruco, Prov. Pedernales, Dominican Republic. Moscosoa 5:159-185. International Union for the Conservation of Na- ture (lUCN). 2006. 2006 lUCN Red List of threatened species, (www.iucnredlist.org). Keith, A. R., J. W. Wiley, S. C. Latta, and J. A. Ottenw ALDER. 2003. The birds of Hispaniola — Haiti and the Dominican Republic. BOU Check- list 21. British Ornithologists’ Union, Tring, Unit- ed Kingdom. Latta, S. C. and R. Lorenzo (Editors). 2000. Results of the national planning workshop for avian con- servation in the Dominican Republic. Direccion Nacional de Parques, Santo Domingo, Dominican Republic. Latta, S. C., C. C. Rimmer, and K. P. McFarland. 2003. Winter bird communities in four habitats along an elevational gradient on Hispaniola. Con- dor 105:179-197. Latta, S., C. Rimmer, A. Keith, J. Wiley, H. Raf- FAELE, K. McFarland, and E. Eernandez. 2006. Birds of the Dominican Republic and Haiti. Princeton University Press, Princeton, New Jer- sey, USA. Lovette, I. J. AND E. Bermingham. 2002. What is a wood-warbler? Molecular characterization of a monophyletic Parulidae. Auk 119:695-714. Wetmore, a. and B. H. Swales. 1931. Birds of Haiti and the Dominican Republic. U.S. National Mu- seum Bulletin 155. Woods, C. A., E E. Sergile, and J. A. Ottenwalder. 1992. Stewardship plan for the national parks and natural areas of Haiti. Elorida Museum of Natural History, Gainesville, USA. The Wilson Journal of Ornithology 120(1): 195— 199, 2008 Foraging and Nesting of the ‘Akikiki or Kaua‘i Creeper {Oreomystis hairdi) Eric A. VanderWerf and Pauline K. Roberts^ ABSTRACT — The ‘Akikiki or Kaua‘i Creeper {Or- eomystis hairdi) is a rare, little-known Hawaiian hon- eycreeper endemic to the island of Kaua‘i. Its range is contracting, the population is declining, and it is a can- ' U.S. Fish and Wildlife Service, 300 Ala Moana Boulevard, Room 3-122, Box 50088, Honolulu, HI 96850, USA. ^ Kaua’i Forest Bird Recovery Projecl, Hawai’i Di- vision of Forestry and Wildlife, P. O. Box 458, Wai- mea, HI 96796, USA. ^ Current address: Pacific Rim Conservation, 3038 Oahu Avenue, Honolulu, HI 96822, USA. Corresponding author; e-mail: ewerf@hawaii.rr.com didate for listing under the U.S. Endangered Species Act. We report an instance of foraging by excavation ob.served on 22 May 2006, a behavior previously un- known in this species, and on parental behavior at two nests ob.served on 24 May 2006 and 27 May 2007, about which there is little previous information. Both parents brought food to the nest, the male provided food for the female, and the female also foraged in- dependently. The nesting pair in 2007 had a juvenile from a previous nest, indicating the ‘Akikiki will at- tempt to rai.se two broods. These observations are of limited extent, but even small facts can contribute to our understanding of the biology of the ‘Akikiki and causes of its decline. Received I Tehruary 2007. Ac- cepted 16 June 2007. 196 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 The ‘Akikiki or Kaua‘i Creeper (Oreomys- tis bairdi) is a small forest bird in the Ha- waiian honeycreeper subfamily (Drepanidi- nae; Fringillidae) endemic to the island of Kaua‘i. It was once found throughout Kaua‘i and was locally abundant in the Alaka‘i and Koke‘e areas into the early 1960s (Richardson and Bowles 1964). It has disappeared from much of the island and is now restricted to high elevation native forests in remote parts of the Alaka‘i Plateau (USFWS 1984, Scott et al. 1986, Foster et al. 2004). Forest bird sur- veys have shown a decrease in distribution from 88 to 36 km^ between 1970 and 2000 and a decline in estimated population from 6,832 ± 966 to 1,472 ± 680 birds over this period (Foster et al. 2004). No field research has focused on the ‘Aki- kiki and we assume the causes of this decline are those typical of other Hawaiian forest birds: diseases carried by non-native mosqui- toes, degradation of native forest habitat by invasive non-native plants, browsing and root- ing by feral ungulates, and nest predation by non-native rodents (Foster et al. 2000, Scott et al. 2001, USFWS 2006). The threat from mosquito-bome diseases may worsen as glob- al warming allows mosquitoes to invade the highest, coldest parts of the island that once provided a refuge from disease (Benning et al. 2002). The ‘Akikiki has been a candidate for listing under the U.S. Endangered Species Act since 1994 due to its small and declining pop- ulation, restricted distribution, and a variety of threats (USFWS 2004). It is considered criti- cally endangered by the International Union for the Conservation of Nature (lUCN 2006). The ‘Akikiki is one the least-known of all extant Hawaiian bird species. Only eight ‘Aki- kiki nests are known and there is no infor- mation about nest success, parental care of nestlings, reproductive rates, survival of adults or juveniles, or movements (Eddinger 1972, Foster et al. 2000, USFWS 2006). The objective of this paper is to report ob- servations on foraging and nesting of the ‘Akikiki from trips to the Alaka‘i Wilderness Preserve from 22 to 24 May 2006 and 24 to 28 May 2007. The primary purpose of these trips was to conduct forest bird surveys and research on the critically-endangered Puaiohi or small Kaua‘i Thrush {Myadestes palmeri). These observations on the ‘Akikiki are of lim- ited extent, but even small facts can contribute to our understanding of the biology of this species and causes of its decline. OBSERVATIONS On 22 May 2006, EAV observed a pair of ‘Akikiki foraging in the forest canopy on a low ridge in the upper Halepa‘akai drainage. The birds visited several trees and spent ~ 1 min for- aging on branches and twigs of a small, partly dead ‘dlapa (Cheirodendron trigynum) tree. They appeared to be gleaning and probing for insects on the bark, but photographs of one of the birds revealed that it had also been exca- vating rotten wood from the center of a twig, presumably for insect larvae (Fig. 1). The fresh- ly excavated area where the bird had been work- ing could be identified by the pale color of the newly exposed wood. Two small pieces of wood were visible on the bird’s bill. There also were areas of darker, more weathered wood on the same twig that appeared to be older exca- vations. We discovered an ‘Akikiki nest on 24 May 2006, -12.5 m above ground level in a 14 m tall ‘ohi‘a {Metrosideros polymorpha) tree on a bank of Halepa‘akai Stream. Our attention was drawn to the tree by songs and calls from the male and by the unusual amount of time spent in the tree by the pair. The nest was positioned in a fork where several epiphyte- encrusted branches -1—3 cm in diameter di- verged from a larger branch (Fig. 2). The composition of the nest could not be examined closely due to its height, but it appeared to be composed of moss, small pieces of bark, bits of lichen, and fine plant fibers (Fig. 2). We observed the nest from 1045 to 1215 hrs HST and made notes about all activity. One individual, assumed to be the female based on its behavior, sat on the nest contin- uously and left the nest once for 20 min. The head of one nestling was visible just over the rim of the nest (Fig. 2). A second adult, as- sumed to be the male, fed the female twice on the nest and the female passed food to the nestling each time. The male also fed the fe- male once -10 m from the nest, after which she left the area, presumably to forage. The female fed the nestling upon her return. The male sang near the nest only once during the timed observation period, but sang more fre- SHORT COMMUNICATIONS 197 FIG. 1. ‘Akikiki foraging by excavation on a dead twig of a ‘blapa tree in the upper Halepa‘akai drainage, Kaua‘i, 22 May 2006. Note the pale color of freshly excavated wood between the birds’s feet and the darker color of older, weathered excavations higher on the twig. Photograph by E. A. VanderWerf. quently earlier in the day when we were oc- cupied with other activities. The nest tree was one of the larger trees in the area and was prominent in the forest can- opy. It had many flowers and was visited often by other birds, including ‘Apapane (Hima- tione sanguinea), Japanese White-eyes (Zos- terops japonicus), and ‘Anianiau (Hemigna- thiis parvus). The female ‘Akikiki tolerated the presence of these birds and did not react to them, even when they approached to within —50 cm of the nest. No ‘Akikiki other than the nesting pair were seen or heard in the vi- cinity of the nest. On 27 May 2007, EAV found an ‘Akikiki nest in construction on Mohihi Ridge. The nest- ing pair was accompanied by a juvenile, indi- cating they were attempting to renest. The nest was —8 m above ground level in a 8.5 m tall ‘ohi‘a tree. The exterior was compo.sed largely of moss and appeared to be complete. One bird, assumed to be the female, brought fine plant fibers to the nest three times from 1410 to 1510 hrs HST and worked on the nest lining. A .sec- ond bird, assumed to be the male, followed the female closely on two of the visits and called near the nest. The male fed the female once —20 m from the nest. During the feeding the female fluttered its wings and gave high-pitched twit- tering calls. The juvenile had pale spectacles typical of immature ‘Akikiki, and had fluffy plumage and foraged clumsily, indicating it hatched earlier in the same season. DISCUSSION The nests we observed were similar to other reported ‘Akikiki nests, but the nest found in 2006 was slightly higher above ground level and in a taller tree. Two ‘Akikiki nests re- ported by Eddinger (1972) were on terminal branches 8.1 and 8.5 m above ground level in the crowns of non-blooming ‘ohi‘a trees, two nests observed during the 2000 Kaua‘i forest bird survey were also in ‘6hi‘a trees, a nest observed in 1999 was 9 m above ground level in a 9.6 m ‘6hi‘a tree (David Kuhn, cited in Foster et al. 2000), two other nests were 4 and 6 m above ground level in ‘ohi‘a (Jim Denny, 198 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. I, March 2008 LIG. 2. ‘Akikiki nest in the crown of a ‘ohi‘a tree in the upper Halepa‘akai drainage, Kaua i, 24 May 2006. The female's head is visible above the nest rim. and downy feathers on the head of a nestling are visible m front of the female. Photograph by E. A. VanderWerf. cited in Foster et al. 2000), and a nest found in 2005 was 8 or 9 m above ground level in a ‘ohi‘a (Brett Hartl, pers. comm.). The nests reported by Eddinger (1972) were composed primarily of moss. The nest observed in 1999 by David Kuhn had an exterior of moss and a rim and lining of strips of ‘ohi‘a bark, plant rootlets, and other fine plant fibers. At least one of the nests observed by Jim Denny in- cluded ‘olapa bark (Foster et al. 2000). The timing of the nesting attempt we observed in 2006 was within the March-June breeding season described by Foster et al. (2000). The nest observed in 2005 by Brett Hartl was un- der construction on 29 January, which is ear- lier than all other known nesting activity. The nest we observed in 2007 would fledge in July, if successful. There is limited information about the breeding biology of the ‘Akikiki. Nest con- struction is reported to be by the male and female (Eddinger 1972), and the male is re- ported to feed the female during nest construc- tion (Foster et al. 2000). Incubation has been observed by the female only, and Eddinger (1972) commented on the tenacity of the fe- male to remain on the nest during incubation despite disturbance. One nest found by Eddin- ger (1972) contained a single egg and a sec- ond nest contained two nestlings. Our obser- vations provide new information. Both male and female ‘Akikiki brought food to the nest, although only the female was observed to broo and feed the nestling. The nestling was fed three times during 1 .5 hrs of observation, twice with food provided by the male and once with food provided by the female. The male provided food for the female during brooding, and it is likely that such provision- ing occurs throughout the nesting cycle. The female also foraged independently and the rel- ative importance of male provisioning remains unknown. These behaviors are common to other Hawaiian honeycreepers (Lepson and Freed 1997. Lepson and Woodworth 2001, Pratt 2005). The nestling provisioning rate (2.0/hr) and relative contribution of each gen- der (67% by the male) were similar to those SHORT COMMUNICATIONS 199 in the Hawaii Creeper (3.7 provisionings/hr, 57% by the male) (VanderWerf 1998). The pair observed in 2007 was at least attempting to raise two broods in a season, which is un- usual among insectivorous Hawaiian honey- creepers. The nesting season is longer than previously reported with activity beginning in January in some years and renesting attempts possibly continuing into July. ‘Akikiki are reported to forage on trunks, branches, and twigs of live and dead trees, primarily ‘ohi‘a and koa {Acacia koa), and oc- casionally in subcanopy shrubs (Foster et al. 2000). They feed on insects, insect larvae, and other arthropods that are taken by gleaning and probing from bark, crevices, and epi- phytes (Foster et al. 2000). There are no pre- vious reports of the excavating behavior that we observed. No other bird on Kaua‘i is known to excavate in this manner and it is improbable the ‘Akikiki was taking advantage of an excavation begun by another species. The medium-length bill of the ‘Akikiki is gen- erally believed to be adapted for gleaning and probing, but our observations indicate it can also be used for excavation. The presence of similar but older and weathered excavation patterns on the same twig suggests it had been excavated previously, perhaps by the same in- dividual ‘Akikiki. Excavation may be a more regular foraging behavior in the ‘Akikiki than is currently realized. ACKNOWLEDGMENTS We thank T R. Savre and D. L. Leonard for cama- raderie in the field, Brett Hartl for information about the nest observed in 2005, D. L. Leonard, C. E. Braun, and two anonymous reviewers for comments that helped improve the manuscript, and the Hawai‘i Di- vision of Forestry and Wildlife for continuing support of the Kaua‘i Forest Bird Recovery Project. LITERATURE CITED Benning, T L., D. LaPointr, C. T Atkinson, and P M. VrrousEK. 2002. Interactions of climate change with biological invasions and land u.se in the Hawaiian Islands: modeling the fate of endem- ic birds using a geographic information system. Proceedings of the National Academy of Science 99:14246-14249. Eddinger, C. R. 1972. Discovery of the nest of the Kauai Creeper. Auk 89:673-674. Foster, J. T, J. M. Scott, and P. W. Sykes Jr. 2000. ’Akikiki {Oreomystis bairdi). The birds of North America. Number 552. Foster, J. T, E. J. Tweed, R. J. Camp, B. L. Wood- worth, C. D. Adler, and T. Telfer. 2004. Long- term population changes of native and introduced birds in the Alaka‘i Swamp, Kaua‘i. Conservation Biology 18:716-725. International Union for the Conservation of Na- ture (lUCN). 2006. 2006 lUCN Red list of threat- ened species, (www.iucnredlist.org) (accessed 17 January 2007). Lepson, j. K. and L. a. Freed. 1997. ‘Akepa {Loxops coccineus). The birds of North America. Number 294. Lepson, J. K. and B. L. Woodworth. 2001. Hawai‘i Creeper {Oreomystis mana). The birds of North America. Number 680. Pratt, H. D. 2005. The Hawaiian honey creepers. Ox- ford University Press, Oxford, United Kingdom. Richardson, F. and J. Bowles. 1964. A survey of the birds of Kauai, Hawaii. B. P. Bishop Museum Bul- letin 227. Scott, J. M., S. Conant, and C. van Riper III. 2001. Evolution, ecology, conservation, and manage- ment of Hawaiian birds: a vanishing avifauna. Studies in Avian Biology 22:1-428. Scott, J. M., S. Mountainspring, F. L. Ramsey, and C. B. Kepler. 1986. Forest bird communities of the Hawaiian Islands: their dynamics, ecology, and conservation. Studies in Avian Biology 9:1- 431. U.S. Fish and Wildlife Service (USFWS). 1984. Kauai forest bird recovery plan. USDI, Fi.sh and Wildlife Service, Region 1, Portland, Oregon, USA. U.S. Fish and Wildlife Service (USFWS). 2004. Re- view of native species that are candidates or pro- posed for listing as endangered or threatened; an- nual notice of findings on resubmitted petitions; annual description of progress on listing actions; notice of review. Federal Register 7L53756- 53835. U.S. Fish and Wildlife Service (USFWS). 2006. Re- vised recovery plan for Hawaiian forest birds. USDI, Fish and Wildlife Service, Region 1, Por{- land, Oregon, USA. Vandi-rWerp, F. a. 1998. Breeding biology and ter- ritoriality of the Hawai'i Creeper. Condor 100: 541-545. 200 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 The Wilson Journal of Ornithology 120( 1):200-201, 2008 First Observation of Duetting in the Olive-backed Euphonia (Euphonia gouldi) Thor Hanson' ABSTRACT — I observed a pair of Olive-backed Euphonias {Euphonia gouldi) in the Caribbean low- lands of Costa Rica singing a prolonged antiphonal duet on 28 February 2006. The male and female were perched near one another within a larger conspecific flock, and exchanged closely-coordinated notes in phrases lasting ~5 sec. This appears to be the first report of duetting behavior for this species or for any member of family Fringillidae. Received 24 January 2007. Accepted 28 April 2007. Duetting behavior in birds has been report- ed for more than 200 species (Farabaugh 1982) and is the subject of increasing research attention (Langmore 1998, Hall 2004). Func- tions of duetting remain unclear, but likely in- clude joint territorial defense, mate guarding, and the formation or strengthening of pair bonds (Morton 1996, Langmore 1998, Hall 2004). Duets vary widely in complexity among species (Farabaugh 1982) and may serve different or even multiple purposes in different settings (Marshall-Ball et al. 2006). Duetting is relatively uncommon in northern temperate regions, but occurs more regularly in the tropics where year-round territoriality and long-term pair bonds are common and may encourage its development (Morton 1996, Slater and Mann 2004). Documenting the extent of this behavior in the tropics is far from complete and remains a research priority (Hall 2004). This paper describes the first re- cord of duetting in the Olive-backed Euphonia {Euphonia gouldi) and discusses the obser- vation in the context of related literature. OBSERVATIONS I observed a pair of Olive-backed Euphon- ias singing a complex, antiphonal duet on 28 February 2006 in Canton Sarapfqui, Costa ' Department of Forest Resources, University of Ida- ho, Moscow, ID 83844, USA; e-mail; thor@rockisland.com Rica. These observations occurred in the un- derstory of unlogged primary lowland rain- forest at the La Selva Biological Station (McDade et al. 1994). I watched a male and female perched —10 cm from one another on the same branch within a conspecific flock of at least eight individuals. The species is di- chromatic and the male and female roles in the duet were identifiable from sex-specific plumage patterns. The female interspersed staccato chips and single-notes seamlessly be- tween the more varied whistled tones of the male. Completed phrases lasted 4-5 sec and were repeated with variations after a brief pause for >2 min. The rest of the flock was dispersed in nearby undergrowth vocalizing with a range of simple whistle tones and calls. No other duets were heard or observed in > 10 min of observation. DISCUSSION The song of the Olive-backed Euphonia has been described as “a rapid, jerky, rambling melody of rolling chrrs, staccato notes, short, clear whistles and more nasal, slurred whis- tles” (Stiles and Skutch 1989:421). In my ob- servations, the ‘staccato notes’ and perhaps what Stiles and Skutch describe as ‘short, clear whistles’ were supplied by the female. The resulting song was continuous and I would have ascribed it to the male alone if both individuals had not been in clear view at close range. Euphonias often frequent the can- opy or dense edges and secondary vegetation (Stiles and Skutch 1989), and can be difficult to view clearly. It remains to be learned whether duets are common in this species and what function they might serve. My obser- vations occurred during the beginning of the breeding season (Stiles and Skutch 1989), a time when duets may be involved in regula- tion of reproductive synchrony (Hall 2004), or the establishment or reaffirmation of pair bonds (Farabaugh 1982). SHORT COMMUNICATIONS 201 I am unaware of reports of duetting from any other member of the subfamily Euphoni- inae or the family Fringillidae as eurrently de- fined (American Ornithologists’ Union 1998). Distinctive and prolonged vocalizations are common in this group, however, and the closely-related Thick-billed Euphonia {E. lan- iirostris) and Violaceous Euphonia {E. viola- cea) are well-documented mimics (Snow 1974, Morton 1976, Remsen 1976). Stiles and Skutch (1989: 420) describe vocalizations for the female Yellow-throated Euphonia {E. hi- rundinacea) as “thin, dry notes, sometimes high-pitched and almost trilled,” but they do not report duetting. Future efforts should ex- plore the frequency and purpose of duets in the Olive-backed Euphonia, and observers should be on the alert for duetting behavior in its congeners. ACKNOWLEDGMENTS This work was supported in part by NSF-IGERT grant 01 14304. The author thanks two anonymous re- viewers for helpful comments on the manuscript. LITERATURE CITED American Ornithologists’ Union. 1998. Checklist of North American birds. Seventh Edition. American Ornithologists’ Union, Washington, D.C., USA. Earabaugh, S. M. 1982. The ecological and social sig- nificance of duetting. Pages 85-124 in Acoustic communication in birds (D. E. Kroodsma and E. H. Miller, Editors). Academic Press, New York, USA. Hall, M. L. 2004. A review of hypotheses for the functions of avian duetting. Behavioral Ecology and Sociobiology 55:415-430. Langmore, N. E. 1998. Functions of duet and solo songs of female birds. Trends in Ecology and Evo- lution 13:136-140. Marshall-Ball, L., N. Mann, and J. B. Slater. 2006. Multiple functions to duet singing: hidden conflicts and apparent cooperation. Animal Be- havior 71:823-831. McDade, L. a., K. S. Bawa, H. A. Hespenheide, and G. S. Hartshorn (Editors). 1994. La Selva: ecol- ogy and natural history of a neotropical rain for- est. University of Chicago Press, Chicago, Illinois, USA. Morton, E. S. 1976. Vocal mimicry in the Thick- billed Euphonia. Wilson Bulletin 88:485-487. Morton, E. S. 1996. A comparison of vocal behavior among tropical and temperate passerine birds. Pages 258-268 in Ecology and evolution of acoustic communication in birds (D. E. Kroodsma and E. H. Miller, Editors). Cornell University Press, Ithaca, New York, USA. Remsen, J. V. 1976. Observations of vocal mimicry in the Thick-Billed Euphonia. Wilson Bulletin 88: 487-488. Slater, J. B. and N. I. Mann. 2004. Why do the fe- males of many bird species sing in the tropics? Journal of Avian Biology 35:289-294. Snow, B. K. 1974. Vocal mimicry in the Violaceous Euphonia. Wilson Bulletin 86:179. Stiles, E G. and A. F. Skutch. 1989. A guide to the birds of Costa Rica. Cornell University Press, Ith- aca, New York, USA. The Wilson Journal of Ornithology 1 20( 1 ):20 1-204, 2008 The Display of a Reddish Hermit (Phaethornis ruber) in a Lowland Rainforest, Bolivia Adam Felton, ' Annika M. Felton,'-^ and David B. Lindenmayer' ABSTRACT — The Reddish Hermit (Phaethornis ruher) is a commonly encountered hummingbird with- in the understory of tropical forests from Venezuela to ' Centre for Resource and Environmental Studies, WK Hancock Building, Australian National University, Canberra, Australia 0200. ^ Institute Boliviano de Investigacion Forestal (IBIF), Avenue 2 de Agosto esq. Cuarto Anillo Casilla Postal 6204, Santa Cruz de la Sierra, Bolivia. ’Corresponding author; e-mail: adamf@cres.anu.edu.au .south-east Brazil. Reddish Hermits, like many hum- mingbirds, perform elaborate displays associated with mate and territory acquisition. We provide the first de- tailed de.scription with supporting illustrations of intri- cate displays between two male P. ruher to which we refer as “Rotation" and “Arc” displays. Received 2H December 2006. Accepted 17 July 2007. The Reddi.sh Hermit (Phacthorni.s ruher) is a common inhabitant of tropical forest under- 202 THE WILSON JOURNAL OF ORNITHOLOGY • Vol 120, No. I, March 2008 story from Venezuela to south-east Brazil and is one of the smallest (~2.4 g) known hum- mingbird species (Oniki 1996). Like many hummingbirds. Reddish Hermits form leks where several males display to females which visit these assemblages to choose a mate (Snow 1973). The displays of Reddish Her- mits at leks and in captivity often involve vi- sually spectacular and elaborate aerial maneu- vers (Mobbs 1971). Davis (1934:732) observed the display of a Reddish Hermit in October 1931 and sug- gested that it “must be seen for its beauty to be appreciated”. Over the ensuing decades, several researchers have provided written ac- counts of Reddish Hermit displays in the wild (Snow 1973) and in captivity (Mobbs 1971). One of these accounts (Snow 1973:171) de- scribes a display performed by a visiting male in front of a temtorial male within a lek. The visiting male is described as hovering “8-10 cm above and in front of the owning male” and, as it hovered, “the bird’s rear swayed side to side by about an inch while the head remained stationary ”. Our objective in this paper is to provide the first detailed descrip- tion, with supporting illustrations, of what ap- pears to be a similar but more extensive dis- play, which we suggest was by a male intrud- ing into the territory of a neighboring male. OBSERVATIONS One of us (AF) was conducting bird sur- veys on 23 August 2004 within the lowland subtropical humid forest (Holdridge Life Zone System) of the Guarayos Forest Reserve, De- partmento Santa Cruz, Bolivia. This conces- sion is -300 km north of the lowland city of Santa Cruz. The forest varies in altitude from 230 to 390 m with an average elevation of 320 m. The mean annual temperature is 25° C with mean annual precipitation of —156 cm. The region experiences a distinct dry season from May to October. One survey point was in a 471-m^ logging gap (15° 40' 07 S, 62° 45' 77 W) created by felling a 60-cm dbh tower tree {Schizolobium parahyha; Caesalpiniaceae). The following observations and sound re- cording were made in the understory of ad- jacent pioneer vegetation of two Reddish Her- mits at 0620 hrs. The temperature was 14° C with 100% cloud cover and no wind. No con- specifics were noted within the vicinity during FIG. 1. Illustration of the rotation display of the Reddish Hermit with the perched individual to the left and the displaying individual to the right. these displays, but the observed situation is not inconsistent with the circumstances ex- pected to be found within a lek (Snow 1973). We observed an individual P. ruber perched on a branch - 1 m above ground. The perching individual would flick its tail up and down, and occasionally fly to a different perch briefly during the displays of the other indi- vidual before returning to the original perch (Fig. 1). The displaying conspecific engaged in an intricate flying display —25 cm to the front and above the perched individual. Rotation Display.— We refer to the first dis- play as the “rotation display” (Fig. 1). The displaying individual conspicuously erected white plumage near its flanks, presumably of its thigh feathers. With its back arched and head raised, the bird increased its stroke am- plitude to what we perceived was the point of contact (Altshuler and Dudley 2003), due to an accompanying continuous rapid droning sound, which was distinct from that produced during normal flight. We were under the im- pression that the wings collided at both the ends of the upstroke and the downstroke. The tail was simultaneously raised and lowered with a speed that blurred its image to the hu- man eye. Concurrently the bird yawed through 300° of movement. This was one of the most mesmerizing aspects of the display due to the SHORT COMMUNICATIONS 203 FIG. 2. Illustration of the arc display of the Red- dish Hermit with the perched individual to the left and displaying individual to the right. individual’s capacity to avoid any correspond- ing rolling, pitching, or translational move- ment. This was done in a punctuated move- ment (at the start and end point of each arc), taking ~ 1 sec to complete an arc. At the same time, it rhythmically opened and closed its bill, thereby displaying the vibrant yellow col- oration of the gape. All of these movements were characterized by extreme speed and pre- cision and continued in repeated bouts that ranged from 3 to over 75 sec (mean = 15.4 sec; n = 5). The rotation display was accompanied by a high pitch repetitious song that was made by either the displaying individual or the perched individual. Like Mobbs (1971), we do not feel confident categorically assigning the song to one individual or the other. It consisted of high pitched ascending and descending indi- vidual notes and trills (zee’zee’zee’zeezezeze ze’zee), similar to that described by Nicholson (1931) as heard from a P. ruber lek in north- ern Guyana. Arc Display. — The individual at times switched to alternate, but equally transfixing behavior between rotation displays, which we refer to as the “arc display” (Fig. 2). The dis- playing individual during this phase remained facing the perched individual and moved lat- erally over a —100° arc, as if tethered at the chest to its perching audience. Tail move- ments may have been related more to con- trolling flight, than to display. We estimate the amplitude (horizontal distance moved) of this display was —30 cm. The displaying individ- ual could complete 3 arcs within 1 sec with each arc punctuated by a “diu” call (Fig. 3). This call was consistently between —1,000 and 2,000 kHz, inconsistent with the frequen- cy of their normal “tsi” call. The arc display was flanked both before and after by the ro- tation display, and was the shorter of the two displays. The arc display ranged from 1.0 to just over 3.3 sec in length (mean = 2.0 sec, n = 8). The entire display ended suddenly with both individuals moving from the field ' h 1 i [ 1 1 t . 0.000 kHz FIG. 3. Spectrogram of “diu” .sounds made by Reddish Hermit while performing the arc display. 204 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 1, March 2008 of view. We are uncertain whether one chased the other away. DISCUSSION We have several reasons for suggesting that both individuals were males and this observa- tion was an antagonistic display similar to that described by Snow (1973). Among the hum- mingbirds, the hermits (Phaethominae) are no- table for their lack of sexually dimorphic char- acteristics (Hoglund 1989). However, it is pos- sible to use plumage characteristics for distin- guishing male P. ruber from females (Hoglund 1989). Both male and female P. ruber have a rufous breast, but the male’s breast is marked by a black ‘V’ with the female lacking this mark altogether or possessing an obscure black blem- ish in its place (Davis 1958, Oniki 1970, Hog- lund 1989). Both the displaying and the perched individuals in our observations had a black ‘V’. We are confident that both individuals were males and the observed behavior was not a courtship display. Notably, it was the individual with the stron- ger male coloration that was perched with the displaying individual’s black chest bar fainter. This may indicate the displaying individual was a less mature male. This would be concordant with the description by Snow (1973) of a male attempting to displace a more established neigh- boring territorial owner. The other indication this display represents antagonistic behavior between two males over what appears to be a lek territory is the season in which the observation occurred. Breeding records for P. ruber in this region of Bolivia are absent, but records from Brazil (Oniki 1970), Guyana (Davis 1934, 1958), and Trin- idad (Snow 1973) consistently show the breeding season overlaps with the local dry season. The observed display occurred during the middle of the dry season in this region of Bolivia and we are confident this was a breed- ing-related display. We believe the most parsimonious expla- nation for the observed behavior is that it con- sisted of an encounter between a resident (perching individual) and intruding (display- ing) male within what was likely a lek terri- tory. We suggest that more systematic obser- vations are needed before we can gain a thor- ough understanding of these intricate displays. ACKNOWLEDGMENTS This project was supported through the generous fi- nancial assistance of the American Ornithologists’ Union, Sennheiser Australia, and the Lincoln Park Zoo, Chicago. We thank the personnel of IBIL, Agroin- dustria Lorestal La Chonta Ltd., Proyecto de Manejo Lorestal Sostenible (BOLLOR), and Birdlife Bolivia (Armonia) for providing logistical support, especially Marielos Pena and Todd Lredericksen. Eugenio Mer- cado provided integral assistance during all phases of the field work and the advice of Ross Cunningham greatly contributed to the design of this project. We thank Dan Brooks and an anonymous reviewer for helping us to improve an earlier version of this man- uscript. We thank the Australian Red Cross without whom this project could not have been completed. All research was approved by the relevant authorities and this study was conducted within the ethical guidelines of Australia and Bolivia. LITERATURE CITED Altshuler, D. L. and R. Dudley. 2003. Kinematics of hovering hummingbird flight along simulated and natural elevational gradients. Journal of Ex- perimental Biology 206:3139-3147. Davis, T. A. W. 1934. Notes on the display in the humming-birds Phaethornis siiperciliosus (Linn.) and Pygmonris ruber (Linn.). Ibis 13:732-738. Davis, T. A. W. 1958. The displays and nests of three forest hummingbirds in British Guiana. Ibis 100. 31-39. Hoglund, J. 1989. Size and plumage dimorphism m lek-breeding birds: a comparative analysis. Amer- ican Naturalist 134:72-87. Mobbs, a. J. 1971. Notes on the Reddish Hermit hum- mingbird. Aviculturalist Magazine 77:160-163. Nicholson, E. M. 1931. Communal display in hum- ming-birds. Ibis 13:74—83. Oniki, Y. 1970. Nesting behavior of Reddish Hermits {Phaethornis ruber) and occurrence of wasp cells in nests. Auk 87:720—728. Oniki, Y. 1996. Band sizes of southeastern Brazilian hummingbirds. Journal of Eield Ornithology 67. 397-391. Snow, B. K. 1973. The behavior and ecology of hermit hummingbirds in the Kanaku Mountains, Guyana. Wilson Bulletin 85:163-177. SHORT COMMUNICATIONS 205 The Wilson Journal of Ornithology 120(1):205— 209, 2008 Home Range and Habitat Preferences of the Banded Ground-cuckoo {Neomorphus radiolosus) Jordan Karubian^’^ and Luis Carrasco^ ABSTRACT — The Banded Ground-cuckoo {Neo- morphus radiolosus) is a rare, endangered, and poorly known species endemic to the Choco Biogeographic Zone. We summarize 7 months of data from radio track- ing an adult in northwestern Ecuador. Home range esti- mates were 42.2 ha (minimum convex polygon) and 49.9 ha (95% kernel analysis); the core area was 3.4 ha (50% kernel analysis). The bird favored undisturbed habitat and avoided secondary forest. It was primarily insectivorous and rarely associated with army ants {Eciton sp.) and not with mammals. Breeding occurred from March through June and the marked bird was seen with an unmarked individual throughout the study. The Banded Ground- cuckoo has a large home range, limited dispersal ability, and apparently depends on undisturbed forest. Defores- tation and habitat fragmentation appear to be the gravest threats facing the species. Received 15 December 2006. Accepted 28 April 2007. Ground-cuckoos {Neomorphus: Cuculidae) are terrestrial, rain-forest-dwelling birds dis- tributed across the Neotropics (Haffer 1977, Payne 1997). They forage on insects, small vertebrates, and fruit often while associating with army ants {Eciton sp.), Tayassu pecca- ries, or Saimiri, Cebus, and Saguinas primates (Sick 1949, Willis and Oniki 1978, Willis 1982, Terborgh 1983, Siegal et al. 1989). Bi- parental care has been reported for one species (Karubian et al. 2007), and none is thought to be a brood parasite (Sick 1949, Haffer 1977, Roth 1981, Payne 1997). Little else is known about any of the four species of Neomorphus, which replace each other geographically. The Banded Ground-cuckoo {Neomorphus radiolosus) is endemic to the Choco rain for- ests of northwestern Ecuador and western Co- lombia. It is one of the rarest birds in either ' Center for Tropical Re.search, In.stitute of the En- vironment, University of California at Los Angeles, La Kretz Hall, Suite 300, Box 951496, Los Angeles, CA 90095, USA. Mavier Zambrano n 16-45 y Rid de Janeiro, Quito, Ecuador. ^Corresponding author; e-mail; jordank@ucla.edu country; in recent years it has been reported from only two localities in Ecuador (Hom- buckle 1997, Lopez-Lanus et al. 1999, Ridge- ly and Greenfield 2001) and three localities in Colombia (Payne 1997). It is considered “vul- nerable to extinction” because of widespread deforestation throughout its range, but a lack of data on habitat requirements, home range, and other basic ecological parameters has hin- dered assessment of its conservation require- ments and status (BirdLife International 2000, Greenfield 2002, Renjifo et al. 2002). We re- port observations on the home range and hab- itat use of the Banded Ground-cuckoo based on a 7-month radio-tracking study of an adult in northwestern Ecuador. METHODS Our study was conducted at Bilsa Biologi- cal Station (79°45'W, 0° 22' N, 330-730 m elevation), a 3,500-ha private reserve within the Mache-Chindul Ecological Reserve in Es- meraldas Province, Ecuador. Bilsa is mostly undisturbed forest but also contains selective- ly-logged and regenerating forests. The sur- rounding area is largely deforested but also contains selectively logged and regenerating forests. We captured an adult Banded Ground-cuck- oo in a mist net in December 2002, obtained standard morphological measures, and applied three colored leg bands. We also attached a radio transmitter (model RI-2C, Holohil Sys- tems, Carp, ON, Canada) using a backpack style harness of thin rubber ligature (Vehren- camp and Halpenny 1981). The 6.0-g radio weighed 1.4% of the bird’s total body mass of 433 g. Total processing time was 43 min. We tracked the bird over a period of 210 days (until the radio failed) using a Telonics TR4 receiver and a RA-2AK “H” antenna. We attempted to record a single fix of the bird’s location during each radio-tracking ses- sion when we first observed the bird or when 206 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 we heard a vocalization in combination with a radio signal that indicated the distance be- tween the bird and the researcher was <20 m. We recorded only one fix per day to ensure independence of points. We recorded 53 fixes, 22 (42%) after observing the bird and 31 (58%) after hearing it. We recorded a single fix at each of two nests but suspended radio tracking during nesting to minimize distur- bance. We recorded UTM coordinates at each fix using a Garmin GPS and plotted the co- ordinates using the Animal Movement Anal- ysis extension in ArcView Geographical In- formation System (Hooge and Eichenlaub 2000). Fixes were visualized as minimum convex polygons (MCP’s) (Mohr 1947), and 95% and 50% fixed kernel isopleths using least-squares cross validation (Worton 1989, Seamann and Powell 1996). MCPs are includ- ed despite their inherent biases (Powell 2000) to allow comparison with other studies. We recorded the presence or absence of army ants and mammals at radio points and opportunis- tically recorded social behavior and foraging. We assessed habitat characteristics at 48 fixes using 10-m radius circular plots which were compared to 10-m radius circular control plots of known history (primary forest, selectively- logged forest, and secondary forest regener- ating 12-20 years). We estimated canopy height and number of trees in the genera Cec- ropia (Cecropiaceae) and Miconia (Melasto- mataceae) in each circular plot. Both are pi- oneer tree species that can be used as an index of habitat disturbance. We counted number of trees with a diameter at breast height (dbh) >50 cm in circular plots of 20 m. We mist netted a second adult on 4 May 2005 near a nest in incubation stage and at- tached a 6.0-g Holohil RI-2C radio transmitter (1.6% of the bird’s body mass of 365 g) with a backpack style harness. Total processing time was 38 min. We used radio tracking and visual observation to confirm this bird re- sumed normal incubation that same afternoon and continued through 7 May. On 4 June 2005 we discovered the remains of this bird (bones, most of which had been fractured, and feath- ers scattered over 3 m-) 420 m from point of capture. We recovered the radio 3 m from the remains. The broken bones and damage to the radio suggest predation by a mammal. We did not obtain any radio points for this bird. We present data as means ± SE. We ana- lyzed habitat data with pairwise Wilcoxon signed rank tests using P < 0.05 as the sig- nificance level because the data were not nor- mally distributed. Analyses were conducted using JMP (SAS 2003). RESULTS We obtained 53 independent fixes between 5 December 2004 and 7 July 2005. The MCP and kernel analysis home range sizes were, respectively, 42.2 ha and 49.9 ha (Fig. 1). The core use area was 3.4 ha (Fig. 1). Vegetation at fixes differed {P < 0.05) from plots in sec- ondary forest in all four parameters consid- ered (Table 1) and from plots in selectively- logged forest in two parameters, but did not differ from primary forest plots in any param- eter. The bird was observed with army ants on five occasions (9.5% of all fixes) but not with mammals. W^e observed the bird eat seven in- sects (4 grasshoppers, 3 unknown) that it flushed up and pursued as would a Roadrun- ner {Geococcyx spp.). On one occasion the bird appeared to eat decaying pulp (no larvae were present) of a fallen Gustavia dodsoiiii (Lecythidaceae) seed. Excluding observations at active nests, the radio-equipped bird was seen or heard with another adult Banded Ground-cuckoo on seven occasions (13.2% of fixes) evenly distributed throughout the study period. Morphological measures for the first and second radio-equipped birds, respectively, were tarsus (73.8 and 77.1 mm), wing chord (150 and 156 mm), tail length (250 and 254 mm), bill depth (19.6 and 16.7 mm), bill width (13.8 and 13.8 mm), culmen from the distal edge of the nare (33.8 and 35.1 mm), and exposed culmen (58.1 and 58.0 mm). The gender of both birds was unclear. The first ra- dio-equipped bird was lightly molting on the body and remiges, and the second bird was heavily molting on the body and remiges, and had an active brood patch. DISCUSSION This is the first report of home range size for any Neomorphinae, a sub-family of cuck- oos with five New World genera including the Ground-cuckoos and Roadrunners. The Band- ed Ground-cuckoo’s home range of ~50 ha, a SHORT COMMUNICATIONS 207 To La Piedrita 50% KA Core Area MCP Home Range Community (La Yesita) To La Y 200m t To Dogola Major Trail Minor Trail 95% KA Home Range Bilsa Stati FIG. 1 . Minimum convex polygon (MCP) home range and Kernel analysis (KA) home range and core use area of an adult Banded Ground-cuckoo during 7 months in Bilsa Reserve, northwestern Ecuador. conservative measure because of the study’s limited duration, is large relative to other ter- restrial rain forest species. Mountain Wood Quail (Odontophorus hyperythrus, 275 g) in the Colombian Andes had home ranges of 5.2 ha (Franco et al. 2006) and Chowchilla (Or- thonyx spaldingii, 150 g) in the Australian Wet Tropics had home ranges of <2 ha (Jan- sen 1999). Banded Ground-cuckoos are cryptic, diffi- cult to observe, unlikely to be captured in mist nets, and do not reliably respond to playback, making censuses using traditional methods unreliable. However, local population size can be estimated with home range data. Assuming Bilsa contains 2,000 ha of suitable habitat (e.g., primary forest; Jatun Sacha Foundation, unpubl. data) and the species forms socially monogamous pairs with 50-ha territories, Bil- sa could contain —40 pairs. This emphasizes the reserve’s importance for this and other en- demics given the widespread deforestation in the Choco (Sierra 1999, Conservation Inter- national 2001). Army ants are common at Bilsa and the Banded Ground-cuckoo occasionally associ- ated with them (9.5% of fixes) but was not an obligate or even frequent army ant follower as has been previously suggested (Willis and On- iki 1978, Willis 1982, Ridgely and Greenheld 2001). It did not associate with mammals, al- though this may be due to the rarity of pec- TABLE 1. Habitat characteristics at Banded Ground-cuckoo radio fix locations versus control plots. Plot type Canopy ht. (m) Trees >50 cm dbh M icon id Cecropia Radio fixes (n = 48) 23.27 ± 0.94 4.31 ± 0.36 0.64 ± 0.16 0.79 ± 0.26 Control in = 170) Primary (n = 98) 25.26 ± 0.42NS 5.08 ± 0.26^'^ 0.42 ± 0.1 5NS 0.48 ± 0.1 4NS Selectively-logged (/? = 34) 22.38 ± 0.9()Ns 3.79 ± 0.58NS 1.79 ± 0.42* 2.88 ± 0.68*** Secondary (/; = 38) 15.42 ± 1.04*** 1.18 ± 0.20*** rj 1 + be * 4.34 ± 0.98*** * = P < O.O.S; ** = p < 0.01; = /’ < 0.(K)l; NS = /> > 0.05. 208 THE WILSON JOURNAL OF ORNITHOLOGY • Vol 120, No. 1, March 2008 caries and primates within its home range (JK, unpubl. data). Our observations here in com- bination with a previous study of the species nesting biology (Karubian et al. 2007) suggest that large insects and small vertebrates are an important part of the diet. The Banded Ground-cuckoo, as a terrestrial omnivore that primarily captures rapid live prey, fills an un- usual ecological niche among neotropical rain forest birds, which may be reflected in its large home range. The radio-equipped bird was with another adult at least seven times from December into May. The adults maintained vocal but not vi- sual contact when together and often foraged >20 m apart. Breeding extended from March into May and birds exhibited bi-parental care at the nest (Karubian et al. 2007). L6pez-La- nus et al. (1999) recorded an apparent juvenile at Bilsa in February 1997, suggesting the breeding season may begin earlier. The color- banded bird that we radio-tracked in the pre- sent study was repeatedly observed in April 2007 close to the 2005 nest sites. The home range contained a mosaic of hab- itat types including ~25% secondary forest, but the Banded Ground-cuckoo showed a strong preference for undisturbed forest. Two nesting attempts and care of a newly fledged chick also occurred in primary forest (Karu- bian et al. 2007). This suggests dependence on primary forest both for nesting and forag- ing, and corroborates previous assessments of the species’ conservation requirements (BirdLife International 2000, Greenfield 2002, Renjifo et al. 2002). Deforestation and frag- mentation appear to be the primary threats to the Banded Ground-cuckoo; protection of the remaining primary Choco rainforest should be high priority for conservation of this species. Other factors contributing to the species’ vul- nerability may include its relatively large home range, low density, presumably poor dispersal ability, and apparent clutch size of one egg (Karubian et al. 2007). ACKNOWLEDGMENTS Jorge Olivo, Domingo Cabrera, Andrew Cook, and Alex Dunn gathered much of the data and J. L. Storey and R. Dufaes assisted with MCP and kernel analyses and maps. We thank each of them for their dedication and excellent work. T B. Smith provided important intellectual and logistical support. The manuscript was improved by comments from Jiigen Haffer and one anonymous reviewer. We also thank the Bilsa Biolog ical Station, especially Carlos Aulestia, Julieta Bir- mingham, and M. B. McColm. This research was sup- ported by the Chicago Zoological Society; Cleveland Metroparks Zoo; Conservation, Food and Health Foundation; Disney Wildlife Conservation Fund; Na- tional Geographic Society; and National Science Foun- dation (OISE-0402137). All research was conducted with permission from Fundacion Jatun Sacha and with approval of the Ecuadorian Ministry of the Environ- ment (Permit 009-CI-FAU-DRE-MA), and the Univer- sity of California, Los Angeles. LITERATURE CITED BiRoLire International. 2000. Threatened birds of the world. BirdLife International and Lynx Edi- ciones, Barcelona, Spain. Conservation International. 2001. Ecosystem pro- file; Choco-Manabi conservation corridor of Cho- c6-Darien- western Ecuador hotspot. Conservation International, Washington, D.C., USA. Franco, R, K. Fierro-Calderon, and G. Kattan. 2006. Population densities and home range sizes of the Chestnut Wood-quail. Journal of Field Or- nithology 77:85—90. Greenfield, P. 2002. Neomorphus radiolosus. in Libro rojo de las aves del Ecuador (T. Granizo, C. Pa- checo, M. B. Ribadeneira, M. Guerrero, and L. Suarez, Editors). Simbioe, Conservation Interna- tional, EcoCiencia, Ecuadorian Ministerio del Ambiente and lUCN, Quito. Haffer, J. 1977. A systematic review of the neo-trop- ical Ground-cuckoos (Aves, Neomorphus). Bon- ner Zoologische Beitraege 28:48-76. Hooge, P. N. and B. Eichenlaub. 2000. Animal move- ment extension for Arcview Version 2.04. USGS, Alaska Science Center, Biological Sciences Of- fice, Anchorage, USA. Hornbuckle, j. 1997. Two sightings of Banded Ground-cuckoo Neomorphus radiolosus in Ecua- dor. Cotinga 8:90. Jansen, A. 1999. Home ranges and group-territoriality in Chowchillas Orthonyx spaldingii. Emu 99:280- 290. Karubian, J., L. Carrasco, D. Cabrera, A. Cook, AND J. Olivo. 2007. Nesting biology of the Band- ed Ground-cuckoo {Neomorphus radiolosus). Wil- son Journal of Ornithology 119:221—227. Lopez-Lanus, B., K. Berg, R. Strewe, and P. G. W. Salaman. 1999. The ecology and vocalisations of Banded Ground-cuckoo Neomorphus radiolosus. Cotinga 11:42-45. Mohr, C. O. 1947. Table of equivalent populations of North American small mammals. American Mid- land Naturalist 37:223-249. Payne, R. B. 1997. Family Cuculidae (cuckoos). In Handbook of the birds of the world. Volume 4. Sandgrouse to cuckoos (J. del Hoyo, A. Elliott, and J. Sargatal, Editors). Lynx Ediciones, Barce- lona, Spain. Powell, R. A. 2000. Research techniques in animal SHORT COMMUNICATIONS 209 ecology: controversies and consequences. Colum- bia University Press, New York, USA. Renjifo, L. M., a. M. Franco-Maya, J. D. Amaya- Espinel, G. H. Kattan, and B. Lopez-Lanus. 2002. Libro rojo de aves de Colombia. Instituto de investigacion de Recursos Biologicos Alexan- dex von Humboldt y Ministerio del Medio Am- biente, Bogota, Colombia. Ridgely, R. S. and P. Greenfield. 2001. The birds of Ecuador. Cornell University Press, Ithaca, New York, USA. Roth, P. 1981. A nest of the Rufous-vented Ground- cuckoo {Neomorphus geojfroyi). Condor 83:388. SAS. 2003. IMP statistical software. SAS Institute Inc., Cary, North Carolina, USA. Seamann, D. E. and R. a. Powell. 1996. An evalu- ation of the accuracy of kernel density estimators for home range analysis. Ecology 77:2075-2085. Sick, H. 1949. Beobachtungen an dem brasilianischen bodenkuckkuck Neomorphus geojfroyi dulcis Snethlage. Pages 229-239 in Ornithologie als biologische Wissenschaft. Stresemann-Eestschrift, Heidelberg, Germany. SiEGAL, C. E., J. M. Hamilton, and N. R. Castro. 1989. Observations of the Red-billed Cuckoo {Neomorphus pucheranii) in association with tam- arins (Saguinas) in northeastern Amazonian Peru. Condor 91:720-722. Sierra, R. 1999. Vegetacion remanente del Ecuador Continental. Circa 1996. 1:1,000,000. Proyecto INEFAN/GEE y Wildlife Conservation Society, Quito, Ecuador. Terborgh, j. 1983. Eive New World primates: a study in comparative ecology. Princeton University Press, Princeton, New Jersey, USA. Vehrencamp, S. L. and L. Halpenny. 1981. Capture and radio-transmitter attachment techniques for Roadrunners. North American Bird Bander 6: 128-132. Willis, E. O. 1982. Ground-cuckoos (Aves: Cuculi- dae) as army ant followers. Revista Brasileira de Biologia 42:753-756. Willis, E. O. and Y. Oniki. 1978. Birds and army ants. Annual Review of Ecology and Systematics 9: 243-246. WoRTON, B. J. 1989. Kernel methods for estimating the utilization distribution of home range studies. Ecology 70:164-168. The Wilson Journal of Ornithology 120(1):209-213, 2008 Comparisons between Juvenile and Adult American Robins Foraging for Mulberry Fruit E. Natasha Vanderhoff ^ 3 and Perri K. Eason’ ABSTRACT — American Robins (Turdus migrato- rius) depend on fruit in fall and winter and, accord- ingly, juvenile robins must quickly learn to acquire this resource. We compared the foraging abilities of juve- niles and adults foraging for mulberries (Morus spp.), the first fruit that ripens after juveniles have become independent. Juveniles were significantly less success- ful than adults at obtaining mulberries; 23% of juve- niles’ attempts to remove fruit were successful com- pared to 69% of attempts by adults to remove fruit. As a result, juveniles consumed 0.4 mulberries/min whereas adults consumed 2.8 mulberries/min. Further investigations of juveniles’ ability to forage for fruit are needed to understand the mechanisms involved in ' Department of Biology, University of Louisville, Louisville, KY 40292, USA. 2 Current address: Department of Biology, Francis Marion University, Florence, SC 29501, USA. ^Corresponding author; e-mail: evanderhoff @ fmarion.edu the development of skills to forage for this vital re- source. Received 20 February 2007. Accepted 17 June 2007. Foraging for fruit is often assumed to be an easily acquired skill, unlike foraging for ani- mal prey (Stevens 1985, Desrochers 1992, Ricklefs 2004). Support for this assumption regarding avian fruit foraging, however, is mixed. Analyses of stomach contents at times show that juveniles consume more fruit than adults, which suggests that fruits are relatively easily obtained (Wheelwright 1986, Eggers 2000). In addition, some studies found no dif- ferences in foraging behavior or rates of fruit consumption between juveniles and adults (European Robin \Erithacus ruhecuUi], Com- mon Blackbird \Turdus merula]., Eurasian 210 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 Blackcap [Sylvia atricapilla]; Hampe 2001, Bas et al. 2006). However, juveniles often consume different types of fruits than adults, which could be a result of differences in for- aging abilities (Boddy 1991), and tests of skill level and acquisition of skill have yet to be thoroughly conducted. If juveniles are less skilled at foraging for fruit, their lower proficiency could be costly, especially for individuals of the many species that depend on fruit for migration and over- wintering (Willson 1986, Blake and Loiselle 1992, Levey and del Rio 2001, McCarty et al. 2002, Heise and Moore 2003, Kwit et al. 2004). Juvenile birds often have lower surviv- al rates than adults and starvation is a major cause of juvenile mortality (Sullivan 1988, Catterall et al. 1989, Donnelly and Sullivan 1998). Accordingly, juveniles of migratory species and those that have high mortality should be expected to develop their fruit-for- aging skills quickly. One species that relies heavily on fruit for migration and overwintering is the American Robin {Turdus migratorius). The diet of the robin is composed primarily of fruit from late summer through winter (Wheelwright 1986L We compared foraging rates on mulberry fruit (Morus spp.) to investigate whether juveniles were less successful at foraging on fruit than adults. Mulberry is the first tree to fruit after juveniles begin to forage independently at our study site in Louisville, Kentucky. According- ly, we were able to 'compare naive juveniles to experienced adults in their ability to forage on this fruit. Our specific objectives were to: (1) quantify and note differences in the out- come of mulberry foraging bouts by juveniles and adults and (2) examine whether or not ju- veniles and adults differ in their rates of suc- cessful and failed attempts to pick mulberries. METHODS We investigated juvenile and adult robins’ ability to forage for mulberry fruit from 25 May until 23 June 2006 at Joe Creason Park in Louisville, Kentucky (38° 12' N, 85° 42' W). Both adult and juvenile robins foraged in fruiting trees simultaneously at our field site and there was little apparent competition be- tween age groups. We conducted continuous focal scans on 64 American Robins (38 adult scans and 26 juvenile scans) foraging for mul- berries on trees scattered throughout the 27.5- ha park. Robins typically flew into the trees, ate fruit, and left relatively quickly; therefore, each scan started when a focal bird entered a mulberry tree and ended when it left the tree or was no longer visible. Samples ranged in length from 10 to 243 sec with the average scan lasting a little over a minute (mean ± SE = 61 ± 5.5 sec). We recorded four behaviors to assess the ability of juveniles and adults to forage for mulberries. First, when a robin unsuccessfully tried to pick a mulberry, we categorized it as a failure. Typically, a failure resulted when a robin was either unable to reach a fruit or un- able to pull sufficiently hard to detach the mulberry. Second, when a robin successfully picked and ate a whole mulberry we recorded it as an ‘eat’. Third, because each mulberry is composed of multiple drupelets, robins at times picked and ate only a part of one mul- berry; we recorded this as ‘piece’. Finally, if a robin picked a fruit, but subsequently dropped it, we called it a ‘drop . We calculated percent success (number of eats/total attempted picks) and percent failure (number of fails/total attempted picks) for each adult and juvenile scan, and compared adult and juvenile scans using a MANOVA followed by individual ANOVAs. We also calculated fruit capture rate (number of eats/ min) and fruit failure rate (number of failures/ min) for all scans, and compared juvenile scans and adult scans using a MANOVA and follow-up ANOVAs. Fruit failure rate was transformed to meet MANOVA assumptions as follows: transformed fruit failure rate = 1/ (1 + failure rate). Few robins dropped fruit and robins seldom picked only a piece of the mulberry. The piece and drop data did not meet MANOVA assumptions, and could not be successfully transformed. We excluded these data from the MANOVA analyses and instead used two Fisher’s exact tests to ex- amine whether a greater proportion of juve- niles engaged in these behaviors than adults. Means ± SE are reported from non-trans- formed data and SAS© (2003) was used for all statistics. RESULTS Juvenile American Robins were significant- ly less skilled at obtaining mulberries than SHORT COMMUNICATIONS 211 80 cc c ‘w 70 iS w 0 0 0 S 0 60 0 w w 0 50 Q. 0 E 40 0 0 0 30 0 c 0 20 0 X) =3 0 CL E 10 Adults Juveniles FIG. 1. Juvenile vs. adult American Robins’ success at obtaining mulberries. adults (MANOVA: Wilks’ Lambda F259 = 15.27, P < 0.001); only 23 ± 6.4% of at- tempts to pick mulberries by juveniles were successful compared to 69 ± 5.2% for adults (ANOVA: F, 6o = 29.55, P < 0.001; Fig. 1). In addition, a significantly larger percentage of attempts by juveniles ended in failure (AN- OVA: F160 = 25.41, P < 0.001; juvenile mean = 64 ± 6.7%, adult mean = 23 ± 4.6%). Foraging rates were also significantly different between juveniles and adults (MANOVA: Wilks’ Lambda p2,6i ~ 24.74, P < 0.001). Ju- veniles consumed significantly fewer mulber- ries per minute than adults (ANOVA: Pj ^2 = 49.96, P < 0.001) with juveniles consuming less than one mulberry/min (0.4 ± 0.14) and adults consuming nearly three mulberries/min (2.8 ± 0.26; Fig. 2). Juveniles also had a high- er rate of failure than adults (ANOVA: Pj 52 FIG. 2. Juvenile vs. adult American = 4.11, P < 0.047) with juveniles, on aver- age, having 1.65 (±0.33) failures/min and adults having 1.04 (±0.22) failures/min. Both juveniles and adults seldom dropped a mul- berry once it was picked; only 4% of juveniles and 5% of adults had drops (Fisher’s exact test, P = 1.00). There was no significant dif- ference in the proportion of adults and juve- niles which consumed only a piece of a mul- berry (Fisher’s exact test, P = 0.52). DISCUSSION Juvenile American Robins were less profi- cient mulberry foragers than adults; they were less successful at obtaining mulberry fruit and consumed significantly fewer mulberries per minute than adults. Our results strongly indi- cate that juvenile robins need to develop skills to forage for fruit. f Juveniles Robins’ fruit capture rale per minute. 212 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 Other studies have found no differences in foraging rates between juveniles and adults foraging for fruit (Hampe 2001, Bas et al. 2006). One reason for the disparity may be that other researchers observed juveniles that had already developed their skills to forage for fruit. Mulberry, in our study, was the first fruit that juvenile robins learned how to obtain, and the disparity between juveniles and adults may be greatest at this time. If experience is a critical factor in development of foraging skills, juveniles should show improvement when foraging on other fruits later in the sea- son. There is some supporting evidence for this at our field site, where juvenile robins had higher success at obtaining and consuming cherries, which fruit later in the season (Van- derhoff and Eason 2007). Fruit type may also be an important factor in why some studies found no differences be- tween juveniles and adults. First, fruit type can influence foraging rates; a previous inves- tigation of robins foraging for fruit found that fruit type and, more specifically, size and clus- tering of fruits within a plant were important in affecting rates of fruit consumption for adults (White and Stiles 1991). Second, type of fruit may also influence the level of skill required to obtain it. Whereas some fruits may be more challenging to acquire and require ex- perience, other fruits may require little or no experience to obtain. For example less expe- rienced Common Whitethroats (Sylvia com- munis) foraged primarily for woody night- shade (Solanum dulcamara) berries, a simple fruit, while older birds were able to forage for aggregate blackberries (Rubus fruticosus) (Boddy 1991). Similarly, in our study less ex- perienced juveniles had difficulty obtaining mulberry fruit which, like blackberry, is not a simple fruit. Studies that have not found dif- ferences between juveniles and adults have fo- cused on simple fruits that may require little or no experience to obtain (glossy buckthorn [Frangula alnus], Hampe 2001; Italian buck- thorn [Rhamnus alaternus], Bas et al. 2006). Juveniles may be less adept at obtaining fruit than adults due to either physical or cog- nitive constraints. Gape size is a limiting fac- tor in fruit selection (Wheelwright 1985, Snow and Snow 1988) and juveniles in some species may not have a sufficient gape to ob- tain certain fruits. However, previous investi- gations have shown that juvenile gape size does not seem to limit fruit choice or ingestion (White and Stiles 1991, Jung 1992). Cognitive abilities of juveniles may not be fully devel- oped leading to less proficient fruit foraging. For instance, juveniles are often less adept at identifying suitable food items (Whitehead 1984, Caro 1994, Estes et al. 2003). This sug- gests that juvenile robins may be less able to recognize ripe fruits and, thus, may attempt to pick a higher percentage of unripe fruits. Fruit ripeness can have a two-fold effect on forag- ing rates; it can result in higher failure rates because fruits cannot be removed or lower consumption rates due to the increased amount of time to remove unripe fruit. Fruit is a vital resource for birds and it is surprising that relatively few studies have in- vestigated foraging for fruit by juveniles, and the mechanisms involved in their foraging de- cisions. Our results show that juvenile robins are less proficient at foraging for fruits than adults casting doubt on the assumption that foraging for fruit requires little or no skill ac- quisition. The assumption that foraging for fruit is easier than other tasks may be accu- rate; we did not examine this aspect of the development of skills of robins. Future re- search should be conducted to understand how the rate at which juveniles develop skills at foraging for fruit compares to the rate at which they learn to forage for other food items, as well as the mechanisms that underlie the learning process. Answering these ques- tions will lead to a better understanding of the factors that affect foraging for fruit and add to our knowledge of animal learning, and plant- animal interactions. ACKNOWLEDGMENTS We thank C. J. Byers, C. J. Hanna, L. B. Rifai, and T C. Pelletier for helpful comments on the manuscript. We also thank Louisville Metro Parks for permission to conduct the study at Joe Creason Park. LITERATURE CITED Bas, j. M., P Pons, and C. Gomez. 2006. Exclusive frugivory and seed dispersal of Rhamnus alternus in the bird breeding season. Plant Ecology 183: 77-89. Blake, J. G. and B. A. Loiselle. 1992. Fruits m the diets of neotropical migrant birds in Costa Rica. Biotropic a 24:200—210. Boddy, M. 1991. Some aspects of frugivory by bird SHORT COMMUNICATIONS 213 populations using coastal dune scrub in Lincoln- shire. Bird Study 38:188-199. Caro, T. M. 1994. Hunting and grouping in adoles- cence. Pages 161-197 in Cheetahs of the Seren- geti. University of Chicago Press, Chicago, Illi- nois, USA. Catterall, C. P, J. Kikkawa, and C. Gray. 1989. Inter-related age-dependent patterns of ecology and behaviour in a population of Silvereyes (Aves: Zosteropidae). Journal of Animal Ecology 58:557-570. Desrochers, a. 1992. Age and foraging success in European Blackbirds: variation between and with- in individuals. Animal Behaviour 43:885-894. Donnelly, R. E. and K. A. Sullivan. 1998. Foraging proficiency and body condition of juvenile Amer- ican Dippers. Condor 100:385-388. Eggers, S. 2000. Compensatory frugivory in migra- tory Sylvia warblers: geographical responses to season length. Journal of Avian Biology 31:63- 74. Estes, J. A., M. L. Riedman, M. M. Staedler, M. T. Tinker, and B. E. Lyon. 2003. Individual varia- tion in prey selection by sea otters: patterns, caus- es, and implications. Journal of Animal Ecology 72:144-155. Hampe, a. 2001. The role of fruit diet within a tem- perate breeding bird community in southern Spain. Bird Study 48:116-123. Heise, C. D. and E R. Moore. 2003. Age-related dif- ferences in foraging efficiency, molt, and fat de- position of Gray Catbirds prior to autumn migra- tion. Condor 105:496-504. Jung, R. E. 1992. Individual variation in fruit choice by American Robins (Turdus migratorius). Auk 109:98-111. Kwit, C., D. j. Levey, C. H. Greenberg, S. F. Pear- son, J. P. McCarty, S. Sargent, and R. L. Mum- ME. 2004. Fruit abundance and local distribution of wintering Hermit Thrushes {Catharus guttatus) and Yellow-rumped Warblers (Dendroica coron- ata) in South Carolina. Auk 121:46-57. Levey, D. J. and C. M. del Rio. 2001. It takes guts (and more) to eat fruit: lessons from avian nutri- tional ecology. Auk 118:819-831. McCarty, J. P, D. J. Levey, C. H. Greenberg, and S. Sargent. 2002. Spatial and temporal variation in fruit use by wildlife in a forested landscape. Forest Ecology and Management 164:277-291. Ricklefs, R. E. 2004. The cognitive face of avian life histories. Wilson Bulletin 116:119-133. SAS. 2003. SAS system. Version 9.1 for Windows. SAS Institute, Cary, North Carolina, USA. Snow, B. and D. Snow. 1988. Birds and berries. T. & A. D. Poyser Limited, Carlton, United Kingdom. Stevens, J. 1985. Foraging success of adult and ju- venile Starlings Sturnus vulgaris: a tentative ex- planation for the preference of juveniles for cher- ries. Ibis 127:341-347. Sullivan, K. A. 1988. Ontogeny of time budgets in Yellow-eyed Juncos: adaptation to ecological con- straints. Ecology 69:118-124. Vanderhoff, E. N. and P. K. Eason. 2007. Disparity between adult and juvenile American Robins Tur- dus migratorius foraging for ground invertebrates and cherry fruits. Ethology 113:1212-1218. Wheelwright, N. T. 1985. Fruit size, gape width, and the diets of fruit-eating birds. Ecology 66:808- 818. Wheelwright, N. T. 1986. The diet of American Rob- ins: an analysis of U.S. Biological Survey records. Auk 103:710-725. White, D. W. and E. W. Stiles. 1991. Fruit harvesting by American Robins: influence of fruit size. Wil- son Bulletin 103:690-692. Whitehead, J. M. 1984. Development of feeding se- lectivity in mantled howling monkeys, Alouatta palliate. Pages 105-1 17 in Primate ontogeny, cog- nition, and social behavior (J. G. Else and P. C. Lee, Editors). Cambridge University Press, Cam- bridge, United Kingdom. Willson, M. E 1986. Avian frugivory and seed dis- persal in eastern North America. Current Orni- thology 3:223-279. 214 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 The Wilson Journal of Ornithology 120(1);214-216, 2008 Previously Unknown Food Items in the Diet of Six Neotropical Bird Species Luis Sandoval,'’^ Esteban Biamonted and Alejandro Solano-Ugalde^ ABSTRACT — We report new food items for six species of Costa Rican birds. This report includes the first egg predation observed for Hoffman’s Woodpeck- er {Melanerpes hoffmannii), the first vertebrate record- ed in the diet of Sooty Thrush (Turdus nigrescens), and records for Groove-billed Ani (Crotophaga sulci- rostris). Black-and-white Owl {Strix nigrolineata). Blue-crowned Motmot {Momotus momota), and Clay- colored Thrush (Turdus grayi). Received 24 January 2007. Accepted 3 May 2007. Basic knowledge of the diet of bird species is fundamental to understanding their ecolog- ical requirements and designing effective con- servation plans for their protection (Kantak 1979, Barrantes and Loiselle 2002). Unfortu- nately, our understanding of the diets of wild tropical birds is poor. New food items are con- tinuously being added (Chacon-Madrigal and Barrantes 2004, Garcia-C. and Zahawi 2006) to the known diets for even well-studied spe- cies, such as motmots (Skutch 1945, 1947, 1964, 1971; Orejuela 1980; Remsen et al. 1993). The results of these studies indicate that many species have a more varied diet than was previously thought. Information on the diets of Costa Rican birds was generally and briefly summarized by Stiles and Skutch (1989). We report here additional diet items for five species which have a wide distribution in the Neotropics (Stiles and Skutch 1989, Stotz et al. 1996) and present more information about their diets than has been available (Ibanez et al. 1992, Gerhardt et al. 1994, Eitniear and Aragon- Tapia 2000). One of these species, the Sooty Thrush {Turdus nigrescens), is endemic in the highlands of Costa Rica and Panama (Stiles ' Escuela de Biologia, Universidad de Costa Rica, San Pedro de Montes de Oca, Costa Rica. 2 Casa #6, Coopecabanas, Santa Ana, San Jose, Cos- ta Rica. 3 Corresponding author; e-mail: biosandoval@hotmail.com and Skutch 1989, Stotz et al. 1996), and has not had detailed study of its diet. Our specific objective is to provide information on undoc- umented food items eaten by six species of Costa Rican birds that we recorded over the last 5 years during fieldwork throughout the country. SPECIES ACCOUNTS Groove-billed Ani {Crotophaga sulciros- tris). We observed an ani holding a Ground- sparrow (Melozone sp.) egg in its bill on 6 July 2005 at Getsemani, Heredia (10°06'N, 84° 02' W; 1,350 m). The eggshell was whit- ish with small reddish spots characteristic of the genus (Stiles and Skutch 1989). We earlier (1 week) saw a White-eared Ground Sparrow (M. leucotis) with nest material in its bill, 50 m from where the ani was first found and it is possible the egg belonged to this species. The ani swallowed the egg contents through a hole in the eggshell. This is the second re- cord of egg predation by the Groove-billed Ani (Eitniear and Aragon-Tapia 2000) and the first for the tropics. The Smooth-billed Ani (C. ani) (Alvarez 1975) and the Groove-billed Ani are the only species in the genus that have been observed eating eggs. Black-and-white Owl (Strix nigrolineata). We observed a Black-and-white Owl on 6 April 2005 at Altamira Station, La Amistad International Park (9° 01' N, 83° 00 W; 1,382 m) fly to a cecropia tree (Cecropia sp.) at 1745 hrs with a Barn Swallow (Hirundo rus- tica) in its claws where another Black-and- white Owl was perched. The first owl passed the swallow from its bill to the bill of the sec- ond owl in ~2 sec. The second owl ate the swallow in one gulp after biting the neck and head while the swallow was held in its claw. Birds are known to be part of the diet of the Black-and-white Owl (Ibanez et al. 1992, Ger- hardt et al. 1994), but predation on Barn Swal- lows has been not previously reported. Barn SHORT COMMUNICATIONS 215 Swallows roost on cables in open areas during migration (Skutch 1944); this behavior can make it easy prey for nocturnal predators (Mcllhenny 1937, LaPorte 1974). Predation of swallows by other species of related neotrop- ical owls has not been reported, although owl predation on birds is not rare (Marks et al. 1999). Blue-crowned Motmot {Momotus momota). We found an individual Blue-crowned Mot- mot on 24 September 2003 eating a shrew {Cryptotis sp.) (estimated length of 6 cm) at Coronado, San Jose (9° 57' N, 83° 59' W; 1,500 m). The motmot was first seen on the ground in a house garden with the prey in its bill. The motmot repeatedly hit the shrew against the ground for about 5 min and then flew to a higher perch where it swallowed the prey after repeated smashing it for ~2 min. We could see the silhouette of the shrew in the bird’s crop after swallowing; the motmot remained motionless for at least 5 min before departing. Several authors report this species of motmot commonly feeds on small verte- brates such as birds, toads, lizards, and snakes (Skutch 1983, Stiles and Skutch 1989, Ridge- ly and Gwynne 1989), but small mammals are taken rarely (Chacon-Madrigal and Barrantes 2004). We suspect the shrew may have been at the upper size limit of prey consumed by this species based on the manner in which the motmot manipulated it. We observed a Blue-crowned Motmot on 4 March 2004 on the Universidad de Costa Rica Campus, in San Jose (9° 54' N, 84°03'W; 1,200 m), eating a portion of wasp nest of about 5 cm in length. The bird struck the wasp nest fragment against the ground in a manner similar to that used for subduing animal prey. It was not possible to identify the content of the nest fragment. The Blue-crowned Motmot is known to consume social insects, such as ants and wasps, as part of its diet (Skutch 1983, Raw 1997, Ridgely and Greenfield 2001), but attacks on wasp nests were previ- ously unknown. Hoffmann’s Woodpecker (Melcmerpes hojf- mcmnii). We observed an individual Hoff- mann’s Woodpecker on 9 March 2004 at Cor- onado, San Jose (9°58'N, 83° 59' W; 1,500 m) eating an unidentified bird egg, which was completely white and no longer than 2 cm. The woodpecker Hew into a single tree in a semi-open area holding the egg in its bill. Af- ter arriving at a perch, the bird climbed the tree to a fork where it placed the egg and be- gan to peck a hole into the egg. The bird then started feeding on the contents by rapidly sticking its tongue in and out of the hole in the egg. Woodpeckers have a worldwide dis- tribution and have previously been observed consuming eggs in four species of Melaner- pes, two species of Picus, the Red-naped Sap- sucker (Sphyrapicus nuchalis). Great Spotted Woodpecker (Dendrocopos major), and Cu- ban Green Woodpeeker (Xiphidiopicus per- cussus) (Bryant 1921, Fajer et al. 1987, Coo- per 1992, Winkler and Christie 2002). Sooty Thrush {Turdus nigrescens). We saw a Sooty Thrush eating a lizard {Mesaspis mon- ticola) on 28 May 2001 at Chirripo National Park (9°27'N, 83° 26' W; 3,600 m). When first observed, the bird was on the ground and had the lizard in its bill, constantly hitting it against the rocky ground surface. The action of hitting the prey on the ground appeared challenging as the thrush took five rests during 3 min of observation. The bird placed the liz- ard on the ground for periods of about 15 sec throughout the resting periods. The thrush then walked into the dense vegetation with the lizard and disappeared where it is presumed the prey was consumed. The Sooty Thrush in- habits an area in which the lizard Mesaspis monticola is abundant (Savage 2002), but has not been previously reported feeding on this lizard. Previous observations indicate that ber- ries and small invertebrates constitute the bulk of the diet of the Sooty Thrush (Stiles and Skutch 1989, Clement 2000). Clay-colored Thrush {Turdus grayi). We observed a Clay-colored Thrush feeding on a house gecko (Hemidactilus frenatus) on 3 1 January 2005 at Golfito, Puntarenas (08° 39' N, 83° 09' W; 5 m) while perched on the lawn in front of the Universidad de Costa Rica stu- dent houses. The Clay-colored Thrush squashed the gecko’s head by repeatedly clos- ing and opening its bill for approximately 40 sec. The bird then moved its head back and with small lateral movements swallowed the gecko. This thrush is mainly frugivorous, but lizards and snakes have been documented as constituting a part of its diet (Feduccia 1971, Stiles and Skutch 1989, Clement 2000); how- ever this gecko species has not previously 216 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 been reported in the diet of the Clay-colored Thrush. The house gecko, despite being a re- cent colonist (Savage 2002), was recently re- ported as part of the diet of the House Wren {Troglodytes aedon) (Barquero and Hilje 2005), indicating this invasive species has be- come a new food resource for some species. ACKNOWLEDGMENTS We thank Julio Sanchez, Joseph Wunderle, C. E. Braun, and an anonymous reviewer for valuable com- ments’on an earlier draft of the manuscript. We also thank Alisson Olivieri, L. D. Holtzman, and C. E. Braun for suggestions that improved our English re- daction. We offer our most sincere thanks to all the persons who supported and encouraged our fieldwork throughout the past years. LITERATURE CITED Alvarez, H. 1975. The social system of the Green Jay in Colombia. Living Bird 14:5-44. Barquero, M. and B. Hilje. 2005. House Wren preys on introduced gecko in Costa Rica. Wilson Bul- letin 117:204-205. Barrantes, G. and B. Loiselle. 2002. Reproduction, habitat use, and natural history of the Black-and- Yellow Silky-flycatcher (Phainoptila melanoxan- tha), an endemic bird of the western Panama-Cos- ta Rica highlands. Ornitologia Neotropical 13: 121-136. Bryant, H. 1921. California Woodpecker steals eggs of Wood Pewee. Condor 23:33. Chacon-Madrigal, E. and G. Barrantes. 2004. Blue-crowned Motmot {Momotus momota) pre- dation on a long-tongued bat (Glossophaginae). Wilson Bulletin 116:108-110. Clement, P. 2000. Thrushes. Princeton University Press, Princeton, New Jersey, USA. Cooper, J. 1992. Egg eating by a Red-naped Sapsuck- er (Sphyrapicus nuchalis). Northwestern Natural- ist 73:59-60. Eitniear, J. and a. Aragon-Tapia. 2000. Red-billed Pigeon (Columba flavirostris) nest predated by Groove-billed Ani {Crotophaga sulcirostris). Or- nitologia Neotropical 11:231-232. Fajer, E., K. Schmidt, and J. Eschler. 1987. Acorn Woodpecker predation on Cliff Swallow nests. Condor 89:177-178. Feduccia, a. 1971. Turdus grayi feeding on a snake. Auk 83:197. Garcia-C., j. M. and R. a. Zahawi. 2006. Predation by a Blue-crowned Motmot {Momotus momota) on a hummingbird. Wilson Journal of Ornithology 118:261-263. Gerhardt, R., D. McAnnis-Gerhardt, C. Flatten, AND N. Bonilla-Gonzales. 1994. The food habits of sympatric Ciccaba owls in northern Guatemala. Journal of Field Ornithology 65:258-264. Ibanez, C., C. Ramo, and B. Busto. 1992. Notes on food habits of the Black and White Owl. Condor 94:529-531. Kantak, G. 1979. Observation on some fruits-eating birds in Mexico. Auk 96:183-186. La Porte, P. 1974. Common Grackle kills a Bam Swallow. Wilson Bulletin 86:477-478. Marks, J. S., R. J. Canning, and H. Mikkola. 1999. Family Strigidae (typical owls). Pages 76-242 in Handbook of the birds of the world. Volume 5. Barn-owls to hummingbirds (J. del Hoyo, A. El- liot, and J. Sargatal, Editors). Lynx Edicions, Bar- celona, Spain. McIlhenny, E. a. 1937. Life history of the Boat-tailed Grackle in Louisiana. Auk 54:274-295. Orejuela, j. 1980. Niche relationships between Tur- quoise-browed and Blue-crowned motmots in the Yucatan Peninsula, Mexico. Wilson Bulletin 92: 229-244. Raw, a. 1997. Avian predation on individual neotrop- ical social wasps (Hymenoptera, Vespidae) out- side their nest. Ornitologia Neotropical 8:89-92. Remsen Jr., j. V., M. Hyde, and A. Chapman. 1993. The diets of neotropical trogons, motmots, barbets and toucans. Condor 95:178-192. Ridgely, R. and P. Greenfield. 2001. The birds of Ecuador. Cornell University Press, Ithaca, New York, USA. Ridgely, R. and J. Gwynne. 1989. A guide to the birds of Panama. Princeton University Press, Princeton, New Jersey, USA. Savage, J. 2002. The amphibians and reptiles of Costa Rica: a herpetofauna between two continents, be- tween two seas. University of Chicago Press, Chi- cago, Illinois, USA. Skutch, a. 1944. Barn Swallow in Costa Rica in July. Auk 61:470-471. Skutch, A. 1945. Life history of the Blue-throated Green Motmot. Auk 62:489-517. Skutch, A. 1947. Life history of the Turquoise-browed Motmot. Auk 64:201-217. Skutch, A. 1964. Life history of the Blue-diademed Motmot Momotus momota. Ibis 106:321—332. Skutch, A. 1971. Life history of the Broad-billed Mot- mot with notes on the Rufous Motmot. Wilson Bulletin 83:74-94. Skutch, A. 1983. Birds of tropical America. Univer- sity of Texas Press, Austin, USA. Stiles, G. and A. Skutch. 1989. A guide to the birds of Costa Rica. Cornell University Press, Ithaca, New York, USA. Stotz, D., j. Fitzpatrick, T. Parker, and D. Mos- KOVITZ. 1996. Neotropical birds: ecology and con- servation. University of Chicago Press, Chicago, Illinois, USA. Winkler, H. and D. A. Christie. 2002. Family Picidae (woodpeckers). Pages 296—555 in Handbook of the birds of the world. Volume 7. Jacamars to woodpeckers (J. del Hoyo, A. Elliot, and J. Sar- gatal, Editors). Lynx Edicions, Barcelona, Spain. SHORT COMMUNICATIONS 217 The Wilson Journal of Ornithology 120(1):2 17-221, 2008 Anvil Use by the Red-cockaded Woodpecker Kristin J. Bondo,‘'^ Lauren N. Gilson,' and Reed Bowman' " ABSTRACT — We observed Red-cockaded Wood- peckers (Picoides borealis) wedging longleaf pine {Pi- nus palustris) seeds into crevices in tree bark in Polk County, Florida from October to December 2004. Five individuals, four uniquely color-banded and one un- identified individual, wedged —14 seeds. Initially, we thought the birds were caching the seeds. Additional observations indicated the seeds were being wedged under the ends of the flaking bark of longleaf pines so seeds could be held firm and opened for consumption. Anvil use, where items are wedged for subsequent ma- nipulation, is known to occur in several avian taxa, but most notably in Piciformes, Corvidae, Passerida, and Sittidae. At least 16 woodpeckers worldwide have been reported using anvils. This is the first report of Red-cockaded Woodpeckers using longleaf pine bark as an anvil to facilitate extracting the seed. Received 29 January 2007. Accepted 22 May 2007. Many species of woodpeckers wedge food items too large or hard to consume whole in tree forks, crevices, or similar structures so they can be hammered or broken into smaller pieces for consumption (Winkler et al. 1995). These sites are referred to as “anvils”. Anvils are classified into three types: (1) occasional anvils — any hard surface (such as a rock or flat surface) where a food item is hammered, (2) proto-anvils — any natural crevice where a food item is wedged and held firm for ham- mering, and (3) true anvils — anvils created by the bird in vertical branches or trunks to hold food items for consumption (Cramp 1985). Anvils also can be classified by their frequen- cy of use: ancillary anvils are usually used opportunistically and only once, whereas main anvils are used regularly and frequently (Ked- ra and Mazgajski 2001). Anvil use has been described for over 50 bird species in 12 different taxa (Sibson 1974, ' Archbold Biological Station, P. O. Box 2057, Lake Placid, FL 33862, USA. ^ Current address: University of Regina, Department of Biology, Regina, SK S4S 0A2, Canada. ^Corresponding author; e-mail: rbowman@archbold-station.org Winkler et al. 1995, Lefebvre et al. 2002); oc- casional and proto-anvil use appears to occur most frequently in Piciformes, Corvidae, Pas- serida, and Sittidae. Anvils often are reused in species specializing in one abundant food source (Cramp 1985). The Great Spotted Woodpecker (Dendrocopos major) consumes primarily conifer seeds in winter (Cramp 1985) and depends on anvils to extract seeds from thousands of cones (Winkler et al. 1995). Corvids feeding extensively on pine seeds, in- cluding the Pinyon Jay {Gymnorhinus cyano- cephalus) (Baida 2002) and Clark’s Nutcrack- er (Nucifraga Columbiana) (Tomback 1998), wedge entire cones into anvils to extract the seeds. The Pinyon Jay often re-uses anvils (Baida 2002), but these are proto-anvils and not true anvils as used by the Great Spotted Woodpecker. Proto-anvils are used to open individual winged pine seeds for consumption. Winged pine seeds are similar in appearance to the sa- maras of ash {Fraxinus) and maple {Acer). The seed is on one end and a thin papery wing encases the seed and extends beyond the seed. White-headed Woodpeckers {Picoides albo- larvatus) (Garrett et al. 1996), Brown-headed Nuthatches {Sitta pusilla) (Morse 1967, 1968), White-breasted Nuthatches {S. carolinensis) (Morse 1968), Pygmy Nuthatches {S. pyg- maea) (Norris 1958), and Pine Warblers {Den- droica pinus) (Morse 1967) wedge winged pine seeds into bark or a crevice to extract the seed. Hairy Woodpeckers {Picoides villosus) consume winged pine seeds by wedging them into ponderosa pine {Pinus ponderosa) cones and using the cones as anvils to extract the seeds (Stallcup 1969). Conner et al. (2001 ) described the seed han- dling behavior of the Red-cockaded Wood- pecker {Picoides borealis). The seed is placed between the trunk and the bird’s breast before the seed wing is detached and the seed eaten; however, neither Conner et al. (2001 ) nor pre- vious authors mention use of anvils. The ob- 218 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 1, March 2008 jective of this paper is to describe anvil use by Red-cockaded Woodpeckers. OBSERVATIONS Observations were made of five individual birds in two different family groups. Both ter- ritories occurred in mesic longleaf pine {Pinus palustris) forests in Polk County, south-cen- tral Florida at the Avon Park Air Force Range (APAFR) (27° 38' N, 81° 17' W). As part of a long-term study, all individuals in the popu- lation are uniquely color-banded and the gen- der, social, and nesting status of all birds are known (Bowman et al. 1999, 2004). We con- ducted observations in the mornings and eve- nings while monitoring the population throughout September 2004-March 2005. We (KJB) observed handling and consumption of pine seeds from October to December. The breeding female from one family group, consisting of a male and female breed- ing pair, emerged from her roost at 0740 hrs EST on 24 October, followed shortly by her mate which was roosting nearby. One bird of the pair, whose identity was unconfirmed be- cause of the similarity of the pair’s color-band combinations, landed on the underside of a longleaf pine cone. It removed a winged seed from the opened pine cone, flew to the trunk of the same tree, and ascended ~2 m by cir- cling the bole to a total height of 5-7 m. It then wedged the winged seed vertically be- hind loose bark and began to peck. While this bird was handling its seed, its mate flew to the same cone, removed another winged seed, flew and climbed to a position above and op- posite its mate on the same tree, and per- formed the same behavior. Following wedging and pecking at each seed, both birds returned to this same cone for subsequent seeds even though the tree bore many cones. In each in- stance the woodpeckers landed on the cone, one at a time, returned to the tree bole wedg- ing each seed vertically behind loose bark and followed the seed insertion behavior by peck- ing. In total, the two birds removed and han- dled seven seeds in the same manner. Initially we thought the birds were caching seeds because after the birds pecked, presum- ably to remove the seed from the wing, we did not observe any seed wings fall to the ground. In addition, the birds did not imme- diately wedge the seed, but circled up and around the bole, presumably searching for a suitable place to cache the seed. We later ob- served additional seed-wedging behavior by other Red-cockaded Woodpeckers and con- firmed that seeds were being wedged to aid in handling and consumption. The bark was used as an “anvil” to facilitate the detachment of the non-edible wing from the seed. We encountered a female fledgling in an- other family group on 30 November at 0719 hrs EST perched on a longleaf pine cone ~2 m above ground. The cone was on a partially uprooted but still living longleaf pine tree that had recently been blown over. She removed a winged seed from the cone, flew to the trunk of another longleaf pine and carried the seed to the side opposite from where we were standing, -1.5 m above ground. Before we could again observe the bird, she had flown away. We found three longleaf pine seeds wedged firmly under the bark on the trunk of the tree where we saw her take the seed. In all cases, the seed had been removed and the remaining portion wedged vertically in the bark with the end formerly containing the seed facing up and the wing facing down. We also observed the breeding male at this territory handle four longleaf pine seeds on two different occasions between early Novem- ber and early December. Each time, the male removed the seed from a cone, flew with it to the upper trunk or a vertical branch where the bark was visibly peeling, wedged it behind bark, and pecked. We observed a bird from this same territory on 4 November, whose col- or-band combination was not observed, han- dle two pine seeds in the same manner. Each bird returned to obtain more seeds from the same cone in all instances. We concluded the seeds had been consumed and this behavior represented “anvil” use rather than caching. DISCUSSION Occasional and proto-anvil use has been re- corded in many avian taxa worldwide (Lefe- brve et al. 2002), and we summarized use of proto-anvils in 16 species of woodpeckers worldwide (Table 1). Only one woodpecker species, the Great Spotted Woodpecker, uses true anvils and frequently re-uses them (Wink- ler et al. 1995). Our observations of Red- cockaded Woodpeckers in Florida suggest use of proto-anvils rather than true anvils. The SHORT COMMUNICATIONS 219 TABLE 1. Woodpeckers (Picidae) documented to use proto-anvils. Taxa Species Anvil location Food items Melanerpes lewis^ Lewis’s Cleft in snag or power pole Acorns, nuts M. erythrocephalus Red-headed Cleft in branch, pow- er pole‘s Acorns,*’ '^ insects” M. formicivorus^'^ Acorn Cleft in limb Acorns M. flavifrons^ Yellow-fronted Base of broken limb Insects, fruits M. hypopolius^ Gray-breasted Clefts, excavated holes Insects, fruits M. carolinus Red-bellied Cleft in bark,' tree'-'^ or posb Seeds,' j nuts'" Sphyrapicus varius' Yellow-bellied Sapsucker Cleft in bark Seeds Dendrocopos minor"' Lesser Spotted Cleft in bark Cones, nuts D. medius'"' Middle Spotted Cleft in bark Cones, seeds D. leucotos"' White-backed Cleft in tree Nuts D. major Great Spotted Cleft in tree,"^ poles” Cones,"’” fruit,"’ nuts,"’ insects,"’ nestlings’" D. syriacus Syrian Cleft in wall,p tree,"^ or branch^! Fruit stones,?"’ walnuts'" Picoides pubescens'^ Downy Cleft in bark, knotholes Seeds P. villosus Hairy Cleft in bark,' cones' Seeds’’ P. albolarvatus^ White-headed Cleft in bark Seeds Picus viridis^ European Green Cleft Cones ^Bock (1970), ‘’Kilham (1958), ‘^Rogers et al. (1979), Moskovits (1978), ^ MacRoberts (1970), ^ MacRoberts and MacRoberts (1976), § Oniki and Willis (1998), h Leonard (2000), ‘ Erlwein (1996), J Mueller (1971), ^ Shackelford et al. (2000), ' Labedz (1980), Cramp (1985), " Tracy (1924), ° Kedra and Mazgajski (2001), P Gorman (1998), ‘5 Bunyard (1924), ■■ Davis (1995), * Garrett et al. (1996), ‘ Stallcup (1969). woodpeckers did not excavate the anvil site nor did they re-use them. Anvil use appears to be relatively common among European woodpeckers (Winkler et al. 1995), but their use seems less common among North American woodpeckers. Wheth- er this is a real behavioral difference, such as differences in diet, differences in forest struc- ture that might reduce the availability of suit- able anvil sites, or simply an artifact of ob- servational or reporting biases is unknown. Despite numerous studies on Red-cockaded Woodpeckers, this behavior has not been pre- viously reported. We do not know if this be- havior is widespread and unreported, or lo- calized to Florida or even this population. Avon Park Air Force Range is near the south- ern limit of longleaf pines and trees may be edaphically limited in size. We often observed Red-cockaded Woodpeckers foraging within 10-12 m of the ground but elsewhere, forag- ing heights might be greater and anvil use more difficult to observe. In foraging studies of Red-cockaded Woodpeckers in longleaf pine forests in Texas, woodpeckers appeared to hold seeds loosely between their breast and the bark while consuming them (R. N. Conner, pers. comm.). This might have been a misin- terpretation of anvil use because, to our knowledge, neither woodpeckers nor nut- hatches are known to hold food items between their breast and an object to facilitate con- sumption. The Great Spotted Woodpecker will hold food between breast and substrate before wedging it (Winkler et al. 1995) and most woodpeckers, particularly the Red-bellied Woodpecker (Melanerpes carolinus), will hunch their shoulders forward and push their belly against the bark when hammering food items so that fallen portions of their food can be trapped (Kilham 1983). Red-cockaded Woodpeckers may place seeds between breast and substrate to help position it before it is inserted vertically into the bark. Anvil use may occur only occasionally even in Florida. The majority of the Red- cockaded Woodpecker diet consists of small arthropod prey (Jackson 1994), which are probably sufficiently small to be consumed without the use of anvils. Inclusion of winged pine seeds in the diet of Red-cockaded Wood- peckers may vary greatly, even when pine mast is plentiful (Hooper and Lennartz 1981), suggesting seed consumption may depend on 220 THE WILSON JOURNAL OF ORNITHOLOGY • VoL 120, No. 1, March 2008 the relative availability of other foods. Use of anvils in birds may be an evolutionary pre- cursor to caching (Richards 1958, Winkler et al. 1995), which may be adaptive if food re- sources vary spatially and/or temporally (Smith and Reichman 1984). Red-cockaded Woodpeckers are not known to cache food (Jackson 1994), except perhaps bone frag- ments (Repasky et al. 1991). Caching occurs in other woodpeckers, most notably among the Melanerpes species, but is not widely re- ported in the genus Picoides. Although the Hairy Woodpecker may cache rarely, it has a relative hippocampus volume similar to the scatter-hoarding Red-bellied Woodpecker (Volman et al. 1997). Some woodpecker species use proto-anvils to break apart large or difficult to consume food items (Winkler et al. 1995) and it is like- ly Red-cockaded Woodpeckers use anvils in the same manner. Caching by Red-cockaded and other woodpeckers (Winkler et al. 1995) should be re-evaluated because anvil use and caching are similar behaviors that can be dif- ficult to distinguish (Winkler et al. 1995). Fu- ture research should address the frequency of anvil use and re-use among woodpeckers, and distinguish those observations from caching. The evolutionary relationship between anvil use and caching within woodpeckers and, spe- cifically within the genus Picoides, warrants further study. ACKNOWLEDGMENTS We thank Archbold Biological Station and Avon Park Air Force Range for funding and support. We also thank the University of Regina library, J. B. Dunning, and H. P. Weeks for help obtaining inter-library loans, and R. J. Fisher, R. N. Conner, and D. C. Rudolph for helpful comments improving the manuscript. We are indebted to the O’ Hanlons for providing hospitality and moral support, and C. Rook-Hobbs, who expedited the manuscript’s completion. LITERATURE CITED Balda, R. P. 2002. Pinyon Jay {Gymnorhinus cyano- cephalus). The birds of North America. Number 605. Bock, C. E. 1970. The ecology and behavior of the Lewis’s Woodpecker {Asyndesmus lewis). Univer- sity of California Publications in Zoology 92:1- 100. Bowman, R., D. L. Leonard Jr., L. K. Backus, and A. R. Mains. 1999. Interspecific interactions with foraging Red-cockaded Woodpeckers in south- central Florida. Wilson Bulletin 111:346—353. Bowman, R., D. L. Leonard Jr., D. Swan, and D. SCHWALM. 2004. Demography and population trends of a small Red-cockaded Woodpecker pop- ulation in south-central Florida. Pages 187-197 in Red-cockaded Woodpecker: road to recovery (R. Costa and S. J. Daniels, Editors). Hancock House Publishers, Blaine, Washington, USA. Bunyard, P. F. 1924. Surry field notes. British Birds 17:202. Conner, R. N., D. C. Rudolph, and J. R. Walters. 2001. Red-cockaded Woodpecker. Surviving in a fire maintained ecosystem. University of Texas Press, Austin, USA. Cramp, S. 1985. The birds of the Western Palearctic. Volume 4. Oxford University Press, New York, USA. Davis Jr., W. E. 1995. Downy Woodpecker and White-breasted Nuthatch use “vice” to open sun- flower seeds: is this an example of tool use? Bird Observer 23:339-342. Erlwein, K. M. 1996. Hairy and Red-bellied wood- peckers use bark crevice to break open seeds. Kingbird 46:200—201. Garrett, K. L., M. G. Raphael, and R. D. Dixon. 1996. White-headed Woodpecker {Picoides albo- larvatus). The birds of North America. Number 252. Gorman, G. 1998. Syrian Woodpecker using wall crevice as “anvil”. British Birds 91:378. Hooper, R. G. and M. R. Lennartz. 1981. Foraging behavior of the Red-cockaded Woodpecker in South Carolina. Auk 98:321—334. Jackson, J. A. 1994. Red-cockaded Woodpecker {Pi- coides borealis). The birds of North America. Number 85. Kedra, a. H. and T D. Mazgajski. 2001. Factors af- fecting anvil utilization by Great Spotted Wood- pecker Dendrocopos major. Polish Journal of Ecology 49:79-86. Kilham, L. 1958. Sealed-in winter stores of Red-head- ed Woodpeckers. Wilson Bulletin 70:107—113. Kilham, L. 1983. Life history studies of woodpeckers of eastern North America. Publications of the Nut- tall Ornithological Club Number 20. Labedz, T. E. 1980. Yellow-bellied Sapsucker feeding on hackberry seeds. Nebraska Bird Review 48:89. Leeebvre, L., N. Nicolakakis, and D. Boire. 2002. Tools and brains in birds. Behaviour 139:939- 973. Leonard Jr., D. L. 2000. Breeding and life history observations of the Gray-breasted Woodpecker {Melanerpes hypopolius). Ornitologfa Neotropical 11:341-348. MacRoberts, M. H. 1970. Notes on the food habits and food defense of the Acorn Woodpecker. Con- dor 72:196-204. MacRoberts, M. H. and B. R. MacRoberts. 1976. Social organization and behavior of the Acorn SHORT COMMUNICATIONS 221 Woodpecker in central coastal California. Ornitho- logical Monographs 21:1-115. Morse, D. H. 1967. Foraging relationships of Brown- headed Nuthatches and Pine Warblers. Ecology 48:94-103. Morse, D. H. 1968. The use of tools by Brown-headed Nuthatches. Wilson Bulletin 80:220-224. Moskovits, D. 1978. Winter territorial and foraging behavior of Red-headed Woodpeckers in Florida. Wilson Bulletin 90:521-535. Mueller, H. C. 1971. Sunflower seed carrying by Red-bellied Woodpeckers {Centurus carolinus). Bird-Banding 42:46-47. Norris, R. A. 1958. Comparative biosystematics and life history of the nuthatches Sitta pygmaea and Sitta pusilla. University of California Publications in Zoology 56:119-300. Oniki, Y. and E. O. Willis. 1998. Nesting of Yellow- fronted Woodpeckers Melanerpes flavifrons (Pi- cidae). Ornitologia Neotropical 9:81-86. Repasky, R. R., R. J. Blue, and P. D. Doerr. 1991. Laying Red-cockaded Woodpeckers cache bone fragments. Condor 93:458-461. Richards, T. J. 1958. Concealment and recovery of food by birds, with some relevant observations on squirrels. British Birds 51:497-508. Rogers, D. T, J. A. Jackson, B. J. Schardien, and M. S. Rogers. 1979. Observations at a nest of a partial albino Red-headed Woodpecker. Auk 96: 206-207. Shackelford, C. E., R. E. Brown, and R. N. Conner. 2000. Red-bellied Woodpecker {Melanerpes car- olinus). The birds of North America. Number 500. SiBSON, R. B. 1974. Rock Wren using an anvil. No- tornis 21:305. Smith, C. C. and O. J. Reichman. 1984. The evolution of food caching by birds and mammals. Annual Review of Ecology and Systematics 15:329-351. Stallcup, P. L. 1969. Hairy Woodpeckers feeding on pine seeds. Auk 86:134-135. Tomback, D. F. 1998. Clark’s Nutcracker (Nucifraga Columbiana). The birds of North America. Num- ber 331. Tracy, N. 1924. Woodpeckers and pine-cones. British Birds 17:276-279. VoLMAN, S. E, T. C. Grubb Jr., and K. C. Schuett. 1997. Relative hippocampal volume in relation to food-storing behavior in four species of wood- pecker. Brain Behavioral Evolution 409:110-120. Winkler, H., D. A. Christie, and D. Nurney. 1995. Woodpeckers. An identification guide to the woodpeckers of the world. Houghton Mifflin Company, New York, USA. The Wilson Journal of Ornithology 120(l):221-225, 2008 Gender Identification of Grasshopper Sparrows Comparing Behavioral, Morphological, and Molecular Techniques Frank K. Ammer,^ '^’^ Petra Bohall Wood,^ and Roger J. McPherson^ ABSTRACT. — Correct gender identification in monomorphic species is often difficult especially if males and females do not display obvious behavioral and breeding differences. We compared gender specific ’ West Virginia Cooperative Fish and Wildlife Re- search Unit, Division of Forestry and Natural Resourc- es, West Virginia University, R O. Box 6125, Morgan- town, WV 26506, USA. 2 U.S. Geological Survey, West Virginia Coopera- tive Fish and Wildlife Research Unit, Division of For- estry and Natural Resources, West Virginia University, R O. Box 6125, Morgantown, WV 26506, USA. Department of Biology, Clarion University of Rennsylvania, Clarion, RA 16214, USA. Current address: Department of Biology, Frostburg State University, 101 Braddock Road, Frostburg, MD 21532, USA. ‘^Corresponding author; e-mail: fammer@frostburg.edu morphology and behavior with recently developed DNA techniques for gender identification in the mono- morphic Grasshopper Sparrow {Ammodramus .savan- narum). Gender was ascertained with DNA in 213 in- dividuals using the 2550F/2718R primer set and 3% agarose gel electrophoresis. Field observations using behavior and breeding characteristics to identify gen- der matched DNA analyses with 100% accuracy for adult males and females. Gender was identified with DNA for all captured juveniles that did not display gender specific traits or behaviors in the field. The mo- lecular techniques used offered a high level of accu- racy and may be useful in studies of dispersal mech- anisms and winter assemblage composition in mono- morphic species. Received H January 2007. Accepted ! 7 June 2007. Gender identification of adult and juvenile birds is often difficult in demographic studies 222 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 DMA fragment size 600 bp - Z 430 bp - W FIG I Gender typing of Grasshopper Sparrows using PCR methods; 3% agarose gel showing gene products amplified with the Fridolfsson and Ellegren (1999) 2550F/2718R primers. Lanes 1, 3, and 5 represent the typical 600 bp and 430 bp band patterns of females. Lanes 2 and 4 represent the typical 600 bp band pattern of males. The size standard lane represents a 100 bp DNA ladder. examining species with subtle or no dimor- phic characters. Gender can be reliably ascer- tained only during the breeding season when behavioral and external sex characteristics can be directly observed in species lacking size or plumage dimorphism (Underwood et al. 2002). Identifying the gender of juveniles pre- sents additional difficulties as these individu- als display no gender related traits or behav- iors. The ability to assign gender is especially useful when examining sex-specific factors such as juvenile sex ratios and gender specific habitat selection (Underwood et al. 2002). The most reliable methods to identify gen- der until recently relied on invasive surgical procedures that require anesthesia and are not a feasible option for most field studies. The introduction of polymerase chain reaction (PCR) techniques targeting the chromo-heli- case DNA (CHD) binding gene (Griffiths et al. 1998, Fridolfsson and Ellegren 2000) pro- vides a noninvasive alternative. The avian CHD-W and CHD-Z genes occur on different sex chromosomes and gender can be ascer- tained by examining the composition of these genes (Boutette et al. 2002). Recently devel- oped techniques rely on PCR primers that an- neal to specific conserved exonic regions and amplify across an intron in both genes (Grif- fiths et al. 1998, Fridolfsson and Ellegren 1999, Dawson et al. 2001). This technique is based on noncoding introns in the CHD-Z and CHD-W genes which vary in length and there- fore can be differentiated by gel electropho- resis (Griffiths et al. 1998). The Grasshopper Sparrow (Ammodramus sa- vannarum) is a widely distributed North Amer- ican grassland species that has shown long-term population declines over much of its range (Brennan and Kuvlesky 2005). Individuals of the eastern subspecies (A. 5. pratensis) display no plumage or size dimorphism making it ex- tremely difficult to identify gender by observa- tion alone. Further, adult females in breeding territories are secretive and commonly spend substantial time on the ground walking to and from the nest area, making it difficult to observe female-specific brooding and feeding behaviors (Vickery 1996). Territorial male behaviors are more obvious because they commonly perch and sing at the periphery of their territory (Vick- ery 1996). Some morphological and behavioral characteristics may be useful in gender identi- fication in this species; however, quantitative ev- idence supporting the accuracy and applicability of these characters is lacking (Delany et al. 1994). We examined the effectiveness of the 2550F/2718R (Fridolfsson and Ellegren 1999) primer set for gender identification of adult and juvenile Grasshopper Sparrows. We then used molecular gender results to compare with morphological characters and breeding behav- ior to examine their efficacy for accurate gen- der assessment in the field. SHORT COMMUNICATIONS 223 q q q 1) 00 00 00 00 00 c 1 1 1 1 (S q in d rn d d B yj U •S W c/o >n cn cn NO CJ G oi c (N o o (U d d d d S ‘■3 = ON ON * * NO O c 1 m o q q cd d d 00 d S 'c ro 'Ct 00 in d d d d 1 1 1 1 II * B •n rn q d q q d in (N E 8 in On cn in CN ■3 W on (N o o cd $ d d d d 'c CQ '5b * * in 00 cn 1 (N n q yj yj W) c (N d d cd u. a * * 3 in m q q o M d d ON yj o in ■G- NO in t? u q O q q o 00 d d d cd Vii 04 PJ 1 1 1 1 cd x: cS q o O in, CJ — < o^ in d "g «) o '5b i o o P nO uu 00 NO O 04 o r~ d d d d & 0 c Ti r-; q q ro s 'It d d d ^ — LU -t in -J — rp p, CQ < _0J 1> -3 C "c .o G E (u O ‘da»°n Wing chord bill width, and bill length were included as criterion vanables m the discriminant analysis. Values are Lmber of individuals (n) and the percent (%) classified correctly into each gender. Juvenile Predicted female Predicted male Actual gender Predicted female Predicted male Actual gender Female Male 27 87.1 29 22.5 4 12.9 100 77.5 31 129 11 73.3 11 32.4 4 26.7 23 67.6 15 34 of Taq polymerase (New England BioLabs Inc., Ipswich, MA, USA). The thermal profile began with an initial de- nature at 94° C for 2 min. A touchdown pro- cedure was performed as follows: (1) preheat 94° C for 2 min; (2) 10 cycles of denaturation, 94° C for 45 sec, annealing 60° C-51° C (- 1° C per cycle) for 45 sec, extension, 72° C for 45 sec; (3) 30 cycles denaturation, 94° C for 30 sec, annealing, 50° C for 30 sec, extension, 72° C for 30 sec; and (4) extend at 72° C for 10 min and stored at 4° C. All thermal cycling was performed in a DNA Engine (PTC 200, Bio-Rad Laboratories Inc. Hercules, CA, USA) configured with a heated lid. PCR prod- ucts were visualized on 3% agarose gels stained with ethidium bromide (Fig. 1). Fe- males had two bands (ZW) while males had one (ZZ). Random samples were electropho- resed on 6% polyacrylamide gel stained with ethidium bromide to ensure adequate resolu- tion was obtained with agarose. Statistical Analysis.— Analyses were con- ducted using the SAS software package (SAS Institute Inc. 1999). Morphological characters were tested independently for differences be- tween adult males and females, and between juvenile males and females with analysis of variance (ANOVA) (Zar 1996). Mass, wing chord, bill width, bill length, and tarsus length were the dependent variables in the models. Differences were considered significant at a = 0.10. Morphological characters with P ^ 0.10 were included as criterion variables in parametric quadratic discriminant function analyses (DISCRIM) to assign adult and ju- venile individuals into gender classes (SAS Institute Inc. 1999). Error rates were based on cross validation procedures. RESULTS We identified gender using molecular tech- niques for all 164 adults sampled. Gender as- signments based on data collected in the field using breeding characteristics and behaviors were 100% consistent with DNA analyses. Gender assignment using morphological data was less consistent. For adults, wing chord = 40.5, P < 0.001), bill width (Fi,„5 = 32.2, P < 0.001), and bill length (Fi 152 = 2.80, P = 0.10) differed by gender, while mass (Fi,53 = 1-73, P = 0.19) and tarsus length (f’u5i = 0.19, P = 0.67) did not (Table 1). Discriminant function analysis with bill width, bill length, and wing chord correctly classified 87.1% of adult females and 77.5% of adult males to the gender groups previously ascertained with molecular and behavioral techniques (Table 2). We successfully assigned gender using mo- lecular techniques to all 49 juveniles/non- breeders sampled in 2002. These 34 males and 15 females did not display sex related traits or behaviors in the field and gender could not be assigned based on behavioral data. No morphological differences were detected be- tween genders in juveniles with univariate analyses (mass: F142 1.38, df = I, P — 0.25, wing chord: F,42 = 2.06, df = I, P = 0.16, bill width: F,39 = 0.05, df = 1, F = 0.83, bill length: F^,42 = 0.37, df = 1, F = 0.55, and tarsus length: F142 = 0.01, df = ^ 7 0.93). However, discriminant function analysis using bill width, bill length, and wing chord as cri- terion variables correctly assigned 73.3% of females and 67.6% of males (Table 2). DISCUSSION Molecular gender assignment for adult Grasshopper Sparrows using methods of Fri- SHORT COMMUNICATIONS 225 dolfsson and Ellegren (1999) was in 100% agreement with behavioral data collected in the field. Gender assignment with this tech- nique was an excellent alternative to more in- vasive procedures normally required to assign gender in juveniles. The gender of adults dis- playing territorial behavior and external char- acters was reliably identified in all individuals examined in the field. However, if gender dis- crimination is not conclusive with breeding characters and behaviors, we suggest using adult wing chord, bill width, and bill length to assist gender identification in the field. These three measurements were 87% accurate in assigning females and 78% accurate for as- signing males. Gender identification of juveniles not dis- playing secondary sexual characteristics and behaviors remains a challenge for researchers in the field. Discrimination of juvenile gender using morphological characters was less ac- curate than for adults in breeding condition. Significant differences in morphometries be- tween juveniles and adults were not surprising because a large percentage (86%) of the ju- veniles sampled were 5-6 day old nestlings and fledglings not yet fully developed. The molecular techniques used were efficient, relatively inexpensive, and less invasive than traditional procedures. These methods do not re- quire extensive training and may be performed in many modestly equipped laboratories. Data obtained with these techniques may be instru- mental in providing answers to questions tar- geting dispersal mechanisms and winter assem- blage composition in monomorphic species. ACKNOWLEDGMENTS We thank those individuals who provided assistance with data collection and analyses, especially George Seidel, Joshua Scullen, Danielle Mortensen, Scott Bos- worth, Amanda Carroll, Nick Ammer, and Amanda Ammer. The Environmental Protection Agency, West Virginia Division of Natural Resources, U.S. Geolog- ical Survey, West Virginia University Research Cor- poration, WVU Division of Forestry, and the USGS West Virginia Cooperative Fish and Wildlife Research Unit provided financial, administrative, and logistic support. Arch Coal Inc., Cannelton Mine, Amherst Corporation, and the Ark Land Company provided lo- gistic support, access to field sites, and field housing. Gregory Forcey, Michael Capp, Dusty Perkins, and an anonymous reviewer provided valuable comments on earlier versions of this manuscript. LITERATURE CITED Boutette, J. B., E. C. Ramsay, L. N. D. Potgieter, AND S. A. Kama. 2002. An improved polymerase chain reaction-restriction fragment length poly- morphism assay for gender identification in birds. Journal of Avian Medicine and Surgery 16:198- 202. Brennan, L. A. and W. P. Kuvlesky. 2005. North American grassland birds: an unfolding crisis? Journal of Wildlife Management 69:1-13. Dawson, D. A., S. Darby, F M. Hunter, A. P Krupa, I. L. Jones, and T. Burke. 2001. A critique of avian CHD-based molecular sexing protocols illustrated by a Z-chromosome polymorphism detected in auklets. Molecular Ecology Notes 1:201-204. Delany, M. F, C. T. Moore, and D. R. Progulske Jr. 1994. Distinguishing gender of Florida Grass- hopper Sparrows using body measurements. Flor- ida Field Naturalist 22:48-51. Fridolfsson, a. K. and H. Ellegren. 1999. A simple and universal method for molecular sexing of non- ratite birds. Journal of Avian Biology 30:1 16-121. Fridolfsson, A. K. and H. Ellegren. 2000. Molecular evolution of the avian CHDl genes on the Z and W sex chromosomes. Genetics 155:1903-1912. Griffiths, R., M. C. Double, K. Orr, and R. J. G. Dawson. 1998. A DNA test to sex most birds. Molecular Ecology 7:1071-1075. Pyle, P, S. N. G. Howell, D. F. DeSante, R. P. Yu- NiCK, AND M. Gustafson. 1997. Identification guide to North American birds. Part 1. Columbi- dae to Ploceidae. Slate Creek Press, Bolinas, Cal- ifornia, USA. SAS Institute Inc. 1999. SAS/STAT user’s guide. SAS Institute Inc., Cary, North Carolina, USA. Seutin, G., B. N. White, and P. T. Boag. 1990. Pres- ervation of avian blood and tissue samples for DNA analyses. Canadian Journal of Zoology 69: 82-90. Underwood, R. M., R. J. Crockett, R. R. Roth, C. L. Keeler Jr., and M. S. Parcells. 2002. A com- parison of flow cytrometry and the polymerase chain reaction as sexing techniques for the Wood Thrush. Journal of Field Ornithology 73:239-245. Vickery, P. D. 1996. Grasshopper Sparrows {Ammo- dramiis savannarum). The birds of North Ameri- ca. Number 239. Zar, j. H. 1996. Biostatistical analysis. Third Pidition. Prentice Hall, Upper vSaddle River, New Jersey, USA. 226 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 The Wilson Journal of Ornithology I20(I):226-228, 2008 Abnormal Eggs of Rio Grande Wild Turkeys on the Edwards Plateau, Texas Kyle B. Melton,' Justin Z. Dreibelbis,' Ray Aguirre,^ Bret A. Collier,' '* T. Wayne Schwertner,^ Markus J. Peterson,' and Nova J. Silvy' ABSTRACT. — We studied the reproductive ecology of Rio Grande Wild Turkeys {Meleagris gallopavo in- termedia) in the Edwards Plateau region, Texas during 2005 and 2006. Runt eggs from a single adult female were observed through three nesting events over 2 years. Mean mass and volume for the runt eggs were 44% of normal Wild Turkey eggs. Production of runt eggs is common in domesticated gallinaceous birds, yet little information is available on runt egg produc- tion in wild gallinaceous birds. To our knowledge, our observations are the first which indicate runt egg pro- duction occurs in Wild Turkeys. Received 20 February 2007. Accepted 11 May 2007. Knowledge of reproductive rates is critical to monitoring long-term dynamics of Wild Turkey {Meleagris gallopavo) populations (Vangilder 1992). Reproductive rates are in- fluenced by multiple components of the repro- ductive process. Runt eggs, those having vol- umes <15% of the average (Koenig 1980), are perhaps the most common egg abnormality documented in domestic fowl (Pearl and Cur- tis 1916, Romanoff and Romanoff 1949). Sev- eral avian species, both domestic and wild, have been known to produce runt eggs (Her- nandez et al. 2006), but occurrence is low for most species; approximately one in every 1,000-2,000 eggs (Mallory et al. 2004, Her- nandez et al. 2006). Documentation of runt eggs in wild populations is rare (Rothstein 1973, Mallory et al. 2004). Our objective is to report what we believe is the first obser- vation of runt egg production by Wild Tur- keys. 1 Department of Wildlife and Fisheries Sciences, Texas A&M University, College Station, TX 77843, USA. 2 Texas Parks and Wildlife Department, Comfort, TX 78013, USA. 2 Texas Parks and Wildlife Department, Mason, TX 76856, USA. Corresponding author; e-mail: bret@tamu.edu OBSERVATIONS We tracked, via triangulation and homing (White and Garrott 1990), 11 radio-marked Wild Turkey hens through two nesting seasons (2005-2006) on a 984-ha ranch ~9.5-km north of Leakey, Texas, USA. One of these radio-marked hens produced runt eggs during three consecutive nesting attempts (2 in 2005, 1 in 2006). We captured this 3.7-kg hen on 24 February 2004 as an adult (>18 months of age). She produced a clutch of five eggs in 2004 with no abnormal appearing eggs. We first observed a set of runt eggs on 13 April 2005 during the hen’s initial 2005 nest- ing attempt. The first nest contained three runt eggs and one normal egg. Normal laying be- havior for turkeys is to lay one egg per day (Healy 1992). Daily checks when the hen was off the nest confirmed that no additional eggs were laid until 18 April when one additional runt egg was laid (total clutch; four runt eggs, one normal egg). ^Ve continued to monitor the nest daily from 18 to 27 April during which time we did not locate the hen on the nest although she was located in the nest area. We observed one additional runt egg, on 28 April, bringing the clutch size to six (five runt eggs and one normal egg). One additional runt egg was deposited between 28 April and 1 May bringing the total clutch to seven (six runt eggs and one normal egg). The normal-sized egg was depredated on 2 May and only six runt eggs remained. We considered the nest depredated on 4 May when the remaining eggs were found hidden under leaf litter (un- damaged) in separate locations away from the nest. We collected the runt eggs, measured size and volume, and ascertained if they were viable. We continued to radio-track the hen and documented a second nesting attempt on 30 May 2005 containing three runt eggs. During SHORT COMMUNICATIONS 227 TABLE 1. Characteristics of normal and runt Rio Grande Wild Turkey eggs on the Edwards Plateau, Texas, 2005-2006. n Min Max Mean SD Normal Length, mm 176 53.4 65.6 61.0 0.19 Width, mm 176 40.7 72.0 47.1 0.24 Weight, g 176 47.4 85.1 68.7 6.46 Volume, ml 161 27.5 55.0 43.6 4.62 Runt Length, mm 7 39.2 49.5 44.7 0.39 Width, mm 7 33.4 37.5 35.5 0.17 Weight, g 7 23.6 34.2 30.3 3.79 Volume, ml 4 16.5 22.0 19.3 3.18 monitoring of the renesting attempt, the hen abandoned the nest. Within 1 week she was located > 1 km from the nest and was not ob- served near the nest again. We collected the three runt eggs on 9 June 2005 to check via- bility and to obtain measurements. We docu- mented a third nest the following year by this hen on 19 April 2006, which contained 12 runt eggs and four normal eggs. We monitored the nest and hen daily through 14 days of in- cubation, and found the nest partly depredated on 3 May 2006. We collected shell remains from 10 depredated eggs (eight runt and two normal eggs). We located 90 nests from 69 individual hens during 2005-2006 and obtained clutch sizes for 70 nests {n = 885 eggs). Based on our data, runt eggs in Wild Turkeys occurred at a frequency of 2.4% (21/885). We measured length (mm), width (mm), mass (g), and vol- ume (ml) using water displacement for un- damaged runt eggs (/? == 7) and undamaged/ unhatched normal eggs {n = 176) collected during 2005-2006 (Table 1). Mean mass and volume of the runt eggs was 44% of normal eggs; 31% smaller than the suggested size for classifying eggs as runts (Koenig 1980). None of the runt eggs contained yolks, making them unviable. DISCUSSION Production of runt eggs is usually thought to be caused by a temporary disturbance to the reproductive system (Pearl and Curtis 1916, Romanoff and Romanoff 1949). More- over, birds under environmental stress may be more prone to produce runt eggs (Mallory et al. 2004). Wild Turkey reproduction and, therefore, egg production is negatively affect- ed by low rainfall and soil moisture (Beasom and Pattee 1980) as well as nutritional limi- tation (Blankenship 1992). Continued produc- tion of runt eggs suggests a congenital defect or permanent injury to the bird’s oviduct (Pearl and Curtis 1916, Mulvihill 1987). The frequency of runt eggs in Wild Turkeys is low; persisting environmental stresses presumably could alter their frequency and reduce the po- tential production of the population. However, the low prevalence of runt eggs in our study suggests the impact of runt egg production on population trajectory is probably limited. ACKNOWLEDGMENTS Funding was provided by the Texas Turkey Stamp Fund through the Texas Parks and Wildlife Depart- ment, the National Wild Turkey Federation Texas State Superfund, and the Department of Wildlife and Fish- eries Sciences, Texas A&M University. This research was conducted under Texas A&M University Animal Use Permit 2005-005. LITERATURE CITED Beasom, S. L. and O. H. Pattee. 1980. The effect of selected climatic variables on Wild Turkey pro- ductivity. Proceedings of the National Wild Tur- key Symposium 4:127-135. Blankenship, L. H. 1992. Physiology. Pages 84-100 in The Wild Turkey: biology and management (J. G. Dickson, Editor). Stackpole Books, Harrisburg, Pennsylvania, USA. Healy, W. M. 1992. Behavior. Pages 46-65 in The Wild Turkey: biology and management (J. G. Dickson, Editor). Stackpole Books, Harrisburg, Pennsylvania, USA. Hernandez, F, J. A. Arredondo, F. Hernandez, F. C. Bryant, and L. A. Brennan. 2006. Abnormal eggs and incubation behavior in Northern Bob- white. Wilson Journal of Ornithology 118:114- 1 16. Koenig, W. D. 1980. The determination of runt eggs in birds. Wilson Bulletin 92:103-107. Mallory, M. L., L. Kief, R. G. Clark, T. Bowman, P Blums, A. Mednis, and R. T. Alisauskas. 2004. The occurrence of runt eggs in waterfowl clutches. Journal of Field Ornithology 75:209- 217. Mulvihill, R. S. 1987. Runt eggs: a discovery, a syn- opsis, and a proposal for future study. North American Bird Bander 12:94-96. Pi;arl, R. and M. R. Ci'RTIS. 1916. Studies on the physiology of reproduction in the domestic fowl — XV. Dwarf eggs. Journal of Agricultural Research 6:977-1042. 228 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 Romanoff, A. L. and A. J. Romanoff. 1949. The avi- an egg. John Wiley and Sons, New York, USA. Rothstein, S. I. 1973. The occurrence of unusually small eggs in three species of songbirds. Wilson Bulletin 85:340-342. Vangilder, L. D. 1992. Population dynamics. Pages 144-164 in The Wild Turkey: biology and man- agement (J. G. Dickson, Editor). Stackpole Books, Harrisburg, Pennsylvania, USA. White, G. C. and R. A. Garrott. 1990. Analysis of wildlife radio-tracking data. Academic Press, San Diego, California, USA. The Wilson Journal of Ornithology 120(1 ):228-230, 2008 Barred Forest Falcon (Micrastur ruficollis) Predation on Relatively Large Prey Fabio Rdhe‘'’“ and Andre Pinassi Antunes^ ABSTRACT. — We describe three successful preda- tion events by the Barred Forest Falcon (Micrastur ruf- icollis) in the Atlantic Forest of coastal southeast Bra- zil. The prey items were a Plumbeous Pigeon (Pata- gioenas plumbea), a Brown Tinamou (Cryptiirellus ob- soletus), and a large toad (Chaunus ictericus). This is the first report of successful attacks on prey heavier than the forest falcon, of which none was successfully carried away. These large prey items represent a trade- off between high nutrient value and safety of carrying prey to a secure perch. Received 16 November 2005. Accepted 23 July 2006. The Barred Forest Falcon {Micrastur rufi- collis) is a small neotropical falconid weigh- ing an average of 168 g for males and 233 g for females (Thorstrom 2000). It lives in dense primary or secondary forest where it is known to feed on a variety of small verte- brates and large invertebrates (Sick 1993). Studies of its diet are few (del Hoyo et al. 1994), the most detailed of which (405 prey identified) found reptiles to be the most im- portant group, in terms of frequency, followed by birds. However, both were equal in bio- ' Coordenagao de Pesquisas em Ecologia (CPEC), INPA — Institute Nacional de Pesquisas da Amazonia. ^Laboratorio de Herpetologia, Departamento de Zoologia, Universidade Estadual Paulista — UNESP, Rio Claro, Sao Paulo, Brazil. 3 Current address: WCS— Wildlife Conservation So- ciety, Av. General Rodrigo Octavio Jordao Ramos, 3000, Japiim. Institute de Ciencias Biologicas, Depar- tamento de Biologia. Bloco Projecto Sauim-de-Colei- raAVeS. 4 Corresponding author; e-mail: fabiorohe@gmail. mass captured and delivered to nests during the breeding season (Thorstrom 2000). There is one example of predation on a humming- bird (Nunnery et al. 2002) and fruit consump- tion has been reported (Thorstrom 1996). Hil- ty and Brown (1986) reported the forest falcon is believed to specialize mostly on small birds, but are not specialists on them (Thorstrom 2000). Forest falcons are regularly captured in Amazonian forest mist nets attacking small birds already caught in the nets (Mario Cohn- Haft, pers. comm.). They also follow swanns of army ants at times, presumably hunting ei- ther fleeing insects or other birds that follow the ants (Willis et al. 1983). The largest prey described weighed 160 g, roughly the mini- mum weight of the predator (Thorstrom 2000). We describe three cases in the Brazil- ian Atlantic rainforest of M. ruficollis suc- cessfully taking prey larger than itself, and briefly discuss the implications for the forag- ing and breeding ecology of the species. The bird prey weights were obtained from Sick (1993, 1997). OBSERVATIONS All of our observations were by chance, during the day, in the course of other research in the Brazilian State of Sao Paulo. The three sightings possibly involve three different in- dividuals, based on the distances between sites of at least 3 km. The three observations were on dirt roads among patches of eucalyptus {Eucalyptus saligna) plantations and Atlantic forest on private farms (23° 55' S, 47° 41' W) com SHORT COMMUNICATIONS 229 at an altitude of 650-980 m in the Serra do Mar coastal mountain range. We observed a forest falcon in May 2001 on the ground eating a Plumbeous Pigeon (Patagioenas plumbea), the breast of which was already mostly consumed. The falcon flushed on our approach trying unsuccessfully to carry the prey, which it dropped after drag- ging it for 1 m. This pigeon has an average weight of 231 g. We observed another forest falcon in January 2002 on the ground eating a large toad (Chaunus ictericus). Adult males of this toad species, approximately the size we observed, weigh 200-250 g (A. P. Antunes, pers. obs.). The raptor was eating from the throat region, possibly avoiding the area of the paratoid glands. In August 2003 we encountered a falcon on the ground stripping the neck feathers from a live Brown Tinamou (Crypturellus obsoletus; weight 480 g). The forest falcon was grasping the tinamou ’s dorsum and flapping the wings slightly. The prey showed no external evi- dence of injury but, after the forest falcon re- leased it, the tinamou remained on its back, apparently unable to move. The raptor flushed upon our arrival trying to carry the prey, which it was unable to move. One additional observation occurred at ap- proximately sea level at Itamambuca Beach, Ubatuba Municipality, in October 2003. A Barred Forest Falcon landed on a perch 1 m above the ground and 1.5 m from a large ter- restrial Black-white Tegu {Tupinambis meri- anae), which we had been observing. This liz- ard was ~1 m in length (including tail) and weighed ~ 1.0-1. 5 kg (A. P. Antunes, pers. obs.). The hawk watched the lizard closely and flushed after noticing our presence. DISCUSSION The possibility of vehicle strikes was dis- carded in the predation events because: ( 1 ) the Plumbeous Pigeon predation site can not be accessed by car, (2) only our vehicle was al- lowed on the farm in the case of the tinamou, and (3) no signs of flattening of the toad were observed. Other reports of toad predation by hawks have included a White-tailed Hawk {Buteo al- bicaudatus) eating Just the legs of Chaunus marinus (Sick 1997) and a Red-tailed Hawk {B. jamaicensis) preying on Anaxyrus boreas (Jones and Stiles 2000). Our observations describe predation by Barred Forest Falcons on two different animal classes (birds and amphibians), reinforcing the importance of these groups in its diet. This is the first report of successful attacks by M. ruf- icollis on prey heavier than 160 g. Thorstrom (2000) documented maximum weight of prey based on estimated weight of items taken and delivered to nests by breeding forest falcons. Prey items must be sufficiently light for the adult to carry in flight. We documented cap- tured prey as much as twice the weight of a Barred Forest Falcon. None of these larger prey items could be carried by the forest fal- con suggesting the threshold weight for car- rying is roughly that of the falcon. Thorstrom (2000) reported the Collared Forest Falcon (Micrastur semitorquatus) (average body mass for females = 869 g) captured an Oc- ellated Turkey (Meleagris ocellata) weighing 3 kg which represented —3.5 times the body mass of the female forest falcon. This female fed and stayed near this kill for several days, and later carried pieces of the carcass to the nest where she was attending two nearly- fledged nestlings (Russell Thorstrom, pers. comm.). These observations suggest that predation on considerably larger prey may not be un- common when forest falcons are not associ- ated with nesting, when pieces of the carcass are carried to the nest. The risk of preying on animals that must be consumed in situ may be compensated by the obviously greater nutrient value these prey represent. ACKNOWLEDGMENTS We thank Mario Cohn-Haft and T M. Sanaiotti for valuable comments and help with revision of the man- uscript. We thank J. A. Sedgwick, R. O. Bierregaard, C. M. White, and Russell Thorstrom for review and suggestions that improved this paper. We thank Euca- tex S/A for logistic support and authorization to de- velop research projects in its area, and FAPESP (Pro- cesses #01/13341-3 and #04/10974-3) for financial support. LITERATURE CITED DEL Hoyo, J., a. Eliot, and J. Saragatal. 1994. Handbook of the birds of the world. Volume 2. New world vultures and guineafowl. Lynx Edi- ciones, Barcelona, Spain. Hilty, S. L. and W. L. Brown. 1986. A guide to the 230 THE WILSON JOURNAL OF ORNITHOLOGY • VoL 120, No. 1, March 2008 birds of Colombia. Princeton University Press, Princeton, New Jersey, USA. Jones, M. S. and B. Stiles. 2000. Bufo boreas (Boreal Toad). Predation. Herpetological Review 31:99. Nunnery, T, L. Linda, and M. R. Welford. 2002. Barred Forest-falcon (Micrastur ruficollis) preda- tion on a hummingbird. Journal of Raptor Re- search 36:239-240. Sick, H. 1993. Birds in Brazil — a natural history. Princeton University Press, Princeton, New Jer- sey, USA. Sick, H. 1997. Ornitologia Brasileira. Nova Fronteira. Rio de Janeiro, Brazil. Thorstrom, R. 1996. Fruit-eating behavior of a Barred Forest-falcon. Journal of Raptor Research 30:44. Thorstrom, R. 2000. The food habits of sympatric forest-falcons during the breeding season in north- eastern Guatemala. Journal of Raptor Research 34:196-202. Willis, E. O., D. Wechsler, and F. G. Stiles. 1983. Forest-falcons, hawks, and a pygmy-owl as ant fol- lowers. Revista Brasileira de Biologia 43:23-28. The Wilson Journal of Ornithology 120(l):231-237, 2008 Ornithological Literature Compiled by Mary Gustafson GEORGE MIKSCH SUTTON: ARTIST, SCIENTIST, AND TEACHER. By Jerome A. Jackson. University of Oklahoma Press, Nor- man, USA. 2007: 239 pages, 29 figs., and 27 color plates. ISBN: 978-0-8061-3745-2. $29.95 (cloth). — Sutton was one of the pre- mier bird illustrators and artists of the 20th century, and a major figure in North American ornithology. This book is a definitive biogra- phy of Sutton and nicely complements Sut- ton’s autobiographical travelogues and auto- biography of his early years {Bird Student, 1980, University of Texas Press, Austin). This story of the life and work of George Miksch Sutton draws upon Jackson’s numerous per- sonal recollections of Sutton as well as in- depth interviews with dozens of Sutton’s con- temporaries, colleagues, family, and friends (65 are listed in the Acknowledgments). It also draws upon the voluminous Sutton cor- respondence archived at the Universities of Oklahoma, Michigan, and Cornell, and the Smithsonian Archives in Washington, D.C. The book follows an essentially chronolog- ical format. The first chapter, “Family Ties,” traces Sutton’s ancestry back through two generations and emphasizes the closeness of the Sutton family as well as Sutton’s early in- terest in birds. The next three chapters trace in more detail Sutton’s growth and develop- ment as an artist, ornithologist, and person, from his birth in rural Bethany, Nebraska (Sutton’s father was ‘Professor of Eloquence’ at Conner College in Bethany, Nebraska), his early encounters with birds, mail-order taxi- dermy lessons, and the family’s move to Beth- any, West Virginia. We read of the publication of Sutton’s first articles in Bird-Lore, and his mentoring by Louis Agassiz Fuertes, to his entry into professional ornithology through working with W. Clyde Todd at the Carnegie Museum in Pittsburgh. Sutton’s interest in birds and art date back to his early years (he was selling bird drawings for a penny or so each by age seven). Chapters follow on Sut- ton’s role as Pennsylvania State Ornithologist, his expedition to Southampton Island in the Canadian Arctic, his doctoral work at Cornell University under the mentorship of Arthur A. Allen, his activities in the army during World War II, and his growing love of the Arctic, the Neotropics, and the southwestern United States that was to play a major role in his art and science. The chapter “The Michigan Years,” traces Sutton’s long relationship with Josselyn Van Tyne that included collaboration on research and the Wilson Ornithological Club (now Society) that led to Sutton’s full- time employment at the University of Michi- gan. Sutton’s complex relationship with Cor- nell University did not endure and he was pleased to move to the greener pastures of the University of Michigan. His relationship with Van Tyne, however, quickly deteriorated, and this led, in 1952, to a shift to the University of Oklahoma, where Sutton would spend the final 30 years of his life. A chapter entitled “Teacher, Conservationist, Philanthropist,” emphasizes these aspects of Sutton’s profes- sional life. “From Illustration to Fine Art” traces Sutton’s struggles and triumphs with bird illustration and bird art. The concluding chapter, “Ovenbird and Golden Eagle,” fol- low Sutton in his last days and includes Sutton poetry, demonstrating yet another facet of Sut- ton’s personality and accomplishments. An Appendix lists the many graduate students that Sutton mentored. A 26-page section of end notes documents the scholarly research that Jackson did for this portrait of George Miksch Sutton. A section, “Works by George Miksch Sutton,” presents an exhaustive bib- liography of Sutton’s published works that runs to more than 450 entries, and includes nearly a dozen books beginning in 1928 with An Introduction to the Birds of Pennsylvania (J. Horace, Harrisburg), followed in 1934 by Eskimo Year (McMillan, New York) and end- ing with the posthumous Birds Worth Watch- ing (1986, University of Oklahoma Press, Norman). A final section, “Works Illustrated by George Miksch Sutton,” lists 40 books and 231 232 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 monographs written by other authors but il- lustrated in part or entirely by Sutton. He, of course, illustrated many of his own written works. Jackson has captured Sutton’s personality in words, and the 27 color plates and more than a dozen line drawings illustrate Sutton’s growth as an artist and illustrator, and dem- onstrate the diversity of his artwork. This thoroughly researched and well-written book is a significant contribution to both the history of ornithology and the history of bird illustra- tion and art.— WILLIAM E. DAVIS JR., Pro- fessor Emeritus, Boston University, 23 Knoll- wood Drive, East Falmouth, MA 02536, USA, e-mail: wedavis@bu.edu IVORYBILL HUNTERS: THE SEARCH FOR PROOF IN A FLOODED WILDER- NESS. By Geoffrey E. Hill. Oxford Univer- sity Press, New York, USA. 2007: 260 pages, 74 black and white photographs, 4 maps, and 1 diagram. ISBN: 978-0-19-532346-7. S24.95 (cloth). — This is another in a series of books that have recently appeared about the Ivory- billed Woodpecker (Campephilus principalis) (e.g., T. Gallagher, 2005, The Grail Bird: Hot on the Trail of the Ivory-billed Woodpecker, Houghton Mifflin, Boston, MA, USA; J. Jack- son, 2006, In Search of the Ivory-billed Wood- pecker, Smithsonian Books, Washington, D.C., USA). Long thought by many to be ex- tinct in the United States (the last fully doc- umented sighting of ivorybills was in the 1930s), recent reports of sightings have rekin- dled interest and hope that this largest of the North American woodpeckers may still be ex- tant. In this book Hill recounts a search for ivorybills in northwest Florida and reports in an Appendix on 15 “. . . Observations of Ivo- ry-billed Woodpeckers” and 43 “. . . sound detections of Ivory-billed Woodpeckers” made between May 2005 to May 2006. Un- fortunately, no photographic evidence was gathered that would offer irrefutable proof that the ivorybill was the species that was seen and heard. The book is primarily a tale of the search for ivorybills. As the author explains in the Preface: ‘T focus primarily on the search and the searchers rather than the wood- pecker. After all, the search for and rediscov- ery of the Ivory-billed Woodpecker is a hu- man tale. The ivorybills of the Choctawhatch- ee River Wetlands in the panhandle of north- ern Florida eat their grubs, cut their cavities, and raise their young as they have always done.” Hill also emphasizes in the Preface that: “This book is story, not a technical re- port” and contains neither footnotes nor in- text citations. It does, however, craft argu- ments directed at convincing the reader that the ivorybills he reports on are real. After an introductory chapter that shares some 20th century ivorybill history, the sec- ond chapter: “A Most Improbable Discov- ery,” recounts the decision to kayak the Choc- tawhatchee River: “No trip could have been conceived more whimsically,” (page 20) and the unexpected hearing of “Loud knocking, and sighting of a bird that could have been an ivorybill: “it was big and mostly black with a lot of white flashing on the top and bottom of the back of its wings” (page 23) was re- markable. The third chapter: “From Possible to Virtual Certainty,” recounts the adventures of Hill and several of his graduate students in subsequent trips, this time sightings included two birds by one field assistant, and Hill was convinced there were really ivorybills there. Chapters follow on planning for a systematic search for the ivorybills, securing of limited funding for the project, the habitat and log- ging history of the Choctawhatchee River lands, and a series of adventures and misad- ventures by Hill and his associates as they searched for Ivory-billed Woodpeckers. In a chapter: “Good Science, Bad Science, or No Science At All” the author emphasizes that searching for Ivorybills is not science, and ex- plores the controversy around recent reported sightings and adds some historical perspec- tive. Another chapter highlights the searches made for tangible evidence that included sys- tematic searches for possible nest or roosting sites, feeding marks on trees, video and audio recordings, and presents the results as a cir- cumstantial case for the existence of ivorybills through spectrograms of “double knocks” and “kent calls” recorded during searches. It is abundantly clear the author is convinced that ivorybills are alive and well in his study area: “By June 2005, Tyler, Brian, and I knew that there was at least one ivorybill near the mouth of Bruce Creek in the Choctawhatchee ORNITHOLOGICAL LITERATURE 233 River bottom” (page 143), and in the Epi- logue: “The discovery of a population of ivo- rybills along the Choctawhatchee River in Florida is an exciting and fun birding event.” This book makes a case for this reality. How- ever, to date (including the 2007 field season) no irrefutable photographic evidence has been produced. An interesting parallel exists be- tween the ivorybills and the Tasmanian Tiger (Thylacinus cyanocephalus), a marsupial car- nivore. The last tiger in captivity died in 1936, but sightings have been reported ever since, some by qualified biologists. However, no road-killed tigers have been found nor has in- disputable photographic evidence been pro- duced. In the absence of irrefutable evidence, and with the intrinsic problems faced by rare animals, much of the scientific community is likely to remain skeptical of the continued ex- istence of either Ivory-billed Woodpeckers or Tasmanian Tigers. Hill’s book is written for a popular audience and is not a scientific work. In any event, it chronicles an interesting adventure story, and I recommend it as a good read. — WILLIAM E. DAVIS JR., Professor Emeritus, Boston University, 23 Knollwood Drive, East Fal- mouth, MA 02536, USA; e-mail: wedavis@ bu.edu HANDBOOK OF AVIAN HYBRIDS OF THE WORLD. By Eugene M. McCarthy. Ox- ford University Press, New York, USA. 2006: 583 pages and 20 figures. ISBN: 0-19- 518323-1. $89.50 (cloth). — Research on avian hybridization played a key role in the devel- opment of the New Synthesis (Mayr 1942), and remains at the forefront of modern evo- lutionary biology (Prager and Wilson 1975, Grant and Grant 1992, Price 2008). Given the importance of hybridization in evolutionary biology, taxonomy, conservation biology, be- havioral ecology, and avicultural and poultry science, it is surprising that the last English language compilation of avian hybrids was published nearly a half century ago (Gray 1958). Handbook of Avian Hybrids of the World fills this vacuum with an exhaustive listing of natural hybrids and those obtained in aviaries or through breeding experiments. McCarthy’s intent (Preface) to “provide basic information about each of the thousands of types of reported avian crosses, to provide ac- cess to documenting literature, and to famil- iarize readers with the nature of avian hybrid- ization,” succeeds in most respects. The book is organized in five main sections: introduction (35 pages), hybrid accounts (299 pages), appendices (15 pages), literature cited (160 pages), and index of English and scien- tific names (69 pages). The introduction out- lines the organizational scheme of hybrid ac- counts, briefly addresses hybrid fertility, tax- onomy, identification of hybrids, expression of hybrid traits, hybrid zones, and causes of hybridization. McCarthy’s review of each of these topics is brief, but the sections that ad- dress the hybrid diagnosis, the process by which parental species may be ascertained from morphology (Graves 1990), are too gen- eral to be of much use. The reader is left with no real idea of the systematic processes that are required to accurately assign the parentage of a hybrid or how alternative hypotheses may be rejected. McCarthy makes a number of poorly supported claims in the introduction — for example, that heterosis or hybrid luxuri- ance (when the measurements of hybrids fall outside the cumulative range of measurements observed in the parental species) has been ob- served in avian hybrids. To my knowledge, there are no well-documented cases of hybrid luxuriance in birds. References that purport such a phenomenon insufficiently sampled the range of morphological variation in the paren- tal species. Hybrid accounts and the cited literature form the core of the book. A typical hybrid account takes the form: Selasphorus ardens (Glow-throated Hum- mingbird) X Selasphorus scintilla (Scintillant Hum- mingbird) NHR. BRO. Panama. Panov 1989; Stiles 1983. McCarthy introduces a coding system that permits the reader to quickly note whether hy- bridization occurs under natural conditions or in captivity, the fertility of hybrids, cross re- versibility, breeding range overlap, etc. This is a significant organizational advance over past compilations of hybrid literature (Gray 1958, Panov 1989). Interesting and useful an- notations accompany many of the hybrid 234 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 crosses, although a small fraction of the ac counts I checked garbled the facts or present- ed erroneous information (e.g., Bogota Sun- angel, Heliangelus zusir. Graves 1993]). The appendix listing aviary hybrids of canaries is particularly fascinating because it amply dem- onstrates how slowly genetic barriers to hy- bridization evolve among oscines. Powerful internet search engines make the task of searching the hybrid literature far eas- ier than it was just a decade ago. Nevertheless, McCarthy’s compilation of the literature is im- pressive. His quest for published and unpub- lished hybrid records probed the nooks and crannies of internet web sites, birding jour- nals, birder’s trip lists, museum data bases, and the largely inaccessible 19th and early 20th century avicultural and ornithological lit- erature housed only in a few museum and uni- versity libraries. However, those who cast a wide net had better be prepared to throw back the by-catch. Unfortunately, McCarthy kept it all the good, the bad, and the vague. The lack of discrimination reduces the usefulness of his work because readers will have to dig deeply into the cited works to evaluate the quality of the documentation. As an example of material that should have been rejected, a simple entry (“Magpie Robin X White-rump Shama; Pai, Thailand’’) in a birder’s trip log posted on an internet web site (Appendix 3, ASIA96) constituted the sole authority for the hybrid cross, Copsychiis malabaricus X C. saularis. Regrettably, considerable numbers of insufficiently documented hybrids extracted from internet sources and print literature were included. Curiously, McCarthy also included records of inferred hybridization where no hy- brid offspring are known to exist. To his cred- it, McCarthy attempted to break the chain of citation for some hybrid crosses that have been erroneously reported in the literature, but he also propagated several hybrid records (e.g., Selasphorus ardens X S. scintilla) cre- ated by Panov (1989) and managed to invent new ones on his own (e.g.. White-bearded Manakin [Manacus manacus] X Crimson- hooded Manakin [Pipra aureola)). “Double entry’’ is another persistent prob- lem in the hybrid accounts. This occurs when two or more parental hypotheses have been proposed for an enigmatic specimen leading to the listing of two hybrid combinations. Considerable background research is often necessary to detect “double entry” errors. Spelling errors are few and far between (e.g., ""Metallura altricularis'"). However, there seemed to be an unusual number of citations assigned inappropriately to hybrid records. In all fairness, keeping track of 5,000+ cited sources and thousands of reported hybrids is no mean feat. Still, a pre-publication appraisal by reviewers familiar with the avian hybrid literature may have eliminated many of the factual errors in the book. The damage caused by indiscriminate pub- lication of insufficiently or incompetently di- agnosed hybrids is threefold. First, it intro- duces unsupported records into the literature that are maddeningly difficult to eradicate once published. Despite the caveats men- tioned in McCarthy’s introduction, less savvy users may naturally assume the documentation backing hybrid records in the book meets some quality threshold. Unfortunately, the quality of hybrid records runs the gamut from wild guesses and unadorned assertions posted on internet web sites to detailed diagnoses published in reputable peer-reviewed journals (Banks and Johnson 1961, Graves 1996). Lastly, inclusion of hybrids based on sight re- cords and photographs of free-living birds en- dorses the notion promoted in many birding publications that accurate diagnosis of hybrids observed in the field is not only possible but probable. While some hybrids may certainly be identified under field conditions (e.g., Ver- mivora chrysoptera [Golden-winged Warbler] X V. pinus [Blue-winged Warbler]), an accu- rate hybrid diagnosis in most taxonomic groups, particularly those with drab plumages (e.g., Empidonax, Contopus, Phylloscopus) or from regions of high species diversity, can sel- dom be accomplished under field conditions or from photographs of free-living birds. De- tailed hybrid diagnoses based on fully vouch- ered specimens (available for repeated mor- phological and genetic analyses) remain the gold standard in the field (Hennache et al. 2003). Whenever possible, hybrid birds (par- ticularly rare or possibly unknown combina- tions) should be collected for scientific study. The scientific value of such specimens is sig- nificant as demonstrated by the voluminous literature on avian hybridization. Every band- ing station should possess a scientific collect- ORNITHOLOGICAL LITERATURE 235 ing permit that allows hybrid birds to be col- lected for scientific study. Museums are more than willing to process and archive such spec- imens for use by the greater scientific com- munity. McCarthy lists museum specimen numbers in the annotations for some hybrid records. Although it would entail a Herculean effort, the value of future editions of this book would be greatly enhanced if the availability of voucher specimens and relative quality of documentation were noted for all hybrid com- binations. The most serious problem with compila- tions of this nature is that synthesists may be tempted to use the data for phylogenetic or macroecological studies of hybridization. In many research programs, it is standard prac- tice to assign students the task of converting catalogs of this type into spreadsheet data files for analyses by senior investigators. However, few senior investigators will have the time or library resources to examine the original ci- tations or the expertise and comparative ma- terial needed to vet supposed cases of hybrid- ization. This would be difficult to do outside a large museum. Macroecological and phylo- genetic analyses based on unexpurgated data extracted from hybrid catalogs constitute poor science, user caveo. Who needs this book? Every museum and university library needs a copy of Handbook of Avian Hybrids of the World. I will use it frequently as a convenient and comprehensive guide to the hybrid literature. Aviculturists may also find the contents useful but I suspect the vast majority of birders would not use it enough to warrant its purchase. — GARY R. GRAVES, National Museum of Natural His- tory, Smithsonian Institution, Washington, D.C. 20013, USA; e-mail: gravesg@si.edu LITERATURE CITED Banks, R. C. and N. K. Johnson. 1961. A review of North American hybrid hummingbirds. Condor 63:3-28. Grant, R R. and B. R. Grant. 1992. Hybridization of bird species. Science 256:193-197. Graves, G. R. 1990. Systematics of the “Green-throat- ed Sunangels” (Aves: Trochilidae): valid taxa or hybrids? Proceedings of the Biological Society of Washington 103:6-25. Graves, G. R. 1993. Relic of a lost world: a new spe- cies of sunangel (Trochilidae: Heliangelus) from “Bogota”. Auk 110:1-8. Graves, G. R. 1996. Hybrid wood warblers, Dendroi- ca striata X Dendroica castanea (Aves: Fringil- lidae: Tribe Parulini) and the diagnostic predict- ability of avian hybrid phenotypes. Proceedings of the Biological Society of Washington 109:373- 390. Gray, A. P. 1958. Bird hybrids. Commonwealth Ag- ricultural Bureaux, Bucks, England. Hennache, a., P. Rasmussen, V. Lucchini, S. Rimondi, AND E. Randi. 2003. Hybrid origin of the Imperial Pheasant Lophura imperialis (Delacour and Ja- bouille, 1924) demonstrated by morphology, hy- brid experiments, and DNA analyses. Biological Journal of the Linnean Society 80:573-600. Mayr, E. 1942. Systematics and the origin of species. Columbia University Press, New York, USA. Panov, E. N. 1989. Natural hybridisation and etholog- ical isolation in birds (in Russian). Nauka, Mos- cow, USSR. Prager, E. M. and a. C. Wilson. 1975. Slow evolu- tionary loss of the potential for interspecific hy- bridization in birds: a manifestation of low regu- latory evolution. Proceedings of the National Academy of Sciences USA 72:200-204. Price, T. 2008. Speciation in birds. Roberts and Com- pany, Greenwood Village, Colorado, USA. Stiles, E G. 1983. Systematics of the southern species of Selasphorus (Trochilidae). Auk 100:311-325. HANDBOOK OF THE BIRDS OF THE WORLD, VOLUME II: OLD WORLD FLY- CATCHERS TO OLD WORLD WAR- BLERS. Edited by Josep del Hoyo, Andrew Elliott, and David Christie. Lynx Edicions, Barcelona, Spain. 2006: 798 pages, 55 color plates, more than 300 photographs, and 723 maps. ISBN: 84-96553-06-X. $250.00 (cloth). — Handbook of the Birds of the World continues with Volume 1 1 in the landmark se- ries. This volume contains seven Families of Passerines including Old World Flycatchers, Batises and Wattle-eyes, Fantails, Monarch- flycatchers, Kinglets and Firecrests, Gnat- catchers, Cisticolas and Allies, and Old World Warblers. Each Family account has sections outlining familial characteristics including systematics, morphology, habitat, habits, voice, food and feeding, breeding, move- ments, relationship with man, status and con- servation, and a general bibliography. This section is lavishly illustrated with exquisite photographs that capture the birds in intimate moments bathing, roosting, nesting, caring for young, feeding, or just being birds. The species accounts start with a color plate 236 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 1, March 2008 followed by individual accounts for the spe- cies on the plate. The plates are up to the ex- cellent standards of the series with accurate shapes and well reproduced colors. The text for each species includes the scientific name as well as the name in English, French, Ger- man, Spanish, and other common names. Many of the same subjects in the family ac- count are addressed for each species individ- ually; other topics include subspecies and dis- tribution, descriptive notes, and a range map. The editors went to great lengths to ensure this volume was up to date at the time it went to press. The book includes two species first described in 2006, the Dark Batis (Batis cryp- ta) and the Odedi Bush-warbler {Cetti had- deni). Also included is a photograph in hand of the Large-billed Leaf Warbler {Acroce- phalus orinus) trapped in March 2006, the first record since the species was discovered in 1867; remarkable for a book with a publi- cation date of 2006. The volume is well-ed- ited. I found one minor formatting fault and one apparent miscount. The series includes a Foreword to each vol- ume on some aspect of ornithology. This vol- ume has a Foreword by ^agan §ekercioglu titled The Ecological Significance of Bird Populations. This paper describes the ecolog- ical processes that birds undertake (pollina- tion, seed dispersal, predation, carrion con- sumption, etc.) and the effect that declines in bird populations have on the ecosystem if those processes are lost. This is a thought-pro- voking article, although I would have pre- ferred having the references with the article rather than mixed with the general references for the entire volume. I highly recommend this series. — MARY GUSTAFSON, Rio Grande Joint Venture, Tex- as Parks and Wildlife Department, 2800 South Bentsen Palm Drive, Mission, TX 78572, USA; e-mail: mary.gustafson@tpwd. state, tx.us. TEXAS QUAILS: ECOLOGY AND MANAGEMENT. Edited by Leonard A. Brennan. Texas A&M University Press, Col- lege Station, USA. 2007: ix + 491 pages, with black and white photographs, tables, graphs, and maps. ISBN-13: 978-1-58544-503-5, $40.00 (cloth). — This hefty, hardback volume addresses four quail species that inhabit the state of Texas; Gambel’s Quail (Callipepla gambelii), Montezuma or Meams’ Quail (Cy- tonyx montezumae). Northern Bobwhite {CoT inus virginianus), and Scaled Quail {Callipe- pla squamata). This book is a collection of essays and pa- pers divided into three sections. One section deals with the life history of the four native quail species occurring in Texas. Another sec- tion addresses the birds’ populations in each of the eco-regions as well as acknowledging the challenges of research; past, present, and future. In addition, this section addresses man- agement opportunities for these species in their particular eco-region. A third section at- tends to the rich cultural heritage of quail and quail hunting in the Lone Star State, as well as the future of these birds in Texas. Within each of these three sections are chapters written by some of the best quail re- searchers in Texas and Oklahoma. Overall, the chapters are written well, enhanced by rele- vant black and white photographs. Tables and graphs often appear on the same or adjacent page which aids in further illustrating the dis- cussion and/or research. Quotations and doc- umentation are cited at the end of the sentence by author(s) and year in parentheses. Of par- ticular note is the extensive bibliography sec- tion toward the back of the book, labeled “References.” For the researcher, the book represents an invaluable source on the history, ecology, management practices, and guide- lines for the present and future. It also pro- vides an appreciable inventory of past works. As valuable as this volume is for the re- searcher, it is not exclusively intended for the professional ornithologist. The book has a number of chapters that address issues appli- cable to wildlife managers and landowners. Chapters describing past methodologies (suc- cessful or failed), and recommendations re- garding future practices are present in abun- dance. Many of the recommendations are bom of mistakes, misunderstanding, and errors from the past. Much of existing research, in particular regarding Scaled, Gambel’s and Montezuma Quail is from other states, most notably Arizona, New Mexico, and western Oklahoma. This collection will be of interest to ecologists, conservationists, citizen-scien- ORNITHOLOGICAL LITERATURE 237 lists, and anyone with interest in quail and wildlife eonservation. Despite the title, this book’s value is not limited to Texas but the wider southwestern United States. Managing for suitable native habitat for quail species is quite often in concert with restoring habitat for “non-game” species as well. This is an impressive comprehensive, up- to-date collection of Texas and Southwest re- search reports detailing quail behavior, life history, ecology, management, native habitat, and its restoration. It will be of interest to those concerned about the fragmentation and loss of wild spaces. This book will help in- form those concerned with the survival, not just of any one particular species, but of the cultural, historical, and aesthetic sensitivity of humans. This book would be a significant reference addition to the library of anyone interested in wildlife. I recommend it. — MATTHEW W. YORK, 9910 Ramblewood Drive, Waco, TX 76712, USA; e-mail: mwy391@yahoo.com Editor’s Comments Production of Volume 119 (2007) went rel- atively smoothly thanks to reviewers, authors, and the staff at Allen Press. We are receiving and publishing more manuscripts, and expect this trend to continue. We are also receiving more manuscripts from outside of North America, especially from Central and South America. We also expect this trend to contin- ue. Publishing manuscripts from across the world has some pitfalls as usage of common and scientific names is frequently not consis- tent. Effective 1 January 2008, we now follow Gill and Wright (Gill, E and M. Wright, 2006. Birds of the world. Recommended English names. Princeton University Press, Princeton, New Jersey, USA and Oxford, United King- dom.) for all non-North American species. This reference source was published on behalf of the International Ornithological Congress. We will listen to rational views on changes in nomenclature (especially newly described species) as they become published. We con- tinue to follow the American Ornithologists’ Union for North American species except for some species that are not primarily residents or migrants in North America. We believe we have outstanding papers in this issue and more are scheduled for later is- sues in Volume 120. We continue to seek re- viewers capable of providing timely and thor- ough reviews of manuscripts. Please help us access current e-mail addresses via updating The Flock or direct communication with our office. Clait E. Braun Editor, 2008 THE WILSON JOURNAL OF ORNITHOLOGY Editor CLAIT E. BRAUN 5572 North Ventana Vista Road Tucson, AZ 85750-7204 E-mail: TWILSONJO@comcast.net Editorial Board RICHARD C. BANKS KATHY G. BEAL JACK CLINTON EITNIEAR SARA J. OYLER-McCANCE Editorial NANCY J. K. BRAUN Assistant Review Editor MARY GUSTAFSON Texas Parks and Wildlife Department 2800 South Bentsen Palm Drive Mission, TX 78572, USA E-mail: WilsonBookReview@aolcom GUIDELINES FOR AUTHORS Please consult the detailed “Guidelines for Authors” found on the Wilson Ornithological Society Web site (http://www.ummz.umich.edu/birds/wos/index.html). All manuscript submissions and revisions should be sent to Clait E. Braun, Editor, The Wilson Journal of Ornithology, 5572 North Ventana Vista Road, Tucson, AZ 85750- 7204, USA. The Wilson Journal of Ornithology office and fax telephone number are (520) 529-0365. The e- mail address is TWilsonJO@comcast.net. NOTICE OF CHANGE OF ADDRESS Notify the Society immediately if your address changes. Send your complete new address to Ornithological Societies of North America, 5400 Bosque Boulevard, Suite 680, Waco, TX 76710. The permanent mailing address of the Wilson Ornithological Society is: %The Museum of Zoology, The University of Michigan, Ann Arbor, MI 48109. Persons having business with any of the officers may address them at their various addresses given on the inside of the front cover, and all matters pertaining to the journal should be sent directly to the Editor. MEMBERSHIP INQUIRIES Membership inquiries should be sent to Timothy J. O’Connell, Department of Natural Resource Ecology and Management, Oklahoma State University, 205 Life Sciences West, Stillwater, OK 74078; e-mail: oconnet® okstate.edu THE JOSSELYN VAN TYNE MEMORIAL LIBRARY The Josselyn Van Tyne Memorial Library of the Wilson Ornithological Society, housed in the University of Michigan Museum of Zoology, was established in concurrence with the University of Michigan m 1930. Until 1947 the Library was maintained entirely by gifts and bequests of books, reprints, and ornithological magazines from members and friends of the Society. Two members have generously established a fund for the purchase of new books; members and friends are invited to maintain the fund by regular contribution. The fund will be administered by the Library Committee. Robert Payne, University of Michigan, is Chairman of the Committee. The Library currently receives over 200 periodicals as gifts and in exchange for The Wilson Journal of Orni- thology. For information on the Library and our holdings, see the Society’s web page at http:// www.ummz.umich.edu/birds/wos/index.html. With the usual exception of rare books, any item in the Library may be borrowed by members of the Society and will be sent prepaid (by the University of Michigan) to any address in the United States, its possessions, or Canada. Return postage is paid by the borrower. Inquiries and requests by borrowers, as well as gifts of books, pamphlets, reprints, and magazines, should be addressed to: Josselyn Van Tyne Memorial Library, Museum of Zoology, The University of Michigan, 1 109 Geddes Avenue, Ann Arbor, MI 48109-1079, USA. Contributions to the New Book Fund should be sent to the Treasurer. This issue of The Wilson Journal of Ornithology was published on 12 March 2008. Continued from outside back cover 1 59 Timing and location of mortality of fledgling, subadult, and adult California Gulls Bruce H. Pugesek and Kenneth L. Diem 167 Effects of predation and food provisioning on Black Tern chick survival Shane R. Heath and Frederick A. Servello 176 Use of clay licks by Maroon-fronted Parrots {Rhynchopsitta terrisi) in northern Mexico Rene A. Valdes-Peha, Sonia Gabriela Ortiz-Maciel, Simon O. Valdez Juarez, Ernesto C. Enkerlin Hoeflich, and Noel E R. Snyder 1 8 1 Cavity number and use by other species as correlates of group size in Red-cockaded Woodpeckers John J. Kappes Jr. Short Communications 190 First description of nests and eggs of two Hispaniolan endemic species: Western Chat-tanager {Calyptophilus tertius) and Hispaniolan Highland-tanager {Xenoligea montana) Christopher C. Rimmer, Lance G. Woolaver, Rina K Nichols, Eladio M. Fernandez, Steven C. Latta, and Esteban Garrido 195 Foraging and nesting of the ‘Akikiki or Kauai Creeper {Oreomystis bairdi) Eric A. VanderWerfand Pauline K Roberts 200 First observation of duetting in the Olive-backed Euphonia {Euphonia gouldi) Thor Hanson 20 1 The display of a Reddish Hermit {Phaethornis ruber) in a lowland rainforest, Bolivia Adam Felton, Annika M. Felton, and David B. Lindenmayer 205 Home range and habitat preferences of the Banded Ground-cuckoo {Neomorphus radiolosus) Jordan Karubian and Luis Carrasco 209 Comparisons between juvenile and adult American Robins foraging for mulberry fruit E. Natasha Vanderhoffand Perri K Eason 214 Previously unknown food items in the diet of six neotropical bird species Luis Sandoval, Esteban Biamonte, and Alejandro Solano-Ugalde 217 Anvil use by the Red-cockaded Woodpecker Kristin J. Bondo, Lauren N Gilson, and Reed Bowman 221 Gender identification of Grasshopper Sparrows comparing behavioral, morphological, and molecular techniques Frank K Ammer, Petra Bohall Wood, and Roger J. McPherson 226 Abnormal eggs of Rio Grande Wild Turkeys on the Edwards Plateau, Texas Kyle B. Melton, Justin Z. Dreibelbis, Ray Aguirre, Bret A. Collier, T. Wayne Schwertner, Markus J. Peterson, and Nova J. Silvy 228 Barred Forest Falcon {Micrastur rujicollis) predation on relatively large prey Fdbio Rohe and Andre Pinassi Antunes 231 Ornithological Literature Compiled by Mary Gustafson 237 Editors Comments The Wilson Journal of Ornithology (formerly The Wilson Bulletin) Volume 1 20, Number 1 CONTENTS March 2008 Major Articles 1 A new White-eye {Zosterops) from the Togian Islands, Sulawesi, Indonesia Mochamad Indrawan, Pamela C. Rasmussen, and Sunarto 10 The WTite-eyed Foliage-gleaner (Furnariidae; Automolus) is two species Kevin J. Zimmer 26 Recent advances in the behavioral ecology of tropical birds Bridget J. M. Stutchbury and Eugene S. Morton 38 Distribution, behavior, and conservation status of the Rufous Twistwing [Cnipodectes superrufus) Joseph A. Tobias, Daniel J. Lebbin, Alexandre Aleixo, Michael]. Andersen, Edson Guilherme, Peter A. Hosner, and Nathalie Seddon 50 Natural history and behavior of the Aldabra Rail {Dryolimnas [cuvieri] aldabranus) Ross M. Wanless and Philip A. R. Hockey 62 Post-fledging movement and spatial habitat-use patterns of juvenile Swainson’s Thrushes Jennifer D. White and John Eaaborg 74 Spring migratory stopover of Swainson’s Thrush along the Pacific Coast of southwest Costa Rica Scott Wilson, Keith A. Hobson, Douglas M. Collister, and Amy G. Wilson 35 Demography of Eastern Yellow Wagtails at Cape Romanzof, Alaska Heather M. Renner and Brian J. McGaffery 92 Breeding ecology of the Narcissus Flycatcher in north China Ning Wang, Yanyun Zhang, and Guangmei Zheng 99 Reproductive success of Fiouse Wrens in suburban and rural landscapes Michael J. Newhouse, Peter P Marra, and L Scott Johnson 105 Habitat selection and reproductive success of Cerulean Warblers in Indiana Kirk L Roth and Kamal Islam 111 Nesting biology of grassland birds at Fort Campbell, Kentucky and Tennessee James J. Giocomo, E. Daniel Moss, David A. Buehler, and William G. Minser 120 Factors affecting home range size and movements of post-fledging grassland birds Kimberly M. Suedkamp Wells, Joshua J. Millspaugh, Mark R. Ryan, and Michael W Hubbard 1 3 1 Ecological factors affecting response of Dark-eyed Juncos to prescribed burning Jinelle H. Sperry, T. Luke George, and Steve Zack 1 39 Winter habitat use by Boreal Chickadee flocks in a managed forest Adam Hadley and Andre Desrochers 146 Long-term effects of wastewater irrigation on habitat and a bird community in central Pennsylvania Adam T Rohnke and Richard H. Yahner 1 53 Long-term trends in breeding birds in an old-growth Adirondack forest and the surrounding region Stacy A. McNulty, Sam Droege, and Raymond D. Masters Continued on inside back cover Wilson Journal of Ornithology Volume 120, Number 2, June 2008 MCZ UBBABV m 19 Published by the Wilson Ornithological Society THE WILSON ORNITHOLOGICAL SOCIETY FOUNDED 3 DECEMBER 1888 Named after ALEXANDER WILSON, the first American ornithologist. President— James D. Rising, Department of Zoology, University of Toronto, Toronto, ON MSS 3G5, Canada; e-mail: rising@zoo.utoronto.ca First Vice-President— E. Dale Kennedy, Biology Department, Albion College, Albion, MI 49224, USA; e-mail: dkennedy@albion.edu Second Vice-President— Robert C. Season, USDA, Wildlife Services, 6100 Columbus Avenue, Sandusky, OH 44870, USA; e-mail: beason@netzero.com Editor Clait E. Braun, 5572 North Ventana Vista Road, Tucson, AZ 85750, USA; e-mail: TWILSONJO@ comcast.net Secretary John A. Smallwood, Department of Biology and Molecular Biology, Montclair State University, Montclair, NJ 07043, USA; e-mail: smallwoodj@montclair.edu Treasurer MelindaM. Clark, 52684 Highland Drive, SouthBend, IN 46635, USA; e-mail: MClark@tcservices.biz Elected Council Members— Carla J. Dove, Greg H. Farley, and Mia R. Revels (terms expire 2009); Rebecca Holberton, Robert S. Mulvihill, and Timothy O’Connell (terms expire 2010); Jameson F. Chace, Sara R. Morris, and Margaret A. Voss (terms expire 2011). Membership dues per calendar year are: Active, $21.00; Student, $15.00; Family, $25.00; Sustaining, $30.00; Life memberships, $1,000.00 (payable in four installments). The Wilson Journal of Ornithology is sent to all members not in arrears for dues. THE WILSON JOURNAL OF ORNITHOLOGY (formerly The Wilson Bulletin) THE WILSON JOURNAL OF ORNITHOLOGY (ISSN 1559-4491) is published quarterly in March June, September, and December by the Wilson Ornithological Society, 810 East 10th Street, Lawrence, KS 66044-8897. The subscription price, both in the United States and elsewhere, is $40.00 per year. Periodicals postage paid at Lawrence, KS. POSTMASTER; Send address changes to OSNA, 5400 Bosque Boulevard, Suite 680, Waco, TX 767 All articles and communications for publications should be addressed to the Editor. Exchanges should be addressed to The Josselyn Van Tyne Memorial Library, Museum of Zoology, Ann Arbor, MI 48109, USA. Subscriptions, changes of address, and claims for undelivered copies should be sent to OSNA, 5400 Bosque Boulevard, Suite 680, Waco, TX 76710, USA. Phone: (254) 399-9636; e-mail: business@osnabirds.org. Back issues or single copies are available for $12.00 each. Most back issues of the journal are available and may be ordered from OSNA. Special pnces will be quoted for quantity orders. All issues of the journal published before 2000 are accessible on a free web site at the University of New Mexico library (http://elibrary.unm.edu/sora/). The site is fully searchable, and full-text reproductions of all papers (including illustrations) are available as either PDF or Dj Vu files. © Copyright 2008 by the Wilson Ornithological Society Printed by Allen Press, Inc., Lawrence, KS 66044, U.S.A. COVER: Wilson’s Storm Petrel {Oceanites oceanicus). Illustration by Don Radovich. 0 This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). MCZ LIBRARY MAY 19 2009 HARVARD UNIVERSITY FRONTISPIECE Male Ruffed Grouse (Bonasa umbelhts) on a display log amidst a dense thicket ot aspen iPopulus tremuloides) in north-central Minnesota. Ruffed Grouse exhibit cyclic population fluctuations over much of their range. These cycles have been thought to be caused by predation, particularly by Northern Goshawks (Accipiter gentiUs). Zimmerman et al. (page 239) report that winter weather conditions explained more variation in spring grouse indices than goshawks, but were not a definitive explanation ot eye es. oto graph by Gordon W. Gullion (with permission of the University of Minnesota). The authors dedicate this paper Wilson Journal of Ornithology Published by the Wilson Ornithological Society VOL. 120, NO. 2 June 2008 PAGES 239-432 The Wilson Journal of Ornithology 120(2):239-247, 2008 NEW INSIGHT TO OLD HYPOTHESES: RUEEED GROUSE POPULATION CYCLES GUTHRIE S. ZIMMERMANJ 2 RICK R. HORTON, DANIEL R. DESSECKER,^ AND R. J. GUTIERREZi 6 ABSTRACT. We examined factors hypothesized to influence Ruffed Grouse {Bonasa umbellus) population cycles by evaluating 13 a priori models that represented correlations between spring counts of male Ruffed Grouse drumming displays and these factors. We used AIC^ to rank the relative ability of these models to fit the data and used variance components analysis to assess the amount of temporal process variation in Ruffed Grouse spring counts explained by the best model. A hypothesis representing an interaction between winter precipitation and winter temperature was the top-ranked model. This model indicated that increased precipitation during cold winters (soft snow cover for roosting) was correlated with higher grouse population indices, but that increased precipitation during warm winters (snow crust effect) was correlated with lower spring counts. The highest ranked model (AIC^, weight = 0.45), explained only 17% of the temporal process variation. The number of migrating Northern Goshawks (Accipiter gentilis), which has been correlated with grouse cycles in previous studies, does not adequately explain, by itself, the variation in annual population indices of Ruffed Grouse. Other factors not considered in our analysis, such as endogenous mechanisms, parasites, or interactions among factors may also be important, which suggest that mechanisms mediating the Ruffed Grouse cycle still require investigation. Received 17 March 2007. Accepted 17 August 2007. The Ruffed Grouse (Boncisa umbellus) is a widespread and important game bird in North America (Rusch et al. 2000). Population size ' Department of Fisheries, Wildlife, and Con.serva- tion Biology, University of Minnesota, 200 Hodson Hall, St. Paul, MN 55108, USA. 2 Current address; U.S. Fish and Wildlife Service, Division of Migratory Bird Management, I 1510 American Holly Drive, Laurel, MD 20708, USA. ^ Ruffed Grou.se Society, P. O. Box 657, Grand Rap- ids, MN 55744, USA. ■^Current address: Minnesota Department of Natural Resources, Grand Rapids, MN 55744, USA. Ruffed Grouse Society, P O. Box 2, Rice Lake, W1 54868, USA. ^Corresponding author; e-mail: gutieO I 2(«)umn.edu of grouse is known to fluctuate through time and, in some areas, these fluctuations follow a cycle of ~10 years (Keith 1963, Rusch et al. 20()0). Their population fluctuations and cycles have been studied extensively (Bump et al. 1947, Keith 1963, Gullion and Marshall 1968, Keith and Ru.sch 1989, Small et al. 1991 ). These studies have generated many hy- potheses about mechanism(s) responsible for grouse cycles. Ideas advanced as mechanisms controlling grouse fluctuations range from the bizarre (e.g., rabbit [Lagomorpha] seasons were too long; Bump et al. [I947|) to the plausible (e.g., predation; Ru.sch et al. 2()()0). Most recently, some researchers have present- ed correlative data that suggest predation by 239 240 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 2, June 2008 Northern Goshawks {Accipiter gentilis) and Great Homed Owls {Bubo virginianus) are most likely causing cyclic population fluctu- ations in Ruffed Grouse (Keith and Rusch 1989, Lauten 1995). However, there have been no studies that simultaneously evaluated competing hypotheses about factors (e.g., pre- dation, weather, competitors) that could poten- tially influence grouse numbers through their effect on survival and reproduction. We used model selection to compare relative support for multiple hypotheses representing potential factors that could influence Ruffed Grouse population fluctuations. Our objective in this study was to revisit a phenomenon (the 10- year grouse cycle) that has been assumed to be resolved scientifically (i.e., predation caus- es these cycles, Rusch et al. 2000) using the comparative model approach. We believe this approach can be used to guide future research about Ruffed Grouse population cycles. METHODS Study Area. — Our study area encompassed a 20,35 8-km2 area around Grand Rapids, Min- nesota, USA (47°13'N, 93° 31' E), defined by a circle having a radius of 80.5 km. This was the approximate maximum distance trav- eled from Grand Rapids by hunters who col- lected grouse for our study. This region was within a transition zone between the decidu- ous hardwood and boreal forests. The weather was variable over the study period and was characterized by warm, moist summers and cold, snowy winters. Response Variable. — We were interested in factors that might influence Ruffed Grouse cy- cles and used spring counts of male Ruffed Grouse dmmming displays (a gross index to population fluctuations; Rusch and DeStefano 1989, Zimmerman and Gutierrez 2007) as our response variable. Male Ruffed Grouse use several displays to attract or court females during the spring, one of which is called the “drumming display” (Rusch et al. 2000:9). This display consists of a male grouse beating its wings in a manner that produces a distinct low frequency nonvocal sound (Rusch et al. 2000). We used data from surveys within the study area that were conducted annually by the Minnesota Department of Natural Re- sources. Surveys occurred along non-random- ly selected, but consistently surveyed, routes that were approximately 14.4 km in length along maintained roads. Wildlife biologists, land managers, and volunteers attempted to survey each route once during the spring (ear- ly Apr-early May). Volunteers conducted sur- veys by stopping every 1.6 km along each route for 4-min intervals to listen for drum- ming displays. The number of drumming dis- plays heard per 4-min interval was recorded at each stop (Minnesota Department of Nat- ural Resources, unpubl. protocol). Grouse were not surveyed on rainy or windy days. We recognize this index may not truly represent population size and may have represented only drumming display activity of male Ruffed Grouse (Anderson 2001). Development of a priori Hypotheses. — We evaluated hypotheses (models) about factors that influence Ruffed Grouse population dy- namics that exist in the literature (Bump et al. 1947, Gullion and Marshall 1968, Rusch et al. 2000) and our own experience working with grouse. We limited our set of hypotheses to those for which data were available. We con- sidered 13 hypotheses, a priori to analysis, which represented predictions about how gos- hawk abundance, weather during the breeding season, weather during the previous winter, exploitative competition (forest tent caterpil- lars, Malacosoma americanum), color phase ratios, mass of male grouse, and age ratios of both genders during the autumn correlated with Ruffed Grouse population indices that were conducted from 1983 to 2004 (Table 1). We limited our analysis to the interval 1983— 2004 because we did not have individual grouse data prior to 1982; the 1982 data were used to predict the population index for the following spring (i.e., spring 1983 counts). Predation Hypothesis.— Grouse are killed by many predators (Bump et al. 1947). Correlative evidence supports a substantial, if not controlling, effect of goshawk and Great Homed Owl predation on the Ruffed Grouse cycle (Keith and Rusch 1989). We did not lo- cate relevant data about Great Homed Owl predation or owl abundance on our study area, but we did acquire population indices of gos- hawks, which are important predators of Ruffed Grouse in Minnesota (Eng and Gullion 1962). We estimated an annual index of gos- hawk abundance from counts of migrating raptors at Hawk Ridge Nature Center in Du- Zimmerman et al. • RUFFED GROUSE POPULATION CYCLES 241 TABLE 1. A priori models representing hypotheses of factors correlating with population cycles in Ruffed Grouse in northern Minnesota, USA, 1983-2004. Model Verbal hypothesis Year^ Season dates Variables/Predictions^-‘^ 1 Goshawk abundance T Fall — Hawk Ridge -GH 2 Ereezing while female lays eggs t-1 20 Apr- 15 May -DF, 3 Cold and wet weather during hatching and brooding t-1 25 May-25 Jun ±Ph + T, ± (P,*T,) 4 Negative effects of snow-less cold winter T 1 Dec-31 Mar +T,. + P,. - (T,*P,) 5 Duration of extreme cold events T 1 Dec-31 Mar 6 Snow depth T 1 Dec-31 Mar +S 7 Duration of adequate snow cover T 1 Dec-31 Mar + SC 8 Consecutive years of adequate snow t-1, t-2, t-3 1 Dec-31 Mar + SC + sc, + SC2 9 Tent caterpillars (cover hypothesis) t-1 Summer -C, 10 Tent caterpillars (immediate effect due to loss of buds or production of secondary compounds) t-1, t-2, t-3 Summer -c, - C2 - C3 11 Color phase ratios (phenotypic hypothesis)^' T Fall hunt ±GR 12 Mass of males T Fall hunt 13 Age ratio of both genders — insight into cycles? t-1 Fall hunt ±Ap ® t - 1 September of previous year to current breeding season. — negative correlation, + = positive correlation, ± = effect unknown — could be positive or negative correlation. Ph - amount of precipitation during the hatching period; Th = mean minimum daily temperature during the hatching period; P„ = amount of precipitation during^ the nesting period; DFi = number of days below freezing during the laying period; = mean minimum temperature during the winter season; P^^. - amount of precipitation during the winter season; Cw = number of nights with min temperature below -15° C; S = average daily snow depth during the winter season; SC = number of days with > 15 cm of snow during previous winter (t); SC] = number of days with > 15 cm of snow during t-1; SC2 = number of days with > 15 cm of snow during t-2; GH = index to goshawk abundance: number of goshawks counted at Hawk Ridge; Cj through C3 = percent of area within 80.5 km of Grand Rapids with forest tent caterpillar damage; = average mass of immature male grouse from RGS hunt data; Ap - age ratio of both genders (calculated as number of young-of-year birds harvested divided by number of adults harvested) from national hunt; GR = proportion of gray-phased birds harvested during the national hunt. luth, Minnesota (—110 km from Grand Rap- ids). Raptor counts were conducted by trained observers from late August through Novem- ber. Volunteers recorded the number of obser- vation hours and used spotting scopes to re- cord counts of migrating raptors by species. We used these data to estimate the average number of goshawks observed per hour per year as our explanatory variable (model 1 , Ta- ble 1). We recognize these counts may not re- flect predation rates on grouse, but they are similar to indices used in previous studies that correlated grouse abundance with goshawk predation (Keith and Rusch 1989, Lauten 1995). Weather Hypotheses.— ^Q 1 5 cm), number of days with extreme low temperatures (minimum temperature <-15° C), and average daily snow depth (cm) for the winter season. We calculated the num- ber of days with a minimum temperature be- low freezing during the egg-laying season, and the average daily minimum temperature (°C) and mean daily precipitation (cm) during the hatching/brood rearing season. Competition Hypotheses. — Forest tent cat- erpillar outbreaks may affect Ruffed Grouse population dynamics (Jakubas and Gullion 1991). Tent caterpillars could have influenced grouse directly by reducing foliage cover (model 9, Table 1) or lowering food supply (i.e., defoliated aspen [Populus spp.] trees do not produce catkins, which are the primary winter food for grouse in our study area; mod- el 10, Table 1); or indirectly by causing the defoliated trees to increase production of com- pounds that discourage defoliation by cater- pillars (model 10, Table 1). In the latter case, these compounds either may have made the aspen catkins unpalatable or inhibited diges- tion of these catkins by grouse. We used data on tent caterpillars collected by the Minnesota Department of Natural Resources’ Forest Health Unit and GIS (ArcView 3.2, ESRI Inc., Redlands, CA, USA) to estimate the amount of caterpillar defoliation within the study area for each year of our study. These data con- tained spatially explicit maps with area poly- gons representing various categories of tent caterpillar defoliation (i.e., heavy, moderate, light, or none) throughout the entire state of Minnesota. We combined these four catego- ries into two groups: “tent caterpillar out- breaks” (i.e., >20% of an area impacted) and “no tent caterpillar outbreaks” (<20% of an area impacted) for each polygon within the annual GIS layers. We then calculated the pro- portion of our study area that was defoliated (i.e., “tent caterpillar outbreaks”) each year. Grouse Hypotheses. — Gullion (1970) and Gutierrez et al. (2003) suggested that survival correlated with color-phases (model 1 1 , Table 1). We also hypothesized that mass of male grouse may correlate with display activity if mass reflected the condition of males entering winter (model 12, Table 1). Finally, we con- sidered whether age ratio of harvested birds reflected overall population momentum (mod- el 13, Table 1). Ruffed Grouse were harvested in the study area by hunters participating in the Ruffed Grouse Society’s annual National Grouse and Woodcock Hunt (RGS HUNT) held during the second week of October each year since 1982. We collected morphometric data on all grouse registered during the RGS HUNT {n = 5,700). We estimated the propor- tion of gray-phased individuals in the popu- lation, the age ratio of both genders (calculat- ed as number of young-of-year harvested di- vided by number of adults harvested), and the average mass of male grouse from the har- vested grouse. We considered these estimates to be characteristics of our study population because the RGS HUNT occurred over the en- tire area. We assumed the Ruffed Grouse har- vest data would not be biased for color phase because we knew of no reason, a priori, why hunters would take one color phase over an- other. We assumed that hunters would take a greater proportion of juveniles than existed in the overall population due to age-specific vul- nerability (Small 1989). However, we did not believe this bias would change among years, and considered the proportion of juveniles as an index to relative proportion of this segment in the population among years. Data Analysis.— We used model selection Zimmerman et al. • RUFFED GROUSE POPULATION CYCLES 243 (Burnham and Anderson 1998) to rank our a priori hypotheses about factors that correlated with fluctuations in Ruffed Grouse popula- tions. We first expressed our a priori hypoth- eses as statistical models (Table 1). We then checked for correlations among variables within a model. We used a repeated measures design because grouse were sampled along survey routes each year, and mixed modeling to estimate the parameters for each of our a priori models. We used a global model to identify the best covariance structure for the display data. Modeling covariance using this process allowed us to account for correlations between grouse indices among years (i.e., the autocorrelation associated with time series data). We used Akaike’s Information Criterion (AICJ to rank the models’ ability to fit the data, and then estimated the amount of tem- poral process variation explained by the best (lowest AIC^) model. We used PROC CORR in SAS (Version 9.1, SAS Institute Inc., Cary, NC, USA) to develop a correlation matrix of all predictor variables included in the a priori hypotheses. If a pair of variables from the same model had a correlation coefficient >0.60, we excluded one variable from the model to avoid multi- collinearity. The variable excluded depended upon our subjective assessment of the poten- tial for biological interpretation of each of the two correlated variables. We used multiple linear mixed modeling (PROC MIXED in SAS) to estimate the pa- rameters of our statistical models. We consid- ered year to be a random effect, predictor var- iables (i.e., weather, goshawk index, tent cat- erpillar damage) to be fixed effects, and sur- vey routes as the sampling units. We assumed that routes were independent units within years. We used a repeated measures design to analyze the mixed models because, although routes were independent within years, samples from routes were not independent among years. The assumptions of linear regression are that error terms are distributed normally and variances are constant. However, estimation techniques based on normal distributions are robust to departures from normality (White and Bennetts 1996). We modeled the vari- ance-covariance structure because we sus- pected dependence among repeated samples. non-constant variances among years, and au- tocorrelation of counts through time. We used restricted maximum likelihood and a global model to estimate five variance-covariance structures described by Littell et al. (1996) in- cluding first-order autoregressive (ARl), het- erogeneous ARl (HARl), compound sym- metric, unstructured, and log-linear. The ARl structure assumes that variances are constant among years and that covariances decline with increasing time between observations. The HARl is similar to the ARl structure, but it allows variance estimates to change among years. The compound symmetric assumes ho- mogeneous variance and covariances. The un- structured model assumes that no pattern in variance and covariances exists. The log-lin- ear structure allows flexibility in modeling variances as a function of experimental con- ditions (e.g., increased variances during years with higher population indices). The autore- gressive covariance structures are particularly applicable to cyclic species because they al- low observations closer in time to correlate more than observations further away in time. We used standard maximum likelihood esti- mation (Littell et al. 1996:498) and AIC,. to rank candidate models once we identified the most parsimonious variance-covariance struc- ture. We used the Satterthwaite approximation method for estimating the degrees of freedom for each a priori model. We estimated the amount of variation ex- plained by the AIQ selected model by first estimating an intercept-only model (i.e., a model containing an intercept parameter with no other fixed effects) and a model containing a variable for each year. The difference in re- sidual variation between these two models represents the total temporal process variation in the population index. Next, we estimated the residual variation from the best annual co- variate model. We calculated the percent of temporal process variance explained as the difference in residual variation between the intercept-only model and the best covariate model divided by the total temporal process variation. RESULTS Twenty-nine permanent survey routes were sampled within the study area. The number of routes surveyed each year (// = 22 years) var- 244 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 FIG. 1 . Plot of Ruffed Grouse population indices (average drums/stop along male display survey routes) collected by the Minnesota Department of Natural Re- sources from 1983 to 2004, Grand Rapids area, Min- nesota, USA. Error bars represent 95% confidence in- tervals. ied from 19 to 29 because inclement weather, accessibility of the route, and availability of volunteers to conduct surveys varied by year. The number of times individual routes were surveyed during the entire study varied from 6 to 22. The routes that were surveyed the least were not initiated until 1996-1998, and not all of these newer routes have been sur- veyed each year since their initiation. The mean number of drums (displays) heard per stop varied from 0.57 in 1993 to 2.27 in 1989. Grouse population indices during our study reflected a ~ 10-year cyclic pattern (Fig. 1). Modeling the covariance structure of the Ruffed Grouse population index indicated the data supported a HARl structure. This co- variance structure included 22 variance esti- mates (i.e., a variance estimate for each year) and a parameter (p) that described the corre- lation between observations separated by 1 year. This structure indicated that years with higher population indices were characterized by greater variance among survey routes and that indices were correlated from 1 year to the next (p = 0.74). Our data most strongly supported the inter- action between winter temperature and precip- itation model (model 4, Tables 1-2, Fig. 2). AIC^ weights indicated there was strong sup- port for this model because it was almost three times as likely as the second ranked model. This model indicated that grouse indices were highest during cold snowy and warm dry win- ters, and lowest following relatively warm snowy and cold dry winters (Fig. 2, Table 3). This interaction model was ranked the highest, but variance components analysis indicated this model only accounted for 17% of the tem- poral process variation. DISCUSSION The Ruffed Grouse cycle has captured the interest of grouse biologists for many years, and has stimulated many studies to investigate its causative mechanisms (Bump et al. 1947, Keith 1963, Gullion 1984, Keith and Rusch 1989, Balzer 1995, Lauten 1995). Rusch et al. (2000) concluded that avian (particularly gos- hawk) predation was most closely linked to the grouse cycle and, hence, may be the mech- TABLE 2. Model selection assessing the influence of ecological factors on Ruffed Grouse population indices in northern Minnesota, USA, 1983—2004. Model -2 Log likelihood AICcb Delta AICc AICc Weight 4 1085.6000 28 532 3 1087.7000 28 532 1 1093.4000 26 532 12 1093.7000 26 532 11 1093.9000 26 532 9 1094.5000 26 532 13 1094.7000 26 532 6 1095.7000 26 532 2 1096.5000 26 532 5 1097.0000 26 532 7 1097.0000 26 532 10 1093.4000 28 532 8 1093.7000 28 532 1144.8286 0.0000 0.4448 1146.9286 2.1000 0.1557 1148.1802 3.3516 0.0833 1148.4802 3.6516 0.0717 1148.6802 3.8516 0.0648 1149.2802 4.4516 0.0480 1149.4802 4.6516 0.0435 1150.4802 5.6516 0.0264 1151.2802 6.4516 0.0177 1151.7802 6.9516 0.0138 1151.7802 6.9516 0.0138 1152.6286 7.8000 0.0090 1152.9286 8.1000 0.0077 K = number of fixed effect, covariance, and variance parameters in models. AICc- = small sample adjustment of Akaike’s Information Criterion. Zimmerman et al • RUFFED GROUSE POPULATION CYCLES 245 FIG. 2. Predicted effects of interaction between precipitation and temperature on Ruffed Grouse population indices in northern Minnesota, 1983-2004. anism responsible for these cycles. Predation has been suggested to operate on Ruffed Grouse population cycles in two ways. In the first case, resident predators, particularly rap- tors, cause grouse cycles when their staple prey (snowshoe hares [Lepus americanus]) decline and they switch to grouse, which pre- cipitates the decline of grouse in Canada and Alaska (Keith and Rusch 1989). In the second case, snowshoe hare declines in northern Can- ada and Alaska may force goshawks to mi- grate south (Keith and Rusch 1989). When these migrating raptors reach wintering areas in the northern United States, they may in- crease predation rates on grouse to the point of initiating the decline phase of the grouse cycle (Keith and Rusch 1989, Lauten 1995). The cycle does not gain upward momentum until either snowshoe hares increase and they become the staple prey of raptors or these rap- tors no longer invade southern areas. The study by Lauten (1995) is particularly com- pelling because he radio-marked large num- bers of birds and recorded substantial mortal- ity from predation. In addition, he showed a negative correlation between the Ruffed Grouse cycle and counts of migrating gos- hawks (from Hawk Ridge, Whitefish Point Bird Observatory, Wisconsin Christmas Bird Counts, and the Wisconsin Checklist Project; goshawk indices from all raptor count sources were correlated during their study) and Great Horned Owls (from Wisconsin Christmas Bird Counts and Wisconsin Checklist Project). In addition to the negative correlation between grouse and raptor indices, Lauten (1995) also found a positive correlation between the gos- hawk index, avian predation, and winter mor- tality rates of Ruffed Grouse. We did not have data on predation rates during our study, but based on the results of our analysis, we spec- ulate that mortality rates of Ruffed Grouse may be buffered by high quality snow roosts (cold and snowy winters) or less energetically stressful winters (warm and dry winters) even when raptor abundance is high (Gullion 1973). In our analysis, the predation hypothesis was neither a competing model nor explana- tory of variation in our grouse population in- dex (drumming display counts). We used gos- hawk indices similar to those used by Lauten 246 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 TABLE 3 Parameter estimates from AIC, best model (effect of winter temperature, precipitation, and their interaction) for assessing changes in Ruffed Grouse population indices in northern Minnesota, USA, 1983-2004. Parameter Estimate SE dP 95% Confidence interval Intercept Winter precip Winter min temp Interaction term -0.2782 37.0183 0.1142 -2.9560 0.3581 10.5500 0.0498 1 .6443 8.53 9.59 7.39 8.00 -1.0950-0.5386 13.3651-60.6716 -0.00226-0.2307 -6.7479-0.8358 ^ Estimated using the Satterthwaite procedure in PROC MIXED. (1995), but found no relationship between grouse and goshawk indices from the previous year, which was used as a surrogate for pre- dation in other studies. Alternatively, the in- fluence of predators may lag due to switching of prey (Tornberg et al. 2005). For example, goshawks may initially prey upon other spe- cies (e.g., snowshoe hares) before switching to Ruffed Grouse. We considered this model initially, but did not include it as an a priori model because our initial model set was large and we felt there was more support for the non-lag predation hypothesis. Reviewers sug- gested including this predator lag model and we estimated this model post hoc to the anal- ysis. This model performed about equally as well as our original predation model (AIC^ = 1,148.10), but provided no additional explan- atory power. In contrast, we found that winter weather explained the most variation among those hy- potheses we considered. Winter weather can influence thermoregulation, cover from pred- ators, and condition of females at the start of the breeding season. Our top model supported the hypothesis that cold, snowy winters fa- vored grouse whereas cold, snow free condi- tions were unfavorable. However, our top model only explained 17% of the variation in grouse population indices, suggesting that oth- er factors must be important in explaining grouse cycles (e.g., interactions among fac- tors, variation in grouse counts). For example, there could be an interaction between the number of goshawks and snow quantity/qual- ity. Such an interaction may explain the dif- ferent observations between our study and Lauten (1995) because high quality snow- roosting may provide protective cover during goshawk invasions. Also, there could be in- teractions between predators, grouse food quality, and snow conditions (Gullion 1970). For example, when forage quality decreases. individual grouse may need to forage for lon- ger periods to find adequate food, which would expose them to predation for longer pe- riods of time (Jakubas and Gullion 1991). We concluded that goshawk counts (pre- dation hypothesis) by themselves did not ex- plain Ruffed Grouse population cycles within our study area. It was possible that predation was interacting with winter weather condi- tions in a way that was unknown. Further, we did not evaluate models that hypothesized bot- tom-up mechanisms for control of cycles. For example, Gullion (1984) hypothesized that food quality changes as a result of intensive foraging on aspen by grouse, which in turn affects either survival or reproduction of grouse. However, Jakubas and Gullion (1991) suggested that food quality alone as manifest- ed by high levels of secondary compounds could not mediate a cycle in Ruffed Grouse. Although weather appeared to correlate with Ruffed Grouse population changes better than some other factors, our study was based on indices of both Ruffed Grouse and goshawk numbers, which could confound our results if the indices did not adequately reflect the pop- ulation size or ecological interactions (Ander- son 2001). In addition, we do not have data on Great Horned Owls, mammalian predators, endogenous factors (Matthiopoulos et al. 2005), or parasites (Mougeot et al. 2005), which may also influence Ruffed Grouse pop- ulation trends. Neither Northern Goshawk counts nor win- ter weather appear to explain population in- dices (relative fluctuations of grouse numbers based on drumming display counts) for our study population based on our simultaneous evaluation of plausible factors affecting grouse numbers. Thus, we believe the mech- anism(s) regulating the Ruffed Grouse cycle (at least in northern Minnesota) is still un- known. We suggest that future studies of Zimmerman et al. • RUFFED GROUSE POPULATION CYCLES 247 Ruffed Grouse cycles use robust estimators of grouse counts (e.g., Zimmerman and Gutier- rez 2007) and counts of local predators rather than indices to evaluate the influence of pre- dation, weather, habitat and their interactions on Ruffed Grouse population fluctuations. ACKNOWLEDGMENTS We thank Jana Albers, A. E. Elling, and M. A. Lar- son for providing the tent caterpillar data, weather data, and Minnesota DNR Ruffed Grouse display sur- vey data, respectively. We appreciate the assistance of the Grand Rapids Chapter of the Ruffed Grouse So- ciety. E J. Nicoletti and the Hawk Mountain Bird Ob- servatory provided goshawk data. L. I. Berkeley, L. G. Erickson, and D. D. Grandmaison provided valuable comments on earlier drafts of this manuscript. J. D. Nichols provided valuable discussion during the de- velopment of our modeling approach. C. E. Braun, R. E. Kenward, and Patrik Byholm provided valuable comments that improved the quality and clarity of the manuscript. Support for this research was provided by the University of Minnesota, Minnesota Agriculture Experiment Station, Leigh Perkins Eellowship to G. S. Zimmerman, Donald Rusch Fellowship to G. S. Zim- merman, and the Ruffed Grouse Society. LITERATURE CITED Anderson, D. R. 2001. The need to get the basics right in wildlife field studies. Wildlife Society Bulletin 29:1294-1297. Balzer, C. C. 1995. Survival, hunting mortality, and natality of Ruffed Grouse in northwestern Wis- consin. Thesis. University of Wisconsin, Madison, USA. Bump, G., R. W. Darrow, F. C. Edminster, and W. E Crissey. 1947. The Ruffed Grouse: life history, propagation, management. New York State Con- servation Department, Albany, USA. Burnham, K. P. and D. R. Anderson. 1998. Model selection and inference. A practical information- theoretic approach. Springer- Verlag, New York, USA. Eng, R. L. and G. W. Gullion. 1962. The predation of goshawks upon Ruffed Grouse at the Cloquet Forest Research Center, Minnesota. Wilson Bul- letin 74:227-242. Gullion, G. W. 1970. Factors influencing Ruffed Grouse populations. Transactions of the North American Wildlife and Natural Resources Confer- ence 35:93-105. Gullion, G. W. 1973. Grouse cycles: are they real? Minnesota Science 29:6-1 I. Gullion, G. W. 1984. Grou.se of the North Shore. Wil- low Creek Press, Oshkosh, Wisconsin, USA. Gullion, G. W. and W. H. Marsiiali.. 1968. Survival of Ruffed Grouse in a boreal forest. The Living Bird 7:1 17-167. Gutierrez, R. J., G. S. Zimmerman, and G. W. Ciui.- LiON. 2003. Daily survival rates of Ruffed Grouse Bonasa umbellus in northern Minnesota. Wildlife Biology 9:351-356. Jakubas, W. j. and G. W. Gullion. 1991. Use of quak- ing aspen flower buds by Ruffed Grouse: its re- lationship to grouse densities and bud chemical composition. Condor 93:473-485. Keith, L. B. 1963. Wildlife’s ten-year cycle. Univer- sity of Wisconsin Press, Madison, USA. Keith, L. B. and D. H. Rusch. 1989. Predation’s role in the cyclic fluctuations of Ruffed Grouse. Inter- national Ornithological Congress 19:699-732. Lauten, D. j. 1995. Survival, demography, and be- havior of Ruffed Grouse of different color phase during a cyclic decline in northwestern Wisconsin. Thesis. University of Wisconsin, Madison, USA. Littell, R. C., G. a. Milliken, W. W. Stroup, and R. D. WoLFiNGER. 1996. SAS system for mixed models. SAS Institute Inc., Cary, North Carolina, USA. Matthiopoulos, j., j. H. Halley, and R. Moss. 2005. Socially induced Red Grouse population cycles need abrupt transitions between tolerance and ag- gression. Ecology 86:1883-1893. Mougeot, E, S. a. Evans, and S. M. Redpath. 2005. Interactions between population processes in a cy- clic species: parasites reduce autumn territorial be- haviour of male Red Grouse. Oecologia 144:289- 298. Moss, R. 1986. Rain, breeding success, and distribu- tion of Capercaillie Tetrao urogallus and Black Grouse Tetrao tetri.x in Scotland. Ibis 128:65-72. Rusch, D. H. and S. DeStefano. 1989. To tally the grouse. Pages 200-206 in Ruffed Grouse (S. At- water and J. Schnell, Editors). Stackpole Books, Harrisburg, Pennsylvania, USA. Rusch, D. H., S. DeStefano, M. C. Reynolds, and D. Lauten. 2000. Ruffed Grouse. The birds of North America. Number 515. Small, R. J. 1989. Tipping the balance. Pages 248- 251 in Ruffed Grouse (S. Atwater and J. Schnell, Editors). Stackpole Books, Hamsburg, Pennsyl- vania, USA. Small, R. J., J. C. Holzwart, and D. H. Rusch. 1991. Predation and hunting mortality of Ruffed Grouse in central Wisconsin. Journal of Wildlife Manage- ment 55:512-520. Swenson, J. E., L. Saari, and Z. Bonczar. 1994. Ef- fects of weather on Ha/.el Grouse reproduction: an allometric perspective. Journal of Avian Biology 25:8-14. Tornberg, R., E. Korpimaki, S. Jungell, and V. Reie. 2005. Delayed numerical response of goshawks to population lluctuations of forest grouse. Oikos 1 1 1:408-415. Wiini;, G. C. and R. E. Bi:NNi;rrs. 1996. Analysis of frequency count data using the negative binomial distribution. Ficology 77:2549-2557. ZiMMiiRMAN, Cl. S. AND R. J. CiiniERRi:/,. 2007. The inlluence of ecological factors on detecting drum- ming Ruffed Grouse. Journal of Wildlife Manage- ment 71:1765-1792. The Wilson Journal of Ornithology 120(2);248-255, 2008 COMMUNAL CALLING AND PROSPECTING BY BLACK-HEADED TROGONS (TROGON MELANOCEPHALUS) CHRISTINA RIEHL‘ 2 ABSTRACT. Many species of trogons (Trogoniformes; Trogonidae) gather in mixed-gender calling assem- blages during the breeding season. These assemblages have been compared to leks and are usually assumed to function in mate choice; however, no field data exist to support this hypothesis. I present five non-mutually exclusive hypotheses for the function of mixed-gender calling assemblages and test the predictions using field data from Black-headed Trogons (Trogon melanocephaliis) in a lowland dry forest in Costa Rica. Adult trogons gathered in mobile assemblages of 3-12 individuals that called frequently and chased one another from perch to perch while moving through the forest. Groups formed at sunrise throughout all stages of the breeding cycle and were not significantly male-biased. Both breeding and non-breeding individuals participated in assemblages, and copulations were not observed in assemblages. Prey capture rates of individuals foraging in groups did not differ from those foraging alone. Males called and chased other individuals more often than did females; however, males chased females and other males at equal rates. Assemblages of trogons frequently prospected (investigated potential nest sites and active nests of conspecifics). These data suggest that communal assemblages of calling trogons do not function solely in social mate choice, nor do they enhance foraging efficiency. They may have a role in maintenance of territorial boundaries and selection of future nest sites. Received 2 February 2007. Accepted 20 July 2007. Trogons (Trogoniformes: Trogonidae) are among the most secretive of tropical forest birds. Many aspects of their behavior, partic- ularly those relating to courtship, mate choice, nesting biology, and social interactions remain poorly understood or undescribed (Johnsgard 2000). Trogons are generally solitary; how- ever, assemblages of calling males have been reported in several species (Table 1). These groups usually consist of 3—10 males, which call repeatedly while chasing each other from perch to perch, and a smaller number of fe- males. Most observers have reported that calling assemblages contain more males than females and form only in the breeding season, leading to the widespread assumption these groups function in courtship and mate choice (e.g., Skutch 1972, Brosset 1983, Howell and Webb 1995, Johnsgard 2000). These assemblages have been termed “quasi-leks” (Brosset 1983: 8), “noisy leks” (Howell and Webb 1995: 432), and “lek-like assemblages” (Johnsgard 2000:113), but virtually no field data exist to support this hypothesis. No studies of individ- ' Museum of Comparative Zoology, Harvard Uni- versity, 24 Oxford Street, Cambridge, MA 02138, USA. 2 Current address: Department of Ecology and Evolu- tionary Biology, Princeton University, 106 A Guyot Hall, Princeton, NJ 08544, USA; e-mail: criehl@princeton. edu ually color-marked trogons have been per- formed and the breeding status of individuals in assemblages is usually not known. The be- havior is described only from anecdotal ob- servations and no quantitative data exist on the sex ratios of groups or on the behavior of females participating in groups. There are no reports of copulations occurring in assemblag- es and no information on whether these groups continue to form throughout the breed- ing season. There are several reasons to doubt that mate choice occurs in the context of group displays. Unlike polygynous lekking species, trogons are monogamous and territorial; both parents excavate the nesting cavity, incubate the eggs, and feed the nestlings until fledging. In the few species that have been well studied, the male chooses a nest site, begins to excavate it if necessary, and advertises it to females by calling repeatedly (Skutch 1983, Hall and Ka- rubian 1996, Johnsgard 2000). The quality of the nest site and territory, therefore, appears to be an important component of female choice. Males remain territorial throughout the breeding season, often advertising their territories by calling in synchrony with males in adjacent territories. The few recorded in- stances of trogons copulating have occurred at the nest site with no extra-pair individuals present (Slud 1964, Hall and Karubian 1996). Finally, calling assemblages are mobile; indi- 248 Riehl • BLACK-HEADED TROGON SOCIAL BEHAVIOR 249 TABLE 1. Group size and composition of the 10 trogon species reported to form calling assemblages. Species Group size and composition^* Source Eared Quetzal (Euptilotis neoxenus) 2 M, 1 F Zimmerman 1978 Hispaniolan Trogon {Priotelus roseigaster) 6^ Wetmore and Swales 1931 Choco Trogon {Trogon comptus) Haffer 1975 Black-headed Trogon {T. melanocephalus) <10 M= Howell and Webb 1995 Amazonian White-tailed Trogon {T. viridis) “several M” Johnsgard 2000 Citreoline Trogon {T. citreolus) <10 M** Howell and Webb 1995 Slaty-tailed Trogon {T. massena) “several”’’ Skutch 1972 Violaceous Trogon {T. violaceus) “a number”’’ Skutch 1972 Narina Trogon {Apaloderma narina) <7 M** Brosset 1983 Bare-cheeked Trogon (A. aequatoriale) <6 M** Brosset 1983 ^ M = male, F = female. No information available on sex ratio. Unspecified number of females also present. viduals move together as a group rather than gathering and displaying at a fixed arena (Cunningham van-Someren 1973, Brosset 1983, Johnsgard 2000). The hypothesis that communal calling as- semblages function in social mate choice gen- erates the following predictions: (1) assem- blages should form most frequently at the on- set of the breeding season, prior to egg-laying; (2) individuals that are tending nests should not participate in calling assemblages; (3) males should call more frequently than fe- males; and (4) the sex ratio of assemblages should be male-biased. Alternatively, group calling assemblages could serve one or more of several functions. These hypotheses are not mutually exclusive, but do generate different predictions concern- ing timing of group formation, behavior of males and females, and presence of breeding individuals in assemblages. Extra-pair Mate Choice. — Communal call- ing assemblages are unlikely to have a role in social mate choice, but they could provide an opportunity for already-mated trogons to gain extra-pair copulations. Recent molecular anal- yses of paternity have revealed the majority of socially monogamous birds are genetically promiscuous; in fact, true genetic monogamy is the exception rather than the rule (Griffith et al. 2002). This hypothesis predicts that: (1) assemblages should form throughout the breeding season; (2) males should call and chase more frequently than females; (3) males should chase females more often than other males; and (4) females that are tending young should not participate in assemblages, where- as males that are tending young should partic- ipate. Foraging Benefits. — Trogons in groups might forage more efficiently than lone indi- viduals if groups are better able to locate prey, watch for predators, exclude interspecific competitors from foraging areas, or invade territories defended by lone pairs (Alexander 1974, Clark and Mangel 1986). This hypoth- esis does not predict behavioral differences among males and females in calling assem- blages, but it does predict that: (1) assemblag- es should not be restricted to the breeding sea- son; (2) foraging efficiency (prey capture rate) should be higher for individuals in groups than for lone individuals; and (3) both breed- ing and non-breeding individuals should par- ticipate in assemblages. Nest-site Prospecting. — Individuals might assemble into groups to investigate potential nest sites and active nests of conspecifics if predation risk is lowered or if individuals ben- efit by following other individuals and sharing information (Patterson and Makepeace 1979). This hypothesis predicts that: (1) assemblages should form throughout the breeding season; (2) groups should be observed investigating potential nest sites; and (3) groups should be composed primarily of non-breeding individ- uals. Maintenance of Territorial Boundaries. — Communal displays might provide an oppor- tunity lor pairs to advertise their presence to their neighbors and to re-enforce territorial boundaries. If so, (1) assemblages should 250 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 form throughout the breeding season; (2) males should call and chase more frequently than females; (3) males should chase males more often than females; and (4) groups should be composed primarily of breeding in- dividuals. In this paper, I document mixed-gender calling assemblages in a breeding population of Black-headed Trogons {Trogon melanoce- phalus) in a lowland dry forest in Costa Rica, and use field observations to test hypotheses for the adaptive benefits of communal calling in this species. Black-headed Trogons form assemblages of up to 10 individuals of both genders during the breeding season (Howell and Webb 1995), but little is known of the reproductive behavior of the species aside from anecdotal accounts of such groups. Skutch’s (1948) study, based primarily on ob- servations at a single nest, remains the most thorough account of the natural history of this species. I monitored 14 Black-headed Trogon nests, followed foraging individuals, and re- corded assemblages of calling trogons during one breeding season, June— August 2004. I re- corded the total number of individuals in each group, the number of males and females, in- stances of prospecting behavior, calling rates, and prey capture rates of focal individuals of both genders and, when possible, the breeding status of individuals. METHODS Barque Nacional Santa Rosa (SRNP) is a semi-forested 10,800-ha preserve on the Pa- cific coastal plain of northwestern Guanacaste Province, Costa Rica (10° 45' 11 00 N, 85 30'— 45' W) and is part of the 120,000-ha Area de Conservacion Guanacaste (ACG). Lowland dry forest predominates, which is highly sea- sonal and largely deciduous (Janzen 1988). I located and monitored 14 Black-headed Tro- gon nests by walking trails and transects through the SRNP, listening for calling tro- gons, and searching for potential nest sites (cavities in arboreal Nasutitermes termitaria). Nests were observed from either a concealed place or a blind at least 20 m from the nest. Each nest with chicks was observed for a min- imum of 1 1 hrs and I followed foraging adults from 10 of the 14 nesting pairs (focal animal sampling; Altmann 1974). Each focal adult was observed for a minimum of 15 hrs; ob- servation time was equally divided between morning (0530-1200 hrs CST) and afternoon (1200-1830 hrs CST) periods. Black-headed Trogons often gather in mo- bile assemblages of 3—12 individuals that perch in the same or adjacent trees, call fre- quently, and chase one another from perch to perch (Howell and Webb 1995, Johnsgard 2000). I often encountered these groups while following focal adults and while locating sing- ing trogons at sunrise. I recorded the time of day for each encounter, time since sunrise, date, total group size, number of individuals of each gender, reproductive status of individ- uals (if known), calling rates of focal individ- uals, and prey capture rates of focal individ- uals. Breeding individuals could often be identified by distinctively damaged rectrices (which become frayed and bent to one side while incubating or brooding in the nesting cavity). When one individual flew at another and supplanted it on a perch, I recorded the gender of both birds. Group members did not assemble at fixed sites, but traveled slowly to- gether through the forest. I followed groups on foot until they dispersed and used a hand- held GPS unit to record the total distance trav- eled by the group (path length) as well as the net distance traveled. Groups were considered to have dispersed when ^ two individuals re- mained in the same tree or when < one in- dividual continued to vocalize. I used the ArcView extension programs ANIMAL MOVEMENT and SPATIAL ANALYST to project coordinates into UTMs, and to calcu- late total distances and net distances traveled by assemblages (ArcView Projection Utility Wizard; Hooge and Eichenlaub 1997). I recorded trogons investigating potential nest sites and active nests while observing nests, lone individuals, and calling groups. Prospecting birds were identified following Doligez et al. (2004): (1) presence of extra- pair individuals near active nests, identified by molt patterns and/or feather wear; (2) simul- taneous presence of two individuals of the same gender near active nests; (3) observa- tions of individuals repeatedly inspecting con- tents of a nest from the entrance of the cavity without entering; or (4) observations of indi- viduals perching on and/or pecking potential nest sites (unexcavated termitaria). Several of these criteria often occurred together. Riehl • BLACK-HEADED TROGON SOCIAL BEHAVIOR 251 All data were tested for normality (Kol- mogorov-Smirnov goodness-of-fit test) and nonparametric tests were used when data were not normally distributed. I used log-likelihood G-tests to test for effect of time of day on the probability of observing assemblages. I used linear regressions with F-tests for significance to examine the effect of date on assemblage size and on the probability of observing as- semblages (Zar 1999). I used replicated G-tests to test for heterogeneity and goodness- of-fit to a sex ratio of unity (Sokal and Rohlf 1995) to examine whether calling assemblages consisted of equal numbers of males and fe- males. Chasing data (incidences of a focal in- dividual supplanting another individual on a perch) were converted to rates (chases/min) for each observation period. All pair-wise combinations (male supplanting male, male supplanting female, female supplanting male, and female supplanting female) were com- pared using ANOVA. I used a Mann-Whitney G-test to examine whether calling rates dif- fered between focal males and females in groups, and a r-test to examine whether prey capture rates differed between focal individ- uals foraging alone and those in groups. All tests were two-tailed and results were consid- ered significant at a = 0.05. Analyses were conducted with SPSS 12.0 (2003). RESULTS 1 encountered 25 calling assemblages of Black-headed Trogons in 79 days (677 hrs) of observations of nests and focal adults. Assem- blages were more likely to be observed during morning than afternoon periods (log-likeli- hood G-test, G - 20.72, df = 1, F < 0.001); all but two assemblages were observed in the early morning soon after sunrise (assembly, x ± SD == 62 ± 27 min after sunrise; dispersal, X = 134 ± 39 min after sunrise). Group size ranged from 3 to 12 individuals (Jc ± SD = 5.6 ± 2.4). There was no difference in average number of adult males and females in the as- semblages (male, T ± SD = 3.45 ± 1.28 in- dividuals/group; female, x = 2.05 ± 0.94 in- dividuals/group; replicated G-tests; Gj = 13.77, df = 25, P = 0.98). Assemblages of Black-headed Trogons did not gather at fixed sites; members of a group moved slowly together through the forest for up to 1,300 m before dispersing (path length X ± SD = 870 ± 340 m; net distance x - 562 ± 299 m). Groups were fairly cohesive; all individuals usually stayed in the same or ad- jacent trees, often in the crown of a large can- opy tree (Enterolobium, Spondias, Pithecel- lobium). Both males and females in the groups foraged gleaning caterpillars and other insects from the canopy foliage, and called continu- ally. Calling assemblages were observed during all 3 months of the study period. Date had no effect on frequency of assemblages (F = 0.0046, df = 11, F, ,0 = 0.046, P = 0.83; Fig. lA), or on group size (F = 0.0014, df = 25, ^1,24 = 0.033, P = 0.86; Fig. IB). Both breed- ing and non-breeding adults participated in as- semblages; at least 9 of 25 groups (36%) con- tained one or more individuals known to be tending nests {x = 0.44 ± 0.65 breeding in- dividuals per group). Not all breeding adults could be identified and these values are con- servative. Focal males chased other individuals from perch to perch significantly more often than did focal females (ANOVA, F376 = 15.52, P < 0.001; Fig. 2). However, males were not significantly more likely to chase other males than females (ANOVA, F, 33 = 0.049, P = 0.83; Fig. 2). Focal males in assemblages also called more often than did focal females (Mann-Whitney U = 128, F < 0.001). On av- erage, prey capture rates for focal individuals foraging in groups (T ± SE = 1.03 ± 0.18 captures/min) did not differ from those for- aging alone {x = 0.94 ± 0.13; /-test, /^ = 0.405, F = 0.69; Fig. 3). I observed trogons prospecting in groups (investigating neighboring nests and potential nest sites) on 12 occasions (Table 2). Pros- pectors included birds of both genders that were attending nests (// = 4 individuals) as well as birds of unknown reproductive status (n = 71). Most observations were of trogons investigating potential nest sites (arboreal Na- siititermes termite nests) that were not exca- vated (// = 10). In addition, extra-pair trogons were observed looking into occupied nests (/? = 2). Lone trogons also prospected (// = 9 individuals; 5 males of unknown reproductive status, 3 females of unknown reproductive sta- tus, and I breeding female). 252 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 5 - o o 4 - (/) V oi (0 3 - n E » 0) 0.4 0.2 Assemblages Lone FIG. 3. Mean number (+ SE) of prey captures/min/observation period for focal Black-headed Trogons foraging in assemblages and alone (Parque Nacional .Santa Rosa, Costa Rica. .hin-Aug 2004) (// = 20 observation periods. 640 min total for each category). 254 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 TABLE 2. Prospecting by groups of Black-headed Trogons (n = 12 prospecting events) in Parque Nacional Santa Rosa, Costa Rica, 2004. Date Number of males in group Number of females in group Status of termitarium being investigated Reproductive status of prospecting birds^ 09 Jun 3 2 Unoccupied Unknown 15 Jun 4 4 Unoccupied 1 hr male, 7 unknown 23 Jun 3 3 Unoccupied Unknown 27 Jun 4 1 Unoccupied Unknown 04 Jul 4 3 Unoccupied 2 br male, 7 unknown 10 Jul 5 3 Unoccupied Unknown 21 Jul 3 4 Unoccupied Unknown 24 Jul 1 1 Occupied (chicks) Unknown 07 Aug 3 2 Unoccupied 1 br female, 4 unknown 18 Aug 4 4 Occupied (chicks) Unknown 26 Aug 6 3 Unoccupied Unknown 29 Aug 1 2 Unoccupied Unknown 2 br = breeding; tending either eggs or chicks. peeling has an important role in the dynamics of dispersal and habitat use within a popula- tion, and is more widespread than previously thought (Boulinier et al. 1996, Reed et al. 1999). Studies in the temperate zone have found that prospecting birds gather while in- vestigating nearby nest sites, which influences their choice of future nest sites (Zicus and Hennes 1989, Part and Doligez 2003, Doligez et al. 2004). The study of prospecting behav- ior in tropical birds is still in its infancy and group prospecting has been reported in only a few species (Patterson and Makepeace 1979, Eadie and Gauthier 1985). It is also possible that calling assemblages allow breeding adults (particularly males) to delineate and reinforce territorial boundaries between adjacent territories. The hypothesis that group displays function in maintenance of territories is supported by the findings that males in assemblages chased other individuals more often than did females, and chased in- dividuals of both genders. Males in assem- blages also called more frequently than did females; individual males at the nest adver- tised their territories by calling in synchrony with males in adjacent territories (with as many as 4 males calling simultaneously; CR, unpubl. data). Males from neighboring terri- tories were also observed mobbing potential nest predators together; five males gathered in response to alarm calls at one nest (CR, un- publ. data). Calling assemblages usually formed at or immediately after sunrise, raising the possi- bility that function of communal calling in tro- gons is similar to that of the dawn chorus of passerines. Like trogon assemblages, the dawn chorus was originally thought to have a role in mate attraction; however, dawn singing also extends throughout the breeding season and may also function in territory defense (Staicer et al. 1996). Prospecting individuals might use a dawn chorus to assess singing residents and territory occupancy; non-territorial male Com- mon Nightingales {Luscinia megarhynchos), for example, prospect for territories primarily in the hour before sunrise, when singing by resident males is at a peak (Amrhein et al. 2004). Chorusing and prospecting may both peak at dawn if foraging is less profitable at this time of day. Alternatively, non-territorial individuals may find it most informative to prospect for vacant territories at a time when most resident males are singing (Hutchinson 2002, Amrhein et al. 2004). Non-territorial trogons might therefore use dawn calling as- semblages to assess the status of occupied nests and prospect for potential nest sites, whereas resident trogons might call to adver- tise their presence and defend their territories against floaters and neighboring territory holders. Future studies of communal calling in tro- gons should confirm the breeding status and age class of individuals participating in as- semblages, and should investigate the effect of prospecting behavior on future nest site se- lection and reproductive success. Observa- tions of color-marked birds are needed to as- Riehl • BLACK-HEADED TROGON SOCIAL BEHAVIOR 255 certain whether groups are composed of the same individuals that gather to display day af- ter day, or whether groups also include non- territorial floaters. ACKNOWLEDGMENTS I thank D. H. Janzen, J. R. Trimble, D. S. Causey, J. A. Klemens, S. J. Agosta, the staff of Parque Na- cional Santa Rosa, and the Museum of Comparative Zoology at Harvard University for logistical support. J. O. Coulson, T. D. Coulson, and two anonymous re- viewers provided invaluable advice and comments on the manuscript. I am grateful to the Harvard College Research Program for providing funding for this proj- ect. LITERATURE CITED Alexander, R. A. 1974. The evolution of social be- havior. Annual Review of Ecology and System- atics 5:325-383. Altmann, j. 1974. Observational study of behavior: sampling methods. Behaviour 49:227-267. Amrhein, V., H. P. Kunc, and M. Naguib. 2004. Non- territorial nightingales prospect territories during the dawn chorus. Proceedings of the Royal Acad- emy of Sciences Series B (Supplement) 271: S167-S169. Brosset, a. 1983. Parades et chants collectives chez les couroucous du genre Apaloderma. Alauda 5 1 : 1-10. Boulinier, T, E. Danchin, J.-Y. Monnat, C. Doutre- LANT, AND B. Cadiou. 1996. Timing of prospect- ing and the value of information in a colonial breeding bird. Journal of Avian Biology 27:252- 256. Clark, C. W. and M. Mangel. 1986. The evolution- ary advantages of group foraging. Theoretical Population Biology 30:45-75. Cunningham van-Someren, G. R. 1973. Calling and flight display assembly in male trogons. Bulletin of the East African Natural History Society 81: 1 17-260. Doligez, B., T. Part, and E. Danchin. 2004. Pros- pecting in the Collared Flycatcher: gathering pub- lic information for future breeding habitat selec- tion? Animal Behaviour 67:457-466. Eadie, j. M. and G. Gauthier. 1985. Prospecting for nest sites by cavity-nesting ducks of the genus Bu- cephala. Condor 87:528-534. Griffith, S. C., I. P. E Owens, and K. A. Thuman. 2002. Extra pair paternity in birds: a review of interspecific variation and adaptive function. Mo- lecular Ecology 11:2195-2212. Haffer, j. 1975. Avifauna of northwestern Colombia, South America. Bonner Zoologische Monogra- phien 7: 1-1 82. Hall, L. S. and J. O. Karubian. 1996. Breeding be- havior of Elegant Trogons in southeastern Arizo- na. Auk 1 13:143-150. Hooge, P. N. and B. Eichenlaub. 1997. Animal move- ment extension to Arcview, Version 1.1. USGS, Alaska Biological Science Center, Anchorage, USA. Howell, S. N. G. and S. Webb. 1995. A guide to the birds of Mexico and northern Central America. Oxford University Press, New York, USA. Hutchinson, J. M. C. 2002. Two explanations of the dawn chorus compared: how monotonically changing light levels favor a short break from singing. Animal Behavior 64:527-539. Janzen, D. H. 1988. Ecological characterization of a Costa Rican dry forest caterpillar fauna. Biotro- pica 20:120-135. JoHNSGARD, P. A. 2000. Trogons and quetzals of the world. Smithsonian Institution Press, Washington, D. C., USA. Part, T. and B. Doligez. 2003. Gathering public in- formation for habitat selection: prospecting birds cue on parental activity. Proceedings of the Royal Society of London, Series B, 270:1809-1813. Patterson, I. J. and M. Makepeace. 1979. Mutual interference during nest prospecting in the Shel- duck, Tadorna tadorna. Animal Behaviour 27: 522-535. Reed, J. M., T. Boulinier, E. Danchin, and L. W. Oring. 1999. Informed dispersal: prospecting by birds for breeding sites. Current Ornithology 15: 189-259. Skutch, a. E 1948. Life history of the Citreoline Tro- gon. Condor 50:137-147. Skutch, A. F. 1972. Studies of tropical American birds. Publications of the Nuttall Ornithological Club, Number 10. Nuttall Ornithological Club, Cambridge, Massachusetts, USA. Skutch, A. E 1983. The birds of tropical America. University of Texas Press, Austin, USA. Slud, P. 1964. The birds of Costa Rica: distribution and ecology. Bulletin of the American Museum of Natural History 128:1-430. SoKAL, R. R. AND F. J. Rohlf. 1995. Biometry: the principles and practice of statistics in biological research. W. H. Freeman and Company, New York, USA. SPSS Institute Inc. 2003. SPSS for Windows. Version 12.0. SPSS Institute Inc., Chicago, Illinois, USA. Staicer, C. a., D. a. Spector, and A. G. Horn. 1996. The dawn chorus and other diel patterns in acous- tic signaling. Pages 426-453 in Ecology and evo- lution of acoustic communication in birds (D. E. Kroodsma and E. H. Miller, Editors). Cornell Uni- versity Press, Ithaca, New York, USA. Wetmore, a. and B. H. Swales. 1931. The birds of Haiti and the Dominican Republic. Bulletin of the U. S. National Museum 155:1-483. Zar, j. H. 1999. Biostatistical analysis. Fourth Edition. Prentice Hall. Upper Saddle River, New Jersey, USA. Zicus, M. C. AND S. K. Hennes. 1989. Nest prospect- ing by Common Goldeneyes. Condor 9 1 :807-8 1 2. Zimmerman. D. A. 1978. Eared Trogon — immigrant or visitor? American Birds 32:135-139. The Wilson Journal of Ornithology 1 20(2):256— 267, 2008 SONG VARIATION IN BUFF-BREASTED ELYCATCHERS (EMPIDONAX FULVIFRONS) M. ROSS LEIN' ABSTRACT. — I examined song variation within and among 23 individual Buff-breasted Flycatchers (Empi- donax fulvifrons) recorded in the Chiricahua and Huachuca mountains of Arizona in 1999. I recorded two distinct song types from each individual during intense pre-dawn singing. I used both spectrographic cross correlation (SPCC) of entire songs and discriminant function analysis (DFA) of temporal and frequency measurements to examine whether songs were individually distinctive, and whether songs differed between the two localities. Similarity values of pairs of songs from SPCC were significantly greater for within-male than for between-male comparisons for both song types. Mean similarity values for the two song types did not overlap between these comparison categories. Similarity values between songs of pairs of males from the same mountain range were not greater than for comparisons between pairs of males from different ranges. All temporal and frequency measures for both song types varied significantly more among than within individuals. DFA of principal com- ponent scores derived from these measures assigned 85% of Type 1 and 86% of Type 2 songs to the correct individual. Only three frequency variables measured from Type 1 songs differed significantly between birds from the two mountain ranges. DFA assigned only 61% of songs of either type to the correct mountain range, not significantly greater than expected by chance. Thus, both techniques demonstrate significant individual distinctiveness in songs of this species, and neither suggests any geographic structuring of song variation between the two mountain ranges. However, SPCC is considerably more efficient and has greater potential to assign unknown recordings to known individuals correctly, and to detect recordings of “new” individuals not included in the reference sample. Received 24 April 2007. Accepted 19 July 2007. Most literature on song variation deals with oscine birds (Passeriformes, suborder Passeri) (Lovell and Lein 2004). Songs are learned in most or all oscines (Kroodsma 1996), which is a major factor generating song variation at individual, population, and geographic scales. Suboscine birds (suborder Tyranni) appear to show lower levels of song variation. The ab- sence of learning during song ontogeny has been demonstrated for a few species of North America tyrant flycatchers (Tyrannidae) (Kroodsma 1996), and this has been general- ized to all suboscines based primarily (and cir- cumstantially) on the limited variation in their songs relative to oscines. Early studies used song variation in subos- cine passerines to help resolve the taxonomic status of populations (e.g., Stein 1963; John- son 1963, 1980; Lanyon 1978) or described how song variants were used in communica- tion systems (Smith 1969, 1970, 1988). None of these studies focused explicitly on the na- ture of song variation among individuals with- in local populations. Song variation among in- dividuals, albeit minor, is apparent in pub- ' Department of Biological Sciences, University of Calgary, Calgary, AB T2N 1N4, Canada; e-mail: mrlein@ucalgary.ca lished spectrograms of a variety of suboscine species (e.g., Stein 1963, Payne and Budde 1979, Kroodsma 1984). However, in contrast to the large, often qualitative, differences which may be obvious in spectrograms of dif- ferent individuals (e.g., Borror 1960) or pop- ulations (e.g., Baptista and King 1980) of os- cines, examination of suboscine song varia- tion requires quantitative analysis of suffi- ciently large samples of recordings. The few analyses conducted to date have demonstrat- ed, for example, that songs of Alder Flycatch- ers (Empidonax alnorum) (Lovell and Lein 2004) and Acadian Flycatchers {E. virescens) (Wiley 2005) are individually distinctive, and that songs of the endangered Southwestern Willow Flycatcher {E. traillii extimus) differ significantly from those of a neighboring sub- species {E. t. adastus) (Sedgwick 2001). How- ever, more studies are needed to demonstrate the generality of such patterns, especially for other groups of suboscines. Most quantitative studies of song variation have used univariate or multivariate analyses of temporal and frequency characters mea- sured from spectrograms. Characters are usu- ally selected on the basis of their ease of mea- surement with the assumption that if enough characters are measured, the analysis will cap- 256 Lein • SONG VARIATION IN BUFF-BREASTED FLYCATCHERS 257 ture the features that may be important. How- ever, it is often difficult to select characters that can be measured in an objective manner for all songs, and characters are often ex- tremely general (e.g., number of notes, max- imum frequency, etc.), capturing little of the “structure” of the vocal signal. Digital processing of acoustic signals pro- vides alternative methods for characterizing variation among songs. Peter Marler and co- workers (Clark et al. 1987) first used spectro- graphic cross correlation (SPCC) to character- ize variation in notes of songs of Swamp Sparrows (Melospiza georgiana). SPCC com- pares two digital spectrograms at successive offsets on the time axis and calculates nor- malized covariance (ranging from - 1 to 1 ) at each offset. The maximum covariance is used as a measure of similarity between the two signals (Baker and Logue 2003). SPCC has the advantage that distribution of sound en- ergy in both frequency and time are compared in a “holistic” manner, probably capturing more of the relevant features of the song than would a small number of quantitative mea- surements (Khanna et al. 1997). Several bioa- coustical analysis software packages include routines that automate such measurements. The CORMAT routine of the SIGNAL digital signal analysis software (Engineering Design, Berkeley, CA, USA) can run SPCC on sam- ples of up to 200 songs simultaneously. The Buff- breasted Flycatcher (Empidonax fulvifrons) is the least known of the 1 1 species of this genus that breed in Canada or the Unit- ed States (Bowers and Dunning 1994). It breeds in montane forests through Mexico south to Guatemala, El Salvador, and Hon- duras (AOU 1998), but has been studied al- most exclusively in extreme southeastern Ar- izona where small numbers of birds, possibly less than 100 pairs in total (Martin 1997), in- habit several isolated mountain ranges. I re- corded songs of male Buff-breasted Flycatch- ers during 1999 in the Chiricahua Mountains and the Huachuca Mountains, the two ranges with the largest populations of this species in Arizona. Male Buff-breasted Flycatchers use two distinctive song types, which I designate as Type 1 and Type 2, during pre-dawn singing and during strong daytime singing (Fig. 1 ). Bowers and Dunning (1994) published spec- trograms of Type 1 songs, which they describe as “chee-lick”, but did not describe or illus- trate the Type 2 song. Type 2 songs are rarely used during the sporadic daytime singing typ- ical of the breeding cycle following pairing. They are similar to Type 1 songs in sound, but the higher frequency is obvious to the ear. My objective was to characterize patterns of song variation within and among individ- uals in the Chiricahua and Huachuca moun- tains. I conducted both spectrographic cross- correlation and multivariate analyses of song characters for the same sample of songs to answer two questions. First, are songs of Buff- breasted Flycatchers individually distinctive? Second, are there detectable differences in songs of birds between the two mountain ranges? METHODS Study Area and Field Methods. — I recorded 23 individual male Buff-breasted Flycatchers at seven sites in two isolated mountain ranges in southeastern Arizona between 7 and 25 June 1999. Specific recording localities in the Chiricahua Mountains (number of individuals in parentheses) included Cave Creek (3), Pin- ery Canyon (2), Rucker Canyon (6), and West Turkey Creek (2). Localities in the Huachuca Mountains were Carr Canyon (3), Garden Canyon (1), and Sawmill Canyon (6). Males at West Turkey Creek and Sawmill Canyon were recorded on multiple dates whereas in- dividuals at other sites were recorded only on one or two dates. Buff-breasted Flycatchers breed in open woodlands dominated by pines {Pinus spp.), live oaks {Quercus spp.), and al- ligator juniper {Juniperus deppeana). Sites within the same mountain range were sepa- rated by maximum distances of 22 km (Chir- icahua) and 7 km (Huachuca), whereas breed- ing populations in the two mountain ranges are separated by more than 100 km of un- suitable desert habitat. Birds arrived on terri- tories in mid-April and paired immediately. However, nesting did not begin until late May. Most pairs were involved in nest-building or egg-laying during the period of recording. Recordings of pre-dawn singing were made between 0436 and 0518 hrs MS'f using a Sony TCD-DIO Proll DAT recorder and a Tel- inga Proll parabolic microphone, or a Sony TC-D5 Proll cassette recorder and a Audio- 258 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 DPIN , BLOG ■ f\k FLOG 0.2 0.4 Time (s) SPRG - w A FIG. I. Audiospectograms of songs of Buff-breasted Flycatchers. Each panel contains examples of Type 1 and Type 2 songs for an individual male. The four males in the left column are from the Chiricahua Mountains and those in the right column are from the Huachuca Mountains, Arizona. Tech ATS 15a “shotgun” microphone. I used recordings of pre-dawn singing because each male sang strongly for 15-20 min prior to dawn each day, even during phases of the breeding cycle when daytime singing was rare and sporadic. Males used relatively low perch- es during pre-dawn singing and could be ap- proached quite closely; recordings made at this time were of high quality. Males at West Turkey Creek and Sawmill Canyon were banded and color-marked for in- dividual identification; males at other sites were unmarked. However, such markings were not visible before the end of pre-dawn singing. Misidentification of males was highly improbable because there were few individu- als at any site and males on neighboring ter- ritories could be heard clearly while recording pre-dawn singing of individuals. Processing of Recordings. — Recordings were acquired as digital files using RTSD Ver- sion 2.0 bioacoustical analysis software (En- gineering Design, Berkeley, CA, USA) with a sample rate of 20,000 Hz and 16-bit amplitude resolution. Analog signals were filtered during acquisition with a Krohn-Hite Model 3550 fil- ter to avoid aliasing. I selected single record- ings for each individual for subsequent sam- pling based on high signal-to-noise ratios and adequate numbers of songs. Eight examples of each song type were extracted as individual sound files from each recording using SIG- NAL 4.0 bioacoustical analysis software (En- gineering Design, Berkeley, CA, USA). Many recordings contained hundreds of songs and I selected samples in quasi-random manner. Only one song of each type was selected from each 20-sec segment of the recording with a Lein • SONG VARIATION IN BUFF-BREASTED ELYCATCHERS 259 high signal-to-noise ratio. One or more seg- ments were skipped between each segment from which songs were selected if more than eight segments met the criterion. I selected songs within each 20-sec segment with mini- mal amounts of reverberation or background noise. SIGNAL 4.0 was used to band-pass fil- ter sampled songs, removing noise outside the frequency range of interest (1,500-7,000 Hz for Type 1 songs and 1,500-8,000 Hz for Type 2 songs), and to normalize all sampled songs to the same root-mean-square ampli- tude. Use of single recordings for each individual may underestimate the amount of variation in songs within individuals, but was necessitated by the limited number of recordings available for many males. However, analyses of songs of individual Alder Flycatchers (Lovell and Lein 2004), Willow Flycatchers {E. traillii) (MRL, unpubl. data), and Dusky Flycatchers (E. oberholseri) (Stehelin 2005) demonstrate little variation among recordings across the breeding season. They also show that patterns of variability within and among individuals are similar when comparisons are made using either single recordings of individuals or mul- tiple recordings made on different dates. Analysis Using Spectrographic Cross Cor- relation.— Spectrographic cross correlation (SPCC) was conducted using the CORMAT routine in SIGNAL 4.0. Similarity values cal- culated by CORMAT are somewhat sensitive to the exact parameters used to generate spec- trograms because of the trade-off between time resolution and frequency resolution in- herent in the Fast Fourier Transform proce- dure. Preliminary analyses indicated that spec- trograms with a Hanning window (WINDOW = HANN), 64-point transforms (XFTLEN = 64), and 500 steps (XFTSTP == 500) provided maximum similarity values. These produced “wide-band” spectrograms with time and fre- quency resolutions of 3.2 ms and 312.5 Hz, respectively. Use of a fixed number of steps is justified because CORMAT adds zero-am- plitude segments to short signals to bring all component signals to the length of the longest signal. However, because of the different du- rations of the two types of songs, transform intervals and overlaps varied slightly between song types (0.45 ms and 86% for Type 1 songs, 0.54 ms and 83% for Type 2 songs). CORMAT produces a lower triangular half- matrix in which each value is the peak cross correlation value for a pair of signals. Values for comparisons of a signal with itself and re- ciprocal comparisons between each pair of signals are omitted. SPCC of 184 signals re- sulted in a half-matrix of 16,836 {n{n — l)/2) comparisons of different pairs of songs of each type. I wrote a Fortran program to cal- culate mean values for the 28 comparisons among the eight songs sampled from each male (within-male similarity) and the 64 com- parisons of songs between each pair of males (between-male similarity), creating a 23 X 23 triangular half-matrix for each song type. Mean similarity values in these two matrices were averaged to create a third matrix con- taining a mean index of similarity for each male-male comparison. Normal parametric tests could not be used to evaluate differences between within-male and between-male similarity values because the similarity values in the matrices were not independent observations (each song of each male was used in multiple comparisons). Ran- domization tests (Manly 1997) using Resam- pling Stats software (Blank et al. 2001) and 1,000 iterations tested whether the difference between mean within-male and mean be- tween-male similarity values in each matrix was greater than expected by chance. I divided the between-male similarity val- ues into those involving comparisons within a single range and those involving comparisons between ranges to examine differences in songs between the two mountain ranges. 1 tested for significant differences between within-range and between-range similarity values using Mantel tests (Sokal and Rohlt 1995). Similarity values were converted to dissimilarity values by subtracting each from 1. The half-matrices of dissimilarity values were compared with a design half-matrix con- taining a 0 in each cell representing a within- range comparison and a 1 in each cell repre- senting a between-range comparison. Analysis Using Multivariate Compari- sons.— A series of temporal and frequency variables was measured or calculated for the same songs used in the SPCC analysis (8 ex- amples of each song type for each of 23 in- dividuals). A representative sample of songs was examined to ascertain which temporal 260 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 TABLE 1. Time and frequency variables measured or calculated from audiospectrograms of Type 1 and Type 2 songs of Buff-breasted Flycatchers. Variables indicated with * were measured only for Type 1 songs. Code Variable (units) DURN PKIT PK2T PK3T* VALT STRF ENDF PKIF PK2F PK3F* VALE P2VFR P1P2FR PIPN P2PN P3PN* VPN Duration of entire song (ms) Duration from start of song of first frequency “peak” (ms) Duration from start of song of second frequency “peak” (ms) Duration from start of song of third frequency “peak” (ms) Duration from start of song of frequency “valley” (ms) Frequency at start of song (Hz) Frequency at end of song (Hz) Maximum frequency at first frequency “peak” (Hz) Maximum frequency at second frequency “peak” (Hz) Maximum frequency at third frequency “peak” (Hz) Minimum frequency at frequency “valley” (Hz) Frequency difference between first frequency peak and frequency valley (Hz) Frequency difference between first and second frequency peaks (Hz) Relative time of first frequency peak (PKIT ^ DURN) Relative time of second frequency peak (PK2T ^ DURN) Relative time of third frequency peak (PK3T DURN) Relative time of frequency valley (VALT DURN) and frequency characteristics could be mea- sured with objectivity and were repeatable pri- or to final measurements. Sound in both song types is modulated up and down in frequency repeatedly, producing a series of “peaks” and “valleys” on the spectrogram. The first peak is identical between song types for each in- dividual. Type 1 songs have one more mod- ulation than Type 2 songs resulting in one ad- ditional peak and valley. The final set of var- iables measured included the duration of the song, and the times and frequencies of the peaks (3 for Type 1 songs, 2 for Type 2 songs) and valleys. No measurements were made for the first valley in Type 1 songs because some individuals exhibited a break in sound pro- duction at this point with no clear inflection in frequency. Temporal and frequency measurements were made in a semi-automated fashion using custom programs written in the SIGNAL command language. Points were measured on a spectral contour generated from the spectro- gram. The spectral contour tracks the frequen- cy with the maximum sound energy at a given time. Sound amplitude increases gradually at the start of songs and fades out gradually at the end, and the apparent locations of these points can be shifted on spectrograms by mod- ifying the parameters controlling spectrogram intensity. Consequently, start and end times of the song were defined arbitrarily as the points at which the spectral contour exceeded (start) or fell below (end) an amplitude threshold of 20 dB below the maximum amplitude of the spectrogram. This procedure resulted in a con- sistent approximation of these times because the amplitude of all songs was normalized pri- or to analysis. Frequencies at the start and end of the song were extracted from the spectral contour at these times. Similarly, times of peaks and valleys were defined as the points of local maxima (peaks) or minima (valleys) of frequency in the spectral contour. Five tem- poral and six frequency variables were mea- sured for Type 1 songs (Table 1). Only four temporal and five frequency variables were measured for Type 2 songs because these songs have one fewer frequency peak. Addi- tional variables (6 for Type 1 songs, 5 for Type 2 songs) were calculated from the mea- sured variables (Table 1). I calculated coefficients of variation (CV) for each variable for each song type to quan- tify the magnitude of variability. I calculated within-male coefficients of variation (CV^) to measure variation within a single recording and among-male coefficients of variation (CVJ from the variable means from each male. A one-way ANOVA was conducted on each variable to compare within-male and among-male variation. I conducted principal components analyses (PCA) on the data sets for each song type be- Lein • SONG VARIATION IN BUFF-BREASTED ELYCATCHERS 261 EIG. 2. Audiospectograms of Type 1 and Type 2 songs of Buff-breasted Flycatchers indicating the reference points for temporal and frequency measurements (STR = Start; PK = Peak; VAL = Valley). cause some of the variables were highly cor- related with one another. This produced a smaller number of uncorrelated variables (PC scores) for each song, which were entered into a discriminant function analysis (DFA) to ex- amine whether songs of different individuals could be distinguished. Results of jack-knifed classifications, in which each song was as- signed to an individual using discriminant functions calculated from all songs in the data set except the one being classified, are report- ed as percentages of songs assigned correctly. This is a conservative estimate of the power of the classification procedure (Manly 1994). I conducted a similar analysis to test wheth- er songs of individuals from the two mountain ranges could be reliably distinguished. This analysis used mean values of variables for the eight songs of each type for each individual to maintain sample independence. SYSTAT 10.2 was used for parametric sta- tistical analyses. Probability plots were used to check variables for an approximate fit to a normal distribution before parametric statisti- cal tests were applied. I report mean values ± SD and use a criterion of a < 0.05 for statis- tical significance. RESULTS Visual examination of spectrograms re- vealed small differences between songs of dif- ferent individuals (Fig. 1). These differences were reflected in the time and frequency var- iables measured (Fig. 2), and were consistent from song to song within a single recording and across recordings made over the course of a season (M. R. Lein, unpubl. data). However, these differences are not detectable by ear in the field or on recordings. Mean similarity values between pairs of songs, calculated using SPCC (Table 2), were significantly higher for comparisons of songs within individuals than for comparisons be- tween individual males for Type 1 songs (ran- domization test, P < O.OOl), Type 2 songs (randomization test, P < 0.001), and average TABLE 2. Similarity values calculated by spectrographic cross correlation for comparison of songs within and between individual Buff-breasted Flycatchers. Mean similarity value * SD (Range) Within-male comparisons (n 23) Between-male comparisons (;i = 2.33) Song Type 1 Song Type 2 Average 0.88 ± 0.03 (0.84-0.93) 0.84 ± 0.04 (0.79-0.93) 0.86 ± 0.02 (0.82-0.91) 0.65 ± 0.10 (0.38-0.86) 0.57 ± 0.10 (0.33-0.79) 0.61 ± 0.08 (0.36-0.81) 262 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 0.6 0.4 0.2 0.0 0.6 >» o S 0.4 3 o- o li: 0.2 0.0 0.6 0.4 0.2 0.0 0.3 0.4 0.5 0.6 0.7 0.8 0.9 Similarity Value FIG. 3. Similarity values from spectrographic cross correlation of different songs from individual male Buff- breasted Flycatchers (within-male) or of songs from different pairs of males (between-male). A. Song Type 1. B. Song Type 2. C. Average for song types. C. Mean for Song Types I 1 Within Male Between Males a similarity values (randomization test, P < 0.001). Both within-male and between-male values had considerable variation (Table 2, Fig. 3) with minor overlap in the ranges for both Type 1 and Type 2 songs. However, when similarity values for the two song types were averaged for each comparison (i.e., for each individual male or for each pair of males), there was no overlap in the ranges (Ta- ble 2). The maximum average similarity value for between-male comparisons was 0.81 while the minimum value for within-male compari- sons was 0.82. There were no significant differences in similarity values for comparisons between males within a single mountain range and comparisons between males from different mountain ranges (Table 3) for Type 1 songs (Mantel test, g = 0.7\4, P = 0.20), Type 2 songs (g = -0.271, P = 0.45) or average sim- ilarity values (g = 0.295, P = 0.34). The rang- es of similarity values for the two types of comparisons overlapped completely (Fig. 4). Univariate analyses indicated that all vari- ables measured or calculated for individual songs of both types varied significantly more among individuals than within individuals (Tables 4, 5; one-way ANOVAs, all P < TABLE 3. Similarity values calculated by spectrographic cross correlation for comparison of songs between individual Buff-breasted Flycatchers within the same mountain range (either Chiricahua Mountains or Huachuca Mountains) and between mountain ranges. Song Type 1 Song Type 2 Average Mean similarity value SD (Range) Within-range comparisons (n = 123) Between-range comparisons (n — 130) 0.65 ± 0.10 (0.38-0.86) 0.64 ± 0.10 (0.39-0.84) 0.57 ± 0.10 (0.36-0.78) 0.57 ± 0.10 (0.33-0.79) 0.61 ± 0.08 (0.39-0.80) 0.61 ± 0.08 (0.36-0.81) Lein • SONG VARIATION IN BUFF-BREASTED FLYCATCHERS 263 0.6 0.4 0.2 0.0 0.6 >> o o 0.4 3 O' 2? U. 0.2 0.0 0.6 0.4 0.2 0.0 0.3 0.4 0.5 0.6 0.7 0.8 0.9 Similarity Value FIG 4. Similarity values from spectrographic cross correlations of songs of pairs of male Buff-breasted Flycatchers from the same mountain range (within-range) or from different mountain ranges (between-range). A. Song Type 1. B. Song Type 2. C. Average for song types. A. Song Type 1 Within Range Between Ranges B. Song Type 2 Within Range Between Ranges C. Mean for Song Types I I Within Range Between Ranges r^rkrlrlrli-wn- TABLE 4. Descriptive statistics and coefficients of variation for eight Type 1 songs for each of 23 male Buff-breasted Flycatchers. 17 variables measured or calculated for Variable Mean ± SD Mean CV^ (Range) CVa F22. 161^ DURN 150.6 ±11.4 3.0 (1. 4-5.5) 7.6 40.0 PKIT 16.7 ± 2.9 1 1.9 (6.1-20.1) 17.2 9.4 PK2T 81.2 ± 6.7 2.7 (0.9-5.3) 8.3 57.1 PK3T 107.4 ± 7.0 2.8 (1. 1-4.7) 6.5 34.9 VALT 94.7 ± 6.7 2.5 (1. 2-4.0) 7.1 54.0 STRF 2825.4 ± 374.4 9.3 (3.3-16.6) 13.3 7.1 ENDF 3383.9 ± 267.8 3.1 (1. 0-6.9) 7.9 40.4 PKIF 4469.7 ± 172.0 1.0 (0.4-2.3) 3.8 96.6 PK2F 5292.7 ± 288.1 1.3 (0.2-2. 8) 5.4 1 12.9 PK3F 4089.8 ± 166.8 1 .2 (0.3-2.5) 4.1 74.3 VALF 2548.5 ± 264.9 2.0 (0.8-3.9) 10.4 187.6 P2VFR 2744.3 ± 371.0 2.7 (0.9-4. 1) 13.5 179.9 P1P2FR 823.0 ± 265.8 10.6 (3.2-20.2) 32.3 70.0 PIPN 1 1. 1 ±2.0 1 1.7 (5.0-21.4) 18.0 12.1 P2PN 54.0 ± 3.8 2.6 (0.9-4.3) 7.0 47.0 P3PN 71.5 ± 4.5 3.0 (1. 0-5.6) 6.2 24.7 VPN 63.0 ± 4.3 2.6 (1. 3-4.0) 6.8 43.5 “ f-values for ANOVAs comparing within- and among-male variation for each variable; all P < ().(X)I 264 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 2, June 2008 TABLE 5. Descriptive statistics and coefficients of variation for eight Type 2 songs for each of 23 male Buff-breasted Flycatchers. 14 variables measured or calculated for Variable Mean ± SD Mean CV^ (Range) CVa Fi2. 16i" DURN 193.2 ± 15.1 2.3 (0.9-3.6) 7.8 78.2 PKIT 18.8 ± 2.7 8.2 (2.9-16.3) 14.3 12.0 PK2T 93.5 ± 7.4 3.7 (1. 8-5.4) 8.0 28.5 VALT 37.8 ± 4.9 3.7 (0.9-7.9) 12.8 65.5 STRF 2288.0 ± 236.9 3.8 (1.4-7.9) 10.4 44.1 ENDL 2912.3 ± 292.6 3.6 (1. 4-6.8) 10.0 48.0 PKIF 4463.6 ± 168.3 0.8 (0.4-1. 9) 3.8 151.5 PK2F 6871.2 ± 390.3 1.4 (0.3-3.2) 5.7 93.8 VALE 2691.8 ± 203.3 3.1 (0.9-6.4) 7.6 32.0 P2VFR 4179.4 ± 379.0 3.2 (1.5-6.3) 9.1 49.8 P1P2FR 2407.6 ± 333.5 4.4 (0.9-9.0) 13.9 54.8 PIPN 9.8 ± 21.5 8.6 (4.0-28.3) 15.7 15.4 P2PN 48.6 ± 4.7 4.0 (2.5-6. 1) 9.6 38.6 VPN 19.7 ± 3.0 4.2 (1. 5-6.9) 15.4 86.9 F-values for ANOVAs comparing within- and among-male variation for each variable; all P < 0.001. 0.001). The ratio of CVa/mean CV^ was >1 for all variables. PCA of the variables generated 5 principal components (PCs) with eigenvalues >1.0 ex- plaining 85.3% of the variation in the original variables for Type 1 songs, and 5 PCs with eigenvalues >1.0, explaining 81.9% of the variation in the original variables for Type 2 songs. MANOVAs on the scores of individual songs on the five PCs, conducted as part of DFA, indicated highly significant differences among multivariate means for different indi- viduals for both Type 1 songs (F, 10,773 54.5, P < 0.001) and Type 2 songs (Fi 10,773 = 56.3, P < 0.001). Jack-knifed classifications as- signed 156 of 184 Type 1 (84.8%) and 159 of 184 Type 2 songs (86.4%) to the correct in- dividual. Univariate analyses of variables averaged over the eight songs of each type for each in- dividual indicated three frequency variables measured or calculated for Type 1 songs dif- fered significantly between males from the Chiricahua and Huachuca mountains (Table 6). There was no difference between mountain ranges in other variables for Type 1 songs (all F121 ^ 2.4, all P > 0.137) or in any variable for Type 2 songs (all F121 ^ 2.9, all P ^ 0.101). PCA of the individual means for the vari- ables generated 5 PCs with eigenvalues >1.0 for each song type explaining 87.4 and 85.6% of the variation in the original variables for Type 1 and Type 2 songs, respectively. MAN- OVAs on the scores for individual males on the five PCs indicated no significant differ- ences in multivariate means between the two mountain ranges for either Type 1 (F5 ,7 = 2.1, P = 0.115) or Type 2 songs (F5 17 = 1.1, P = 0.419). Jack-knifed classifications assigned 14 of 23 individuals (60.9%) to the correct moun- tain range in DFAs for each song type. This does not differ (x^ = 0.91, P = 0.34) from the frequency expected if individuals were as- signed randomly to mountain ranges. DISCUSSION Both analyses indicate that both song types of Buff-breasted Flycatchers are individually TABLE 6. Variables measured or calculated for Type 1 songs of Buff-breasted Llycatchers that showed significant differences between birds from the Chiricahua Mountains and the Huachuca Mountains, Arizona. Mean ± SD (Hz) Variable Chiricahua Mtns (n = 13) Huachuca Mtns (n = 10) ^i.2i ^ 3508.0 ± 70.2 3222.6 ± 133.0 10.4 0.004 2649.1 ± 253.9 2417.7 ± 228.3 5.1 0.034 2612.8 ± 411.7 2915.1 ± 229.7 4.3 Q-Q^O ENDL VALE P2VLR Lein • SONG VARIATION IN BUFF-BREASTED ELYCATCHERS 265 distinctive. SPCC showed almost no overlap in similarity values for within-male and be- tween-male comparisons (Fig. 3), while DFA showed a high level of accuracy in assignment of songs to the correct individual. Most mea- sured or calculated variables showed little var- iation within individuals (CV^ values <5%, Tables 4, 5). Many different features of songs had high CV^/mean CV^ ratios (Tables 4, 5) indicating their potential to differentiate among individuals (Robisson et al. 1993, Bee et al. 2001, Vignal et al. 2004). The “errors” in classification of songs in the DFA are also instructive. Nineteen of 28 Type 1 songs assigned to an incorrect individ- ual were associated with “reciprocal errors” in which one song of male A was assigned to male B and one song of male B was assigned to male A. Many of the misclassifications were also consistent with the findings of the SPCC analysis. For example, DFA assigned five Type 1 songs of male SPRG (from Saw- mill Canyon in the Huachuca Mountains) to male SNAG (from West Turkey Creek in the Chiricahua Mountains), and three Type 1 songs of SNAG to SPRG. This pair had an extremely high between-male similarity value (0.84), indicating the close similarity of their Type 1 songs despite their geographical sep- aration. Examination of the results for Type 2 songs showed similar patterns. Both techniques failed to demonstrate sig- nificant differences in songs of males between the two mountain ranges. There was no sep- aration of between-male similarity values cal- culated for comparisons within a single moun- tain range from those for comparisons be- tween ranges (Fig. 4). Three of 31 variables showed significant univariate differences be- tween the two ranges (Table 6). This did not result in a correct classification of songs by DFA that was significantly better than ex- pected by chance. This finding demonstrates the potential danger of basing analyses of song variation on a small number of variables that are often chosen subjectively. Selected variables may not accurately reflect overall patterns of variation. The absence of differentiation in songs be- tween the two mountain ranges is also appar- ent from examination of the SPCC results for pairs of males with between-male similarity values that approach those typical of within- male comparisons. Six of the nine between- male similarity values >0.8 for Type 1 songs, and four of six between-male similarity values >0.75 for Type 2 songs, were for comparisons between males from different ranges. There appears to be no relation between song simi- larity of pairs of males and geographic prox- imity. It is reassuring that both analytical tech- niques produced qualitatively similar results. However, they approach the question of vari- ation differently and, therefore, each is best suited to address different kinds of questions. SPCC is rapid, but provides only a single val- ue for similarity between each pair of songs. It provides no information about the ways in which the two songs are similar or different from each other. Multivariate analysis of a large set of temporal and frequency measure- ments is time consuming, and has an element of subjectivity in choosing appropriate fea- tures for measurement. However, it does per- mit identification of the acoustic features in which different samples of songs differ. It has been suggested that DFA may be used to assign individual identities to “un- known” recordings (Terry et al. 2001). Dis- criminant functions generated from songs of known reference individuals are used to clas- sify the unknown songs. These songs will be assigned to the correct individuals if there is sufficient information in the song features used in the DFA. However, this technique has a serious potential drawback. It is poorly-suit- ed to handle “true unknowns”, that is, songs of “new” individuals that are not included in the sample of reference individuals used to generate the discriminant functions. DFA forces assignment of all songs to one of the pre-existing categories (individuals). SPCC does not have this drawback. “Un- known” songs from an individual in the ref- erence sample will be identified correctly. They will have high similarity values with known songs from that individual, and rela- tively low similarity values with other indi- viduals in the reference sample. Songs of a “true unknown” will have low similarity val- ues with all males in the reference sample and may be identilied as coming from a “new" individual. SPCC does have a serious potential prob- lem that must be recogni/ed. It works best for 266 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 relatively simple sounds without silent inter- vals separating different elements. If two songs had identical elements, but differed in the timing of the intervals separating them, the elements would show progressively less tem- poral overlap during the duration of the songs. This would result in a low cross correlation value even though the sound elements were identical (Khanna et al. 1997). This problem is probable with songs more complex than those of Empidonax flycatchers. However, it is possible to divide complex songs into a se- ries of elements, to conduct SPCCs on the dif- ferent elements (Nelson et al. 1995), and to combine or average the cross correlation val- ues for the different elements of two songs into a single measure of similarity. This would parallel the procedure used to generate an av- erage similarity value across the two song types of Buff-breasted Flycatchers. ACKNOWLEDGMENTS Valerie Haines provided able assistance in the field and made many of the original recordings. Stephen Russell, Josiah and Valere Austin, and Sheridan Stone provided valuable advice or logistical support for the field work. Tara Stehelin acquired the digital sound files from the recordings. Critical comments from Val- erie Haines, Sarah Hechtenthal, and Scott Lovell great- ly improved the final manuscript. Field work was con- ducted under permits from the USDA-Coronado Na- tional Forest and the U.S. Army-Fort Huachuca Mil- itary Reservation. The research was evaluated and approved by the Life and Environmental Sciences An- imal Care Committee of the University of Calgary (Protocol 96-006) and was supported by a grant from the Natural Sciences and Engineering Research Coun- cil of Canada. LITERATURE CITED American Ornithologists’ Union (AOU). 1998. Check-list of North American birds. Seventh Edi- tion. American Ornithologists’ Union, Washing- ton, D.C., USA. Baker, M. C. and D. M. Logue. 2003. Population differentiation in a complex bird sound; a com- parison of three bioacoustical analysis procedures. Ethology 109:223-242. Baptista, L. F. and J. R. King. 1980. Geographical variation in song and song dialects of montane White-crowned Sparrows. Condor 82:267-284. Bee, M. a., C. E. Kozich, K. J. Blackwell, and H. C. Gerhardt. 2001. Individual variation in ad- vertisement calls of territorial male green frogs, Rana clamitans: implications for individual dis- crimination. Ethology 107:65-84. Blank, S., C. Seiter, and P. Bruce. 2001. Resampling Stats add-in for Excel user’s guide. Resampling Stats Inc., Arlington, Virginia, USA. Borror, D. j. 1960. The analysis of animal sounds. Pages 26-37 in Animal sounds and communica- tion (W. E. Lanyon and W. N. Tavolga, Editors). American Institute of Biological Sciences, Wash- ington, D.C., USA. Bowers Jr., R. K. and J. B. Dunning Jr. 1994. Buff- breasted Flycatcher {Empidonax fulvifrons). The birds of North America. Number 125. Clark, C. W., P. Marler, and K. Beeman. 1987. Quantitative analysis of animal vocal phonology: an application to Swamp Sparrow song. Ethology 76:101-115. Johnson, N. K. 1963. Biosystematics of sibling species of flycatchers in the Empidonax hammondi-ober- holseri-wrightii complex. University of California Publications in Zoology 66:79-238. Johnson, N. K. 1980. Character variation and evolu- tion of sibling species in the Empidonax difftcilis- flavescens complex (Aves: Tyrannidae). Univer- sity of California Publications in Zoology 112:1- 153. Khanna, H., S. L. L. Gaunt, and D. A. McCallum. 1997. Digital spectrographic cross-correlation: tests of sensitivity. Bioacoustics 7:209-234. Kroodsma, D. E. 1984. Songs of the Alder Flycatcher {Empidonax alnorum) and Willow Flycatcher {Empidonax traillii) are innate. Auk 101:13-24. Kroodsma, D. E. 1996. Ecology of passerine song de- velopment. Pages 3—19 in Ecology and evolution of acoustic communication in birds (D. E. Kroods- ma and E. H. Miller, Editors). Cornell University Press, Ithaca, New York, USA. Lanyon, W. E. 1978. Revision of the Myiarchus fly- catchers of South America. Bulletin of the Amer- ican Museum of Natural History 161:427—628. Lovell, S. E and M. R. Lein. 2004. Song variation in a population of Alder Flycatchers. Journal of Field Ornithology. 75:146-151. Manly, B. J. E 1994. Multivariate statistical methods: a primer. Second Edition. Chapman and Hall, London, United Kingdom. Manly, B. J. F. 1997. Randomization, bootstrap and Monte Carlo methods in biology. Second Edition. Chapman and Hall, London, United Kingdom. Martin, J. A. 1997. Distribution, abundance, and hab- itat characteristics of the Buff-breasted Flycatcher in Arizona. Thesis. University of Arizona, Tucson, USA. Nelson, D. A., P. Marler, and A. Palleroni. 1995. A comparative approach to vocal learning: intra- specific variation in the learning process. Animal Behaviour 50:83-97. Payne, R. B. and P. Budde. 1979. Song differences and map distances in a population of Acadian Fly- catchers. Wilson Bulletin 91:29-41. Robisson, P, T. Aubin, and J.-C. Bremond. 1993. In- dividuality in the voice of the Emperor Penguin Aptenodytes forsteri: adaptation to a noisy envi- ronment. Ethology 94:279-290. Lein • SONG VARIATION IN BUFF-BREASTED ELYCATCHERS 267 Sedgwick, J. A. 2001. Geographic variation in the song of Willow Flycatchers: differentiation be- tween Empidonax traillii adastus and E. t. exti- mus. Auk 118:366—379. Smith, W. J. 1969. Displays of Sayornis phoebe (Aves, Tyrannidae). Behaviour 33:283—322. Smith, W. J. 1970. Song-like displays in Sayornis spe- cies. Behaviour 37:64-84. Smith, W. J. 1988. Patterned daytime singing of the Eastern Wood-Pewee, Contopus virens. Animal Behaviour 36: 1111-1123. Sokol, R. R. and F. J. Rohlf. 1995. Biometry. Third Edition. W. H. Ereeman and Company, New York, USA. Stehelin, T. E. 2005. Variation in syllable structure and singing behaviour in a population of Dusky Flycatchers {Empidonax oberholseri). Thesis. University of Calgary, Calgary, Alberta, Canada. Stein, R. C. 1963. Isolating mechanisms between pop- ulations of Traill’s Flycatcher. Proceedings of the American Philosophical Society 107:21-50. Terry, A. M. R., P. K. McGregor, and T. M. Peake. 2001. A comparison of some techniques used to assess vocal individuality. Bioacoustics 11:169- 188. VlGNAL, C., N. MATHEVON, AND S. MOTTIN. 2004. Au- dience drives male songbird response to partner’s song. Nature 430:448-451. Wiley, R. H. 2005. Individuality in songs of Acadian Flycatchers and recognition of neighbours. Ani- mal Behaviour 70:237-247. The Wilson Journal of Ornithology 1 20(2);268-276, 2008 PHYLOGENETIC RELATIONSHIP AND SONG DIEFERENCES BETWEEN CLOSELY RELATED BUSH WARBLERS {CETTIA SEEBOHMI AND C. DIPHONE) SHOJI HAMAO,' 5 maria J. S. VELUZ,^ TAKEMA SAITOH,^ AND ISAO NISHIUMP ABSTRACT. — We investigated the phylogenetic relationship and differences in the song structure between the Philippine Bush Warbler {Cettia seebohmi) and the Japanese Bush Warbler (C. diphone). We compared complete sequences of the mitochondrial cyt-b gene of C. seebohmi to those of other Cettia taxa from GenBank and found C. seebohmi formed a monophyletic group with C. haddeni and C. diphone. The phylogenetic tree also suggests that C. seebohmi is more closely related to C. haddeni than to C. diphone although this was not strongly supported due to the low bootstrap values. The estimated nucleotide differences between C. seebohmi and C. haddeni (4.37%), and between C. seebohmi and C. diphone (3.87-4.37%) were larger than the inter- subspecific difference between C. diphone borealis and C. d. cantans (2.44%). Cettia seebohmi, C. haddeni, and C. diphone diverged prior to the subspecies divergences of C. diphone. The basic structure of songs was similar in C. seebohmi and C. diphone', all songs consisted of pure monotone whistles followed by variably modulated warbles. However, sonagraphic parameters showed statistically significant differences between spe- cies. It is reasonable to regard C. seebohmi and C. diphone as separate species. Received 26 February 2007. Accepted 8 September 2007. Bush Warblers {Cettia spp.) radiated in southeast Asia and the southwest Pacific, al- though one species, C. cetti (Cetti’s Warbler), is distributed in Europe. Molecular phyloge- netics suggest the southwestern Pacific species (i.e., C. ruficapilla on the Fiji Islands, C. an- nae on the western Caroline Islands, C. parens on the southern Solomon Islands, and C. had- deni on the northern Solomon Islands) are monophyletic, and the continental C. diphone (Japanese Bush Warbler) belongs to a differ- ent clade from these island forms (Lecroy and Barker 2006). That study, however, did not include C. seebohmi (Philippine Bush War- bler), which is distributed between the island and continental forms. Cettia seebohmi is similar to C. diphone in morphology and plumage, although slight dif- ferences occur in plumage color (Delacour ' Institute for Nature Study, National Museum of Nature and Science, 5-21-5 Shiroganedai, Minato-ku, Tokyo, 108-0071 Japan. 2 Zoology Division, National Museum of the Phil- ippines, P. Burgos Street, Manila, 1000, Philippines. 3 Department of Life Sciences, Rikkyo University, 3-34-1 Nishi-Ikebukuro, Toshima-ku, Tokyo, 171-8501 Japan. ^ Department of Zoology, National Museum of Na- ture and Science, 3-23-1 Hyakunin-cho, Shinjuku-ku, Tokyo, 169-0073 Japan. Corresponding author; e-mail: hamao @ kahaku . go . j p 1942, Kennedy et al. 2000), wing shape, and male body size (Hamao et al. 2006a). It is endemic to northwestern Luzon, Philippines (Grant 1894, Kennedy et al. 2000), while C. diphone breeds in eastern China, southern Us- suriland, Korea, and Japan (Ornithological So- ciety of Japan 2000). These two Bush War- blers are considered different species (Grant 1894, Dickinson et al. 1991, Kennedy et al. 2000), although some authors consider both as C. diphone (Delacour 1942, Baker 1997). The Ornithological Society of Japan (2000) also describes the breeding range of C. diphone as including the Philippines, indicating the two Bush Warblers are regarded as the same spe- cies. Thus, it is important to investigate the genetic relationship between C. seebohmi and C. diphone to understand the taxonomy of these closely related forms and the evolution of Cettia in southeast Asia and the southwest Pacific. Differences in song structure between re- lated species contribute to reproductive isola- tion between them, because vocalization is a recognition cue between species (de Kort et al. 2002, Matyjasiak 2005). Birds can avoid heterospecific matings by discriminating be- tween songs of conspecifics and heterospecif- ics. Therefore, differences in song structure between C. seebohmi and C. diphone are use- ful for understanding their evolutionary rela- 268 Hamao et al. • RELATIONSHIP BETWEEN CETTIA SPECIES 269 tionship. Our objectives are to show: (1) the phylogenetic relationship and divergence time between C. seebohmi and C. diphone using mitochondrial DNA sequences, and (2) differ- ences in the acoustic structure of their songs. METHODS Study Sites and Fieldwork. — The study site for C. seebohmi was an open forest with dense bush at Ambangeg (16°31'N, 120° 50' E; 1,355 m elevation) in the Cordillera Moun- tains of northwestern Luzon, Philippines. We recorded songs of 20 males on 26-27 April 2005 and collected 20-30-p.L blood samples by brachial venipuncture from two males and a female for DNA analysis. The study sites for C. diphone were decid- uous secondary forests that included patches of previously cultivated but presently aban- doned lands dominated by dwarf bamboo {Pleioblastus chino) in Saitama, central Hon- shu, Japan: one at Furusato (36° 06' N, 139° 18' E; 70 m elevation) and the other at Sho- gunsawa (36° 01' N, 139° 20' E; 50 m eleva- tion). We recorded songs of 38 males; 16 males at Furusato on 16 June 2003, and 5 and 17 males at Shogunsawa on 26 May and 2 June 2003, respectively. We recorded each singing male for at least 3 min using a Sony TCD-D8 DAT recorder with a Sony ECM-G3M directional micro- phone. Each recording was sufficient to record all song types of the male because the reper- toires were small and the birds sang frequently (reported for C. diphone by Hamao 1992, 1993; Momose 1999). Sequence Analysis. — DNA was extracted from blood samples using standard phenol/ chloroform extraction. Polymerase chain re- action (PCR) amplifications of entire mito- chondrial cytochrome b (cyt-b) gene and par- tial NADH dehydrogenase subunit V (ND5) were performed using primers (H == heavy strand; L = light strand; numbers give the po- sition of the 3 '-end in the Callus gallus mi- tochondrial genome): mt-F (H- 16065), 5'- GGA GTC TTC AGT TTT TGG TTT ACA AGA C-3' (Helbig and Seibold 1999) and L14()80ND5P, 5'-TCA ACY CAY GCM TTC TTC AAA GC-3' (modified from Sorenson et al. 1999). These two primers gave an approx- imately 2.0-kb product. Primers mt-F and L14764ND5HCZ, 5'-TGA TTT AAR CTM MTA GGA CCA GAA GG-3' (Nishiumi and Kim 2004) were used for sequencing. PCR was performed in volumes of 10 p.L containing 10 ng of total genomic DNA, 0.2 mM of each dNTP, 1.5 mM MgCl2, 0.4 p,M of each primer, and 0.2 units of Taq DNA polymerase (Takara Ex-taq). Amplifications were performed using a Takara PCR Thermal Cycler MP (Takara) under the following con- ditions: 30 sec at 94° C, 30 sec at 52° C, and 90 sec at 72° C (35 cycles). Samples were incubated at 94° C for 3 min before the cyclic reactions and, after completion, at 72° C for 5 min. We placed 2 p,L of the reaction product on a 1.5% agarose gel in 0.5 X TBE buffer to check the success of the reaction. The re- mainder of the PCR product was purified us- ing ExoSAP-IT (Amersham Bioscience, Pis- cataway, NJ, USA). PCR products were sequenced with the BigDye Terminator Cycle Sequencing v3.1 Cycle sequencing Kit (Applied Biosystems, Foster, CA, USA) on an ABI PRISM® 3100- Avant (Applied Biosystems, Foster, CA, USA). Sequencing reactions were performed in volumes of 10 p.L containing 50 ng of PCR product, 1.5 p.L of Terminator Ready Reaction Mix, 1.25 p,L of 5 X Sequencing Buffer, and 1.6 pmol of the sequencing primer. The re- actions were conducted under the following conditions using a Takara PCR Thermal Cy- cler MP: 10 sec at 96° C, 5 sec at 50° C, and 4 min at 60° C (25 cycles). Samples were in- cubated at 96° C for 1 min before the cyclic reactions. Products were purified using etha- nol/EDTA/sodium acetate precipitation ac- cording to the sequencing kit manual. Genetic and Phylogenetic Analyses. — We compared complete sequences of the mito- chondrial cyt-b gene of three individuals of C. seebohmi to those of four other Cettia taxa from GenBank (C. diphone borealis from Ko- rea: ABI 59 198; C. d. cantans from Japan and Korea: AB 1 59 1 94-AB 1 59 1 97; C. haddeni from the northern Solomon Islands: DQ06645 1 ; C. fortipes from Taiwan: L77122), which are all of the Cettia species that had complete cyt-b sequences available from GenBank. We used three other species from GenBank (Urosphena squameiceps (Asian Stubtail]: AB159177, AB159179; Syl- via atricapilla (Eurasian Blackcap(: Z73494; Acrocephalus orientalis (Oriental Reed War- 270 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 bier]: AB 159 186) as an outgroup for the phy- logenetic analysis. Sequences were aligned by eye with ATGC V4.0.8 and Genetix-Mac vlO.l (Genetix). An appropriate substitution model for the data set was estimated via the Akaike information cri- terion (AIC; Akaike 1974) and a hierarchical likelihood ratio test (Posada and Crandall 1998), both calculated in MODELTEST 3.7 (Posada and Crandall 1998). The choice of cyt-b model was based on the best-fit model according to the AIC. The sequences of Cettia and Urosphena were included in the product data set. The selected model for cyt-b was a transversional model (Posada and Crandall 1998) with a gamma distribution (TVM + G; settings: empirical base frequencies of ^ 0.2814, TTc = 0.3570, ttg = 0.1233, = 0.2382, transition/trans version ratio = 7.7216, gamma distribution shape parameter a = 0.1037). Phylogenetic trees were constructed using different approaches: neighbor joining (NJ; Saitou and Nei 1987), maximum parsi- mony and maximum likelihood (MP and ML; PAUP v4.0bl from Swofford 2003), and Bayesian inference of phylogeny (MrBAYES 3.1.2; Huelsenbeck and Ronquist 2001). The robustness of clades was estimated by 1,000 bootstrap replicates in NJ and MP (Eelsenstein 1985), 100 bootstrap replicates in ML, and by Bayesian posterior probabilities in Bayesian inference using Markov chain Monte Carlo (MCMC). Eour Metropolis-coupled MCMC chains with incremental heating temperature 0.1 were run for 1,000,000 generations and sampled every 100 generations. Two analyses were run simultaneously starting from random trees. The first 25% of generations were dis- carded as “bum-in” and the posterior proba- bility was estimated for the remaining gener- ations. Samples from the stationary phases of the independent mns were pooled to obtain the final results. The nucleotide divergence between popu- lations is shown as the Kimura two-parameter distance (%) calculated by MEGA v.3.1 (Ku- mar et al. 2004). We used 1.6% per million years (MY) and 2.0% per MY of the cyt-b substitution rates for small birds to estimate divergence time (Fleischer et al. 1998, and Moore and DeFilippis 1997, respectively). Song Analysis. — All recorded sounds were analyzed with computer software (SAS-Lab Pro. Version 4.2, Raimund Specht). Sounds were displayed as sonagrams. FFT-lengths of 256 and 1024 were used to produce plots of temporal and frequency measures, respective- ly. Songs of both the Philippine and Japanese Bush warblers consisted of an initial constant frequency (CF) part and a subsequent fre- quency modulated (EM) part (Momose 1999). We measured nine sonagraphic parameters of each song type, which represented song stmc- ture (Fig. 1): CF part, (1) frequency, (2) num- ber of notes; FM part, (3) maximum frequen- cy, (4) minimum frequency, (5) frequency range, (6) duration, (7) number of notes, (8) number of frequency inflections, i.e., number of changes in the sign of the derivative (slope) of the frequency in the sonagram, and (9) modulation rate, i.e., number of frequency in- flections/sec. We also measured the number of song types of each male, as males used a small number of stable song types (reported for C. diphone by Momose 1999, Hamao and Ueda 2000). We performed Mann-Whitney f/-tests with a sequential Bonferroni correction (Rice 1989) to assess the table-wide type I error rate when comparing song parameters between the two species. We used the average values for indi- vidual males and considered them indepen- dent data points. RESULTS Phylogenetic Relationships and Genetic Distances. — Complete mitochondrial cyt-b se- quences from three individuals of C. seebohmi were obtained and deposited in GenBank (ac- cessions AB28 1094-6). They were compared to the sequences available in GenBank of C. haddeni, C. fortipes, and two subspecies of C. diphone. Phylogenetic analysis showed C. seebohmi formed a monophyletic group with C. haddeni and C. diphone (the bootstrap values were 69, 100, 69, and 100 by NJ, MP, ML, and Bayes, respectively) (Fig. 2). All four different mod- els inferred that C. seebohmi is more closely related to C. haddeni than to C. diphone, al- though the monophyly of C. seebohmi and C. haddeni was not strongly supported (the boot- strap values were 76, 63, 56, and 56 by NJ, MP, ML, and Bayes, respectively). The estimated nucleotide difference be- tween C. seebohmi and C. diphone borealis Hamao et al. • RELATIONSHIP BETWEEN CETTIA SPECIES 271 FM max FM min FIG. 1. Parameters used for song structure measurements of Cettia seebohmi and C. diphone. Constant frequency (CF frequency), frequency modulated maximum (FM max), minimum frequency (FM min), duration, number of frequency inflections (FM inflections), and number of notes (FM notes). (3.87%) was smallest among those between the species (Table 1). However, the difference between C seebohmi and C diphone cantons (4.37%) was equivalent to that between C. seebohmi and C. haddeni. These differences were larger than the inter-subspecific differ- ence in C. diphone (2.44%). The divergence time of C. seebohmi from C. diphone borealis was estimated to be 1 .94- 2.42 million years ago (MYA), which is older than between subspecies C. d. borealis and C. d. cantons (1.22-1.53 MYA), but equivalent to that between C. seebohmi and C. haddeni or C. diphone cantons (2.19-2.73 MYA) (Ta- ble 2). Song Structure. — A male C. seebohmi (male #2) had four song types (Fig. 3). The songs consisted of pure monotone whistles (CF parts) followed by variably modulated warbles (FM parts), except for one song type lacking the CF part (Fig. 3). A male C. di- phone (male #34) also had four song types, and the CF and FM parts in the songs (Fig. 3). The basic structure of songs was similar in other males of both species. All sonagraphic parameters had statistically significant differences between species (Table 3). The number of notes in CF parts was larger in C. diphone than in C. seebohmi. The fre- quency was higher in the songs of C. seeboh- mi than in C. diphone in both the CF and FM parts. The frequency range of FM parts was wider in C. seebohmi than in C. diphone. The FM parts of C. seebohmi had a longer dura- tion, and more notes and inflections, although the modulation rate was lower. Individual males of C. seebohmi had fewer song types than those of C. diphone. Thus, songs of C. seebohmi had higher frequency and a wider frequency range, and were more complexly modulated than those of C. diphone. However, males of C. seebohmi had less variable songs. DISCUSSION Genetic distance seemingly suggests that C. seebohmi is most closely related to C. diphone borealis (Table 1). However, the distance be- tween C. seebohmi and the other subspecies of C. diphone (i.e., C. d. cantons) was equiv- alent to that between C. seebohmi and C. had- deni. Therefore, from the results of genetic distances, it is equivocal which species, C. di- phone or C. haddeni, is more closely related to C seebohmi. The molecular phylogenetic tree suggests that C. seebohmi is more closely related to C haddeni than to C. diphone, but this was not strongly supported due to the low bootstrap values (Fig. 2). Thus, further ex- amination is needed to identify the phyloge- netic position of C. seebohmi. 272 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 r AB 1 59 1 98 Cettia diphone borealis NJ/MP ML/Bayes -/86 73/98 82/100, 92/100 69/100 69/100 76/63, 56/56 65/- 84/98 84/100 78/74 AB159197 AB159195 Cettia diphone cantans — DQ066451 Cettia haddeni 78/100 75/100 AB281094 AB281095 AB281096 Cettia seebohmi L77122 Cettia fortipes 99/100 83/100 AB159177 AB159179 Urosphena squameiceps Z73494 Sylvia atricapilla A B 1 5 9 1 86 Acrocephalus oriental is 0.05 substitutions/site FIG. 2. Neighbor-joining tree of cyt-b sequences from Cettia and Urosphena. The tree was rooted using Sylvia atricapilla and Acrocephalus orientalis as outgroups. Numbers at each terminal are GenBank accessions; numbers at each node indicate bootstrap support (above, neighbor joining/maximum parsimony; below, maxi- mum likelihood/Bayes). TABLE 1. Kimura two-parameter distances (%) between Cettia spp. in southeast Asia and the southwest Pacific. Taxonomic unit 1 1 Cettia diphone borealis 2 Cettia diphone cantans 3 Cettia haddeni 4 Cettia seebohmi 5 Cettia fortipes 6 Urosphena squameiceps 2.44 4.74 5.24 3.87 4.37 4.37 10.60 10.11 13.04 9.91 10.40 11.57 4 5 11.17 9.90 11.77 6 Hamao et al. • RELATIONSHIP BETWEEN CETTIA SPECIES 273 TABLE 2. Estimated divergence time (MYA) between Cettia spp. in southeast Asia and the southwest Pacific under the assumption of 1.6 and 2.0%/MY substitution rates. Taxonomic unit 1 2 3 4 1 Cettia diphone borealis 2 Cettia diphone cantans 3 Cettia haddeni 4 Cettia seebohmi 1.22-1.53 2.37-2.96 1.94-2.42 2.62-3.28 2.19-2.73 2.19-2.73 The estimation of divergence time suggests that C. seebohmi and C. haddeni diverged 2.19-2.73 MYA, and that C. seebohmi and C. diphone diverged 1.94-2.73 MYA. This sug- gests that Cettia species in southeast Asia and the southwest Pacific radiated at nearly the same time, at the end of the Pliocene. Mon- arch flycatchers (Family Monarchidae) diver- sified across the Pacific archipelagos in a sin- gle radiation (Filardi and Moyle 2005). Evo- lution of Cettia species shows a similar pat- tern, indicating that in some avian taxa, this is a common pattern when birds radiate on Pacific islands. The estimated nucleotide difference be- tween the Korean and Japanese subspecies of C. diphone {borealis and cantans, respective- ly) was 2.44%, which is similar to that esti- mated by Kajita (2002): 2.3% difference in the cyt-b gene between Japanese subspecies (i.e., C. d. cantans [throughout Japan], C. d. diphone [Bonin Islands], and C. d. restricta [Minami-Daito Island, extinct but recovered on Okinawa Island by Kajita et al. 2002]), and the continental subspecies (i.e., C. d. borealis [Korea] and C. d. canturians [China]). The di- vergence time of C. d. borealis with C. d. can- tans was estimated to be 1.22-1.53 MYA. These results suggest the continental ancestor of C. diphone dispersed to the main and pe- ripheral islands in Japan, and formed subspe- cies in the Pleistocene. The songs of both C. seebohmi and C. di- phone consist of CF and FM parts (Fig. 3). This basic song structure was also found in other Cettia species (i.e., C. fortipes in Taiwan [Wells 1982, Orenstein and Pratt 1983, Roz- endaal 1987], C. carolinae in the Tanimbar Islands [Rozendaal 1987], C. ruficapilla in the Fiji Islands [Orenstein and Pratt 1983], and C. annae in the Caroline Islands [Orenstein and Pratt 1983]). The similarity in song structure may be caused by the similar habitats of the species. The initial CF part in songs of C. di- sec FIG. 3. Basic song structure of Cettia seehohmi (left) and C. (liphone (right). Four song types of one male are shown for each species. 274 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 TABLE 3. Song features (means number of males. ± SD) of Cettia seebohmi and C. diphone. Numbers in parentheses are C. seebohmi (20) C. diphone (38) P Song structure CF frequency (kHz) 1.67 ± 0.10 1.25 ± 0.15 <0.001* CF notes 1.00 ± 0.00 2.70 ± 0.52 FM max (kHz) 5.79 ± 0.23 3.35 ± 0.36 <0.001* EM min (kHz) 1.62 ± 0.12 1.25 ± 0.11 <0.001* FM width (kHz) 4.12 ± 0.27 2.10 ± 0.39 <0.001* FM duration (sec) 0.81 ± 0.10 0.30 ± 0.06 <0.001* FM notes 5.2 ± 0.5 2.9 ± 0.5 <0.001* FM inflections 8.6 ± 1.3 3.9 ± 1.2 <0.001* FM modulation rate 10.56 ±1.51 12.82 ± 3.77 <0.01* Repertoire size No. of song types 2.3 ± 0.8 3.0 ± 0.9 <0.01* * t/-test with Bonferroni correction. ** U-iesl was not performed because of the large difference in the variances between the species. phone may be an adaptation to a bushy en- vironment (Momose 1986). The habitats of both C. seebohmi and C. diphone are roughly the same: thickets, shrubs, and forest under- growth (Kennedy et al. 2000, Ornithological Society of Japan 2000) in which pure mono- tonal whistles may have high carrying ability (Morton 1975). The sonagraphic parameters of songs dif- fered between C. seebohmi and C. diphone (Table 3). The frequency was higher in C. see- bohmi that had smaller body size than C. di- phone (Hamao et al. 2006a). Smaller birds produce songs with higher frequencies (Ryan and Brenowitz 1985), which may affect the difference in frequency. The other possible factor related to these differences is vegetation density. The habitat of C. diphone is dense bamboo thickets and shrubs at the forest edge and forest floor (Nakamura and Nakamura 1995, Ornithological Society of Japan 2000), while the habitat of C. seebohmi is thickets and the understory of open forests (Kennedy et al. 2000). The songs of C. seebohmi had a higher frequency, a wider frequency range, and a more complex modulation pattern. These vocalization features are connected to open habitats (e.g.. Hunter and Krebs 1979, Shy 1983, Anderson and Conner 1985) due to sound transmission properties of the habitats (Morton 1975). Cettia diphone males did not show statis- tically significant differences in their respons- es to C. seebohmi and C. diphone songs in a song playback experiment (Hamao et al. 2006b), but this does not directly indicate they are the same species. Males may lose their territory and paternity if they regard conspe- cific rivals as heterospecifics. In contrast, if females regard heterospecific males as con- specifics, they may suffer a greater cost by producing hybrid offspring. There is a differ- ence in the cost of misidentifying potential competitors and mates. Therefore, males and females should differ in species recognition (Searcy and Brenowitz 1988). Variable songs are more likely to be acceptable for males. Cettia seebohmi diverged from C. haddeni and C. diphone prior to the subspecies diver- gences of C. diphone. Its songs have the same basic structure as those of C. diphone, but ap- parently differ in sonagraphic parameters. It is reasonable to regard C. seebohmi and C. di- phone as separate species. ACKNOWLEDGMENTS We thank the Department of Environment and Nat- ural Resources (DENR), Parks and Wildlife Bureau (PAWB), and the DENR-CAR (Cordillera Administra- tive Region) for permits and logistical support. We es- pecially thank Teber Dionisio, Yolly Ruperto, and Emerita Tamiray of Mt. Pulag. We also thank Hiroyuki Morioka and Manabu Kajita for assistance with the literature search, and Francis Veluz and Nicky Icar- angal for field assistance. This study was planned as one of the “Natural History Researches of the Island Arcs in the Western Pacific” by the National Museum of Nature and Science, Tokyo, and was partly sup- ported by the Japan Ministry of Education, Culture, Hamao et al. • RELATIONSHIP BETWEEN CETTIA SPECIES 275 Sports, Science, and Technology (grant 17770076 to IN). Finally, we thank three anonymous reviewers and C. E. Braun for improving the manuscript. LITERATURE CITED Akaike, H. 1974. A new look at the statistical model identification. IEEE Transactions on Automatic Control 19:716-723. Anderson, M. J. and R. N. Conner. 1985. Northern Cardinal song in three forest habitats in eastern Texas. Wilson Bulletin 97:436-449. Baker, K. 1997. Warblers of Europe, Asia and North- ern Africa. Helm identification guides. Christo- pher Helm Publishers, London, United Kingdom. DE Kort, S. R., P. M. Den Hartog, and C. Ten Cate. 2002. Diverge or merge? The effect of sympatric occurrence on the territorial vocalizations of the Vinaceous Dove Streptopelia vinacea and the Ring-necked Dove S. capicola. Journal of Avian Biology 33:150-158. Delacour, C. J. 1942. The Bush Warblers of the gen- era Cettia and Bradypterus, with notes on allied genera and species. Ibis 84:509-519. Dickinson, E. C., R. S. Kennedy, and K. C. Parkes. 1991. The birds of the Philippines, an annotated check-list. British Ornithologists’ Union, Oxford, United Kingdom. Felsenstein, j. 1985. Confidence limits on phyloge- nies: an approach using the bootstrap. Evolution 39:783-791. Filardi, C. E. and R. G. Moyle. 2005. Single origin of a pan-Pacific bird group and upstream coloni- zation of Australasia. Nature 438:216-219. Fleischer, R. C., C. E. Mcintosh, and C. L. Tarr. 1998. Evolution on a volcanic conveyor belt: us- ing phylogeographic reconstructions and K-Ar- based ages of the Hawaiian Islands to estimate molecular evolutionary rates. Molecular Ecology 7:533-545. Grant, W. R. O. 1894. On the birds of the Philippine Islands. Part II. The highlands of north Luzon, 5000 feet. Ibis (6th Series) 6:501-522. Hamao, S. 1992. Lack of pair-bond: a polygynous mating system of the Japanese Bush Warbler Cet- tia diphone. Japanese Journal of Ornithology 40: 51-65. Hamao, S. 1993. Individual identification of male Jap- anese Bush Warblers Cettia diphone by song. Jap- anese Journal of Ornithology 41:1-7. Hamao, S. and K. Ueda. 2000. Simplified song in an island population of the Bush Warbler Cettia di- phone. Journal of Ethology 18:53-57. Hamao, S., M. J. S. Veluz, and I. Nishiumi. 2006a. Difference in morphology between two closely re- lated Bush Warblers, Cettia diphone and C. .see- hohmi, in Japan and the Philippines. Miscella- neous Reports of the Institute for Nature Study 37: 1 7-26. Hamao, S., M. J. S. Veluz, and I. Nishiumi. 2006b. Species recognition by song between the Luzon Bush Warbler, Cettia .seehohrni, and the Japane.se Bush Warbler, C. diphone. Memoirs of the Na- tional Science Museum 44:21-29. Helbig, a. j. and I. Seibold. 1999. Molecular phylog- eny of Palearctic-African Acrocephalus and Hip- polais warblers (Aves: Sylviidae). Molecular Phy- logenetics and Evolution 1 1 :246-260. Huelsenbeck, j. P. and E R. Ronquist. 2001. Mr- Bayes: Bayesian inference of phylogenetic trees. Bioinformatics 17:754-755. Hunter, M. L. and J. R. Krebs. 1979. Geographical variation in the song of the Great Tit, Pams ma- jor, in relation to ecological factors. Journal of Animal Ecology 48:759-785. Kajita, M. 2002. Within-species phylogeny and bio- geography in Cettia diphone. Iden 56:42-46. Kajita, M., T. Mano, and E Sato. 2002. Two forms of Bush Warbler Cettia diphone occur on Okina- wajima Island: re-evaluation of C. d. riukiuensis and C. d. restricta by multivariate analyses. Jour- nal of Yamashina Institute for Ornithology 33: 148-167. Kennedy, R. S., P. C. Gonzales, E. C. Dickinson, H. C. Miranda Jr., and T. H. Fisher. 2000. A guide to the birds of the Philippines. Oxford University Press, Oxford, United Kingdom. Kumar, S., K. Tamura, and M. Nei. 2004. MEGA3: Integrated software for molecular evolutionary ge- netics analysis and sequence alignment. Briefings in Bioinformatics 5:150-163. Lecroy, M. and E K. Barker. 2006. A new species of Bush-warbler from Bougainville Island and a monophyletic origin for southwest Pacific Cettia. American Museum Novitates 3511:1-20. Matyjasiak, P. 2005. Birds associate species-specific acoustic and visual cues: recognition of hetero- specific rivals by male Blackcaps. Behavioral Ecology 16:467-471. Momose, H. 1986. Mechanism of maintaining territo- ries by acoustic communication. Pages 127-157 in Reproductive strategies of birds (S. Yamagishi, Editor). Toukaidaigaku-shuppankai, Tokyo, Japan. Momose, H. 1999. Structure of territorial songs in the Japanese Bush Warbler {Cettia diphone). Memoirs of the Faculty of Science, Kyoto University (Se- ries Biology) 16:55-65. Moore, W. S. and V. R. deFilippis. 1997. The window of taxonomic resolution for phylogenies based on mitochondrial cytochrome b. Pages 83-113 in Avian molecular evolution and systematics (D. P. Mindel, Editor). Academic Press, San Diego, Cal- ifornia, USA. Morton, E. S. 1975. Ecological sources of selection on avian sounds. American Naturalist 109:17-34. Nakamura, T. and M. Nakamura. 1995. Birds’ life in Japan with color pictures: birds of mountain, woodland and field. Hoikusya, Osaka. Japan. Nishiumi, I. and C.-H. Kim. 2004. Little genetic dif- ferences between Korean and Japanese popula- tions in songbirds. National Science Museum Monographs 24:279-286. Oren.stein, R. I. AND H. D. Prait. 1983. The relation- 276 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 ships and evolution of the southwest Pacific war- bler genera Vitia and Psamathia (Sylviinae). Wil- son Bulletin 95:184-198. Ornithological Society of Japan. 2000. Check-list of Japanese birds. Sixth and Revised Edition. Or- nithological Society of Japan, Obihiro, Japan. Posada, D. and K. A. Crandall. 1998. Modeltest: testing the model of DNA substitution. Bioinfor- matics 14:817-818. Rice, W. R. 1989. Analyzing tables of statistical tests. Evolution 43:223-225. Rozendaal, F. G. 1987. Description of a new species of Bush Warbler of the genus Cettia Bonaparte, 1834 (Aves: Sylviidae) from Yamdena, Tanimber Islands, Indonesia. Zoologische Mededelingen 61: 177-202. Ryan, M. J. and E. A. Brenowitz. 1985. The role of body size, phylogeny, and ambient noise in the evolution of bird song. American Naturalist 126: 87-100. Saitou, N. and M. Nei. 1987. The neighbor-joining method: a new method for reconstructing phylo- genetic trees. Molecular Biology and Evolution 4: 406-425. Searcy, W. A. and E. A. Brenowitz. 1988. Sexual differences in species recognition of avian song. Nature 332:152-154. Shy, E. 1983. The relation of geographical variation in song to habitat characteristics and body size in North American tanagers (Thraupinae: Piranga). Behavioral Ecology and Sociobiology 12:71-76. Sorenson, M. D., J. C. Ast, D. E. Dimcheff, T. Yuri, AND D. P. Mindell. 1999. Primers for a PCR- based approach to mitochondrial genome se- quencing in birds and other vertebrates. Molecular Phylogenetics and Evolution 12:105-114. SwoFFORD, D. L. 2003. Phylogenetic analysis using parsimony (and other methods). Version 4. Sin- auer Associates, Sunderland, Massachusetts, USA. Wells, D. R. 1982. Biological species limits in the Cettia fortipes complex. Bulletin of the British Ornithologists’ Club 102:57-62. The Wilson Journal of Ornithology 120(2):277-285, 2008 AUTUMN STOPOVER NEAR THE GULF OF HONDURAS BY NEARCTIC-NEOTROPIC MIGRANTS ANDREW B. JOHNSON‘ 2.3 AND KEVIN WINKER' ABSTRACT. — The southeastern Yucatan Peninsula hosts high numbers of transient Nearctic-Neotropic mi- grants during autumn migration, but the importance of this region during migratory stopover has not been addressed. We studied autumn stopover body mass gains among passerine migrants in tropical lowland forest 20 km inland from the Gulf of Honduras. Most individuals captured had some subcutaneous fat. Ten of 15 taxa with sufficient sample sizes had significant positive diel (24 hr) gains in a body condition index. Estimates of net mass gains in these 10 taxa suggested they all were depositing fat; average individuals in four of these taxa were depositing sufficient fuel to undertake an entire night of migration after only 1 day of fattening; Empidonax spp.. Red-eyed Vireo (Vireo olivaceus). Gray Catbird (Dumetella carolinensis), and Northern Waterthrush (Seiu- rus noveboracensis). Two (Wood Thrush [Hylocichla mustelina] and Common Yellowthroat [Geothlypis trichas]) of the four species apparently not gaining mass at the study site migrate late in the season and occurred only after Hurricane Iris severely altered the habitat. Four other species (Gray Catbird, Magnolia Warbler [Dendroica magnolia], American Redstart [Setophaga ruticilla], and Indigo Bunting [Passerina cyanea]) had significant gains in mass after the hurricane. These data demonstrate the importance of the region as an autumn stopover site for some species and suggest that stopover areas farther north are also important to migrants passing through the southeastern part of the Yucatan Peninsula. Received 19 January 2006. Accepted 25 September 2007. The geography of North America causes Nearctic-Neotropic migrants that breed across thousands of square kilometers of boreal and temperate forest to funnel through a small fraction of the land area in the forests of Cen- tral America during the nonbreeding season. This concentration of migrants likely makes forests in Central America vital as both win- tering and stopover habitat. Many migrant species have winter ranges that extend far be- yond the Gulf of Honduras (AOU 1998), and forests adjacent to the Gulf of Honduras are probably important stopover habitat. Pub- lished accounts of high volumes of transient Nearctic-Neotropic migrants in Panama (Gal- indo et al. 1963, Galindo and Mendez 1965) suggest there should be similar numbers of migrants farther north. Monroe (1968) docu- mented a large spring migration northward across the Gulf of Honduras and speculated that a large autumn migration also occurs. However, there has been no information pub- lished about how transient migrants use stop- over sites in this region. Few studies have addressed energetic needs ' University of Alaska Museum, 907 Yukon Drive, Fairbanks, AK 99775, USA. ‘ Current address: Museum of Southwestern Biolo- gy, University of New Mexieo, Albuquerque, NM 87131, USA. ^Corresponding author; e-mail: ajohnson@unm.edu of migrants in Central America (Rogers and Odum 1966; Child 1969; Winker 1995a, b), and there are no published studies of migrant stopover ecology near the Gulf of Honduras. We chose a site near the Gulf of Honduras in lowland tropical forest to examine the autumn migration of Nearctic-Neotropic migrants. We provide the first extensive data on autumn stopover by woodland migrants in this region and address the following questions: (1) what levels of fat are carried in the region; (2) do migrants refuel at this site and, if so, to what extent; and (3) do species that have farther to migrate fuel more? METHODS Study Area. — Our study site was a 25-year- old, second-growth, lowland forest adjacent to a citrus orchard in the floodplain of the Rio Grande (16° 16' N, 88° 52' W) near Big Falls Village, Toledo District, Belize (Fig. 1). The site was in a matrix of human-altered habitats (primary forest remnants, citrus orchard, fields, and habitats in stages of regrowth) that had a canopy height of —20 m with some gaps filled with dense woody vegetation and vine tangles 3 m in height. Trapping. — We established a 1.26-ha study site in August 2001 and placed 30 12-m mist nets in two parallel rows of 15 nets spaced 30 111 apart. Each net was spaced 30 ni apart from its neighbor within each row, and we alter- 277 278 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 FIG. 1. Big Falls, Toledo District study site in Be- lize in northern Central America on the Yucatan Pen- insula. nated mesh sizes between 30 and 36 mm. Nets were opened beginning on 11 August and were open all day when conditions permitted. We accrued 8,805 net hrs until 7 October, when we removed nets from the study site in anticipation of Hurricane Iris, which struck on 8 October. The effect of Hurricane Iris on the site changed the habitat from a nearly closed, 20-m high canopy to a 5-m high tangle of up- rooted and broken trees, broken branches, and vines. We re-opened 15 nets placed on the original site on 19 October. Five nets were placed in their original net lanes and others were placed as close as possible to their orig- inal positions. Placement of nets was con- strained by the drastically altered forest struc- ture (e.g., fallen trees and dense tangles of vines). We only netted during mornings and evenings during this period due to lack of shade on the site. We accrued 1,114 more net hrs and concluded the effort on 1 5 November. Each bird captured was identihed to spe- cies, age class, and whether it was male or female when possible. Time and net of cap- ture, body mass, and chord of closed wing (hence ‘wing chord,’ as traditionally used, see Stiles and Altshuler 2004) were recorded for all birds captured; tail, tarsometatarsus, and bill lengths were recorded for most individu- als. Subcutaneous fat deposits were scored following Helms and Drury (1960). All Em- pidonax were treated as a single taxon for this study, and included five Nearctic-Neotropic migrant species verified by museum speci- mens {traillii, alnorum, flaviventris, virescens, and minimus). Daily Mass Gains. — The extent of mass gain was estimated at the species level follow- ing Winker et al. (1992) and Winker (1995a). This method calculates a condition index for each individual and examines the relationship between condition index and time of capture using simple linear regression. Rising and Somers (1989) and Freeman and Jackson (1990) suggested that no single linear external measurement of a bird is a good cor- relate of body size, although these studies sug- gested that tarsometatarsus (“tarsus”) or tib- iotarsus was the best univariate indicator of body size. However, Connell et al. (1960) and Rogers and Odum (1964, 1966) showed that wing length was a good predictor of fat-free mass in some passerines during migration, al- though they did not examine other variables. Winker (1995 a) found a condition index based on wing chord or tail to be a better predictor of fat content than one based on tarsus in a sample of fat-extracted Tennessee Warblers {Vermivora peregrina). We used wing chord to standardize body mass by calculating a con- dition index for all species except Kentucky Warbler {Oporornis formosus). A condition index for each individual was calculated by dividing its mass by its wing chord (Winker 1995a). The condition index was examined for each species with respect to time of day (equivalent to time of capture of each individual; hereafter referred to simply as ‘time’) using simple linear regression. Re- gressions of fat scores, mass, and condition indices using other morphological characters (tail, tarsometatarsus, and bill lengths) with Johnson and Winker • MIGRATION STOPOVER NEAR THE GULF OF HONDURAS 279 respect to time were used to corroborate wing chord condition index trends. A condition in- dex constructed using the first principal com- ponent of wing chord, tail length, tarsometa- tarsus, and bill lengths combined was also ex- plored as an option, but the estimates were not an improvement over the use of a single mor- phological character. Multiple regression was also tried (Dunn 2000), but the estimates were similar and did not justify the added complex- ity of the analyses. Estimates of net 24-hr mass changes were made for species having regression slopes of condition index with respect to time signifi- cantly different from zero (a = 0.05). The slope of the condition index regression was converted to estimated daily gross mass gains for the average individual in a species by mul- tiplying by 12.42 hrs (average length of daily bird activity from field notes) and multiplying by the sample’s mean wing chord. Net mass gains were calculated by subtracting estimates of nightly metabolic demands from the esti- mated gross daily gains. Our estimate of noc- turnal loss was a mass-specific existence me- tabolism estimate (Kendeigh 1970:60) using both fat-free mass data when available (Dun- ning 1993) and average “lean mass” (Dunn 2002) from birds captured at the site. We used a value of 30.2 kJ energy per g fuel (Pennycuick 2003) for our flight capacity estimates, which assumes the fuel used during migration is 95% fat. Our taxon-averaged flight capacity estimates would have to be re- vised downward from our present energetic estimates if it is found that migrants in this region are depositing protein to the extent as trans-Sahara Eurasian Blackcaps (Sylvia atri- capilla) (Karasov and Pinshow 1998). Fuel composed of 30% protein would reduce our flight capacity estimates by about 50% (Pen- nycuick 2003: equation 12). We assumed mass gain was not the result of rehydration after migration (Nisbet et al. 1963). Catabolism of lipids produces water that helps to maintain water balance, and Rog- ers and Odum (1966) showed that, even in emaciated post-flight birds in Panama, water content was not different from that of fat birds. Bauchinger and Biebach (2000) also showed that water content did not differ among pre-migratory, immediately post-mi- gratory, or post-migratory birds that had 7 days to recover with free access to food and water. Morphological data for each taxon were checked for normality. Mass and morpholog- ical variables used to estimate net mass gains in each species were normally distributed ex- cept for wing chord in American Redstart (Se- tophaga ruticilla). Transformations of wing chord for American Redstart failed to nor- malize these data. Tail length was not record- ed in this species. Residuals were checked for normality using quantile-quantile plots fol- lowing regression analyses and examined vi- sually to ascertain whether there were any pat- terns that indicated unequal variance. Resid- uals were normal and no patterns indicating heteroscedasticity were found in any of these taxa. Hurricane Iris struck on 8 October 200 1 and the drastic habitat change that resulted may have had an effect on fat deposition by mi- grants. No taxon had sufficiently large sample sizes (n ^ 30) pre- and post-hurricane to con- duct separate analyses before and after the storm. Consequently, each migrant taxon was based on a sample from either entirely before or entirely after the hurricane; samples were not pooled between pre- and post-hurricane efforts. Flight Capacity Estimates. — Maximum hours of flight possible for the average indi- vidual gain in each taxon at the site were cal- culated using net gain estimates from this study, published values for the energetic con- tent of fuel (30.2 kJ per g; Pennycuick 2003), and size-specific rate of energy use during mi- gration (Tucker 1974). These values were used to estimate the duration an average in- dividual of each species could fly in a night. The methodology of Pennycuick (1989) was not applicable because we did not record wing span measurements. Mass Comparisons. — Condition index anal- yses provide a species-level estimate of daily gains in mass, but do not consider the amount of fat already carried by individuals of a spe- cies. Average mass from each species was compared to its fat-free mass to examine the fat load the average individual of each species was carrying (Dunning 1993), and to average “lean mass” (mean mass of individuals from our site with fat scores of zero; Dunn 2002) using two-sample r-tests. The paucity of in- 280 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 dividuals of some species that had fat scores of zero at our site made the latter comparison tenuous (e.g., no Red-eyed Vireos [Vireo oli- vaceus] or Veeries [Cathams fuscescens] had fat scores of zero). RESULTS We captured 30 or more individuals of 14 species and the genus Empidonax, and present summary statistics of morphological charac- ters to allow comparisons with other studies (Table 1). We encountered no emaciated in- dividuals. We rarely observed actively grow- ing feathers and assumed that energetic de- mands for the vast majority of these birds were limited to migration and maintenance costs. Comparisons of Mass with Fat-free and Lean Mass. — All taxa but Wood Thrush (Hy- locichla mustelina) were significantly heavier than the species’ average fat-free mass (Table 2), and most individuals were carrying visible subcutaneous fat (82% of the individuals of the 15 taxa had fat scores >0). Many Veeries and Swainson’s Thrushes {Catharus ustulatus) had heavy fat loads (Table 1; all Veeries had fat scores >0). Few taxa were significantly heavier than lean mass, but all taxa except Wood Thrush were significantly heavier than fat-free mass (Table 2). We observed frequent defecation by birds during handling that sug- gested they were feeding at this site. The peak of migration, especially of Catharus thrushes, corresponded with ripening of large amounts of fruits on our site, particularly the understo- ry tree, Dendropanax arboreus (Araliaceae). Estimates of Daily Mass Gain. — Eleven taxa had significant slopes of mass with re- spect to time (Table 3). Nine taxa had regres- sions of wing chord condition index with re- spect to time with significant F-values (slopes different from zero); all slopes were positive (Table 3). Kentucky Warblers and Veeries did not show significant gains in wing chord con- dition index, even though regressions of mass on time had significant positive slopes, and fat score regressions with respect to time also showed significant positive slopes for Ken- tucky Warblers (Table 3). Both species had marginal F-values for gain in wing chord con- dition index with respect to time (Kentucky Warbler, P = 0.064; Veery, P = 0.068). Four taxa had significantly positive slopes a 0s0Nt^O(NsDaN3’0s3’s0'or^0N3' 0 r^m'0 osm^-mos cosininosinr-sOsDin o Z in 3 ^ooddddddd +1 +1 +1 +1 +1 S +1 +1 +1 +1 +1 +1 +1 +1 +1 (Nt^inos — E'sj-'sOm^'ooso — cMm lnsoooa^O ,.^(Nr-ooinpinsD— ;oq 7 C/3 i/SFo6sod.3FdF — — ^dosso 3 M) -H— 'CM(NmZ — — ^(NCNCN--— ' — 3 < —1 ^ r-v sOOsr^OOs — sOOO — (M'ltON — 0>n OJ oomsom — — r^nooor'r-r^_oo + 1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 3 £ > oo(NO(NsD(nso — inoocNOsONino ooosr^cNmin3'r^oo(NCNinONin3‘_ d d d d d d d d d d d — ' d sor^ONOsOooininsor^r^'Oinsovo c £ 3 3 ^^^in ^00 ooaNr^oinsooo'3"^3'inor^ooc O oomsomTtcNr^mcnTtoosO'^inoo ^ ^ ^ ' ^ ^ ^ s; osooNcnr-oor-ooin-^— 'Os(N^ mincN^— '00'^inoo(N3;^_sO'pp c/3 — ‘ddddddddd_ — dd — (U CJ +1 +i +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 3 mooinsominr^mr'3'cnr'-^-^0 3 r^oo— ■mosm-^oscN'ctinr^rn'^in .3 dddoNdddsoddddoNONin "3 O 'Sd _o 03 "o 3 & 0 ^ d ^ 0 £ Q a c d d c t ? S b TABLE 1. Mean ± SI Species idonax spp. eyed Vireo, Vireo olivaceus y, Catharus fuscescens inson’s Thrush, C. ustulatus d Thrush*’, Hylocichla mustelin ’ Catbird*’, Dumetella carolinen nolia Warbler*’, Dendroica mag ;rican Redstart*’, Setophaga ruti m-eating Warbler, Helmitheros abird, Seiurus aurocapillus hern Waterthrush, S. noveborac tucky Warbler, Oporornis form< imon Yellowthroat*’, Geothlypis ded Warbler, Wilsonia citrina go Bunting*’, Passerina cyanea l'SSi|2|£|>oS§o| tSc^>cn^OS<^OZ;4UP:>£ 8 Johnson and Winker • MIGRATION STOPOVER NEAR THE GULE OE HONDURAS 281 TABLE 2. Body mass 2001. of woodland migrants mist-netted in Big Ealls, Belize, 11 August-15 November Lean mass^ Mean mass Fat-free mass^ Lean mass f Fat-free mass f Red-eyed Vireo'^ 16.88 14.59 10. 14*** Veery'^ 32.15 26.66 g 92*** Swainson’s Thrush 26.5 29.36 24.18 3.34* 23.50*** Wood Thrush 42.0 42.93 42.21 1.12 1.23 Gray Catbird 32.4 34.35 31.80 4.17* g Magnolia Warbler 7.0 7.17 6.92 1.87 2.63** American Redstart 6.75 6.93 6.49 1.11 5 Q3*** Worm-eating Warbler 12.0 12.27 10.79 1.12 Ovenbird 16.7 17.44 15.52 1.94 Northern Waterthrush 15.3 15.53 13.68 0.698 9 Q2*** Kentucky Warbler 12.0 12.77 11.36 2.87* 9 25*** Common Yellowthroat 9.15 9.31 8.36 0.836 5 38*** Hooded Warbler 9.5 9.41 8.2 -0.692 9 37*** Indigo Bunting 13.1 13.50 12.34 1.23 8.76*** ^ Mean mass of individuals with fat score of zero. Dunning (1993). *P < 0.05; ** P < 0.005; *** P < 0.0005. No Veeries or Red-eyed Vireos had a fat score of zero. of fat scores with respect to time (Table 3), and all of these species also had significant slopes of mass with respect to time. Kentucky Warblers had significant slopes of mass and fat score with respect to time, and it appeared this species was fattening at the site even though the wing chord condition index re- gression with respect to time was not signifi- cant (Table 3). Tail length condition index in this species had a slope significantly different from zero (Table 3), which was used to obtain an estimate of net daily mass gain. Veeries did not show any other evidence of mass gain be- sides a significant positive slope in mass with respect to time, and they were excluded from further analyses. Regressions of wing chord with respect to time for Swainson’s Thrushes and Red-eyed TABLE 3. Relationships between morphological characters and time of capture of individual woodland migrants in Big Ealls, Belize mist-netted between 11 August and 15 November 2001. Values are F-statistics from linear regressions of the character against time and test the null hypothesis that slope of the linear model was not significantly different from zero. Asterisks indicate P < 0.05 for the F-test. Species Mensural characters Condition indice.s^ Wing chord Tail length Mass Fat Wing chord Tail length Empidonax 0.2 3.1 8.8* 7.7* 9.8* 4.8* Red-eyed Vireo 5.3* 2.3 8.3* 0.5 4.1* 3.0 Veery 1.9 1.2 4.8* 3.0 3.5 2.8 Swainson’s Thrush 8.5* 4.0* 7.9* 13.1* 4.2* 3.2 Wood Thrush 0.2 0.2 0.1 2.8 0.2 Gray Catbird 2.2 0.2 24.2* 0.6 18.8* 5.0* Magnolia Warbler 0.4 0.4 6.5* 3.4 7.0* * oc American Redstart 2.2'’ 6.7* 6.6* 6.6* Worm-eating Warbler 1.1 1.1 8.6* 1.2 6.5* 4.0 Ovenbird 0.0 0.0 2.1 2.0 2.9 1.6 Northern Waterthrush 0.1 1.8 8.2* 2.8 8.8* 3.1 Kentucky Warbler 3.9 0.5 7.4* 4.4* 3.6 5.2* Common Yellowthroat 0.2 0.0 0.8 0.4 0.5 1.8 Hooded Warbler 0.0 0.2 1.8 0.2 2.3 1.4 Indigo Bunting 0.0 0.9 4.3* 0.1 5.0* 0.9 “ Mass divided by morphological character. Variable not normally distributed. 282 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 TABLE 4. Linear models for diurnal change in condition index for woodland migrants^ at Big Falls, Toledo District, Belize, 11 August-15 November 2001. Condition Species n b m SE m F p r2 gain/day^ Empidonax 88 0.14936 0.00260 0.00083 9.8 0.0025 0.11 0.03228 Red-eyed Vireo 39 0.19128 0.00276 0.00136 4.1 0.0504 0.10 0.03427 Veery 66 0.31382 0.00287 0.00154 3.4 0.0675 0.05 C Swainson’s Thrush 306 0.29956 0.00133 0.00065 4.2 0.0418 0.01 0.01651 Wood Thrush 45 0.43273 -0.0012 0.00251 0.2 0.6334 0.01 C Gray Catbird 269 0.36757 0.00406 0.00094 18.8 <0.0001 0.07 0.05041 Magnolia Warbler 77 0.11762 0.00106 0.00040 7.0 0.0100 0.09 0.01316 American Redstart 31 0.11017 0.00086 0.00034 6.6 0.0157 0.18 0.01071 Worm-eating Warbler 31 0.17130 0.00160 0.00063 6.5 0.0165 0.19 0.01987 Ovenbird 43 0.23247 0.00146 0.00086 2.9 0.0955 0.07 C Northern Waterthrush 84 0.19323 0.00250 0.00085 8.8 0.0041 0.10 0.03104 Kentucky Warbler^ 59 0.25142 0.00243 0.00106 5.2 0.0259 0.09 0.03017 Common Yellowthroat 48 0.17598 0.00360 0.00052 0.5 0.4916 0.01 C Hooded Warbler 51 0.14421 0.00085 0.00056 2.3 0.1372 0.05 c Indigo Bunting 87 0.20401 0.00138 0.00062 5.0 0.0274 0.06 0.01714 a Equations are Y = b + mX, where Y is condition (g/mm), m is slope (condition change/hr), b is the Y intercept, and X is time (hr). F-statistic and corresponding P-value indicate whether the slope differed from zero; r2 is the coefficient of determination and is a measure of the strength of the relationship between time (X) and condition (10. ^ Units are g/mm for the average day length of 12.42 hrs. c Slope not significantly different from zero and gains were not estimable. Tail condition index used to estimate mass gain. Vireos revealed a significant positive slope (Table 3), suggesting that larger individuals of these two species were more likely to be cap- tured later in the day. We divided the sample of Swainson’s Thrushes in half by wing chord to examine the possibility that wing chord might impart, rather than remove, a size bias when calculating individual condition index values. We tested whether the longer- winged half of our sample had greater condition in- dices than the shorter-winged half using a one- tailed Mest. There was no difference in con- dition indices between the two groups (t = 0.62, df = 303, P = 0.27), suggesting the sig- nificant relationship of wing chord with time of day in these species was not causing the significant relationship between our wing chord-based condition index and time of day. Linear models (Table 4) for the 10 taxa that had significant slopes of condition index with respect to time (Table 3) were used to estimate average gross and net daily mass gains (Table 5). Net estimates of mass gain were variable among those species showing gains. For ex- ample, average gains for Swainson’s Thrush were less than 4% of lean body mass, but in four taxa (Empidonax, Gray Catbird [Dume- tella carolinensis]. Red-eyed Vireo, and Northern Waterthrush [Seiurus noveboracen- sis]) our estimates showed net diel mass gains of more than 10% of mean body mass (Table 5). Flight Capacity Estimates. — Estimates of flight times possible with taxon-average gains varied from <3 to >10 hrs. Average individ- uals of four taxa {Empidonax, Gray Catbird, Red-eyed Vireo, and Northern Waterthrush) were estimated to be capable of flying for be- tween 8 and 1 1 hrs if all mass gained was fat (i.e., an entire night after just 1 day of fatten- ing; Table 5). DISCUSSION Our data suggest that 1 0 of the 1 5 taxa stud- ied used this site to acquire resources and gain mass, and that no taxa had significantly neg- ative daily mass gain estimates. The extent of mass gain varied among taxa at the site. All taxa with significant mass gains gained suffi- cient mass to more than offset estimated noc- turnal losses. The substantial net gain estimates in some taxa {Empidonax, Red-eyed Vireo, Gray Cat- bird, and Northern Waterthrush) suggest that habitat at this site was sufficiently favorable to allow for nearly a full night of migration after only 1 day of feeding; 90% of the taxa had a net gain of at least 5% of lean mass per Johnson and Winker • MIGRATION STOPOVER NEAR THE GULF OF HONDURAS 283 TABLE 5. Estimates of daily net increases in uses tail condition index) for woodland migrants November 2001. Units are g, except where noted. mass using wing chord condition index (Kentucky Warbler captured in Big Falls, Toledo District, Belize, 11 August-15 Species Gross gain per day^* Existence metabolism'^ Existence metabolism as percentage of lean mass Net gain per flay‘d Increase as percentage of lean mass Elight cost (g/hr)d Hours of flight^ Empidonax 2.13 0.53 1.60 11.43 0.15 10.82 Red-eyed Vireo 2.57 0.44 2.82 2.13 13.64 0.21 10.10 Swainson’s Thrush 1.54 0.61 2.30 0.93 3.51 0.36 2.57 Gray Catbird 4.31 0.69 2.13 3.62 11.18 0.42 8.61 Magnolia Warbler 0.74 0.27 3.86 0.47 6.76 0.09 5.17 American Redstart 0.62 0.26 3.85 0.36 5.31 0.09 4.05 Worm-eating Warbler 1.29 0.37 3.08 0.92 7.65 0.15 5.95 Northern Waterthrush 2.21 0.43 2.81 1.78 11.64 0.19 9.17 Kentucky Warbler 1.38 0.37 3.08 1.01 8.42 0.16 6.29 Indigo Bunting 1.07 0.39 2.98 0.68 5.18 0.17 4.01 ^ Mass gain during 1 day for average individual using average size for wing chord (Kentucky Warbler uses tail length). Mass loss due to existence metabolism (Kendeigh 1970). Net 24-hr mass gain after subtraction of nightly mass loss of existence metabolism from gross gain/day. ^ Cost of flight in g of fat/hr calculated following Tucker (1974:306) using average mass of captured individuals. ® Hours of flight possible calculated from Tucker (1974) using an energy value of 30.2 kJ per g (Pennycuick 2003). day (Table 5). Our estimates are conservative for individuals that stopped for only 1 day and did not spend a night at our site because, as nocturnal migrants, they would lose little to nocturnal resting metabolism, and the amount available for migration would be closer to the gross mass gain estimates. Only 40% of species with significant mass gain estimates also had significant fat score trends. Fat scores are subjective ordinal esti- mates of the fat content in birds, not absolute measures, and they estimate only visible sub- cutaneous deposits on the venter of a bird, even though fat is also deposited in other ar- eas (King and Famer 1965). The association between lipid index (g lipid per g lean dry mass) from fat extractions and fat scores for wintering Dark-eyed Juncos (Junco hyemalis) taken by one experienced observer on the same birds was high (r^ = 0.974; Rogers 1991). Rogers (1991) showed that an experi- enced observer could detect small changes in fat stores by scoring visible fat, but he cau- tioned that inter-observer differences in fat scores could affect analyses of fat score data. Fat scores were taken by two observers during our study, and one had not previously used the technique. It is likely the lack of signifi- cance in fat score regressions when so many species showed significant mass and condition index gains was, at least in part, due to inter- observer variation and the confounding factor of observer inexperience. Thus, we emphasize these are estimates, but assume that gains in condition index reflect gains in mass. If the significance of the relationship of wing chord with respect to time could impart a bias to wing chord condition index values, as Winker (1995a) suggested might occur, one would expect a significant difference between the mean condition indices of the larger (lon- ger-winged) individuals and the smaller (shorter-winged) individuals. Our analysis of Swainson’s Thrushes, showing the longer- winged half of our sample did not have dif- ferent condition indices than the shorter- winged half, demonstrated concern about a possible failure of this methodology is un- warranted. The year-to-year variability at this site can- not be addressed with these data, especially since the site was severely altered by Hurri- cane Iris. Species with significant daily gains in 1 year in Minnesota during a 3-year study often showed significant gains in more than 1 year (Winker 1995a). However, Dunn (2000) has shown significant differences in mass gains among years at a single site in a single species. Habitat damage caused by Hurricane Iris may have changed the suitability of the site for some migrants, although it is notable that four of the six species in our sample still had significant mass gains at this site after Hurricane Iris. 284 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 Species such as Veery and Swainson’s Thrush, whose winter ranges are entirely or largely in South America, might have been expected to be fattening as much as Red-eyed Vireos, which also winter entirely in South America. A migration strategy that would ac- count for low levels of fattening is that higher resource certainty in the tropics compared to temperate habitats might not require deposit- ing large amounts of fat; individuals may even be less likely to fatten if they are already close to their winter destinations (Winker 1995a). However, this hypothesis does not seem to ap- ply to Veeries and Swainson’s Thrushes, which were generally carrying large amounts of fat at our site (Table 1). This may explain why Veeries apparently were not fattening at the site (Table 3) and Swainson’s Thrushes had the lowest estimate of all fattening species (Table 5), even though they were among the species with the farthest minimum distance yet to travel. The significant differences in several species between lean or fat-free mass and mean mass at our site (and the generally heavy fat loads for several species; Table 1) demonstrate that stopover sites farther north are important areas for fattening for some spe- cies. Where the species that arrive fat obtain this fat, and why some species arrive fatter than others, remain important ecological and evolutionary questions. The high level of fattening by some species and the high level of fat already carried by others is similar to the amount of fattening observed in migrants preparing to make west- ern trans- Atlantic flights from New England (e.g., Nisbet et al. 1963) and trans-gulf flights from the central coast of the Gulf of Mexico (Woodrey and Moore 1997). This is in con- trast to the generally low levels of fattening observed in other, mid-continental studies (Winker et al. 1992, Dunn 2002), and is dif- ferent from the low levels of fattening ob- served by Winker (1995a) on the Isthmus of Tehuantepec. The high levels of fat observed at our site may be in preparation for a long- distance migratory flight, rather than shorter flights south through Central America. The emaciated migrants observed by Rogers and Odum (1966) in Panama would seem to sug- gest they arrived from a great distance, also supporting the possibility of longer flights over Central America, rather than shorter flights through it. Mass gain data from farther south in Central America would be useful in elucidating this situation. Our study suggests that lowland forests in the southeastern Yucatan Peninsula are im- portant to many species of transient migrants as an area to build fuel stores. Some species showed average trends of depositing sufficient fat in 1 day for an entire night of migration. However, some species arrived at this site car- rying substantial fat loads apparently from far- ther north. This study combined with other studies of fattening and stopover ecology of Nearctic-Neotropic migrants is helpful for de- veloping hypotheses about geographic varia- tion and interspecific differences in fuel de- position strategies in the Nearctic-Neotropic migration system. Much work remains to be done in many of these same areas during spring migration, and there are few or no pub- lished data collected during any season from South America, Central America between Be- lize and Panama, or the southeastern United States. It does appear that more than one strat- egy for fattening in northern Central America has evolved among passerine migrants. ACKNOWLEDGMENTS This study was funded by the University of Alaska Museum, the U.S. Department of Agriculture (SCA 58-6612-8-022), an anonymous donor, and the Uni- versity of Alaska Fairbanks Graduate School. Able field assistance was provided by Santos Hun. Critical support in Belize was provided by Donald Owen-Lew- is, the Francisca Bardalez family, the Cantel Hun fam- ily, Peter Dunham, and K. M. Prufer. Personnel in the Conservation Division of the Forest Department in Be- lize provided permits to conduct the research and working with them was a pleasure. A. N. Powell, D. L. Thomas, E. C. Murphy, E. H. Dunn, and two anon- ymous reviewers provided helpful editorial sugges- tions and statistical advice. D. M. Johnson served as our botanical consultant. LITERATURE CITED American Ornithologists’ Union (AOU). 1998. Check-list of North American birds. Seventh Edi- tion. American Ornithologists’ Union, Washing- ton, D.C., USA. Bauchinger, U. and H. Biebach. 2000. Differential catabolism of muscle protein in Garden Warblers (Sylvia borin): flight and leg muscle act as a pro- tein source during long distance migration. Jour- nal of Comparative Physiology B 171:293-301. Child, G. I. 1969. A study of nonfat weights in mi- Johnson and Winker • MIGRATION STOPOVER NEAR THE GULF OF HONDURAS 285 grating Swainson’s Thrushes (Hylocichla ustula- ta). Auk 86:327-388. Connell, C. E., E. P. Odum, and H. Kale. 1960. Fat- free weights of birds. Auk 77:1-9. Dunn, E. H. 2000. Temporal and spatial patterns in daily mass gain of Magnolia Warblers during mi- gratory stopover. Auk 117:12-21. Dunn, E. H. 2002. A cross-Canada comparison of mass change in birds during migration stopover. Wilson Bulletin 114:368-379. Dunning Jr., J. B. 1993. CRC handbook of avian body masses. CRC Press, Boca Raton, Florida, USA. Freeman, S. and W. M. Jackson. 1990. Univariate metrics are not adequate to measure avian body size. Auk 107:69-74. Galindo, P. and E. Mendez. 1965. Banding of thrush- es and catbirds at Almirante, Panama. Second year of observations. Bird-Banding 36:233-239. Galindo, P, E. Mendez, and A. J. Adames. 1963. Banding of migrant thrushes in Almirante, Pana- ma. Bird-Banding 34:203-209. Helms, C. W. and W. H. Drury. 1960. Winter and migratory weight and fat field studies of some North American buntings. Bird-Banding 31:1-40. Karasov, W. H. and B. Pinshow. 1998. Changes in lean mass and in organs of nutrient assimilation in a long-distance passerine migrant at a spring- time stopover site. Physiological Zoology 71:435- 448. Kendeigh, S. C. 1970. Energy requirements for exis- tence in relation to size of bird. Condor 72:60- 65. King, J. R. and D. A. Earner. 1965. Studies of fat deposition in migratory birds. Annals of the New York Academy of Sciences 131:422-440. Monroe Jr., B. L. 1968 A distributional survey of the birds of Honduras. Ornithological Monographs 7. Nisbet, I. C. T, W. H. Drury Jr., and J. Baird. 1963. Weight-loss during migration. Part I. Deposition and consumption of fat by the Black-poll Warbler Dendroica striata. Bird-Banding 34:107-138. Pennycuick, C. j. 1989. Bird flight performance: a practical calculation manual. Oxford University Press, Oxford, United Kingdom. Pennycuick, C. J. 2003. The concept of energy height in animal locomotion: separating mechanics from physiology. Journal of Theoretical Biology 224: 189-203. Rising, J. D. and K. M. Somers. 1989. The measure- ment of overall body size in birds. Auk 106:666- 674. Rogers, C. M. 1991. An evaluation of the method of estimating body fat by quantifying visible subcu- taneous fat. Journal of Field Ornithology 62:349- 356. Rogers, D. T. and E. P. Odum. 1964. Effect of age, sex, and level of fat deposits on major body com- ponents in some wood warblers. Auk 81:505-513. Rogers, D. T. and E. P. Odum. 1966. A study of au- tumnal postmigrant weights and vernal fattening of North American migrants in the tropics. Wilson Bulletin 78:415-433. Stiles, E G. and D. L. Altshuler. 2004. Conflicting terminology for wing measurements in ornitholo- gy and aerodynamics. Auk 121:973-976. Tucker, V. A. 1974. Energetics of natural avian flight. Publications of the Nuttall Ornithological Club 15:298-333. Winker, K. 1995a. Autumn stopover on the Isthmus of Tehuantepec by woodland Nearctic-Neotropic migrants. Auk 112:690-700. Winker, K. 1995b. Neotropical stopover sites and Middle American migrations: the view from southern Mexico. Pages 150-163 in Conservation of neotropical migratory birds in Mexico (M. Wil- son and S. Sader, Editors). Miscellaneous Publi- cation. Maine Agricultural and Forestry Experi- ment Station, Orono, USA. Winker, K., D. W. Warner, and A. R. Weisbrod. 1992. Daily mass gains among woodland migrants at an inland stopover site. Auk 109:853-862. WooDREY, M. S. AND E R. MooRE. 1997. Age-related differences in the stopover of fall landbird mi- grants on the coast of Alabama. Auk 1 14:695— 707. The Wilson Journal of Ornithology 1 20(2):286— 295, 2008 NUMBERS OF MIGRATORY BIRDS STOPPING OVER IN NEW ORLEANS, LOUISIANA, USA IN RELATION TO WEATHER PETER H. YAUKEY'^ AND SHAWN C. POWELL' ABSTRACT. Migratory birds were surveyed at stopover sites in New Orleans, Louisiana, during seven spring and fall seasons in relation to weather conditions. Weather was classified into four synoptic scenarios distinguished primarily by positions of fronts, wind directions, and pressure characteristics. Overall numbers of migrants in spring were more numerous during the synoptic scenarios in which a cold front was about to arrive (Frontal Gulf Return) or had already passed (Post Frontal), than when airflow was off the Gulf or from the east (Coastal Return, which often preceded frontal passage). Counts of White-eyed {Vireo griseus) and Red-eyed vireos (V. olivaceus). Indigo Buntings (Passerina cyanea), and thrushes (Hylocichla and Catharus combined) showed differences among weather scenarios, mostly with high counts under Post Frontal conditions and/or low counts under Coastal Return. Post Frontal values were higher in fall than Coastal Return for overall migrants. Common Yellowthroats (Geothlypis trichas), and Indigo Buntings. However, four individual species failed to show any variation with weather. Counts after cold front passages in both spring and fall were higher than those preceding the fronts for all migrants combined. Received 29 July 2004. Accepted 27 September 2007. Studies of the ecology and behavior of mi- gratory birds during passage are important, given the significance of the migratory jour- ney in their annual cycles. A key component of the migratory strategies of migrants is use of stopover sites for resting, re-hydrating, and refueling during their journeys, as well as finding protection from predators. The number of birds using a given site can vary dramati- cally from day to day; weather has long been thought to have a key role in daily variations in numbers of birds migrating and occurring at stopover sites (reviewed by Lack 1960; Richardson 1978, 1990). The number of birds occurring at a stopover site on a given day is a function of several variables including the volume of migration on the preceding night or nights, and decisions made by birds using the site on previous dates whether to depart or remain another day. Both are influenced by weather; good migratory weather can bring migrants in and stimulate others to depart. Substantial research has been devoted to the impact of weather on migration and researchers have arrived at a consensus regarding basic weather conditions that favor migration (e.g.. Lack 1960; Richardson 1978, 1990; Alerstam 1990; Berthold 1993). For in- stance, fall migration is heaviest under con- ditions of rising or high pressure and gentle or following (northerly) winds, as often occur ' Department of Geography, University of New Or- leans, New Orleans, LA 70148, USA. 2 Corresponding author; e-mail: pyaukey@uno.edu following cold front passage. These studies have associated spring migration with tail winds as well, but these are southerly and of- ten associated with decreasing pressure. Pre- cipitation is generally unfavorable for migra- tion. This fall scenario applies to New Or- leans, Louisiana, but spring migration is somewhat different due to the location of the city, 100 km inland from the Gulf of Mexico. First, the southerly latitude of New Orleans causes new arrivals to approach across tropi- cal and low subtropical latitudes over the Gulf of Mexico where weather patterns differ from those of the mid latitudes where most other studies have been conducted. Second, most trans-Gulf migrants in spring do not arrive in the vicinity of New Orleans until well into the daylight hours; they have been shown to mostly fly well inland past New Orleans be- fore alighting (Lowery 1945, Gauthreaux 1971). The lack of birds stopping over in near- coastal areas of Louisiana has led to these ar- eas being labeled the “coastal hiatus” (Low- ery 1945). It is reasonable to expect that mi- gration over the Gulf is strongest when there are tailwinds (southerly) and no rain. Gulf Coast observers (e.g., Duncan 1994) have generally perceived that numbers of birds alighting in near-coastal areas such as New Orleans are greatest when a cold front pene- trates the Gulf, presenting inclement weather for northbound trans-Gulf migrants causing them to stop near the coast rather than travel farther inland as usual. 286 Yaukey and Powell • STOPOVER OF MIGRANTS AND SYNOPTIC WEATHER 287 Our overall objective was to examine the effect of weather on number of migratory birds stopping over in New Orleans, Louisi- ana. We compared numbers of migrants among different synoptic conditions before and after passage of cold fronts based upon an existing weather classification system (Muller 1977). Peak numbers in spring were predicted under conditions that cause north- bound migrants to “stall” in New Orleans (cold front passage, bringing inclement weath- er into the paths of northbound trans-Gulf mi- grants). We did not predict which scenarios would have the highest counts in fall, because cold front passages could either elevate stop- over bird counts by stimulating arrival of birds, or lower them by stimulating departure. METHODS Bird Surveys. — Migrant birds were counted on 1 5 survey transects in two urban residential neighborhoods and two city parks within New Orleans, Louisiana. One neighborhood (6 transects) was typical of New Orleans and had relatively sparse vegetation density, while the other (3 transects) was among the most-veg- etated residential areas of the city. Residential transects were along quiet roads and alleys. One park (2 transects, 2.5 ha) had an exten- sive tree canopy, but no understory vegetation or ground cover; the other park site (4 tran- sects) was a 13-ha woodlot with a basically intact understory, especially of camphor trees (Cinnamomum camphora), exotic elms (Ul- mus spp.), and giant ragweed {Ambrosia tri- fida). The dominant canopy trees on the four sites were live oak (Quercus virginiana), wa- ter oak (Q. nigra), Nuttall oak {Q. nuttallii), hackberry {Celtis laevigata), Chinese tallow {Sapium sehiferum), pecan (Cary a illinoen- sis), magnolia (Magnolia spp.), sweetgum (Liquidamhar styraciflua), loblolly pine (Fi- nns taeda), and slash pine (P. ellioti). The most distant sites were 6 km apart. The urban context of this study may have caused the sampling area to be avoided by some migrants if they perceived the area to be low quality stopover habitat. Transects were straight lines 200 m in length; 3 min were spent walking each and all migratory birds seen or heard within 25 m were recorded. The primary author conducted all bird surveying. This relatively rapid pace of coverage was designed to maximize the number of migrants encountered, which pilot surveys indicated might be accomplished by using a relatively fast pace to maximize the area surveyed. The fast pace allowed all 15 transects to be covered on each survey date, eliminating coverage differences among days and maximizing comparison of migrant counts made on different dates. Transects were surveyed before 4.5 hrs after sunrise, except on 10 days when surveying occurred later (ending within 7 hrs after sun- rise) due to rain. Gauthreaux (1971) detected some spring trans-Gulf migrants arriving over New Orleans in the early morning, but nor- mally peaking after 1200 hrs CST The large majority of birds counted in spring were prob- ably arrivals from previous days. The se- quence of visitation of the four sites was ro- tated each visit. Serial autocorrelation was minimized by allowing 2-3 days to elapse be- tween consecutive surveys, a time span simi- lar to previously reported mean multiple-day stopover lengths of captured migrants in spring in non-urban habitats on the Gulf Coast (Moore and Kerlinger 1987, Loria and Moore 1990). The number of instances of the same species being recorded on the same transect on consecutive visits was calculated at the end of the season to conservatively assess the maximum possible rate of repeat observations of the same individuals. Surveying was conducted from ~ 1 5 March to 15 May each spring and 20 August to 31 October each fall from 1994 to 2000. Birds were not surveyed during high winds or heavy rains. All species that breed on or near the study sites were excluded to restrict the spring analysis to passage migrants — including Mis- sissippi Kite (let ini a mississippiensis). Yel- low-billed Cuckoo (Coccyzus americanus). Great Crested Flycatcher (Myiarchus crini- tus). Eastern Kingbird (Tyranniis tyrannus). Purple Martin (Prague suhis), and Bronzed Cowbird (Molothrus aeneus). Most species that winter in the southeastern United States were also excluded from both spring and fall analyses because they wintered in or near the study area, although the absence of wintering Blue-headed Vireo (Vireo solitarius). Winter Wren (Troglodytes troglodytes). Gray Catbird (Dumetella carol inensis). Blue-gray Gnat- catcher (Polioptila caeridea). Common Yel- 288 THE WIl.SON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 FIG. 1. Synoptic weather categories (after Muller 1977). Map by Eric Bock. lowthroat {Geothylpis trichas), Wilson’s War- bler (Wilsonia pusilla), and Eastern Towhee (Pipilo erythrophthalmus) allowed their inclu- sion. Weather Condition Categories. — Muller (1977) devised an eight-category classifica- tion system for describing weather conditions in Louisiana and the Louisiana Monthly Cli- mate Review has published daily classifica- tions of the 0600 and 1500 hrs CST conditions for several cities in Louisiana on an ongoing basis. Muller’s categories are based upon po- sitions of fronts, wind directions, and other criteria (Fig. 1). Six categories are commonly related to the sequence of weather changes as a cold front approaches: an easterly airflow Yaukey and Powell • STOPOVER OF MIGRANTS AND SYNOPTIC WEATHER 289 occurring while an approaching cold front is still distant is typical of the “Coastal Return” scenario, which transitions into the “Gulf Re- turn” scenario with more southerly flow as the front draws closer. “Frontal Gulf Return” and “Frontal Overrunning” scenarios occur im- mediately before and after the frontal passage, respectively, and are followed by “Continen- tal High” conditions in the fair weather and high pressure after the front. When the ridge is (infrequently) centered west of the Rockies, the post-frontal high pressure is termed a “Pa- cific High.” These synoptic types at times oc- cur outside this frontal sequence, but the Con- tinental High and Frontal Overrunning are closely associated with post-frontal condi- tions; in both spring and fall, >90% of the occurrence of each in this study was imme- diately after a cold frontal passage (i.e., before the appearance of other weather types, but in- frequently occurring with Pacific High). Gulf Return and Coastal Return are the scenarios most different from those following a cold front passage (i.e., the least likely to have northerly winds). They are also the conditions most likely to precede a cold front passage. Two additional categories are used that do not fit into a scheme of frontal passage: “Gulf High” when weather is dominated by a high pressure cell in the Gulf, and “Gulf Tropical Disturbance” when influenced by a tropical low or storm in the Gulf. These eight cate- gories were simplified to four for our analysis: Frontal Overrunning and Continental High were combined into a category labeled “Post Frontal,” Gulf and Coastal Return were com- bined and labeled “Coastal Return,” and Frontal Gulf Return and Gulf High were used without alteration. Pacific Highs and Gulf Tropical Disturbances were infrequent and omitted from statistical analyses. Statistical Analyses. — Numbers of migrant birds were compared among four synoptic cat- egories of climate using Krukal-Wallis tests with accompanying multiple comparison tests (Zar 1984). The number of migrants passing through New Orleans changes as spring or fall progresses, which could complicate simple comparison of bird numbers among synoptic types (whose numbers and magnitude of day to day variations also might vary through the migration seasons). Thus, numbers of birds re- corded on each daily survey were “corrected" for these factors by subtracting from each the “normal” seasonal abundance level (estimat- ed by calculating 5-day means for each date from the 7 years of data in the study), and dividing this residual by the standard devia- tion of those same counts. Autumnal migrants in New Orleans might be influenced by weath- er conditions to the north which they fly through during their nocturnal migratory flight. Thus, these “corrected” migrant num- bers were compared among different synoptic categories of weather occurring at Monroe, Louisiana, 350 km NNW of New Orleans, on the afternoon before the survey date. This was the most northerly station at which synopti- cally classified weather data were available. The most abundant species of migrants for spring and fall seasons were analyzed inde- pendently in addition to analyzing total num- bers of migrants. Analyses of individual spe- cies were confined to surveys within their nor- mal times of passage, excluding dates earlier and later than their extremes of occurrence over the 7 years. Weather analyses for Yellow Warblers (Dendroica petechia) and American Redstarts (Setophaga ruticilla) were confined to on or before 10 September and on or after 1 October, respectively, to avoid potential confusion of their vocalizations during their period of overlap (assessed by visual obser- vations of the species). Bird numbers were also compared before and after passage of cold fronts depicted in the publication Daily Weather Maps in addi- tion to Muller’s (1977) synoptic categories (U.S. Department of Commerce, 1994-2()()0). Any frontal boundary that passed New Or- leans from north to south was included in the analysis even if it was designated a stationary front in Daily Weather Maps. We compared the raw (uncorrected) number of migrants for the last survey prior to each front with the first survey following it using a paired Student's t- test. Correction of numbers for seasonal var- iations in abundance was not considered nec- essary for this analysis since the two counts compared were on similar dates in each case. Neither front was included in the analysis it more than one front (including warm fronts) passed New Orleans between consecutive mi- grant surveys. 290 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 TABLE 1. Synoptic weather classification system for New Orleans, Louisiana, USA (after Muller 1977) with number of survey dates («) on which each occurred over the 7 years of study. Type Spring (n) Fall («) Description Coastal Return"* 13 46 Winds veer from northeast to southeast as ridge drifts eastward (spring) or have easterly component when Bermuda High extends into southeastern U.S. (fall). Gulf Return"* 62 38 Winds veer to southeast and south as air is drawn off Gulf by ridge drifting farther east or in front of low advancing from Texas panhandle; front distant. Frontal Gulf Return 23 13 Front nearby; return flow in warm sector of system. Frontal Overunning^ 24 17 Polar front stationary along Gulf Coast/northern Gulf. Continental High^* 27 52 High pressure center usually east of Rockies; northerly flow; fair and cool to cold weather associated with core of high. Pacific High 4 2 High center usually over eastern Pacific or west of Rock- ies; mild and dry air after cold front. Gulf High 8 14 Southwestern flow associated with western extension of high over Gulf; in summer often extension of Bermuda High. Gulf Tropical Disturbance 1 18 Influenced by tropical disturbances ranging from easterly waves to hurricanes. ® Combined for statistical analyses into a single category, “Coastal Return”. ^Combined for statistical analyses into a single category, “Post Frontal”. RESULTS Synoptic Weather During Spring Pas- sage.— During the seven springs of the study, 998 migrants were recorded on the 173 survey dates. Only 4.9% of the birds counted were on the same transect as an individual of the same species on the previous census date. This total includes all conspecifics not just du- plicate observations of a single individual. It is not known how the mean stopover length of birds compared to the interval between vis- its but, because the number of potential repeat observations of the same individual birds was small, it was assumed that serial autocorrela- tion was unimportant. The most common synoptic scenarios at New Orleans in spring were the Coastal Re- turn (air being drawn from the east or off the Gulf, often as a cold front approached) and Post Frontal (Table 1). The frequency of oc- currence of each scenario varied markedly from year to year; Coastal Return conditions prevailed on 26 to 65% of days in any year. Frontal Gulf Return conditions (cold front near but not yet passed) varied from 4 to 24%, Post Frontal ranged from 17 to 48%, and Gulf High from 0 to 14%. The typical pattern was for migrants to be scarce on most days, punc- tuated by occasional days with increased counts, and on some occasions by spectacular numbers. Two or fewer migrants were record- ed on 51% of the days with none on 23%; counts of >30 migrants were recorded on only 4% of survey dates. The latter group typically represented “fallout” days when hundreds of individuals were present on and near the tran- sects, even though the highest daily count used for statistical analyses only reached 49 because of the strict survey protocol, and be- cause many migrants were excluded because their species also wintered on the site (e.g., Yellow-rumped Warbler [Dendroica corona- taV>. The number of spring migrants (corrected using mean and standard deviation) varied de- pending on the synoptic scenario present in New Orleans on the morning of the survey (Kruskal-Wallis, = 23.19, n = 155, P < 0.001). However, the synoptic conditions pres- ent the previous afternoon showed an even stronger pattern with respect to corrected count numbers (x^ = 30.06, n = 155, P < 0.001; Table 2). Trans-Gulf migrants typically arrive in New Orleans well into the daylight period and early morning surveys largely counted birds that arrived the previous day. Thus, the following analyses were based upon the synoptic conditions present the afternoon before each survey. Yaukey and Powell • STOPOVER OF MIGRANTS AND SYNOPTIC WEATHER 291 TABLE 2. Tests for differences in the number of spring migrants recorded when different synoptic weather scenarios existed at 1500 hrs the previous day. Dates n- Kruskal-Wallis p Significantly different scenarios^ White-eyed Vireo 17 Mar- 12 May 153 18.82 <0.001 Post>Coastal, Post>FGR Red-eyed Vireo 28 Mar- 14 May 127 14.00 0.003 Post>Coastal Thrush spp. 1 Apr- 12 May 112 15.60 0.001 GH>Coastal, FGR>Coastal Hooded Warbler 27 Mar-4 May 103 4.06 0.255 None Indigo Bunting 28 Mar- 14 May 127 12.80 0.005 Post>GH, FGR>GH Post>Coastal, All migrants 17 Mar- 14 May 157 30.06 <0.001 FGR>Coastal ^ Number of survey days within period of occurrence. ^ Coastal = Coastal Return (air flow from Gulf Coast); FGR = Frontal Gulf Return (cold front approaching); Post = Post Frontal (following passage of a cold front); GH = High pressure cell in Gulf. Coastal Return bird numbers (corrected) were significantly lower than for both Frontal Gulf Return (Q = 3.84; Qcritical = 2.64) and Post Frontal (Q = 4.84) conditions. These are normally conditions in which a front is arriv- ing at or has passed New Orleans. The four most numerous species recorded in spring were Indigo Bunting {Passerina cy- aneo). Red-eyed Vireo (Vireo olivaceous). White-eyed Vireo {V. griseus), and Hooded Warbler (Wilsonia citrina). Numbers of each of these species, except Hooded Warbler, var- ied significantly among the different synoptic scenarios (Table 3). Frontal Gulf Return (Q = 2.69) and Post Frontal (Q = 2.89) conditions recorded higher corrected counts for Indigo Bunting than did Gulf High. Coastal Return conditions were also lower than Frontal Gulf Return and Post Frontal, but not significantly. Coastal Return counts for Red-eyed Vireo were significantly lower than Post Frontal counts (Q = 3.44). White-eyed Vireo counts were higher under Post Frontal conditions than either Coastal Return (Q = 3.80) or Fron- tal Gulf Return (Q = 3.02). These scenarios recorded similar numbers, suggesting numbers of this species did not increase when a front was arriving. Coastal Return numbers were significantly lower than Frontal Gulf Return (Q = 2.69) and Gulf High (Q = 3.07) num- bers for thrushes. Synoptic Weather During Fall Passage. — A total of 1 , 1 97 migrants was recorded on 2 1 1 survey dates over the seven fall periods. Dur- ing fall, only 14.7% of individuals were on the same transect as a bird of the same species on the previous survey. Thus, it was presumed that serial autocorrelation was not a serious problem, because these can represent different individuals. The most common synoptic scenarios in fall were Coastal Return and Post Frontal (Ta- ble 1). The number of Coastal Return days varied annually between 35 and 54%, and Frontal Gulf Return between 4 and 15%. Be- tween 26 and 44% of days had Post Frontal TABLE 3. Tests for differences in numbers of fall migrants recorded when different synoptic weather .sce- narios existed at 1500 hrs the previous day. Dates Kru.skal-Wallis p Significantly different .scenarios^ Eastern Wood- Pe wee 21 Aug-28 Oct 172 3.34 0.342 None Gray Catbird 23 Sep-29 Oct 100 5.47 0.140 None Yellow Warbler 21 Aug- 10 Sep 48 1.39 0.707 None American Redstart 1 Oct-27 Oct 14 1.78 0.620 None Common Yellowthroat 25 Aug-29 Oct 168 15.74 0.001 Post>Coastal Indigo Bunting 22 Sep-29 Oct 130 10.63 0.014 Post>Coastal All migrants 2 1 Aug-29 Oct 176 10.60 0.014 Post>Coastal “ Number of survey days within perit)d of iKcurrence. Coastal = Coastal Return (air flow from Gulf Coast); FGR = Frontal Gulf Return (cold front approaching); F\)st ' Post F-rontal (following passage of a cold front); GH = High pressure cell in Gulf. 292 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 conditions each year, and between 0 and 19% were Gulf High days. Numbers of migrants in fall were highly variable from day to day with only 6% of survey dates exceeding 30 system- atically-counted migrants (corresponding to large numbers at the site). However, there were many fewer days with ^2 (10%) or none ( 1 %) detected than in spring. The norm in fall was to have small numbers, but not as small as in spring. Numbers of fall migrants did not vary when different synoptic conditions were present at 0600 hrs in New Orleans (x^ = 6.54, n = 178 dates; P = 0.088), but did vary when different scenarios had been present at 1500 hrs the previous afternoon (less strongly than in spring, P = 0.014, Table 3). More migrants were found under Post Frontal than Coastal Return conditions (Q = 3.19; Qcritical = 2.64). Bird numbers did not vary in associa- tion with the 1500 hrs conditions the previous afternoon in Monroe, Louisiana, 350 km NW of New Orleans (x^ = 7.05, n = 188 dates; P = 0.07). The most numerous species in fall were Eastern Wood-Pewee {Contopus virens). Gray Catbird, Yellow Warbler, American Redstart, Common Yellowthroat, and Indigo Bunting. In contrast to spring, only two species had sig- nificant variation in numbers among scenarios of synoptic weather (Table 3). Both Common Yellowthroat (Q = 3.72) and Indigo Bunting (Q = 3.06) were more numerous in Post Fron- tal than Coastal Return situations. Cold Fronts in Spring and Fall. — The total number of migrant individuals detected in spring was greater in the first survey follow- ing each cold front passage than in the survey prior to each front (mean = 62% increase; paired t = 2.04, n = 53, P = 0.023). Approx- imately 40% of these fronts were weak and showed signs of stalling before or after pass- ing New Orleans. Weak fronts might be ex- pected to have less effect on migrants, but their omission did not support this assumption (mean for strong fronts = 48% increase; paired t = 1.15, n = 31, P = 0.13). The number of migrant individuals in fall was greater following cold front passage than prior to it (mean = 48% increase; paired t = 2.16, n = 43, P = 0.018). Of these fronts, 50% showed signs of stalling in fall and omis- sion of these weak fronts revealed an even stronger pattern for the strong fronts (mean for strong fronts = 93% increase; paired t = 3.14, n = 22, P = 0.003). Gulf Tropical Disturbances. — Numbers of birds were greater under tropical disturbance conditions (0600 or 1500 hrs) than any other scenario. However, these differences were not significant, and patterns of individual species were complicated. Mornings following tropi- cal disturbance conditions in New Orleans ranked lower than any other synoptic scenario for Common Yellowthroats {n = 12 dates with Gulf Tropical Disturbances) and Yellow War- blers {n = 8), but second highest for Eastern Wood-Pewees {n — 16 dates). Too few Gulf Tropical Disturbances occurred during the rel- atively late migratory periods of Gray Cat- birds and Indigo Buntings to suggest patterns for those species. Big Days in Spring and Fall. — The occur- rence of migrants in New Orleans in both spring and fall was punctuated by occasional days when exceptional numbers stopped over. Daily transect counts in spring exceeded 30 migrants on seven dates, peaking at 49 on 4 May 1995. Three of these dates were in 1995; none occurred in 1998 or 1999. All 7 days were associated with frontal activity in the Gulf. These counts were only a small sample of the hundreds of birds that appeared to be present on each day. Peak counts in fall were somewhat higher than in spring and exceeded 40 on seven dates, peaking at 64 on 21 October 1998. All of these dates were associated with frontal passages and/or tropical disturbances. Interaction of Synoptic Scenario Frequency with Bird Numbers. — The mean number of birds recorded in fall was highest on days with Post Frontal conditions. However, this did not translate into larger season-long migrant counts during falls with more Post Frontal days. The percentage of Post Frontal days each fall (which ranged from 26 to 44% of the survey dates) was negatively correlated with the season-long mean daily count of migrants (each day adjusted for means and standard de- viations; Spearman’s r = —0.85, P = 0.015). Driving this finding was a strong negative cor- relation between the percentage of Post Fron- tal days each fall and the mean count of mi- grants on these days (r^ = -0.93, P = 0.003). The low seasonal counts for fall surveys with Yaukey and Powell • STOPOVER OF MIGRANTS AND SYNOPTIC WEATHER 293 many Post Frontal days, despite the high av- erage counts on Post Frontal days overall, might be explainable by variations in the du- ration of high pressure following cold front passages from year to year. If high counts oc- cur on the first few days after the onset of Post Frontal conditions, followed by reduced num- bers as the high pressure continues, fall sur- veys with long stretches of Post Frontal con- ditions might record a large number of low counts. This might happen if the number of arrivals declines over time (perhaps having been concentrated immediately behind the front), while favorable winds for departing across the Gulf continue. Numbers of mi- grants recorded early in strings of Post Frontal weather were higher (counts for days 1-3 av- eraged 0.27-0.57 standard deviations above their seasonal means) than numbers on days 4-6 of such strings (ranging from —0.13 to —0.69 standard deviations below their means). No other synoptic scenario recorded bird counts that varied with their number of days of occurrence, either when considering total seasonal bird counts for all weather types, or just counts on days with their own synoptic conditions, in spring or fall (all P > 0.15). Gulf High was not included in this portion of the analysis, because it was unrecorded in three spring and two fall periods. DISCUSSION We found differences in the number of birds present under different synoptic weather scenarios near the north coast of the Gulf of Mexico despite the increased complexities in- volved in relating weather to stopover num- bers near the edges of ecological barriers (Si- mons et al. 2004). In spring, most birds were recorded stopping over under two synoptic scenarios: one of which was associated with approaching cold fronts and the other with conditions following frontal passage. The post-frontal scenario also recorded the largest bird counts in fall. In both seasons, numbers were smaller under a weather scenario with winds from the east or off the Gulf. The results of our study are consistent with perceptions in the literature (Lowery 1945, Dennis 1954, Gauthreaux 1971, Duncan 1994) and more anecdotal accounts that largest num- bers of migrant birds stop over along the northern Gulf Coast when there is inclement weather to impede passage over or near the Gulf in spring, and in the tail winds immedi- ately following cold frontal passages in au- tumn. We provide confirmation of these pat- terns through systematic quantification of bird numbers over several seasons, which has been generally lacking previously. These patterns are generally consistent with the findings of other studies of the effects of weather on mi- gration (Lack 1960; Richardson 1978, 1990), such as increased movement during tail winds or avoidance of rain. Further studies are need- ed to compare stopover patterns in an urban setting such as ours, to a broader variety of habitats and geographical contexts. Previous authors have noted that weather conditions over the Gulf of Mexico are gen- erally favorable for northward migration in spring, but less frequently favorable for south- bound migration in fall (Able 1972, Gauth- reaux et al. 2005). This may be expected to make stopover bird counts more clearly im- pacted by weather in fall, when bird move- ments may be more tightly clustered in the fewer days of favorable migratory conditions. In contrast, migrant counts in our study dif- fered more among weather scenarios in spring than in fall. The greatest impact of weather on stopover numbers near the Gulf Coast may be in grounding migrants with inclement weather (when frontal systems intercept northbound birds in spring), rather than in stimulating mi- gration to begin (as suggested for fall). It is difficult to imagine situations in which fall mi- grants might encounter inclement weather that might cause grounding, but tropical distur- bances may be a good candidate. Gulf Tropi- cal Disturbances in our study showed mixed patterns — associated with relatively high numbers of birds overall and with some of the big days in fall, but also with low numbers of particular species. The winds and rain of a Gulf disturbance may well be capable of pro- ducing significant fallouts, but only when oth- er conditions are favorable (tail winds to bring birds into the region, etc.). Numbers of migrants in spring varied more strongly with weather conditions of the pre- vious afternoon than weather existing at the time of the field survey. This is consistent with the general understanding that spring mi- grants arrive well into the daylight hours, hav- ing still been over the Gulf when daylight ar- 294 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 rived (Gauthreaux 1971). Thus, an early morning survey will largely record birds that arrived the previous day and spent the night. The numbers of migrants in fall likewise var- ied more strongly with weather conditions the previous afternoon, presumably because this reading indicated what weather conditions (e.g., frontal passages) were established in time for the nocturnal flight. The morning of the census might reflect weather that was not yet present when the birds were migrating at night. One advantage of stopover studies relative to radar studies of migration is the ability to distinguish individual species. Three of the six species examined in fall gave no hint of weather-based differences in counts, despite an overall strong variation of the total migrant count among synoptic features. This high- lights the need for more species-specific stud- ies involving stopover observations and the perils of pooling the migratory behavior of all species together, as is necessary in radar stud- ies. One source of interspecific differences in weather-dependence could be the different mi- gration schedules of different species. Yellow Warblers showed no variation among weather patterns; the overwhelming majority of their records were in August when few cold fronts passed New Orleans. It is logical that early migrants would have migratory strategies not heavily dependent upon frontal systems, which are infrequent early in the fall. White- eyed Vireo counts suggested a lack of elevat- ed numbers in spring under Frontal Gulf Re- turn conditions, and Eastern Wood-Pewee numbers appeared to be unusually high during Gulf High conditions in fall, trends atypical of the overall migrant pool and worthy of further investigation. The likelihood of encountering threatening weather conditions at the end of a migratory flight may also be an important factor affect- ing migrant decisions (Nisbet and Drury 1968). The challenge of meeting energetic de- mands in cold or rainy weather might be ex- acerbated by the elevated numbers of migrants present in stopover habitats under such weath- er conditions, if the migrants compete with one another for food or other resources. The population-level impacts of this possibility may be ameliorated by generally higher fat reserves in migrants that stopover in inclem- ent weather (Moore and Kerlinger 1987, Si- mons et al. 2004). More study is needed on the consequences of inclement weather on the ability to continue migration in good body condition. ACKNOWLEDGMENTS The authors thank the many homeowners of the study area who tolerated our presence in front of their house on so many mornings. The manuscript benefited from the comments of F. R. Moore and two anonymous reviewers. LITERATURE CITED Able, K. 1972. Fall migration in coastal Louisiana and the evolution of migration patterns in the Gulf Re- gion. Wilson Bulletin 84:231-242. Alerstam, T. 1990. Bird migration. Cambridge Uni- versity Press, Cambridge, United Kingdom. Berthold, P. 2001. Bird migration: a general survey. Second Edition. Oxford University Press, Oxford, United Kingdom. Dennis, J. V. 1954. Meteorological analysis of occur- rence of grounded migrants at Smith Point, Texas, April 17-May 17, 1951. Wilson Bulletin 66:102- 111. Duncan, R. A. 1994. Bird migration, weather, and fall- out, including the migrant traps of Alabama and northwest Florida. Self-published, Gulf Breeze, Florida, USA. Gauthreaux Jr., S. A. 1971. A radar and direct visual study of passerine spring migration in southern Louisiana. Auk 88:343-365. Gauthreaux Jr., S. A., J. E. Michi, and C. Belser. 2005. The temporal and spatial structure of the atmosphere and its influence on bird migration strategies. Pages 182—193 in Birds of two worlds: the ecology and evolution of migration (R. Green- berg and P. P. Marra, Editors). Johns Hopkins Uni- versity Press, Baltimore, Maryland, USA. Lack, D. 1960. The influence of weather on passerine migration: a review. Auk 77:171—209. Loria, D. E. and F. R. Moore. 1990. Energy demands of migration on Red-eyed Vireos (Vireo oliva- ceous). Behavioral Ecology 1:24-35. Lowery Jr., G. H. 1945. Trans-Gulf spring migration of birds and the coastal hiatus. Wilson Bulletin 57: 92-121. Moore, F. and P. Kerlinger. 1987. Stopover and fat deposition by North American wood-warblers (Parulinae) following spring migration over the Gulf of Mexico. Oecologia 74:47-54. Muller, R. A. 1977. A synoptic climatology for en- vironmental baseline analysis: New Orleans. Jour- nal of Applied Meteorology 16:20-34. Nisbet, I. C. T. and W. H. Drury Jr. 1968. Short-term effects of weather on bird migration: a field study using multivariate statistics. Animal Behaviour 16:496-530. Yaukey and Powell • STOPOVER OF MIGRANTS AND SYNOPTIC WEATHER 295 Richardson, W. J. 1978. Timing and amount of bird migration in relation to weather: a review. Oikos 30:224-272. Richardson, W. J. 1990. Timing of bird migration in relation to weather: updated review. Pages 79-101 in Bird migration: physiology and ecophysiology (E. Gwinner, Editor). Springer- Verlag, Berlin, Germany. Simons, T. R., F. R. Moore, and S. A. Gauthreaux Jr. 2004. Mist-netting trans-Gulf migrants at coastal stopover sites: the influence of spatial and temporal variability on capture data. Studies in Avian Biology 29:135-143. U.S. Department of Commerce. 1994-2000. Daily weather maps. Climate Prediction Center, Wash- ington, D.C., USA. Zar, j. H. 1984. Biostatistical analysis. Second Edition. Prentice-Hall, Englewood Cliffs, New Jersey, USA. The Wilson Journal of Ornithology 120(2):296— 303, 2008 MASS CHANGES OF MIGRATORY LANDBIRDS DURING STOPOVERS IN A NEW YORK CITY PARK CHAD L. SEEWAGENi 23 ^ND ERIC J. SLAYTON^ ABSTRACT. — We measured rates of mass change of eight species of migratory passerines in a New York City park during three consecutive spring and autumn migrations to evaluate the quality of an urban habitat as a stopover site. We also examined seasonal differences in body condition. Linear regressions of a condition index on time of day detected significant hourly mass gain by Magnolia Warbler (Dendroica magnolia). Black- throated Blue Warbler (D. caeridescens), Ovenbird (Seiurus aurocapilla), and Northern Waterthrush {S. nove- horacensis) during spring, and Common Yellowthroat (Geothlypis trichas) during autumn. Swainson’s Thrush iCatharus ustulatus) showed significant mass loss during autumn. Significant spring mass gain rates ranged from 0.99 to 2.46% of mean body mass/hr. Common Yellowthroat gained 1 .28% of mean body mass/hr during autumn. Most species were heavier and fatter in spring than autumn. The significant mass gain rates were comparable to those in similar studies in more pristine areas. Our results suggest the urban stopover site we examined is a place where migrants can sufficiently replenish energy stores. This highlights the importance of conserving and properly managing remaining green spaces in urban areas along major migratory bird flyways. Received 1 June 2007. Accepted 17 August 2007. Nearctic-neotropical migratory passerines engage in episodes of intense exercise during nocturnal migratory flights followed by peri- ods of hyperphagia and rapid fat deposition during stopovers. Stopover habitats in which depleted energy stores can be promptly re- plenished are necessary for successful and timely migrations (Moore et al. 1995). The fate of birds during this stage of their life- cycle can limit population sizes (Sillett and Holmes 2002), and meeting the habitat re- quirements of migration has become a pri- mary component of current migratory land- bird conservation strategies (e.g., Dettmers and Rosenberg 2000). Despite the rapidly growing interest over the last 20 years in stopover ecology (Moore and Kerlinger 1987, Woodrey and Moore 1997, Wang et al. 1998, Carlisle et al. 2005), stopover site quality (Russell et al. 1994; Dunn 2000, 2001; Rimmer and McFarland 2000), and stopover site conservation (Mc- Cann et al. 1993, Moore et al. 1995, Mehlman et al. 2005), migrant use of urban habitats has received little attention. The Atlantic Coast migration routes overlap with the most urban- ized region of North America and the habitats ' Department of Ornithology, Wildlife Conservation Society, Bronx, NY 10460, USA. 2 Department of Biology, University of Western On- tario, London, ON N6A 5B7, Canada. 2 Corresponding author; e-mail; cseewagen@wcs.org remaining within cities may have an important role in landbird migration (Mehlman et al. 2005). However, it is unknown whether urban parks provide migrants with resources they need during stopovers. The high densities of birds, predominance of exotic invasive plants, and overall degraded habitat conditions often present in such areas may elevate competition for resources and constrain the ability of mi- grants to sufficiently refuel. City parks and similar habitat fragments will represent an in- creasing proportion of the stopover sites avail- able to migrants as urbanization proceeds. Thus, the stopover ecology of birds in these areas is deserving of greater attention. Our ob- jective was to measure the mass change of migrants during stopovers in Bronx Park, New York City, and gauge the quality of an urban habitat as a migratory bird stopover site. We also examined seasonal differences in body condition to learn if migrants carried greater fat loads in spring than autumn, as has often been observed in non-urban areas (e.g.. Wink- er et al. 1992b, Morris and Glasgow 2001). METHODS Site Description. — Bronx County, New York, has a human population of ~ 1 .4 million and a total land area of 109 km2, equaling a population density of 12,844 people/km2 (pj.S. Census Bureau 2004). Bronx Park is a 229-ha municipal park at the center of the county. The park is composed almost entirely of the 296 Seewagen and Slayton • URBAN STOPOVER ECOLOGY 297 campuses of the Bronx Zoo and New York Botanical Garden. The Bronx River bisects Bronx Park before joining the East River to the south. Our study area was an approxi- mately 4.9-ha section of riparian and upland forest on the grounds of the Bronx Zoo (40° 85' N, 73° 87' W). The site does not contain any animal exhibits and is not open to zoo visitors. The west side of this area is a mature dry upland deciduous forest dominated by red oak {Quercus rubra) and sweet gum (Liquid- ambar styracijiua) with some white ash {Fraxinus americana), mockernut {Cary a to- mentosa), black cherry {Prunus serotina), and American elm {Ulmus americana) also pres- ent. The upland forest transitions rapidly down a steep rocky gradient to the east into a riparian zone with wet or seasonally wet soils. The most common riparian species are wil- lows (Salix spp.) and swamp dogwood {Cor- nus foemina). Two non-native invasive plants, Japanese knotweed {Polygonum cuspidatum) and Oriental bittersweet {Celastrus orbicula- tus), are prevalent. We note that urban habitats can vary con- siderably with regard to size, vegetation com- position, extent of human use/recreation, his- torical and current management practices, abundance of birds using them as stopover sites, and other factors that may influence their quality as migratory bird stopover habi- tat. Therefore the birds captured in Bronx Park should not be considered representative of birds that use dramatically different types of habitats within cities during migration (e.g., Seewagen 2008). Data Collection. — Birds were captured in 10-14 mist nets during spring and autumn of 2004-2006. Nets were operated from sunrise until ~ 1200 hrs EST, 5 days/week (weather permitting). Autumn sampling began on 4-6 September each year and ended on 17-18 Oc- tober. Spring sampling began on 28 April-2 May and ended on 1-2 June. Nets were set in 10 locations (some nets were doubled) throughout the site and remained in place for the duration of the study. Six net locations were within close proximity to the river’s edge (<10 m), while the remaining four locations were ~40 m from the river in the upland area. Nets were checked approximately every 30 min. Captured birds were measured (tail and Linflattened wing length in mm), classified to age and gender when possible (Pyle 1997), weighed to the nearest 0.1 g (Ohaus 400 g digital balance), and banded with federal met- al bands. Visible subcutaneous fat in the fur- cular hollow was rated on a six-point scale (Moore and Kerlinger 1987). Pat was scored by the same observer throughout the study with the exception of 5 days during autumn 2004 and 2 days during autumn 2005 to min- imize inter-observer variation (Krementz and Pendleton 1990). Focal Species. — We selected eight focal species based on their sample sizes (>25 total individuals of each in spring and autumn) and a lack of large breeding or over-wintering populations in or near our study site: Swain- son’s Thrush {Catharus ustulatus). Wood Thrush {Hylocichla mustelina). Magnolia Warbler {Dendroica magnolia). Black-throat- ed Blue Warbler {D. caerulescens), Yellow- rumped Warbler (D. coronata), Ovenbird {Seiurus aurocapilla). Northern Waterthrush {S. noveboracensis), and Common Yellow- throat {Geothlypis trichas). Common Yellow- throat and Yellow-rumped Warbler are excep- tions to our criteria in that the former breeds locally and the latter often over-winters in nearby coastal areas (NYSBBA 2000; De- Candido and Allen 2005; CLS, pers. obs.). However, because sample sizes for both spe- cies were relatively large, we assumed the ma- jority of individuals that we captured were transients. Only one Common Yellowthroat was recaptured more than 1 week following initial capture during spring, and no Yellow- rumped Warblers were recaptured during au- tumn suggesting the individuals in our sam- ples did not remain in the study site to breed or over-winter. There were no between-season or between-year captures of Common Yellow- throats or Yellow-rumped Warblers, unlike in- dividuals of other species that nested, over- wintered, or resided year-round in the study site during 2()()4-2()06 (e.g.. Warbling Vireo \Vireo gilviis], American Robin \Turdus mig- ratorius]. Gray Catbird \Duftietella carolinen- .v/.v), wSong wSparrow \Melospiza melodia], and Baltimore Oriole [Icterus galbula]). Statistical Analyses. — We used Mann-Whit- ney 67-tests to compare fat scores between seasons (Hailman 1965). Medians and means are both presented, as the median is the most appropriate measure of central tendency for 298 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 fat score data but the mean is more commonly reported in other studies (Hailman 1965, Ben- son and Winker 2005). We calculated a condition index (mass X 100/wing length) to correct body mass for body size variation (Winker 1995, Dunn 2002, Bonter et al. 2007). We used r-tests to examine differences in condition indices (Carlisle et al. 2005) between seasons. We first tested the re- lationship between condition index and fat score in spring and autumn with Pearson’s product-moment correlations because differ- ences in body mass may not be entirely at- tributable to differences in fat content (Piers- ma 1990, Scott et al. 1994, Karasov and Pin- show 1998). Simple linear regressions of condition index and time of day (converted to hours since sun- rise at time of capture) were used to estimate hourly changes in condition (Morris et al. 1996; Dunn 2001, 2002; Carlisle et al. 2005; Bonter et al. 2007). Changes in condition were converted to changes in mass using the mean spring and autumn wing lengths for each spe- cies. We first regressed wing length on time of day to reveal any existing temporal biases in body size that might distort mass change estimates when using this method (Winker 1995). Statistical tests were performed with SYS- TAT, Version 10.0 (Systat Software Inc. 2000), and PRISM, Version 4.0 (GraphPad Software Inc. 2005) software packages. We considered results statistically significant when P ^ 0.05. RESULTS Capture Rates and Species Richness. — Our sampling effort totaled 2,599 net hrs in spring and 3,764 in autumn. The average spring cap- ture rate of Nearctic-neotropical species {n = 49) equaled 53.9 birds/100 net hrs; the aver- age spring capture rate of all species {n = 63) equaled 59.5 birds/ 100 net hrs. Nearctic-neo- tropical species {n = 48) were captured during autumn at an average rate of 37.3 birds/ 100 net hrs and all species {n = 70) were captured at an average rate of 49.8 birds/ 100 net hrs. Fat Score and Condition Index. — Condition indices were significantly correlated with fat scores in each species in both seasons (Pear- son’s product-moment correlation tests, all P < 0.001) suggesting that differences in con- dition indices reflected differences in lipid stores. Fat scores and condition indices were higher in spring than autumn for Black-throat- ed Blue Warbler, Common Yellowthroat, Magnolia Warbler, and Yellow-rumped War- bler (all P < 0.005; Table 1). Swainson’s Thrushes had higher fat scores in spring than autumn, but showed no seasonal difference in condition indices. Fat scores for Northern Wa- terthrush did not differ between seasons and condition indices were higher in spring than autumn. Wood Thrush was the only species for which fat scores and condition indices were higher in autumn than spring (Table 1). Rate of Mass Change. — Linear regressions of wing length on time of day were negative for Yellow-rumped Warbler during spring and autumn (spring: = 0.03, P = 0.020; au- tumn: F = 0.04, P = 0.027) and Swainson’s Thrush during autumn {F — 0.22, P = 0.01 1). No other species showed significant interac- tions between body size and time of day. The relationship between condition index and time of day was significant in six (all P < 0.05), and nearly significant in one {P = 0.067) of 16 combinations of species and sea- son. Slopes were positive in six of these seven regressions. During spring. Magnolia Warbler, Black-throated Blue Warbler, Ovenbird, and Northern Waterthrush showed significant mass gains, ranging from 0.99 to 2.46% of mean body mass/hr. Common Yellowthroat gained 1.28%/hr during autumn. Swainson’s Thrush in autumn was the only species to lose sig- nificant mass (—4.46%/hr; Table 2). Swain- son’s Thrush was the only species for which spring and autumn mass change rates differed (F, 78 = 8.80, P = 0.004). However, linear re- gressions were not significant for any species in both seasons, preventing meaningful com- parisons of mass change rates among spring and autumn. DISCUSSION Fat Score and Condition Index. — Most of the species we examined were heavier and fat- ter in spring than autumn, similar to another recent study in New York City (Seewagen 2008). These results are consistent with find- ings in non-urban areas (e.g.. King et al. 1963, Winker et al. 1992b, Morris and Glasgow 2001, Dunn 2002, but see Benson and Winker 2005). Greater fat loads during spring are be- Seewagen and Slayton • URBAN STOPOVER ECOLOGY 299 >> OJ C C O 3 p C/2 O !/3 X "2 ^ ^ - § X o 3 bH M.S O o o . .s r^xosr^xooxor^o— ^cNOs^ooo 00 ^ O 00 ^ 00 DO O (N — _ so On — ; p nO'^r-OOX'^n^opppppppp — "sdddoNfNind^rnxomooin inmr\ioo-HOOON'-’OOX'^oor';^pON O O >n p p p p p p p p p p p dddddddddddddddd ci — m— 'ooinmooxpppppppp 1 rorJddoNoodoNrn^’ddr^c^oo mr'-'iinin — „^(N|(n — — -^—' E E ox> r^r^minooON2.5 weeks. We set the 2.5 week interval because it allowed us to include eagles in the analysis when satellite data were irregular or inaccurate during onset of migration. We as- signed the arrival date by using the first day the eagle stopped directional movement and began a period of localized movements (<100 km radius) for >31 days. Eagles could have arrived up to 6 days earlier because of satellite transmission intervals. We separated spring migration data for first-year (juvenile) eagles because migration initiation of recently fledged eagles can be delayed compared to older sub-adults (Gerrard et al. 1978). There was no difference in fall migration date among age classes and data were pooled. We compared differences between the number of days traveled for first-year eagles during north and south migration by a paired comparison >test {n = 40, Sokal and Rohlf 1995). We delineated migration flyways using un- filtered satellite data to define general routes taken by first-year eagles. The small sample size of second through fifth-year eagles did not allow analysis of movements of older ea- gles. We plotted satellite locations during mi- gration in Arc View 3.2 (ESRI 1999) with an Albers Equal Area Conic projection and con- nected locations on a map of eastern North America. We categorized migration routes by proximity to the Appalachian Mountains, i.e., any movements within the mountain boundary (Sure!Maps Raster 1995) were classified as mountain and routes east of the mountains were classified as coastal. We compared the changes between mountain and coastal routes to examine route fidelity for an individual by age class ( 1 year, 2 year, 3-5 year) using a Chi-square test (Sokal and Rohlf 1995). The Mississippi River Valley flyway was not in- cluded because of low sample size. We cal- culated distance traveled during migration by adding the linear distances by locations be- tween winter and summer areas for each ea- gle. We compared migration distance between routes, by gender, and calendar year using three-way ANOVA (SAS Institute Inc. 1999). We defined a stopover site as an area where an eagle made localized movements for 1 to 4 weeks during migration using two or more locations. We recorded the closest body of wa- ter using United States and Canadian hydrol- ogy data layers in ArcView (Natural Resourc- es Canada 2001, USGS 2003). We measured the distance between stopover sites and esti- mated number of days spent at stopovers for each eagle. We averaged the distance traveled between stopovers if an eagle used multiple stopovers. We used one-way ANOVA to com- pare distance traveled between their first northbound and southbound migrations. Stop- over site locations and durations are approxi- mations because of gaps in data between scheduled satellite transmissions. RESULTS Recently fledged Bald Eagles began their first migration from April to June (T ± SE = 30 May ± 3 days, median - 24 May, n = 52). Older sub-adults began migrating north from Elorida earlier in spring, from late March to August (jc ± SE = 1 May ± 4 days, median = 30 Apr, n — 36). The return trip south be- gan in late July-late December for all age classes (jc ± SE = 17 Sep ± 3 days, median = 21 Sep, n = 43, Fig. 1). The number of days eagles traveled on the first migration was greater for southbound than for northbound (paired r-test, x difference = 1 1 days, df = 38, t = 2.2, P = 0.032). Eagles migrated north along the Coastal Plain {n = 24) and Appalachian Mountain (/? = 26) routes during their first year (Fig. 2). Two eagles used the lower Mississippi River Valley to migrate to Mississippi and Missouri. Eagles changed their migration routes less on northbound and southbound movements as they became older (y^ = 13.7, df = 2, P < 0.001). One-year-old eagles changed routes between yearly spring and fall migrations 57% of the time, 2-year-olds 30%, and 3-5- year-olds changed only 17% of the time. We found no difference for migratory dis- tance traveled by gender (E, 0.86, P = 0.36), year {Py^,= 0.25, P = 0.86) or their interaction terms. Route distances, however. 306 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 FIG. 1. Initiation dates of northbound and southbound migration for sub-adult Bald Eagles showing late departure of first-year eagles from Florida, 1997-2004. Dates calculated as means for individual Bald Eagles (n) using the number of years they survived. did differ. Coastal migrants traveled less (x = 1,397 km) than mountain migrants (x = 2,112 km, Fi 3i = 9. n, P = 0.005). One eagle spent the summer 4,146 km north of Florida in coastal Labrador, Canada (51° N) and another visited NW Labrador at 55° N. Another eagle traveled only 340 km north of Florida and spent the summer on Lake Marion, South Car- olina (33° N). We documented 25 of 54 (46%) eagles us- ing migratory stopovers. Eagles used stop- overs 1-3 times during migration, staying 1-4 weeks (x = 14.8 days, median = 14 days, 95% Cl: 12.8-16.8, n = 54, Fig. 3). All but one stopover duration was ^31 days. One fe- male spent the summer in New York and, dur- ing the first migration south, remained at a West Virginia stopover for 54 days, before continuing south to Florida. Mean distance traveled between stops on an eagle’s first mi- gration north (Y = 1,405 km; 95% Cl: 901— 1,908; n = 9) was greater than their first mi- gration south (x = 119 km; 95% Cl: 402— 1,156; n = 16; one-way ANOVA F123 = 4.23, P = 0.051). We detected 20 northbound and 26 southbound stopover sites. Only one north- bound stopover at Caledon State Park on the Potomac River was detected as a stopover during southbound migration. DISCUSSION Bald Eagles migrate in fall and spring to exploit seasonally available food, and to avoid areas where food becomes unavailable. Eagles feed primarily on exothermic fish near the sur- face that are sensitized to seasonal tempera- ture fluctuations (Gerrard and Bortolotti 1988). Eagles in Florida are thought to mi- grate to higher latitudes in summer to escape high temperatures and to forage in areas with increased prey (Wood et al. 1998). The aver- age fall and spring migration dates in this study correspond to previous reports from banding stations and telemetry studies (Broley 1947, Buehler et al. 1991, Weidensaul 2000L Wide variation in both departure and return dates suggests that migration behavior may not be attributed solely to prey abundance or air temperature. We found undescribed migratory routes of Bald Eagles along the Appalachian Mountains and Mississippi Valley. Earlier studies in Mojica et al. • BALD EAGLE MIGRATION 307 FIG. 2. Migration routes of sub-adult Bald Eagles {n = 54) hatched in Florida along the Appalachian Mountains {n = 26), Coastal Plain (n = 24), and Mississippi Valley (n = 2) in eastern North America, 1997- 2004. Return routes southbound are similar to northbound routes shown by directional arrows. Eight of 69 sub- adult eagles did not migrate from Florida. 308 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 2, June 2008 LIG. 3. Stopover sites used by sub-adult Bald Eagles (n — 25, ages 1—4) from Llorida during northbound and southbound migration in eastern North America, 1997—2004. Mojica et al. • BALD EAGLE MIGRATION 309 southwest and north central Florida indicated that eagles used coastal routes predominantly (Broley 1947, Wood and Collopy 1994); how- ever, we found equal use of coastal and moun- tain routes. Both males and females, and all sub-adult age classes used these migratory routes. We suspect that route choice may de- pend on prevailing winds at time of migration, location of natal areas, or genetic pre-pro- gramming (Broley 1947; Hunt et al. 1992; M. W. Collopy, pers. comm.). Biologists in the 1940s hypothesized that Florida eagles found in midwestern states in summer used the Mis- sissippi Valley during migration (Broley 1947). Two eagles (4%) in this study used the Mississippi Valley flyway, but only the lower portion of that flyway. One female spent two consecutive summers in southern Illinois on the Mississippi River and winters in the Flor- ida panhandle. Another female spent four summers in northeastern Alabama and south- ern Tennessee, and winters in west-central Florida. In addition, a male migrated north along the Appalachian Mountains before mov- ing west to spend >1 summer on Lake Erie. The Coastal Plain, Appalachian Mountain, and Mississippi River migratory routes we de- scribe may be specific to Florida Bald Eagles. An increase in route fidelity as eagles age sug- gests that young eagles learn to migrate effi- ciently. As an eagle gains more experience, it may seek the same stopover sites on a partic- ular route. We recorded migration distances >1,000 km farther than those reported in previous studies of Bald Eagles (range: 1,450-3,032 km; Grubb et al. 1994, McClelland et al. 1994, Wood and Collopy 1994). We attribute this to the ability to follow PTT-equipped ea- gles into remote areas of Canada >4,000 km from Florida. Two eagles, however, stopped their northern migration in Georgia and South Carolina, while the remaining eagles traveled several thousand additional kilometers. Some of the difference in distance traveled can be explained by migration routes, because coastal migrants traveled less distance overall than mountain migrants. We hypothesize that mi- grants using the mountain route may be able to travel farther by soaring in and gliding from abundant updrafts that are not found in the Coastal Plain (Heintzelman 1975, Hunt et al. 1992). In addition, the Chesapeake Bay region may serve to short-stop coastal migrants with abundant food resources and concentrations of other eagles. Stopover sites may be important areas along migration routes where most eagles spend 1-4 weeks replenishing energy re- serves. The estimated time spent at stopovers in this study was consistent with previous studies of hatching year sub-adults {x= 14.5, range 1-25 days; Restani 2000, Laing et al. 2005). The small percentage of overlap be- tween north and south stopover sites may in- dicate variation in migration resulting from seasonal weather patterns or prey availability. Approximately half of the eagles were unde- tected using stopovers, but this could be caused by our sampling interval of one trans- mission/week. Eagles may have used stop- overs <6 days and therefore were undetected. Distances traveled between stopover sites were similar on northbound and southbound migration, but low sample size may have re- duced the power of the test. Our results provide substantial information on Bald Eagle migratory timing, routes, and stopovers not described previously from band- ing and VHF telemetry studies. Detailed knowledge of eagle migration is important for managing this wide-ranging species. Eagles may benefit from a national or international management plan rather than current regional eagle plans. Conserving migratory stopover habitat should contribute to the continued health of the Bald Eagle population in Florida. ACKNOWLEDGMENTS We thank the following funding partners for making this research possible: Florida Nongame Wildlife Trust Fund; U.S. Fish and Wildlife Service, Atlanta, Georgia (Section 6 of the Endangered Species Act); USGS, Pa- tuxent Wildlife Research Center; The University of Georgia, Warnell School of Forestry and Natural Re- sources; and Georgia Ornithological Society. T F. Breen, Anthony Steffer, L. M. Phillips, N. J. Douglass, and S. K. Taylor contributed significantly to the initial phase of this study providing experti.se in the field and office. G. W. Barrett, S. B. Castleberry, M. W. Collopy, Marco Restani, R. J. Warren, and two anonymous re- viewers provided comments and suggestions that im- proved the manuscript. Movement, foraging and roost- ing locations, and stopover data are accessible on a web site (http://www.myfwc.com/eagle/) for managers to use in con.serving Bald Fiagle habitat in eastern North America. 310 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 LITERATURE CITED Broley, C. L. 1947. Migration and nesting of Florida Bald Eagles. Wilson Bulletin 59:3—20. Buehler, D. a., T. J. Mersmann, J. D. Fraser, and J. K. D. Seegar. 1991. Differences in distribution of breeding, non-breeding, and migrant Bald Eagles on the northern Chesapeake Bay. Condor 93:399- 408. ESRI. 1999. Arc View 3.2. Environmental Systems Re- search Institute, Redlands, California, USA. Gerrard, j. M. and G. R. Bortolotti. 1988. The Bald Eagle: haunts and habits of a wilderness monarch. Smithsonian Institution Press, Washing- ton, D.C., USA. Gerrard, J. M., D. W. A. Whitfield, P. Gerrard, P. N. Gerrard, and W. J. Maher. 1978. Migratory movements and plumage of subadult Saskatche- wan Bald Eagles. Canadian Field-Naturalist 92: 375-382. Grubb, T. G., W. W. Bowerman, and P. H. Howey. 1994. Tracking local and seasonal movements of wintering Bald Eagles Haliaeetus leucocephalus from Arizona and Michigan with satellite telem- etry. Pages 347-358 in Proceedings of the IV World Conference on Birds of Prey and Owls (B.-U. Meyburg and R. D. Chancellor, Editors). Pica Press, Berlin, Germany. Harmata, a. R. 2002. Vernal migration of Bald Ea- gles from a southern Colorado wintering area. Journal of Raptor Research 36:256-264. Harmata, A. R., J. E. Toepfer, and J. M. Gerrard. 1985. Fall migration of Bald Eagles produced in northern Saskatchewan. Blue Jay 43:232-237. Heintzelman, D. S. 1975. Autumn hawk flights: the migrations in eastern North America. Rutgers University Press, Rutgers, New Jersey, USA. Hunt, W. G., R. E. Jackman, J. M. Jenkins, C. G. Thelander, and R. N. Lehman. 1992. Northward post-fledging migration of California Bald Eagles. Journal of Raptor Research 26:19-23. Laing, D. K., D. M. Bird, and T. E. Chubbs. 2005. First complete migration cycles for juvenile Bald Eagles {Haliaeetus leucocephalus) from Labrador. Journal of Raptor Research 39:11-18. McClelland, B. R., L. S. Young, P. T. McClelland, J. G. Crenshaw, H. L. Allen, and D. S. Shea. 1994. Migration ecology of Bald Eagles from au- tumn concentrations in Glacier National Park, Montana. Wildlife Monographs 125. Millsap, B., T. Breen, E. McConnell, T. Steeper, L. Phillips, N. Douglass, and S. Taylor. 2004. Comparative fecundity and survival of Bald Ea- gles fledged from suburban and rural natal areas in Florida. Journal of Wildlife Management 68: 1018-1031. Natural Resources Canada. 2001. National topo- graphic database: hydrography features. Centre for Topographic Information, Natural Resources Can- ada, Sherbrooke, Quebec, Canada. Restani, M. 2000. Age-specific stopover behavior of migrant Bald Eagles. Wilson Bulletin 112:28-34. SAS Institute Inc. 1999. SAS/STAT User’s Guide. Version 8. SAS Institute Inc., Cary, North Caro- lina, USA. SOKAL, R. R. AND E J. Rohlf. 1995. Biometry. W. H. Freeman and Company, New York, USA. Steenhof, K., M. R. Fuller, M. N. Kochert, and K. K. Bates. 2005. Long-range movements and breeding dispersal of Prairie Falcons from south- west Idaho. Condor 107:481-496. SurelMaps Raster. 1995. Eastern U.S.: NH, VT, NY, MA, PA, RI, CT, NJ, DE, MD, DC, VA, WV, NC, SC, GA, EL, ME. Horizons Technologies Inc., San Diego, California, USA. U.S. Geological Survey (USGS). 2003. U.S. Nation- al Atlas water feature areas: aqueducts, canals, dams, intercoastal waterways, rivers, and streams. Reston, Virginia, USA. Weidensaul, S. 2000. The raptor almanac: a compre- hensive guide to eagles, hawks, falcons, and vul- tures. Lyons Press, New York, USA. Wood, P. B. and M. W. Collopy. 1994. Population ecology of sub-adult southern Bald Eagles in Flor- ida: post-fledging ecology, migration patterns, habitat use, and survival. Final Report. Nongame Wildlife Program Project NG87-026. Florida Game and Fresh Water Fish Commission, Talla- hassee, USA. Wood, P. B., M. W. Collopy, and C. M. Sekerak. 1998. Postfledging nest dependence period for Bald Eagles in Florida. Journal of Wildlife Man- agement 62:333—339. The Wilson Journal of Ornithology 120(2):3 1 1-3 19, 2008 WETLAND FEATURES THAT INFLUENCE OCCUPANCY BY THE ENDANGERED HAWAIIAN DUCK KIMBERLY J. UYEHARA,''* ANDREW ENGILIS JR.,^ AND BRUCE D. DUGGERS ABSTRACT. — Habitat loss, introduced predators, and hybridization with feral Mallards {Anas platyrhynchos) continue to threaten the existence of the endangered Hawaiian Duck or Koloa maoli (A. wyvilliana). Protection and management of core breeding areas is a recovery objective, but lack of quantitative information on the species’ habitat needs hinders recovery efforts. We conducted bi-monthly surveys of 48 wetlands on private lands on the Island of Hawaii from March 2002 to June 2003. We compared Koloa use between seasons, wetland types, and study regions and modeled how use varied with 14 site and landscape variables. Koloa occupied 54% of the surveyed wetlands; use was higher on wetlands enhanced or created for Koloa primarily through the USD As Wetlands Reserve Program (WRP) than on ponds created for agriculture (81 vs. 41%) and on wetlands in the Kohala than in the Mauna Kea region (93 vs. 38%). Koloa were more likely to use wetlands farther (>600 m) from houses, larger (>0.23 ha) wetlands, and those surrounded by more wetlands area (>1 ha). Our results (1) indicate WRP wetlands provide suitable habitat and (2) support wetlands enhancement or creation far from human disturbance. Habitat improvements combined with feral Mallard control may reduce extinction threats to Koloa. Received 11 December 2006. Accepted 16 July 2007. The Hawaiian Duck or Koloa maoli {Anas wyvilliana), hereafter referred to as Koloa, is a monochromatic, non-migratory, endangered species allied with the North American Mal- lard {A. platyrhynchos) complex (Browne et al. 1993). With an approximate population of 2,200 birds (Engilis and Pratt 1993, Engilis et al. 2002), the Koloa is the only endemic duck species to remain in the main Hawaiian Is- lands of more than 1 1 members of Anatidae reported in the fossil record (Olson and James 1991, Burney et al. 2001). The breeding sea- son is year-round with peaks in December to May on Kaua‘i (Swedberg 1967) and April to June on Hawai‘1 (Giffin 1983). Koloa use di- verse habitats from sea level to 3,000 m in- cluding palustrine emergent and riverine (Swedberg 1967, Giffin 1983, Engilis and Pratt 1993), but rarely estuarine (Engilis et al. 2002). Threats to Koloa persistence include depredation by introduced predators, habitat loss, and hybridization with feral Mallards (USDI 2005). Genetically-pure Koloa popu- lations were believed to occur on the islands of Kaua‘i (Browne et al. 1993, Rhymer 2001 ), '73-1270 Awakea Street, Kailua-Kona, HI 96740, USA. 2 Department of Wildlife, Fi.sh, and Conservation Biology, University of California, Davis, CA 95616, USA. ’ Department of Fisheries and Wildlife, Oregon State University, Corvallis, OR 97331, USA. ‘‘Corresponding author; e-mail: kjukem@lava.net Ni‘ihau, and highlands of Hawai‘i with hybrid swarms on 0‘ahu and Maui (Engilis and Pratt 1993), but there is now evidence of hybrid- ization within pure populations (Engilis et al. 2002). Habitat management of core breeding areas is a primary recovery objective for Koloa (USDI 2005). However, little is known about the basic ecology and habitat requirements of Koloa (Rhymer 2001, Engilis et al. 2002); thus, little specific information is available to guide habitat restoration. Wetlands were en- hanced or created between 1999 and 2002 to improve habitat conditions on agricultural lands on the Island of Hawai‘1 through the USDA Natural Resources Conservation Ser- vice’s Wetland Reserve Program (WRP) and other incentive programs (Ducks Unlimited’s private lands initiative. North American Wet- lands Conservation Act, and Partners for Fish and Wildlife). Wetland designs were based on limited field observations and earlier studies that indicated Koloa use should be higher in wetlands: (1) adjacent to other wetlands used by Koloa, (2) distant from human disturbance, (3) adjacent to streams, (4) containing an ap- proximately 50:50 ratio of emergent plants to open water, (5) with an irregular shoreline, (6) free of introduced aquatic vertebrates, and (7) with controlled livestock grazing (Schwartz and Schwartz 1953; Swedberg 1967; Giffin 1983; K. A. Goebel, pers. comm.). These fea- tures are also considered important for North 312 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 American dabbling ducks (Fredrickson and Dugger 1993, Weller 1999). No previous work has attempted to test the relative importance of these criteria on Koloa use or identify other criteria that might influence habitat suitability. This study was designed to: (1) characterize patterns of wetland use by Koloa, (2) compare characteristics of used and unused wetlands, and (3) examine how site and landscape char- acteristics influence wetland use by Koloa to guide future habitat enhancement projects. METHODS Study Area.— Out study was conducted in the Kohala-Mauna Kea region (20° 00' N, 155° 25' W) of Hawaii, the largest and youn- gest island of the Hawaiian archipelago. The area studied was on the windward or northeast slopes of the Kohala and Mauna Kea moun- tains where the majority of mid-elevation de- pressional wetlands, agricultural ponds, and perennial streams occur because older volca- nic substrates prevent percolation. A popula- tion of —200 Koloa was re-established here by the State of Hawai‘i captive propagation and release program (1958—80) after near ex- tirpation of Koloa in the 1950s (Schwartz and Schwartz 1953, Giffin 1983, Engilis and Pratt 1993). Field Methods. — We surveyed all known wetlands on two properties each in the Kohala and Mauna Kea regions consisting of 14 and 34 wetlands, respectively {n = 48) between 305 and 1,219 m in elevation (Fig. 1). The study sites included 16 WRP wetlands (6 in Kohala and 10 in Mauna Kea) and 32 agri- cultural ponds (8 in Kohala and 24 in Mauna Kea) which were stock ponds and small res- ervoirs constructed between 1900 and 2001. Wetland vegetation consisted primarily of Schoenoplectus spp. and Myriophyllum aqua- ticum. Upland vegetation consisted of live- stock forage grasses (e.g., Urochloa mutica, Pennisetum clandestinum) and small patches or stands of trees (e.g., Psidium spp., Metro- sideros polymorpha). Land uses included cat- tle ranching and macadamia nut, coffee, and koa {Acacia koa) farming. We measured 14 characteristics for each wetland (42 sites characterized in Mar-Apr 2002; 6 in Sep 2002; Table 1). We used a Global Positioning System or digital U.S. Geological Survey topographic quadrangles to calculate wetland size, shoreline development (Lind 1974), distance to house, distance to stream, surrounding wetlands, stream length, and building density. We quantified emergent and aquatic plant cover using the point-inter- cept method (Higgins et al. 1996) with points taken on a 5 X 1 m, 5X5 m, or 10 X 10 m grid for wetlands <0.10, 0.10—0.30, or >0.30 ha, respectively. Data on the presence or ab- sence of non-native waterfowl, fishes, bull- frogs {Rana catesbeiana), and livestock graz- ing were gathered during initial wetland char- acterization visits and subsequent bird sur- veys. We classified livestock grazing by the length of the grazing period: continuous (live- stock were free ranging and had year-round access to wetlands), seasonal (wetlands were fenced and livestock had access during part of the year; e.g., non-breeding season), or no grazing. We surveyed each wetland for Koloa at least twice per month from 1 March 2002 to 25 June 2003 (6 wetlands were added in Sep 2002), which included two peak nesting pe- riods (Apr-Jun; Giffin 1983). Surveys were conducted during mornings (sunrise until 4 hrs after sunrise) and afternoons (4 hrs before sunset to sunset). We established 1-2 survey points at each wetland at spots that minimized disturbance and maximized visibility of water birds. We observed wetlands for 5-10 min with 10 X 42 binoculars and a 25-56 X spot- ting scope from each point. We recorded the number of Koloa present on each wetland for each survey. We recorded direction of move- ment to minimize duplicate counts if birds flew from an occupied wetland. We walked the wetland shoreline if no Koloa were ob- served from our survey points. We assumed all ducks matching morphological and behav- ioral descriptions in Engilis et al. (2002) were Koloa. Mallard X Koloa hybrid phenotypes have been recorded on the Island of Hawai‘i (Engilis et al. 2002), but none was recorded at our study sites (1 Mallard male and 1 hy- brid-like female observed in May 2003 in lowlands outside our survey area; KJU, pers. obs.). There were no field methods to distin- guish Koloa from hybrids and few data on the prevalence of hybrids. Statistical Analyses. — Birds on a wetland were unlikely to be undetected during a sur- vey given the small size (0.01—1.3 ha) and Uyehara et al. • WETLAND FEATURES OF HAWAIIAN DUCK 313 FIG. !. Study sites (black dots) in the Kohala-Mauna Kea region of HawaiM where wetlands were surveyed for presence of Koloa relative to the known range of Koloa on Hawai‘i (gray line). sparse vegetation of most wetlands, and we assumed a detection probability of 1.0. We as- sumed that birds were not concealed in veg- etation or flushed without detection. A wet- land was defined as “occupied” if Koloa were observed at least once during a survey. Mul- tiple surveys (19-35 visits per wetland) were used to identify the status of a wetland and minimize assigning a wetland to the wrong category. Bias from undetected birds could re- sult in conservative estimates of occupancy. However, we strived to identify wetlands with regular diurnal use and a wetland with an un- detected bird would likely represent extremely low or nocturnal use, which was beyond the scope of our study. Data were untransformed. 314 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 TABLE 1. Site and landscape characteristics measured for 48 study wetlands surveyed for Koloa on the Island of Hawai‘i during 2002. Variable Description Site Wetland size Wetland vegetation structure (PC A of three variables) i. Open water ii. Emergent iii. Aquatic Invasive species (presence of any one of the following) i. Non-native waterfowl ii. Fishes iii. Bullfrogs Grazing regime Shoreline development Landscape Distance to house Distance to stream Surrounding wetlands Stream length Building density Estimated wetland area using GPS and GIS (ha) Percent area without vascular vegetation Percent area with emergent vegetation (e.g., Cyperus, Schoenoplectus) Percent area with submergent or floating vegetation (e.g., Myriophyl- lum, Nymphaea) Ornamental or barnyard waterfowl in a domestic or semi-feral state (e.g., domestic Mallard breeds, such as Khaki Campbell, Pekin) Cyprinids, poeciliids, or cichlids Rana catesbeiana Length of cattle or sheep grazing period (none, seasonal, continuous) Measurement of shoreline regularity; ratio of the shoreline length to the circumference of a circle with the same area as the wetland (1 = perfect circle) Linear distance (m) from center of study wetland to nearest occupied house Linear distance (m) from center of study wetland to nearest intermit- tent or perennial stream Percent wetland area within a 1-km radius from center of study wet- land Intermittent or perennial stream km within a 1-km radius from center of study wetland (m/ha) Number of houses and other structures within a 1-km radius from center of study wetland (n/ha) We calculated occupancy rate as the propor- tion of wetlands occupied at least once by Ko- loa. We calculated percent occurrence as the number of surveys Koloa were present on a wetland divided by the total number of sur- veys. We assigned each survey to dry (May- Oct) or wet (Nov— Apr) and breeding (Jan— Jun) or nonbreeding (Jul-Dec) seasons. We used a likelihood ratio Chi-square test to com- pare occupancy rates and the Wilcoxon signed-rank test to compare percent occur- rences between wetland categories (e.g., WRP vs. agricultural site). We used logistic regression (Hosmer and Lemeshow 2000; SAS Institute 2001, 2005) in conjunction with an information-theoretic approach (Burnham and Anderson 2002) to examine habitat characteristics most closely associated with Koloa use. We examined bi- variate relationships between explanatory var- iables using Spearman’s rank correlation co- efficient prior to modeling. Distance to house and building density (g — —0.88, P < 0.0001), distance to stream and stream length (r^ = -0.74, P < 0.0001), and surrounding wetlands and stream length (r, = 0.72, P < 0.0001) were strongly correlated. Thus, we excluded one variable from each of the three pairs and retained distance to house, distance to stream, and surrounding wetlands. Mea- sures of open water and emergent vegetation were correlated (r^ = —0.82, P < 0.0001); therefore, we used the first principal compo- nent (66% of total variance) from a PCA of these two variables and percent area with aquatic vegetation to represent vegetation structure in the wetland (eigenvectors = -0.71 for open water, 0.51 for emergent, and 0.48 for aquatic). We combined non-native waterfowl, fishes, and bullfrogs into one var- iable (invasive species) because introduced vertebrates may decrease habitat quality for Uyehara et al. • WETLAND FEATURES OF HAWAIIAN DUCK 315 TABLE 2. Logistic regression identifying factors that influence occupancy of wetlands by Koloa on the Island of Hawai‘i, 2002-2003, in order of increasing AIQ value and decreasing Akaike weight (w,). All models within 10 AIC values of the best model are presented. Model -logiU)^ Kb AICc AAICc H'y Distance to house, Surrounding wetlands 21.787 3 50.119 0.00 0.338 Distance to house. Wetland size 22.248 3 51.041 0.92 0.213 Distance to house 23.627 2 51.521 1.40 0.168 Distance to house. Grazing regime 22.968 3 52.481 2.36 0.104 Distance to house. Invasive species 23.323 3 53.191 3.07 0.073 Distance to house. Distance to stream 23.591 3 53.182 3.61 0.056 Wetland size. Invasive species 23.970 3 54.485 4.37 0.038 Surrounding wetlands. Invasive species 26.120 3 58.785 8.67 0.004 Intercept only 33.104 1 68.295 18.18 0.000 ® —log-likelihood. *’ Number of variables. waterfowl through competition for resources (Weller 1999). Our selection of explanatory variables was a priori, but we could not reasonably develop a subset of multivariable models to compare. Thus, we examined all combinations of two- variable models {a posteriori) using variables from the first analysis that performed as well or better than the null model. We evaluated the relative strength of one- and two-variable models using AAIQ and model weights (w,), and the predictive ability of the best approx- imating models using concordance values. We considered models within 2.0 AIC values of the best model as competitive (Burnham and Anderson 2002). We evaluated parameter es- timates within each model by investigating if confidence limits around the mean included zero and summing model weights for all mod- els containing that variable (i.e., variable im- portance weight; Burnham and Anderson 2002). We used regression for variables as- sociated with occupancy status to compare percent occurrence against the explanatory variable and t-tests to evaluate how slope es- timates differed from zero, restricting our analysis to wetlands occupied by Koloa. Means ± SE and 95% confidence intervals are reported. RESULTS The wetlands had an occupancy rate of 54% and a mean percent occurrence of 9 ± 12% (range 0-46%) for Koloa. WRP wetlands had a higher occupancy rate (81%, n = 16) than agricultural ponds (41%, ti = 32; = 6.21, df = 1, P = 0.01) and a higher percent oc- currence (13 ± 3%) than agricultural ponds (7 ± 2%; Z - 2.23, P < 0.03). Wetlands at Ko- hala had both a higher occupancy rate (93 vs. 38%; x" = 13.77, df = 1, P < 0.001) and percent occurrence (20 ± 3% vs. 4 ± 1%; Z = 4.13, P < 0.001) than wetlands at Mauna Kea. Occupancy rates did not differ between wet versus dry season (P = 1 .00) or breeding versus nonbreeding season (P = 0.09). How- ever, Koloa had a tendency to occupy more wetlands during the breeding (46%, n = 22) than the nonbreeding (29%, n = \A) season. We combined data for all seasons when mod- eling the influence of habitat features on wet- land use, and used all surveys of each wetland when calculating occupancy status. Six variables performed as well or better than the null model including distance to house, wetland size, surrounding wetlands, grazing regime, invasive species, and distance to stream. The best approximating model, us- ing these six variables in the multivariate anal- ysis, was distance to house and surrounding wetlands (w, = 0.338; c = 84.4) which re- ceived 1.6 times more support than the next best model (Table 2). Distance to house oc- curred in all six models and received the high- est variable importance weight (0.952), fol- lowed in decreasing order by simounding wetlands (0.344), wetland size (0.254), inva- sive species (0.115), grazing regime (0.105), and distance to stream (0.056). Only two mod- els ranked above the single-variable model of distance to house, and the confidence intervals for parameter estimates included zero for all explanatory variables except distance to house. 316 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 Koloa used wetlands farther from houses (2,500 ± 261 m for occupied vs. 855 ± 209 m for unoccupied wetlands), larger wetlands (0.23 ± 0.06 ha for occupied vs. 0.06 ± 0.02 ha for unoccupied wetlands), and those sur- rounded by more wetland habitat (0.32 ± 0.04% occupied vs. 0.19 ± 0.03% for unoc- cupied). No wetland within 600 m of a house was used by Koloa suggesting human distur- bance influenced use, but these wetlands (n = 12) also contained invasive species. All wet- lands >600 m from houses that had invasive species {n = 8) were also used by Koloa. We discriminated between the influence of inva- sive species versus distance to house by con- trolling for distance to house, rerunning all six single-variable models using only wetlands >600 m from a house (n = 36). The only competitive model was wetland size (w, = 0.952) and the confidence limits around the parameter estimate did not include zero. Per- cent occurrence for occupied wetlands (n = 26) was independent of surrounding wetlands {P = 0.23), increased with distance to house = 3.04, P < 0.006; Fig. 2), and was not linearly related to wetland size {P = 0.16). However, percent occurrence increased line- arly with wetland size (t = 4.20, P < 0.0004; Fig. 2) when we restricted our analysis to wet- lands <0.50 ha (n = 24) because the distri- bution of wetland size was inadequate to test for a relationship across the full range of sizes (only 2 of 26 wetlands were >0.50 ha and both were outliers being >10 SEs above the mean). Wetland characteristics associated with Ko- loa use varied by wetland type and study re- gion (Table 3). WRP wetlands were similar to agricultural ponds in size and the amount of surrounding wetlands, but averaged more than twice as far from houses. Wetlands at Kohala were on average almost three times farther from houses, larger, and in landscapes with more wetlands than those at Mauna Kea. DISCUSSION Koloa occupancy of WRP wetlands was high and our estimate of 81% is likely con- servative since we also saw sign (feathers, tracks, droppings) at wetlands where birds were not observed. Koloa responded to new WRP wetlands within days of completion at Kohala and 3 months at Mauna Kea (KJU, pers. obs.). Higher Koloa use on WRP versus agricultural ponds indicates that WRP wet- lands provided suitable Koloa habitat. Greater wetland remoteness and area contributed to higher Koloa use at Kohala than at Mauna Kea (Table 3). The low percent occurrence at study sites likely reflects the small Koloa pop- ulation on Hawai‘i. Population size may be constrained by a lack of sufficient wetlands habitat, but also may be a consequence of oth- er limiting factors. Koloa have been described as sensitive to disturbance (Swedberg 1967, Giffin 1983, Chang 1990, Engilis et al. 2002), consistent with our characterization. Koloa did not use any wetland within 600 m of a house. Dis- tance to houses beyond 600 m did not influ- ence wetlands occupancy; however, percent occurrence increased as distance increased. Distance to house was associated with other anthropogenic factors that may have discour- aged use by Koloa. However, when we con- trolled for distance to house, wetland occu- pancy was not associated with invasive spe- cies or grazing regime suggesting human dis- turbance was most important. Koloa in other areas (e.g., river valleys on Kaua‘i) may ha- bituate to human activity. However, this is more common in landscapes characterized by larger wetland area intermixed with agricul- ture and dwellings (Engilis et al. 2002; KJU, pers. obs.). Our study wetlands were small (0.15 ± 0.04 ha, range 0.01-1.3 ha), kettle- shaped (shoreline development: 1.2 ± 0.02, range 1.0-1. 7), and most were recently con- structed or grazed and contained sparse emer- gent vegetation. This combination of features may make waterfowl more sensitive to distur- bance than in larger wetlands, which generally provide greater escape cover that may buffer against human disturbance (Diefenbach and Owen 1989; KJU, pers. obs.) and have higher resource diversity (Brown and Dinsmore 1986, Weller 1999). Koloa occurred in wet- lands as small as 141 m^ indicating they will use small wetlands. However, both occupancy rate and percent occurrence increased with wetland size indicating larger wetlands up to 0.5 ha were used more consistently. WRP sites were constructed near agricul- tural ponds already used by Koloa under the principle that a wetland complex is more at- tractive and productive than a similar but iso- Uyehara et al • WETLAND FEATURES OF HAWAIIAN DUCK 317 Distance to house (m) Wetland size (ha) FIG. 2. Relationship between the percent occurrence of Koloa versus distance to house (A) and wetland size (B). lated wetland (Forman 1995, Weller 1999). Koloa occupancy increased with wetland abundance supporting the concept of devel- oping wetland complexes. This model re- ceived limited support, but Koloa did not abandon existing agricultural ponds following completion of WRP wetlands. Koloa were regularly observed moving among agricultur- al, stream, and new WRP habitats (KJU, pers. obs.). Occupancy was not associated with shore- line development or vegetation structure as predicted because basin topography and veg- etation communities of study sites were sim- ilar. Our measures for livestock grazing and invasive species were relatively crude, and a 318 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 TABLE 3. Wetland characteristics by wetland type (WRP vs. Kea) on the Island of Hawaii, 2002-2003. agricultural) and study area (Kohala vs. Mauna Distance to house (m) Surrounding wetlands (%) Wetland size (ha) Wetland type WRP {n = 16) 2,686 ± 309 Agricultural (n = 32) 1,276 ± 230 0.23 ± 0.04 0.28 ± 0.03 0.11 ± 0.02 0.18 ± 0.05 Study area Kohala {n = 14) 3,202 ± 1,317 Mauna Kea (n = 34) 1,147 ± 990 0.48 ± 0.19 0.18 ± 0.09 0.39 ± 0.36 0.05 ± 0.04 more focused study may reveal a stronger re- sponse by Koloa. Alternately, these variables may influence factors such as duckling sur- vival instead of adult occupancy. We predicted juxtaposition of wetlands to streams may be important as, historically, high densities of Koloa were reported in streams (Schwartz and Schwartz 1953, Swedberg 1967, Giffin 1983). Koloa use did not vary by distance to stream, but our study sites were all within 1 km of a stream which was probably too small a scale to show variation in duck use. Thus, addition- al research is needed at a larger scale to eval- uate if proximity to streams affects use. Our models are descriptive, designed to provide baseline data about wetland charac- teristics that relate to Koloa use, and guide habitat restoration and management strategies for depressional wetlands on Hawai‘i. Our findings indicate that managers should en- hance or create wetlands >600 m from human dwellings, >0.23 ha in size, and in areas with >1 ha of surrounding wetland area within a 1-km radius. Beyond those minima, larger wetlands (up to 0.5 ha) and those farther from human disturbance received more frequent use (Fig. 2). These conditions, however, may be difficult to achieve in the relatively steep to- pography of mid-elevation Hawai‘i; wetlands remote from human activity can present chal- lenges for wetland managers. Agricultural and small wetlands at Mauna Kea received less use by Koloa, but these wet- lands may provide functional habitat, partic- ularly when part of a wetland complex (Fred- rickson and Dugger 1993, Semlitsch and Bod- ie 1998, Naugle et al. 2000, Engilis et al. 2002). Information on the daily and seasonal movements of Koloa is unknown, as is the optimal dispersion of wetlands. Tracking stud- ies may be used to reveal use patterns and test detection probability assumptions. Studies that identify factors currently limiting Koloa pop- ulation size are also needed; this information would allow for more effective conservation strategies and efficient use of conservation dollars. Private landowners, through WRP and other incentives, should be encouraged to en- hance wetlands to increase habitat availability for Koloa. However, efforts to increase wet- lands abundance will benefit not only Koloa, but feral Mallards and Mallard X Koloa hy- brids, as evidenced by the rising population trends of Mallards and hybrids on some bird reserves on 0‘ahu and Maui, and recent sight- ings of hybrid phenotypes on Hawai‘i (USDI 2005). Thus, it is essential that habitat im- provements proceed in parallel with efforts to identify and remove Mallards and hybrids. Fi- nally, occupancy is only one measure of hab- itat suitability. Information on how birds use wetlands would provide a better understand- ing of how wetlands are functioning to meet- ing life history needs of Koloa at Kohala and Mauna Kea. ACKNOWLEDGMENTS Financial and technical assistance was provided by Ducks Unlimited Inc., Delta Waterfowl Foundation, Bard College, USDA Natural Resources Conservation Service, USDI Fish and Wildlife Service, and the Ha- waii State Department of Land and Natural Resourc- es. We are grateful to J. W. and M. T. Bailey, Bennett Dorrance Jr., Thuy Fujimoto, M. S. Gomes, E. L. and H. T. Gunther, and D. M. Matsuura for providing ac- cess to their land and logistical support. We thank K. M. Dugger, T. A. Erickson, K. A. Goebel, J. M. Hi- gashino, Erik Kiviat, A. P. Marshall, and K. E. Moore for technical support and review of drafts of the man- uscript. This publication was improved by the com- ments of M. H. Reynolds and an anonymous peer re- viewer. Uyehara et al. • WETLAND FEATURES OF HAWAIIAN DUCK 319 LITERATURE CITED Brown, M. and J. J. Dinsmore. 1986. Implications of marsh size and isolation for marsh bird manage- ment. Journal of Wildlife Management 50:392- 397. Browne, R. A., C. R. Grieein, R R. Chang, M. Hub- ley, AND A. E. Martin. 1993. Genetic divergence among populations of the Hawaiian Duck, Laysan Duck, and Mallard. Auk 110:49-56. Burney, D. A., H. F. James, L. P. Burney, S. L. Olson, W. Kikuchi, W. L. Wagner, M. Burney, D. Mc- Closkey, D. Kikuchi, F. V. Grady, R. Gage II, AND R. Nishek. 2001. Fossil evidence for a di- verse biota from Kaua‘i and its transformation since human arrival. Ecological Monographs 71: 615-641. Burnham, K. P. and D. R. Anderson. 2002. Model selection and multimodel inference: a practical in- formation-theoretic approach. Second Edition. Springer- Verlag, New York, USA. Chang, P. R. 1990. Strategies for managing endan- gered waterbirds on Hawaiian National Wildlife Refuges. Thesis. University of Massachusetts, Amherst, USA. Dieeenbach, D. R. and R. B. Owen Jr. 1989. A model of habitat use by breeding American Black Ducks. Journal of Wildlife Management 53:383-389. Engilis Jr., A. and T. K. Pratt. 1993. Status and pop- ulation trends of Hawaii’s native waterbirds, 1977-1987. Wilson Bulletin 105:142-158. Engilis Jr., A., K. J. Uyehara, and J. G. Gieein. 2002. Hawaiian Duck (Anas wyvilliana). The birds of North America. Number 694. Forman, R. T. T. 1995. Some general principles of landscape and regional ecology. Landscape Ecol- ogy 10:133-142. Fredrickson, L. H. and B. D. Dugger. 1993. Man- agement of wetlands at high altitudes in the Southwest. Technical Report. USDA, Forest Ser- vice, Albuquerque, New Mexico, USA. Gieein, J. G. 1983. (1) Abundance and distribution of Koloa on the Island of Hawaii. (2) Movements, survival, reproductive success and habitat of Ko- loa on the Island of Hawaii. Pittman-Robertson Projects W-18-R-7 and W-18-R-8, Job R-lll-H. Technical Report. Hawaii Division of Fish and Game, Honolulu, USA. Higgins, K. E, J. L. Oldemeyer, K. J. Jenkins, G. K. Clambey, and R. E Harlow. 1996. Vegetation sampling and measurement. Pages 567-591 in Re- search and management techniques for wildlife and habitats (T. A. Bookhout, Editor). Fifth Edi- tion. The Wildlife Society, Bethesda, Maryland, USA. Hosmer, D. W. and S. Lemeshow. 2000. Applied lo- gistic regression. Second Edition. John Wiley and Sons, New York, USA. Lind, O. T. 1974. Handbook of common methods in limnology. C. V. Mosby Company, St. Louis, Mis- souri, USA. Naugle, D. E., K. F. Higgins, M. E. Estey, R. R. Johnson, and S. M. Nusser. 2000. Local and landscape-level factors influencing Black Tern habitat suitability. Journal of Wildlife Manage- ment 64:253-260. Olson, S. L. and H. E James. 1991. Descriptions of thirty-two new species of birds from the Hawaiian Islands. Part I. Non-Passeriformes. Ornithological Monographs 45:1-88. Rhymer, J. M. 2001. Evolutionary relationships and conservation of the Hawaiian anatids. Studies in Avian Biology 22:61-67. SAS Institute Inc. 2001. SAS®. Version 8.2. SAS In- stitute, Cary, North Carolina, USA. SAS Institute Inc. 2005. JMP-IN®. Version 5. SAS Institute, Cary, North Carolina, USA. Schwartz, C. W. and E. R. Schwartz. 1953. Notes on the Hawaiian Duck. Wilson Bulletin 65:19-25. Semlitsch, R. D. and J. R. Bodie. 1998. Are small, isolated wetlands expendable? Conservation Bi- ology 12:1129-1133. SwEDBERG, G. E. 1967. The Koloa. Technical Report. Hawaii Division of Fish and Game, Honolulu, USA. U.S. Department of the Interior (USDI). 2005. Draft revised recovery plan for the Hawaiian water- birds. Second Revision. USDI, Fish and Wildlife Service, Portland, Oregon, USA. Weller, M. W. 1999. Wetland birds: habitat resources and conservation implications. Cambridge Uni- versity Press, New York, USA. The Wilson Journal of Ornithology 120(2);320— 330, 2008 HABITAT FEATURES ASSOCIATED WITH BARROW’S GOLDENEYE BREEDING IN EASTERN CANADA MICHEL ROBERT,'-3 bRUNO DROLET,' AND JEAN-PIERRE L. SAVARD^ ABSTRACT. We investigated environmental variables linked to presence of Barrow’s Goldeneye (Bucephala islandica) pairs from the eastern North American population on 412 lakes of the Sainte-Marguerite River wa- tershed, Quebec, Canada. We analysed habitat relationships at two spatial scales (i.e., considering all lakes surveyed and high elevation lakes only) and predetermined the high elevation lakes as those including 90% of Barrow’s Goldeneye occurrences. Barrow’s Goldeneye were found on 59 lakes, all of which were >490 m elevation (maximum = 822 m) with 90% at >610 m. Six variables tested using multivariate logistic regressions contributed to explain the occurrence of goldeneyes. Four were significant (P < 0.10) in both the complete and the high elevation data sets; nest boxes ( + ) and brook trout {Salvelinus fontinalis) (-) occurrences, altitude ( + ), and the interaction between altitude and mean slope ( + ). The models explained only a small proportion of Barrow’s Goldeneye occurrence for both data sets (P^ = 0.27 and 0.23, respectively). The negative relationship between Barrow’s Goldeneye and brook trout occurrences, and the positive relationship with altitude probably reflect a positive relationship between goldeneye and highly productive aquatic ecosystems. Barrow’s Goldeneye from eastern North America primarily use high altitudinal, productive lakes during the breeding season, which emphasizes the importance of fishless lakes for that population at risk. Received 16 January 2007. Accepted 7 September 2007. Barrow’s Goldeneye {Bucephala islandica) has a discontinuous distribution; it occurs mostly in western North America (—180,000 individuals) with smaller populations in east- ern North America (—6,000) and Iceland (-2,000) (Eadie et al. 2000, Einarsson et al. 2006, Robert and Savard 2006). Birds from the western North American population dur- ing the breeding season are predominantly as- sociated with alkaline lakes in parkland areas, but also with alpine and sub-alpine lakes, American beaver {Castor canadensis) ponds, and small sloughs (Eadie et al. 2000). No landscape-scale study has investigated habitat attributes associated with these breeding sites, but evidence suggests they are usually highly productive and invertebrate-rich (Savard et al. 1994, Evans 2003). Similarly, birds from the Icelandic population are largely confined to the invertebrate-rich areas of Lake Myvatn and Laxa River (Einarsson 1990, Gardarsson and Einarsson 2004). Birds from the eastern North American pop- ulation were thought to breed above tree-line in ' Canadian Wildlife Service, Environment Canada, 1 141 Route de I’Eglise, P. O. Box 10100, Quebec, QC, GIV 4H5, Canada. ^ Science and Technology, Environment Canada, 1141 Route de I’Eglise, P. O. Box 10100, Quebec, QC, GIV 4H5, Canada. ^ Corresponding author; e-mail; michel.robert@ec.gc.ca northeastern Canada (Palmer 1976, AOU 1998), but recent studies have shown their core breed- ing area to be much further south in large tracts of boreal forest north of the St. Lawrence River Estuary and Gulf in southeastern Canada (Rob- ert et al. 2000b, 2002). The birds reported by Robert et al. (2000b) were mainly on small (<10 ha), high elevation (>500 m), headwater lakes in forests dominated by black spruce {Pi- cea mariana) and feather moss, or by balsam fir {Abies balsamea) and white birch {Betula pa- pyrifera). Nothing is known about the specific habitat requirements of this small population, which may comprise <2,000 pairs (Robert and Savard 2006) and is on the List of Wildlife Spe- cies at Risk of Canada under the Species at Risk Act (Government of Canada 2007). We conducted helicopter surveys of gold- eneye pairs in the core breeding area of the eastern North American population of Bar- row’s Goldeneye with the specific objectives to identify: (1) habitat features, and (2) envi- ronmental variables associated with this pop- ulation. These data will be used to contribute in the production of a management plan iden- tifying conservation needs for this population at risk and its habitats. METHODS Study Area. — The area studied was in the Sainte-Marguerite River drainage basin (1,600 kmQ, Quebec, Canada on the north shore of 320 Robert et al. • BARROW’S GOLDENEYE BREEDING HABITAT 321 FIG. 1. Sainte-Marguerite River watershed in the balsam fir- white birch bioclimatic domain of the boreal forest in Quebec, Canada. the St. Lawrence River Estuary north of the Saguenay River (Fig. 1). It comprises 968 lakes with a mean (± SD) surface area of 13.2 ± 5.3 ha. The area is rugged with mountains (mean altitude = 607 ± 1 87 m) intersected by steep river valleys. The climate is cold, wet, and strongly seasonal. Mean annual tempera- ture is 0° C with annual precipitation between 100 and 130 cm. Annual snowfall is between 350 and 400 cm. The area is in the balsam fir- white birch bioclimatic domain (Wiken 1986, Robitaille and Saucier 1998) of the Laurentian Highlands physiographic region (Desponts 1996), and is primarily used for commercial timber production and recreational fishing. Goldeneye nest boxes were installed around 62 of the lakes of the study area in fall 1998 and 1999 (Savard and Robert 2007). Lake Sampling and Goldeneye Surveys. — We randomly selected 412 lakes in the Sainte- Marguerite River watershed based on surface area and elevation (Robert et al. 2000b): 58% were lakes 1-10 ha and >300 m, 22% were 10-100 ha and >300 m, 10% were 1-100 ha and 0-300 m, and 10% were 0. 2-1.0 ha at all elevations. Lakes were surveyed once by he- licopter (Bell 206 L equipped with bubble- type rear windows) between 28 May and 1 June 2001 noting the presence of Barrow’s and Common (Bucephala clangula) golden- eyes; a lone male or a pair of each species was considered as a pair occurrence for a giv- en lake. Flights over lake margins were at an altitude 10-50 m and flight speed varied from 50 to 100 km/hr, depending on the complexity of aquatic habitats, atmospheric conditions, and topography. Goldeneyes were identified by two observers, a navigator in front of the helicopter and a compiler at the rear of the cabin on the same side as the navigator. The same observers were used in all surveys. Bar- row’s Goldeneye were distinguished from Common Goldeneye from the air by their dis- tinctive, darker upperwing pattern and, when 322 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 possible, by the males’ facial crescent or the females’ bill color (Tobish 1986, Eadie et al. 2000). Environmental Variables. — Considering our sample size and remote study area, most envi- ronmental variables were derived from satellite imagery and topographic maps. This under- scores the exploratory namre of the study and lack of a priori models. We investigated the re- lationships between goldeneye occurrence and 20 environmental variables, including the pres- ence of nest boxes at some lakes {n = 32), which were considered as a co variable. Fish oc- currence was estimated using data for brook trout (Salvelinus fontinalis) from the Quebec Ministere des Ressources naturelles et de la Fau- ne (Jean Tanguay, pers. comm.). Fakes without trout were considered as fishless, as this is typ- ically the only fish species inhabiting lakes of the study area (Facasse and Magnan 1994). The surface area (ha), altitude (m), and number of lakes upstream were measured for each lake us- ing 1:50,000 topographic maps from the Na- tional Topographical Data Base (NTDB) (Nat- ural Resources Canada 2005). Relative abun- dance of emergent rocks and riparian wetlands were visually estimated during helicopter sur- veys, using a semi-quantitative scale ranging from one (little or none) to three (many). Mean slope (%) was measured in a 500-m zone ra- diating from the lake’s perimeter using maps from the NTDB and the pixel interpolation method based on a 25 -m^ pixel grid and a 7X7 pixel calculation window using a smoothed diagonal search (module GRDINT) (Carrara 1988, PCI 2001). All other variables were obtained from a Fandsat TM5 multispec- tral satellite image of the study area taken on 4 June 1999 (VIASAT 2001), which contained ra- diometric information in seven spectrally de- fined channels. Habitats were mapped using a supervised classification approach with classes defined a priori. The spectral signature (radi- ance) of a given habitat class was assigned by sampling pixels known to belong to a given class. Habitat classes were validated using Que- bec Ministere des Ressources naturelles et de la Faune 1:20,000 forest maps dating from 1999 and field data collected on 20—21 September 2001 {n = 35 plots). We used the software ENVI (2005) to evaluate differences between spectral signatures of habitat classes. A first fil- ter was used to exclude all clear-cut areas and reduce confusion between mixed forest and re- generation stands. A second filter was used to exclude all burned areas to reduce confusion be- tween recent bums and recent dear-cuts. Each pixel was assigned to one of 15 different habitat classes using a maximum likelihood approach with a minimum probability level of 80% re- semblance. Eight homogeneous patch types were developed after classification: open water (deep water and shallow water combined), wet- lands (marshes, exploited and natural peatlands), exposed mineral (mainly logging roads), recent dear-cuts (0-5 years or filter 1 and 5-10 years post dear-cuts combined), recent bums (filter 2), regeneration (10-20 years post clear-cuts and deciduous regeneration combined), mature de- ciduous and mixed forests combined, and ma- ture conifer forests (dense and scarce conifer stands combined). The classified image was smoothed using a 3 X 3 filter (Majority 3X3) (ENVI 2005). The raster version of the classi- fied image was analysed using FRAGSTAT (McGarigal and Marks 1995). We calculated the surface area of all patch types for each lake in a 2-km zone radiating from its perimeter and converted these surfaces areas into percentages without considering the portion of the image ob- scured by cloud cover (mean ± SD - 0.09 ± 0.4%). We measured four lake configuration in- dices in a 16-km2 (4x4 km) plot centered on each surveyed lake: number of lakes, mean sur- face area of lakes (ha), lakes mean nearest- neighbor distance (m), and lake mean shape in- dex (1 for round lakes and >1 for irregularly shaped lakes). Habitat Relationships. — We verified the lin- earity of relationships between the presence of Barrow’s Goldeneye pairs and environmental variables individually using a graph of the logit(p) = log(p/[l - p]) according to five bal- anced classes of the explanatory variable, where p is the frequency of Barrow’s Gold- eneye for each of the five classes. Ten vari- ables had non-linear relationships with gold- eneye occurrence. Fake surface area, mean slope, and coniferous forests had quadratic re- lationships and were transformed. The other variables (number of lakes upstream [<2 lakes or more], riparian wetlands [class 1 or class 2 and 3], wetlands [<0.7% or greater], exposed mineral [<0.2% or greater], recent bums [<2% or greater], regeneration [<7% or greater], and number of lakes [<18 lakes or Robert et al. • BARROW’S GOLDENEYE BREEDING HABITAT 323 900 - 800 - 700 - 600 - B 500 - a> hoo- 300 - 200 - 100 - 0 ^ ^ ^ ^ 1 ^ ^ ^ ^ ^ 1 48.25 48.30 48.35 48.40 48.45 48.50 48.55 48.60 48.65 48.70 48.75 48.80 Degree (N) FIG. 2. Distribution by latitude and altitude of 412 lakes surveyed in the Sainte-Marguerite River drainage basin, Quebec, Canada. Black and white dots denote lakes with (observed occurrences) or without (observed absences) Barrow’s Goldeneye, respectively. Gray squares identified lakes that were predicted as occupied by Barrow’s Goldeneye according to a logistic probability level obtained from a ROC analysis. greater]) were transformed as categorical var- iables in a binary format. We used univariate and multivariate (step- wise selection) logistic regressions to assess relationships between the occurrence of Bar- row’s Goldeneye pairs and environmental var- iables (Proc Logistic) (SAS Institute Inc. 2001). Highly redundant variables (variance inflation >20) and variables with a probability level >0.25 calculated from univariate anal- yses were excluded from the multivariate analysis. Because of the large sample size (n = 412) and to minimize the risk of spurious significant correlations, significance levels were set at 0.15 and 0.10 to include or keep, respectively, a variable in multivariate analy- sis. The maximum rescaled coefficient of de- termination (r~) (Nagelkerke 1991) was cal- culated to evaluate the contribution of each variable to the multivariate analysis. We analyzed habitat relationships at two spatial scales considering all lakes surveyed and high elevation lakes only. We predeter- mined high elevation lakes as those including 90% of Barrow’s Goldeneye occurrences. Models resulting from both data sets were as- sessed using the same data for training and testing (resubstitution). We used the crossover of misclassification rates (sensibility vs. 11 - specificity]) in a receiver-operating character- istic (ROC) analysis to examine the optimum probability threshold at which Barrow’s Gold- eneye should occur (Cody and Smith 1997). A confusion matrix was built using this threshold to evaluate sensitivity (proportion of true presence predicted), overall prediction success (proportion of true presence and ab- sence predicted), and Kappa, a measure of the proportion of all possible cases of presence or absence that are predicted correctly after ac- 324 THE WILSON JOURNAL OF ORNITHOLOGY • Vol 120, No. 2, June 2008 (U 'O ^ S I 2 I ^ Al § - (U (U c > ^ (U T3 -r (D o ^ o c/3 (U u a- ci( c/5 (U >• i3 • •a g o s: o w IT) ■B «« ■> c > ^ ^ n T3 jO c/3 (U c/3 a ^ "S 3 o cat o c - c P3 o t3 O (U o x: (U £ n 2 3 O n (U (U c 00 o ^ C 0^ (u li! 13 i c ^ (U oc -2 u + 1 + 1 + 1 + 1 + 1 + in 00 CO X (N CO in (N (N CO CO T o o q q q 00 0 0 0 q d d d d d d d d d d d ' rt CO -— ' ■3" (N — m 00 CO r- r- CO CM (N CM If o o — o o q 0 0 CO q q q — 0 — — 0 0 00 a. d d d d d d d) d> d> d> <0 d> d d d d d d d d V V V 3 3 X X X CO m >n -ct OOOONOOrf in ON \D — 0 ON CV in If cn 0 00 0 CO 3- 0 in CO in CO in o in q 9 Meat n/a n/a 10 724 1. 1. — (N d CN X — 00 00 CO CO CM in 202 1 q q X QS ^ 00 ■3- CO CO ON >n ■3- in CO X 00 CM 1 00 00 — 00 (N — X CO in ON 00 CO if q q ^ 0 0 X X Mir CO ^ d 0 0 — ' CO — (N X 0 0 0 0 z I 0 d d if X d 67 1 3 3 00 ^ n ~2 'c ^ ij" ON CN 00 If CO lO -- 104 ON in ON 1^ > U Tf CO in 0 ON 00 ON CO X CO CO in 9 c 33 9 9 > '3 On 0 — C = cn r- CO 9 9 q 1 — ’ in d — in d CM (N CO (M 31 20 in 230 1 00 Max (N If 0 n X (N — m 1 59 (M 00 X X 00 CO - 531 1 — 9 On 0 0 — 9 9 I (N 0 d 0 0 — Tt CO — 0 0 0 0 1 0 d (M X 124 1 X U + 1 + 1 + 1 1 + + X 00 0 m If CN X CO 'L 00 -- 0 0 0 0 0 d d d d d d d d d 0 0 CO 0 ' (N CO in — 3- CN CM — 0 — CO On — 0 0 — 000 0 If 00 q — _ q \D 0 00 CO in 0 X d d d d d d d> d> d> d> d d> d d d d d d d d V V V V V V V V 3 3 in 00 0 in in (N r- 3- — 0 ON if m (N r-- 0 CO if in X CO ON > u in if >n 0 On (N If 0 in — ON CO in if ON >n — CO NO (N ON 00 tT3 CO CO 00 r- — (N 00 O CO O O O — 00 CO in CO — ^ c« JS s: ^ 3 £ S C c/3 ^ C ^ c ^ c3 a| c2 5 ^ ^ u •- CU C r- 3 3 <>->. c3 3 t--. On r-- o — < NO CO 1/1 T 9 T — '(NOOOOOd 00 d ^ CN Tj- r- o o c t 3 ^ 3 c; ^ S 3 -o 3 S' ^ fli a H O 3 3 o 3 ^ £ w c3 13 ^ (U ^ B g -s z ■£ C3 3 3 _ J J hJ U Dii o S o ^ ’C 3 r C ^ 3 ^ "o X) (U w eg C C O (U (U O- CJ o X (U o a C2, ^ 2 -O ^ o ^ 3 C 13 .2 o w (U 3 "O s> x: L != o c/3 (J o (U ^ O c o c W) C2, S Robert et al. • BARROW’S GOLDENEYE BREEDING HABITAT 325 counting for chance effects (Fielding and Bell 1997). RESULTS Barrow’s Goldeneye were found on 59 (14.3%) lakes, Common Goldeneye on 32 (7.8%), and both species on 6 (1.5%); 79.4% {n — 327) of the lakes surveyed had no gold- eneye. Only one lake had more than one (i.e., 2) Barrow’s Goldeneye pair. All Barrow’s Goldeneye were on lakes >490 m (maximum = 822 m) elevation and 90% were at an al- titude >610 m (i.e., the threshold we used to create the high elevation data set [n = 306, 17.6% of lakes with Barrow’s Goldeneye]) (Fig. 2). Common Goldeneye were found at a broader range of altitudes (190—822 m) and 31% were on lakes <610 m. We chose not to analyze Common Goldeneye data considering low occurrence of the species and the lake sampling design, which tended to favor Bar- row’s Goldeneye. Univariate Logistic Regressions. — Barrow’s Goldeneye occurrence was higher {P < 0.10) on lakes with nest boxes, no fish, none or few lakes upstream, many emergent rocks, none or few riparian wetlands, and at high elevation in landscapes with low proportion of wetlands, high proportion of regeneration, and high mean nearest-neighbor distances. The same environ- mental variables were selected {P < 0.10) when considering only the high elevation data set. The occurrence of Barrow’s Goldeneye was also higher (P < 0.10) on high elevation lakes with high mean slopes and in landscapes with low number of lakes (Table 1). Multivariate Logistic Regressions. — Six variables tested explained the occurrence of Barrow’s Goldeneye (Table 2). Four were sig- nificant {P < 0.05) in both the complete and the high elevation data sets: nest boxes ( + ), fish occurrence ( — ), altitude ( + ), and the in- teraction between altitude and mean slope ( + ). The number of lakes upstream ( — ) was significant {P = 0.07) only in the complete data set, and riparian wetlands ( — ) only in the high elevation data set (P = 0.03). Most re- lationships were highly significant (P 0.01 ), but models explained only a small proportion of Barrow’s Goldeneye occurrence for both the complete and high elevation data sets (P^ = 0.27 and R^ = 0.23, respectively) (Table 2). The cross-over of misclassification rates in- S II ... ~ oc :3 S. I S — II > — q ^ 5 u ■§ y ^ 3 E -3 = d T ir. z — ^ ^ ? 3 “ ri ^ 'S c T cT: > ir; 2 (N O s S d O o ir; rr — O O o O o o o' o d d d d V V V ir-, V 9 q rn r*', d C (N vC II c IT-, CN oc (N O r- d d g o T d sC 1 1 5 C O 0 9 M c c 3 ITi is Al (N o u Q d d o d 1 1 — IT-, 1 _ 9 oc ^ < IT! o m rr ■c o' d d q d d 1 (N ' T r- oc q ir, oc d OC 9 d ^ O r', (N rT* o — o O O d d d X - o o r- o O o d d d d d V V V IT; V 9 m - ri o _ II O o OC o o q oc (N 1 d vC 1 d 1 7 1 0 9 1 1 2 B u B B z >J^4 oc Q r- — o < o (N o q — r-, — T d d d 1 d 1 ir, r' 9 i/', o LXJ ir, rf < it; o sC rr-, T d ir, d 1 d 1 oc oc ir, 1/", (N < d rj 3 i- 3 "3 1 h y L o 1) 3 3 y; 4 y y 3 y Z 1 _3 y -3 y t 3 I ■•3 X y H = E. £ y 3 ”7 r~. 1 3 g- y y. z ~jr. y y: 7 Mcai 3 3 z d 2 5 = 326 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 dicated that lakes with a logistic probability level >0.173 could be predicted to be occu- pied by Barrow’s Goldeneye. The confusion between predicted and observed occurrences indicated, for both the complete and high el- evation data sets, a low sensitivity (31 and 32%, respectively) and Kappa (30 and 25%, respectively) values, but a high overall pre- diction success (72 and 71%, respectively) be- cause of the high prevalence of absences well predicted (Fig. 2). DISCUSSION This study is the first to explore habitat re- lationships for Barrow’s Goldeneye breeding in eastern North America. The multivariate models explained only a small portion of the variance, but there was a strong and signifi- cant link between some variables and the oc- currence of Barrow’s Goldeneye for both the complete data set and high elevation lakes. We found a clear negative relationship with brook trout occurrence. Abundance, diversity, and assemblage characteristics of aquatic insects vary with respect to fish populations as fish predation may, for example, eliminate or re- duce the abundance of nekton or cause them to mature at smaller size (Pope et al. 1973, Pope and Carter 1975, Bendell and McNicol 1987, Chess et al. 1993). Mallory et al. (1994) found that fish status of a lake was a reliable cue to invertebrate abundance in boreal wet- lands. Goldeneyes rely on aquatic inverte- brates for food (Bendell and McNicol 1995, Eadie et al. 2000) and food competition by fish may affect their use of habitat (Eriksson 1979, Eadie and Keast 1982, DesGranges and Gagnon 1994, Poysa et al. 1994). We believe the negative relationship between Barrow’s Goldeneye and brook trout occurrences re- flects a positive relationship between golden- eye and highly productive aquatic ecosystems. Barrow’s Goldeneye from other populations, worldwide, have also been linked to produc- tive habitats, as they occur on small, shallow, alkaline fishless lakes in the Cariboo Park- lands of central British Columbia (Topping and Scudder 1977, Boyd and Savard 1987, Boyd and Smith 1989, Evans 2003) and in- vertebrate-rich volcanic areas of Lake Myvatn and upper Laxa River in Iceland (Einarsson 1988, 1990; Gardarsson and Einarsson 2004). The drainage basin where we conducted our study historically had at least 185 fishless lakes >2 ha (Jean Tanguay, Ministere des Ressources naturelles et de la Faune du Que- bec, unpubl. data), and is an important area of the Laurentian Highlands for brook trout al- lopatry and fishless lakes. Its high altitude and rugged topography precluded many lakes from being recolonized by brook trout after the Wisconsin glaciation (Pope et al. 1989, Lacasse and Magnan 1994). Barrow’s Goldeneye occurrence was clearly associated with altitude, as the species was ab- sent below 490 m. Barrow’s Goldeneye re- ported by Robert et al. (2000b) were also on high elevation lakes, 54% of them above 500 m. It is difficult to dissociate elevation from fishless lakes, as a large proportion occurs at the head of watersheds at high elevation in our study area. The information we obtained on occurrence of brook trout may have been im- perfect, because illegal stocking of fishless lakes is known to occur in our study area (Jean Tanguay, pers. comm.). This may have enhanced the relative importance of altitude, because lakes at higher altitudes are less read- ily accessible for fish stocking than those at lower elevation. The relationship between goldeneyes and altitude may also be related to competition, considering the larger, more ag- gressive Barrow’s Goldeneye may dominate the smaller Common Goldeneye (Savard 1984, Savard and Smith 1987), which is also known to favor lakes without fish (Eriksson 1979, Eadie and Keast 1982, Poysa and Vir- tanen 1994). Common Goldeneye were nu- merous in the valley below our study area and Barrow’s Goldeneye did not occur there (Se- nechal 2003) supporting the association of Barrow’s Goldeneye with high elevation lakes. Similar habitat segregation has been ob- served in the Columbia Valley, British Colum- bia, where Barrow’s Goldeneye occurred only on alkaline lakes of plateaus whereas Com- mon Goldeneye were more numerous on freshwater valley lakes (Savard 1984). We considered nest boxes as a covariate be- cause goldeneyes may use artificial cavities for nesting (Eriksson 1982, Evans et al. 2002, Poysa and Poysa 2002) and because boxes were installed on some lakes in our study area in 1998 and 1999 (Savard and Robert 2007) with the intention of studying Barrow’s Gold- eneye. We believe the relationships between Robert et al. • BARROW’S GOLDENEYE BREEDING HABITAT 327 Barrow’s Goldeneye occurrence and nest box presence is partly biased, as we erected nest boxes on lakes which either had pairs or were close to lakes that had pairs of Barrow’s Gold- eneye during ground surveys conducted in 1998; no nest boxes were established on lakes below 518m. We used multivariate regression without considering nest boxes and found sig- nificant (P < 0.06) relationships between Bar- row’s Goldeneye and the exact same variables (i.e., fish occurrence, altitude, interaction of altitude and slope and, depending on data set, number of lakes upstream [all lakes] or ripar- ian wetlands [high elevation lakes]). Coeffi- cients were also similar (0.24 and 0.19, re- spectively), which supports our conclusion that nest boxes should not be considered as an important variable in this study. Savard (1988) found nest sites up to 2.9 km from pair terri- tories indicating that presence of a nest site (contrary to food resources) was not essential in the selection of a pair territory. The biological significance of relationships between Barrow’s Goldeneye occurrence and other habitat features are less obvious. The mean slope in a 500-m zone radiating from lakes with the low number of lakes upstream may be associated with lake isolation (com- plete or partial) from fish colonization, in- creasing invertebrate productivity. It may also be related to cavity availability considering that Barrow’s Goldeneye from eastern North America use large, decaying trees for nesting rather than abandoned woodpecker cavities as in western North America (Vaillancourt 2006; MR, unpubl. data). Trees, plus natural rock crevices that could also be used for breeding as in Iceland (Eadie et al. 2000), may be more common around lakes surrounded by moun- tainous and steep areas, considering that areas >4 ha with trees on slopes >40% cannot be harvested by the timber industry in Quebec (Ministere des Ressources naturelles et de la Faune 2003). Barrow’s Goldeneye were also associated with lakes with rocks and cluster of lakes, which may reflect shallow productive waters (Savard et al. 1994, Eadie et al. 2000) and/or provide resting and preening sites (Robert et al. 2006). Barrow’s Goldeneye in British Columbia were also associated with lakes that had few riparian marshes or sub- mergent vegetation (Savard et al. 1994, Evans 2003). A variety of factors may have contributed to the low proportion of variability explained by our models. Our estimate of Barrow’s Goldeneye occurrence was based on a single helicopter survey and we may have missed some individuals or observed birds on lakes where they did not breed. Variables such as fish occurrence and the number of lakes up- stream probably represent surrogate indicators of lake invertebrate productivity, and actual measures of goldeneye prey availability would probably have yielded better results (Evans 2003). Stocked fishless lakes may remain at- tractive to goldeneyes for some time as gold- eneyes are strongly philopatric (Eadie et al. 2000). Our data set had limitations as most variables used in our analysis were derived from remote sensing and topographic maps. The potentially important variables such as lake depth, availability of nesting cavities, or lake productivity were not considered (John- son and Winter 2005). These variables were important in lake selection by Barrow’s Gold- eneye in western North America (Boyd et al. 1989, Savard et al. 1994, Evans 2003). Bar- row’s Goldeneye is an uncommon species in eastern North America and its population is estimated at <2,000 pairs (Robert and Savard 2006) distributed primarily across Quebec’s eastern boreal forest during the breeding sea- son (Robert et al. 2000b, 2002). This is a vast area with over 165,000 lakes, >50,000 of which are between 1 and 10 ha in size (NTDB) (Natural Resources Canada 2005). Helicopter surveys conducted yearly from 1990 to 2006 in the central part of the Bar- row’s Goldeneye breeding area yielded mean pair densities almost 1 1 times lower for Bar- row’s Goldeneye than for Common Golden- eye (1.2 vs. 13.0 pairs/100 km^, respectively) (Daniel Bordage, Canadian Wildlife Service, unpubl. data). The low density and extensive distribution of Barrow’s Goldeneye in eastern North America may partly explain the diffi- culty of predicting its occurrence at the breed- ing area level even though some variables, such as altitude (Fig. 2) may help in identi- fying potential breeding areas at the landscape scale. CONSERVATION IMPLICATIONS This study has implications for conserva- tion of Barrow’s Goldeneye in eastern North 328 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 America. It confirms the findings of Robert et al. (2000b) (i.e., Barrow’s Goldeneye from eastern North America primarily use high al- titudinal, productive lakes during the breeding season) and emphasizes the importance of fishless lakes for that population at risk. Un- fortunately, many lakes in our study area have been stocked with brook trout during the last decades (n > 67, by government agencies alone), especially since logging roads have in- creased accessibility. The breeding range of Barrow’s Goldeneye in Quebec covers many areas managed for fishing and hunting activ- ities in which efforts are made to introduce brook trout into originally fishless lakes (Rob- ert et al. 2000a). Stocking of fishless lakes for recreational activities should cease as it may affect Barrow’s Goldeneye and have a nega- tive impact on other taxa such as macroinver- tebrates and amphibians (Chess et al. 1993, Bradford et al. 1998, Pilliod and Peterson 2001, Drouin et al. 2006). Our study indicates that environmental variables including lake al- titude, mean slope, and lake position in the watershed may be useful in predicting Bar- row’s Goldeneye occurrence at the landscape scale, but not at the lake scale. We believe that management measures should be extended to potential Barrow’s Goldeneye breeding areas rather than focussing on lakes with known oc- currences, at least until additional, more spe- cific variables can be identified. ACKNOWLEDGMENTS This study was funded by the Species at Risk Di- vision of the Canadian Wildlife Service, thanks to Is- abelle Ringuet. We sincerely thank Christian Marcotte for participating in helicopter surveys and data com- pilation. We thank Rejean Benoit for participating in the planning of the helicopter survey as well as the initial data analysis. We are grateful to Helene Crepeau from Universite Laval for data analysis, and Sylvain Deslandes, Marcelle Grenier, and Martine Benoit for GIS analysis. We express our gratitude to Jean Tan- guay, from Quebec’s Ministere des Ressources natu- relles et de la Faune, for providing data and informa- tion on fishless lakes in our study area. We thank Lise Deschenes, also from Quebec’s Ministere des Res- sources naturelles et de la Faune, for providing details regarding timber harvesting regulations. Michel Me- lan9on kindly produced Figure 1 and calculated the number of lakes distributed within the Barrow’s Gold- eneye core breeding area. Daniel Bordage kindly pro- vided goldeneye density data from the Waterfowl Sur- vey of Southern Quebec Uplands. We also thank Syl- vie Gauthier, Stephane Legare, Laurent Dufour, and three anonymous reviewers for helpful comments on this paper. LITERATURE CITED American Ornithologists’ Union (AOU). 1998. Check-list of North American birds. Seventh Edi- tion. American Ornithologists’ Union, Washing- ton, D.C., USA. Bendell, B. E. and D. K. McNicol. 1987. Fish pre- dation, lake acidity and the composition of aquatic insect assemblages. Hydrobiologia 150:193-202. Bendell, B. E. and D. K. McNicol. 1995. The diet of insectivorous ducklings and the acidification of small Ontario Lakes. Canadian Journal of Zoology 73:2044-2051. Boyd, W. S. and J.-R L. Savard. 1987. Abiotic and biotic characteristics of wetlands at Riske Creek, British Columbia: a data report. Technical Report Series Number 16. Canadian Wildlife Service, Delta, British Columbia, Canada. Boyd, W. S. and D. W. Smith. 1989. Summary of aquatic invertebrate data collected from wetlands at Riske Creek, British Columbia, 1984 and 1985. Technical Report Series Number 60. Canadian Wildlife Service, Delta, British Columbia, Cana- da. Boyd, W. S., J.-P. L. Savard, and G. E. J. Smith. 1989. Relationships between aquatic birds and wetland characteristics in the Aspen Parkland, central British Columbia. Technical Report Series Number 70. Canadian Wildlife Service, Delta, British Columbia, Canada. Bradford, D. F, S. D. Cooper, T. M. Jenkins Jr., K. Kratz, O. Sarnelle, and A. D. Brown. 1998. Influences of natural acidity and introduced fish on faunal assemblages in California alpine lakes. Canadian Journal of Fisheries and Aquatic Sci- ences 55:2478-2491. Carrara, A. 1988. Drainage and divide networks de- rived from high-fidelity digital terrain models. Pages 581-597 in Proceedings of the NATO Ad- vanced Study Institute on Statistical Treatments for Estimation of Mineral and Energy Resources. D. Reidel Publishing Company, Dordrecht, The Netherlands. Chess, D. W., F. Gibson, A. T. Scholz, and R. J. White. 1993. The introduction of Lahontan cut- throat trout into a previously fishless lake: feeding habits and effects upon the zooplankton and ben- thic community. Journal of Freshwater Ecology 8: 215-225. Cody, R. P. and J. K. Smith. 1997. Applied statistics and the SAS programming language. Fourth Edi- tion. Prentice Hall, Upper Saddle River, New Jer- sey, USA. DesGranges, J.-L. and C. Gagnon. 1994. Duckling response to changes in the trophic web of acidified lakes. Hydrobiologia 279/280:207-221. Desponts, M. 1996. The biogeography of Quebec: ma- jor physiographic regions. Pages 18—20 in The breeding birds of Quebec (J. Gauthier and Y. Au- Robert et al. • BARROW’S GOLDENEYE BREEDING HABITAT 329 bry, Editors). Association quebecoise des groupes d’ornithologues. The Province of Quebec Society for the Protection of Birds, Societe quebecoise pour la protection des oiseaux, Canadian Wildlife Service, Environment Canada, Quebec Region, Montreal, Canada. Drouin, a., P. Sirois, and P. Archambault. 2006. Structure des communautes d’invertebres et des especes d’amphibiens dans les lacs avec et sans omble de fontaine {Salvenilus fontinalis) en foret boreale. Canadian Technical Report of Eisheries and Aquatic Sciences Number 2628. Institut Mau- rice-Lamontagne, Mont-Joli, Quebec, Canada. Eadie, J. M. and a. Keast. 1982. Do goldeneye and perch compete for food? Oecologia 55:225-230. Eadie, J. M., J.-P. L. Savard, and M. L. Mallory. 2000. Barrow’s Goldeneye (Bucephala islandica). The birds of North America. Number 548. Einarsson, a. 1988. Distribution and movements of Barrow’s Goldeneye Bucephala islandica young in relation to food. Ibis 130:153-163. Einarsson, A. 1990. Settlement into breeding habitats by Barrow’s Goldeneyes Bucephala islandica: ev- idence for temporary oversaturation of preferred habitat. Ornis Scandinavica 21:7-16. Einarsson, A., A. Gardarsson, G. M. Gislason, and G. Gudbergsson. 2006. Populations of ducks and trout of the River Laxa, Iceland, in relation to var- iation in food resources. Hydrobiologia 567:183- 194. ENVI. 2005. Environment for visualizing image. RSI Corporate, Boulder, Colorado, USA. Eriksson, M. O. G. 1979. Competition between fresh- water fish and goldeneyes Bucephala clangula (L.) for common prey. Oecologia 41:99-107. Eriksson, M. O. G. 1982. Differences between old and newly established goldeneye Bucephala clangula populations. Ornis Fennica 59:13-19. Evans, M. R. 2003. Breeding habitat selection by Bar- row’s Goldeneye and Bufflehead in the Cariboo- Chilcotin region of British Columbia: nest sites, brood-rearing habitat, and competition. Disserta- tion. Simon Fraser University, Burnaby, British Columbia, Canada. Evans, M. R., D. B. Lank, W. S. Boyd, and E Cooke. 2002. A comparison of the characteristics and fate of Barrow’s Goldeneye and Bufflehead nests in nest boxes and natural cavities. Condor 104:610- 619. Fielding, A. H. and J. E Bell. 1997. A review of methods for the assessment of prediction errors in conservation presence/absence models. Environ- mental Conservation 24:38-49. Gardarsson, A. and A. Einarsson. 2004. Resource limitation of diving ducks at Myvatn: food limits production. Aquatic Ecology 38:285-295. Government of Canada. 2007. Species at Risk Act Public Registry. Canadian Wildlife Service, Fish- eries and Oceans Canada, and Parks Canada Agency, www.sararegistry.gc.ca/default_e.cfm (accessed 9 January 2007). Johnson, D. H. and M. Winter. 2005. Planning for bird conservation: a tale of two models. Pages 1205-1212 in Bird conservation implementation and integration in the Americas: Proceedings of the Third International Partners in Flight Confer- ence (C. J. Ralph and T. D. Rich, Editors). General Technical Report PSW-GTR-191, USDA, Forest Service, Pacific Southwest Research Station, Al- bany, California, USA. Lacasse, S. and P. Magnan. 1994. Distribution post- glaciaire de 1’ omble de fontaine dans le bassin hy- drographique du fleuve Saint-Laurent: impact des interventions humaines. Universite du Quebec a Trois-Rivieres, Trois-Rivieres, Quebec, Canada. Mallory, M. L., P. J. Blanchard, P. J. Weatherhead, AND D. K. McNicol. 1994. Presence or absence of fish as a cue to macroinvertebrate abundance in boreal wetlands. Hydrobiologia 279/280:345- 351. McGarigal, K. and B. J. Marks. 1995. FRAGS- TATS: spatial pattern analysis program for quan- tifying landscape structure. General Technical Re- port PNW-GTR-351, USDA, Forest Service, Pa- cific Northwest Research Station, Portland, Oregon, USA. Ministere des Ressources naturelles et de la Faune. 2003. Manuel d’amenagement forestier. Gouver- nement du Quebec, Direction des programmes fo- restiers, Quebec, Quebec, Canada. Nagelkerke, N. j. D. 1991. a note on a general def- inition of the coefficient of determination. Bio- metrika 78:691-692. Natural Resources Canada. 2005. National topo- graphical data base (NTDB). Centre for Topo- graphic Information, Earth Sciences Sector, Nat- ural Resources Canada, Ottawa, Ontario, Canada. Palmer, R. S. 1976. Handbook of North American birds. Volume 3. Yale University Press, New Ha- ven, Connecticut, USA. PCI. 2001. EASI PACE Software. Release 6.3 PCI Geomatics Inc., Richmond Hill, Ontario, Canada. PiLLiOD, D. S. AND C. R. Peterson. 2001. Local and landscape effects of introduced trout on amphibi- ans in historically fishless watersheds. Ecosystems 4:322-333. Pope, G. E and J. C. H. Carter. 1975. Crustacean plankton communities of the Matamek River sys- tem and their variation with predation. Journal of the Fisheries Research Board of Canada 32:2530- 2534. Pope, G. E, J. C. H. Carter, and G. Power. 1973. The influence of fish on the distribution of Chao- horus spp. (Diptera) and density of larvae in the Matamek River system, Quebec. Transactions of the American Fisheries Society 4:707-714. Pope, G., M.-C. Tarissants, J.-E Frenette, G. Ver- REAULT, AND J.-L. Desgranges. 1989. Effets des facteurs biotiques et abiotiques sur la structure et les relations trophiques des communautes planc- toniques et benthiques: revue des hypotheses. Rapport technique canadien des sciences halieu- 330 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 tiques et aquatiques Number 1687. Ministere des Peches et des Oceans, Institut Maurice-Lamon- tagne, Mont-Joli, Quebec, Canada. POYSA, H. AND S. POYSA. 2002. Nest-site limitation and density dependence of reproductive output in the Common Goldeneye Bucephala clangula: impli- cations for the management of cavity-nesting birds. Journal of Applied Ecology 39:502-510. PoYSA, H. AND J. ViRTANEN. 1994. Habitat selection and survival of Common Goldeneye {Bucephala clangula) broods — preliminary results. Hydro- biologia 279/280:289-296. POYSA, H., M. Rask, and P. Nummi. 1994. Acidifica- tion and ecological interactions at the higher tro- phic levels in small forest lakes: the perch and the Common Goldeneye. Annales Zoologici Fennici 31:397-404. Robert, M. and J.-P. L. Savard. 2006. The St. Law- rence River Estuary and Gulf: a stronghold for Barrow’s Goldeneyes wintering in eastern North America. Waterbirds 29:437-450. Robert, M., R. Benoit, and J.-P. L. Savard. 2000a. Status report on the eastern population of the Bar- row’s Goldeneye {Bucephala islandica) in Cana- da. Committee on the Status of Endangered Wild- life in Canada (COSEWIC), Ottawa, Ontario, Canada. Robert, M., R. Benoit, and J.-P. L. Savard. 2002. Relationship among breeding, molting, and win- tering areas of male Barrow’s Goldeneye in east- ern North America. Auk 119:676-684. Robert, M., B. Drolet, and J.-P. L. Savard. 2006. Effects of backpack radio-transmitters on female Barrow’s Goldeneyes. Waterbirds 29:115-120. Robert, M., D. Bordage, J.-P. L. Savard, G. Fitz- gerald, AND F. Morneau. 2000b. The breeding range of the Barrow’s Goldeneye in eastern North America. Wilson Bulletin 112:1—7. Robitaille, a. and J.-P. Saucier. 1998. Paysages re- gionaux du Quebec meridional. Les Publications du Quebec, Sainte-Foy, Quebec, Canada. SAS Institute Inc. 2001. The SAS System for Win- dows. Release 8.02 TS Level 02MO. SAS Insti- tute Inc., Cary, North Carolina, USA. Savard, J.-P. L. 1984. Territorial behaviour of Com- mon Goldeneye, Barrow’s Goldeneye and Buffle- head in areas of sympatry. Ornis Scandinavica 15: 211-216. Savard, J.-P. L. 1988. Winter, spring and summer ter- ritoriality in Barrow’s Goldeneye: characteristics and benefits. Ornis Scandinavica 19:119-128. Savard, J.-P. L. and M. Robert. 2007. Use of nest boxes by goldeneyes in eastern North America. Wilson Journal of Ornithology 119:28-34. Savard, J.-P. L. and J. N. M. Smith. 1987. Interspe- cific aggression by Barrow’s Goldeneye: a de- scriptive and functional analysis. Behaviour 102: 168-184. Savard, J.-P. L., W. S. Boyd, and G. E. J. Smith. 1994. Waterfowl-wetland relationships in the As- pen Parkland of British Columbia: comparison of analytical methods. Hydrobiologia 279/280:309- 326. Senechal, H. 2003. Etude comparative de I’ecologie de nidification et d’elevage des couvees de Garrot a ceil d’or {Bucephala clangula) et de Harle cou- ronne {Lophodytes cucullatus) dans un habitat de riviere. Thesis. Universite Laval, Quebec, Quebec, Canada. Tobish, T. 1986. Separation of Barrow’s and Common goldeneyes in all plumages. Birding 18:17—27. Topping, M. S. and G. G. E. Scudder. 1977. Some physical and chemical features of saline lakes in central British Columbia. Syesis 10:145—166. Vaillancourt, M.-A. 2006. Caracterisation de la dis- ponibilite des arbres potentiels a la nidification du Garrot d’Islande dans la foret boreale de I’Est du Quebec. Thesis. Universite du Quebec a Montreal, Montreal, Quebec, Canada. VIASAT. 2001. Landsat TM5 satellite imaging, scene identification number: 5013026009915510. Via- sat-Geotechnologie Inc., Montreal, Quebec, Can- ada. WiKEN, E. B. 1986. Terrestrial ecozones of Canada. Ecological Land Classification Series Number 19. Environment Canada, Hull, Quebec, Canada. The Wilson Journal of Ornithology 120(2):33 1—338, 2008 DISTRIBUTION, ABUNDANCE, AND NEST-SITE CHARACTERISTICS OE BLACK SWIETS IN THE SOUTHERN ROCKY MOUNTAINS OF COLORADO AND NEW MEXICO RICHARD G. LEVAD,i 6 KIM M. POTTER, ^ ^ CHRISTOPHER W. SHULTZ, ^ CAROLYN GUNN,4 AND JOSEPH G. DOERR^^ ABSTRACT. — We surveyed 366 historical and potential nesting sites for Black Swifts {Cypseloides niger) from 1997 to 2005 in the Southern Rocky Mountains, evaluated their suitability for nesting, and searched for evidence of occupancy. Our surveys located 70 previously undocumented occupied sites, increasing the inventory of sites in the region from 33 to 103. Our results provide a preliminary estimate of Black Swift population size. Comparison of observed colony sizes with those reported in earlier studies suggests little or no change in population levels over the past 50 years. We rated each nest site on conformance to characteristics described in earlier studies. Analysis of 291 site evaluations support a priori assumptions that increasing stream flow, number of potential nest platforms, amount of available moss, shading of potential nest niches, topographic relief of surrounding terrain, and ease of aerial access to potential nest niches contributed to a higher probability the site would be occupied by Black Swifts. Received 14 March 2007. Accepted 11 September 2007. Observers in the late 19^*^ century believed Black Swifts {Cypseloides niger) nested in Colorado (Drew 1881, Bendire 1895), but the first breeding confirmation in the Southern Rocky Mountains did not occur until 1949 when O. A. Knorr found nests at two water- falls near Silverton in San Juan County (Knorr 1950). Knorr searched the mountains of Col- orado from 1949 through 1958 for Black Swift colonies and located —80 nests at 25 sites in 10 counties (Knorr 1953, 1961). In the 38 years (1959-1997) following Knorr’s in- vestigations, only seven additional colonies were located in Colorado, six at waterfalls (Bailey and Niedrach 1965; Boyle 1998; Hur- tado 2002; Robert Righter, pers. comm.) and one in a limestone resurgence cave (Davis 1964). During this same time period, a single colony was located in the Southern Rocky ' Rocky Mountain Bird Observatory, West Office, 337 25 3/4 Road, Grand Junction, CO 81503, USA. 2 White River National Forest, Rifle Ranger Di.strict, 0094 County Road 244, Rifle, CO 81650, USA. ^ San Juan National Forest, Columbine Ranger Dis- trict and Field Office, 367 South Pearl Street, Bayfield, CO 81122, USA. P. O. Box 791, Dolores, CO 81323, USA. ^ White River National Forest, Eagle Ranger Dis- trict, P. O. Box 720, Eagle, CO 81631, USA. ^Current address: 480 Casey Way, Grand Junction, CO 81504, USA. ’Current address: Willamette National Forest, 211 East 7th Avenue, Eugene, OR 97401, USA. ** Corresponding author; e-mail: kmpotter@fs.fed.us Mountain Region in northern New Mexico (Johnson 1990). An ecological pattern emerged from Knorr’s investigations leading him to suggest that “five physical factors were found to be present to a greater or lesser degree in all the colonies” (Knorr 1961:167). These five fac- tors included (1) the presence of water — “from a rushing torrent to a mere trickle,” (1961:168), (2) high relief — “a commanding position above the surrounding terrain,” (1961:168), (3) inaccessibility to terrestrial marauders, (4) darkness — Knorr “never found an occupied nest upon which the sun shone,” (1961:168), and (5) unobstructed aerial ac- cess— the birds “will not fly through a maze of trees to reach their nests” (1961:168). Lat- er, after evaluating some apparently suitable but unoccupied sites, Knorr added a sixth physical factor of nest sites: the presence of niches in rock for nest placement (Knorr 1993). Subsequent studies have generally con- curred with Knorr’s analysis of Black Swift nesting characteristics (Hunter and Baldwin 1962, Foerster 1987). The objectives of our paper are to: present ( 1 ) the results of exten- sive investigations of potential Black Swift nest sites in the Southern Rocky Mountains, and (2) a model for predicting Black Swift occupancy based on characteristics of known occupied sites. METHODS Field Methods. — The U.S. Forest Service (UwSFS) designated the Black Swift as a sensi- 331 332 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 2, June 2008 TABLE 1. Black Swift surveys in Colorado and northern New Mexico, 1997-2005. Prior to these in- vestigations, 32 colony sites had been located in Col- orado and one in New Mexico. Multiple evaluations were conducted at many sites. State Year Sites in data base New sites surveyed Total sites surveyed New colonies found Colorado 1997 NA 2 4 2 Colorado 1998 NA 41 52 5 Colorado 1999 NA 37 56 5 Colorado 2000 NA 60 106 5 Colorado 2001 311 58 119 21 Colorado 2002 352 76 96 18 Colorado 2003 417 71 149 9 Colorado 2004 426 6 12 3 Colorado 2005 427 1 3 0 Subtotal 427 352 597 68 New Mexico 2003 16 9 9 1 New Mexico 2004 21 7 8 1 Subtotal 21 17 17 2 Totals 448 369 614 70 tive species in the Rocky Mountain Region in 1993 citing the need for more information about its distribution and abundance. The USFS sub- sequently, in cooperation with the Colorado Natural Heritage Program, canvassed museums and available literature and interviewed knowl- edgeable individuals to compile an inventory of historically active nest sites, currently active nest sites, potential nest sites, and incidental Black Swift observations. The list of potential nest sites included waterfalls from Conly’s Wa- terfalls of Colorado (1993), Bums’ Colorado Ice Climber’s Guide (1997), 7.5-minute USGS topographic maps, and National Forest visitor maps. Surveyors discovered many more poten- tial sites as they traveled to the waterfalls on the initial list. Field verification began on a small scale in 1997 and, in 1998, the Rocky Mountain Bird Observatory (RMBO) joined USFS and great- ly expanded the survey effort (Table 1). From 1998 through 2005, USFS and RMBO per- sonnel and trained volunteers surveyed and evaluated potential Black Swift sites in Col- orado. In 2003 and 2004, RMBO, under con- tract with the New Mexico Department of Game and Fish, conducted surveys in New Mexico. We included sites in northern New Mexico that are within the Southern Rocky Mountain Physiographic Region. We did not conduct any surveys in the small portion of this physiographic region that extends into southern Wyoming. Surveyors field-tested a Black Swift survey and monitoring protocol during 1998 and 1999 to evaluate it as a status and population trend- monitoring tool. This field protocol, with minor revisions through the course of investigations (Schultz and Levad 2001, 2003), guided our sur- veys. Surveyors evaluated each site for suitabil- ity for Black Swift occupancy using a rating scale based upon, but not identical to, the char- acteristics described by Knorr (1953, 1961, 1993). Six habitat parameters at each site were individually assigned a numeric rating ranging from 1 to 5 based on increasing abundance or extent (Table 2). The six habitat parameters were: FLOW (the amount of flowing surface water during late summer), RELIEF (the extent of relief or prominence over surrounding terrain from the top of the site), ACCESS (the extent of aerial access to or from the nest niches), SHADING (the extent of shading of nest nich- es), NICHES (the number of suitable nest niches inaccessible to ground predators), and MOSS (the amount of moss available at the nest site). We conducted an intensive training session for surveyors each year to standardize interpretation of the subjective descriptors of nest site char- acteristics. Surveyors searched each potential site for nests and other evidence of Black Swift oc- cupancy, such as the presence of urates. They recorded the maximum number of adult swifts seen simultaneously flying about or roosting at the site to develop an estimate of minimum colony size. Observers remained at the site until dark when feasible to detect and/or count adults arriving to roost. Sites were rated as occupied if any of three conditions were met: (1) a nest containing an incubating Black Swift adult, egg, or chick was observed; (2) a used Black Swift nest was observed; or (3) Black Swifts were observed coming to roost at a suitable nest location. We cautiously applied the second criteria because Cordilleran Flycatcher {Empidonax occiden- talis) and Townsend’s Solitaire (Myadestes townsendi) nests can be confused with those of Black Swifts. Nests of these three species were distinguished by structure, placement, and materials. The third criterion does not meet the standard for confirmation of nesting for most species, but we designated these sites TABLE 2. Rating scale for Black Swift colony site characteristics. Levad et al. • BLACK SWIFTS IN SOUTHERN ROCKY MOUNTAINS 333 C/D •S ^ C t+H o ^ C n ^ C O o 1) O O XU < 6 S 3 Cd O U (/) ^ ^ V o S O o ^ S .2 V. a c T3 c/3 (U O c/3 o c/3 rt O .£ £ CJ C/5 OJ c/) _C c/5 = ^ (U o I = n (U 3 tr (U >> c R o O- ,H OJ I .3 ^ -J c/3 — C/3 "3 0.50. In parentheses is the percent of correc predictions with P > 0.80. J wd JhnBtSm and Anderson limy The ratio of shows the odds that Model « is better than Model b. given the independent variables and data set selected for testing. although there is some uncertainty concerning the location of a few of those sites due to lack of documentation. We confirmed continued occupancy at 23 of these 24 sites nearly 50 years after their discovery. We surveyed all of the Southern Rocky Mountain sites discov- ered by others between 1950 and 1996 and confirmed continued occupancy at five of the seven sites in Colorado as well as at the lone site in New Mexico. We were unable to con- duct evening watches at any of the three his- torical sites where we failed to detect swift occupancy and cannot conclude these sites are vacant. The historically documented size of Black Swift colonies in the Southern Rocky Moun- tains ranges from a single pair to 18 pairs. The largest known colony, averaging 1 1 active nests, is in Box Canyon Falls in Ouray Coun- ty, Colorado, a site that was discovered by Knorr in 1950. Some small colonies of one or two nests appear to be intermittently occupied, but larger colonies seem to be occupied an- nually and to remain relatively stable in size. Our surveys produced an estimate of 2.5 pairs per occupied site based on counts of adults and nests. If those counts are accurate, the total Black Swift breeding population at documented occupied sites in the Southern Rocky Mountains (/? = 103) is -500 breeding adults. If the rate of occupancy at presently identified potential sites that have not yet been surveyed {n — 58) is similar to those that have already been surveyed, those sites may sup- port an additional 75 breeding adults. If our model is accurate in predicting occupancy at sites where surveys were inconclusive {n — 76), those sites could support another 300 breeding adults. Although our catalog of po- tential nest sites contains all published and mapped waterfalls in the region, our field ex- perience indicates there are a number of po- tential sites that appear neither in publications nor on maps. A review of topographical maps in the region, focusing on permanent steams flowing through steep gradients, produced es- timates ranging from 105 to 420 additional potential sites that remain to be catalogued. These sites may support another 200-750 birds. These calculations produce an upper population estimate of —1,000—1,600 breed- ing birds, which is quite similar to the esti- mate of 1,400-1,600 breeding birds in Colo- rado provided by Boyle (1998). Our analysis of the six habitat characteris- tics at Black Swift nesting sites supports Knorr’s conclusions about their relationship to swift occupancy. Some of these characteristics may be by-products of nest site selection rath- er than nesting requirements (Marm and Stiles 1992, Marm 1997). Our most parsimonious model supported the a priori assumptions that volume of stream flow, number of potential nest platforms, amount of moss cover, extent of shading of potential nest platforms, extent of topographic relief, and ease of aerial access to the potential nest sites each significantly contributed to a higher probability the site would be occupied by Black Swifts. Discus- sion of nest requirements by Marm (1997) suggests a simpler model using only FLOW and NICHES. This model correctly predicted occupancy for 84.2% of the sites in our study. The Akaike weights suggest the global model LevacI et al. • BLACK SWIFTS IN SOUTHERN ROCKY MOUNTAINS 337 was about 10-' times more likely to be the correct model than this two-variable model based on our data set. The best model predicted that three occu- pied sites had <20% chance of being used. These are small sites, each hosting a single nest. Occupancy at one site appears to be in- termittent and the other two were the only oc- cupied sites that received a rating of “2” (of a maximum possible score of “5”) for NICH- ES. All three sites have other, larger colonies nearby and may represent sub-optimal habitat sites occupied only when all available nest niches at the primary sites are in use. It is possible that some similar sites we counted as unoccupied are actually used intermittently and we failed to detect empty nest structures. Black Swift nest site selection provides sev- eral advantages. The inaccessibility of nest niches makes nests and eggs nearly invulner- able to terrestrial predation. This low preda- tion rate enables relatively high nest success (72%; Hirshman et al. 2007) and maintenance of population stability (Hunter and Baldwin 1962, Foerester 1987, Lowther and Collins 2002, Hirshman et al, 2007) despite a rela- tively low reproductive rate (1 egg/clutch and 1 clutch/year). Temperature parameters have not been measured at the nest, but the shaded, damp sites are cool and likely have minimal temperature fluctuation. These factors may slow the metabolism of nestlings and permit adults to leave them unattended for long pe- riods during their wide-ranging foraging flights (Lowther and Collins 2002). The high relief above surrounding terrain typical of most nest sites enables adults to reach high altitudes at which they forage quickly and with little effort. CONSERVATION IMPLICATIONS A common but largely untested hypothesis in avian ecology is that animals tend to select habitats that increase fitness, survival, and re- productive success. Our best model predicted that <2% of the 191 waterfalls where we failed to detect Black Swifts had a greater than 80% probability of occupancy. If the sites sampled in our study are representative of all potential sites in the region, our model results suggest that most of the highest quality swift nesting habitat is already occupied, and pop- ulation size of Black Swifts may be limited by suitable nesting habitat. Locating and pro- tecting the limited number of suitable nesting sites that are currently available, and identi- fying their occupancy status, should remain a high priority for conservation of this species. In Colorado, low population numbers have resulted in the Black Swift being designated a sensitive species on National Forest lands and listed as a vulnerable breeder by the Col- orado Natural Heritage Program. The U.S. Forest Service clearly carries the greatest con- servation responsibility for this species in the Southern Rocky Mountains. Within Colorado, 73% of all occupied Black Swift sites are on National Forest lands. Colorado’s San Juan National Forest hosts 26% of the state’s known occupied sites while the White River National Forest and Uncompahgre National Forest each host 18%. Two of the three con- firmed sites in New Mexico are on National Forest lands. Our model should be helpful for identifying which sites are most likely to pro- vide nesting habitat and for prioritizing survey efforts among sites where occupancy has not been ascertained. Our field survey protocol appears to be effective and the model predicts the probability of site occupancy with confi- dence. ACKNOWLEDGMENTS Colorado data for this study were collected through the Rocky Mountain Bird Observatory’s Monitoring Colorado’s Birds program, which is supported by the Colorado Division of Wildlife, U.S. Forest Service, Bureau of Land Management, and National Park Ser- vice. The Rocky Mountain Bird Observatory collected New Mexico data under contract with the New Mexico Department of Fish and Game. David Khaliqi provided early assistance with statistical analysis. Data were col- lected by 78 field technicians and volunteers; the au- thors especially thank J. P. Beason and K. D. Behrens for many days of held work, and arduous hiking and climbing, as well as for providing insights and advice on the manuscript. The authors thank Allison Cariveau and J. R. Young for reviewing earlier drafts of the manuscript; their careful reading and suggestions were invaluable. LITERATURE CITED Baii.hy, a. M. and R. j. NiiiDRACii. 1965. Birds of Colorado. Volume 2. Denver Museum of Natural History, Denver, Colorado, U.SA. BtiNDiKt;, C. 1895. Life histories of North American birds. U.S. National Museum Special Bulletin 3: 17.5-177. Bovi.h, S. 1998. Black Swift (Cypscloidcs nigcr). Pag- 338 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 es 236-237 in Colorado breeding bird Atlas (H. E. Kingery, Editor). Colorado Bird Atlas Partner- ship and Colorado Division of Wildlife, Denver, USA. Burnham, K. P. and D. R. Anderson. 2002. Model selection and multimodel inference: a practical in- formation-theoretic approach. Second Edition. Springer- Verlag, New York, USA. Burns, C. M. 1997. Colorado ice climber’s guide. Chockstone Press Inc., Evergreen, Colorado, USA. Collins, C. T. and K. S. Foerster. 1995. Nest site fidelity and adult longevity in the Black Swift {Cypseloides niger). North American Bird Bander 20:11-14. CONLY, M. 1993. Waterfalls of Colorado. Pruett Pub- lishing Company, Boulder, Colorado, USA. Davis, D. G. 1964. Black Swifts nesting in limestone cave in Colorado. Wilson Bulletin 76:295—296. Drew, F. M. 1881. On the birds of San Juan County, Colorado. Bulletin of the Nuttall’s Ornithological Club 6:127-148. Foerster, K. S. 1987. The distribution and breeding biology of the Black Swift {Cypseloides niger) in southern California. Thesis, California State Uni- versity, Long Beach, USA. Hirshma.n, S. E., C. Gunn, and R. G. Levad. 2007. Breeding phenology and success of Black Swifts in Box Canyon, Ouray, Colorado. Wilson Journal of Ornithology 119:678-685. Hunter, J. E. and P. H. Baldwin. 1962. Nesting of the Black Swift in Montana. Wilson Bulletin 74: 409-416. Hurtado, P. 2002. Probable Black Swift {Cypseloides niger) nesting colony in the Wet Mountains, Pueb- lo County. Journal of the Colorado Field Orni- thologists 36:60-61. Johnson, P W. 1990. Black Swift {Cypseloides niger) nesting in the Jemez Mountains of New Mexico. New Mexico Ornithological Society Bulletin 18: 13-15. Kleinbaum, D. G. 1994. Logistic regression: a self- learning text. Springer- Verlag, New York, USA. Knorr, O. a. 1950. First breeding record of the Black Swift {Nephoecetes n, borealis) in Colorado. Auk 67:516. Knorr, O. A. 1953. The geographical and ecological distribution of the Northern Black Swift in Colo- rado. Thesis, University of Colorado. Boulder, USA. Knorr, O. A. 1961. The geographical and ecological distribution of the Black Swift in Colorado. Wil- son Bulletin 73:155—170. Knorr, O. A. 1993. Black Swift nesting site charac- teristics: some new insights. Avocetta 17:139- 140. Lowther, P. E. and C. T. Collins. 2002. Black Swift {Cypseloides niger). The birds of North America. Number 676. SAS Institute. 2001. SAS enterprise guide. Version 8. Second Edition. Cary, North Carolina, USA. Marin, M. 1997. Some aspects of the breeding biology of the Black Swift Wilson Bulletin 109:290—306. Marin, M. and E G. Stiles. 1992. On the biology of five species of swifts {Apodidae, Cypseloidinae) in Puerto Rico. Proceedings Western Foundation of Vertebrate Zoology. 4:287-351. Schultz, C. and R. Levad. 2001. Black Swift moni- toring protocol. Unpublished report. USDA, For- est Service, San Juan National Forest, Bayfield, Colorado, USA. Schultz, C. and R. Levad. 2003. Black Swift moni- toring protocol. Unpublished report. USDA, For- est Service, San Juan National Forest, Bayfield, Colorado, USA. The Wilson Journal of Ornithology 120(2):339— 344, 2008 NEST REUSE BY VERMILION FLYCATCHERS IN TEXAS KEVIN S. ELLISON ' 2 ABSTRACT— Vermilion Flycatchers (Pyrocephalus rubinus) were documented to reuse their nests within a single breeding season in south Texas. I recorded a consistent, low frequency (12%, n = 250 clutches) of nest reuse during each of four seasons. Nest survival was greater {P < 0.001) for reused nests than for newly constructed nests and the abundance of ectoparasites was low overall for both types of nests. The main advantage for Vermilion Flycatchers reusing nests was greater nesting success as nest reuse was associated with reduced nest predation. Nest reuse also increased time available for nesting attempts as a flycatcher would save 8 days (9%) (during three attempts) of the time typically available for nesting in south Texas. Reported nest reuse was common (>25% of a family’s members) among 5 families of passerines. Aspects of the life histories of these groups support hypotheses for the observed nest reuse behavior among Vermilion Flycatchers. Received 3 Feb- ruary 2007. Accepted 29 September 2007. Most passerine birds construct a new nest for each breeding attempt despite the potential for reusing their own nests. Strategies of nest construction or reuse differ in costs as con- struction of new nests requires time and en- ergy. Reuse of nests or nest sites can be as- sociated with costs such as greater exposure to ectoparasites (Mpller 1994, Brown and Brown 1996, Mpller and Erritzpe 1996, Stan- back and Dervan 2001), predators (Styrsky 2005, Yeh et al. 2007) or adverse microcli- mates (e.g., where conditions differ since time of construction). Most studies of nest reuse have focused on birds that use stable and/or complex nesting structures (Lindell 1996) with an emphasis on reuse of nesting struc- tures between years (Barclay 1988, Cavitt et al. 1999, but see Styrsky 2005). My objectives were to: (1) compare the suc- cess of reused and newly constructed nests of Vermilion Flycatchers {Pyrocephalus rubi- nus), and (2) explore whether ectoparasites af- fected nest reuse. I also reviewed the literature on nest reuse within a single season to ex- amine the relative frequency of this behavior and to identify characters associated with nest reuse (e.g., complex nest structures, habitats, etc.). METHODS Study Site. — I monitored 250 Vermilion Flycatcher nests at Fort Clark Springs, a pri- ' Department of Biological Sciences, University of Manitoba, Winnipeg, MB R3T 2N2, Canada. ^ Current address: Department of Forest and Wildlife Ecology, 1630 Linden Drive, University of Wisconsin, Madison, WI 53706, USA; e-mail; ksellison@wisc.edu vate residential community in Kinney County, Texas (29° 18' N, 100° 43' W) during 1999- 2002. The site consists of manicured lawns, huisache {Acacia minuata), and honey mes- quite {Prosopis glandulosa), as well as ripar- ian woodland of pecan {Carya illinoensis) and live oak {Quercus virginiana). Field Methods. — Nests were located sys- tematically by searching a 40-ha area with the entire area searched approximately every 3 days. Success of nest searching was calibrated with bi-weekly spot map censuses (IBCC 1970) of all birds within a 27-ha subset of the area. This method allowed targeting territories of males for which a nest had not been located as well as facilitated accurate estimation of densities of singing males and nesting pairs. Nest Monitoring. — Each Vermilion Fly- catcher nest was checked daily until the end of laying after which nest checks were at 1-4 day intervals. Nests were checked from the ground with a mirror attached to an ex- tendable pole or from a ladder. Birds were not banded and I could not directly identify whether nest reuse involved the birds that con- structed the nest. Thus, I relied upon detailed, bi-weekly spot-maps of songbird territories (IBCC 1970) and egg laying dates to examine the likelihood the same female reused the nest. I used the difference of the earliest and latest hrst egg dates for each season to esti- mate breeding season length. I calculated the interval between nest attempts, for reused nests, as the difference between the date of fledging and the first egg date of the subse- quent clutch. I excluded data for two reused nests in 2()()() because a second attempt may 339 340 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 have been missed; the intervals were 38-45 days. I also excluded data from third attempts (n = 3) using the same nest. Daily probability for nesting success at the egg and nestling stages was calculated using the Mayfield (1975) method with the Hensler and Nichols (1981) correction. Nest success rates for each nesting strategy (reuse vs. new nest) were compared using nest data from all 4 years and program CONTRAST (Hines and Sauer 1989). Nests reused more than once were considered as independent data points for reuse. I also compared clutch sizes of re- used and new nests to identify any tradeoff with construction and egg production. A random sample of nests was examined for ectoparasites or their presence (e.g., pu- paria or eggs). Nests were collected 2-5 days after they became inactive. Twenty-one nests were collected; four which had been reused and 17 new nests. Three of the reused nests had been successful and one failed at the egg stage. One of the new nests failed at the egg stage, two failed at the nestling stage, and 14 were successful. Each nest was sealed in a plastic bag upon collection and frozen until the contents were examined in the laboratory. Each bag was weighed before disassembling each nest in a Petri dish. A 320X dissecting scope was used to scan all nest materials and contents. Nest materials were separated by categories of twigs, feathers, grass, and moss/ lichen. Voucher specimens of all arthropod materials were collected and preserved in eth- anol. Specimens were identified to Order and/ or Family by T. A. Galloway, Department of Entomology, University of Manitoba, Winni- peg. Nest mass was compared between reused and new nests to ascertain whether reused nests were larger. Literature Review.— relative frequency of nest reuse among passerine bird families was examined by surveying the literature for records of nest reuse. Both the “Nesting-reuse of old nests” section in The Birds of North America series (1999—2002, Volumes 1—18) and the Searchable Ornithological Research Archive (SORA) website (http;//elibrary. unm.edu/sora) were searched. Analysis by bird families was limited to those with two or more species common to North America and considered only species that build open- cupped nests. Species were excluded for which adequate (>20 nests) published nesting data did not exist. RESULTS Vermilion Flycatchers reused nests, pre- sumably their own, which had been successful that season for 12% of reproductive attempts {n = 250 clutches). I recorded the ultimate fate for 27 of these attempts; 19 were suc- cessful and 8 failed due to predation. Vermilion Flycatchers laid eggs shortly af- ter young from a previous brood fledged at 26 nests (3 nests were used 3 times). First egg dates ranged from 1 April to 17 July; mean (± SE) breeding season length was 94.0 ± 6.6 days {n = 3). The mean span was 6.6 ± 3.3 days at the 25 reused nests where the length between fledging and the first egg of the next clutch was known. Detailed maps of territories and comparison of laying dates suggested nests were reused by the same females that constructed the first nests. No nests were known to be used between years and nests from the previous year were not found. The daily survival rate for reused nests was higher (x^ = 30.49, P < 0.001) than for those newly constructed (Fig. 1). Nest failure was primarily due to predation (98%, 47 nests of known fate) as brood parasitism was infre- quent (3% of 250 nests were parasitized by Brown-headed Cowbirds [Molothrus ater]). Clutch size (x ± SE) did not vary between seasons among nests used once (1999: 3.07 ± 0.73 eggs, n = 14, range = 2-5; 2000: 2.95 ± 0.50 eggs, n = 41, range = 2-5; 2001: 3.10 ± 0.54 eggs, ^2 = 51, range = 2-4; 2002: 2.93 ± 0.37 eggs, n = 29, range = 2-4; Kruskal- Wallace x^3 = 3.59, P = 0.31). Mean clutch size did not differ between nests used once and those used again (Mann- Whitney U14135 = 891.50, P = 0.63). Three of the 17 reused nests where clutch size was known contained larger subsequent clutches. There was little evidence of activity of ec- toparasites in the 21 nests I examined. One nest was infested by mites (>2,500 individu- als, Dermanyssus spp.) and three nests con- tained single Dipteran puparia. The mite-in- fested nest fledged at least three young. Both nests that failed at the egg stage contained no arthropods. Mite infestation (>500 mites) was detected at one Vermilion Flycatcher nest dur- ing nest monitoring. Neither nest success nor Ellison • VERMILION FLYCATCHER NEST REUSE 341 m CO 0.990 n (j) (/) CD O O 0.988 - 25% of members of the: Lanidae (50%, n = 2), Turdidae (33%, r? = 10), Mim- idae (33%, n = 10), Bombycilidae (50%, n = 2), and Icteridae (30% of 20 spp.) (Table 1). DISCUSSION Twenty-nine (12%) of 250 Vermilion Fly- catcher clutches were in reused nests during the same season the nests were constructed. Nest reuse by this species had not been re- corded previously (Wolf and Jones 2000:10). Reused nests fared better than those newly constructed. Only successful nests were re- used and there was little evidence for costs associated with ectoparasites at reused nests. Reuse apparently offers a beneht of time for producing more young and nests were reused up to three times per season. Vermilion Fly- catcher nests require a minimum of 4 days for construction (K. S. Ellison, unpubl. data). Thus, nest reuse saves at least 4 days per nest attempt; a total savings of 8 days (8-9% of a typical breeding season) when attempting three clutches. Any benefit of time savings would fluctuate between years, as breeding season length generally reflects rainfall, which is extremely variable (±5 to 6 on the Palmer Drought Severity Index, 1698-1980) in south Texas (Stable and Cleavland 1988). This could be important as the Vermilion Flycatch- er has been suspected of being double-brood- ed (Wolf and Jones 2000). The habitat suitability hypothesis (Fretwell and Lucas 1969) predicts that fecundity should be greater in the central area of a spe- cies range. Texas represents the northeastern extent of breeding for the Vermilion Flycatch- er (Wolf and Jones 2000). Thus, nest reuse in Texas may reflect greater selection for more nesting attempts than elsewhere. Limited nest- ing data from Argentina in the central portion of the Vermilion Flycatcher breeding range suggest that daily nest survival is lower (0.932, n = 23 nests) (Mason 1985). More complete fecundity data, particularly those in- corporating breeding season length and num- ber of renesting attempts by individuals, are required to fully test this hypothesis. Regional comparisons of data for reuse of nest materials for construction of new nests (Taylor and Hanson 1970, Carothers 1974) may clarify factors influencing renesling. Nest reuse, although rare among open cup nesting passerines (Clark and Mason 1985), is more common among the Lanidae, Turdidae, Mimidae, Bombycillidae, and Icteridae than for most other avian families in North Amer- ica (Table 1 ). Shrikes and waxwings stand out 342 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 TABLE 1. Species reported reusing nests within a breeding season. Family/Common name Scientific name Reference(s) Tyrannidae (27 species) Western Wood-Pewee Acadian Flycatcher Willow Flycatcher Least Flycatcher Vermilion Flycatcher Contopus sordidulus Empidonax virescens E. train a E. minimus Pyrocephalus rubinus (Curson et al. 1996) (Whitehead and Taylor 2002) (Yard and Brown 1999) (Briskie and Sealy 1988) This paper Laniidae (2 species) Loggerhead Shrike Lanius ludovicianus (Grimes in Bent 1950:135) Corvidae (19 species) Mexican Jay Aphelocoma ultramarina (Brown 1994) Polioptilidae (4 species) Black-tailed Gnatcatcher Polioptila melanura (Farquhar and Ritchie 2002) Turdidae (10 species) Northern Wheatear Townsend’s Solitaire Wood Thrush Oenanthe oenanthe Myadestes townsendi Hylocichla rnustelina (Kren and Zoerb 1997) (Bowen 1997) (Friesen et al. 1999) Mimidae (10 species) Bendire’s Thrasher Curve-billed Thrasher Le Conte’s Thrasher Toxo stoma bend i re i T. curvirostre T. lecontei (Gilman 1915) (Gilman 1909) (Sheppard 1996) Bombycillidae (2 species) Cedar Waxwing Bombycilla cedrorum (Mountjoy and Robertson 1988) Parulidae (51 species) Hooded Warbler Wilsonia citrina (L. J. Petit, unpubl. data) Emberizidae (49 species) Spotted Towhee California Towhee Field Sparrow Lark Sparrow Dark-eyed Junco Song Sparrow Painted Bunting Pipilo maculatus P. crissalis Spizella pusilla Chondestes grammacus Junco hyemalis Melospiza melodia Passerina ciris (Greenlaw 1996) (Kunzmann et al. 2002) (Allaire 1972, Nicholson 1981) (McNair 1984, 1985; K. S. Ellison, unpubl. data) (Nolan et al. 2002, Yeh et al. 2007) (Nice 1937) (Lowther et al. 1999) Icteridae (20 species) Red-winged Blackbird Yellow-headed Blackbird Boat-tailed Grackle Great-tailed Grackle Hooded Oriole Scott’s Oriole Agelaius phoeniceus Xanthocephalus xanthocephalus Quiscalus major Q. mexicanus Icterus cucullatus 1. parisorum (Harms et al. 1991) (Harms et al. 1991) (Bancroft 1987) (Peer and Sealy 2000) (K. S. Ellison, unpubl. data) (Flood 2002) Fringillidae (16 species) House Finch Evening Grosbeak Carpodacus mexicanus Coccothraustes vespertinus (van Riper 1976, Hill 1993) (Gillihan and Byers 2001) among these groups due to their diets and prevalence at relative high latitudes. Both fac- tors can restrict breeding season length and increase selection for nest reuse. Nest reuse among Icteridae was rare relative to sampling effort (Harms et al. 1991) when compared to Tyrannidae and Mimidae. However, few pub- lished studies present data on frequency of nest reuse and it is difficult to assess the num- ber of studies that lacked the capability to de- Ellison • VERMILION FLYCATCHER NEST REUSE 343 tect reuse. I suggest that nest reuse is more common than recorded, especially for the Tyr- annidae. Members of this group differ little in behavior and it is likely that closely related species also reuse nests. For instance, three species of Empidonax flycatchers occasionally reuse their nests suggesting this behavior may occur throughout the genus. Nest building may be costly for females in terms of time and energy that could otherwise be devoted to egg production and provisioning (Conrad and Robertson 1993). Cavitt et al. (1999) assessed between-season nest reuse by Brown Thrashers (Toxostoma rufum) and found those reusing nests saved 3-8 days of nest construction. Few data exist on the en- ergetic demands of nest construction among open-cup nesting songbirds (Biedenneg 1983). However, most species likely expend far less energy building nests than for incu- bation or rearing young (Weathers 1992, 2001). Biedenneg (1983) reported that North- ern Mockingbirds {Mimus polyglottos) did not elevate their energy expenditures during nest building relative to other reproductive stages (pre-building, incubating, etc.). I conclude the main benefit of nest reuse by Vermilion Fly- catchers is enhanced nest success. ACKNOWLEDGMENTS I thank the residents of Ft. Clark Springs for allow- ing this research in their community. I also thank T A. Galloway for laboratory space and insect identifi- cation, R S. Warren and E. D. Klepper for assistance with site logistics, and M. D. Boyd, S. J. A. Coles, N. L. Marino, H. R. McGaha, and Patricia Sullivan for field assistance. Funds for this research were provided by a graduate fellowship and the George A. Lubinsky Memorial Scholarship from the University of Mani- toba, Frank M. Chapman Memorial Awards, and a Discovery Grant from the Natural Sciences and En- gineering Research Council to S. G. Sealy. LITERATURE CITED Allaire, P. N. 1972. Field Sparrow uses abandoned nest for August brood. Auk 89:886. Bancroft, G. T. 1987. Mating system and nesting phe- nology of the Boat-tailed Grackle in central Flor- ida. Florida Field Naturalist 15:1-18. Barclay, R. M. R. 1988. Variation in the costs, ben- efits, and frequency of nest reuse by Barn Swal- lows (Hirundo nistica). Auk 105:53-60. Bent, A. C. 1950. Life histories of North American wagtails, shrikes, vireos, and their allies. U.S. Na- tional Museum Bulletin Number 197. Biedenneg, D. W. 1983. Time and energy budgets of the Mockingbird {Mimus polyglottos) during the breeding season. Auk 100:149-160. Bowen, R. V. 1997. Townsend’s Solitaire (Myadestes townsendi). The birds of North America. Number 269. Briskie, J. V. AND S. G. Sealy. 1988. Nest re-use and egg burial in the Least Flycatcher, Empidonax minimus. Canadian Field-Naturalist 102:729-731. Brown, C. and M. B. Brown. 1996. The evolution of coloniality and Cliff Swallows. University of Chi- cago Press, Chicago, Illinois, USA. Brown, J. M. 1994. Mexican Jay {Aphelocoma ultra- marina). The birds of North America. Number 118. Carothers, S. W. 1974. Breeding ecology and time- energy budget of male Vermilion Flycatchers and comments on the social organization of south- western riparian birds. Dissertation. University of Illinois, Chicago, USA. Cavitt, J. F, A. T. Pearse, and T. A. Miller. 1999. Brown Thrasher nest reuse: a time saving re- source, protection from search-strategy predators, or cues for nest-site selection? Condor 101:859- 862. Clark, L. and J. R. Mason. 1985. Use of nest material as insecticidal and antipathogenic agents by the European Starling. Oecologia 67:169-176. Conrad, K. F. and R. J. Robertson. 1993. Clutch size in Eastern Phoebes {Sayornis phoebe). I. The cost of nest building. Canadian Journal of Zoology 7 1 : 1003-1007. CuRSON, D., C. B. Goguen, and N. E. Mathews. 1996. Nest site re-use in Western Wood Pewees. Wilson Bulletin 108:378-379. Farquhar, C. C. and K. L. Ritchie. 2002. Black-tailed Gnatcatcher {Polioptila melanura). The birds of North America. Number 690. Flood, N. J. 2002. Scott’s Oriole {Icterus parisorum). The birds of North America. Number 608. Fretwell, S. D. and H. L. Lucas Jr. 1969. On terri- torial behavior and other factors influencing hab- itat distribution in birds. I. Theoretical develop- ment. Acta Biotheoretica 19:16-36. Friesen, L. E., V. E. Wyatt, and M. D. Cadman. 1999. Nest reuse by Wood Thrushes and Rose- breasted Grosbeaks. Wilson Bulletin 111:132- 133. Gillihan, S. W. and B. Byers. 2001. Evening Gros- beak {Coccothraustes verspertinu.s). The birds of North America. Number 599. Gilman, M. F 1909. Among the thrashers in Arizona. Condor 1 1 :49-54. Gilman, M. F 1915. A forty-acre bird census at Sac- aton, Arizona. Condor 17:86-90. Greenlaw, J. S. 1996. Spotted Towhee {Pipilo ma- cula t us). The birds of North America. Number 263. Harms, K. F., L. D. Beletsky, and G. H. Orians. 1991. Conspecific nest parasitism in three species of New World blackbirds. Condor 93:967-974. Hensler, G. L. and j. D. Nichols. 1981. The Mayfield 344 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 2, June 2008 method of estimating nesting success: a model, estimators and simulation results. Wilson Bulletin 93:42-53. Hill, G. E. 1993. House Linch {Carpodacus mexican- us). The birds of North America. Number 46. Hines, J. E. and J. R. Sauer. 1989. Program CON- TRAST. USGS, Patuxent Wildlife Research Cen- ter, Laurel, Maryland, USA. http://www. mbrpwrc.usgs.gov/software.html (accessed 1 Leb- ruary 2005). International Bird Census Committee (IBCC). 1970. Recommendations for an international stan- dard for a mapping method in bird census work. Swedish Natural Resources Council, Stockholm. Bulletin of Ecological Research Committee 9:49- 52. Kren, j. and a. C. Zoerb. 1997. Northern Wheatear {Oenanthe oenanthe). The birds of North Ameri- ca. Number 316. Kunzmann, M. R., K. Ellison, K. L. Purcell, R. R. Johnson, and L. T. Haight. 2002. California To- whee (Pipilo crissalis). The birds of North Amer- ica. Number 632. Lindell, C. 1996. Patterns of nest usurpation: when should species converge on nest niches? Condor 98:464-473. Lowther, P. E., S. M. Lanyon, and C. W. Thompson. 1999. Painted Bunting {Passerina ciris). The birds of North America. Number 398. Mason, P. 1985. The nesting biology of some passer- ines of Buenos-Aires, Argentina. Ornithological Monographs 36:954-972. Mayfield, H. E 1975. Suggestions for calculating nest success. Wilson Bulletin 87:456-466. McNair, D. B. 1984. Reuse of other species’ nests by Lark Sparrows. Southwestern Naturalist 29:506- 509. McNair, D. B. 1985. A comparison of oology and nest record card data in evaluating the reproductive bi- ology of Lark Sparrows, Chondestes grammacits. Southwestern Naturalist 30:213-224. M0LLER, A. P. 1994. The Barn Swallow and sexual selection. Oxford University Press, Oxford, Unit- ed Kingdom. M0LLER, A. P. AND J. Erritz0e. 1996. Parasite viru- lence and host immune defense: host immune re- sponse is related to nest reuse in birds. Evolution 50:2066-2072. Mountjoy, D. j. and R. J. Robertson. 1988. Nest construction tactics in the Cedar Waxwing. Wilson Bulletin 100:128-130. Nice, M. M. 1937. Studies of the life history of the Song Sparrow. Volume 1 . Transactions of the Lin- naean Society. New York, USA. Nicholson, C. P. 1981. Likely nest reuse by a Eield Sparrow. Journal of Field Ornithology 52:235. Nolan Jr., V., E. D. Ketterson, D. A. Crista, C. M. Rogers, E. D. Clotfelter, R. C. Titus, S. J. SCHOECH, AND E. Snajdr. 2002. Dark-eyed Junco (Junco hyemalis). The birds of North America. Number 716. Peer, B. D. and S. G. Sealy. 2000. Conspecific brood parasitism and egg rejection in Great-tailed Grack- les. Journal of Avian Biology 31:271-277. Sheppard, J. M. 1996. Le Conte’s Thrasher (Toxosto- ma lecontei). The birds of North America. Num- ber 230. Stahle, D. W. and M. K. Cleavland. 1988. Texas drought history reconstructed and analyzed from 1698 to 1980. Journal of Climate 1:59-74. Stanback, M. T. and a. F. Dervan. 2001. Within- season nest site fidelity in Eastern Bluebirds: dis- entangling the effects of nest success and parasite avoidance. Auk 118:743-745. Styrsky, j. N. 2005. Influence of predation on nest- site reuse by an open-cup nesting neotropical pas- serine. Condor 107:133-137. Taylor, W. K. and H. Hanson. 1970. Observations on the breeding biology of the Vermilion Fly- catcher in Arizona. Condor 82:315—319. VAN Riper III, C. 1976. Aspects of House Finch breed- ing biology in Hawaii. Condor 78:224-229. Weathers, W. W. 1992. Scaling nestling energy re- quirements. Ibis 134:142-153. Weathers, W. W. 2001. Erratum. Ibis 143:708. Whitehead, D. R. and T. Taylor. 2002. Acadian Fly- catcher (Empidonax virescens). The birds of North America. Number 614. Wolf, B. O. and S. L. Jones. 2000. Vermilion Fly- catcher (Pyrocephalus ruhinus). The birds of North America. Number 484. Yard, H. K. and B. T. Brown. 1999. Willow Flycatch- er nest reuse in Arizona. Journal of Field Orni- thology 70:211-213. Yeh, P. j., M. E. Hauber, and T. D. Price. 2007. Al- ternative nesting behaviors following colonization of a novel environment by a passerine bird. Oikos 116:1473-1480. The Wilson Journal of Ornithology 1 20(2):345-352, 2008 NATURAL HISTORY AND BREEDING BIOLOGY OF THE RUSTY-BREASTED ANTPITTA (GRALLARICULA FERRUGINEIPECTUS) ALINA M. NIKLISON,>'5 JUAN I. ARETA,' « ROMAN A. RUGGERA,' KARIE L. DECKER," CARLOS BOSQUE,^ AND THOMAS E. MARTIN" ABSTRACT. — We provide substantial new information on the breeding biology of the Rusty-breasted Antpitta {Grallaricula ferrugineipectus ferrugineipectus) from 40 nests during four consecutive breeding seasons at Ya- cambu National Park in Venezuela. Vocalizations are quite variable in G. ferrugineipectus. Nesting activity peaked in April when laying began for half of all nests monitored. The date of nest initiation pattern suggests this species is single-brooded. Both parents incubate and the percent of time they incubate is high (87-99%) throughout the incubation period. The incubation period averaged (± SE) 17.0 ± 0.12 days, while the nestling period averaged 13.37 ± 0.37 days. G. f ferrugineipectus has the shortest developmental time described for its genus. Time spent brooding nestlings decreased as nestlings grew, but was still greater at pin feather break day than observed in north temperate species. The growth rate constant based on mass {k == 0.41) and tarsus length {k = 0.24) was lower than the k for north temperate species of similar adult mass. All nesting mortality was caused by predation and overall daily survival rate (± SE) was relatively low (0.94 ± 0.01 ) yielding an estimated 15% nest success. Received 13 January 2007. Accepted 25 July 2007. Breeding biology and life history traits of most neotropical birds are poorly known and antpittas are no exception (Krabbe and Schu- lenberg 2003, Rice 2005). The small antpitta genus Grallaricula comprises eight species (Krabbe and Schulenberg 2003). Nest descrip- tions and scanty breeding information are available for G. ferrugineipectus, G. nana, G. flavirostris, and G. peruviana (Krabbe and Schulenberg 2003). The most detailed study to date is that by Schwartz (1957) for the Rus- ty-breasted Antpitta (G. f. ferrugineipectus) where the description of nest and breeding bi- ology was based on only three nests in a forest (850-900 m elevation) south of Petare, Edo. Miranda, Venezuela. Information on the exact length of incubation and nestling stages, pa- rental behavior, predation rates, and nestling growth rates remain unknown. Our objective ' uses, Montana Cooperative Wildlife Research Unit, University of Montana, Missoula, MT 59812, USA. 2 CICyTTP-CONICET, Materi y Espana, Diamante (3105), Entre Rfos, Argentina. Grupo FALCO, Calle 1 17 Nro 1725 e/66 y 67, La Plata (1900), Argentina. Departamento de Biologfa de Organismos, Univ- ersidad Simon Bolfvar, Apdo. 89()0(), Caracas 1080- A, Venezuela. Corresponding author; e-mail; alinanik@yahoo.com is to provide data on these previously undoc- umented life history traits and other natural history information for G. f. ferrugineipectus based on detailed field observations. Field work was conducted during four consecutive breeding seasons from 2003 to 2006 at Ya- cambu National Park, Edo. Lara, a montane cloud forest area in north-central Venezuela (09°42'N, 69°42'W; 500-2,200 m eleva- tion). Means are presented with one standard error (SE) unless otherwise noted. OBSERVATIONS Distribution and Habitat. — The Rusty- breasted Antpitta has a disjunct distribution ranging from Venezuela and northern Colom- bia (subspecies ferrugineipectus and rara) to northern Peru and western Bolivia (subspecies leymebambae) (Ridgely and Tudor 1994, Krabbe and Schulenberg 2003). G. f. ferrugi- neipectus in Venezuela has a wide altitudinal distribution, ranging from 250 to 2,200 m (Gi- ner and Bosque 1998) and inhabits tropical and subtropical areas of the Andes of Merida and western Lara, Falcon, Yaracuy, Distrito Federal, and Miranda (De Schauensee and Phelps 1978). We worked on this species in Yacambii from its upper distributional limit at 1,600 m elevation down to 1,300 m. We found it mostly in secondary growth wet forest with a closed upper story (7—15 m) and an under- 345 346 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 N c-> S cr u pH 4i B / ■uiitftnft. 1 2 1 2 Time (sec) FIG. 1. Vocalization spectrograms of G. f. ferrugineipectus. (A) Adult main song (probably male), recorded by J. I. Areta, 13 June 2006. (B) Voice on the nest (unknown gender) from a videotape by L. A. Biancucci, 22 April 2005. Both recordings were during the breeding season at Yacambu National Park, Lara, Venezuela. Spectrograms on the right are reduced in scale to show overall pattern. growth occupied mostly by vine tangles, small tree saplings, and Chamaedorea spp. palms. Foraging. — Individual Rusty-breasted Ant- pittas were consistently observed foraging be- tween 0.5 and 2.0 m above ground, but not hopping on the ground like many Grallaria spp. (Stiles 1992; A. M. Niklison, pers. obs.). We did not quantify foraging behavior, but most foraging involved hopping along branch- es, peering at leaves, and sallying to gather flying insects. Schwartz (1957) also reports birds descending from perches to the ground to catch insects and immediately flying back to the perch. Additionally, we had a single observation of a curious foraging behavior where a bird perched on a small branch and vigorously shook it with both feet to flush in- sects which were then caught by sallying. Vocalizations. — The Rusty-breasted Antpit- ta was vocal throughout the day during the breeding season. Singing activity peaked in the early morning and late afternoon. Adults (presumably only males sing, Schwartz 1957) were recorded singing under natural condi- tions and after playback with their own voice. The loudsong of G. f. ferrugineipectus con- sisted of a series of 14-17 plaintive and whis- tled chevron shaped notes (on the sonogram) that increase slowly in pitch, and then descend at a faster rate and ending in one or two dis- tinctive lower pitched and flatter notes (Fig. lA). The unsolicited loudsong lasted 2 sec and ranged from 1.2 to 2.6 kHz with a high frequency harmonic documented at 14 kHz in some individuals. The large variation in tim- ing and rate of pitch increase prompted Schwartz (1957) to distinguish two loudsongs, and our recordings {n = 8) concurred. The loudsongs seem to grade into each other through many intermediates spanning the two extremes (Schwartz 1957 provides a phonetic notation of the voices), and they might be bet- ter understood as extremes within a single ex- tremely variable loudsong. Spontaneous vo- calizations usually consisted of 14 notes and remained constant after playback or when ex- Niklison et al. • RUSTY-BREASTED ANTPITTA NATURAL HISTORY 347 Bi-weekly date FIG. 2. Percent of nests that were initiated (first egg laid) by date for G. f ferrugineipectus at Yacambu National Park, Lara, Venezuela, March to July, 2003-2006. Only nests for which initiation date was observed are included (n = 24). cited, unlike G. lineifrons in which the num- ber of notes decreased from 21 in unsolicited vocalization to 13-15 notes after playback (Robbins et al. 1994). The loudsong of the nominal subspecies at Yacambu differs strik- ingly from that of the allopatric G. f. leyme- bambae whose loudsong is a monotonous se- ries of whistles (Mayer 2000, Krabbe and Schulenberg 2003). The loudsong of G. f fer- rugineipectus recalls that of G. nana and G. lineifrons in basic pattern, note shape, and fre- quency range differing most notably in pitch, speed, and length (Robbins et al. 1994; J. I. Areta, pers. obs.). We recorded a previously unreported short- er, softer, and relatively unstructured and var- iable song emitted by a bird on the nest during the 2 min that preceded leaving the nest for an off-bout (Fig. IB). This bird was not im- mediately replaced by any other and we sus- pect this might be a contact call eliciting pa- rental switch. A similar soft voice was uttered in a parental switch during incubation in more than one occasion (video recording data). Schwartz (1957) also describes a short single- note call, but we did not hear or record this call at our study site. Nesting Chronology. — The nesting period for this species was thought to start in mid May and limited to the earliest part of the rainy season (Schwartz 1957). However, in Yacambu the rainy season spans from mid- April until mid-July. We worked at this site from the beginning of March until early July each year and found that nest building and egg laying started several weeks before the rainy season and continued through June. The peak in nesting activity occurred at the beginning of the rainy season in the last 2 weeks of April when 54% of nests were initiated (Fig. 2). We found 40 nests between March and July for all years combined. The earliest active nest was found empty on 27 March 2006 and had two eggs by 31 March 2006. The latest date of nest initiation was 7 June 2006. On two occasions, nest building activity was observed within 3 days of predation events and within 10 m of the depredated nests. Given the proximity and timing, we as- sume these reflect re-nesting attempts, but the pattern of initiation observed (Fig. 2) suggests this species is single-brooded; pairs that suc- cessfully fledged young were not observed initiating a new nest in the same season. Clutch Size and Eggs. — All nests that were occupied (/? = 37) had two eggs or nestlings similar to that reported by wSchwartz (1957). Clutch size in Grallaricula varies between one in G. nana and G. peruviana, and two in G. ferrugineipectus, and between one and two 348 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 2, June 2008 LIG. 3. Top view of the nest and egg of G. f. ferrugineipectus, Yacambu National Park. Photograph by A. M. Niklison, 19 April 2005. in G. flavirostris (Holley et al. 2001; Greeney et al. 2004a, 2004b; Greeney and Sornoza 2005). Eggs were short, sub-elliptical to oval in shape, and usually pale greenish or grayish with variously shaped markings in brown shades over the entire surface. The greenish background color of G. f. ferrugineipectus eggs (Fig. 3) is different from the pale coffee- brown with darker brown blotches described for G. nana (Greeney and Sornoza 2005), G. cucullata (Schonwetter 1967), G. flavirostris (Holley et al. 2001), and G. peruviana (Gree- ney et al. 2004a, 2004b). Two clutches had eggs with brownish background color. The pattern of markings closely resembles that of G. nana (Fig. 3; and Fig. 1 in Greeney and Sornoza 2005). We weighed 15 eggs at eight nests between day zero and day 2 of the incubation period using an ACCUFAB portable electronic scale (precision 0.001 g). Individual eggs varied from 2.40 to 3.02 g among the eight nests and averaged 2.78 ± 0.05 g, which was —17% of average adult weight (16.48 ± 0.43 g, n = 10). Egg dimensions averaged 1.88 ±0.13 cm in length and 1.73 ± 0.15 cm in width (Mi- tutoyo digital caliper, precision 0.01 mm, n = 3). Volume was estimated using Hoyt’s (1979) formula (0.51 X length X [width^]) and av- eraged 2.84 ± 0.29 cm^. Nests and Nest Placement. — Nests were un- stable and difficult to remove without destroy- ing them, as is common for G. f. ferruginei- pectus and other Grallaricula (Greeney et al. 2004b, Greeney and Sornoza 2005). Nests were in secondary growth areas in microhab- itats with abundant vine tangles. Construction was simple consisting of a slightly concave twig platform on top of which black rootlets were knitted in a shallow cup finished with thin and brown rootlets without moss. They were usually supported by slender vines or placed on top of a palm leaf from 0.5 to 2.0 m in height, averaging 1.30 ± 0.14 m high {n = 17). Unlike G. peruviana (Greeney et al. 2004b), only one adult participated in nest construction. We measured size of nests for outer diam- eter (from edge to edge), inner diameter (cup), outer height (exterior bottom-to-top), and in- ner height (bottom-to-top of cup) of 20 nests Niklison et al. • RUSTY-BREASTED ANTPITTA NATURAL HISTORY 349 (Fig. 3). Inner diameter averaged 6.0 ± 0.16 cm, outer diameter 10.5 ± 0.30 cm, inner height 2.2 ± 0.15 cm, and outer height 4.36 ± 0.31 cm. Nest shape, size, building mate- rials, and poor support are much like those of G. nana (Fig. 1 in Greeney and Somoza 2005), but differ from G. flavirostris and G. peruviana which attach their mossy nests to branches (Holley et al. 2001; Greeney et al. 2004a, 2004b). Incubation Period and Behavior. — We mea- sured duration of incubation periods, when possible, as the number of days between the last egg laid and the last egg hatched (Briskie and Sealy 1990, Martin 2002). Eggs hatched synchronously in all cases, suggesting that in- cubation began when the last egg was laid contradicting Schwartz’s (1957) suggestion that incubation began with the first egg. The incubation period varied from 16 to 17.5 days and averaged 17.0 ± 0.12 days {n = 4) (also see Schwartz 1957). G. f ferrugineipectus has the shortest incubation period described for the genus; G. flavirostris has an incubation pe- riod of 17-21 days (Holley et al. 2001) and G. peruviana has an incubation period of 20 days (Greeney et al. 2004b). We measured parental nest attentiveness (percent time on the nest incubating) and du- ration of incubation bouts by videotaping nests for 6 to 8 hrs starting within 30 min of sunrise (Martin and Ghalambor 1999, Martin 2002). Both adults share incubation and they exchange bouts on the nest. Nest attentiveness increased mildly (r = 0.47, P = 0.06, df = 16) from day 2 to day 17 of incubation, av- eraging 94.6 ± 1.4% overall. However, the in- crease was due to two nests on day 2 in which both parents stayed off the nest for a long pe- riod in the afternoon and caused lower nest attentiveness estimates (82.3 ± 0.75%). Vari- ability in attentiveness in the first few days of incubation is common in tropical birds (T. E. Martin, unpubl. data). If these two nests are excluded, nest attentiveness did not change from day 2 to day 17 of the incubation period (r = 0.20, P = 0.47, df = 14) and averaged 96.3 ± 0.95%. On and off-bout durations were unrelated to incubation age (both r < 0.3, P > 0.05). On bouts averaged 66.53 ± 4.44 min, n = 17. Nests were nearly contin- uously occupied. A general comparison of nest attentiveness and on-bout/off-bout duration among species of Grallaricula is not possible due to the lack of published data. However, information for G. nana indicates lower incubation attentive- ness (65%, Greeney and Somoza 2005), short- er on-bouts averaging (± SD) 53.8 ± 44.3 min, and longer off-bouts averaging (± SD) 10.6 ± 15.7 min than for G. f. ferrugineipec- tus. The incubation day of these measure- ments was not reported and we cannot provide a more precise comparison. Adults during incubation performed nest maintenance by arranging edge sticks or inner cup rootlets and by bringing material to the nest when switching, as also observed by Schwartz (1957). Adults frequently rotated their bodies and stood to examine and possi- bly rotate the eggs, as also observed in G. nana (Greeney and Somoza 2005). Nestling Period and Brooding Behavior. — We measured the nestling period as the num- ber of days from hatching to fledging. The nestling period varied from 12 to 14 days and averaged 13.37 ± 0.37 days {n — 4) similar to the 13-day period reported by Schwartz (1957) but relatively shorter than the reported 14-16 days for G. flavirostris (Holley et al. 2001). Nestlings are naked at hatching and the gape is deep orange. The juvenile plumage is a thick wooly down that is hair-like rather than feather-like. The plumage tract of the head during development is reduced to a sin- gle narrow track that resembles a scalp-lock. Adults provisioned young an average of 3.77 trips/hr (n = 1) on the first day of the nestling period, 3.91 ± 1.06 (n = 3) on day 2, 4.48 ± 0.07 (n = 2) on day 6, and on day 8 (day the pin feather of primary 8 breaks its sheath) they fed 9.08 times/hr {n = 1). Both parents brought spiders, butterflies, crickets, cockroaches, larvae, and beetles as feeding items (this study; Schwartz 1957). The percent of time spent brooding during the nestling period was averaged across bouts for each nest and then averaged across nests, with the SE reflecting this cross-nest mean. Brooding time averaged 98% (n = 24 bouts, 1 nest) on day 1, 84 ± 3% (n = 68, 3 nests) on day 2, 90 ± 3% (n = 52, 2 nests) on day 6, and 61% (/? = 9, 1 nest) on day 8, pin- break day. Time spent brooding is higher (P < 0.001) than published for some north-tem- 350 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. J20, No. 2, June 2008 (A) Nestling age (days) FIG. 4. Nestling measurements of G. f. ferrugi- neipectus at Yacambii National Park plotted against age: (A) Mass, (B) Tarsus length, and (C) Wing-chord. Growth rate constant (A:) and asymptote (A) are indi- cated for mass and tarsus. perate birds; for example the Song Thrush (Turdus philomelos) broods only 29% of the time on the day that pin feathers break their sheaths (Hill et al. 1999). Nestling Growth. — We measured nestling tarsus length and wing chord (Mitutoyo digital calipers, 0.01 mm precision) and weighed nestlings (ACCULAB digital electronic scales, 0.001 g precision). We estimated growth rate using the formula W(0 = A/{ 1 + where W{t) refers to mass or tarsus length at day t, k is the growth rate constant, A is the estimated asymptotic mass or tarsus length, and is the inflection point of change from accelerating to decelerating growth, fol- lowing Ricklefs (1967). The growth rate con- stant for mass (Fig. 4) was lower than the k for unrelated north temperate passerine spe- cies that averaged k - 0.52 ± 0.05 {n = 4) for a similar adult weight (16.2-17.5 g) based on data in Remes and Martin (2002). Nest- lings (10-13 days of age) did not “play dead” during measurements in contrast to Schwartz (1957). Nestling weight on day 12, the day before normal fledging averaged 14.09 ± 0.39 g = 4) which was 85% of adult body mass. The Rusty-breasted Antpitta fledged with a larger relative body mass (^3 = 3.81, P — 0.032) than similar-sized but unrelated north temperate species 75 ±5%,n = 4 (Remes and Martin 2002). This may reflect the longer (^3 = 10.33, P = 0.002) time spent in the nest compared with north temperate species. Nest Predation and Success. — We did not find all nests during nest-building; 16 nests were found during nest-building, 19 during in- cubation, and 5 during the nestling stage. Con- sequently, we calculated daily survival rates using the Mayfield method (Mayfield 1975, Hensler and Nichols 1981). Three nests were abandoned during building and one was lost due to investigator disturbance; these were ex- cluded from the Mayfield estimates. Twenty- five of the remaining 36 nests were lost to predation and the rest were successful. One nest was depredated during the laying period, 18 during incubation, and 6 during the nest- ling period. Overall daily survival rate was 0.94 ± 0.012 {n = 409.5 exposure days, 36 nests). Daily survival rates were 0.931 ± 0.067 for the laying period, 0.929 ± 0.016 for the incubation period, and 0.958 ± 0.017 for the nestling period. Only 15% of G. f. ferru- gineipectus nests were successful as estimated by the Mayfield method. North temperate pas- serine species have a higher estimated nesting success (tn 6.876, P < 0.001), averaging 42.43 ± 0.040% (n = 18) (Martin 1992). DISCUSSION Our data are typical for many tropical spe- cies, where they exhibit a slower pace of life Niklison et al. • RUSTY-BREASTED ANTPITTA NATURAL HISTORY 351 compared with north temperate species, in- cluding small clutch size, slow development during both incubation and nestling stages de- spite high parental care, and lower nesting suceess (Martin 1996, Martin et al. 2000, Martin 2002, Martin et al. 2007). Contrary to views that these traits might reflect long breeding seasons (reviewed by Martin 1996), the breeding season was similar in length to north temperate systems, and the species did not appear to be multi-brooded suggesting that alternative explanations need to be sought. The behavior of sharing ineubation duties by both males and females is relatively uncom- mon in north temperate passerines but com- mon in a number of endemie tropieal groups (Martin et al. 2007). The high brooding atten- tiveness late into the nestling stage may also be common in tropical endemics, but these be- haviors have been neglected and deserve fur- ther attention. ACKNOWLEDGMENTS We thank M. J. Foguet, L. A. Biancucci, Alison Cox, W. A. Cox, J. R. Lang, Andrew Hoar, A. A. Ma- jewska, and William Goulding for valuable help in the field. We are also grateful to Muttulingam Sanjayan, J. L. Hierro, R. J. Fletcher Jr., and two anonymous reviewers for helpful comments on this manuscript. This study was supported by NSF grants DEB- 9981527 and DEB-0543178 to T. E. Martin. Permit numbers are DM/0000237 from FONACIT; PA-INP- 005-2004 from INPARQUES; and 01-03-03-1147 from Ministerio del Ambiente. LITERATURE CITED Briskie, j. and S. G. Sealy. 1990. Evolution of short incubation periods in the parasitic cowbirds, Mol- othrus spp. Auk 107:789-793. De Schauensee, R. M. and W. H. Phelps. 1978. A guide to the birds of Venezuela. Princeton Uni- versity Press, Princeton, New Jersey, USA. Giner, S. and C. Bosque. 1998. Distribucion altitu- dinal de las subfamilias Grallarinae, Formicariinae y Thamnophilinae (Aves, Formicariidae) en Ve- nezuela. Boletin de la Sociedad Biologica de Chile 69:1 15-121. Greeney, H. F. and F. Sornoza. 2005. The nest and egg of the Slate-crowned Antpitta (Grallaricula nana) with observations on incubation behavior in southern Ecuador. Ornitologia Neotropical 16: 137-140. Greeney, H. F, E. C. Hannelly, and M. Lysinger. 2004a. First de.scription of the nest and vocaliza- tions of the Peruvian Antpitta {Grallaricula pe- ruviana) with a northward range extension. Cotin- ga 21:14-17. Greeney, H. E, R. A. Gelis, E. C. Hannelly, and P. J. DeVries. 2004b. The egg and incubation period of the Peruvian Antpitta {Grallaricula peruviana). Ornitologia Neotropical 15:403-406. Hensler, G. L. and j. D. Nichols. 1981. The Mayfield method of estimating nesting success: a model, estimators and simulation results. Wilson Bulletin 93:42-53. Hill, I. E, B. Cresswell, and R. E. Kenward. 1999. Comparison of brooding patterns between Black- bird {Turdus merula) and Song Thrush {Turdus philomelos). Bird Study 46:122-126. Holley, D. R., C. A. Lindell, M. A. Roberts, and L. A. Biancucci. 2001. First description of the nest, nest site, and eggs of the Ochre-breasted Antpitta. Wilson Bulletin 113:435-438. Hoyt, D. F. 1979. Practical methods of estimating vol- ume and fresh weight of bird eggs. Auk 96:73- 77. Krabbe, N. and T. S. Schulenberg. 2003. Family Formicariidae (ground-an thirds). Pages 682-731 in Handbook of the birds of the world. Volume 8. Broadbills to tapaculos (J. del Hoyo, A. Elliot, and D. Christie, Editors). Lynx Edicions, Barcelona, Spain. Martin, T. E. 1992. Breeding productivity consider- ations: what are the appropriate habitat features for management? Pages 455-473 in Ecology and conservation of neotropical migrants (J. M. Hagan and D. W. Johnston, Editors). Smithsonian Insti- tution Press, Washington D.C., USA. Martin, T. E. 1996. Life history evolution in tropical and south temperate birds: what do we really know? Journal of Avian Biology 27:263-272. Martin, T. E. 2002. A new view for avian life history evolution tested on an incubation paradox. Pro- ceedings of the Royal Society of London, Series B 269:309-316. Martin, T. E. and C. K. Ghalambor. 1999. Males helping females during incubation. Required by microclimate or constrained by nest predation? American Naturalist 153:131-139. Martin, T. E., P. R. Martin, C. R. Olson, B. J. Hei- DiNGER, AND J. J. FoNTAiNE. 2000. Parental care and clutch sizes in North and South American birds. Science 287:1482-1485. Martin, T. E., S. K. Auer, R. D. Bassar, A. M. Nik- lison, AND P. Lloyd. 2007. Geographic variation in avian incubation periods and parental influenc- es on embryonic temperature. Evolution 61: 2558-2569. Mayer, S. 2000. Birds of Bolivia 2.0. Bird Songs In- ternational, Enschede, The Netherlands. Mayfield, H. 1975. Suggestions for calculating nest success. Wilson Bulletin 87:456-466. Reme.^, V. AND T. E. Martin. 2002. Environmental in- fluences on the evolution of growth and devel- opmental rates in passerines. Evolution 56:2505- 2518. Rice, N. H. 2005. Phylogenetic relationships of ant- 352 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 pitta genera. (Passeriformes: Formicariidae). Auk 122:673-683. Ricklefs, R. E. 1967. A graphical method of fitting equations to growth curves. Ecology 48:978-983. Ridgely, R. S. and G. Tudor. 1994. Birds of South America. Volume 2. University of Texas Press, Austin, USA. Robbins, M. B., N. Krabbe, G. H. Rosenberg, R. S. Ridgely, and F. Sornoza M. 1994. Notes on the natural history of the Crescent-faced Antpitta. Wilson Bulletin 106:169-173. Stiles, F. G. 1992. A new species of antpitta (Formi- cariidae: Grallaria) from the eastern Andes of Co- lombia. Wilson Bulletin, 104:389-399. ScHONWETTER, M. 1967. Handbuch der Oologie. Aka- demie-Verlag, Berlin, Germany. Schwartz, P. 1957. Observaciones sobre Grallaricula ferrugineipectus . Boletin de la Sociedad Venezo- lana de Ciencias Naturales 18:42-62. The Wilson Journal of Ornithology 120(2);353-365, 2008 FORAGING ECOLOGY OF PARROTS IN A MODIFIED LANDSCAPE: SEASONAL TRENDS AND INTRODUCED SPECIES GREG D. MATUZAK,' -* M. BERNADETTE BEZY,^ AND DONALD J. BRIGHTSMITH3 ABSTRACT. — We studied the diet and foraging ecology of a community of six psittacines in western Costa Rica. All had a varied diet with clear seasonal changes in preferred food items, mostly due to changes in plant phenology. There was a significant relationship between parrot mass and food types: larger-bodied parrots con- sumed more seeds and smaller-bodied parakeets consumed more fruit pulp. Leaves, bark, and lichen were also consumed by most psittacines. Most parrots consumed more plant species in the dry season when food avail- ability was at its peak. Levins’ niche breath showed varying levels of diet specialization among species and, for some species, variation among seasons. There was less similarity in seasonal psittacine diets when compared to overall diets. Scarlet Macaws {Ara macao) under study were captive raised and released which may have contributed to their narrow diet breadth as they may have lacked the knowledge or experience to exploit addi- tional food sources. Non-native and cultivated species comprised 76% of the diet of Scarlet Macaws, and averaged 28% for all other species. This suggests that foraging parrots may have increased conflicts with humans as landscapes become increasingly modified. Forest restoration strategies should augment the abundance of food species consumed when overall food supply is at its annual low. Received 20 February 2007. Accepted 4 September 2007. Knowledge about diets is fundamental for understanding species’ niches, roles in com- munities, and potential impacts on other spe- cies (Moegenburg and Levey 2003, French and Smith 2005, Munshi- South and Wilkinson 2006). Knowledge of diet is also needed to design effective conservation and manage- ment strategies, and to predict how landscape level changes may affect species (Fitter and Christiansen 1995, Bennett and Owens 1997). Approximately one third of all psittacines are threatened with extinction and with anthro- pogenic changes, many species have become locally or regionally extinct (Collar 1997). However, in some cases parrot species in- crease in abundance with landscape conver- sion and become agricultural pests (Forshaw 1989, Bucher 1992). Overall, the natural his- tory of psittacines is poorly known with little or no information on the diet of over 75% of the recognized species (Collar 1998). This lack of basic diet information poses problems for those who work to understand and con- ' 10090 Skyline Drive, Grass Valley, CA 95945, USA. ^ University of Costa Rica, Department of Biology, San Jose, Costa Rica. ^ Schubot Exotic Bird Health Center, Department of Veterinary Pathobiology, Texas A&M University, Col- lege Station, TX 77843, USA. * Corresponding author; e-mail: gmatuzak@hotmail.com serve threatened parrots, mitigate agriculture damage, and understand the impacts of parrots on vegetative communities (Collar 1998, Moegenburg and Levey 2003). Parrots feed predominantly on seeds, fruit pulp, and flowers along with variable amounts of leaves, bark, nectar, and insects (Forshaw 1989, Sazima 1989, Pizo et al. 1995). Parrots eat ripe and unripe fruits, and many species consume large amounts of immature and ma- ture seeds, making some psittacines effective pre-dispersal seed predators (Desenne 1994, Pizo et al. 1995). Pollination and seed dis- persal have rarely been recorded for parrots (Fleming et al. 1985, Cotton 2001) with most parrot feeding leading to some reduction in plant fitness (Galetti 1993, Ragus-Netto 2005). Identifying and conserving key food resources and their habitats may be vital to the long term conservation of some parrot com- munities. We conducted a 2-year study on the for- aging ecology and diet of a community of six sympatric parrot species in western Costa Rica. Our objectives were to document: (I) the plant species and plant parts consumed by each parrot species, (2) the level of seed pre- dation and frugivory for each species, (3) sea- sonal diet changes, and (4) key resources used by each psittacine species. We also discuss the importance of introduced and cultivated plant 353 354 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 species to foraging parrots in view of the changing landscape throughout the Neotrop- ics. METHODS Study Area.— This study was conducted on the southern Nicoya Peninsula in the Province of Puntarenas, western Costa Rica. The pri- mary research site was Curu Wildlife Refuge (09° 47' N, 84° 56' W). Curu is a private wild- life refuge and working farm of 1 ,492 ha with 70% forested habitats, and 30% pastures and plantations (Schutt and Vaughan 1995). Curu is part of the Peninsular Biological Corridor, a loosely connected network of small forested fragments on the southern Nicoya Peninsula (Vaughan et al. 1994). We collected additional data on Tortugas and Negritos islands, ~2 and 5 km, respectively, from Curu in the Nicoya Gulf. Rainfall totals -200 cm per annum and is strongly seasonal with a wet season from May to November, and a dry season from Decem- ber to April (months with <10 cm of precip- itation). The average temperature throughout the year is 27.3° C (Vaughan et al. 1994). The site is at the boundary of tropical dry forest, tropical moist forest, and tropical premontane life zones (Tosi 1969), and contains a mosaic of mangroves {Avicennia, Rhizophora, Lagun- cularia), dry deciduous forest, semi-decidu- ous forests, mixed coconut (Cocos) forest, beaches, evergreen forest, pastures, and plan- tations of Tectona grandis and Mangifera indica (Vaughan et al. 1994). Six psittacine species occur on southern Ni- coya Peninsula: Scarlet Macaw (Ara macao — 900 g), Yellow-naped Amazon (Amazona au- ropalliata — 480 g), Red-lored Amazon (A. autumnalis — 420 g). White-fronted Amazon (A. albifrons — 230 g). Orange-fronted Para- keet (Aratinga canicularis — 80 g), and Orange-chinned Parakeet (Brotogeris jugular- is — 65 g) (weights from Stiles and Skutch 1989). Scarlet Macaw is considered an endan- gered species (Appendix 1 of the Convention on International Trade in Endangered Species [CITES]). The small population of macaws in the project area was established in an area ab- sent of a wild population in 1999 through a reintroduction program. Macaws observed were a group of 9—12 birds that had been re- leased 4 years prior to the onset of our study (Brightsmith et al. 2005). Yellow-naped Parrot is a threatened species (CITES 2002) and the area contains one of the largest known roost- ing populations of the species in Costa Rica with a minimum of 300 individuals (Matuzak and Brightsmith 2007). Parrot Foraging Observations. — We estab- lished six transects in areas known to be fre- quented by parrots to document psittacine di- ets. Transects averaged 1,000 m in length and were in deciduous and semi-deciduous forests, mixed coconut forest, evergreen (gallery) for- est, mixed mangroves, open pastures, and plantations of T. grandis, M. indica, and mixed citrus (Citrus spp.). We established one transect in each habitat type; however, some transects crossed multiple habitat types. Each transect was surveyed a minimum of three times per week. Each habitat and season re- ceived the same survey effort throughout the study. We walked transects in the morning (630- 1030 hrs) and early afternoon (1400-1800 hrs) during known parrot foraging peaks (GDM, unpubl. data). Data were collected from August 2003 to July 2005. We also re- corded opportunistic observations of parrots foraging at any time of day. The following data were noted whenever parrots were found feeding: date, time, habitat type, species of parrot, number of parrots, species of plant, and plant part consumed. Eruit was considered fruit pulp and not whole fruits. We ascertained whether parrots were consuming fruit pulp, seeds, or both based on evaluating dropped fruits when whole fruits were being con- sumed. Cultivated tree species include native and non-native trees planted by humans (usu- ally in plantations) for the consumption of fruits and seeds, whereas non-native species refer to any species introduced to Costa Rica from another country and can include orna- mentals planted for their flowers. Therefore, some species such as M. indica can be native and cultivated for their fruit. The two groups of islands and adjacent areas to Curu were vis- ited sporadically, and foraging observations in these areas were recorded opportunistically. An observation of one or more parrots feeding was recorded as a single feeding bout; how- ever, if a parrot or group of parrots flew to and fed upon another plant of the same or dif- ferent species, an additional feeding bout was Matuzak et al. • PARROT DIETS IN A MODIFIED LANDSCAPE 355 recorded (Galetti 1993, Wermundsen 1997, Renton 2001). Statistical Analyses. — The standardized Levins’ (1968) niche breadth index was cal- culated from the number of parrots observed feeding on each plant species consumed. Val- ues close to 0 indicates dietary specialization and a value close to 1 indicates a broad diet (Colwell and Futuyama 1971). We analyzed seasonal changes in diet by comparing the Levins’ diet breadth index and number of food plant species consumed between the wet (Jun- Nov) and dry seasons (Dec-May) (Levins 1968; Renton 2001, 2006). Pearson product-moments and simple linear regression were used to test the relationship between body mass and percent seeds, fruit pulp, flowers, and leaves in the diet using the Data Desk software package (Data Descrip- tion Inc. 2006). Jaccard similarity coefficients (7), a statistic used to compare the similarity of plant species in the diets between psitta- cines, were estimated to compare the overall and seasonal diets between the five main par- rot species. All data are presented as mean ± SD. All statistical tests used a 0.05. RESULTS Foraging Ecology and Diet. — We recorded 1,159 foraging bouts by six species represent- ing all of the psittacines known from the area: Scarlet Macaw (52%), Yellow-naped Parrot (10%), White-fronted Parrot (9%), Orange- fronted Parakeet (10%), Orange-chinned Par- akeet (19%), and Red-lored Parrot (<1%; Ta- bles 1-3). Red-lored Parrots were rare at the site and only observed foraging three times; this species was eliminated from further anal- ysis. We observed psittacines foraging on 61 plant species from 25 families (x = 3\ ±2 plant species and x = 18 ± 1.3 plant families per psittacine species, n = 1,159 foraging bouts) (Tables 1-3). Twelve plant species were non-native and/or cultivated in Costa Rica (54% of all feeding bouts) (Tables 1-3). The number of feeding bouts per food plant species ranged from 1 to 238 {x = 19 ± 35, n = 1,159). Terminalia catappa (Family: Combretaceae) represented 2 1 % of the for- aging bouts. The nine next most common diet species together represented an additional 46% of the total: Delonix regia — 9% (Family: Caesalpiniaceae), Mangifera indica — 6% (Family: Anacardiaceae), Tectona grandis — 6% (Family: Verbenaceae), Pithecellobium saman — 5% (Family: Fabaceae), Cocos nuci- fera — 5% (Family: Palmae), Guazuma ulmi- folia — 4% (Family: Sterculiaceae), Elaeis gui- neensis — 4% (Family: Palmae), Bombacopsis quinata — 4% (Family: Bombacaceae), and Spondias mombin — 3% (Family: Anacardi- aceae). Psittacines ate seeds (54%), fruit pulp (24%), flowers (10%), leaves (7%), bark (4%), and lichen (<1%). Excluding Scarlet Macaws, the proportions were fruit pulp (38%), seeds (34%), and flowers (16%). All parrots con- sumed both ripe and unripe seeds and they fed on both cultivated and non-native species. The parrot community consumed 41 species dur- ing the wet season {n = 572 foraging bouts. Levins’ diet breadth = 0.364) and 50 species during the dry season {n — 587 bouts. Levins’ diet breadth = 0.205), an 18% increase in the dry season. Scarlet Macaw. — This species ate 32 food plant species from 15 families {n = 600 for- aging bouts. Levins’ diet breadth = 0.118) (Table 1). T. catappa (38%) and D. regia (16%) were the most commonly consumed species; both are non-native. The diet was comprised of seeds (73%), fruit pulp (10%), bark (6%), flowers (5%), leaves (5%), and li- chen (<1%). Non-native and cultivated plant species represented 76% of macaw foraging bouts. Macaws ate 27 food plant species in the dry season {n = 262 foraging bouts. Lev- ins’ diet breadth = 0.139) and 18 in the wet season {n = 338 foraging bouts. Levins’ diet breadth = 0.216), a 33.3% increase in the dry season. Amazona Parrots. — Yellow-naped Parrots ate 34 food plant species from 21 families (/? == 121 foraging bouts. Levins’ diet breadth = 0.388) (Table 2) and White-fronted Parrots ate 36 food plant species in 21 families (/? = 108 foraging bouts. Levins’ diet breadth = 0.225) (Table 2). The diet of Yellow-naped Parrots was comprised of seeds (61%), fruit pulp (23%), flowers (7%), leaves (7%), bark (<1%), and lichen (<1%) while the diet of White-fronted Parrots was comprised of seeds (37%), fruits (31%), flowers (26%), leaves (5%), and bark (2%). Non-native and culti- vated plant species represented 37% of Yel- 356 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 2, June 2008 TABLE 1. Loraging observations for Scarlet Macaws. Species marked with vated, and *** are both non-native and cultivated. Parts eaten are coded as S = bark, L = leaves, L = flowers, and Li = lichen. * are non-native, ** are culti- = seeds, LP = fruit pulp, B = Family/Species Part eaten Total # individuals Feeding bouts Months Anacardiaceae Anacardium excelsum Spondias mombin Spondias purpurea Mangifera indica** s LP,B LP LP 10 66 2 10 4 27 1 6 Leb-Mar Jul-Sep Mar-Apr Leb-May Bignoniaceae Tabebuia rosea L 2 1 Leb-Mar Bombacaceae Ceiba aesculifolia Ceiba pentandra Bombacopsis quinata Ochroma pyramidale L,S,B,L S,B L,S,B L 9 4 7 12 5 2 5 4 Leb-Apr Leb-Apr Dec-Leb Leb-May Caesalpiniaceae Schizolobium parahybum Delonix regia^ Cassia grandis S,L,B L,S,L L,S,B 9 206 8 6 98 4 Apr-May Sep-Dec Dec Combretaceae Terminalia catappa"^ L,S,L 498 230 All year Lythraceae Lagerstroemia speciosa* S 31 11 Oct-Jan Meliaceae Cedrela odorata Swietenia macrophylla s s 2 2 1 1 Leb-Mar Jun Labaceae Pseudosamanea guachapele Lysiloma divaricatum Enterolobium cyclocarpum Inga vera Inga sp. Pithecellobium saman L,L,B S B,Li L,S,L S L,S,L,B 23 7 4 18 5 56 9 3 2 4 6 42 Nov Jan-Mar Leb-Mar Mar-May Mar-Apr Jan-Mar Moraceae Ficus sp. L,S,PP 2 1 Oct Myrtaceae Psidium guajava^* P,S,LP,L 19 5 Mar-Apr Palmae Cocos nucifera* Elaeis guineensis**"^ S L 92 30 55 14 All year Jul-Sep Rhizophoraceae Rhizophora mangle B,L 5 4 Sep Sterculiaceae Sterculia apetala Guazuma ulmifolia S L 2 12 1 7 Mar Dec-Leb Tiliaceae Luehea seemannii L,S 9 4 Jan-Leb Verbenaceae Tectona grandis"^ Avicennia germinans Total # individuals and bouts S,L S 76 6 1,244 31 6 600 Sep-Nov Sep Matuzak et al. • PARROT DIETS IN A MODIFIED LANDSCAPE 357 low-naped Parrot foraging bouts, while non- native and cultivated plant species represented 24.1% of White-fronted Parrot foraging bouts. Yellow-naped Parrots ate 18 plant species in the wet season {n = 61, Levins’ diet breadth = 0.373) and 23 in the dry season {n = 60, Levins’ diet breadth = 0.370), a 21.7% increase in the dry season. White-fronted Par- rots ate 22 plant species in the wet season (n = 44, Levins’ diet breadth = 0.160) and 26 in the dry season {n = 64, Levins’ diet breadth ~ 0.417), a 15.4% increase in the dry season. Parakeets. — Orange-fronted Parakeets ate 24 food plant species in 16 families {n = 113 foraging bouts. Levins’ diet breadth = 0.551) (Table 3) and Orange-chinned Parakeets ate 30 food plant species in 17 families {n = 214 foraging bouts. Levins’ diet breadth = 0.244) (Table 3). The diet of Orange-fronted Para- keets was comprised of fruits (47%), seeds (25%), flowers (20%), leaves (6%), and bark (3%) while the diet of Orange-chinned Para- keets was comprised of fruits (47%), seeds (20%), leaves (15%), flowers (14%), and bark (4%). Non-native and cultivated plant species represented 33% of Orange-chinned Parakeets foraging bouts, while non-native and cultivat- ed plant species represented 17% of the for- aging bouts of Orange-fronted Parakeets. Orange-fronted Parakeets ate 14 plant spe- cies in the wet season {n = 39, Levins’ diet breadth = 0.574) and 17 in the dry season {n = 74, Levins’ diet breadth = 0.580), a 17.6% increase in the dry season. Orange-chinned Parakeets ate 19 plant species in the dry sea- son {n = 124, Levins’ diet breadth = 0.175) and 23 in the wet season {n — 90, Levins’ diet breadth = 0.361), a 17.4% decrease in the dry season. Diet Similarities. — Overall diet overlap was greater (7 range 52-71.4) than diet overlap when compared in the dry and wet seasons (7 dry range 26.1-57.1, 7 wet season range 26.3- 46.2). The greatest overall similarity in psit- tacine diets was between Orange-chinned Par- akeets and Scarlet Macaws (7 = 71.4) and Orange-fronted Parakeets (7 = 66.7), and be- tween both Amazon parrots (7 = 65.2). Sea- sonally, the greatest similarity in psittacine di- ets was in the wet season between Orange- chinned Parakeets and White-fronted Parrots (7 = 56) and Orange-fronted Parakeets (7 = 57.1), and between White-fronted Parrots and Orange-fronted Parakeets (7 = 52.4). In the dry season, the greatest similarity in diets was between Orange-chinned Parakeets and Scar- let Macaws (7 = 46.2) and White-fronted Par- rots (7 = 45), and between Orange-fronted Parakeets and Scarlet Macaws (7 = 43.5). Body Type and Food Type Preference. — There was a significant positive correlation between body mass and percent seeds in the diet (r = 0.97, P = 0.008), and a significant negative correlation between body mass and percent fruit pulp in the diet (r = —0.96, P = 0.009, Fig. 1). Body mass did not correlate with either the percent of flowers (r = —0.71, P = 0.179) or percent of leaf consumption (r = -0.48, P = 0.42) in the diet of this parrot community. DISCUSSION Foraging Ecology and Diet. — The Costa Rican parrot community studied ate predom- inantly seeds (54% of all foraging observa- tions). This confirms the general characteriza- tion of parrots as major seed predators (De- senne 1994, Renton 2001). However, when Scarlet Macaws were excluded, parrots in our study consumed more fruit pulp and flowers than seeds. This was due to the smaller-bodied species consuming more fruit and flowers than seeds, as reported in other parrot communities (Desenne 1994, Pizo et al. 1995). All parrots in Curu foraged on flowers (5- 26% of their diets). This is a larger percentage of flowers than reported by most studies: Ren- ton (2001) found that Lilac-crowned Parrots (Amazona finschi) do not consume flowers in Mexico; flowers comprised only 4.1% of the diet of the Pacific Parakeet {Aratinga strenua) in Nicaragua (Wermundsen 1997); and flow- ers were only 2.5% of the diet of Puerto Rican Amazons {Amazona vittata) (Snyder et al. 1987). Desenne (1994) and Pizo et al. (1995) documented that flowers comprised 18% of all foraging bouts in two parrot communities, while Galetti (1993) found that flowers com- prised 20% of the diet of Scaly-headed Parrots (Pionus maximiliani) and Pizo et al. (1995) reported that flowers comprised 25% of the diet of Reddish-bellied Parakeets (Pyrrhnra frontalis). Flower consumption is important because some psittacines reportedly act as pollinators (Cotton 2001), while many destroy the flowers they eat (Ragusa-Netto 2005; 358 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 TABLE 2. Foraging observations for Yellow-naped and White-fronted parrots. Species marked with * are non-native, ** are cultivated, and *** are both non-native and cultivated. Parts eaten are coded as S — seeds, FP = fruit pulp, B = bark, L = leaves, F = flowers, and Li = lichen. Yellow-naped Parrot White-fronted Parrot Family/Species Part eaten # ind. # bouts # ind. # bouts Months Anacardiaceae Anacardium excelsum S Spondias mombin FP Mangifera indica** FP,L Astronium graveolens S Annonaceae Annona sp. FP Bignoniaceae Tabebuia rosea F Bombacaceae Ceiba pentandra S,L,B Bombacopsis quinata F,FP,S Ochroma pyramidale F Boraginaceae Cordia alliadora S,F Burseraceae Bursera simaruba S Caesalpiniaceae Schizolobium parahybum F,L Senna reticulata S Tamarindus indica S Chrysobalanaceae Licania platypus Combretaceae Terminalia catappa^ S Terminalia oblonga S Combretum sp. F Elaeocarpaceae Muntingia calabura S,FP Euphorbiaceae Sapium glandulosum S,FP Leguminocae Erythrina poeppigiana F Erythrina costaricensis F Lorantaceae Psittacamthus sp. S,F Meliaceae Cedrela odorata S,FP Fabaceae Lysiloma divaricatum S Enterolobium cyclocarpum S,L,B,Li Inga sp. S,F,B Pithecellobium saman S,F,L Leucaena leucocephala S 2 1 6 3 Mar-Apr 40 6 10 2 Aug-Sep 15 6 18 8 Mar 0 0 2 1 Mar 4 1 0 0 Feb 4 2 4 2 Feb-Mar 0 0 4 3 Feb 33 11 0 0 Feb-Apr 0 0 2 1 Jan 2 1 3 2 Feb 50 8 9 4 Feb-Mar 9 4 5 2 Apr-May 0 0 5 1 Feb 0 0 2 1 Feb 2 1 0 0 Jan 8 1 17 4 Feb-Mar 2 1 3 1 Feb 3 2 0 0 Feb 12 1 0 0 Jan 9 3 10 1 Sep-Oct 4 2 35 9 Dec-Jan 5 1 2 1 Jan-Feb 0 0 10 3 Oct-Dec 29 5 2 1 Feb-Mar 4 2 4 2 Mar-Apr 19 7 5 2 Feb-Apr 0 0 10 4 Jul 0 0 7 3 Mar 21 7 5 2 Jun-Sep Matuzak et al. • PARROT DIETS IN A MODIFIED LANDSCAPE 359 TABLE 2. Continued. Family/Species Part eaten Yellow-naped Parrot # ind. # bouts White-fronted Parrot # ind. # bouts Months Moraceae Brosimum alicastrum S,FP 0 0 45 1 Oct Ficus insipida S,FP 12 2 19 2 Oct Myrtaceae Psidium guajava^* S,FP 3 1 0 0 Apr Palmae Scheelea rostrata FP 4 2 0 0 Jun, Sep Elaeis guineensis^^"^ FP 0 0 145 8 May-Jun Rhizophoraceae Rhizophora mangle F 0 0 4 1 Jan Rubiaceae Calycophyllum candidissimun S 6 1 0 0 Feb Rutaceae Citrus aurantifolia*^*^ s 10 2 0 0 Oct-Nov Citrus limeta*** s 6 2 0 0 Nov Citrus aurantium*^"^ s 49 17 17 2 Oct-Dee Citrus paradise*"^* s 2 1 0 0 Nov Zanthoxylum sp. S,FP 8 1 2 1 Sep Sterculiaceae Sterculia apetala s 0 0 2 1 Feb-Mar Guazuma ulmifolia S,FP 1 1 23 9 Nov-Jan Tiliaceae Luehea seemannii S,F 4 1 33 11 Dec-Jan Verbenaceae Tectona grandis**"^ S 67 16 15 4 Aug-Nov Avicennia germinans S,L 4 1 11 4 Apr-Sep Unknown Vine sp. S 0 0 5 1 Feb Total # individuals and bouts 453 121 501 108 GDM, unpubl. data). Given that parrots and other flower eaters can destroy the entire flow- er crop of individual trees (Galetti 1993, Ra- gusa-Netto 2005), it is important for research- ers to be aware of the possible ecological im- pacts of lowering fitness of these plant species in heavily modified landscapes where tree abundances of some species may have been drastically lowered by habitat clearing. The relationship between parrot body size and the percentage of seeds or fruit in the diet may be related to nutritional requirements of parrot species. Seeds are generally high in protein (Gilardi 1996) and larger-bodied avian species may require greater amounts of pro- tein for maintenance (Klasing 1998). Smaller- bodied species may require more fruit pulp, which can be high in sugars, as their metab- olism may require additional energy since they have higher energy needs than larger- bodied species. Therefore, the abundance and ratios of larger-bodied and smaller-bodied par- rots in a psittacine community may be a good predictor of what plant species may be re- quired to sustain populations in certain areas. The small-bodied parakeets and Amazona parrots showed the most similarity in overall psittacine diets, suggesting that congeneric and similar-sized species forage on a large subset of the same plant species. The greater similarity in psittacine diets in the wet season as opposed to the dry season could be related to fewer species of trees producing parrot food in the wet season (Matuzak and Brightsmith 2()07) making it more likely that psittacines would forage on the same plant species. 360 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 TABLE 3. Foraging observations for Orange-fronted and Orange-chinned parakeets. Species marked with * are non-native, ** are cultivated, and *** are both non-native and cultivated. Parts eaten are coded as S = seeds, FP = fruit pulp, B = bark, L = leaves, F = flowers, and Li = lichen. Orange-fronted Orange-chinned Parakeet Parakeet Family/Species Part eaten # md. # bouts # md. # bouts Months Anacardiaceae Anacardium excelsum Spondias mombin Spondias purpurea Mangifera indica** Bignoniaceae Tabebuia rosea Bombacaceae Ceiba aesculifolia Ceiba pentandra Bombacopsis quinata Caesalpiniaceae Schizolobium parahybum Delonix regia* Senna sp. Cecropiaceae Cecropia sp. Combretaceae Terminalia catappa* Terminalia oblonga Laguncularia racemosa Euphorbiaceae Sapium glandulosum Leguminocae Gliricidia sepium Erythrina poeppigiana Lorantaceae Psittacamthus sp. Meliaceae Cedrela odorata Swietenia macrophylla Fabaceae Enterolobium cyclocarpum Inga vera Pithecellobium saman Moraceae Brosimum alicastrum Eicus insipida Myrtaceae Psidium guajava** Palmae Elaeis guineensis*** Rutaceae Zanthoxylum sp. F 0 0 FP 12 2 L 3 1 FP 4 1 F,L,B 8 4 S,F 0 0 S,F,L,B 0 0 S,F,L,B 23 5 L 0 0 F 8 1 F 2 1 S,F 0 0 F 28 3 S 2 1 L 0 0 FP 7 2 S,F 15 7 F 4 1 S,F 6 2 S,FP,L,B 34 8 S 0 0 L,B 12 1 F 0 0 S,F,L 10 2 S,FP 0 0 S,FP,B 0 0 S,FP 25 7 FP 18 3 S,FP 16 3 2 1 Feb 21 2 Sep 0 0 Oct 532 47 May-Aug 17 5 Apr, Jun 14 8 Jan- Apr 172 11 Feb-Apr 100 22 Jan- Apr 8 1 Jul 2 1 Jun 20 1 Mar 6 2 Jul 0 0 Dec 0 0 Feb 73 8 Feb-May 15 2 Aug-Oct 0 0 Jan-Feb 0 0 Jan 31 6 Sep-Oct 11 2 Sep-Oct 4 1 Nov 67 8 Mar-Aug 3 1 Feb 88 11 Jan-Feb 14 4 Jun 209 12 May-Jul 60 10 Mar-Apr 156 19 Jun-Jul 16 5 Jun-Aug Matuzak et al. • PARROT DIETS IN A MODIFIED LANDSCAPE 361 TABLE 3. Continued. Orange-fronted Orange-chinned Parakeet Parakeet Family/Species Part eaten # ind. # bouts # ind. # bouts Months Sterculiaceae Sterculia apetala Guazuma ulmifolia Tiliaceae Luehea seemannii Verbenaceae Tectona grandis^** Verbenaceae Avicennia germinans Unknown Vine sp. Total # individuals and bouts s 22 2 S,FP 115 32 S,F 5 2 S,F 36 11 S,L 57 11 S 0 0 472 113 46 7 Jan-Feb 3 2 Dec-Apr 4 1 Jan, Jun 15 3 Jun-Sep 29 5 Aug-Sep 31 6 Jan-Feb 1,869 214 Smaller parrot species such as those in Bro- togeris and Aratinga often increase in abun- dance in modified landscapes, as larger parrot species decline (Karubian et al. 2005). Thus, anthropogenic impacts can drastically reduce the ratio of large to small psittacines in a com- munity. Smaller-bodied parrot species may even disperse some smaller-seeded trees (Jan- zen 1981, Fleming et al. 1985). Plant repro- duction and regeneration in modified land- scapes could be altered by removing the larg- est-bodied seed predators, and increasing flo- ral predators and potential dispersers of small seeds. Diet Specialization and Seasonal Shifts. — The number of plant types consumed by each parrot species ranged from 24 to 36 species from 15 to 21 plant families. The numbers of food items consumed by the two Amazona species in Guru are similar to those for other parrot species: Scaly-headed Parrots in Brazil ate 38 plant species from 18 families (Galetti 1993) and Lilac-crowned Parrots in Mexico ate 33 plant species from 14 families (Renton 2001). Orange-fronted Parakeets in Guru for- aged on 24 food plant species in 16 families. This is greater than the 15 species from 12 families recorded for the congeneric Pacific Parakeet in Nicaragua (Wermundsen 1997). Scarlet Macaws foraged on 32 food plant spe- cies in 15 families in Guru, while studies of wild Scarlet Macaws reported 15, 43, and 52 food species in Belize; Garara, Gosta Rica; and Peru, respectively (Gilardi 1996, Renton 2006, Vaughan et al. 2006). Diet composition of each species of parrot varied seasonally. Four of five parrot species ate more plant species during the dry season than the wet season. This trend corresponds with an increase in the number of food species available, and the percentage of trees bearing seeds and flowers during the dry season (GDM, unpubl. data). The number of food species and trees bearing large quantities of fruit pulp such as M. indica, S. mombin, E. guineensis, Psidium guajava, and Scheelea rostrata increased during the wet season. The wet season was also when leaves became a larger part of the diet of each species. Diet breadth among parrots usually increas- es with increasing food abundance and diver- sity of available food items (Wermundsen 1997, Renton 2001). This pattern held for only one species in our study. White-fronted Parrot, which had a 260% increase in diet breadth when food availability peaked in the dry sea- son (Matuzak and Brightsmith 2007). The in- crease in diet breadth of Orange-chinned Par- akeets during the wet season could also be related to food abundance, as fruit pulp (the species preferred food part) abundance peaks at this time of year (GDM, unpubl. data). Two other species. Scarlet Macaws and Orange- chinned Parakeets, had more specialized diets during the dry season food peak (36 and 52% reductions in diet breadth, respectively). Dur- ing the dry season in Belize, Scarlet Macaws have a less specialized diet (Levins’ diet breadth = 0.394; Renton 2006) than did Scar- let Macaws in our study (Levins’ diet breadth 362 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 FIG. 1. Body mass versus percent of seeds (A) and fruits (B) in the diets of parrots in Guru Wildlife Refuge on the Nicoya Peninsula of Costa Rica. = 0.139). In Cum, Scarlet Macaws in the dry season specialize on two food items, P. saman and T. catappa, both of which produce abun- dant resources in the dry season and combine for 53% of the macaws’ dry season diet. The more specialized diet of macaws in this study may be due to several factors, including less food diversity and availability due to a human modified landscape, a preference for T. catap- pa over other available resources, and/or a lack of knowledge and training to find local resources after their release to the site. The diet breadth of Lilac-crowned Amazons in Mexico was 0.22 in the dry season and 0.55 in the wet season when the number of food species available peaked (Renton 2001). The diet breadth of the two species of Amazona parrots in Curu was larger in the dry season Matuzak et al. • PARROT DIETS IN A MODIFIED LANDSCAPE 363 and lower in the wet season when compared to Lilac-crowned Amazons; however, food availability peaked in the dry season in Guru. Yellow-naped Parrots and Orange-fronted Par- akeets exhibited no major difference in diet breadth between seasons. Native versus Non-native Food Plant Spe- cies.— Use of non-native food resources by parrots is widespread and important in sus- taining some species during times of low food availability (Forshaw 1989, Pitter and Chris- tiansen 1995). Over half of all foraging bouts in our study were on non-native and cultivated tree species; however, when macaws are not included, only 29% of foraging bouts were on non-native and cultivated species. Non-native species consumed included two naturalized species that occur along the coasts of Costa Rica (C. nucifera and T. catappa), seven spe- cies cultivated for fruit (4 species of Citrus, M. indica, P. guajava, and E. guineensis), two non-native ornamentals {D. regia and Lager- stroemia speciosa), and T. grandis, a tree planted for timber (Holdridge et al. 1997). Two species (C. nucifera and T. catappa) were used mainly by Scarlet Macaws and rep- resented 48% of their diet. Wild populations of macaws at other sites in Costa Rica forage heavily on T. catappa as this plant species is highly abundant and is in seed all year along coastal areas and beaches where wild macaws are found. The macaws in Guru exploit a sim- ilar range of food resources in similar envi- ronments as wild macaws along the central coast of Costa Rica (Vaughan et al. 2006; GDM, unpubl. data). Citrus trees (lime, lemon, grapefruit, and orange) were an important food resource for Yellow-naped Parrots from November to De- cember, a period of increasing food abundance prior to onset of nesting (Matuzak and Brightsmith 2007). Two species, E. guineensis and T. grandis, provided important food re- sources for the entire parrot community. These species were important during declining food availability and low food diversity, April-June and September-October, respec- tively (Matuzak and Brightsmith 2007). CONSERVATION IMPLICATIONS The ability of psittacines to exploit non-na- tive resources may be important for their fu- ture survival, not only in Costa Rica, but across Latin America. The high rates of hab- itat conversion in Costa Rica over the past several decades have produced millions of hectares dominated by forest fragments, small farms, and plantations (Kleinn et al. 2002). Parrots have greatly expanded their available food base by using introduced and cultivated food plants. As more areas of the Neotropics are converted from native habitats to small farms, persistence of parrots may become linked to their ability to exploit introduced and cultivated species in modified landscapes. Use of non-native and cultivated species brings psittacines into direct competition with humans, making them a perceived pest of crops (Bucher 1992). Psittacines in many parts of the world are trapped or killed due to real or perceived damage to crops (Bucher 1992). The benefits of feeding on introduced and cul- tivated species for most psittacines apparently outweigh the mortality inflicted by humans and, as a result, some species become abun- dant in modified landscapes (Pitter and Chris- tiansen 1995, Moegenburg and Levey 2003). However, for uncommon psittacines with low reproductive rates like macaws or large ama- zons, conflicts with humans could threaten the species’ persistence (Bucher 1992). Deforestation in Costa Rica has removed up to 60-70% of the nation’s original forest cov- er (Kleinn et al. 2002). In the study area, many hectares are regenerating forest cover as farms and pastures are being abandoned (Kleinn et al. 2002). Targeted restoration of G. ulmifolia, B. quinata, S. mombin, and L. seemannii in our study area would be feasible and of great benefit as these native tree species provide key food resources to psittacines. G. ulmifolia and S. mombin are of special importance since both fruit during the time of lowest overall food availability and may turn out to be key- stone species in these dry forest environments. Increasing the abundance of native species should also decrease the dependence of psit- tacines on introduced and cultivated species, and decrease the potential for negative inter- actions with humans. ACKNOWLEDGMENTS We thank Agustina Arcos, Alejandro Solano, Laura Falasz, Sigrid Collado, Olivier Horiol, Jerome Hillaire, Cynthia Shafer, and all project assistants and volun- teers for their dedication to assisting with the held- 364 THE WILSON JOURNAL OF ORNITHOLOGY • VoL 120, No. 2, June 2008 work. We thank Jason May for helping with statistical analyses and the Ministry of Environment and Energy (MINAE) in Costa Rica and the Guru Wildlife Refuge for their support of the project and access to all areas of the wildlife refuge. Funding was provided by Ami- gos de las Aves USA Parrot Conservation and Re- search Fund, The Amazona Society, Loro Parque Foundation, Lincoln Park Zoo, Bird Clubs of Virginia, Columbus Zoo, Kaytee Avian Foundation, Cleveland Zoological Society, and the following private donors: Tracey Coryell, Jeff Kidston, and Dr. Janice Boyd. LITERATURE CITED Bennett, P. M. and I. P. F. Owens. 1997. Variation in extinction risk among birds: chance or evolution- ary predisposition? Proceedings of the Royal So- ciety of London B. 264:401-408. Brightsmith, D., J. Hilburn, A. del Campo, J. Boyd, M. Frisius, R. Frisius, D. Janik, and F. Guillen. 2005. The use of hand-raised psittacines for re- introduction: a case study of Scarlet Macaws (Ara macao) in Peru and Costa Rica. Biological Con- servation 121:465-472. Bucher, E. H. 1992. Neotropical parrots as agricultural pests. Pages 201-219 in New World parrots in cri- sis: solutions from conservation biology (S. R. Beissinger and N. F. R. Snyder, Editors). Smith- sonian Institute Press, Washington, D.C., USA. CITES. 2002. Transfer the Yellow-naped Amazon {Amazona auropalliata) from Appendix II to Ap- pendix I. Proponent: Costa Rica. Conference of the Parties to the Convention of International Trade in Endangered Species of Wild Fauna and Flora (CITES). Proposition 16. Collar, N. J. 1997. Family Psittacidae. Pages 280- 479 in Handbook of the birds of the world (J. del Hoyo, A. Elliott, and J. Sargatal, Editors). Lynx Edicions, Barcelona, Spain. Collar, N. J. 1998. Information and ignorance con- cerning the word’s parrots: an index for twenty- first century research and conservation. Papa- geienkunde 2:201-235. Colwell, R. K. and D. J. Futuyama. 1971. On the measurement of niche breadth and overlap. Ecol- ogy 52:567-576. Cotton, P. A. 2001. The behavior and interactions of birds visiting Erythrina fusca flowers in the Col- ombian Amazon. Biotropica 33:662-669. Data Description Inc. 2006. Data Desk statistical software. Ithaca, New York, USA. Desenne, P. 1994. Estudio preliminar de la dieta de 15 especies de psitacidos en un bosque siempreverde, cuenca del Rio Tawadu, Reserva Forestal El Cau- ra, Edo. Bolivar. Pages 25-42 in Biologia y con- servacion de los psitacidos de Venezuela (G. Mo- rales, I. Novo, D. Bigio, A. Luy, and F. Rojas- Suarez, Editors). Graficas Giavimar, Caracas, Ve- nezuela. Fleming, T. H., C. F. Williams, F. J. Bonaccorso, and L. H. Herbst. 1985. Phenology, seed disperal, and colonization in Muntingia calabura, a neotropical pioneer tree. American Journal of Botany 72:383- 391. Forshaw, j. M. 1989. Parrots of the world. Third Edi- tion. Landsdowne Editions, Melbourne, Australia. French, A. R. and T. B. Smith. 2005. Importance of body size in determining dominance hierarchies among diverse tropical frugivores. Biotropica 37: 96-101. Galetti, M. 1993. Diet of the Scaly-headed Parrot (Pionus maximiliani) in a semi-deciduous forest in southeastern Brazil. Biotropica 25:419-425. Gilardi, j. D. 1996. Ecology of parrots in the Peruvian Amazon: habitat use, nutrition and geophagy. Dis- sertation. University of California, Berkeley, USA. Holdridge, L. R., L. j. Poveda, and Q. Jimenez. 1997. Arboles de Costa Rica. Volume 1. Tropical Sci- ence Center, San Jose, Costa Rica. Janzen, D. H. 1981. Seed predation by Orange- chinned Parakeets in Costa Rica. Auk 98:841- 844. Karubian, j., j. Fabara, D. Yunes, J. P. Jorgenson, D. Romo, and T. B. Smith. 2005. Temporal and spatial patterns of macaw abundance in the Ecu- adoran Amazon. Condor 107:617-626. Klasing, K. C. 1998. Comparative avian nutrition. CAB International, Wallingford, Connecticut, USA. Kleinn, C., L. Corrales, and D. Morales. 2002. For- est area in Costa Rica: a comparative study of tropical forest cover estimates over time. Environ- mental Monitoring and Assessment 73:17-40. Levins, R. 1968. Evolution in changing environments. Princeton University Press, Princeton, New Jer- sey, USA. Matuzak, G. D. and D. j. Brightsmith. 2007. Roost- ing of Yellow-naped Parrots in Costa Rica: esti- mating the size and recruitment of threatened pop- ulations. Journal of Field Ornithology 78:159- 169. Moegenburg, S. M. and D. J. Levey. 2003. Do fru- givores respond to fruit harvest? An experimental study of short-term responses. Ecology 84:2600- 2612. Munshi-South, j. and G. S. Wilkinson. 2006. Diet influences life span in parrots (Psittaciformes). Auk 123:108-118. PlTTER, E. AND M. B. CHRISTIANSEN. 1995. Ecology, status and conservation of the Red-fronted Macaw {Ara rubrogenys). Bird Conservation International 5:61-78. Pizo, M. A., I. SiMAO, AND M. Galetti. 1995. Diet and flock size of sympatric parrots in the Atlantic for- est of Brazil. Ornitologia Neotropical 6:87-95. Ragus-Netto, j. 2005. Extensive consumption of Ta- bebuia aurea (Manso) Bentham & Hooker (Big- noniaceae) nectar by parrots in a tecoma savanna in the southern Pantanal (Brazil). Brazilian Jour- nal of Biology 65:339-344. Renton, K. 2001. Lilac-crowned Parrot diet and food Matuzak et al. • PARROT DIETS IN A MODIFIED LANDSCAPE 365 resource availability: resource tracking by a parrot seed predator. Condor 103:62-69. Renton, K. 2006. Diet of adult and nestling Scarlet Macaws in southwest Belize, Central America. Biotropica 38:280-283. Sazima, I. 1989. Peach-fronted Parakeet feeding on winged termites. Wilson Bulletin 101:656-657. ScHUTT, A. AND C. Vaughn. 1995. Incorporating wild- life into development: the case of the Cum wild- life refuge and farm, Costa Rica. Pages 250-254 in Integrating people and wildlife for a sustainable future. Proceedings of the First International Wild- life Management Congress (J. A. Bissonette and P. R. Krausman, Editors). The Wildlife Society. Bethesda, Maryland, USA. Snyder, N. R. E, J. W. Wiley, and C. B. Kepler. 1987. The parrots of Luquillo: natural history and conservation of the Puerto Rican Parrot. Western Foundation of Vertebrate Zoology, Los Angeles, California, USA. Stiles, F. G. and A. Skutch. 1989. A guide to the birds of Costa Rica. Cornell University Press, Ith- aca, New York, USA. Tosi, J. 1969. Ecological life zone map of Costa Rica. Tropical Science Center, San Jose, Costa Rica. Vaughan, C., N. Nemeth, and L. Marineros. 2006. Scarlet Macaw, Ara macao, (Psittaciformes: Psit- tacidae) diet in central Pacific Costa Rica. Inter- national Journal of Tropical Biology 54:919-926. Vaughan, C., J. F. Gamboa, D. Hernandez, J. M. Mora, P. E. Sanchez, and H. C. Kiel. 1994. Ev- aluacion de los recursos biologicos y diseno de corredores para los distritos de Lepanto, Paquera, y Cobano del Canton Central de la Provincia de Puntarenas. Informe Final. Universidad Nacional, Heredia, Costa Rica. Wermundsen, T. 1997. Seasonal change in the diet of the Pacific Parakeet Aratinga strenua in Nicara- gua. Ibis 139:566-568. The Wilson Journal of Ornithology 120(2):366— 370, 2008 THE SIGNAL FUNCTION OF A MELANIN-BASED PLUMAGE ORNAMENT IN GOLDEN-WINGED WARBLERS EMILY ANNE McKINNON'-^^ AND RALEIGH J. ROBERTSON' ABSTRACT. — The decline of Golden-winged Warblers (Vermivora chrysoptera) has been attributed in part to hybridization with a sister species, the Blue-winged Warbler (V. pinus), which lacks the black throat patch typical of the Golden-winged Warbler. Understanding the signal function of male plumage ornaments in Golden- winged Warblers may provide insight into the mechanisms driving hybridization. If Golden-winged Warbler males use the black throat patch for interspecific signaling. Blue-winged Warblers or hybrids may be leading hybridization between the species. We examined the signal function of the melanin-based throat patch in a population of Golden-winged Warblers on the edge of the hybrid zone. Males with increased ultra violet chroma in their throat patches were older and their mates had significantly earlier first egg dates. This suggests the black throat patch of Golden-winged Warblers may be an age-related indicator of quality. Female Golden-winged Warblers should not mate preferentially with males which lack the black throat patch if it functions as an indicator of age and or male quality. Received 13 January 2007. Accepted 22 September 2007. Females, in many birds that exhibit sexual dimorphism, prefer and benefit from choosing more ornamented and older mates (Andersson 1994). Sexually selected plumage may have a role in maintaining species separation in sym- patric populations of related species (Saetre et al. 1997). Golden-winged {Vermivora chry- soptera) and Blue-winged (V. pinus) warblers are sister species that regularly hybridize in areas of range overlap (Gill 1997), yet hybrid- ization seems to result in the eventual extir- pation of the Golden-winged Warbler (Gill 1980). Reasons for this are unclear; recent work (Leichty and Grier 2006) suggests that non-assortative mate choice based on plumage may contribute to genetic swamping of Gold- en-winged Warblers by Blue- winged Warbler genes. Golden-winged Warblers have a black, mel- anin-based throat patch and auriculars, and a carotenoid-based yellow cap and wingbar. Blue- winged Warblers are yellow overall with grayish-blue wings and a thin black eye-line. Hybrid females generally resemble female Golden- winged Warblers and not Blue- winged Warblers. Thus, male Golden-winged War- blers may pair with hybrid females, while male Blue-winged Warblers are less likely to ' Department of Biology, Queen’s University, King- ston, ON K7L 3N6, Canada. 2 Current address: Faculty of Forestry and Environ- mental Management, University of New Brunswick, Fredericton, NB E3B 6E1, Canada. ^ Corresponding author; e-mail: emily_anne.mckinnon@unb.ca pair with a hybrid (Confer 2006). Leichty and Grier (2006) experimentally found that female Golden-winged Warblers do not pair with sim- ulated hybrids (males with facial pattern ex- perimentally removed). This suggests the black facial pattern is an important species- specific signal for mate choice. Leichty and Grier (2006) also reported that males lacking the black facial pattern lost their territory more often than untreated Golden-winged Warblers, suggesting the throat patch may function in intrasexual interactions. Our objectives were to: (1) describe reflec- tance characteristics of the Golden-winged Warbler throat patch, and (2) ascertain if these characteristics correlated with indices of male quality, which would suggest a possible signal function. Reproductive success and age were used as indicators of male quality. We record- ed the date a male’s mate first laid an egg as a proxy for reproductive success. We expected ultra violet (UV) chroma (proportion of re- flectance in the UV range) and darkness to be important components of the melanin-based patch (Mennill et al. 2003). If the black throat patch functions as a signal of fitness, high quality throat patches should correlate with in- creased reproductive success. Throat patch quality should also increase with age (Brooks and Kemp 2001). METHODS We captured 34 male Golden-winged War- blers (18 after-second year, ASY; and 16 sec- ond-year, SY) with mist nets from 6 May to 366 McKinnon and Robertson • MELANIN-BASED PLUMAGE IN GOLDEN-WINGED WARBLERS 367 15 July 2005 at the Queen’s University Bio- logical Station (44° 34' N, 76° 19' W; 40 km north of Kingston, Ontario, Canada). This population of Golden-winged Warblers is on the edge of the hybrid zone (Dabrowski et al. 2005). Arriving Golden-winged Warblers were monitored at known breeding sites and first egg date was recorded for each pair’s first nest {n = 11) for those territories where we were able to find the nest. Three throat patch feathers were collected from each male and stored in opaque enve- lopes until analysis. Feathers from each bird were taped to dark black paper with less than 10% reflectance (CANSON #425 Stygian black) overlapping in a way that mimicked how they would lay on the bird. Plumage col- oration was measured in a darkened labora- tory room with no windows (to minimize var- iation in ambient light) using an Ocean Optics S2000 spectrometer with a PX-2 pulsed xenon lamp set to a white spectralon standard (WS- 1) (Ocean Optics Inc., Dunedin, FL, USA) and dark standard (black felt in a box). A rub- ber sheath fitted over the spectrometer probe excluded light and ensured the probe was at a fixed distance, perpendicular to the sample. Five readings from each sample (lifting and lowering the probe between readings) were re- peatable (r = 0.61, F34140 = 8.95, P < 0.001) (Lessells and Boag 1987) and were averaged for analyses. Analysis was limited to wave lengths from 300 to 700 nm, the approximate visual range of passerines (Cuthill et al. 1999). Average reflectance for each 10-nm interval from 300 to 700 nm was calculated to give 41 reflectance measures per bird. Reflectance data were summarized using principal components analysis (PCA). All data conformed to assumptions of parametric sta- tistics. Principal components (PCs) were com- pared to first egg date with linear regressions. Throat patch components were compared be- tween age classes (ASY and SY) with two- sample Gtests. All statistics were calculated using IMP 5. 1 . RESULTS Principal component I proved to be a mea- sure of “darkness” of the throat patch. It rep- resented reflectance across all wave lengths equally, as confirmed by a strong positive re- lationship when plotted against mean reflec- tance (r = 0.99, P < 0.001, n = 34). Principal component II was a measure of “UV chroma” of the throat patch and correlated positively (r = 0.96, P < 0.001, n = 34) with the propor- tion of reflectance in the UV range. The ei- genvalue for PC II was < 1, but we retained PC II in the analyses. Principal component III explained 0.5% of the variation, representing variation at short (UV) and long (red) wave lengths. Only PCs I and II were considered in the following tests due to the small influence of PC III on variation in spectra, and with no clear biological reason to support addition of PC III to the analyses. Throat patch plumage had low reflectance with little variation across the spectrum ex- amined (Fig. 1). PC I scores of older (ASY) males did not differ significantly from PC I scores of second-year (SY) males {t = 2.03, P = 0.20, n = 18 and 16) although, on av- erage, their throat patches were darker than those of younger males (Fig. 2A). Older males had higher PC II scores than younger males {t = 2.03, P = 0.034, n = 18 and 16) (Fig. 2B) indicating they had a greater proportion of UV reflectance in their throat patch than younger males. Principal component I did not correlate with first egg date (r == 0.22, P - 0.51, n = 11); however, PC II and first egg date had a significant negative trend (r = —0.65, P — 0.029, n = 11) (Fig. 3) indicating that males with higher UV reflectance in their throat patch generally had mates which laid eggs earlier in the season. DISCUSSION Throat patch feathers of male Golden- winged Warblers had a reflectance curve sim- ilar to that of Black-capped Chickadees {Poe- cile atricapillus) (Mennill et al. 2003) with some variation at UV wave lengths and little variation across the rest of the spectrum (Fig. 1 ). Our PCA results suggest that UV chroma (PC II) and darkness (PC I) account for the majority of the variation across male Golden- winged Warbler throat patches. Males with the highest UV chroma had mates that laid eggs significantly earlier in the season. Earlier first egg dates are positively correlated with in- creased reproductive success in other species (Verhulst et al. 1995). This suggests UV chro- 368 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 5 4.5 4 V W 3.5 t5 ^ « 2.5 & 2 # 1.5 1 0.5 0 FIG. 1. Golden-winged Warbler plumage reflectance curves for the throat patch of ASY (solid line) and SY (dashed line) males. Curves were calculated from average reflectance values. Wave length (nm) ma and reproductive success may be positive- ly correlated in Golden-winged Warblers. The darkness of the throat patch did not correlate significantly with first egg date or age. It may be that darkness is not as impor- tant as a signal as UV chroma. However, anal- ysis of feathers in the laboratory has limita- tions in how it can be applied to reflectance of feathers on a living bird. The background paper may have shown through the feathers influencing darkness measurements by in- creasing their reflectance. This problem would have affected darkness more than UV chroma, as the paper did not reflect in UV wave lengths. Mounting feathers on different back- grounds with varying reflectance, although beyond the scope of this study, would allow the magnitude of this effect to be examined. Throat patches of older (ASY) males had significantly higher UV chroma and tended to be darker than throat patches of SY males. Elaboration of UV plumage ornaments in old- er males has been documented (Andersson and Amundsen 1997, Siefferman et al. 2005). This trend could be further clarified with re- liable estimates of the exact age of individu- als. Age-based indicators are simple and hon- est signals of quality because older males have survived longer than younger males, and tend to have increased reproductive success due to experience (Brooks and Kemp 2001). Golden- winged Warblers have a high return rate in the population studied (60% of males identified in 2005 were banded in previous years; E. A. McKinnon, unpubl. data) and often occupied the same territory. Males which return to known ‘good quality’ territories should be in a better position to breed earlier than younger males which may have to expend more energy fighting or searching for territories. Older males which have better territories than younger males are probably also in better physiological condition (Marra and Holberton 1998) suggesting a proximate mechanism for age-related variation in throat patch quality. Poston et al. (2005) reported that poor nutri- tion has a dulling effect on the black bib of House Sparrows {Passer domesticus). Corre- lations between increased UV reflectance of melanin-based plumage and individual quality have also been documented (Siitari and Huhta 2002, Doucet et al. 2005). Golden-winged Warblers molt their throat patch in breeding areas at the end of the breeding season (Pyle 1997). Younger males with poor quality ter- ritories or less foraging experience may be in poorer nutritional condition at time of molt than older birds and, thus, have paler and less UV reflective throat patches. Provided Golden-winged Warbler females use appearance of the throat patch to assess age and quality of a male, it is unlikely they would mate with males lacking the throat patch, such as a Blue- winged Warbler or the hybrid ‘Brewster’s Warbler’ {Vermivora leu- cobronchialis). Leichty and Grier (2006) pre- sented results showing that Golden-winged Warbler males whose black face pattern was McKinnon and Robertson • MELANIN-BASED PLUMAGE IN GOLDEN-WINGED WARBLERS 369 (A) (B) o Q. 2 1.5 1 0.5 0 -0.5 -1 -1.5 ASY ^ SY FIG. 2. Box plots showing the difference in PC I (A) and PC II (B) between ASY and SY males. Hor- izontal lines show 25th, 50th and 75th percentiles with lines extending to minimum and maximum values. An open circle indicates an outlier. removed (simulating the ‘Brewster’s Warbler’ facial pattern) were at a mating disadvantage in a pure Golden-winged Warbler population. Leichty and Grier (2006) also found that female Golden-winged Warblers chose to mate with males which had the black throat patch. Our study found that females lay earlier when paired with males with more UV reflec- tive throat patches. It is unclear what infor- mation, if any, the black throat patch conveys to female Blue-winged Warblers or hybrids. A study of mate preferences of Blue-winged Warblers and hybrids could further elucidate the role of plumage as a potential mechanism driving this hybridization system. -1 1 -1.5 . . ■ ■ I 138 140 142 144 146 148 150 152 First egg date FIG. 3. Relationship between UV chroma and first egg date (ordinal date 139 represents 19 May 2005). ACKNOWLEDGMENTS Rachel Vallender is credited with the initial idea for this study and her research facilitated the fieldwork for this project. We are grateful to S. M. Doucet for help with spectrometry. K. C. Fraser provided assistance in the field and in reviewing this manuscript. S. L. Harp- er, Anna Yunnie, and E. A. Stancu provided field as- sistance. We thank the staff at the Queen’s University Biological Station for field support. Laurie Smaglick- Johnson helped with field assistance and Golden- winged Warbler photographs. Robert Montgomerie provided spectrometry equipment and helpful com- ments. Funding was provided by the Queen’s Univer- sity Summer Work Experience Program, Alexander and Cora Munn Summer Research Award to EAM, and a Natural Sciences and Engineering Research Council of Canada Discovery Grant to RJR. We thank K. J. McGraw and an anonymous reviewer for helpful comments on the original manuscript. LITERATURE CITED Andersson, M. 1994. Sexual selection in animals. Princeton University Press, Princeton, New Jer- sey, USA. Andersson, S. and T. Amundsen. 1997. Ultraviolet colour vision and ornamentation in Bluethroats. Proceedings of the Royal Society of London Se- ries B 264:1587-1591. Brooks, R. and D. J. Kemp. 2001. Can older males deliver the good genes? Trends in Ecology and Evolution 16:308-313. Confer, J. L. 2006. Secondary contact and introgres- sion of Golden-winged Warblers {Vermivora chry- soptera). Auk 123:958-961. CuTuiu., I. C., A. T. D. Benneti, J. C. Partridge, and E. J. Maier. 1999. Plumage reflectance and the objective assessment of avian sexual dichroma- tism. American Naturalist 160:183-200. 370 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 Dabrowski, a., R. Fraser, J. L. Confer, and I. J. Lovette. 2005. Geographic variability in mito- chondrial introgression among hybridizing popu- lations of Golden-winged (Vermivora chrysop- tera) and Blue-winged {V. pinus) warblers. Con- servation Genetics 6:843—853. Doucet, S. M., D. j. Mennill, R. Montgomerie, R T. Boag, and L. M. Ratcliffe. 2005. Achromatic plumage reflectance predicts reproductive success in male Black-capped Chickadees. Behavioural Ecology 16:218-222. Gill, F 1980. Historical aspects of hybridization be- tween Blue-winged and Golden-winged warblers. Auk 97:1-18. Gill, F. 1997. Local cytonuclear extinction of the Golden-winged Warbler. Evolution 51:519—525. Leichty, E. R. and j. W. Grier. 2006. Importance of facial pattern to sexual selection in Golden- winged Warbler {Vermivora chrysoptera). Auk 123:962-966. Lessells, C. M. and P. T. Boag. 1987. Unrepeatable repeatabilities: a common mistake. Auk 104:116- 121. Marra, P. P. and R. L. Holberton. 1998. Corticoste- rone levels as indicators of habitat quality: effects of habitat segregation in a migratory bird during the non-breeding season. Oecologia 116:284—292. Mennill, D. J., S. M. Doucet, R. Montgomerie, and L. M. Ratcliffe. 2003. Achromatic color varia- tion in Black-capped Chickadees, Poecile atricap- illa: black and white signals of sex and rank. Be- havioral Ecology and Sociobiology 53:350-357. Poston, J. P, D. Hasselquist, I. R. K. Stewart, and D. E Westneat. 2005. Dietary amino acids influ- ence plumage traits and immune responses of male House Sparrows, Passer domesticus, but not as expected. Animal Behavior 70:1171-1181. Pyle, P. 1997. Identification guide to North American birds. Part I. Columbidae to Ploceidae. Slate Creek Press, Bolinas, California, USA. Saetre, G. P, T. Moum, S. Bures, M. Kral, M. Ada- mian, AND J. Moreno. 1997. A sexually selected character displacement in flycatchers reinforces premating isolation. Nature 387:589—592. SlEFFERMAN, L., G. E. HiLL, AND E S. DOBSON. 2005. Ornamental plumage coloration and condition are dependent on age in Eastern Bluebirds Sialia sial- is. Journal of Avian Biology 36:428-435. SIITARI, H. AND E. Huhta. 2002. Individual color var- iation and male quality in Pied Flycatchers {Fi- cedula hypoleuca): a role of ultraviolet reflec- tance. Behavioral Ecology 13:737-741. Verhulst, S., j. H. van Balen, and J. M. Tinbergen. 1995. Seasonal decline in reproductive success of the Great Tit: variation in time or quality? Ecol- ogy 76:2392-2403. The Wilson Journal of Ornithology 120(2):371-377, 2008 FACTORS INFLUENCING FIDELITY OF HOUSE FINCHES TO A FEEDING STATION ANDREW K. DAVIS ‘ 2 ABSTRACT. — House Finches (Carpodacus mexicanus) in North America are a commonly-studied species, but basic aspects of their life history remain poorly understood. I banded and observed marked House Finches at a backyard feeding site in a suburban neighborhood in Atlanta, Georgia from August 2002 through July 2004 to address how age, gender, mycoplasmal conjunctivitis, and time of year affected fidelity of individual House Finches to this site. Of 386 House Finches banded, I recaptured 77 and recorded 1,210 reobservations. More than half (55%) of all birds banded were not seen again and, of those that were, almost half (44%) were seen for less than 2 months. A House Finch’s age, gender, and month of capture significantly affected how many times it was subsequently encountered (recaptured or reobserved), but not the duration of time it spent at the site of banding. Young birds (HYs) were encountered more often than adults (AHYs), and females more than males. Young birds with mycoplasmal conjunctivitis were encountered less often than those without, but this was not true for adults. These data indicate high site fidelity of adults during the breeding season and low site fidelity of juveniles early in the summer that becomes higher in late summer. Most birds captured in fall were encountered for up to 3 months. These results are discussed in relation to previous studies and their implications for transmission of Mycoplasma gallisepticum among House Finches. Received 1 February 2007. Accepted 8 September 2007. House Finches {Carpodacus mexicanus) in North America have become the subject of in- tense study in the past decade. This species has become susceptible to a newly emerged disease, mycoplasmal conjunctivitis (e.g., Hartup et al. 2001, Roberts et al. 2001a, Farm- er et al. 2002, Altizer et al. 2004a, Faustino et al. 2004). This disease, caused by the bac- terium Mycoplasma gallisepticum (MG), can cause infected House Finches to develop eas- ily recognizable swellings around their eyes (mycoplasmal conjunctivitis) with outbreaks occurring annually during fall and winter (Al- tizer et al. 2004a, 2004b). House Finches are also studied because of natural variation in male plumage color and, thus, have become a focal species for study of sexual selection and male quality (e.g.. Hill 1992b, 1998; Hill et al. 1999; Hill 2002). Finally, House Finches in eastern North America originated from the southwestern United States where they are non-migratory, and there is now interest in de- velopment of migratory tendencies in the ex- panding eastern population (Able and Belthoff 1998, Egbert and Belthoff 2003). Ironically, despite the large number of stud- ' Department of Environmental Studies, Emory Uni- versity, Atlanta, GA 30322, USA. ^ Current address: Warnell School of Forestry and Natural Resources, The University of Georgia, Athens, GA 30602, USA; e-mail: akdavis@uga.edu ies involving this species, basic aspects of its’ life history remain poorly understood. For ex- ample, few studies have examined movements of birds at a single site, such as at a backyard feeding station. Thus, little is known of their daily activity patterns at feeding stations, and how this changes throughout the year. What little is known comes from summaries and re- ports from long-term banding stations (Mc- Clure 1989, Hamilton 1992, Hilton 1994). These studies have shown that House Finches can be characterized by large-scale move- ments and seasonal turnover (Hamilton 1991, 1992; Hilton 1994). Previous studies contain important information, but most involved comparing rates of recapture, which may not accurately estimate fidelity to a feeding station (Hamilton 1992), and most did not examine factors affecting site fidelity, such as demo- graphic aspects. Given that transmission of MG is likely to occur where birds gather in large numbers, such as at bird feeders, learn- ing how feeders are used by House Finches is important. Further, since mycoplasmal con- junctivitis is a newly emerged disease, the ef- fect of this disease on feeder use and local movements has yet to be clarified. 1 report results of a 2-year study in which House Finches were trapped, uniquely band- ed, and reobserved at a single suburban back- yard bird feeding station in an effort to elu- cidate the factors influencing their site fidelity. 371 372 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 My Specific objectives were to learn: (1) if the number of times individual House Finches captured or reobserved (‘encountered’) varied with time of year, age or gender, or whether or not they had conjunctivitis; and (2) if these same factors influence the duration of each in- dividual’s presence at the trap site (as mea- sured by the time between initial capture and last encounter). METHODS Site Description. — The study site was the backyard of the author’s house in a suburb of Atlanta, Georgia (33° 39' N, 84° 26' W). The yard was approximately 30 X 40 m, and con- tained three tube-style bird feeders on metal poles in the middle of the yard. All feeders were filled with black-oil sunflower seed, the preferred food of House Finches (Hill 1993). Trapping and Banding. — I trapped House Finches from the beginning of August 2002 through the end of July 2004 (24 months). I conducted trapping sessions at least once a week for a minimum of 3 hrs each morning. I used a combination of mist nets (9 m long, with 30 mm mesh) placed around the feeders and two walk-in cage traps following Hill (2002). The cage traps were cylindrical and made of hardware mesh; each contained a standard bird feeder filled with sunflower seed. Two entrances in the hardware mesh near the bottom of the trap allowed birds to walk in (via a wooden perch placed outside each entrance). House Finches were readily captured by setting this trap in place of the lure feeders on trapping days. I banded each bird upon capture with a numbered USGS metal band and three color bands in a unique color-metal band combina- tion for later reobservation (Hill 1992a). I re- corded each bird’s age as either after-hatch year (AHY) or hatch-year (HY) based on skull ossification, plumage, or retrix fault bars (Pyle 1997). Similarly, I assigned gender of each bird based on the dimorphic adult plumage of the species (Hill 1993). Most HY birds were still in juvenile plumage and gender could not be assigned. I recorded the presence or ab- sence of conjunctivitis in all birds based on outward signs of conjunctival swelling and/or ocular discharge following Altizer et al. (2004b). This same procedure was performed for all newly captured as well as recaptured birds. Reobservations. — I watched for banded in- dividuals at the feeding station, usually in the mornings, when activity at the feeders was high. I recorded the band combination, gender, date and time of observation, and the presence or absence of conjunctivitis whenever a band- ed House Finch was reobserved at or near the feeders. House Finches often visit feeders in large flocks, especially during winter (Hill 1993) and I recorded the number of other finches in the flock with the reobserved indi- vidual. Every effort was made to watch for banded individuals on most days each month. Data Analyses.— I used the pooled trapping and reobservation data from the 24-month pe- riod for this study. I first ensured the reobser- vation data contained no erroneous band com- binations (Milligan et al. 2003) and created two variables for testing site fidelity of House Finches. I summed all of the encounters (first capture + any subsequent recaptures or reob- servations) for every bird. If the bird was not encountered after initial capture, it was as- signed a ‘ r . I assumed this value represented the approximate frequency at which the indi- vidual used the feeding station. Second, I cal- culated the time in days between the first (ini- tial capture) and last (recapture or reobserva- tion) encounter for each individual, using only those individuals that were encountered two or more times. This value was assumed to rep- resent the time span over which the individual visited the feeding station. I used univariate analyses of variance, using either of the above variables as dependents (after log-transfor- mation) with age (AHY or HY), gender, con- junctivitis status (with, without), month (of initial capture), and all two-way interactions between age, gender, and conjunctivitis as in- dependent variables. Significance was accept- ed when P < 0.05 but results were considered nearly significant when P < 0.1. RESULTS I banded 386 separate individuals over the 24-month study period (Table 1), and later made 77 recaptures and 1,210 reobservations of these individuals for a total of 1,673 en- counters. Overall, 55.7% of all House Finches banded were not recaptured or reobserved af- ter initial capture (Fig. lA). Another 24.7% Davis • HOUSE FINCH FIDELITY TO FEEDING STATIONS 373 TABLE 1. Trapping totals (new captures only) and age ratios of House Finches by month at a feeding station in Atlanta, Georgia (trapping data pooled over 2 years). House Finches could not be reliably assigned to age classes in December. Month % AHY % HY Total banded Jan 100.0 0.0 21 Feb 100.0 0.0 16 Mar 100.0 0.0 8 Apr 100.0 0.0 10 May 16.3 83.7 43 Jun 8.3 91.7 36 Jul 6.1 93.9 49 Aug 18.2 81.8 40 Sep 33.3 66.7 62 Oct 48.1 51.9 34 Nov 38.9 61.1 35 Dec N/A N/A 32 Totals 34.9 65.1 386 were recorded up to five times. Combined, 80.4% of all House Finches banded were en- countered between one and five times at the feeding station, and 19.6% of all House Finch- es were encountered more than five times. Only 6.7% of all birds were encountered 20 or more times (Fig. lA). Examination of the time period between first and last encounter for all individuals encountered two or more times indicated 44% of the House Finches were present less than 50 days (Fig. IB), while only 14% were encountered over a span longer than 1 year. There were significant or nearly significant effects of age, gender, month, and an age*conjunctivitis interaction effect on num- ber of times an individual was encountered (Table 2). There were no significant effects or significant interaction effects in the second ANOVA using time span between first and last encounter as the dependent variable. Birds captured in April, August, September, Octo- ber, November, and December were all en- countered more than the mean of 4.15 en- counters (Fig. 2A). Birds captured in March, April, August, October, and December were subsequently encountered for longer time spans than the mean of 79.5 days (Fig. 2B). Birds captured in May, June, and July (mostly HYs; Table 1) were encountered less than 50 days. Birds captured in June were encountered only 1 .5 times on average, for fewer than 50 days, but yet in June there were large flocks TABLE 2 ANOVA examining factors influencing number of times House Finches were encountered at a feeding station in Atlanta, Georgia. The nonsignificant interaction of Gender*Conjunctivitis is not shown. Only birds of known age and gender were used in the analysis. Variable df Mean square F p Age 1 0.784 3.513 0.063 Gender 1 0.969 4.345 0.039 Conjunctivitis 1 0.427 1.912 0.168 Month 11 0.408 1.828 0.052 Age * Conjunctivitis 1 0.998 4.474 0.036 Gender * Age 1 0.273 1.223 0.270 Error 176 0.223 at the feeding station (Fig. 2C). Conversely, in April, there were few House Finches ob- served at the feeders, but those that were cap- tured in this month (all adults; Table 1) were later frequently encountered (~5 times) for ~5 months later (150 days). The effects of age and gender on number of encounters per individual varied (Fig 3 A). Young House Finches were encountered more often on average than adults, and females were encountered more often than males. Overall, young females were encountered most often. Birds initially captured with con- junctivitis were not subsequently encountered differently than those without (Table 2, Fig. 3B). However, there was a significant inter- action of age and conjunctivitis (Table 2, Fig. 3B); adults with conjunctivitis had a slightly higher encounter rate than those without, but young birds with conjunctivitis had a lower rate than those without. DISCUSSION There was high seasonal turnover of House Finches at this suburban site in Georgia. High turnover rates of this species have been doc- umented in a number of studies in eastern (Hamilton 1992) and western North America (McClure 1989). Only a small percentage of House Finches at my site frequented the feed- ing station over 20 times (6.7%) and for lon- ger than 1 year (14%). These numbers indi- cate that only a small fraction of the House Finches observed at backyard bird feeders in suburban Atlanta, Georgia remain at individ- ual sites year-round. Thus, most House Finch- es at my (and perhaps other) feeders must 374 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 X) 1 5 10 15 20 25 30 35 40 More Number of times encountered Number of days between first and last encounter FIG. 1. A. Distribution of all House Finch encounters over 24 months. Values above bars show percentages of the total number of banded individuals (n = 386). B. Number of days between first (initial capture) and last encounters (final recapture or reobservation) of House Finches over 24 months. Only individuals with two or more encounters are included. Values above bars show percentages of the total number of banded individuals with two or more encounters (n = 172). have been either transient individuals or young of the year that rarely remained at the site. I trapped 386 separate House Finches at this one site over the 2-year period and many new birds were trapped each month, even though the average number of birds seen at the site was <6 (Fig. 2C). Thus, it is likely that most House Finches observed at backyard feeders are infrequently the same individuals. My data provide additional information on factors influencing feeder fidelity in House Finches. For example, there was a seasonal component to the turnover rate at this site. House Finches trapped in April, which were invariably adults (Table 1), tended to frequent the site throughout the summer, but not much longer. These probably represented adults nesting in the area throughout the summer. Flock sizes increased in May and June, which could be explained as adults and their fledg- lings frequenting the feeding station. Birds trapped in these months (mostly young; Table 1) were rarely seen after their initial trapping. Encounter rates increased in August and re- mained high through the fall. Most of the birds trapped during these months were HYs, Davis • HOUSE FINCH FIDELITY TO FEEDING STATIONS 375 FIG. 2. A. Mean number of encounters of normal House Finches per month. B. Mean length of time between first and last encounter of normal House Finches per month. C. Monthly variation in flock sizes of House Finches observed at feeders over 24 months. Dashed lines indicate 24-month averages. although they would have been independent of their parents and perhaps traveling in loose flocks (Hill 1993). Seasonal turnover and fidelity to this site were characterized by high fidelity of adults during the breeding season, low fidelity of ju- veniles early in the summer, but becoming higher in late summer. Most birds captured in fall were encountered for up to 3 months. These might represent wintering individuals. Fidelity to this site declined in late winter and, by February, most birds captured were rarely 376 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 AHY-No Conj. (106) AHY-Conj. (6) HY-No Conj. (185) HY-Conj. (24) FIG. 3. A. Mean number of encounters of normal House Finches in all age and gender classes. B. Mean number of encounters of House Finches in each age/conjunctivitis category. Sample sizes for each class shown in parentheses in both charts. seen again. Adult breeders arrived in April and remained throughout the summer. Direct comparisons of encounter rates by age and gender classes indicated young House Finches were encountered more frequently than adults and females were encountered more than males (Fig. 3A). If the number of encounters represents relative feeder use, this result has important implications for transmis- sion of MG. Roberts et al. (2001b) found that young birds were more likely to be clinically infected than adults. Moreover, young female House Finches were found to have conjunc- tivitis more than any other age and gender class (Altizer et al. 2004b). My results indi- cate that young females are encountered most often at a feeding station. Therefore it may be possible that younger individuals, especially females, contract this disease more often be- cause they visit bird feeders more than other individuals, thereby coming in contact with any infected individuals or contaminated sur- faces more frequently. Also of importance to MG dynamics is the extremely high turnover of individuals I ob- served as 386 separate birds were trapped at a single feeding station during this 2-year study. If this feeder site typifies the use of other feeders in this geographic region, this turnover rate is likely a factor in the trans- mission of MG among House Finches. More- over, increased feeder use by individuals dur- ing fall compared to other seasons (Fig. 2A) occurs in the same time period as annual out- breaks of MG in this area (Altizer et al. 2004b), thereby increasing the likelihood of transmission between individuals through contact at feeders. The results of this study serve as an ex- ample of how the practice of bird feeding can influence the life history of a bird species. House Finches clearly have adapted to using Davis • HOUSE FINCH FIDELITY TO FEEDING STATIONS 377 feeders in eastern North America (Dhondt et al. 1998) and this study shows they vary in their fidelity to these feeders by age, gender, and time of year. Increased feeder use during certain time periods or by select age/gender classes could be an important driver of the spread of Mycoplasma gallisepticum. ACKNOWLEDGMENTS Sonia Altizer provided assistance on all aspects of the project. Katherine Cook and Elizabeth Lindsey helped with trapping finches and recording data. Fund- ing was partly provided by an NSF grant to Andre Dhondt (DEB -009445 6) and by an Emory University Research grant. LITERATURE CITED Able, K. R and J. R. Belthoff. 1998. Rapid ‘evolu- tion’ of migratory behaviour in the introduced House Finch of eastern North America. Proceed- ings of the Royal Society of London Series B 265: 2063-2071. Altizer, S. M., W. M. Hochachka, and A. A. Dhondt. 2004a. Seasonal dynamics of mycoplas- mal conjunctivitis in eastern North American House Finches. Journal of Animal Ecology 73: 309-322. Altizer, S., A. K. Davis, K. C. Cook, and J. J. Cher- ry. 2004b. Age, sex, and season affect the risk of mycoplasmal conjunctivitis in a southeastern House Finch population. Canadian Journal of Zo- ology 82:755-763. Dhondt, A. A., D. L. Tessaglia, and R. L. Slothow- ER. 1998. Epidemic mycoplasmal conjunctivitis in House Finches from eastern North America. Jour- nal of Wildlife Diseases 34:265-280. Egbert, J. R. and J. R. Belthoff. 2003. Wing shape in House Finches differs relative to migratory habit in eastern and western North America. Con- dor 105:825-829. Farmer, K. L., G. E. Hill, and S. R. Roberts. 2002. Susceptibility of a naive population of House Finches to Mycoplasma gallisepticum. Journal of Wildlife Diseases 38:282-286. Faustino, C., C. S. Jennelle, V. Connolly, A. K. Da- vis, E. C. SwARTHOUT, A. Dhondt, and E. G. CoocH. 2004. Mycoplasma gallisepticum infection dynamics in a House Finch population: analysis of seasonal variation in survival and transmission rate. Journal of Animal Ecology 73:651-669. Hamilton, T. R. 1991. Seasonal movement of House Finches in the Midwest. North American Bird Bander 16:119-122. Hamilton, T. R. 1992. Turnover within a population of House Finches in the Midwest. North American Bird Bander 17:116-118. Hartup, B. K., a. a. Dhondt, K. V. Sydenstricker, W. M. Hochachka, and G. V. Kollias. 2001. Host range and dynamics of mycoplasmal con- junctivitis among birds in North America. Journal of Wildlife Diseases 37:72-81. Hill, G. E. 1992a. An inexpensive source of colored leg bands. Journal of Field Ornithology 63:408- 410. Hill, G. E. 1992b. Proximate basis of variation in ca- rotenoid pigmentation in male House Finches. Auk 109:1-12. Hill, G. E. 1993. House Finch (Carpodacus mexican- us). The birds of North America. Number 46. Hill, G. E. 1998. Plumage redness and pigment sym- metry in the House Finch. Journal of Avian Bi- ology 29:86-92. Hill, G. E. 2002. A red bird in a brown bag. The function and evolution of colorful plumage in the House Finch. Oxford University Press, New York, USA. Hill, G. E., P. M. Nolan, and A. M. Stoehr. 1999. Pairing success relative to male plumage redness and pigment symmetry in the House Finch: tem- poral and geographic constancy. Behavioral Ecol- ogy 10:48-53. Hilton, B. 1994. Carpodacus finches in South Caro- lina’s piedmont: migration, sex ratios, site fidelity, and longevity. North American Bird Bander 19: 1-11. McClure, H. E. 1989. Epizootic lesions of House Finches in Ventura County, California. Journal of Field Ornithology 60:421-430. Milligan, J., A. K. Davis, and S. M. Altizer. 2003. Errors associated with using colored leg bands to identify wild birds. Journal of Field Ornithology 74:111-118. Pyle, P. 1997. Identification guide to North American Birds. Part 1. Slate Creek Press, Bolinas, Califor- nia, USA. Roberts, S. R., P. M. Nolan, and G. E. Hill. 2001a. Characterization of Mycoplasma gallisepticum in- fection in captive House Finches {Carpodacus mexicanus) in 1998. Avian Diseases 45:70—75. Roberts, S. R., P. M. Nolan, L. H. Lauerman, L.-Q. Li, and G. E. Hill. 2001b. Characterization of the mycoplasmal conjunctivitis epizootic in a Hou.se Finch population in the southeastern USA. Journal of Wildlife Disea.ses 37:82-88. The Wilson Journal of Ornithology 120(2):378-383, 2008 GENDER IDENTIFICATION OF CASPIAN TERNS USING EXTERNAL MORPHOLOGY AND DISCRIMINANT FUNCTION ANALYSIS JOSHUA T. ACKERMAN,'* JOHN Y. TAKEKAWA,^ JILL D. BLUSO,'-^ JULIE L. YEE,'' AND COLLIN A. EAGLES-SMITH' * ABSTRACT— Caspian Tern {Sterna caspia) plumage characteristics are sexually monochromatic and gender cannot easily be distinguished in the field without extensive behavioral observations. We assessed sexual size dimorphism and developed a discriminant function to assign gender in Caspian Terns based on external mor- phology. We collected and measured Caspian Terns in San Francisco Bay, California, and confirmed their gender based on necropsy and genetic analysis. Of the eight morphological measurements we examined, only bill depth at the gonys and head plus bill length differed between males and females with males being larger than females. A discriminant function using both bill depth at the gonys and head plus bill length accurately assigned gender of 83% of terns for which gender was known. We improved the accuracy of our discriminant function to 90% by excluding individuals that had less than a 75% posterior probability of correctly being assigned to gender. Caspian Terns showed little sexual size dimorphism in many morphometries, but our results indicate they can be reliably assigned to gender in the field using two morphological measurements. Received 19 April 2007. Accepted 15 August 2007. Many seabird species are sexually mono- morphic in plumage characteristics and gender cannot easily be assigned without extensive behavioral sampling (such as gender-specific breeding behaviors or vocalizations) (Coulter 1986, Chardine and Morris 1989, Phillips and Furness 1997, Casaux and Baroni 2000), lap- arotomy (Quinn 1990, Stem and Jarvis 1991) or subsequent laboratory genetic analyses (Jodice et al. 2000, Genovart et al. 2003, Quintana et al. 2003, Devlin et al. 2004, Se- tiawan et al. 2004). Discriminant analysis of external morphometries is a mathematical ap- proach that has been widely used to assign gender of seabirds, including several gulls {Larus spp.) (Banners and Patton 1985, Evans et al. 1993, Mawhinney and Diamond 1999, Torlaschi et al. 2000), terns {Sterna spp. and ' uses. Western Ecological Research Center, Davis Field Station, One Shields Avenue, University of Cal- ifornia, Davis, CA 95616, USA. 2 uses. Western Ecological Research Center, San Francisco Bay Estuary Field Station, 505 Azuar Drive, Vallejo, CA 94592, USA. 3 Flumboldt State University, Wildlife Department, One Harpst Street, Areata, CA 95521, USA. ^ uses. Western Ecological Research Center, 3020 State University Drive East, Modoc Hall, Room 3006, Sacramento, CA 95819, USA. 5 U.S. Fish and Wildlife Service, 2800 Cottage Way, Suite W-2605, Sacramento, CA 95825, USA. ^ Corresponding author; e-mail; jackerman@usgs.gov Chlidonias spp.) (Coulter 1986, Stern and Jar- vis 1991, Devlin et al. 2004, Fletcher and Hamer 2003, Bluso et al. 2006), petrels {Ful- marus spp., Thalassoica spp., Daption spp., and Pagodroma spp.) (Van Franeker and Ter Braak 1993, Weidinger and Van Franeker 1998), jaegers {Stercorarius spp.) (Phillips and Furness 1997), shearwaters {Puffinus spp.) (Genovart et al. 2003), noddy (Anous spp.) (Chardine and Morris 1989), shags (Phalacrocorax spp.) (Casaux and Baroni 2000, Quintana et al. 2003), kittiwakes (Rissa spp.) (Jodice et al. 2000), and Yellow-eyed Penguins (Megadyptes antipodes) (Setiawan et al. 2004). Discriminant analysis produces a linear combination of morphological variables that best describes the distinction between known males and females (Khattree and Naik 2000). One prior study examined gender differenc- es in Caspian Tern {Sterna caspia) morphol- ogy (Quinn 1990), but the discriminant func- tion, which accurately assigned gender to 77% of the terns, did not incorporate head plus bill length. The distance from the back of the head to the tip of the bill (hereafter head-to-bill length) is widely used in discriminant function analyses to differentiate gender in many Lar- idae (Hanners and Patton 1985, Stem and Jar- vis 1991, Phillips and Furness 1997, Fletcher and Hamer 2003, Devlin et al. 2004). This single measurement has distinguished gender 378 Ackerman et al. • GENDER IDENTIFICATION OF CASPIAN TERNS 379 with 88-98% accuracy in some gull species (Coulson et al. 1983, Mawhinney and Dia- mond 1999, Jodice et al. 2000, Torlaschi et al. 2000) and linear combinations of head-to-bill length with other morphological characteris- tics has increased the ability to accurately as- sign gender of several tern species (Stern and Jarvis 1991, Fletcher and Hamer 2003, Devlin et al. 2004). Our objective was to examine sexual size dimorphism of Caspian Terns in San Francisco Bay, California, USA, and de- velop a discriminant function incorporating head-to-bill length and several other external measurements. METHODS Study Area. — San Francisco Bay, California (37.8° N, 122.3° W) is the largest estuary on the west coast of North America. Caspian Terns breed at as many as 13 different sites in San Francisco Bay (Strong et al. 2004). We sampled Caspian Terns near three breeding colonies, ineluding a site in the North Bay (Napa River near the Carquinez Strait) and two sites in the South Bay on the Don Ed- wards San Francisco Bay National Wildlife Refuge (Alviso Slough and Coyote Hills Pond 2A). Collections and Measurements. — We col- lected terns by shotgun under California De- partment of Fish and Game Scientific Collec- tion permits (SC-801034-05 and SC-007250), U.S. Fish and Wildlife Service permit (MB 120 154-2), and guidelines of the USGS, Western Ecological Research Center Animal Care and Use Committee. We measured cul- men length, bill depth at the gonys, head-to- bill length, tarsus length (tarsometatarsus bone), wing length (carpal joint to the end of the longest straightened primary), length of rectrices R1 and R6 (R1 was the central most rectrix on right side, R6 was the outer most rectrix on right side), and body mass for each tern. We measured tern morphology to the nearest 0.01 mm with digital calipers (Fisher Scientific, Hampton, NH, USA), except wing chord and tail measurements, which were measured to the nearest 1 .0 mm with a stopped wing rule. We measured body mass to the nearest 1.0 g with a 1-kg Pesola spring scale (Pesola AG, Baar, Switzerland). One of two researchers measured terns to reduce any observer related variation in measurement er- ror. We examined the gonads of each tern via necropsy and verified gender by genetic anal- ysis. A drop of blood was collected from each tern for gender analysis of the chromo-heli- case-DNA binding protein gene (e.g., Jodice et al. 2000, Quintana et al. 2003) at Zoogen Services Inc.®, Davis, California, USA. Analyses. — We used analysis of variance to test differences in morphological measure- ments between male and female terns. We ex- amined sexual size dimorphism (SSD) in terns by calculating the absolute value of the dif- ference between the mean morphological measurement for females and males, and di- viding this quantity by the mean value for males. We calculated the best measurements for classifying gender of Caspian Terns with a forward stepwise discriminant function anal- ysis using PROC STEPDISC in SAS software (Morrison 1990, SAS Institute 2004). We used a criterion based on an F-test of Wilks’ lamb- da (A) at each step of the analysis to enter the variable contributing the most or to remove the variable contributing the least diserimina- tory power to the model until no further var- iables could be entered or removed at the 0.15 significance level. We believed that a random- ly captured tern was equally likely to be fe- male or male in the absence of morphometric data; therefore, we used a prior probability of 50% for the likelihood of being female. We used five of the eight morphological measure- ments recorded in our discriminant function analysis; we excluded body mass and retrices R1 and R6, because they are more pliable and can vary over time (Kaufman 1983, Voelker 1997). We defined discriminant scores (here- after D) as D = — 0.5(jc - |jl)'S“'(jc - |jl), where (x — |jl)'S“^(jc — jjl) represented the squared distance from a tern with measure- ments jc to a subpopulation with mean |jl and variance matrix X (Khattree and Naik 2000, SAS Institute 2004). We calculated discrimi- nant scores, and using the mean and variance of the respective genders. We classified terns into the gender for which the smallest squared distance, or the largest D score was measured. We classified terns as males if D^aie ■Female' defined as was >0 and as females if PMaie-remaie We simplified discriminant scores into linear expressions without changing the effect of scoring between genders by assuming the var- 380 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 TABLE 1. Morphological measurements (mean ± SD) and sexual in San Francisco Bay, California, 2006. size dimorphism (SSD) of Caspian Terns Measurement Female® Male® ^1,38 p SSD (%) Bill depth (mm) 15.19 ± 0.66 16.60 ± 0.99 27.87 <0.0001 8.49 Head-to-bill (mm) 134.00 ± 3.24 137.22 ± 3.23 9.92 0.003 2.35 Tarsus (mm) 45.30 ± 2.36 46.45 ± 3.22 1.66 0.21 2.48 Culmen length (mm) 67.27 ± 2.28 68.29 ± 2.96 1.49 0.23 1.49 Wing chord (mm) 429.10 ± 6.95 430.80 ± 10.60 0.36 0.55 0.39 Tail retrix 1 (mm) 105.90 ± 4.46 109.65 ± 7.44 3.74 0.06 3.42 Tail retrix 6 (mm) 150.10 ± 8.60 151.50 ± 9.76 0.23 0.63 0.92 Tail R6~ Ri 44.20 ± 8.55 41.85 ± 6.48 0.96 0.33 5.62 Mass (g) 669.94 ± 69.85 662.02 ± 45.13 0.18 0.67 1.20 a Gender confirmed by necropsy and genetic analysis. iance S was constant (SAS Institute 2004). We produced classification error rates with a re- substitution analysis in SAS and validated our discriminant functions using a cross-valida- tion procedure (Lachenbruch and Mickey 1968) where each tern was classified using a function derived from the total sample ex- cluding the tern in question (e.g., Chardine and Morris 1989, Phillips and Furness 1997). We also calculated posterior probabilities of a tern being female (Probability = 1/[1 + and plotted these values against their corresponding discriminant score. This allowed us to calculate cut-off points for discriminant scores that had a 75% probability of being a female or male. RESULTS We collected 40 Caspian Terns from 19 April to 15 June 2006 in San Francisco Bay. Twenty were females and 20 were males by necropsy with gender confirmed using genetic analysis. Gender assigned via necropsy was 100% in concordance with that from genetic analysis. Two of the eight morphological char- acteristics we measured differed between males and females with bill depth at the gonys and head-to-bill length showing the least amount of overlap between genders (Table 1). The largest proportional differences in struc- tural measurements between males and fe- males were also greatest in bill depth, fol- lowed by tail retrix R1 length, tarsus length, and head-to-bill length (Table 1). Bill depth at the gonys and head-to-bill length were the best structural measurements separating male and female Caspian Terns (Wilks’s A = 0.50: ^2,37 = 18.62, P < 0.0001). The discriminant function with these two mor- phometries correctly classified 80% of known females and 85% of known males (83% cor- rect classification rate overall). The leave-one- out cross-validation test (Lachenbruch and Mickey 1968) also correctly classified gender of 83% of the terns. The discriminant scores were: Dp^^^ie ~ t)ill depth (22.1297) + head- to-bill length (12.8779) - 1,030.9089, and Dj^aie “ t)ill depth (24.1306) + head-to-bill length (13.1927) - 1,105.3964. Thus, Function 1 was: D^ale-Female - t)ill depth (2.0008) + head-to-bill length (0.3148) - 74.4875 and we classified terns as males when DMaie-Femaie was >0 and as females when DMale-Female WaS <0. There was some overlap in morphological measurements between males and females where the probability of correctly classifying gender was reduced (Fig. 1). Terns with dis- criminant scores from -1.10 to 1.10 had <75% probability of being correctly assigned to gender (Fig. 2). Twenty-five percent (10 of 40) of the terns we collected were within this overlapping range. Of the 30 terns with dis- criminant scores outside these cutoff points, the discriminant function correctly classified 90% of the individuals (Fig. 2). A second discriminant function using only bill depth at the gonys also was successful in classifying gender (Wilks’ A = 0.58: Fi 3g = 27.87, P < 0.0001). The discriminant scores for this function were: ^ t)ill depth (21.4758) - 163.1193, and D^aie = bill depth (23.4606) - 194.6647. Thus, Function 2 was: D^ale-Femaie = bill dcpth (1.9849) ~ 31.5454. This simplified discriminant function cor- rectly classified 80% of known females and Ackerman et al. • GENDER IDENTIFICATION OF CASPIAN TERNS 381 FIG. 1. Discriminant function using bill depth at the gonys and head-to-bill length to classify female (below solid line) and male (above solid line) Caspian Terns in San Francisco Bay, California. Area between the stippled lines indicates morphological overlap where the discriminant function had <75% probability of correctly classifying gender. 75% of known males (78% correct classifi- cation rate overall) indicating that bill depth was the most important characteristic for dif- ferentiating gender. The cross-validation test also correctly classified 78% of the terns of known gender for Function 2. DISCUSSION Caspian Terns in San Francisco Bay had only slight sexual size dimorphism in most morphological characteristics we measured (Table 1). However, males were significantly larger than females in bill depth and head-to- bill length. Bill depth at the gonys was the most important morphological measurement in classifying gender of Caspian Terns and this variable alone correctly classified 78% of the terns. A discriminant function incorporat- ing both bill depth and head-to-bill length im- proved our ability to accurately assign gender of Caspian Terns to 83%. Head-to-bill length and bill depth often are the most important measurements in discriminant functions for seabirds, especially larids (Hanners and Patton 1985, Van Franeker and Ter Braak 1993, Phil- lips and Furness 1997, Mawhinney and Dia- mond 1999, Torlaschi et al. 2()()0, Genovart et al. 2003, Devlin et al. 2004). Our accuracy is slightly better than that reported by Quinn (1990) who developed an earlier discriminant function for Caspian Terns in Texas. He found that 77% of Caspian Terns could be accurately Discriminant score FIG. 2. Probability of being female in relation to the discriminant function scores based on bill depth and head-to-bill length of Caspian Terns in San Fran- cisco Bay, California. All Caspian Terns with discrim- inant function scores <0 were classified as females and >0 as males in validation procedures. Lines indicate the cutoff points for discriminant scores of 1.10 and — 1.10 if the probability of being female was set to 0.25 and 0.75, respectively. assigned to gender using a function incorpo- rating several morphological measurements including bill depth and culmen length, but not head-to-bill length. Quinn (1990) found that head-to-bill length was significantly larg- er in male than female Caspian Terns, but this morphometric was not incorporated into his discriminant function. Our accuracy of 83% for Caspian Terns is comparable to other studies of Laridae. Our discriminant function based on bill depth and head-to-bill length was 5-11% more accurate than functions of multiple morphological characteristics developed for several other tern species (Coulter 1986, Quinn 1990, Stern and Jarvis 1991, Fletcher and Hamer 2003, Devlin et al. 2004), but 3% lower than the function we developed for Forster’s Terns {Sterna for- steri) (Bluso et al. 2006). For example, using model validation procedures, 73 and 74% of Arctic Terns {S. paraciisaea) (Fletcher and Hamer 2003, Devlin et al. 2004), 72 and 78% of Common Terns (S. hirundo) (Coulter 1986, Fletcher and Hamer 2003), 78% of Black Terns (Chlidonias niger) (Stern and Jarvis 1991), and 86% of Forster’s Terns (Bluso et al. 2006) were correctly classified as male or female. Discriminant functions that have been de- veloped for gulls are, in general, more accu- 382 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 rate than functions derived for terns. Using multiple morphological measurements, dis- criminant functions for gulls have accurately classified gender of 90-99% of individuals (Mills 1971, Manners and Patton 1985, Evans et al. 1993, Mawhinney and Diamond 1999, Torlaschi et al. 2000). The difference in clas- sification rates between Caspian Terns and other Laridae suggests that Caspian Terns ex- hibit greater sexual size dimorphism than sev- eral other terns, but are not as dimorphic as most gull species. We observed some overlap in morphologi- cal measurements between male and female Caspian Terns where the probability of cor- rectly classifying an individual was reduced (Fig. 1). Posterior probabilities can be calcu- lated and individuals with probabilities <75% can be excluded to increase the probability of correctly classifying a tern as male or female. Ten of 40 Caspian Terns we collected were within this range (Fig. 2). Our accuracy in- creased to 90% after excluding these terns. We caution that excluding terns within this over- lapping size range could introduce bias into some studies because researchers would nec- essarily be classifying only the largest males and smallest females. ACKNOWLEDGMENTS This research was funded by the CALFED Ecosys- tem Restoration Program (grant # ERP-02D-C12) with additional support from the USGS, Western Ecological Research Center. We thank Sarah Stoner-Duncan, Ter- ry Adelsbach, John Henderson, Mark Ricca, Cathy Johnson, Carolyn Marn, Angela Rex, Joe Northrup, Lani Stinson, Maliheh Nakhai, Stacy Moskal, Eli French, Kristen Dybala, Brooke Hill, and Susan Wain- wright-De La Cruz for field assistance. We also thank Clyde Morris, Joy Albertson, Mendel Stewart, Joelle Buffa, Marge Kolar, Eric Mruz, John Krause, Tom Huffman, Larry Wyckoff, Carl Wilcox, Karen Taylor, Rick Morat, Lew Allen, Carol Atkins, Donna Podger, Nicole Athearn, Can Duck Club, California Depart- ment of Fish and Game, and staff at the Don Edwards San Francisco Bay National Wildlife Refuge (Special Use Permit 11640-2006-006) for logistical support, technical assistance, and access. Early versions of the manuscript were improved from comments by Cheryl Strong and two anonymous reviewers. The use of trade, product, or firm names in this publication is for descriptive purposes only and does not imply endorse- ment by the U.S. Government. LITERATURE CITED Bluso, J. D., J. T. Ackerman, J. Y. Takekawa, and J. L. Yee. 2006. Using morphological measurements to sex Forster’s Terns. Waterbirds 29:511-516. Casaux, R. and a. Baroni. 2000. Sexual size dimor- phism in the Antarctic Shag. Waterbirds 23:489- 493. Chardine, j. W. and R. D. Morris. 1989. Sexual size dimorphism and assortative mating in the Brown Noddy. Condor 91:868—874. CouLSON, J. C., C. S. Thomas, J. E. L. Bufferfield, N. Duncan, P. Monaghan, and C. Shedden. 1983. The use of head and bill length to sex live gulls Laridae. Ibis 125:549-557. Coulter, M. C. 1986. Assortative mating and sexual dimorphism in the Common Tern. Wilson Bulletin 98:93-100. Devlin, C. M., A. W. Diamond, and G. W. Saunders. 2004. Sexing Arctic Terns in the field and labo- ratory. Waterbirds 27:314-320. Evans, D. R., E. M. Hoopes, and C. R. Griffin. 1993. Discriminating the sex of Laughing Gulls by lin- ear measurements. Journal of Field Ornithology 64:472-476. Fletcher, K. L. and K. C. Hamer. 2003. Sexing terns using biometrics: the advantage of within-pair comparisons. Bird Study 50:78—83. Genovart, M., M. McMinn, and D. Bowler. 2003. A discriminant function for predicting sex in Bal- earic Shearwater. Waterbirds 26:72-76. Hanners, L. a. and S. R. Patton. 1985. Sexing Laughing Gulls using external measurements and discriminant analysis. Journal of Field Ornitholo- gy 56:158-164. JODICE, P. G. R., R. B. Lanctot, V. A. Gill, D. D. Roby, and S. A. Hatch. 2000. Sexing adult Black-legged Kittiwakes by DNA, behavior, and morphology. Waterbirds 23:405-415. Kaufman, K. E. 1983. Tail moult of Forster’s Tern. British Birds 76:357. Khattree, R. and D. N. Naik. 2000. Multivariate data reduction and discrimination with SAS Software. SAS Institute Inc., Cary, North Carolina, USA. Lachenbruch, P. a. and M. R. Mickey. 1968. Esti- mation of error rates in discriminant analysis. Technometrics 10:1-11. Mawhinney, K. and T. Diamond. 1999. Sex determi- nation of Great Black-backed Gulls using mor- phometric characters. Journal of Field Ornitholo- gy 70:206-210. Mills, J. A. 1971. Sexing Red-billed Gulls from stan- dard measurements. Journal of Marine and Fresh- water Research 5:326-328. Morrison, D. F. 1990. Multivariate statistical methods. Third Edition. McGraw Hill Inc., New York, USA. Phillips, R. A. and R. W. Furness. 1997. Predicting the sex of Parasitic Jaegers by discriminant anal- ysis. Waterbirds 20:14-23. Quinn, J. S. 1990. Sexual size dimorphism and paren- tal care patterns in a monomorphic and dimorphic larid. Auk 107:260-274. Quintana, E, G. Somoza, M. Uhart, C. Cassara, P. Gandini, and E. Frere. 2003. Sex determination of adult Rock Shags by molecular sexing and Ackerman et al. • GENDER IDENTIFICATION OF CASPIAN TERNS 383 morphometric parameters. Journal of Field Orni- thology 74:370-375. SAS Institute Inc. 2004. SAS OnlineDoc®. Version 9.1.2. SAS Software, SAS Institute Inc., Cary, North Carolina, USA. http://v8doc.sas.com/ sashtml/ (accessed 28 September 2006). Setiawan, a. N., J. T. Darby, and D. M. Lambert. 2004. The use of morphometric measurements to sex Yellow-eyed Penguins. Waterbirds 27:96-101. Stern, M. A. and R. L. Jarvis. 1991. Sexual dimor- phism and assortative mating in Black Terns. Wil- son Bulletin 103:266-271. Strong, C. M., L. B. Spear, T. P. Ryan, and R. E. Dakin. 2004. Forster’s Tern, Caspian Tern, and California Gull colonies in the San Francisco Bay: habitat use, population sizes and trends, 1982- 2003. Waterbirds 27:411-423. Torlaschi, C., P. Gandini, E. Frere, and R. M. Peck. 2000. Predicting the sex of Kelp Gulls by external measurements. Waterbirds 23:518-520. Van Franeker, J. A. and C. J. F. Ter Braak. 1993. A generalized discriminant for sexing fulmarine petrels from external measurements. Auk 110: 492-502. VoELKER, G. 1997. The molt cycle of the Arctic Tern, with comments on aging criteria. Journal of Field Ornithology 68:400-412. Weidinger, K. and j. A. Van Franeker. 1998. Ap- plicability of external measurements to sexing of the Cape Petrel Daption capense at within-pair, within-population and between-population scales. London Journal of Zoology 245:473-482. The Wilson Journal of Ornithology 120(2):384-389, 2008 EFFECTS OF TRAFFIC NOISE ON AUDITORY SURVEYS OF URBAN WHITE-WINGED DOVES JEFFREY B. BREEDEN, '•'* FIDEL HERNANDEZ,'-’ RALPH L. BINGHAM,' NOVA J. SILVY,2 AND GARY L. WAGGERMAN’ ABSTRACT. We investigated the effects of urban noise on auditory surveys of White-winged Doves {Zen- aida asiatica) in two major cities in Texas. We conducted auditory point counts throughout the morning in San Antonio (n = 6) and Austin (n = 10) during week days (when traffic noise is higher) and weekends. We categorized survey points as near or far from roads (<0.8 and >0.8 km, respectively) for comparison. We documented no difference in density estimates in Austin between week days (46 ± 10 pairs/ha) and weekends (52 ± 10 pairs/ha- P = 0.23); however, weekend estimates were consistently higher throughout the morning. Weekend density estimates in San Antonio were higher after 0620 hrs (P < 0.04), the time coinciding with beginning of the morning commute during week days in this city. We documented that weekend estimates (45 ± 5 pairs/ha) were higher than week day estimates (33 ± 5 pairs/ha) for points near roads (within 0.8 km; P = 0.02) but not for points far from roads {P = 0.16). Our results indicate that traffic noise can bias auditory surveys. Survey methods that account for probability of detection should be used to correct for potential noise bias. Received 24 May 2007. Accepted 20 September 2007. Auditory point-count indices are commonly used to estimate abundanee of avian popula- tions. The assumption is the number of birds heard calling provides an accurate index to abundance (Sisson 1968, Keppie et al. 1970); however, the validity of indiees recently has been questioned (Anderson 2001, Thompson 2002). Faetors sueh as observer variability, habitat variables, and weather ean affect de- tection of calling birds potentially biasing sur- veys (LaPerriere and Haugen 1972, Shields 1977, Basket! et al. 1978). Human-induced disturbances such as traffic noise can also bias auditory surveys. Assessing potential bias in auditory surveys due to traffic noise is im- portant considering the need to aceurately as- sess the status of wildlife populations and the likelihood that traffic noise will increase with an expanding human population. Little research exists regarding the potential influenee of traffic noise on auditory survey ' Caesar Kleberg Wildlife Research Institute, Texas A&M University-Kingsville, Kingsville, TX 78363, USA. 2 Department of Wildlife and Fisheries Sciences, Texas A&M University, College Station, TX 78840, USA. 2 Texas Parks and Wildlife Department, 1221 Ranch Road 12, San Marcos, TX 78666, USA. Current address: Department of Animal Sciences, Box T-0070, Tarleton State University, Stephenville, TX 76402, USA. ^ Corresponding author; e-mail: fidel.hernandez@tamuk.edu methods in urban areas. Due to their calling frequency and urban populations, the White- Winged Dove (Zenaida asiatica) is a good species to study to gain insight into the effects of noise disturbance on auditory counts. West (1993) evaluated the accuracy of the White- winged Dove call-count index in San Antonio, Texas, but did not assess the effects of noise. Emlen and DeJong (1981) suggested that sur- veys should be conducted only when back- ground-noise level was within acceptable lim- its with correction factors applied for noise bias, if necessary. Dawson (1981), however, reported that environmental noise had little ef- fect on number of Grey Gerygone {Gerygone igata) counted in New Zealand forests, but suggested that it could act through effects on observers and birds. Historically, White-winged Doves occurred in rural environments in southern Texas and northern Mexico during the breeding season. The species has been expanding its distribu- tion northward and increasing in density in Texas, and has become common in urban ar- eas outside its historic range (George et al. 1994, Waggerman 2001). Urban surveys were initiated by the Texas Parks and Wildlife De- partment (TPWD) in San Antonio and Austin during the late 1980s in response to the range expansion. TPWD biologists assumed that traffic volume, and thus traffic noise, was greater during early morning on week days, because of commuter traffic, rather than on 384 Breeden et al. • WHITE- WINGED DOVE AUDITORY SURVEYS 385 weekends when designing these urban sur- veys. Protocol therefore mandated that urban surveys be conducted only during early morn- ing on weekends in an effort to reduce poten- tial noise bias. However, this constrained sam- pling period allowed little time to complete urban surveys. Thus, more effort was required by biologists and volunteers to complete ur- ban surveys within the specified time period making the process less efficient. Additional urban areas need to be moni- tored in the future with the continuing north- ward range expansion of White-winged Doves. Thus, knowledge of the potential im- pacts of urban noise on auditory counts is needed. The objective of this study was to ex- amine if abundance estimates obtained through auditory counts for urban populations of White-winged Doves varied between time periods (i.e., week day vs. weekend) when traffic noise differed. METHODS We selected a sample of survey points from the total used in the annual survey of White- winged Doves from Austin {n = \0 selected; n = 122-204 total) and San Antonio (n = 6 selected; n = 219-241 total), Texas as de- scribed by Breeden et al. (2004). Research has indicated that calling frequency of some col- umbids may be more variable at lower den- sities (Cohen et al. 1960, Keppie et al. 1970, Viers 1970). Thus, we chose points at which previous surveys estimated dove densities were >25 pairs/ha to reduce potential bias as- sociated with lower densities. We conducted auditory counts during 15 May- 15 June 2004 following Rappole and Waggerman (1986). We calculated density estimates (pairs/ha) at each survey point based on observed calling intensity following Uzzell and Kiel (1950) and Sepulveda et al. 2006. We surveyed each point twice: once during a week day (Mon- day-Friday) and once during a weekend (Sat- urday-Sunday) to examine if survey period influenced auditory counts. Observers con- ducted auditory counts once per 20-min inter- val at each survey point throughout the morn- ing during 0540-1 100 hrs CST We chose this time period because peak White-winged dove calling occurs during morning hours (Sepul- veda et al. 2006). Each observer sampled two survey points during a morning. We compared points near (<0.8 km) and far (>0.8 km) from high-traffic roads (i.e., >4 traffic lanes) to further examine the possible effects of traffic noise on calling intensity. We used 0.8 km as a threshold value for traffic noise because, based on our experience, traffic noise affected observers’ hearing ability with- in this distance. Statistical Analysis. — We used a repeated measures ANOVA model to test for differ- ences between week day and weekend density estimates in Austin and San Antonio using SAS (SAS Institute 2001). We used the same procedure to test for differences between week day and weekend density estimates for points near and far from roads. Counts for a partic- ular time interval occasionally were not re- corded because the observer was unable to sample a point during the chosen time interval due to traffic delays. The corresponding ob- servation in the same time interval of the comparison trial was omitted if a call count was not recorded for a particular time interval. We present results as means ± SE and report density estimates as pairs/ha. We use the term “period” to refer to survey period on week days or weekends. RESULTS We did not find an interaction between pe- riod and time of day in Austin = 0.72, P = 0.77) and therefore pooled across time of day (Fig. 1). There was no difference in den- sity estimates between week days (46 ± 10 pairs/ha) and weekends (52 ± 10 pairs/ha; Fj 9 = 1.63, P = 0.23). However, weekend esti- mates were consistently higher throughout the morning (Fig. 1). We did find an interaction for San Antonio between period and time of day (F,5 ,32 = 2.49, P < 0.01). Thus, we com- pared week day and weekend density esti- mates in San Antonio for each 20-min interval between 0540—1 100 hrs. There was no differ- ence in density estimates between week day and weekend estimates before 0620 hrs (P > 0.20) (Fig. 1) after which weekend estimates were higher (Table 1, P < 0.04). We did not find an interaction between pe- riod and time of day for points near roads 15.249 “ 1.18, P = 0.29) or for points far from roads (P,5.,.34 = 0.45, P = 0.96). Week- end estimates (45 ± 5 pairs/ha) were higher than week day estimates (33 ± 5 pairs/ha) for 386 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 (A) I (B) s — • — Week day — » — Weekend FIG 1 White-winged Dove density estimates (mean ± SE) during week day- and weekend-morning audi- tory counts in (A) Austin (n = 10 survey points) and (B) San Antonio, Texas (n = 6 survey points), May-June 2004. points near roads (Fi 9 = 7.46, P — 0.02) (Fig. 2). However, we did not find any difference between week day (52 ± 16 pairs/ha) and weekend (57 ± 16 pairs/ha) density estimates for points far from roads (Fj 5 = 2.76, P = 0.16) (Fig. 2). DISCUSSION We obtained higher White-winged Dove density estimates during all weekend surveys compared to week day surveys. We docu- mented statistically greater White-winged Dove densities on weekends in San Antonio only after 0640 hrs, the time coinciding with the beginning of the morning commute during week days in this city. Differences between week day and weekend estimates were partic- ularly evident when data were analyzed by distance from roads. We observed differences in density estimates between week day and weekend surveys at points that were within 0.8 km of roads but not at points greater than 0.8 km. Thus, our data suggest there is noise- induced bias for auditory surveys conducted in an urban environment. Noise may affect auditory surveys through Breeden et al. • WHITE-WINGED DOVE AUDITORY SURVEYS 387 (A) Time — • — Week day — o — Weekend FIG. 2. White-winged Dove density estimates (mean ± SE) during week day- and weekend-morning audi- tory counts at (A) survey points (n = 9) within 0.8 km of a high-traffic road of four lanes or greater, and (B) survey points (n — 7) >0.8 km of a high traffic road of four lanes or greater, San Antonio and Austin, Texas, May-June 2004. several mechanisms including masking ef- fects, altered behavior, and density changes. Several studies have documented that in- creased noise often causes birds to alter their calling behavior (Slabbekoorn and Smith 2002, Slabbekoorn and Peet 2003, Brumm 2004). For example, Brumm (2004) docu- mented that male Common Nightingales (Lus- cinia megarhynchos) in noisier territories sang at a higher volume than birds at less noisy locations to increase detection of song by po- tential mates. Slabbekoorn and Peet (2003) re- ported that Great Tits {Purus major) also ad- justed to high human noise disturbance by singing at a higher mean frequency. Slabbe- koorn and Smith (2002) reported similar ad- justments by Little Greenbuls {Andropadus vi- rens), where males living in quiet rain forests sang at low-frequency notes that were not used by those inhabiting ecotone forests where high levels of environmental noise masked low-pitched notes. Noise also may result in changes in actual bird density. Reijnen et al. (1995, 1997) re- ported that stress due to traffic noise was the most important factor in reducing bird activity and density along roadways. Forman et al. (2002) documented that bird presence and breeding were reduced up to 1,200 m from a busy road. Reijnen and Foppen (1994) docu- 388 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 2, June 2008 TABLE 1. White-winged Dove density estimates during week day and weekend auditory counts conducted during morning (0540-1100 hrs) in San Antonio, Texas (n = 6 points), May-June 2004. X ± SE3 Time Week day Weekend df r-value P-value 0540-0559 12 ± 7 9 ± 7 54.7 0600-0619 17 ± 6 24 ± 6 34.9 0620-0639 32 ± 6 46 ± 6 34.9 0640-0659 33 ± 6 57 ± 6 34.9 0700-0719 38 ± 6 54 ± 6 34.9 0720-0739 36 ± 6 52 ± 6 34.9 0740-0759 42 ± 6 58 ± 6 34.9 0800-0819 39 ± 6 60 ± 6 34.9 0820-0839 35 ± 6 63 ± 6 34.9 0840-0859 35 ± 6 59 ± 7 43.1 0900-0919 40 ± 7 66 ± 7 54.7 0920-0939 36 ± 7 60 ± 7 54.7 0940-0959 28 ± 7 63 ± 7 54.7 1000-1019 27 ± 8 52 ± 8 71.6 1020-1039 22 ± 9 59 ± 9 94.0 1040-1100 18 ± 7 48 ± 7 54.7 0.45 0.653 -1.23 0.227 -2.20 0.035 -3.69 0.001 -2.46 0.019 -2.39 0.022 -2.46 0.019 -3.30 0.002 -4.40 0.000 -3.60 0.001 -3.49 0.001 -3.16 0.003 -4.64 0.000 -2.93 0.005 -3.75 0.000 -3.98 0.000 3 Least squares means shown for density estimates (pairs/ha). merited that male Willow Warblers {Phyllos- copus trochilus) close to highways experi- enced difficulties in attracting or keeping a fe- male, possibly because traffic noise along roadways could distort their breeding song. Our study and prior research suggests that urban surveys should consider the effect of urban noise on auditory counts to minimize potential bias. Abundance estimators that ac- count for differences in detection such as dis- tance sampling (Buckland et al. 2001) may be an effective alternative to reduce error asso- ciated with decreased detection resulting from traffic noise disturbance (Rosenstock et al. 2002, Anderson 2003, Breeden 2005). ACKNOWLEDGMENTS We thank Texas Parks and Wildlife Department summer interns, Susanne Contreras and Steve Cortez for conducting these surveys. D. G. Hewitt, M. J. Pe- terson, J. A. Roberson, and 2 anonymous reviewers provided helpful comments on an early version of this manuscript. Lunding was provided by the Texas Parks and Wildlife Department’s White-winged Dove Stamp Lund. This paper is Caesar Kleberg Wildlife Research Institute manuscript 05-121. LITERATURE CITED Anderson, D. R. 2001. The need to get the basics right in wildlife field studies. Wildlife Society Bulletin 29:1294-1297. Anderson, D. R. 2003. Response to Engeman: index values rarely constitute reliable information. Wild- life Society Bulletin 31:288-291. Baskett, T. S., M. j. Armbruster, and M. W. Sayre. 1978. Biological perspectives for the Mourning Dove call-count survey. Transactions of the North American Wildlife and Natural Resources Confer- ence 43:163-180. Breeden, J. B. 2005. An assessment of survey meth- ods for White-winged Dove populations in urban environments. Dissertation. Texas A&M Univer- sity-Kingsville, Kingsville, USA. Breeden, J. B., L. Hernandez, N. J. Silvy, R. L. Bing- ham, AND G. L. Waggerman. 2004. An evaluation of sampling methods for White-winged Dove sur- veys in urban areas. Proceedings of the South- eastern Association of Pish and Wildlife Agencies 58:274-281. Brumm, H. 2004. The impact of environmental noise on song amplitude in a territorial bird. Journal of Animal Ecology 73:434-440. Buckland, S. T, D. R. Anderson, K. P. Burnham, J. L. Laake, D. L. Borchers, and L. Thomas. 2001. Introduction to distance sampling: estimating abundance of biological populations. Oxford Uni- versity Press, New York, USA. Cohen, A., H. S. Peters, and L. E. Poote. 1960. Call- ing behavior of Mourning Doves in two midwest life zones. Journal of Wildlife Management 24: 203-212. Dawson, D. G. 1981. Counting birds for a relative measure (index) of density. Studies in Avian Bi- ology 6:12—16. Emlen, j. T. and M. j. DeJong. 1981. The application of song detection threshold distance to census op- erations. Studies in Avian Biology 6:346-352. Breeden et al. • WHITE-WINGED DOVE AUDITORY SURVEYS 389 Forman, R. T T, B. Reineking, and A. M. Hersper- GER. 2002. Road traffic and nearby grassland bird patterns in a suburbanizing landscape. Environ- mental Management 29:782-800. George, R. R., R. E. Tomlinson, R. W. Engel-Wil- soN, G. L. Waggerman, and a. G. Spratt. 1994. White-winged Dove. Pages 29-52 in Migratory shore and upland game bird management in North America (T. C. Tacha and C. E. Braun, Editors). International Association of Fish and Wildlife Agencies, Washington D.C., USA. Keppie, D. M., H. M. Wight, and W. S. Overton. 1970. A proposed Band-tailed Pigeon census — a management need. Transactions of the North American Wildlife and Natural Resources Confer- ences 35:157-171. LaPerriere, a. J. and A. O. Haugen. 1972. Some factors influencing calling activity of wild Mourn- ing Doves. Journal of Wildlife Management 36: 1193-1198. Rappole, J. H. and G. L. Waggerman. 1986. Calling males as an index of density for breeding White- winged Doves. Wildlife Society Bulletin 14:151- 155. Reijnen, R. and R. Foppen. 1994. The effects of car traffic on breeding bird populations in woodland. I. Evidence of reduced habitat quality for Willow Warblers (Phylloscopus trochilus) breeding close to a highway. Journal of Applied Ecology 31:85- 94. Reijnen, R., R. Foppen, and G. Veenbass. 1997. Dis- turbance by traffic of breeding birds: evaluation of the effect and considerations in planning and man- aging road corridors. Biodiversity and Conserva- tion 6:567-581. Reijnen, R., R. Foppen, C. T. Braak, and J. Thissen. 1995. The effects of car traffic on breeding bird populations in woodland. III. Reduction of density in relation to the proximity of main roads. Journal of Applied Ecology 32:187-202. Rosenstock, S. S., D. R. Anderson, K. M. Giesen, T. Leukering, and M. F. Carter. 2002. Landbird counting techniques: current practices and an al- ternative. Auk 119:46-53. SAS Institute. 2001. SAS software release 8.2. SAS Institute Inc., Cary, North Carolina, USA. Sepulveda, M., E Hernandez, D. G. Hewitt, W. P. Kuvlesky Jr., G. Waggerman, and R. L. Bing- ham. 2006. An evaluation of auditory counts for estimating breeding populations of White-winged Doves. Journal of Wildlife Management 70:1393- 1402. Shields, W. H. 1977. The effect of time of day on avian census results. Auk 94:380-383. Sisson, L. H. 1968. Calling behavior of Band-tailed Pigeons in reference to a census technique. Thesis. Oregon State University, Corvallis, USA. Slabbekoorn, H. and M. Peet. 2003. Birds sing at a higher pitch in urban noise. Nature 424:267. Slabbekoorn, H. and T. B. Smith. 2002. Habitat-de- pendent song divergence in the Little Greenbul: an analysis of environmental selection pressures on acoustic signals. Evolution 56:1849-1858. Thompson, W. L. 2002. Towards reliable bird surveys: accounting for individuals present but not detect- ed. Auk 119:18-25. UzzELL, P. B. AND W. H. KiEL Jr. 1950. Unpublished progress report. Texas Parks and Wildlife Depart- ment, Austin, USA. ViERS Jr., C. 1970. The relationship of calling behav- ior of White-winged Doves to population and pro- duction in southern Arizona. Dissertation. Univer- sity of Arizona, Tucson, USA. Waggerman, G. L. 2001. White- winged Dove density, distribution, movement, and harvest. Performance Report. Federal Aid Project W-128-R, Job 9. Tex- as Parks and Wildlife Department, Austin, USA. West, L. M. 1993. Ecology of breeding White-winged Doves in the San Antonio metropolitan area. The- sis. Texas Tech University, Lubbock, USA. Short Communications The Wilson Journal of Ornithology 1 20(2):390-392, 2008 Specimen Shrinkage in Cinnamon Teal Robert E. Wilson^’^ and Kevin G. McCracken^ ABSTRACT. — Body size measurements from fresh- ly collected birds and dried museum specimens were used to evaluate specimen shrinkage in Cinnamon Teal {Anas cyanoptera). Six of seven body measurements of female Cinnamon Teal differed significantly after specimen drying, whereas five of seven male body measurements differed. The largest amount of shrink- age was in bill height, bill width, and tarsus length. Bill length at nares showed no significant shrinkage suggesting it is a more conservative measurement than exposed culmen and, therefore, a more reliable method for accurately measuring bill length. Correction values for body size measurements are reported for future wa- terfowl studies combining measurements of both live birds and museum specimens. Received 1 March 2007. Accepted 6 September 2007. Specimen shrinkage during the process of drying is common. Shrinkage can cause ana- lytical problems if not properly corrected in studies involving live or freshly killed birds and museum specimens (e.g.. Winker 1996). Correction for shrinkage is needed before ap- plying to live birds when developing classifi- cation criteria for gender, subspecies, or spe- cies based on morphological features from museum specimens (Greenwood 1979, Jenni and Winkler 1989, Winker 1993). For exam- ple, Mueller (1990) reported that a shrinkage value of 1.72% would comprise 34% of wing length differences between male and female Northern Saw-whet Owls {Aegolius acadicus). In addition, the amount of shrinkage varies among body parts and species (Winker 1993). Shrinkage values from one taxon may have limited use outside that particular taxon or similar morphological species because of the morphological diversity of birds and variety of preparation techniques (Jenni and Winkler 1989, Winker 1993). Specimen shrinkage in waterfowl has yet to be investigated. This pa- * Institute of Arctic Biology, Department of Biology and Wildlife, and University of Alaska Museum, Uni- versity of Alaska, Fairbanks, AK 99775, USA. -Corresponding author; e-mail: ftrewl@uaf.edu per reports shrinkage values for Cinnamon Teal (Anas cyanoptera), which may be used to develop correction values for similar size (—350-550 g) waterfowl. METHODS Cinnamon Teal are widespread throughout the Western Hemisphere and five subspecies currently are recognized (Snyder and Lums- den 1951, Delacour 1956, AOU 1957, Johns- gard 1978, Gammonley 1996). The three most widespread subspecies of Cinnamon Teal (A. c. cyanoptera, A. c. orinomus, and A. c. sep- tentrionaliunr, 26 females, 80 males) were collected in Argentina (2003), Peru (2002), and western United States (2002-2003) as part of a larger population genetic and mor- phological study. Even though subspecies are distinct in overall body size, there is overlap in measurements among subspecies (R. E. Wilson, unpubl. data). Therefore, different subspecies were pooled for each gender to as- certain the extent of shrinkage for each mea- surement. Seven body measurements were recorded for each bird (±0.1 mm unless otherwise in- dicated; Baldwin et al. 1931); wing chord length (carpal joint to longest primary feather unflattened; ± 1 mm), tail length (± 1 mm), exposed culmen length (edge of forehead feathers to posterior edge of nail), bill length at nares (posterior edge of nares to posterior edge of nail), total tarsus length (top of bent knee to bottom of foot), bill height (height of upper mandible at nares), and bill width (width of upper mandible at nares). Measure- ments were taken the same day specimens were collected prior to preparation as museum specimens (fresh measurements), and subse- quently 9 months to 2 years after preparation (dry measurements from standard museum round skins) by the same individual (R. E. Wilson) with the same set of calipers. The right wing and tarsus were used for fresh and dry measurements of each specimen. Voucher 390 SHORT COMMUNICATIONS 391 TABLE 1 . Effects of shrinkage on body measurements of Cinnamon Teal with correction values from dried specimens to live birds. Mean length (mm) Shrinkage Correction Gender Fresh SE Dry SE p % SE factor Males Wing chord 80 194.51 1.57 191.00 1.49 8.44 <0.001 2.12 0.25 1.018 Total tarsus 80 41.92 0.26 40.46 0.23 11.21 <0.001 3.43 0.30 1.036 Tail 80 83.04 0.74 93.66 0.95 -0.91 0.366 -0.83 0.83 0.887 Bill nare 80 35.07 0.20 35.08 0.19 -0.04 0.968 -0.03 0.18 1.000 Bill culmen 80 45.63 0.23 45.10 0.26 4.29 <0.001 1.16 0.27 1.012 Bill height 80 13.72 0.09 12.97 0.10 8.01 <0.001 5.36 0.66 1.058 Bill width 80 16.89 0.08 15.95 0.11 11.98 <0.001 6.44 0.54 1.059 Females Wing chord 26 189.69 2.88 185.46 2.59 5.25 <0.001 2.17 0.39 1.023 Total tarsus 26 41.56 0.42 39.89 0.42 4.64 <0.001 3.94 0.84 1.042 Tail 26 81.54 1.58 84.49 1.89 -2.35 0.027 -3.73 1.53 0.965 Bill nare 26 32.79 0.28 32.7 0.30 0.71 0.483 0.28 0.38 1.003 Bill culmen 26 43.04 0.38 42.55 0.38 2.54 0.018 1.11 0.44 1.012 Bill height 26 13.06 0.17 12.6 0.20 2.48 0.020 3.45 1.42 1.037 Bill width 26 16.34 0.19 15.3 0.20 5.29 <0.001 6.21 1.15 1.068 ^ r-value from paired sample t-test. specimens are archived at the University of Alaska Museum (Fairbanks). A paired utest was used to compare differences between fresh and dry measurements. Pearson corre- lation values were used to examine the rela- tionship between body mass and percent shrinkage. RESULTS Five of the seven measurements for males had significant differences after drying (Table 1). There was no significant change in tail length or bill length at nares, but both mea- surements showed an increase after drying. All other measurements had >1% decrease with bill height and bill width having the larg- est shrinkage. Percent shrinkage of total tarsus length (Pearson correlation = 0.255, P = 0.023) and culmen length (Pearson correlation = —0.356, P = 0.001) had a significant rela- tionship with body mass. Six of the seven body measurements for fe- males had significant differences after drying (Table 1). Bill length at nares had no signifi- cant difference. All measurements except tail length decreased after drying with bill width and total tarsus having the greatest amount of shrinkage. There were no significant relation- ships between any of the shrinkage measure- ments and body mass. DISCUSSION Cinnamon Teal had significant changes af- ter specimen preparation for most measure- ments. Specimen preparation may have con- tributed to differences between measurements besides the drying process. The bills of spec- imens were tied to keep them closed during the drying process in the field. Tying of bills may have squeezed the bill together, slightly decreasing bill width. Tail length increased af- ter drying for House Sparrows {Passer do- mesticus) and was attributed to the retraction of the intercalamal skin (Bjordal 1983). Bill length is an important descriptor for studying feeding ecology (Borras et al. 2000) and subspecies classification (e.g., Ridgway 1902, Hall 1996). Therefore, it is critical to have a bill measurement that is repeatable and accurate. There are several ways to measure bill length with the three main alternatives be- ing total culmen length, exposed culmen length, and length from the nares (Baldwin et al. 1931). Fjeldsa (1980) suggested the amount of shrinkage of the exposed culmen will vary according to bill anatomy and, thus, one universal correction factor would not be applicable to all bird species. This has led to the suggestion that bill length from the pos- terior edge of the nares is the most reliable 392 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 bill measurement as both end points are easily defined (Winker 1998, Borras et al. 2000). This study confirms the recommendation, in particular for waterfowl, that bill length should be measured from the nares, especially if no correction factors are available. The range of shrinkage values of -3.73 to 6.44% for Cinnamon Teal is comparable to other studies, which report values ranging from -1.5 to 4.0% depending on the body measurement. Correction values to convert dry measurements ranged from 1 .000 to 1 .068 for measurements that decreased and 0.887 (males, not significant) and 0.965 (females) for tail lengths which increased (Table 1). Winker (1993) suggested correction values that ranged from 0.960 to 0.996 (fresh to dry) which converts to 1.004-1.040 (dry to fresh). No previous data describing specimen shrink- age have been reported for waterfowl to our knowledge; the values reported here provide general correction factors for future studies of morphology in similar sized waterfowl. ACKNOWLEDGMENTS We thank Rosario Acero, Raul Clarke, Adrian Con- treras, Alejandro Gonzalez, Daniel Ramadori, Karina Ramirez, Sergio Goldfeder, Direccion de Fauna Santa Cruz, Ministerio de la Produccion Chubut, Direccion de Fauna Silvestre - Secretaria de Ambiente y Desar- rollo Sustentable de la Republica Argentina, Instituto Nacional de Recursos Naturales del Peru, U.S. Fish and Wildlife Service, California Department of Fish and Game, Edwards Air Force Base, Sonny Bono Sal- ton Sea National Wildlife Refuge (NWR), Oregon De- partment of Fish and Wildlife, Malheur NWR, Utah Division of Wildlife Resources, Colorado Division of Wildlife, Browns Park NWR, North Dakota Game and Fish Department, and South Dakota Department of Game, Fish, and Parks for authorization to collect wa- terfowl specimens. We thank Delta Waterfowl Foun- dation, T. H. Valqui, Larry Janke, and Ambassador Duck Club for aid in specimen collection. Expedition and laboratory costs were funded by the Institute of Arctic Biology at the University of Alaska Fairbanks, National Science Foundation (NSF EPS-0092040, EPS-0346770, and DEB-0444748), grants from the Delta Waterfowl Foundation, Frank M. Chapman Fund at the American Museum of Natural History, and Da- vid Burnett Memorial Award to REW. Carla Dove and Sarah Sonsthagen provided helpful comments on this manuscript. LITERATURE CITED American Ornithologists’ Union (AOU). 1957. Check-list of North American birds. Fifth Edition. American Ornithologists’ Union. Baltimore, Maryland, USA. Baldwin, S. P, H. C. Oberholser, and L. G. Worley. 1931. Measurements of birds. Scientific Publica- tions of the Cleveland Museum of Natural History 2:1-165. Bjordal, H. 1983. Effects of deep freezing, freeze-drying and skinning on body dimensions of House Spar- rows {Passer domesticus). Cinclus 6:105—108. Borras, A., J. Pascual, and J. C. Senar. 2000. What do different bill measures measure and what is the best method to use in granivorous birds? Journal of Field Ornithology 74:606-611. Delacour, j. 1956. The waterfowl of the world. Vol- ume 2. Country Life Limited, London, United Kingdom. Fjeldsa, j. 1980. Post-mortem changes in measure- ments of grebes. Bulletin of the British Ornithol- ogists’ Club 100:151-154. Gammonley, j. H. 1996. Cinnamon Teal {Anas cyanop- tera). The birds of North America. Number 217. Greenwood, J. G. 1979. Post-mortem shrinkage of Dunlin Calidris alpina skins. Bulletin of the Brit- ish Ornithologists’ Club 99:143-145. Hall, G. A. 1996. Yellow-throated Warbler {Dendroi- ca domincia). The birds of North America. Num- ber 223. Jenni, L. and R. Winkler. 1989. The feather-length of small passerines: a measurement for wing-length in live birds and museum skins. Bird Study 36:1-15. JoHNSGARD, P. A. 1978. Ducks, geese, and swans of the world. University of Nebraska Press, Lincoln, USA. Mueller, H. C. 1990. Can Saw-whet Owls be sexed by external measurements? Journal of Field Or- nithology 61:339-346. Ridgway, R. 1902. The birds of North and Middle Amer- ica. U.S. National Museum Bulletin 50, Part 2. Snyder, L. L. and H. G. Lumsden. 1951. Variation in Anas cyanoptera. Occasional Papers of the Royal Ontario Museum of Zoology 10:1-18. Winker, K. 1993. Specimen shrinkage in Tennessee Warblers and “Traill’s” Flycatchers. Journal of Field Ornithology 64:331-336. Winker, K. 1996. Specimen shrinkage versus evolu- tion: I’iwi morphology. Conservation Biology 10: 657-658. Winker, K. 1998. Suggestions for measuring external characters of birds. Ornithologia Neotropical 9: 23-30. SHORT COMMUNICATIONS 393 The Wilson Journal of Ornithology 120(2):393— 395, 2008 Breeding Range Extension of the Coastal Plain Swamp Sparrow Bryan D. Watts, Michael D. Wilson,^ Fletcher M. Smith, ^ Barton J. Paxton,^ and J. Bill Williams^ ABSTRACT — The Coastal Plain Swamp Sparrow (Melospiza georgiana nigrescens) is morphologically distinct, restricted to a narrowly-defined habitat type, and geographically isolated within the mid-Atlantic Coastal Plain. The breeding range has been considered to extend from the Nanticoke River in Maryland north to the Hudson River. We report a previously undocu- mented population near Warsaw, Virginia that extends the known range south and west into a region of the Chesapeake Bay with extensive tidal fresh and brack- ish marshes consistent with the habitat requirements of this form, but for which there has been no documented breeding. A broader investigation of occurrence within appropriate habitat seems warranted given the small global population size and uncertain status within the southern portion of its range. Received 9 November 2006. Accepted 2 August 2007. The Coastal Plain Swamp Sparrow {Melo- spiza georgiana nigrescens) is restricted to tidal fresh and brackish marshes of the mid- Atlantic coast. The form is distinctive in hav- ing a larger bill, grayer plumage, and more black in the crown and nape compared to oth- er Swamp Sparrows (Bond and Stewart 1951, Greenberg and Droege 1990). Breeding pop- ulations appear to use a limited range of hab- itats that contain a mix of marsh vegetation and shrubs with a characteristic structure that typically forms along the marsh-upland inter- face (Meanley 1975, Beadell et al. 2003) The known breeding distribution of the Coastal Plain Swamp Sparrow has been ex- tended over time as expanded survey efforts have discovered new breeding locations. The subspecies was initially described from sev- eral specimens from the Nanticoke River near Vienna, Maryland (Bond and Stewart 1951). Investigations have extended the range west to near Washington, D.C. and north to the mouth of the Hudson River (Stewart and Rob- ' Center for Conservation Biology, College of Wil- liam and Mary, Williamsburg, VA 23187, USA. ^Corresponding author; e-mail; bdwatt@wm.edu bins 1958, Greenberg and Droege 1990, Droe- ge and Blom 1996). The breeding range since 1957 has been described to extend from the Nanticoke River in Maryland north to the Hudson River (AOU 1957, Mowbray 1997). Beadell et al. (2003), in a recent evaluation of the breeding population, established a center of abundance around Delaware Bay and the Tuckahoe and Mullica rivers in coastal New Jersey. Greenberg and Droege (1990), in an earlier review of the breeding range, reported no evidence of breeding south of Maryland. Clapp (1997) reported no confirmed breeding records for coastal Virginia. Here, we report on a previously undocumented population of M. g. nigrescens near Warsaw, Virginia. METHODS Mulberry Point Marsh (37° 59' 30" N, 76° 53'35"E) is 178 ha along the oligohaline reach of the Rappahannock River near War- saw, Virginia. Vegetation within the marsh is dominated by big cordgrass (Spartina cyno- suroides), salt meadow hay (S. patens), marsh hibiscus {Hibiscus moscheutos), cattail {Typha augustifolia), and olney threesquare {Scirpus olneyi) with a diverse mixture of other marsh species. The marsh contains numerous tree and shrub hummocks. Historically, parts of the marsh have been used for agriculture and cattle grazing, and the marsh is disturbed by a roadway, a dike system, and limited ditch- ing. These alterations along with the natural hummocks create topographic variation, hab- itat diversity, and an extensive network of marsh-upland ecotones. Normal tidal variation along this reach of the river is ~().5 m. We received a report of 14 birds within the marsh during June 2004 (F. T. Atwood, pers. comm.) and conducted a spot-mapping effort between 24 May and 13 June 2005 to: (1) es- timate population size, and (2) document breeding. We systematically walked through- out the entire marsh surface twice and mapped 394 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 locations of all singing males on aerial pho- tographs. The overall number of singing males and their approximate locations were consis- tent between surveys. We assumed the num- ber of singing males provided a good estimate of the breeding population but made no at- tempt to identify their pairing status. We spent ~2 hrs on 24 May observing birds to locate nests for a limited number of pairs. We did not follow nests through time to ascertain breeding success. OBSERVATIONS We mapped 41 singing male Coastal Plain Swamp Sparrows that appeared to be defend- ing discrete territories. Birds were not evenly distributed throughout the marsh but were concentrated within patches containing threes- quare and scattered groundsel {Baccharis hal- imifolia) trees. All patches of any size (>0.25 ha) of this habitat supported birds. Patches comprised entirely of taller vegetation (e.g., big cordgrass) were not occupied. Singing males perched on shrubs within the meadow vegetation or on trees or shrubs along the edge of meadows. We located five active nests dur- ing the morning of 24 May. These included two nests under construction and one nest each with 2, 3, and 4 eggs, respectively. All nests were at or near the base of groundsel trees and hidden in clusters of newly emerging leaves. DISCUSSION Documentation of the Mulberry Point Marsh population of Coastal Plain Swamp Sparrows extends the known breeding range south into Virginia. The site is 90 km west and 30 km south of the Nanticoke River in Maryland. More importantly, this site is with- in the lower western shore of Chesapeake Bay, a region with extensive tidal fresh and brackish marshes consistent with the habitat requirements of this form, but for which there has been no documented breeding. The size of the population places it among the largest cur- rently known in the Chesapeake Bay portion of the range. A cursory survey of Island Farm Marsh, 8 km down river from Mulberry Point on 14 July 2005 resulted in the location of 5 singing males. Documentation of birds at a second site suggests the possibility that breed- ing may be more extensive than currently known. The Coastal Plain Swamp Sparrow appears to be morphologically distinct, geographically isolated, and specialized in a narrowly-defined habitat type. These are characteristics of other sparrow forms such as the Dusky Seaside Sparrow {Ammodramus maritimus nigrescens) (Sykes 1980), Cape Sable Seaside Sparrow {A. maritimus mirabilis) (Pimm et al. 1996), and Ipswich Savannah Sparrow {Passerculus sandwichensis princeps) (Smith et al. 2003) that were extirpated or have been of high con- servation concern in recent decades. A recent assessment of the population of Coastal Plain Swamp Sparrow within Maryland, Delaware, and New Jersey resulted in a conservative es- timate of 28,000 pairs and suggested a decline in both abundance and distribution along the western shore and lower eastern shore of the Chesapeake Bay (Beaded et al. 2003). Given the uncertain status of this form in the south- ern portion of its range, a broad investigation of occurrence within appropriate habitat seems warranted. Consideration of listing of the Coastal Plain Swamp Sparrow by appro- priate agencies and jurisdictions as a species of conservation concern may facilitate con- servation efforts. ACKNOWLEDGMENTS We thank E T. Atwood for early reports of this pop- ulation, A. M. France for allowing access to the site, M. U. Watts for assisting with surveys on Mulberry Point Marsh, and S. C. Spencer for assistance with surveys of Island Farm Marsh. R. B. Clapp and Sam Droege made helpful comments on an earlier draft of this manuscript. This study was funded by the Center for Conservation Biology at the College of William and Mary. LITERATURE CITED American Ornithologists’ Union (AOU). 1957. Check-list of North American birds. Fifth Edition. American Ornithologists’ Union, Washington, D.C., USA. Beadell, J., R. Greenberg, S. Droege, and J. A. Roy- LE. 2003. Distribution, abundance, and habitat af- finities of the Coastal Plain Swamp Sparrow. Wil- son Bulletin 115:38-44. BON’D, G. M. AND R. E. Stewart. 1951. A new Swamp Sparrow from the Maryland Coastal Plain. Wilson Bulletin 63:38-40. Clapp, R. B. 1997. Egg dates for Virginia birds. Vir- ginia Avifauna 6. Virginia Society of Ornithology, Lynchburg, USA. SHORT COMMUNICATIONS 395 Droege, S. and E. a. T. Blom. 1996. Swamp Sparrow (Melospiza georgiana). Pages 408-409 in Atlas of the breeding birds of Maryland and the District of Columbia (C. S. Robbins, Senior Editor). Univer- sity of Pittsburgh Press, Pittsburgh, Pennsylvania, USA. Greenberg, R. and S. Droege. 1990. Adaptations to tidal marshes in breeding populations of the Swamp Sparrow. Condor 92:393-404. Meanley, B. 1975. Birds and marshes of the Chesa- peake Bay country. Tidewater Publishers, Cam- bridge, Maryland, USA. Mowbray, T. B. 1997. Swamp Sparrow {Melospiza georgiana). The birds of North America. Number 279. Pimm, S., J. Curnutt, J. Lockwood, L. Manne, A. Mayner, M. Nott, and K. Balent. 1996. Popu- lation ecology of the Cape Sable Sparrow {Arn- modramus maritimus mirabilis): annual report, 1996. USDI, National Biological Survey/National Park Service, Everglades National Park, Home- stead, Florida, USA. Smith, J. S., Z. Lucas, and W. T. Stobo. 2003. Esti- mate of the Ipswich Sparrow population on Sable Island, Nova Scotia, in 1998, using random-tran- sect survey design. Canadian Journal of Zoology 81:771-779. Stewart, R. E. and C. S. Robbins. 1958. Birds of Maryland and the District of Columbia. North American Fauna 62. USDI, Fish and Wildlife Ser- vice, Washington, D.C., USA. Sykes, P. W. 1980. Decline and disappearance of the Dusky Seaside Sparrow from Merritt Island, Flor- ida. American Birds 34:728-737. The Wilson Journal of Ornithology 120(2):395-398, 2008 Polyandry and Sex Ratio in the Song Sparrow Michael H. Janssen,^ Peter Arcese,^’^ Mark S. Sloan,' and Kelly J. Jewell' ABSTRACT. — Polyandry occurs when females form social bonds and gain simultaneous parental care from multiple male mates. It is thought to be rare in birds and to occur more often in territorial species when the Operational Sex Ratio (OSR; ratio of mature males to females) exceeds one. We asked if variation in the OSR affected the rate of polyandry in Song Sparrows {Melospiza melodia) over 30 years on Man- darte Island, British Columbia, Canada. We found no correlation between OSR and polyandry {R^ = 0.04, df = 28, P = 0.86), but positive correlations between OSR and percent females with more than one social mate {R^ = 0.44, df = 28, P = 0.01), and percent females sharing a territory with a replacement male and her dependent young {R^ - 0.46, df = 28, P = 0.009). We suggest that polyandry in Song Sparrows is limited by the intolerance of territorial males to- wards intruders, but that it occurs when females oc- cupy the territories of two or more males and gain their simultaneous care for the dependent young of a single brood as a consequence. Received 2 October 2006. Ac- cepted 25 July 2007. Cooperative polyandry, wherein two or more males mate with one female and con- tribute care to a single brood, is rare in birds ' Centre for Applied Con.servation Research, Uni- versity of British Columbia, Forest Sciences Centre, 2424 Main Mall, Vancouver, BC V6T 1Z4, Canada. ^ Corresponding author; e-mail: peter.arcese@ubc.ca (Jenni 1974, Oring 1986). However, in species that compete to defend breeding territories, the frequency of polygamy is often related to the ‘Operational Sex Ratio’ (OSR, ratio of mature males to females in a population; Em- len and Oring 1977, Davies and Lundberg 1984, Lank et al. 1985, Reynolds 1987, Wik- tander et al. 2000, Pilastro et al. 2001). Thus, it is possible that polyandry is rare in some species either because the OSR is rarely skewed sufficiently to favor it or because ter- ritory defense by males prevents females from receiving the care of two males at one nest simultaneously (Arcese 1989a). Song Sparrows {Melospizxi melodia), often considered to be monogamous (e.g., Verner and Willson 1966), are now known to engage regularly in genetic polyandry (O’Connor et al. 2006) and social polygyny, the latter being more common when OSR favors females (Smith et al. 1982, Arcese 1989a). We de- scribe the first cases of cooperative polyandry in Song Sparrows and ask if variation in the OSR predicted their occurrence over 30 years on Mandarte Island, British Columbia. Cana- da. We expected that OwSRs favoring males would increase the rate of polyandry, unless territoriality prevented multiple males from feeding one brood simultaneously. 396 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 METHODS Study Area. — Mandarte is a ~6-ha islet about 10 km southeast of Sidney, British Co- lumbia where Song Sparrows have been stud- ied continuously since 1975 (Smith et al. 2006). Song Sparrows on Mandarte Island nest mainly in shrub patches that cover about a third of the islet, are individually marked, and defend territories year-round with peaks in aggression in fall and late winter through spring. Territories were mapped annually and the breeding status of all adults was assessed each 3-7 days from March or April to the end of breeding from 1975 to 2005 (except 1980; see Smith et al. 2006). Mated and Territorial Status of Males. — We recorded instances from 1975 to 2005 where more than one male was observed feed- ing young from one brood simultaneously or sequentially in the course of a female’s nest- ing cycle. We choose feeding by males as the main criteria for polyandry because ‘pair bonds’ are difficult to quantify and genetic ev- idence was available in only 4 years. Al- though extra-pair fertilization and mate switching are common in Song Sparrows (Ar- cese et al. 2002, O’Connor et al. 2006), nei- ther constitute polyandry (Oring 1986). In contrast, cases in which females gain parental care from and maintain social contact with two males simultaneously are generally con- sidered to be polyandrous (Jenni 1974, Oring 1986), and were defined by us as engaging in ‘cooperative polyandry’. In contrast, we de- fined cases wherein males fed young from sin- gle broods sequentially, such as after a terri- tory take-over (Arcese 1989b), as ‘sequential polyandry’. We also recorded instances wherein a male neighbor or replacement with- out young of their own carried food to young on a neighboring or newly acquired territory or were followed persistently by begging young, as suspected cases of sequential or co- operative polyandry. OSR equaled the ratio of sexually mature adult males to females alive in late April each year. Statistical Analyses. — We used SYSTAT 11.0 for all analyses (Systat Software Inc., Point Richmond, CA, USA) and Spearman’s rank correlation to test for links between OSR and the frequency of polyandry, the percent of females with more than one social mate in a year, the percent of females whose young came into contact with a replacement male, and rate of ‘adoption’ (feeding) by replace- ment males. Frequency data were analyzed by Chi-square with the Yates correction applied. RESULTS On average (± SD), 16 ± 9% of females (range = 0-37%) occupied territories with two or more social mates annually from 1975 to 2005. Females gained new social mates by occupying two or more territories within a breeding season (55 cases) or forming social bonds with replacement males after a territory take-over (111 cases); the latter occurring more often (xVtes “ df = 1, P < 0.001). Eighteen of 72 males (25%) were known or suspected to have fed fledglings that remained on their natal territory after the replacement male’s arrival. Cooperative polyandry was confirmed once and suspected twice, but sequential polyandry was more common with seven confirmed and eight suspected cases. Overall, 3.7% of 492 females and 6.5% of 505 males that bred dur- ing 1975 to 2005, bred at least once in a poly- androus group. Three males were recorded in polyandrous associations twice in their life- times. One case of cooperative polyandry was ob- served in detail in May 2005. On 24 April, two young hatched in a nest on the territory of male Ml and female El. On 4 May, Ml lost part of this territory to neighbor M2. Rather than remaining in Ml’s territory, how- ever, El occupied both territories and inter- acted with both Ml and M2, which each fed one of El’s fledglings to independence on about 23 May. El laid the first egg of her sec- ond nest inside M2’s territory on 12 May. Two suspected cases of cooperative polyandry also occurred when females expanded their range to occupy the territories of unpaired males with two fledglings tended by her new social mate and one by the previous social mate. Sequential polyandry occurred in two ways. In three cases, females moved to a neighbor’s territory with all her dependent fledglings. In 12 cases, replacement males occupied the ter- ritories of females with dependent young. In all 15 cases, both males fed young from a sin- gle female’s brood sequentially. Contrary to expectation, OSR and annual SHORT COMMUNICATIONS 397 percent of females in suspected or confirmed cases of polyandry were unrelated = 0.04, df = 28, P = 0.86). OSR was related posi- tively to the percent of females that had more than one social mate in a year (R^ = 0.44, df = 28, P = 0.01), and to the percent of females that mated with replacement males while still caring for the fledglings of a prior mate (R^ = 0.46, df = 28, P — 0.009). The proportion of replacement males that fed dependent young alive at the time they occupied their new ter- ritory was also unrelated to the OSR (R^ = 0.04, df = 28, P = 0.83). Fifteen of 18 (83%) males known or sus- pected to have fed their new mate’s young were previously the female’s neighbor, but three polyandrous males were non-territorial floaters. Thus, many polyandrous males prob- ably interacted socially with females whose young they later fed prior to the time they began sharing a territory. However, in 43 cas- es where polyandry could have been detected, neighbors were not more likely to feed the young of a previous male than were floaters (43% of 35 vs. 38% of 8 cases, respectively). DISCUSSION Polyandry was rare among Song Sparrows on Mandarte Island but did occur when fe- males with dependent fledglings occupied the territories of adjacent males and received si- multaneous parental care from those males as a result. Over 30 years, 3.7% of females and 6.5% of males bred in cooperative or sequen- tially polyandrous groups at least once, com- pared to 21% of females and 13% of males exposed to polygyny (Arcese 1989a). We found no relation between the OSR and frequency of polyandry, perhaps because it occurred so rarely. However, percent females paired to more than one male in a season, and percent fledglings occupying a territory with a replacement male, each increased with the OSR. These results suggest the availability of unmated males and level of male competition affect female breeding dispersal and male re- placement, but they are contrary to the idea that OSR affects polyandry. Emlen and Oring (1977) predicted that polyandry will occur more often when adult sex ratios favor males; an idea well-supported in Prunella, where polyandry is relatively common at OSR’s as extreme as 1 .48 {P. inod- ularis; Davies and Lundberg 1984) and 1.38 (P. collaris\ Davies et al. 1995). In contrast, polyandry was rare in Song Sparrows on Mandarte Island even though the OSR aver- aged 1.67 ± 0.66 (SD, n = 30 years) and ex- ceeded 2.50 in 3 years. We suggest this dif- ference between species arises as a conse- quence of intolerance by territorial male Song Sparrows toward intruders. Whereas males of many polyandrous spe- cies feed young from a single brood within the bounds of one territory (Davies and Lund- berg 1984, Hartley and Davies 1994, Briskie et al. 1998, Goetz et al. 2003), intolerance by male Song Sparrows may prevent genetically polyandrous males from simultaneously feed- ing their young except when they are divided among adjacent territories. O’Connor et al. (2006) noted that —27% of nestling Song Sparrows on Mandarte Island were fathered by males other than the social mate and that most extra-pair males were neighbors. Most (86%) polyandrous males we identified were also neighbors of the female whose young they later fed. Thus, it is possible that male Song Sparrows fed young from adjacent ter- ritories according to their confidence of pater- nity (e.g., Davies and Hatchwell 1992). If true, extra-pair males might also be expected to en- gage more often in vigilance and alarm calling while their extra-pair mates incubate or tend nestlings in an adjacent territory. ACKNOWLEDGMENTS Our work was funded by the National Science and Engineering Research Council of Canada, U.S. Na- tional Science Foundation, and the generous support of Werner and Hildegard Hesse. Many people contrib- uted to data collection, including most recently Laura Sampson, and Scott and Amy Wilson. Patricia Janssen and two anonymous reviewers provided helpful com- ments, and the Tswaout and Tseycum bands kindly fa- cilitated our work on Mandarte Island. LITERATURE CITED Arcese, P. 1989a. Intra.sexual competition and the mat- ing system in primarily monogamous birds: the case of the Song Sparrow. Animal Behaviour 38: 96-111. Arcese, P. 1989b. Territory acquisition and loss in the Song Sparrow. Animal Behaviour 37:45-3.‘S. Arcese, P, M. K. Soc.ge, A. B. Marr, and M. A. Patfen. 2002. Song Sparrow {Melospiza melo- cUa). The birds of North America. Number 704. Briskie, .1. V., R. Montgomerie, T. Poldmaa, and P. 398 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 T Boag. 1998. Paternity and paternal care in the polygynandrous Smith’s Longspun Behavioral Ecology and Sociobiology 43:181 — 190. Davies, N. B. and B. J. Hatchwell. 1992. The value of male parental care and its influence on repro- ductive allocation by male and female Dunnocks. journal of Animal Ecology 61:259-272. Davies, N. B. and A. Lundberg. 1984. Food distri- bution and a variable mating system in the Dun- nock, Prunella modularis. Journal of Animal Ecology 53:895-912. Davies, N. B, I. R. Hartley, B. J. Hatchwell, A. Des- ROCHERS, J. Skeer, AND D. Nebel. 1995. The poly- gynandrous mating system of the Alpine Accentor, Prunella collaris. I. Ecological causes and reproduc- tive conflicts. Animal Behaviour 49:769-788. Emlen, S. T. and L. W. Oring. 1977. Ecology, sexual selection and the evolution of mating systems. Science 197:215-223. Goetz, J. E., K. P. McFarland, and C. C. Rimmer. 2003. Multiple paternity and multiple male feed- ers in Bicknell’s Thrush {Catharus hicknelli). Auk 120:1044-1053. Hartley, I. R. and N. B. Davies. 1994. Limits to polyandry in birds. Proceedings of the Royal So- ciety of London, Series B 257:67-73. Jenni, D. a. 1974. Evolution of polyandry in birds. American Zoologist 14:129-144. Lank, D. B., L. W. Oring, and S. J. Maxson. 1985. Mate and nutrient limitation of egg-laying in a polyandrous shorebird. Ecology 66:1513-1524. O’Connor, K. D., A. B. Marr, P. Arcese, L. F. Kel- ler, K. J. Jeffery, and M. W. Bruford. 2006. Extra-pair fertilization and effective population size in the Song Sparrow {Melospiza melodia). Journal of Avian Biology 37:572—578. Oring, L. W. 1986. Avian polyandry. Current Orni- thology 3:309-351. PiLASTRO, A., L. Biddau, G. Marin, and T. Mingozzi. 2001 . Female brood desertion increases with num- ber of available mates in the Rock Sparrow. Jour- nal of Avian Biology 32:68-72. Reynolds, J. D. 1987. Mating system and nesting biol- ogy of the Red-necked Phalarope Phalaropus loha- tus: what constrains polyandry? Ibis 129:225-242. Smith, J. N. M, Y. Yom-Tov, and R. Moses. 1982. Polygyny, male parental care, and sex ratio in Song Sparrows: an experimental study. Auk 99: 555-564. Smith, J. N. M., L. E Keller, A. B. Marr, and P. Arcese. 2006. Biology of small populations: the Song Sparrows of Mandarte Island. Oxford Uni- versity Press, New York, USA. Verner, j. and M. F. Willson. 1966. The influence of habitats on mating systems of North American passerine birds. Ecology 47:143—147. WiKTANDER, U., O. OLSSON, AND S. G. NiLSSON. 2000. Parental care and social system in the Lesser Spot- ted Woodpecker Dendrocopus minor. Journal of Avian Biology 31:447-456. The Wilson Journal of Ornithology 120(2):398^01, 2008 Novel Courtship Behavior in the Little Greenbul (Andropadus virens) Alexander N. G. KirscheP ABSTRACT. — The Little Greenbul {Andropadus vi- rens) is a common African forest bird that is thought to form monogamous pair bonds. In March 2007, I observed a male hanging below a presumed female perched on a branch, apparently inspecting her cloaca and clinging to her when she flew from that branch, while singing throughout. This apparent mate guarding behavior suggests that extra-pair fertilizations may oc- cur in this species. Received 3 May 2007. Accepted II September 2007. Extra-pair paternity (EPP) is a widespread phenomenon in socially monogamous birds (Birkhead et al. 1987) and has been docu- ‘ Department of Ecology and Evolutionary Biology, University of California, Los Angeles, CA 90095, USA; e-mail: kirschel@ucla.edu mented for many temperate bird species, such as Great Reed Warbler {Acrocephalus arun- dinaceus) (Hasselquist et al. 1996). It has also been reported for tropical species, such as Blue-black Grassquit (Volatinia jacarina) (Carvalho et al. 2006) and Red-backed Eair- ywren {Malurus melanocephalus) (Karubian 2002). There appears to be no general rule that explains the occurrence of EPPs across so- cially monogamous species (Griffith et al. 2002, Westneat and Stewart 2003). Males also sire young of multiple females in species with polygynandrous mating sys- tems such as Dunnock {Prunella modularis) (Davies 1983), the related Alpine Accentor {P. collaris), and the unrelated Smith’s Longspur {Calcarius pictus) (Birkhead et al. 1993). Po- SHORT COMMUNICATIONS 399 lygynandrous species also have much larger cloacal protuberances than socially monoga- mous species (Briskie 1993). Davies (1983) described the unusual behav- ior of cloaca-pecking in the polygynandrous Dunnock in which males peck female cloacas to induce release of sperm from other males. To my knowledge this is the only example of male birds stimulating the ejection of sperm when its paternity is in doubt. Females, how- ever, eject sperm from subordinate males in feral fowl (Gallus gallus) (Pizzari and Birk- head 2000), and when sperm is old and de- graded in Black-legged Kittiwakes {Rissa tri- dactyla) (Wagner et al. 2004). Males of sev- eral bird species are known to eject eggs from nests when EPPs are suspected (e.g.. Sand Martins [Riparia riparia] [Alves and Bryant 1998] and Blue-footed Boobies [Sula neboux- n] [Osorio-Beristain and Drummond 2001]). The Little Greenbul (Andropadus virens, Pycnonotidae) is a common African forest passerine that occurs in primary and second- ary tropical rainforest, forest edges, and gal- lery forest in sub-Saharan Africa. Previous work has documented variation in Little Greenbul morphology and song across an eco- logical gradient (Smith et al. 1997, Slabbe- koom and Smith 2002), but less is known about its breeding behavior. It has traditionally been described as a territorial breeder (Fish- pool and Tobias 2005) with males competing for territories by singing all day during the breeding season. Playback experiments indi- cate that males respond strongly to the poten- tial threat of conspecific intruders by ap- proaching close to or flying noisily above or past a loudspeaker projecting conspecific song. Males may also change their song per- formance in response to playback by singing back more continuously or falling completely silent (A. N. G. Kirschel, unpubl. data). The objective of this paper is to describe courtship behavior in the Little Greenbul that could im- ply the occurrence of EPPs or an alternative mating strategy such as polygynandry. OBSERVATIONS I observed an apparently novel behavior be- tween two Little Greenbuls on the morning of 23 March 2007 while recording the song of a male at the Limbe Botanic Gardens in Cam- eroon (4°()' N, 9° 12' E). A singing male was hanging upside-down from a horizontal branch behind a perched bird, presumed to be a female. The male was flapping his wings continuously to remain suspended from the branch. His head was facing the rear end of the perched bird and he appeared to be in- specting its cloaca (Fig. 1). The male re- mained in this suspended position, singing in- termittently, for several minutes while the perched bird remained on the branch appar- ently unperturbed. Both birds flew to one branch and then another after 5 min of re- cording with the singing male clutching on to the back of the suspected female in flight. Shortly thereafter the two birds separated. DISCUSSION This was a single observation of a behavior not previously reported. I posit the territorial male was inspecting the female as part of a mate guarding strategy, perhaps investigating for evidence of extra-pair copulations. It be- came more difficult to observe the two birds when they started moving to other perches and I did not observe copulation. This observation suggests the mating system of the Little Green- bul may be more complex than previously un- derstood. Male Little Greenbuls spend so much time singing that females could have opportu- nities to move to neighboring territories to seek extra-pair fertilizations from males they per- ceive to be superior. This behavior has been reported for Great Reed Warblers where fe- males moved through several territories before choosing a male; male reproductive success was correlated with song repertoire size (Has- selquist et al. 1996). The cloaca inspection be- havior could represent high levels of EPP and social monogamy in the Little Greenbul or it could represent an alternative mating strategy, such as polygynandry. The Little Greenbul’s presumed sister tax- on, the Yellow-whiskered Greenbul {Andro- padus latirostris), is known to have a complex mating system. Males display in leks where they occur at high densities, at least in parts of the species’ range, but are territorially mo- nogamous at lower densities (Fishpool and Tobias 2005). The only record of courtship be- havior in the Little Greenbul involved a sing- ing male beating its wings, slightly lowering its tail, and puffing its throat as a suspected female approached through the foliage (Keith 400 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 FIG. 1. Male Little Greenbul suspended from a branch inspecting the rear of a presumed female, while singing. The observed birds were in approximately this position at the Limbe Botanic Gardens m Cameroon with the male clinging either to the branch or perhaps the toes of the upper bird, while moving its head up and out of view behind the perched bird. 1992). Gatherings of 6-10 singing males are not thought to represent lekking behavior be- cause there is no evidence of site fidelity in these congregations (Brosset and Erard 1986, Fishpool and Tobias 2005). Brosset and Erard (1986) also described the Little Greenbul as nomadic and rarely territorial, suggesting song might have a greater role in sexual selection than territory defense. Further behavioral studies, and perhaps a comparison of cloacal protuberances with other species, are needed to gain a better understanding of the Little Greenbul’s mating system. Moreover, work is needed to ascertain whether the courtship be- havior observed in Cameroon is a novel be- havior in birds or convergent to the cloaca- pecking strategy of mate guarding. ACKNOWLEDGMENTS I thank Neil Losin for his illustration of the Little Greenbul behavior I observed, and Ndje Francis for assistance in the field at Limbe Botanic Gardens. I thank D. T. Blumstein and T. B. Smith for assistance with funding, and Hans Slabbekoorn, Jordan Karubian, Brittany Enzmann, Neil Losin, Lincoln Fishpool, and one anonymous reviewer for helpful comments on the manuscript. LITERATURE CITED Alves, M. A. S. and D. M. Bryant. 1998. Brood par- asitism in the Sand Martin, Riparia riparia: evi- dence for two parasitic strategies in a colonial pas- serine. Animal Behaviour 56:1323-1331. Birkhead, T. R., L. Atkin, and A. P. M0ller. 1987. Copulation behavior of birds. Behavior 101:101- 138. Birkhead, T. R., J. V. Briskie, and A. P. M0ller. 1993. Male sperm reserves and copulation fre- quency in birds. Behavioral Ecology and Socio- biology 32:85-93. Briskie, J. V. 1993. Anatomical adaptations to sperm competition in Smith’s Longspurs and other po- lygynandrous passerines. Auk 110:875-888. Brosset, A. and C. Erard. 1986. Les Oiseaux des region forestieres du nord-est de Gabon. Volume 1. Ecolo- gie et comportement des especes. Societe Nationale de Protection de la Nature, Paris, France. Carvalho, C. B. V., R. H. Macedo, and J. A. Graves. 2006. Breeding strategies of a socially monogamous neotropical passerine: extra-pair fertilizations, behav- ior, and morphology. Condor 108:579—590. SHORT COMMUNICATIONS 401 Davies, N. B. 1983. Polyandry, cloaca-pecking and sperm competition in Dunnocks. Nature 302:334. Fishpool, L. D. C. and J. A. Tobias. 2005. Family Pycnonotidae (bulbuls). Pages 124-250 in Hand- book of the birds of the world. Volume 10 (Cuck- oo-shrikes to thrushes) (J. del Hoyo, A. Elliot, and D. A. Christie, Editors). Lynx Edicions, Barce- lona, Spain. Griffith, S. C., I. P. F. Owens, and K. A. Thuman. 2002. Extra pair paternity in birds: a review of interspecific variation and adaptive function. Mo- lecular Ecology 11:2195-2212. Hasselquist, D., S. Bensch, and T. von Schantz. 1996. Correlation between male song repertoire, extra-pair paternity and offspring survival in the Great Reed Warbler. Nature 381:229-232. Karubian, j. 2002. Costs and benefits of variable breeding plumage in the Red-backed Fairy-wren. Evolution 56:1673-1682. Keith, S. 1992. Pycnonotidae, bulbuls. Pages 279-377 in The birds of Africa. Volume 4. (S. Keith, E. K. Urban, and C. H. Fry, Editors). Academic Press, London, United Kingdom. Osorio-Beristain, H. and H. Drummond. 2001. Male boobies expel eggs when paternity is in doubt. Behavioral Ecology 12:16-21. Pizzari, T. and T. R. Birkhead. 2000. Female feral fowl eject sperm of subdominant males. Nature 405:787-789. Slabbekoorn, H. and T. B. Smith. 2002. Habitat-de- pendent song divergence in the Little Greenbul: an analysis of environmental selection pressures on acoustic signals. Evolution 56:1849-1858. Smith, T. B., R. K. Wayne, D. J. Girman, and M. W. Bruford. 1997. A role for ecotones in generating rainforest biodiversity. Science 276:1855-1857. Wagner, R. H., F. Helfenstein, and E. Danchin. 2004. Eemale choice of young sperm in a genet- ically monogamous bird. Proceedings of the Royal Society of London Series B-Biological Sciences 27LS134-S137. Westneat, D. F. and I. R. K. Stewart. 2003. Extra- pair paternity in birds: causes, correlates, and con- flict. Annual Review of Ecology, Evolution, and Systematics 34:365-396. The Wilson Journal of Ornithology 120(2):40 1-403, 2008 A Recording of a Type B Song of the Yellow-throated Warbler Bailey D. McKay’ ABSTRACT — On 28 April 2006 in Ohio and again on 19 May 2006 in North Carolina, I observed and, on the April occasion, recorded a Yellow-throated Warbler (Dendroica dominica) singing a type B song. A sound spectrogram of this rare song is compared with the more common type A song of this species. I present evidence that my recording is a type B song and speculate on the rare condition and function of this song. Received 15 February 2007. Accepted 19 July 2007. The Yellow-throated Warbler {Dendroica dominica) is a common neotropical migrant that breeds in the southeastern United States (Hall 1996). Its song is described by Peterson (2002: 230) as “a series of clear slurred notes dropping slightly in pitch.” Most males sing only one song, but there are a few reports of Yellow-throated Warblers singing a second song (Hall 1996). The second song has been ' Department of Biological Sciences, Auburn Univer- sity, Auburn, AL 36849, USA; e-mail: mckaybd@ aubum.edu described as more musical than the type A song and consisting of four notes on the same pitch, three descending, and ending with one on a higher pitch (Nice 1931). It is not known whether this second song functions like a type B song {sensu Spector 1992). The objective of this paper is to: (1) describe the type B song of the Yellow-throated Warbler and (2) report on the context in which this song was ob- served. OBSERVATIONS On 28 April 2006 in Lawrence County, Ohio and again on 19 May 2006 in Graham County, North Carolina I observed Yellow- throated Warblers, one at each site, singing an atypical song. In accordance with the infre- quent reports of a .second Yellow-throated Warbler song, 1 only heard two individuals singing this song during an estimated 400 hrs of fieldwork with this species during two sea- sons. Both observations occurred at dusk around 1930 hrs and, in both cases, the sing- 402 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 8 6 4 2 8 6 4 2 1 t , U‘\ 1 - L ..fc ^ I ; 1 '- S V ' *■ j ' 1 A A. I \ I 1 B. 0.5 s FIG. I . Spectrograms of the two Yellow-throated Warbler song types as sung by a breeding male in Lawrence County, Ohio. A. Type A song. B. Type B song. Sound below 2 kHz was removed from the spectrogram to eliminate visually distracting background noise. ing birds mixed chip-like call notes with the song. I recorded the Ohio bird singing this song, which, to the best of my knowledge, is the first recording of the type B song of the Yellow-throated Warbler. I made a sound spectrogram of the type B song (Fig. IB) using program Syrinx and compared it to the more common type A song as sung by the same bird (Fig. lA). Compar- ison of the two song types reveals the repeated descending note that dominates the type A song is incorporated into the second song. For example, the last phrase, consisting of several descending notes followed by an ascending note, is similar in both songs. The introduc- tion, however, differs between songs. The type B song is less accented with a single descend- ing introductory note followed by two slightly down-slurred notes and a note that is again similar to the descending note in the type A song. This phrase is repeated and followed by the last phrase which is similar to the last phrase of the type A song. Raw recordings of all the songs discussed are available from the author. DISCUSSION The type B song (Fig. IB) is consistent with the only description of a Yellow-throated War- bler second song (Nice 1931) in that it con- tains several notes on the same pitch (although these are mixed with slurred notes) and four descending notes followed by one that is an upsweep, or a rising note. In addition to matching the only description of the Yellow- throated Warbler second song, evidence the observations I am reporting are type B songs include details common to both observations which fit the pattern of type B song use out- lined by Spector (1992). For example, both instances occurred at dusk when the type B song is more frequent and the type A song is usually absent (Spector 1992). The birds in both cases sang type B songs in rapid succes- sion and mixed chip-like call notes between songs, which are characteristics associated with type B songs for Dendroica species (Spector 1992). Type B songs typically in- crease in frequency later in the breeding sea- son (Spector 1992) and both of my observa- SHORT COMMUNICATIONS 403 tions came rather late in the breeding season for Yellow-throated Warblers, which begin nesting earlier than most neotropical migrants (Hall 1996). The context of my observations also agrees with a general pattern of type B song function (Spector 1992). The Yellow-throated Warbler is relatively habitat-specific and, in the eastern portion of its range, prefers the canopy of ma- ture pine {Firms spp.) stands consisting of ~a dozen trees (pers. obs.). This habitat configu- ration seems to be most common in isolated patches within young deciduous forests. In the west, this bird prefers sycamore (Platanus spp.) trees along stream bottoms or mature bald cypress trees (Taxodium spp.) within a swamp (Hall 1996). All of these circumstanc- es cause Yellow-throated Warbler occurrence to be patchy and it is rare to find more than 2-3 singing individuals within earshot of one another (pers. obs.). The type B song func- tions in male-male interactions in most Den- droica species and I propose the patchy dis- tribution of breeding males within appropriate habitat make Yellow-throated Warbler male- male interactions less frequent. This has made the need for a type B song uncommon in this species. This idea appears to be corroborated by the context in which I observed this rare type B song. In both cases, the song was ob- served where Yellow-throated Warbler density was unusually high. In Ohio, I could hear six singing Yellow-throated Warbler males from the spot I recorded the type B song. Likewise, in North Carolina, Yellow-throated Warbler density was the highest I have observed, and 1 could hear seven singing males from where I heard the type B song. I conducted field work with Yellow-throated Warblers at eight other sites in seven states and did not hear this song at any of these locations, where I could not hear more than four, and usually no more than two, males singing from any one loca- tion. The bird in Ohio belonged to the western D. d. albilora subspecies whereas the bird in North Carolina belonged to the eastern D. d. dominica subspecies indicating that a second singing behavior in the Yellow-throated War- bler continues to function in both of its major subspecies. I predict the type B song is more common in the western albilora group as the densities of this warbler are generally higher there than in the east. ACKNOWLEDGMENTS This research was funded through a Frank M. Chap- man grant from the American Museum of Natural His- tory. The Hill Laboratory made helpful comments on the manuscript as did reviewers Jill Soba and D. A. Spector. LITERATURE CITED Hall, G. A. 1996. Yellow-throated Warbler {Dendroi- ca dominica). The birds of North America. Num- ber 223. Nice, M. M. 1931. The birds of Oklahoma. Revised Edition. University of Oklahoma Press, Norman, USA. Peterson, R. T. 2002. Peterson’s field guides: eastern birds. Fifth Edition. Houghton Mifflin Company, Boston, Massachusetts, USA. Spector, D. A. 1992. Wood-warbler song systems: re- view of Paruline singing behaviors. Current Or- nithology 9:199-238. 404 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 The Wilson Journal of Ornithology 120(2):404^07, 2008 Substrate and Vegetation Selection by Nesting Piping Plovers Jonathan B. Cohen,’'^ Elizabeth H. Wunker,^ and James D. Fraser' ABSTRACT. — We studied substrate composition and vegetation cover at Piping Plover (Charadrius melodus) nests and paired random plots on New York beaches that had been widened by renourishment (de- position of dredged sand). Most nests (59.4%, n = 32) were in unvegetated plots, mean ± SE vegetative cov- er around nests was 7.5 ± 1.7%, and all plovers nested in <47% cover. Most nests (59.4%) were on pure sand and mean coarse grain cover (pebble and cobble-sized objects) on nest plots was 9.1 ± 2.6%. Nest plots were more likely to be vegetated than paired random plots. Coarse substrate also was of high relative importance in distinguishing nests and random plots. Beach man- agement projects can reduce sparse vegetation and coarse substrate, which may affect Piping Plover nest site selection. Received 9 December 2006. Accepted 1 August 2007. Piping Plovers {Charadrius melodus) se- lecting nest sites on Atlantic Coast beaches must balance the risks of tidal flooding and egg predation (Burger 1987). They scrape nests in bare or sparsely-vegetated sand, grav- el, or shell substrates on beaches and inter- dune areas, at times adjacent to vegetation clumps, and rely on camouflage and an un- obstructed view of their surroundings to avoid predation (Haig and Elliot-Smith 2004). They may nest close to large objects such as stones or logs (Haig and Elliot-Smith 2004), which may serve as windbreaks or nest markers (Wamock et al. 2002). Atlantic Coast beaches often are altered by people who widen them (beach renourish- ment), build dunes, and plant vegetation (USDI 1996). These alterations are intended to protect human property (Finkl et al. 1988), but also may affect shorebird habitat. Plant- ings that lead to high stem densities may pre- clude shorebird nesting, and substrate or cover modifications may affect nest distribution and ' Department of Fisheries and Wildlife Sciences, Virginia Tech, Blacksburg, VA 24061, USA. 2 Department of Environmental Studies, Warren Wilson College, Asheville, NC 28815, USA. ^ Corresponding author; e-mail: jocohenl@vt.edu susceptibility to predation, weather, and flood- ing. Reducing the negative impacts of beach projects on Piping Plovers depends on under- standing landscape traits that affect nest site selection. We studied Piping Plover nest site selection on renourished beaches in New York in 2004 at sites where future renourishment was planned. Our objective was to learn if Piping Plover nest sites differed from nearby random sites in vegetative cover, coarse materials in the substrate, and number of large objects such as stones and woody debris. METHODS Study Area.— We studied Piping Plovers on Westhampton Island (40° 46' N, 72°43'W), Long Island, New York, from March to Au- gust 2004. Our site consisted of a 6-km long ocean beach on the western end of the island and 3-km long ocean beach on the east end. The two sites were 15 km apart. The habitat contained several parallel cover type zones elongated east to west: ocean intertidal zone, bare backshore, sparsely vegetated backshore/ dune, densely vegetated backshore/dune, and human land use. The U.S. Army Corps of En- gineers renourished the western beach in win- ter 2000/2001 and 50% of the eastern beach in winter 2003/2004. Field Methods.— We estimated substrate composition and vegetation cover around each nest in the first week of incubation using 1 X 1 m sampling quadrats with 36 grid intersec- tions. We placed quadrats over nests with one side parallel to the shore and the nest under the central grid square. We recorded the pres- ence of sand (rock grains <2 mm), pebble (2- 64 mm), cobble (>64 mm), small shells (<2 mm), medium shells (2—64 mm), large shells (>64 mm), small dead wood (<64 mm), large dead wood (>64 mm), beach grass or other plant species (if a stem was emerging from the substrate at the intersection), and other material (such as wrack or debris) under each SHORT COMMUNICATIONS 405 TABLE 1. Characteristics at Piping Plover nests and random sites, Westhampton Island, New York, 2004 {n = 32, l-m^ nest and paired random plots within 50 m). Vegetation Coarse grains Cover (%) Cover (%) Plots near^ > 1 large object Plot type X SE Max Plot.s in“ >5% cover x SE Max Plots in“ >5% cover Nest 7.6 Random 4.4 1.7 45.7 1.9 48.6 0.41 9.1 0.19 5.5 2.6 2.2 55.6 47.2 0.41 0.22 0.28 0.16 “ Mean of the index used in our nest site selection model (0 = < 5% cover, 1 = 5-~100% cover). ^ Mean of the large objects (rock, shell, wood >64 mm in at least one dimension) index used in our nest site selection model (0 = no large objects, 1 = > 1 large object). Where large objects were present, counts ranged from 1 to 5. grid intersection. We recorded the total num- ber of large (>64 mm) objects within the quadrat, regardless of whether they were un- der an intersection. We sampled a paired site at a randomly generated bearing (0-360°) and distance (1- 50 m) from each nest. If we encountered an unsuitable cover type while walking from a nest to its paired point, we turned at a right angle and continued walking. If more than one right angle directed us into potential habitat, we flipped a coin to decide which way to turn. Data Analyses. — We calculated percent veg- etative cover as the number of grid points over an emerging plant/total grid points. We calcu- lated percent cover of each substrate class as the total grid points over that class/total grid points without plants. We modeled selection of nest versus random plots using matched-pair logistic regression (Hosmer and Lemeshow 1998). We evaluated the global model fit using residual and influence plots (Hosmer and Lemeshow 1998). The data were too sparse for logistic regression at high levels of vegetative and coarse grain cover, and number of large objects. Therefore, we created categorical indices where 0 indicated <5% cover and 1 indicated 5-100% cover. Plots with >5% cover had at least two stems or coarse grains within the grid, and the 5% cutoff yielded adequate cell sizes for model fitting. We defined a large object index where 0 indicated no large objects and 1 indicated >one large ob- ject. We performed regression analyses of all possible subsets, and identified the most par- simonious models using corrected Akaike’s Information Criteria (AIC^) and related infor- mation-theoretic criteria (Burnham and An- derson 2002). We inferred the importance of the explanatory variables’ effects using mod- el-averaged unconditional 95% confidence in- tervals (Burnham and Anderson 2002), and a relative importance (/?,) index (Burnham and Anderson 2002). RESULTS We collected data on 32 first nest attempts. Most nest and random plots were unvegetated. TABLE 2. Piping Plover nest site selection model variables and information-theoretic critei subsets of matched-pairs logistic regression models, Westhampton Island, New York, 2004 (/! and paired random plots within 50 m). ria for all possible ' = 32, l-m^ nest Variables in model Number of parameters A1C,.“ AAlC.-t’ coi'-- Model likelihood^ In vegetation (y/n), on coarse substrate (y/n) 2 38.883 o.ooo 0.568 1 .000 In vegetation (y/n), on coarse substrate (y/n), large objects (y/n) 3 40.560 2.157 0.193 0.340 In vegetation (y/n), large objects (y/n) 2 41.424 3.984 0.077 0. 1 36 In vegetation (y/n) 1 4 1 .670 4.21 1 0.069 0.122 On coarse substrate (y/n) 1 42.640 5.531 0.036 0.063 Null 1 43.587 6.187 0.026 0.045 Large objects (y/n) 1 44.494 6.829 0.019 0.033 On coarse substrate (y/n), large objects (y/n) 2 45.130 7.578 0.013 0.023 “Corrected Akaike'.s Int'ormation Criterion. ^ Difference between the AIC, of a particular model and that of the best model. Model likelihood/ilikelihood of all models, d O.SxiAlC, ) 406 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 TABLE 3. Model -averaged effects of vegetation cover, coarse grain cover, and number of large objects on Piping Plover nest site selection {n = 32, 1-m^ nest and paired random plots within 50 m), Westhampton Island, Long Island, New York, 2004. Variable Estimate SE 95% Lower CL Upper R» Vegetation cover >5% 1 .749*^ 0.820 0.141 3.356 0.907 Coarse grain cover > 5% 1.364 0.734 -0.074 2.802 0.809 Large objects present 0.175 0.263 -0.340 0.690 0.302 ^ Relative importance of variable = Scoj of all models containing the variable. Significant effect (unconditional confidence intervals do not contain 0). and most were on pure sand that contained no large objects (Table 1). Three of the eight models we tested had some support (model likelihood >0.125, Table 2). Nest plots were more likely to be vegetated than random plots based on model-averaged statistics using all eight models (Table 3). Percent cover by coarse grains was of high relative importance (Table 3), providing some evidence that plo- vers were more likely to nest in coarse sub- strates than in pure sand. DISCUSSION All Piping Plover nests in our study were in <50% cover with most on bare ground, but they occurred in vegetation more often than ex- pected based on its availability. Apparent se- lection of vegetated nest sites may result if plo- vers nest near vegetated upland zones away from the water line to avoid flooding (Burger 1987, Espie et al. 1996). Average cover around nests in our study was less than in Maryland and Virginia (average 11%, range 9—16%; Pat- terson 1991) and Massachusetts (15.2 ± 23.6%; Macivor 1990). Average cover at ran- dom points in Massachusetts was twice as high as in our study (8.7 ± 16%; Macivor 1990) suggesting that vegetation was less available in our study area than in Massachusetts. Plant loss due to recent renourishment (Looney and Gibson 1993) at our site may have explained why vegetation was relatively sparse. Nests in vegetation can experience more predation than those in the open (Prin- diville Gaines and Ryan 1988, Espie et al. 1996). Thus, renourishment may benefit Pip- ing Plovers by suppressing vegetation. How- ever, renourishment often is followed by beach grass (Ammophila spp.) planting (French 2001) with initial ground cover rang- ing from 3 to 25%. Once beach grass becomes dense, it may have to be thinned each growing season because it can fully recover within 1 year of thinning (U.S. Army Corps of Engi- neers 1967). Piping Plovers in Massachusetts (Jones 1997) and New Jersey (Burger 1987) selected territories with mixed substrate. The preva- lence of nests in pure sand in our study may have been related to birds nesting away from the water, near the sandy dune. Coarse sub- strate may also not have been widely avail- able; percent cover in random plots was ap- proximately half of that at random sites in Massachusetts in the late 1980s (—11%; Macivor 1990). This may also be a result of recent renourishment, which often creates fine-grained beaches (Atlantic States Marine Fisheries Council 2002). Coarse substrate was associated with high hatching success in North Dakota, most likely through camouflage of adults and eggs (Prindiville Gaines and Ryan 1992). Thus, reduced availability of pebbles and cobble may be a negative con- sequence of renourishment that could be mit- igated by replacement of coarse grains. ACKNOWLEDGMENTS The U.S. Army Corps of Engineers, New York Dis- trict, funded this research. The U.S. Fish and Wildlife Service, Long Island Field Office and Region 5 Office; New York Department of Environmental Protection; Suffolk County Parks; Town of Southampton; villages of West Hampton Dunes and Westhampton Beach, New York; and the Long Island Chapter of The Nature Conservancy provided logistical support. Field data were collected by Ellen Creveling, Bradley Pickens, and Dawn Wiley. D. H. Catlin and S. M. Karpanty reviewed the manuscript. LITERATURE CITED Atlantic States Marine Fisheries Council. 2002. Beach nourishment; a review of the biological and physical impacts. ASMFC Habitat Management Series # 7. Atlantic States Marine Fisheries Coun- cil. Washington, D.C., USA. SHORT COMMUNICATIONS 407 Burger, J. 1987. Physical and social determinants of nest-site selection in Piping Plover in New Jersey. Condor 89:811-819. Burnham, K. P. and D. R. Anderson. 2002. Model selection and multimodel inference: a practical in- formation-theoretic approach. Second Edition. Springer- Verlag, New York, USA. Espie, R. H. M., R. M. Brigham, and P. C. James. 1996. Habitat selection and clutch fate of Piping Plovers (Charadriiis melodiis) breeding at Lake Diefenbaker, Saskatchewan. Canadian Journal of Zoology. 74:1069-1075. Einkl Jr., C. E, J. Walker, and I. Watson. 1988. Shoreline erosion: management case history from southeast Elorida. Ocean Shoreline Management 11:129-144. Erench, P. W. 2001. Coastal defences: processes, prob- lems, and solution. Routledge, London, United Kingdom. Haig, S. M. and E. Elliott-Smith. 2004. Piping Plo- ver {Charadrius melodus). The birds of North America. Number 2. Hosmer, D. and S. Lemeshow. 1989. Applied logistic regression. John Wiley and Sons, New York, USA. Jones, L. K. 1997. Piping Plover habitat selection, home range and reproductive success at Cape Cod National Seashore, Massachusetts. Thesis. Uni- versity of Massachusetts, Amherst, USA. Looney, P. B. and D. J. Gibson. 1993. Vegetation mon- itoring related to beach renourishment of the Gulf Islands National Seashore, Perdido Key, Elorida. Pages 226-241 in Coastal Zone 1993: Proceedings of the Eighth Symposium on coastal and ocean man- agement (O. T. Magoon, W. S. Wilson, H. Converse, and L. T. Tobin, Editors). American Society of Civil Engineers, Reston, Virginia, USA. MacIvor, L. H. 1990. Population dynamics, breeding ecology, and management of Piping Plovers on outer Cape Cod, Massachusetts. Thesis. Univer- sity of Massachusetts, Amherst, USA. Patterson, M. E., J. D. Fraser, and J. W. Roggen- BUCK. 1991. Factors affecting Piping Plover pro- ductivity on Assateague Island. Journal of Wild- life Management 55:526-531. Prindiville Gaines, E. M. and M. R. Ryan. 1988. Piping Plover habitat use and reproductive success in North Dakota. Journal of Wildlife Management 52:266-273. U.S. Army Corps of Engineers. 1967. Dune stabili- zation with vegetation on the Outer Banks of North Carolina. Technical Memorandum 22. De- partment of the Army, Corps of Engineers, U.S. Army Coastal Engineering Research Center, Washington, D.C., USA. U.S. Department of Interior (USDI). 1996. Piping Plover {Charadrius melodus), Atlantic Coast Pop- ulation. Revised Recovery Plan. USDI, Fish and Wildlife Service, Hadley, Massachusetts, USA. Warnock, N., C. Elihick, and M. A. Rubeja. 2002. Shorebirds in the marine environment. In Shore- birds. Breeding behavior and populations (E. A. Schreiber and J. Burger, Editors). CRC Press, Boca Raton, Florida. USA. The Wilson Journal of Ornithology 120(2):407-409, 2008 Nest Site Selection by a Male Black-capped Vireo Andrew J. Campomizzi,'’^ Shannon L. Farrell,' and Jerrod A. Butcher' ABSTRACT— We observed a male Black-capped Vireo {Vireo atricapilla) exhibiting nest site selection in east-central Texas. The paired male was observed to de- construct the nest the female was assembling. To our knowledge, male nest site selection has not been observed and reported in the literature for vireos. Received II De- cember 2006. Accepted 6 September 2{)07. It is not uncommon for male pas.serines to begin building nests before they form pair bonds (James 1978, Collias and Collias 1984). Coop- ' Department of Wildlife and b'isheries Sciences, Texas A&M University, College Station, TX 77843, USA. ‘Corresponding author; e-mail: acampomi/.zi@ neo.tamu.edu erative and male-initiated nest-building behav- iors are related to pair bonding, parental quality, and the ability of males to invest in reproduction (Collias and Collias 1984, Hoi et al. 1996, Soler et al. 1998). Nest building by males prior to pairing has been observed in several vireo spe- cies including Plumbeous Vireo {Vireo plutii- heus) (DeMarco et al. 2()00), Bell's Vireo {V. hellii) (Brown 1993), and Yellow-throated Vireo {V. jlavif rolls) (James 1978, Rodewald and James 1996). James ( 1978) observed male Blue- headed Vireos {V. solitcirius) building nests be- fore pairing. James (1978) also noted that com- mencement of paired nest building occuned soon after pairing and hypothesized that males were selecting nest sites. Graber (1961) ob- 408 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 served that male Black-capped Vireos (V. atri- capilla) began building nests before pairing. She reported that, after pairing, a nest would be built by the pair and concluded it was the female that selected the nest site. We observed a male Black-capped Vireo exhibiting nest site selec- tion in April 2006. We also observed the male deconstructing the nest the female was assem- bling. To our knowledge, these behaviors have not been observed and reported for vireos. OBSERVATIONS We conducted nest searching as one com- ponent of research to monitor occupancy and population trends of Black-capped Vireos on private land in Coryell and Hamilton counties in east-central Texas. We monitored 17 pairs of Black-capped Vireos for -864 hrs from 16 March to 31 July 2006. We monitored each nesting pair 2 days per week. We first observed the male Black-capped Vireo on 17 April as it carried nest building material twice to a fork in a branch of a shin oak (Quercus sinuata) —0.5 m from a large downed Ashe juniper {Juniperus ashei). The male was observed for >3 hrs and a female was not detected. We observed the male sing- ing within 20 m of this first nest (nest A) and up to 250 m distant. Nest A consisted of strands of webbing in a slightly cup-shaped platform typical of a partially built vireo nest. We observed the male within 5 m of nest A on 20 April. Nest A was unchanged and a female was not detected. We observed the male for 3 hrs singing within 20 m of nest A and up to 250 m distant on 24 April. We did not observe a female on 24 April and the con- dition of the nest A was unchanged. We observed a female Black-capped Vireo with the male on 27 April. We observed the male singing within 10 m of the female at 0920 hrs. The female carried nesting material to nest A at 0935 hrs which was —15% built. At 1000 hrs the female carried nesting mate- rial north to a different location where nest B was later found. The female carried nesting material to nest B again at 1115 hrs. While the female was arranging nesting material at nest B, the male landed next to her and the female flushed from the nest. The male pulled material from nest B and flew in the direction of nest A. Nest B was - 10% built at that time. We observed the male and female at nest A at 1130 hrs. Between 1130 and 1200 hrs the male flew to nest A without nesting material and perched -1 m from the nest five times. During the same period, the female carried nesting material to nest A twice. Nest A was completed on 1 May and we ob- served the male singing within 10 m of the nest. The condition of nest B was unchanged from 27 April. Nest A was empty on 4 May and still in completed condition. We observed four Black-capped Vireo eggs and one Brown-head- ed Cowbird {Molothrus ater) egg in nest A on 8 May. Nest A was empty and undamaged on 11 May. We observed this pair for approxi- mately 29 hrs throughout the breeding season. During this time, the pair built and laid eggs in three additional nests. We did not locate any other breeding pairs of Black-capped Vireos within -1.5 km of their territory. DISCUSSION Our observations suggest the male Black- capped Vireo preferred one nest over another, selecting it while the female was actively building both nests. Our observations may re- flect behaviors involved in the process of nest site selection. It is possible the male began building both nests and then selected one. It is not uncommon for male Black-capped Vir- eos to begin building nest rims prior to pairing (Grzybowski 1995). These beginning nests are often displayed to females as part of courtship and may subsequently serve as an active nest with approval of the female. We have ob- served Black-capped Vireo pairs deconstruct- ing old nests for material to be used in new nests. Bent (1950) and Nolan (1960) described aggressive interactions between male and fe- male vireos during courtship and nest build- ing. However, we were unable to find reports in the literature of males selecting nests sites to the contrary of the female’s choice or males deconstructing a nest being built by a female by actively removing nest material. It is pos- sible this behavior occurs occasionally but is not observed or reported. Grzybowski (1990) reported that Black-capped Vireos often nest in aggregations and smaller aggregations con- tain a greater proportion of second-year males. We did not detect other Black-capped Vireos nesting within 1.5 km of this pair. We did not ascertain the age of these birds and speculate the male may have been in his second-year. SHORT COMMUNICATIONS 409 the behavior we observed may be the result of a laek of breeding experience or inequality in experience between the male and female. It is difficult to draw any conclusions con- cerning male nest site selection in Black- capped Vireos based on the observations of one pair. We suggest that researchers consider this behavior to document the frequency of male nest site selection. If male nest site se- lection is not uncommon, there may be op- portunities for researchers to examine differ- ences in male initiated nests versus female ini- tiated nests including nest success, nest site characteristics, and initiation dates. ACKNOWLEDGMENTS Funding was provided by a grant from the U.S. De- partment of Defense, Environmental Readiness Pro- gram. We thank the private land owners who allowed us to work on their properties, M. L. Morrison and R. N. Wilkins for guidance in our research and construc- tive comments on this manuscript, and our field tech- nicians for their efforts. LITERATURE CITED Bent, A. C. 1950. Life histories of North American wagtails, shrikes, vireos and their allies. U.S. Na- tional Museum Bulletin Number 197. Brown, B. T. 1993. Bell’s Vireo (Vireo bellii). The birds of North America. Number 35. COLLIAS, N. E. AND E. C. CoLLiAS. 1984. Mate selec- tion and nest building. Pages 57-73 in Nest build- ing and bird behavior. Princeton University Press, Princeton, New Jersey, USA. DeMarco, T. E., C. B. Goguen, D. R. Curson, and N. E. Mathews. 2000. Breeding behavior of the Plumbeous Vireo in New Mexico. Western North American Naturalist 60:394-402. Graber, J. W. 1961. Distribution, habitat requirements, and life history of the Black-capped Vireo {Vireo atricapilla). Ecological Monographs 31:313-336. Grzybowski, J. a. 1990. Population and nesting ecol- ogy of the Black-capped Vireo in Texas — 1988- 1989. USDI, Fish and Wildlife Service, Ecologi- cal Services, Arlington, Texas, USA. Grzybowski, J. A. 1995. Black-capped Vireo {Vireo atri- capillus). The birds of North America. Number 181. Hoi, H., B. Schleicher, and F. Valera. 1996. Nest size variation and its importance for mate choice in Penduline Tits, Remiz pendulinus. Animal Be- haviour 5 1 :464-466. James, R. D. 1978. Pairing and nest site selection in Solitary and Yellow-throated vireos with a de- scription of a ritualized nesting display. Canadian Journal of Zoology 56:1153-1159. Nolan, V. 1960. Breeding behavior of the Bell’s Vireo in southern Indiana. Condor 62:225-244. Rodewald, P. G. and R. D. James. 1996. Yellow- throated Vireo {Vireo flavifrons). The birds of North America. Number 247. SOLER, J. J., A. P. M0LLER, AND M. SOLER. 1998. NeSt building, sexual selection and parental investment. Evolutionary Ecology 12:1153-1 159. The Wilson Journal of Ornithology 120(2):409^12, 2008 Nests of Black-throated Green Warblers in Tree Cavities Douglas C. Tozer' ABSTRACT. — Black-throated Green Warblers {Dendroica virens) typically place their nests within the dense foliage of a limb or in a branch fork against the trunk of a living conifer. 1 report four unusual nests from Algonquin Provincial Park, Ontario, Canada: three in feeding cavities of Pileated Wewdpecker {Dry- ocopiis pileatiis) in snags (i.e., dead trees), and one in a sugar maple borer {Glycohius speciosus, Coleoptera: Cerambycidae) scar in a tree of declining health. These data are the first documentation of this species nesting ' Watershed Ecosystems Graduate Program, Envi- ronmental Science Centre, 1600 West Bank Drive, Trent University, Peterborough, ON K9J 7B8, Canada; e-mail: dtozer@trentu.ca within cavities and enhance our understanding of the importance of snags and trees in declining health for wildlife. Received 15 February 2007. Accepted 20 September 2007. Black-throated Green Warblers {Dendroica virens) are common breeding birds in a vari- ety of forest types ranging from mostly conif- erous to predominantly deciduous across the southern half of Canada east of Alberta, and in parts of the eastern United States (Morse 1993). Throughout its range, the species is 410 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 FIG 1 Nests of Black-throated Green Warblers in Pileated Woodpecker feeding cavities (A C) and in a sugar maple borer scar (D) in Algonquin Provincial Park, Ontario, Canada. Photographs taken in May or June 2006 while nests were active, except B (Oct 2006). closely tied to conifers for nesting (Robichaud and Villard 1999) but, depending on geo- graphical location and habitat, it may prefer either coniferous or deciduous substrates for foraging (Parrish 1995). The species normally selects a territory that contains groups of east- ern hemlock {Tsuga canadensis) and other co- nifers (e.g., eastern white cedar [Thuja occi- dentalism and white spruce [Picea glauca]) in forests dominated by sugar maple {Acer sac- charum) in central Ontario (Grins 1987, Peck and James 1987); some territories contain only one or two canopy-sized conifers (Rob- ichaud and Villard 1999) or none at all (Bent 1953). Even within territories that contain only a few conifers, it appears to prefer to place its nest in coniferous (83%) versus de- ciduous (17%) trees (n = 41 nests in Ontario; Peck and James 1987), typically within the dense foliage of a limb or in a branch fork against the trunk (Harrison 1975, 1979). The objective of this paper is to describe four unusual nests of Black-throated Green Warblers that were built in cavities in trees. To my knowledge, no previous study has doc- umented use of cavities for nesting by this species (e.g.. Chapman 1907, Bent 1953, Morse 1993) and none of the records currently in the Ontario Nest Records Scheme mention use of cavities for nesting {n = 74 nests; Peck and James 1987; M. K. Peck, pers. comm.). OBSERVATIONS I conducted extensive nest searching be- tween early-May and late-July 2006 in 22 sug- ar maple-dominated forest stands (15—40 ha each) scattered across the western portion (-3,500 km^) of Algonquin Provincial Park, Nipissing District, Ontario, Canada (45 37 N, 78° 21' W). The canopy of each stand was -85% sugar maple with the remainder pre- dominantly American beech (Fagus grandi- folia), eastern hemlock, and yellow birch (Bet- ida alleghaniensis). The stands were at differ- ent stages within the single-tree selection (Hunter 1990) harvest cycle (-25 years be- tween cuts) or were uncut reference stands that had not been cut for 60 years. SHORT COMMUNICATIONS 411 I incidentally located six nests of Black- throated Green Warblers; four of these nests were unusual as three were at the bottom of Pileated Woodpecker (Dryocopus pileatus) feeding cavities in snags (i.e., dead trees; Fig. lA-C) and the fourth was at the bottom of a sugar maple borer (Glycobius speciosus, Co- leoptera: Cerambycidae) scar (Fig. ID). The scar was healing, but had protruding bark at the bottom that supported and concealed the nest from below, above, and three-quarters of the sides (Fig. ID) allowing the adults to enter the well-concealed nest through the remaining open quarter. The nests were in three tree spe- cies and were 1.5— 4.5 m above the ground; one nest successfully fledged young, one was depredated by an unknown predator, and the fates of the others were unknown (Table 1). One nest was inactive when found, but the material, size, and structure of the cup matched the rest of the nests where adults were observed. There were at least five ma- ture, canopy-sized eastern hemlocks within a 75-m radius of each of the four nests sug- gesting the absence of conifers (the favored nesting substrate) was not a factor influencing the birds to place their nests in cavities. One to four male Black-throated Green Warblers were on territory in each of the 22 stands, es- pecially where eastern hemlocks provided a coniferous canopy component. DISCUSSION Previous reports have described unusual nesting locations for Black-throated Green Warblers. Three nests were in open areas: 0.2 m above ground in a small red cedar (Jimi- perus virginiana) in a reclaimed pasture (Bent 1953); in a barberry (Berheris spp.) —100 m from the nearest forest (Brewster 1906); and well-concealed within a grapevine (Vitis spp.) “some distance from any woods” (Brown 1889:74). Two nests were at low heights in swampy forest: 0.3 m above ground in the stems of a skunk cabbage {Symplocarpus foe- ticlus), catbriar {Smilax rotundifolia), and a dead bush (Bowdish 1906); and on the ground among ferns (Brewster 1895). The geograph- ically isolated subspecies I), v. waynei of the southeastern United States coastal plain reg- ularly nests at a variety of heights in cypress {Taxodiiitn spp.), oak (Quercus spp.), and magnolia {Magnolia spp.) (Morse 1993). OJ CJ (U D. •o I § ■5 u z ° CQ 2 •= H I qI: % C3 E 0) 'S •o OJ •a D 5 c « c I r- S= u C ■5 D. C C (D C CQ D Q D > y; O C ^ U = 1 ^ c ^ I 5 c c 'c "c D D cj T O -s — > > rs u u u T3 ^ — L. iJ V V rs V V V u Li, X U- 00 < cc u ^ 1 1 i i i t I f I -2 II ^ c ci ^ 5 - u 15 s ? ^ ^ ii 5: ri £ j! c “ ,il Z „ _ -3 -3 g # i Z — X ^ = - •= = 5-^ 7 i ' ^ y z C t = C. ,> ^ II T3 412 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 2, June 2008 It is perhaps not surprising that this species, at least occasionally, nests in cavities. At least two other wood-warblers (Parulidae) that nor- mally do not place their nests in cavities have been documented using them; Orange- crowned Warbler (Vermivora celata), which normally nests in well-concealed locations among ground detritus in small depressions or crevices and only occasionally builds elevated nests in dense vegetation (Sogge et al. 1994), has nested 3.2 m above ground in a natural cavity of a canyon maple (Acer grandidenta- tum) (Zyskowski 1993); and Common Yel- lowthroat (Geothlypis trichas), which normal- ly nests on or up to 0.1 m above ground in thick vegetation (Guzy and Ritchison 1999), has nested 0.6 m above ground in a shoe that was hung on the side of a house (Brockway 1899). Cavity-use by nesting Black-throated Green Warblers may be a regular but infre- quent occurrence. Future research is warranted to identify the frequency and importance of cavity-nesting by Black-throated Green Warblers. It is not known if tree harvesting activities may or may not be influencing this little-known nesting behavior by reducing the number of snags and trees of declining health that might have served as cavity nesting sites. ACKNOWLEDGMENTS I thank Kris Vande Sompel and Peter Burke for find- ing three of the unusual nests, and Karla Falk, Meagh- an Noad, Andrea Orr, Eleanor Proctor, Sonya Rich- mond, and Philip Wilson for help with monitoring. Kris Vande Sompel (Fig. lA, C) and Peter Burke (Fig. ID) provided photographs. Kristof Zyskowski, an anonymous reviewer, Ken Elliott in particular. Dawn Burke, Karla Falk, Erica Nol, Pat Tozer, and Ron Tozer provided helpful comments that improved earlier drafts of the mansucript. Keith Fletcher, Brad Steinberg, Paul Gelok, Andrew Trant, and the Algonquin Park Wildlife Research Station helped with logistics. Funding and in- kind support were provided by Algonquin Forestry Authority, Bancroft Minden Forest Co., Canadian Wildlife Service, Canadian Forest Service, Enhanced Forest Productivity Science Program, Mazinaw-Lanark Forest Inc., Natural Sciences and Engineering and Re- search Council of Canada, Ontario Ministry of Natural Resources, Ottawa Valley Forest Inc., Tembec Inc., Trent University, and Westwind Forest Stewardship. LITERATURE CITED Bent, A. C. 1953. Life histories of North American wood warblers. Part 1. U.S. National Museum Bulletin Number 203. Bowdish, B. S. 1906. Some breeding warblers of De- marest, N.J. Auk 23:16—19. Brewster, W. 1895. A ground nest of the Black- throated Green Warbler. Auk 12:184-185. Brewster, W. 1906. The birds of the Cambridge re- gion of Massachusetts. Memoirs of the Nuttall Or- nithological Club 4:1—426. Brockway, W. 1899. Odd nesting of Maryland Yel- low-throat. Auk 16:360-361. Brown, J. C. 1889. Unusual nesting site of Dendroica virens. Auk 6:74. Chapman, F. M. 1907. The warblers of North America. D. Appleton, New York, USA. Crins, W. J. 1987. Black-throated Green Warbler. Pag- es 382-383 in Atlas of the breeding birds of On- tario (M. Cadman, P. Eagles, and F M. Helleiner, Compilers). University of Waterloo Press, Water- loo, Ontario, Canada. Guzy, M. J. and G. Ritchison. 1999. Common Yel- lowthroat (Geothlypis trichas). The birds of North America. Number 448. Harrison, H. H. 1975. A field guide to birds’ nests: United States east of the Mississippi River. Peter- son Field Guide Series. Houghton Mifflin, Boston, Massachusetts, USA. Harrison, H. H. 1979. A field guide to western birds’ nests: United States west of the Mississippi River. Peterson Field Guide Series. Houghton Mifflin, Boston, Massachusetts, USA. Hunter, M. L. 1990. Wildlife, forests, and forestry. Prentice Hall, Englewood Cliffs, New Jersey, USA. Morse, D. H. 1993. Black-throated Green Warbler (Dendroica virens). The birds of North America. Number 55. Ontario Ministry of Natural Resources [OMNR]. 2000. A silvicultural guide to managing southern Ontario forests. Ontario Ministry of Natural Re- sources. Queen’s Printer, Ontario, Toronto, Cana- da. Parrish, J. D. 1995. Experimental evidence for intrin- sic microhabitat preferences in the Black-throated Green Warbler. Condor 97:935-943. Peck, G. K. and R. D. James. 1987. Breeding birds of Ontario: nidiology and distribution. Volume 2. Passerines. Royal Ontario Museum, Toronto, Can- ada. Robichaud, I. AND M. ViLLARD. 1999. Do Black- throated Green Warblers prefer conifers? Meso- and microhabitat use in a mixedwood forest. Con- dor 101:262-271. Sogge, M. K., W. M. Gilbert, and C. van Riper III. 1994. Orange-crowned Warbler (Vermivora cela- ta). The birds of North America. Number 101. Zyskowski, K. 1993. Nest-site selection in Orange- crowned and Virginia’s warblers in high-elevation forests of the Mogollen Rim (Arizona): variation in nest placement, phenology, and microclimate. Thesis. University of Arkansas, Fayetteville, USA. SHORT COMMUNICATIONS 413 The Wilson Journal of Ornithology 120(2):413-416, 2008 Nest and Fledglings of the Red-ruffed Fruitcrow (Pyroderus scutatus) Mercival R. Francisco/ '^ Paulo R. R. Oliveira Jr.,^ and Vitor O. Lunardi^ ABSTRACT — Information on the breeding behav- ior of the Red-ruffed Fruitcrow {Pyroderus scutatus) is scarce and restricted to the subspecies P. s. grana- densis and P. s. orenocensis. We found the first nest of the nominate subspecies (P. s. scutatus) in an At- lantic Forest area in southeastern Brazil on 28 Novem- ber 2004. The nest contained two nestlings and was built on a horizontal fork, 16.7 m above ground. It was cup-shaped with a substantial base composed of twigs: outside diameter 38 cm, cup diameter 16.5 cm, outside height 1 1.3 cm, and inside height 5 cm. The nestlings were thickly covered with brownish down. Only one unknown gender adult visited the nest. Although lo- cally endangered, the breeding cycle of this species remains poorly known. Received 10 February 2007. Accepted 12 September 2007. The cotingas comprise a diverse group of birds endemic to the Neotropics (Snow 1982). They inhabit tropical and subtropical forests, and most species are arboreal. These birds are of great interest because different species ex- hibit unusual courtship displays while others have bright plumage or bizarre ornaments. However, information on their breeding biol- ogy is scarce (Snow 1982, Ridgely and Tudor 1994, Sick 1997). The Red-ruffed Fruitcrow {Pyroderus scu- tatus) is a large cotingid which inhabits humid and montane forests, forest borders, and lo- cally somewhat drier regions. Five subspecies are recognized and our paper focuses on the nominate subspecies (from southeastern Bra- ' Univer.sidade Federal de Sao Carkxs, Campus de Sorocaba, CEP 18043-970, Caixa Postal 3031, Soro- caba, SP, Brazil. ^ Universidade Estadual Paulista, Campus de Assis, Departamento de Ciencias Biologicas, Laboratorio de Comportamento de Vertebrados, Avenue Dom Anto- nio, 2100, CEP 19806-900, A.ssis, SP, Brazil. Universidade de Brasilia, Campus Universitario Darcy Ribeiro, ICC-Sul, Instituto de Ciencias Biolo- gicas, Programa de Pos-Gradua^ao em Ecologia, CEP 70910-900, Brasilia, DE Brazil. Corresponding author; e-mail: mercival @power.ufscar.br zil, eastern Paraguay, and northeastern Argen- tina) (Snow 2004). Information on the breeding behavior of the Red-ruffed Fruitcrow is scarce and restricted to the subspecies P. s. granadensis and P. s. orenocensis. Nests described for P. s. grana- densis in Colombia are open shallow cups with the base made mainly of twigs and the cup composed exclusively of the fronds of ferns, placed 5-8 m above ground in slender tree branches (Serrano 1994, Snow 2004). The objectives of our paper are to: (1) describe the nest of the nominate subspecies, P. s. scutatus, from southeastern Brazil, and (2) describe the nestlings. OBSERVATIONS The nest was found on 28 November 2004 at Carlos Botelho State Park, Sao Paulo State, southeastern Brazil (24° 04' 14.6" S, 47° 57' 34.3" W, altitude 851 m). This park contains 37,664 ha of coastal Atlantic Forest at an al- titude of 22 to 1,000 m above sea level. An- nual rainfall can exceed 160 cm in higher el- evations. The nest contained two nestlings and was in a primary forest area where the canopy reached 30-35 m. It was built in a honey wood tree {Alchornea triplinenia, Euphorbi- aceae), 18 m in height at an overhanging branch, 4 m from the trunk and 16.7 m above ground. The nest was in a horizontal fork with branches ~5 cm in diameter. Ferns were abun- dant at the tree crown making the nest almost undetectable from the ground, as well as from the top of the tree. The nest was cup-shaped with a substantial base composed of twigs and the cup was thinly built with dry stripped ferns (Fig. 1). The measurements of the nest were: outside diameter 38 cm, cup diameter 16.5 cm, outside height 11.3 cm, and inside height 5 cm. The nestlings were thickly cov- ered with brownish down, which was concen- trated on their heads and hung over their eyes, that were already open. The bill and gape 414 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 FIG. I. Nest and young Red-ruffed Fruitcrow in Carlos Botelho State Park, SP, Brazil (photograph by M. R. Francisco). were yellow, and the throat was poorly feath- ered with a remarkable pinkish skin. The pri- mary pin feathers were visible, but were bro- ken only at the tip, suggesting the nestlings were fairly young. They were similar in size and were sedentary, not moving or making sounds when handled. An adult repeatedly ap- proached the nest to within 1.5 to 2 m while the nest was being examined. We conducted 12 hrs of focal observations from 0530 to 1730 hrs (UTC-3) on 1 Decem- ber 2004 during which only one unknown gender adult (identified by a broken feather in its tail) visited the nest 40 times. The nestlings were not brooded after being fed except when it was raining. The adult spent 26 and 114 min at the nest, respectively, during two heavy storms. We do not know if the adult brooded the nestlings at night. Visits were 9-192 sec in length (46.52 ± 46.09 sec, n = 27) (mean ± SD) when it was not raining and the time between visits averaged 13.71 ± 9.9 min (range: 4-43 min, n = 35). The adult perched -10-15 m from the nest before provisioning the young and carefully checked the surround- ing area for 10—30 sec, possibly to avoid at- tracting potential predators to the nest. Routes taken by the adult to and from the nest varied. Upon leaving the nest, the adult frequently flew directly to the ground, apparently to cap- ture insects. The adult was observed carrying large insects in its bill in 14 of the 40 visits. Regurgitated seeds of palmito {Euterpe edulis {n = \) and Virola oleifera {n = 3) were found on the top of the branch on which the nest was placed. No fecal material was found near the nest, suggesting the adult removed fecal sacs. The parent was observed chasing SHORT COMMUNICATIONS 415 another adult Red-ruffed Fruiterow that had approached the nest to within 5-6 m on five occasions. DISCUSSION Many large and medium-sized cotingas, such as species in the genera Procnias, Li- paugus, Xipholena, Perissocephalus, Cephal- opterus, and Carpodectes build extraordinari- ly small and flimsy open nests, usually sup- ported by twigs or small branches. Some spe- cies are known to use lichens, mosses, and similar materials for camouflage (Snow 1982, 2004; Sanchez 2002). The nest of the Red- ruffed Fruiterow we observed was quite large and did not present materials for camouflage. However, it was hidden in the middle of ferns on an inaccessible branch. Chasing potential predators appears to be a method of nest de- fense in this species as T. K. Salmon (Snow 1982) and Serrano (1994) observed Red- ruffed Fruitcrows leaving the nests to chase hawks that approached their territories. A base composed of twigs and the cup built with stripped ferns was also observed for P. s. granadensis (Serrano 1994). Snow (2004) noted that Red-ruffed Fruiterow nests are shallow cups that allow light to pass through, but the base of the nest we observed was bulky and substantial. We observed two nest- lings, but clutch sizes in P. s. granadensis were invariably one (n = 6 nests) (Serrano 1994, Snow 2004). The nest we observed was found during the rainy season in southeastern Brazil and coin- cided with the peak of ripening of palmito fruits in the area. Four to five fruiting palmito trees could be seen from the nest, as it was the most abundant tree with fruit at that lo- cation. The presence of a regurgitated palmito seed close to the nest suggests the young were fed this fruit. The relationship between nesting activities of the Red-ruffed Fruiterow and availability of palmito fruits should be ex- amined as entire populations of this tree have been lost to palm heart harvesting (Galetti and Aleixo 1998). In some areas, P. sciitatns are only seen during the fruiting period Euterpe edulis (Galetti et al. 1999). The Red-ruffed Fruiterow is not globally threatened, but P. s. scutatiis is listed as en- dangered in many Brazilian states including Rio Grande do Sul, Sao Paulo, Minas Gerais, and Rio de Janeiro. Many populations of this species have declined, presumably due to hab- itat fragmentation and destruction of the At- lantic Forest ecosystem (Willis 1979, Ribon et al. 2003, Snow 2004). Only 7.5% of the orig- inal 1 million km^ area of Atlantic Forest re- mains distributed in several small and discon- nected fragments, and a few large forest tracts (Myers et al. 2000). Large frugivores, such as the Red-ruffed Fruiterow are among the first species that disappear in small woodlots (Wil- lis 1979, Ribon et al. 2003). The data in this paper provide a starting point for understand- ing the breeding beahvior of P. s. scutatus and may be useful in the future conservation of the species. ACKNOWLEDGMENTS We are grateful to L. F. Silveira, H. L. Gibbs, Cesar Sanchez, and an anonymous referee for important comments on early versions of this manuscript. We also thank Instituto Florestal do Estado de Sao Paulo (IF) for permits for field work at Carlos Botelho State Park. LITERATURE CITED Galetti, M. and A. Aleixo. 1998. Effects of palm heart harvesting on frugivores in the Atlantic For- est of Brazil. Journal of Applied Ecology 35:286- 293. Galetti, M., V. B. Zipparro, and P. C. Morellato. 1999. Eruiting phenology and frugivory on the palm Euterpe edulis in a lowland Atlantic Forest of Brazil. Ecotropica 5:1 15-122. Myers, N., R. A. Mittermeier, C. G. Mittermeier, G. A. B. Fonseca, and J. Kent. 2000. Biodiver- sity hotspots for conservation priorities. Nature 403:853-858. Ribon, R., J. E. Simon, and G. T. Mattos. 2003. Bird extinctions in Atlantic Forest fragments of the Vi- gosa region, southeastern Brazil. Conservation Bi- ology 17:1827-1839. Ridgely, R. S. and G. Tudor. 1994. The birds of South America. Volume 2. University of Texas Press, Austin, USA. Sanchez, C. 2002. Nest, egg, and nesting biology of the Snowy Cotinga {Carpodectes nitidus). Wilson Bulletin 114:517-519. Serrano, V. H. 1994. Generalidades sobre la seleccidn de habitat, el cicio reproductivo y el sistema lek de apareamiento de Pyroderus scutatus (toro de monte). Pages 343-355 in Ucumari: un caso tfpico de la diversidad bidtica Andina (.). (). Rangel. PaI- itor). Carder Corporacidn Autdnoma and Univer- sidad Nacional de Colombia, Cali. Sick. H. 1997. Ornitologia Brasileira. Editora Nova P'ronteira, Rio de .laneiro, Brazil. Snow. D. 1982. Phe cotingas: bellbirds, umbrellabirds 416 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 and other species. Cornell University Press, Itha- ca, New York, USA. Snow, D. B. 2004. Family Cotingidae (cotingas). Pag- es 32-108 in Handbook of the birds of the world. Volume 9. Cotingas to pipits and wagtails (J. del Hoyo, A. Elliott, and D. Christie, Editors). Lynx Edicions, Barcelona. Spain. Willis, E. O. 1979. The composition of avian com- munities in remanescent woodlots in southern Brazil. Papeis Avulsos de Zoologia 33:1-25. The Wilson Journal of Ornithology 120(2):416-418. 2008 Palila {Loxioides bailleui) Fledgling Fed by Hawai‘i ‘Amakihi {Hemignathus virens) Chris Farmer,!'^ Bridget A. Frederickd’^ Paul C. Bankori Robert M. Stephens/ and Carter W. Snow^ ABSTRACT. — We observed an adult male Hawai‘i ‘Amakihi {Hemignathus virens) repeatedly feed a fledgling Palila {Loxioides bailleui). We observed lb- 19 food provisions during 14 hrs of observation be- tween 21 and 29 June 2006. The presumed biological parents were frequently seen nearby, but adult Palila were not observed feeding the fledgling. Received 28 April 2007. Accepted 26 July 2007. Interspecific parental care among non- brood parasitic birds is rare and appears to have no ultimate, evolutionary benefits (Shy 1982, McNair and Duyck 1991, Yoerg and O’Halloran 1991, Lozano and Lemon 1998, Drozdz et al. 2004). Shy (1982) proposed nu- merous possible proximate causes for such be- havior. These can be encompassed by a few, underlying ultimate causes such as the high cost of ignoring one’s own offspring, adult re- sponse to a super-normal juvenile stimulus or gaining experience in chick-rearing. Interspecific feeding within the Hawaiian honeycreepers (Drepanidinae: Fringillidae) is unknown. Palila {Loxioides bailleui) of all age 1 U.S. Geological Survey, Hawai‘i Cooperative Studies Unit, University of Hawai‘i at Hilo/PACRC, KTlauea Field Station, P. O. Box 44, Hawai‘i National Park. HI 96718, USA. 2 U.S. Geological Survey, Pacific Island Ecosystems Research Center. KTlauea Field Station, P. O. Box 44, Hawai‘i National Park, HI 96718, USA. 2 Current address: P. O. Box 6712, Hilo, HI 96720, USA. ^ Corresponding author; e-mail: chris_farmer@usgs.gov classes primarily consume the unhardened seeds of the mamane {Sophora chrysophylla) and have relatively low reproductive capacity (van Riper 1980, Banko et al. 2002). Palila typically have one brood per year (mean = 2 eggs), in part because fledglings do not be- come independent until they are 3-4 months of age (Miller 1998). Conversely, Hawaifi ‘Amakihi {Hemignathus virens; hereafter ‘Amakihi) are generalists that typically feed on arthropods and nectar, have two clutches per year (mean = 2.5 eggs per clutch), and provide parental care for 2-3 months (Lindsey et al. 1998). The dissimilar natural histories of the two species have resulted in considerable differences in their abundance and distribu- tion. Approximately 2,000-3,000 Palila re- main in the subalpine mamane and naio (My- oporum sandwicense) forest on the western slope of Mauna Kea on the island of Hawai‘i (Banko et al. 2002, Johnson et al. 2006). In contrast, ‘Amakihi are abundant throughout a wide range of habitats and elevations in Ha- wai‘i (Scott et al. 1986, Lindsey et al. 1998). The objective of our paper is to document in- terspecific provisioning between these two Drepanidinae species. METHODS Our observations occurred at —2,350 m el- evation on the northern slope of Mauna Kea Volcano, Hawai‘i Island (19°53"N, 155 26 W). This habitat is dry, open woodland dom- inated by mamane with scattered naio and an exotic grass understory (e.g., Dactylis glom- SHORT COMMUNICATIONS 417 erata, Holcus lanatus, Anthoxanthum odora- tum, and Poa pratensis). A small population of Palila has been created at this site through translocations from the main population on the western slope of Mauna Kea and releases of captive-reared birds by the Zoological So- ciety of San Diego. The population consisted of —30 Palila at the time of our observations (21-29 Jun 2006); nearly all were marked with colored plastic and U.S. Geological Sur- vey aluminum leg bands. The ‘Amakihi is the most abundant native or alien passerine found in this habitat (Scott et al. 1986; USGS, un- publ. data). Ornithologists familiar with the local bird community used 8 X 32 binoculars to observe the birds, usually within a distance of 10 m. Behavioral observations were usually made by 1-2 observers, although on 29 June four people watched the fledgling-adult interac- tions. OBSERVATIONS Our observations were of an adult male ‘Amakihi feeding a fledgling Palila, apparent- ly providing the fledgling with most of its food during a 9-day period in June 2006. Two unbanded Palila fledglings of approximately the same age (25-28 days after hatch) were seen 0.3 m apart in a naio tree on 15 June 2006. A banded adult male and female Palila flew into the tree and fed at least one fledg- ling, which we assume was their offspring. This pair was engaged with their third repro- ductive attempt of the season at the time of our observations, having produced a 4-month- old fledgling prior to the 1 -month-old fledg- ling and its sibling described above. The fe- male was incubating an egg (later found to be infertile) in a third nest, 200 m from the fledg- ling, which was eventually abandoned, and the pair subsequently started, but did not fin- ish, a fourth nest. We witnessed a banded (alu- minum band only) adult male ‘Amakihi feed- ing a Palila fledgling on 21 June. The fledg- ling chased after the ‘Amakihi, which twice placed its bill down the fledgling’s throat in a manner suggesting food regurgitation. We banded a Palila fledgling on 22 June that we presumed to be the one that was fed by the ‘Amakihi, based on its subsequent in- terspecific association. The fledgling weighed 29 g, 6 g less than the average hatch-year weight, but its subcutaneous furcular fat (score = 3) was relatively high for its age class (mode = 2; n = 357; USGS, unpubl. data). We relocated the Palila fledgling in the same area on 27 June and observed it repeat- edly begging and being fed by a banded male ‘Amakihi. We saw two feedings and heard seven others, based on changes in the begging calls, over 5 hrs of observation. The Palila fledgling pursued the ‘Amakihi through the trees, ignoring several nearby Palila adults, in- cluding a male that was the presumptive fa- ther. Similar behavior was observed during 3 hrs on 28 June when we saw the fledgling being fed by a banded male ‘Amakihi at least four times and heard a possible fifth feeding. The fledgling fed independently upon the flowers of a native mint {Stenogyne rugosa) and young leaves of mamane and naio, but was not observed eating mamane pods. The Palila fledgling was last seen on 29 June when it was fed up to three times over 5 hrs of ob- servation. We observed 16-19 feedings in 14 hrs (—1.1 provisionings/hr) over 9 days during which the juvenile began its third week after fledging. DISCUSSION Our repeated observations of an adult male ‘Amakihi feeding a fledgling Palila are unique for the Drepanidinae and are the only record of feeding interactions between Palila and any other bird species in 30 years of Palila re- search (Banko et al. 2002). We believe that only one ‘Amakihi fed the Palila fledgling, be- cause the bird we observed was banded with an aluminum band, his appearance was con- sistent among observations, and all feeding occurred in the same area (50 m diameter). ‘Amakihi are territorial during the nesting sea- son (van Riper 1987, Lindsey et al. 1998), which suggests that only one male would oc- cupy this small area. We know nothing about the reproductive effort of this ‘Amakihi prior to or during its interactions with the Palila fledgling. We observed the presumed parents of the Palila fledgling, both of which were wild- translocated birds, attempt at least four nests during 2006. Palila typically raise <1 lledg- ling per year and there are no records of more than three nests per female in 1 year (Banko 418 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 et al. 2002). Thus, the parents might have been unable to provide sufficient food to sus- tain three fledglings and themselves. Parental neglect could have encouraged the fledgling to actively pursue interspecific feeding oppor- tunities (e.g., Verhulst and Hut 1996). We did not see the fledgling beg for food or otherwise interact with its parents or other nearby (<50 m) Palila, although the putative male parent fed the fledgling’s presumed nest-mate on at least five occasions. We have not detected evidence of brood parasitism between bird species on Mauna Kea (USGS, unpubl. data) and it seems likely the Palila we observed was adopted by the ‘Amakihi after it fledged. The Palila fledgling was slightly more than twice the weight of an average male ‘Amakihi (13.4 g) (Lindsey et al. 1998) and provisioning such a large fledg- ling would have been energetically demand- ing. The fledgling’s high level of furcular fat suggests the ‘Amakihi was devoting a large portion of its time to provisioning the Palila. Miller (1998) estimated four provisionings/hr for Palila fledglings during the first post-fledg- ing week while van Riper (1980) observed 0.9 provisionings/hr. The rate of interspecific feeding we observed (1.1 /hr) was at the lower boundary for intraspecific feeding rates at this stage of fledgling development. The rarity of interspecific feeding suggests the behavior of the ‘Amakihi was misdirected, either because of the super-normal stimulus provided by the Palila fledgling’s size and begging behavior or by the loss of its own nest or offspring (Shy 1982). It seems unlikely the ‘Amakihi benefited from its behavior, unless it gained needed experience in provisioning a fledgling. ACKNOWLEDGMENTS This research was made possible in part thanks to support from the Federal Highway Administration; the U.S. Army Garrison, Hawaii; and U.S. Geological Survey wildlife research programs. We thank the Ha- waii Division of Forestry and Wildlife, and the U.S. Fish and Wildlife Service for access to study sites, and other assistance. We are grateful for administrative support from Hawaii Cooperative Studies Unit (Uni- versity of Hawaii at Hilo) and Pacific Island Ecosys- tems Research Center. We also thank four anonymous reviewers for their helpful suggestions. We are espe- cially grateful to the research interns who helped us in the field: M. R. Fredricks, M. A. Cacciapaglia, E. M. Olimpi, and B. S. O’ Hare. Any use of trade, product, or firm names in this publication is for descriptive pur- poses only and does not imply endorsement by the U.S. Government. LITERATURE CITED Banko, P. C., L. Johnson, G. K. Lindsey, S. G. Fancy, T. K. Pratt, J. D. Jacobi, and W. E. Banko. 2002. Palila (Loxioides bailleui). The birds of North America. Number 679. Drozdz, R., M. Hromada, and P. Tryjanowski. 2004. Interspecific feeding of a Great Grey Shrike (Lan- ius excubitor) fledgling by adult Yellowhammers (EmberizcL citrinella). Biological Letters 41:185— 187. Johnson, L., R. J. Camp, K. W. Brinck, and P. C. Banko. 2006. Long-term population monitoring: lessons learned from an endangered passerine in Hawai’i. Wildlife Society Bulletin 34:1055—1063. Lindsey, G. D., E. A. VanderWerf, H. Baker, and P E. Baker. 1998. Hawaii {Hemignathus virens), Kauai {Hemignathus kauaiensis), 0‘ahu {Hem- ignathus chloris), and Greater ‘Amakihi {Hemig- nathus sagittirostoris). The birds of North Amer- ica. Number 360. Lozano, G. A. and R. E. Lemon. 1998. Adoption of Yellow Warbler nestlings by Song Sparrows. Wil- son Bulletin 110:131-133. McNair, D. B. and B. Duyck. 1991. Interspecific feeding among some oscines. Chat 55:9-11. Miller, L. J. 1998. Behavioral ecology of juvenile Palila {Loxioides bailleui): foraging development, social dynamics, and helping behavior. Thesis, University of Maryland, College Park, USA. Scott, J. M., S. Mountainspring, F. L. Ramsey, and C. B. Kepler. 1986. Forest bird communities of the Hawaiian Islands: their dynamics, ecology, and conservation. Studies in Avian Biology 9:1- 431. Shy, M. M. 1982. Interspecific feeding among birds: a review. Journal of Field Ornithology 53:370- 393. VAN Riper III, C. 1980. Observations on the breeding of the Palila Psittirostra bailleui of Hawaii. Ibis 122:462-475. VAN Riper III, C. 1987. Breeding ecology of the Ha- wai‘i Common ‘Amakihi. Condor 89:85-102. Verhulst, S. and R. A. Hut. 1996. Post-fledging care, multiple breeding and the costs of reproduction in the Great Tit. Animal Behaviour 51:957-966. Yoerg, S. I. AND J. O’Halloran. 1991. Dipper nest- lings fed by a Gray Wagtail. Auk 108:427-440. SHORT COMMUNICATIONS 419 The Wilson Journal of Ornithology 120(2):4 19^22, 2008 Adoption: Adaptation or Reproductive Error in Eastern Bluebirds? Daniel R WetzeR’^’^ and C. Ray Chandler* ABSTRACT — We describe an adoption event by a female Eastern Bluebird (Sialia sialis) and suggest a mechanism by which adoption could be adaptive. The foster female adopted three young and we were able to quantify two measures of parental care (nestling provisioning and nest defense). The foster female suc- cessfully reproduced with the father of the young in the same location later in the breeding season. The adopting female may be increasing her fitness by gain- ing future reproduction with a fertile mate at a pro- ductive nest box through parental care for young that are not related. Received 12 March 2007. Accepted 6 October 2007. Adoption is the parental care of presumed non-linear offspring (PNLO) by an individual that is replacing the original parent and care- giver (Plissner and Gowaty 1988). The adap- tive significance of adoption in birds is not well understood because it is thought to occur infrequently within populations, although 150 species of birds have documented cases of adoption (A vital et al. 1998). It is commonly assumed there is no direct benefit to an indi- vidual adopting another’s young. Thus, adop- tion is usually referred to as “reproductive er- ror”. Reproductive error may be maladaptive as it occurs when an individual has strong en- vironmental and behavioral cues to care for young, but their efforts are displaced to un- related young (Rohwer 1986). It is assumed the adopting individual is not discriminating between its own and unrelated young. Avital et al. (1998) suggested that, although adoption seems to occur infrequently, there may be suf- ficient benefit for this trait to remain in a pop- ulation. Individuals may adopt PNLO to gain access to mates or nest-sites, or to display their parental care to potential new mates ' Department of Biology, Georgia Southern Univer- sity, Box 8042, Statesboro, GA 30460, USA. ^ Current address: Department of Biology, 101 T. H. Morgan Building, University of Kentucky, Lexington, KY 40506, USA. ^Corresponding author; e-mail: dan.wetzel@uky.edu (Gori et al. 1996). A new parent that can dem- onstrate its parental care may have an in- creased chance of re-nesting with a new mate (Rohwer 1986). Mate replacement and adoption by the East- ern Bluebird {Sialia sialis) have been docu- mented after male disappearance or experi- mental removal of males (Hamilton 1943, Kreig 1971, Pinkowski 1978, Gowaty 1983a). However, adoption by female Eastern Blue- birds is rare and the extent of parental care by adopting females is unknown (Gowaty 1983b, Rohwer 1986). This species is an obligate sec- ondary cavity nester that breeds from late March through early August in the southeast- ern United States (Gowaty and Plissner 1998). Bluebirds are monogamous with bi-parental care of nestlings. Females typically lay 4-5 eggs in each clutch and lay up to three clutch- es per breeding season (Gowaty and Plissner 1998). OBSERVATIONS We describe a case of adoption by a female Eastern Bluebird in southeastern Georgia on a study area with several color-banded bluebirds and multiple nesting pairs in summer 2005. The study area was established in 2005 on 44.5 ha of open fields and contained 18 nest boxes. Nest box #4, with female F4 and male M4, hatched three young on 19 April and nest box #1 with female El and male Ml, hatched two young on 23 April. The last known ob- servation of female F4 at nest box #4 was on 17 April and the last known observation of male Ml from nest box #1 was on 24 April. On 28 April, nest box #1, ~250 m from nest box #4, contained two dead nestlings, and two cold eggs, with no sign of either parent (FI or Ml). The deceased nestlings of female FI would have been 6 days of age at that time, and the nestlings of male M4 at nest box #4 were 10 days of age. On that same day, pro- visioning and nest defense observations were performed at nest box #4. We quantified pro- 420 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 visioning rates by observing the nest with a 20x spotting scope for 1 hr from a blind at ~30 m distance. Female FI from nest box #1 was observed at nest box #4 with the original nest box #4 male, M4, provisioning the adopt- ed nestlings at a rate of 1.67 times/chickThr, while the original male M4 provisioned the nestlings 6.67 times/chick/hr. Mean (it SE) nestling provisioning rates for females and males on the study site were 1.91 (± 0.21) and 1.86 (± 0.24) provisions/chick/hr, respec- tively. We quantified the defense responses of the new pair by placing a rubber eastern rat snake (Elaphe alleghaniensis) 0.5 m from the base of the nest box under a camouflaged cloth. The cloth was attached to a string operated from a blind ~20 m distance from the nest box. At a random time, after placing the snake under the camouflaged cloth and the parents had returned to normal activities, the string was pulled and the snake was revealed. Be- haviors of both male and female bluebirds were observed for 14 min. The foster female, FI, performed investigative and nest defense behaviors at nest box #4 by landing on the top of the nest box three times to inspect and po- tentially deter the potential nest predator. Dur- ing the same time, male M4 performed many defensive behaviors including chattering at the snake, wing flashing, and landing on top of the nest box to inspect and possibly deter the potential predator. The three nestlings fledged from nest box #4 on 6 May. The new pair (female FI and male M4) re-nested and initiated egg-laying in nest box #4 on 12 May; the pair fledged three young on 14 June. In the final nesting attempt at nest box #4, initiated on 27 June, three nest- lings survived until 13 days of age, at which time they died with no signs of trauma or dep- redation. Foster female FI returned to the study site in summer 2006 and nested suc- cessfully with another male on a territory ad- jacent to nest box #4. Male M4 was not ob- served at the study site in 2006. DISCUSSION We can identify three possible reasons for this adoption event by the female bluebird. First, the adopted nestlings may have been re- lated to the foster female in some way. In this case, the female, with her own young dead. would increase her inclusive fitness by adopt- ing related nestlings. This act is favored by selection, as the shared genetic material from the foster female would still be passed on in the related young. This is an unlikely scenario because, although breeding individuals do show some site fidelity from one season to the next (39-56% of breeding adults return to their previous year’s nesting site), nestling dis- persal from natal sites is much greater with only 11.9% of females and 14.6% of males returning to the natal study site (Gowaty and Plissner 1998). This suggests there is only a small chance the foster female was related to the original breeding pair (F4 and M4) or the nestlings she adopted. Inclusive fitness does not seem to be the reason for this adoption event. Conspecific brood parasitism occurs in Eastern Bluebirds where females lay eggs in nests of other bluebird pairs but provide no parental care to the young (Gowaty and Pliss- ner 1998). It is possible that female FI para- sitized the other breeding pair while female F4 was in the egg-laying cycle (Eastern Blue- birds usually lay 1 egg/day). By adopting the young at nest box #4, the foster female would be providing parental care to at least one of its own nestlings. Although possible, it seems unlikely that conspecific brood parasitism is the reason for the adoption event. The egg- laying cycle of female FI had no overlap with female F4 (female F4 began incubating 2 days before female FI began egg-laying), conspe- cific brood parasitism occurs infrequently in this species (1—2% of nestlings produced), and it has been suggested that females most likely to adopt the strategy of brood parasitism are females without their own nest boxes (Meek et al. 1994, Gowaty and Plissner 1998). The second reason for adoption, reproduc- tive error, is a likely scenario as most of the few cases of adoption in bluebirds occur when an individual has lost young of a similar age to the PNLO (Plissner and Gowaty 1988). Fe- male FI lost her nestlings, which would have been 6 days of age at the time of adoption, after the presumed desertion by her mate. The lone female joined with the lone male at nest box #4 and adopted the three 10-day-old nest- lings. This female was possibly responding to physiological and behavioral cues evoked by her own young, which as a byproduct induced SHORT COMMUNICATIONS 421 the adoption of PNLO. This can be concluded because the foster female had young at ap- proximately the same age as the adopted nest- lings, there were other available, unused box- es in which to nest, and there was potential for ‘floater’ males in the area. Anecdotal ev- idence of Eastern Bluebird females that re- place original breeding females suggests that replacement females will only adopt young when they are between 7 and 14 days of age (Plissner and Gowaty 1988, Gowaty and Pliss- ner 1998). Nestlings that are <7 days of age tend to be left to perish or are killed and ul- timately removed from the nest so the new pair can re-nest; nestlings that are >14 days of age are treated indifferently (Gowaty and Plissner 1998). The replacement female adopted PNLO that were 10 days of age, sug- gesting she could have misdirected parental behavior toward another’s offspring. The fos- ter female did not feed or defend the nestlings to the extent of the original male, but did feed an appropriate amount (not less than the mean female provisioning rate) for the number of nestlings. The male was feeding at a high rate, potentially the result of the perceived loss of the original mate. The other likely cause of this adoption is the lone female was demonstrating her paren- tal abilities to a new mate, and thereby gaining access to the male and a nest box. For adop- tion to be beneficial, the cost of the parental care to the female must be less than the benefit derived from the male genes and the quality of the breeding site. This is an adaptive cause for adoption, as the female would presumably be ensuring future reproduction with a fertile mate at a productive nest box by helping to care for PNLO. In areas where pairs have multiple broods, adoption may be a strategy that is favored early in the breeding season by individuals searching for a new mate (Rohwer 1986, Gori et al. 1996). Within-season mate replacement may be common in Eastern Blue- birds as only 35-60% of unsuccessful pairs and 70-85% of successful pairs will re-nest together in the same season (Pinkowski 1977). Bluebirds are socially monogamous in the breeding season suggesting that a lone bird may not have an available mate or a place to nest (Gowaty and Plissner 1998). There were several unused and available nest boxes in close proximity, but nest sites are historically limiting and this behavior may be a strategy to gain access to a potentially limited resource (Siefferman and Hill 2005b). Territory quality can vary widely and nest box #4 may have been on a particularly good territory. The fos- ter female may have fed and defended the PNLO early in the breeding season to dem- onstrate her abilities as a parent and gain ac- cess to a mate, nest box, and/or territory. There is evidence that male mate choice may occur in Eastern Bluebirds (Siefferman and Hill 2005a). This is a potential cause for the adoption, as the new male and female pair successfully re-nested, produced six eggs, and fledged three young. Many would assume that adoption is mal- adaptive, as it may simply be an act of displaced parental care. However, infanticide, which has been observed in many bird species including Eastern Bluebirds, is clearly not displaced pa- rental care and is likely an adaptive behavior. Rohwer (1986) argues that although mate re- placement may occur infrequently, it must have occurred often to select for infanticide, which suggests that mate replacement has been suffi- ciently frequent to eliminate adoption if it was maladaptive. Alternatively there may be strong selective pressure on parental care under normal contexts so selection against it in rare circum- stances (like adoption) is relatively weak (Plis- sner and Gowaty 1988). The described event is anecdotal and we il- lustrate the possible and reasonable interpre- tations of this adoption. We further suggest a mechanism for adoption to be adaptive in Eastern Bluebirds. It may be difficult to iden- tify directly if and how adoption is adaptive; however, it is possible that studies examining this behavior could focus on the apparent quality of nest boxes, territory, and mates in- volved. ACKNOWLEDGMENTS We thank J. R. Brzyski, L. M. Siefferman, D. F Westneat, and two anonymous reviewers for providing helpful comments and suggestions on the manuscript. We also thank Howard Lumber Company for donating lumber to construct nest boxes. LITERATURE CITED AvITAL, E., E. JaBI.ONKA, and M. l.ACtlMANN. 1998. Adopting adoption. Animal Behaviour 55:1431- 1459. Gori, D. E, R. Sifwf.kt, and J. Casf:i.i.h. 1996. Ac- 422 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 cepting unrelated broods helps replacement male Yellow-headed Blackbirds attract mates. Behav- ioral Ecology 7:49-54. Gowaty, R a. 1983a. Male parental care and apparent monogamy among Eastern Bluebirds (Sialia sial- is). American Naturalist 121:149-157. Gowaty, P. A. 1983b. Overlap of two broods of East- ern Bluebirds in the same nest and brood reduc- tion. Wilson Bulletin 95:148-150. Gowaty, P. A. and J. H. Plissner. 1998. Eastern Blue- bird (Sialia sialis). The birds of North America. Number 381. Hamilton Jr., W. J. 1943. Nesting of the Eastern Blue- bird. Auk 60:91-94. Kreig, D. C. 1971. The behavioral patterns of the East- ern Bluebird. New York State Museum and Sci- ence Service Bulletin Number 415. Meek, S. B., R. J. Robertson, and P. T. Boag. 1994. Extrapair paternity and intraspecific brood para- sitism in Eastern Bluebirds revealed by DNA fin- gerprinting. Auk 111:739-744. PiNKOWSKi, B. C. 1977. Breeding adaptations of the Eastern Bluebird. Condor 79:289—302. PiNKOWSKi, B. C. 1978. Two successive male Eastern Bluebirds tending the same nest. Auk 95:606-608. Plissner, J. H. and P. A. Gowaty. 1988. Evidence of reproductive error in adoption of nestling Eastern Bluebirds (Sialia sialis). Auk 105:575-578. Rohwer, S. 1986. Selection for adoption versus infan- ticide by replacement “mates” in birds. Current Ornithology 3:353-395. SiEFFERMAN, L. AND G. E. HiLL. 2005a. Evidence for sexual selection on structural plumage coloration in female Eastern Bluebirds (Sialia sialis). Evo- lution 59:1819-1828. SiEFFERMAN, L. AND G. E. HiLL. 2005b. UV-bluc struc- tural coloration and competition for nestboxes in male Eastern Bluebirds. Animal Behaviour 69:67- 72. The Wilson Journal of Ornithology 120(2):422-424, 2008 An Intraspecific Killing in Adult Pacific Reef Egrets (Egretta sacra) Christa Beckmann ABSTRACT. — Intraspecific killing without canni- balism is rare in birds. I report an observation of an adult Pacific Reef Egret (Egretta sacra) killing an adult conspecific at One Tree Island, Great Barrier Reef, Australia. The motivation and context for the killing were not apparent. To the best of my knowl- edge, this is the first report of intraspecific killing in Pacific Reef Egrets. Received 18 June 2007. Accepted 12 September 2007. Intraspecific killing in birds occurs occasion- ally and may or may not represent cannibalism. Whereas cannibalism of eggs and chicks is not uncommon in birds (Stanback and Koenig 1992), killing without cannibalism is rare. Kill- ing without cannibalism may result from terri- torial disputes (Flux and Flux 1992), siblicide (Fujioka 1985), and infanticide (Parkes 2005). In the rare situation of an adult killing a con- specific adult, there are indications that victims ' Department of Environmental Sciences, University of Technology, Sydney, P. O. Box 123, Broadway NSW 2007, Australia. ^ Current address: School of Biological Sciences, University of Sydney, NSW 2006, Sydney, Australia; e-mail: cbec6408@mail.usyd.edu.au may be in poor physical condition (Franklin’s Gull [Lams pipixcan], Brose 2006; Ruddy Turn- stone [Arenaria interpres]. Crossland 1995). These instances are not well described and why it occurs is uncertain. Most reports are based on anecdotal evidence (House Wren [Troglodytes aedon\, Belles-Isles and Pieman 1987; Smooth- billed Ani [Crotophaga ani\, Loflin 1982; Great Gray Owl [Strix nebulosa], Fisher 1975; Red- tailed Hawk [Buteo jamaicensis], Clevenger and Roest 1974; Tree Swallow [Tachycineta bicol- or], Lombardo 1986; Lesser Snow Goose [Chen caerulescens caerulescens]. Hick and Cooke 1988) rather than direct observations (North- western Crows [Corvus caurinus], Andersen 2004; House Sparrows [Passer domesticus], Grubbs 1977). This note describes the first ob- servation of an adult Pacific Reef Egret (Egretta sacra) killing an adult conspecific and suggests possible reasons why killing may occur. OBSERVATIONS I observed a dark morph Pacific Reef Egret attacking a white morph egret at One Tree Is- land, Great Barrier Reef, Australia (23° 27' N, 152° 3' E) on 9 February 2007 at 0746 hrs. The event occurred on the island between two SHORT COMMUNICATIONS 423 of the field station buildings on and beside a foot path. The area consists of coral rubble substrate and is sparsely vegetated with Pi- sonia (Pisonia grandis) trees, —1.5 m tall. The event was observed from both inside one of buildings from the window and when the birds were out of view from the window, I contin- ued observations from outside, standing close to the building. Observations were made ei- ther with or without 10 X 42 binoculars at distances of —15 m (from window) to 30 m (outside) from the birds. I observed the dark morph egret stabbing the white morph bird on and around the head with its bill while the white morph was sitting on its legs with its head down and covered in blood. A second dark egret was perched on the railing of a nearby building, watching. The aggressor took two steps back and then ran at the white egret, stabbing it in the head. At 0748 hrs the dark morph left and returned 2 min later. It stood with one foot on the back of the white egret, essentially pinning it to the ground while stabbing it repeatedly. It then stood on the back of the white egret with both feet while attacking the white egret around the head. Six minutes later the white egret tried to escape, running with head down and wings outstretched. The dark egret chased closely af- ter it, pulling on its wing feathers. The white egret moved — 10 m away and the dark egret then left at 0755 hrs. At 0800 hrs the two dark morph egrets returned. One bird attacked the white egret stabbing it twice in the head and both left again, 3 min later. At 0833 hrs, one dark egret returned and perched on the bal- cony of a nearby building. At 0836 hrs the dark egret flew to the ground and stabbed the white egret four times in the head. A second dark egret arrived 1 min later and stood near- by, observing. The white egret walked —4 m away and both dark egrets left at 0838 hrs. A dark egret was seen in the area at 0955 hrs, but was not observed attacking the white egret. In total, 52 min elapsed between when observations began (at which point the victim was already covered in blood), and when the last attack by the dark morph egret was ob- served. I checked the condition of the white egret at 1500 hrs. It was still alive, but was lying on the ground with eyes closed. It later died and I then examined the carcass. The bird was an adult and not in breeding condition. It was not in molt and had wounds on and about the head only, which was covered in blood. The bird appeared to have been in poor con- dition as the keel was totally exposed with no fat or muscle tissue. I assumed that only one dark morph egret was responsible for the at- tack(s). DISCUSSION This note documents an adult Pacific Reef Egret killing an adult conspecific and, to the best of my knowledge, is the first report of an adult Ardeidae species killing a conspecific adult. Cannibalism did not appear to be the motivation of this killing as the dark egret did not feed upon the white egret. Disputes over resources, (particularly food, offspring, mates, and territory) seem likely underlying reasons for intraspecific killing (without cannibalism) among adult birds (Flux and Flux 1992). It is possible the killing oc- curred over a dispute, perhaps over territory. The white egret did not show signs of being in breeding condition (i.e., no brood patch, cloacal protuberance, or breeding plumage) and was in poor body condition, and likely starving. The dark egret was not in breeding plumage, nor was the dark morph bird which observed only. Therefore, it is not likely the dark egret could have been defending breed- ing territory, eggs or chicks. Individuals are known to defend foraging territory; however, aggressive displays between birds are simple and brief (Marchant and Higgins 1990). Sub- missive birds either leave the area or adopt the withdrawn crouch display (sitting on their legs with head tucked in) and the aggressor then halts its attack (Marchant and Higgins 1990). The white morph, being in poor condition, may have been unable to leave the territory of the dark egret before being severely damaged. The dark egret may have continued the attack until the white egret was outside its territory. That the second dark egret was not attacked by the aggressor places doubt on the theory that it was defending a foraging territory. Lit- tle is known about the territorial behavior of this species (Marchant and Higgins 1990) and, as my observations began when the attack was underway, any conclusions on this issue are speculative. In two other reported observations of intra- specific killing, the victim was in poor body 424 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 2, June 2008 condition (Crossland 1995, Brose 2006). In both cases, there was more than one aggressor as well as other individuals watching. It is possible the motive for this attack was some- how related to the poor condition of the vic- tim. Future research into when and why adult intraspecific killing occurs should involve more observations of intra-specific aggres- sion, and could involve presentation of taxi- dermy models of conspecifics and documen- tation of aggressive behavior. ACKNOWLEDGMENTS I thank R A. Biro and two anonymous reviewers for helpful comments on earlier drafts of this manuscript. LITERATURE CITED Andersen, E. M. 2004. Intraspecific predation among Northwestern Crows. Wilson Bulletin 116:180- 181. Belles-Isles, J. C. and J. Picman. 1987. Suspected adult intraspecific killing by House Wrens. Wilson Bulletin 99:498. Brose, C. 2006. Lranklin’s Gulls kill wounded indi- vidual. Blue Jay 64:215. Clevenger, C. T. and A. I. Roest. 1974. Cannibalism in Red-tailed Hawk. Auk 91:639. Crossland, A. C. 1995. A probable case of intraspe- cific killing in turnstones {Arenaria interpres). Notonis 42:281-282. Lisher, B. M. 1975. Possible intraspecific killing by a Great Gray Owl. Canadian Field-Naturalist 89:71- 72. Flux, J. E. C. and M. M. Flux. 1992. Nature red in claw: how and why starlings kill each other. No- tornis 39:293-300. Fujioka, M. 1985. Sibling competition and siblicide in asynchronously-hatching broods of the Cattle Egret Bubulcus ibis. Animal Behaviour 33:1228- 1242. Grubbs, V. L. 1977. An intraspecific mortal attack. Wilson Bulletin 89:477. Hick, D. S. and F. Cooke. 1988. A possible case of intraspecific killing in the Lesser Snow Goose. Wilson Bulletin 100:665-666. Loflin, R. 1982. Ani male apparently killed by other Anis while attempting to parasitize nest. Auk 99: 787-788. Lombardo, M. P. 1986. A possible case of adult intra- specific killing in the Tree Swallow. Condor 88: 112. Marchant, S. and P. j. Higgins (Editors). 1990. East- ern Reef Egret. Pages 1002—1009 in Handbook of Australian, New Zealand and Antarctic birds. Vol- ume 1, Part B. Ratites to ducks, Oxford University Press, Melbourne, Australia. Parkes, M. L. 2005. Inter-nest infanticide in ardeids. Waterbirds 28:256-257. Stanback, M. T. and W. D. Koenig. 1992. Cannibal- ism in birds. Pages 277-298 in Cannibalism: ecol- ogy and evolution among diverse taxa (M. A. El- gar and B. J. Crespi, Editors). Oxford University Press, New York, USA. The Wilson Journal of Ornithology 120(2):425-431, 2008 Ornithological Literature Compiled by Mary Gustafson BIRDS OF THE ALEUTIAN ISLANDS, ALASKA. By Daniel D. Gibson and G. Vernon Byrd. Nuttall Ornithological Club and American Ornithologists’ Union, Series in Ornithology Number 1. 2007. ISBN: 978-0-943610-73-3. 351 pages, 18 color plates, 2 figures. $40.00 (cloth). — The Aleutian Islands, Alaska, covers some 1,800 km of longitude, bridging the gap between Nearctic and Palearctic faunas. Also spanning some 475 km of latitude, their location between 50-55° N — at the boundary between the North Pacific Ocean and Bering Sea — places them in the middle of an exceptionally dynamic zone of cyclogenesis, an area dominated for much of the year by “the Aleutian Low” re- sulting in a cool, wet, maritime climate fre- quented by major storms. At least two of these islands have coined themselves as the “The Birthplace of the Winds.” These winds are largely the cause of an archipelago-wide vege- tation type dominated by grasses, herbs, and dwarf shmbs; only in the most protected pockets are taller shmbs to be found. These remote islands are home to a fasci- nating blend of western and eastern birdlife, as well as a unique, localized “Beringian,” avifauna. They support staggering nesting populations of millions of alcids and other seabirds. The isolated location of much of the chain results in its not being on the principal migration routes of many species with the ex- ception of some waterfowl and shorebirds. The Aleutians are a fascinating area for study of vagrancy in both Asian and North Ameri- can species, including many landbirds. A large number of birds may be carried east from Japan and the Russian Far East to these islands by major storm systems that track east and northeast in both spring and fall from the Asian coast, across the Aleutian chain, and onward into the southern Bering Sea or Gulf of Alaska. Two authorities have studied the avifauna of the Aleutian Islands in greater depth than anyone else: Daniel D. Gibson and G. Vernon Byrd — Gibson primarily through his many years as collections manager at the University of Alaska Museum, and Byrd through his po- sition as supervisory wildlife biologist for the Alaska Maritime National Wildlife Refuge. For some 35 years they visited nearly every one of the Aleutian Islands, combining this intimate knowledge with a thorough review of the published literature and an exhaustive mu- seum search for archived specimens and pho- tographs. The result is an exceptionally thor- ough inventory of the avifauna of the Aleutian region, including all the islands of the chain as well as data from waters offshore. Following short introductory sections de- scribing the physiography, climate, habitats, and history of ornithological study in the re- gion, the lion’s share of the book is taken up with detailed individual species accounts which cover the 271 species (plus 28 addi- tional subspecies) recorded through 2005 and substantiated by specimens or photographs. (Those species documented only by written details are included in an appendix.) Each ac- count is organized by dividing the Aleutians in to three sub-groups: the Eastern Aleutians (Fox Islands and Islands of Four Mountains), Central Aleutians (Andreanof and Rat is- lands), and Western Aleutians (Buldir and Near islands). Within each of these groups, species’ occurrence is presented by season, and the information includes data on seasonal abundance, early and late dates, maximum counts, habitat, and, when relevant, estimated nesting population sizes. In addition, there are color photographs grouped in the center of the book which depict a few birds and the differ- ent Aleutian habitats. In the back of the book, there is brief analysis of the seasonal move- ment of birds within the island chain, and more thorough looks and helpful tabular com- parisons of habitat use, egg-laying dates, and early arrival and late departure dates. Also in- cluded is a gazetteer. There is discussion of the substantial changes to habitats and bird populations brought on by hunting, military activities, and the widespread introduction of mammalian predators including rats and fox- es, as well mention of the recent projects to 425 426 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 2, June 2008 eradicate some of these harmful species and the results to date. This is an exhaustive treatise. The depth of the ornithological record presented is excep- tional. Details concerning many of the unique and fascinating taxonomic problems and un- certainties encountered in this region at both the species and subspecies level are covered. This reviewer could not find anything of sub- stance included within the book’s covers with which to find fault. However, the chosen font and format of the accounts make them some- what hard to read and it is difficult to quickly scan an account to find the specific section desired. Hopefully, future volumes in this new Series in Ornithology will have a more user- friendly print and layout. In regards to the book’s content, I wish that the authors had in- cluded more. Yes, it is easy to stand on the sidelines and want writers to include yet more information in many of their books. In the case of Birds of the Aleutian Islands, Alaska it might seem particularly undeserving to ask for more given the amazing depth and breadth of information presented. Yet, as the volume describes itself, it is indeed an “inventory.” There is little discussion on theories of oc- currence such as vagrancy to the Aleutians and the possible effects of major meteorolog- ical events. Nor do more than a handful of the species accounts place records into a larger perspective comparing occurrences to those elsewhere in Alaska, East Asia, or similar lat- itude. Each account does include a brief, sin- gle-sentence description of the species’ over- all range and, for rare visitors from the Alaska mainland, mention is made of how close to the Aleutians they nest. An account of the questionable occurrence of Willow Ptarmigan (Lagopus lagopus) anywhere but on the east- ern-most island (Unimak) includes a fine dis- cussion of other records of Willow and Rock (L. muta) ptarmigan making major sea/ice crossings in both eastern Russia and western Alaska. But, such discussions are rare. Many accounts of other migrant and vagrant species would have profited by just brief mention of how these Aleutian records fit into the broader picture of avian occurrence in at least the Be- ring Sea and northernmost Pacific Ocean re- gions. For example, the single, exceptional Aleutian record for Sooty Tern {Sterna fus- cata) involved remains of an adult salvaged on Attu Island in September 1997, but the ac- count does not include mention of the previ- ous passage of extra-tropical typhoons (prob- ably the most likely mechanism of transport) or if there exist any other records of this spe- cies northeast of Japan. The account of the single Aleutian record of Purple Martin (Progne subis) would have benefited from mention of its status in Alaska. The single rec- ord of Wood Warbler (Phylloscopus sibilatrix) from October 1978, which established the first North American record of this Western Pale- arctic species, would have been placed in bet- ter perspective with mention of a subsequent (October 2004) record from the nearby Prib- ilof Islands and of there being multiple au- tumn records from Japan. These comments would help the reader in gaining a much better perspective on such occurrences. Overall, this is a truly monumental, scholarly work. Its thoroughness in treating the fascinat- ing avifauna of the Aleutian Islands will be in- valuable to researchers and to students of bird distribution throughout North America and the Pacific Rim. — PAUL LEHMAN, 11192 Porto- belo Drive, San Diego, CA 92124, USA; e-mail: lehman.paul@verizon.net ATLAS OF BIRD DISTRIBUTION IN NEW ZEALAND 1999-2004. By C. J. R. Robertson, P. Hyvonen, M. J. Fraser, and C. R. Pickard. The Ornithological Society of New Zealand, Nelson, New Zealand. 2007. 533 pages, 2,100+ maps. ISBN: 0958248656. $125.00 (cloth).— As a member of the Orni- thological Society of New Zealand, I received my copy of this impressive tome gratis; this is just as well, as it is a large (30.5 X 21.5 cm), heavy (2.5 kg) book with a substantial price tag. When I describe this book as im- pressive, the word is well-chosen. This is a quality product with clearly-printed maps, in- teresting commentary on each species, useful interpretation of the data, and extensive use of appendices which present data associated with the atlas effort. Atlas efforts began in New Zealand in 1969 when field observers were encouraged to pro- vide lists of species from 3,675 squares cov- ering the country (North, South, and Stewart islands), each square containing 9,144 m^ (10,000 yds2). This effort was summarized in ORNITHOLOGICAL LITERATURE 427 a provisional atlas published in 1978 which included data collected through 1976 (Bull et al. 1978). This small book was followed by a more substantial version published in 1985 (Bull et al. 1985) which covered the years 1969-1979 and incorporated the data included in the 1978 provisional atlas. The 1985 atlas contained 19,049 species lists from the 3,675 squares. The book, subject of this review, summarizes data contained in 31,817 species lists from 3,192 squares, now 10 km^ each, throughout North, South, and Stewart islands with the addition of the Chatham Islands. It is described in the Introduction (page 8) thus: “This bird distribution atlas demonstrates dra- matic and rapid change in bird distribution in all parts of the country. These data provide the challenge of questions that need urgent an- swers ...” A wide range of information is presented in the 533 pages in addition to the expected dis- tribution maps. The Introduction (8 pages) provides a good overview of the atlas process, in particular collection of field data and the technical aspects of the data processing sys- tems used. An important section of the Intro- duction discusses various problems inherent in the effort, such as nomenclature, treatment of rarities, and treatment of breeding records. These issues are clearly obvious targets for critics, and I believe these potential short-falls are countered, or at least explained, well in this section. The authors are cognizant throughout that there are well-known prob- lems with distributional atlases for which data are mostly collected by amateurs and espe- cially those, like this one, that are based on simple species lists. The basic goal of any at- las should be the simple presentation of dis- tribution data. Users of the data must temper their efforts by acknowledgment of the inher- ent short-falls. The Introduction concludes with an interesting series of potential uses for the atlas data. Prominent among these are the “significant observed changes in gross distri- bution . . . for 57% of the birds” included in this atlas and its 1985 predecessor (detailed in Appendix K). This result is the key reason that an atlas is undertaken; the authors note these data provide an important baseline for further study. Prominent in New Zealand are its fas- cinating endemic species; these are of primary concern in conservation, efforts conducted in often spectacular and well-publicized fashion in New Zealand (Black Robin [Petroica trav- ersi], Kakapo [Strigops habroptila], and Kiwi [Apteryx spp.]). Among endemics, increases in gross distribution were found for 15 spe- cies, with 25 reduced, and 26 without change. Native and introduced taxa generally had in- creases, some rapid, over their distributions in the 1985 atlas. It is important to remember that the data presented in the atlas do not al- low conclusions regarding numbers of indi- viduals present. There follows a relatively short (7 pages) discussion on aspects of analysis of the data; weaknesses are discussed, mostly inherent bi- ases resulting from variations between observ- ers in the process of data collection and the details provided, especially related to habitat and breeding behavior descriptions. This sec- tion provides useful information and caveats for future users of the data. The bulk of the atlas is the almost 300 pag- es of species distribution maps, followed by some 40 pages of maps depicting habitats re- ported by observers. A nice touch is inclusion with the atlas of a transparent plastic book- mark-sized overlay tool inscribed with the three differing-scale matrixes of grid squares used in the atlas. This tool is useful in helping the reader differentiate and identify squares in a locality of interest. Each species has a full page distribution map for North Island and South Island (Stew- art Island is included with South Island); sev- eral species occur on only one of the major islands. The maps also include distributions for the Chatham Islands, an addition since the 1985 publication. Each square where the spe- cies was found is marked with a red spot su- per-imposed on clearly-printed relief maps. An attempt to indicate relative abundance is made by using a range of four sizes of spots, the smallest indicating the species was found on <10% of the lists submitted for that square, and largest spot for over 60%. The introduction to the map section (page 19) re- ports potential inaccuracies in this method, noting its non-use for taxa for which there are fewer than 50 reports in all; interpretation of these variable-sized spots requires care, as the authors indicate: “This division of results pro- duces a probable over-emphasis for single- sheet reports for a square ...” At best, this 428 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 method provides only a rough indication of relative abundance. Each map page presents additional infor- mation of hve types with varying usefulness. Most significant is a brief (rarely more than 4 lines) commentary at the bottom of each page which discusses changes from the 1985 atlas, suggests reasons for these changes where known, includes taxonomic notes, and provides a useful set of literature citations for the spe- cies. I found these text entries to be the most valuable part of the overall effort. Taken to- gether with the maps, these entries provide the reader with an excellent overview of the cur- rent status of each species. Not so useful are reproductions on each map page of the distri- bution map for that species in the 1985 atlas. Presumably this was done to allow a quick comparison with the 1985 distribution but, un- fortunately, these reproductions are small and I found it difficult to discern changes in distri- bution except for a few large-scale expansions or contractions; the information provided in the text comments is far more useful. Leading off each map page are English, Latin, and Maori names for the species, its status, its 6-letter code (as compared to the 4-letter banding codes generally used in the United States), and the number and percentage of squares and list sheets where the species was observed. English and Latin names pre- sented the authors with a problem, as the most recent Checklist of New Zealand birds (Tur- bott 1990) is clearly dated, both taxonomically and with respect to the recently-controversial issue of standardization of English names in- ternationally. Although a new Checklist is im- minent, the authors chose to follow Turbott (1990) with changes or updates noted in the text on each page. I found these latter notes useful and informative. The topic of English names is hotly-debated, but I lean to use of the name traditional in the country where the species is endemic as a first choice. Generally this is the name used on the map pages. As with many names used by indigenous peoples, Maori names vary regionally. The authors in- dicate they have used a list provided to them by the Maori Language Commission, but have included additional names that have been used in the past in ornithological works. The de- scriptors used to describe status are simply- defined; Endemic (breeds only in New Zea- land territories). Native (breeds in New Zea- land territories and elsewhere). Migrant (“a reasonable number” migrates to New Zealand territories but does not breed). Straggler (or Vagrant; “not a regular migrant”, or few mi- grate to New Zealand territories, but do not breed), and Introduced (“introduced by Homo sapiens''). While these definitions could be tightened, I must say as a tour operator cater- ing to bird-listers, that the definition of “En- demic” as a species that breeds only in New Zealand greatly simplifies the determination of which species are New Zealand endemics, notably seabirds that breed on a few islands in New Zealand but have large oceanic distri- butions (i.e.. Black Petrel [Procellaria parkin- soni]. Cook’s Petrel [Pterodroma cookii]). It should be noted that most definitions of “En- demic” (such as: unrecorded away from the country in question) would not cover these two procellariids. In addition, each map page has four small maps, one for each season, purporting to show differences in seasonal distribution, and an- other small map depicting breeding records. Migration among land birds in New Zealand is assumed to be minimal with the exception of possible altitudinal movements; these maps confirm this knowledge, but I had difficulty discerning whether there were differences or not. To test my ability to discern these, I checked the maps for Shining Bronze and Pa- cific Long-tailed cuckoos {Chrysococcyx lu- cidus, Urodynamis taitensis), both of which essentially depart New Zealand during the austral winter. These maps show significant seasonal differences, although it is important to note the seasons are defined as 3-month pe- riods, a method that does not consider varia- tions in breeding phenology and migration timing between species. Shorebirds which breed in New Zealand are strongly migratory in many cases, and these movements are shown well by the seasonal maps, as are pres- ence and absence of seabirds that breed around New Zealand. I found the maps for migratory shorebirds visiting New Zealand less informative; for most there were few squares involved and those shown as occupied probably reflect the presence of suitable hab- itat (this information is useful, but is depicted in the habitat maps shown later in the atlas). The remaining data presented on each map ORNITHOLOGICAL LITERATURE 429 page, again as small maps, show the squares in which breeding was reported. Perhaps the least useful of the information presented, the authors state (page 38): “For most taxa the availability of breeding records from the sur- vey is very poor, or poorly distributed, and the maps record only what was reported.” Atlases with which I am most familiar are those from Iowa and Nebraska; these are usually de- scribed as “breeding bird” atlases, and put far greater emphasis on a hierarchy of firmness of evidence indicating breeding. The New Zea- land atlas does not purport to be a “breeding bird” atlas, but is entitled a “bird distribu- tion” atlas. I found the maps informative and easy to use. The maps are best for land birds, but are weak in depicting distribution of seabirds, oth- er than those which breed on islands near the coast of the main islands. The data are too few to be meaningful for other seabird species. It seems likely the distributions shown for sev- eral of these reflect the availability of off- shore boat trips rather than true distributions of the seabirds. Good examples are the distri- butions shown for Shy Albatross {Thalassar- che cauta steadi, T. c. salvini), both of which contain excellent representations of the route followed by the Cook Strait ferries. Following the distribution maps is a series of 40 pages of habitat maps. These show hab- itats reported by observers for surveyed squares and are subdivided to be useful indi- cators of habitat distribution. A notable ex- ample is the sub-division of “Native Forest” into Kauri, Podicarp, Tawa, Beech, Rata/Ka- mahi. Mixed Hardwoods, Manuka/Kanuka, and Shrubland. Maps I found somewhat nov- el, but informative, are sub-alpine scrub, scree/rock and herbfield, introduced pine, or- chard, and market garden/crops. Coastal hab- itat is subdivided into several categories, in- cluding estuary/mudflat, lagoon, harbor, and rocky coast, but some are so widely-distrib- uted and close to each other that it is difficult to obtain useful information from the maps. Finally, there are 7 maps using red dots of lour sizes to indicate the number of taxa re- ported in squares in “grouped habitats”. 1 lound these maps useful; the grouped habitats are native forest, exotic plantation, farmland, residential, wetland, and coastal. Most atlases are content to present distri- butional data in the form of maps with more or less textual commentary. This atlas, how- ever, does not stop there. To complete the main body of the book are two interesting sec- tions, Data Use Examples, and Biodiversity. Both provide interesting examples of use of the atlas data. The 20-page Data Use Example section has a series of maps that show squares covered by the most active observers; observer #2211 covered almost the entire North Island! Also in this section are interesting “Comparative” maps; one shows squares with grouped alpine habitat using an orange dot. New Zealand Rockwren (Xenicus gilviventris) in pink, and squares with both in green. The green dot squares indicate the limited distribution of Rockwren as well as their apparent absence from large areas of alpine habitat. I couldn’t find any pink dots; apparently there weren’t any (Rockwrens shouldn’t occur in non-alpine habitat), or perhaps the pink color was too similar to the orange to be easily discerned. There follow two articles reporting innovative uses of atlas data, which make for interesting reading as well. The final 31 pages of the main body of the atlas (except for References) cover Biodiver- sity. This section uses data in the species map section and modifies them using smoothing software. These maps allow easier visualiza- tion of combinations of data such as occur- rence by square and number of taxa per square. Thus, in areas where a number of squares are occupied by higher numbers of taxa, the color hue is darker; the hue lightens as these numbers decline. Most interesting among these maps are those showing distri- bution of endemic and introduced taxa, and a series showing distribution of endangered taxa ranging from most threatened to least threat- ened. The final map in this section (page 410) is a thought-provoking inversion of the smoothed map showing current distribution of extant endemic taxa; it is based on the premise that “the areas most affected by human influ- ences were those with the richest and most biodiverse environments.” This map suggests that before humans arrived, taxa were most numerous in the eastern South Island, partic- ularly southward, where a wider range of hab- itats exist. Unfortunately, at least 50% of these pre-human taxa are now extinct. 430 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 2, June 2008 The Reference section concludes the main body of the atlas; it is extensive and valuable. The only complaint I have is that it is subdi- vided (General References, Atlas References, Classified Summarized Notes, Taxon Refer- ences); this often leads to extra time locating a citation. The atlas presents 12 Appendices and an Index. Appendices A through E essentially summarize the data provided on field observer list cards, including monthly data and grouped habitat data. Appendix F is an interesting ranking by grouped habitat of species ob- served; thus New Zealand Pipit {Anthus no- vae seelandiae) was the most-often observed species in alpine habitats. Grey Gerygone {Gerygone igata) in Native Forest, European Starling (Sturnus vulgaris) in Farmland, and Common Chaffinch {Fringilla coelebs) in Ex- otic Plantations. Appendices G and I were of interest from a bird listing viewpoint; Appen- dix G lists distribution records not shown on a map, being the rarest of the species reported. They range from oddities such as Emu {Dro- maius novaehollandiae), through locally-dis- tributed rarities like South Georgia Diving-pe- trel {Pelecanoides georgicus), and vagrants such as Ruff {Philomachus pugnax) and Com- mon Greenshank {Tringa nebularia). Appen- dix I lists species unrecorded during the 1999-2004 atlas field surveys but, which were included in the 1985 atlas (e.g.. Hoary-headed Grebe [Poliocephalus poliocephalus]. South Island Bushwren [Xenicus longipes longipes], and two species of Wood-swallow \ArtamusY) or are now known to breed in New Zealand but were not reported in 1999-2004, mostly island endemics outside the geographic scope of this atlas. Appendix H is a list of observers and the numbers of list sheets, squares, and taxa reported by each; I discovered here that observer #2211 is Tadeusz Wnorowski. Ap- pendix J reproduces Instructions to Contribu- tors as presented on field cards. Appendix K is an important summary of observed changes in distribution between this atlas and the 1985 version; this Appendix summarizes the results of the 1999-2004 atlas survey. The 137 spe- cies included provide much food for thought and grist for further investigation. The final Appendix, L, lists total number of lists re- turned, and top squares for lists returned, number of taxa, and taxa per square. I recommend this book to anyone interested in the birds of New Zealand. Although it is large and expensive, bordering on being quite acceptable and impressive as a coffee-table exhibit, it has something for everyone, from the researcher to the dedicated bird-lister. It will no doubt be of most use as a spring-board for further study of the distributional changes revealed within, especially those hitherto un- suspected. Both positive and negative changes in distribution are significant; conservationists and land developers would do well to study this atlas with a view to consequences of their activities. — ROSS SILCOCK, R O. Box 57, Tabor, lA 51653, USA; e-mail: silcock@ rosssilcock.com LITERATURE CITED Bull, P. C., P. D. Gaze, and C. J. R. Robertson. 1978. Bird distribution in New Zealand, a Provisional Atlas. The Ornithological Society of New Zealand Inc., Wellington, New Zealand. Bull, P C., P. D. Gaze, and C. J. R. Robertson. 1985. The Atlas of bird distribution in New Zealand. The Ornithological Society of New Zealand Inc., Wellington, New Zealand. Turbott, E. G. 1990. Checklist of the birds of New Zealand and the Ross Dependency, Antarctica. Random Century, Auckland, New Zealand. RAPTORS OF EASTERN [WESTERN] NORTH AMERICA (2 volumes). By Brian K. Wheeler. Princeton University Press, Prince- ton, New Jersey, USA. 2003: 439 pages, 562 photographs, and 36 maps [2003: 544 pages, 630 photographs, and 54 maps]. ISBN 0-691- 13476-5 [ISBN 0-691-13477-2]. Price per volume $29.95 (paper).— We have come a long way in identifying raptors in the field since the first modem field guide, Roger Tory Peterson’s A Field Guide to the Birds, was published in 1934. Peterson’s guide illustrated 23 species of eastern birds of prey with 33 images on four plates and one line drawing. Wheeler’s “east- ern” volume of the currently reviewed two- volume work illustrates the same 23 species, plus three additional mainly western raptors, with 562 photographic images spread across 147 pages. Peterson’s book devoted 16, 11.5 X 18. 5 -cm pages of text and illustrations to raptors. Wheeler’s “eastern” volume devotes ORNITHOLOGICAL LITERATURE 431 439, 14.5 X 23.5-cm pages of text and illus- trations to the group. The difference is readily apparent when one compares a single species. Wheeler’s “eastern” volume spends 43 pages, 82 color photographs, and 5 maps on Red- tailed Hawk (Buteo jamaicensis). Peterson spent 25 lines of text and one black-and-white illustration on the same species. To say the two books are different is an understatement. Wheeler’s two volumes also differ signifi- cantly in approach from the highly acclaimed, 254-page Hawks in Flight by Pete Dunne, Da- vid Sibley, and Clay Sutton (1988), which, at the time, adopted a holistic, gestalt, or “gizz” attitude to identifying flying birds of prey. Ac- cording to Wheeler “Identifying raptors is much like assembling a puzzle: you do it piece by piece.” And piece by piece, subspe- cies by subspecies, and color morph by color morph, is, in fact, the approach that Brian Wheeler takes throughout this meticulously well-written, well-researched, and well-illus- trated field guide and handbook. For those who have yet to see the work, Wheeler’s two volumes most closely resemble Dick Fors- man’s similarly detailed. The Raptors of Eu- rope and the Middle East: A Handbook of Field Identification (1999), albeit with consid- erably more text. Wheeler begins his “eastern” volume with 34 pages of introductory material in seven chapters that cover anatomy, plumage, molt, age differences, perching, and flight behavior, together with some background on his pho- tography. The section includes five glossaries detailing more than 250 terms, including, for example, four types of albinism and four types of flight-profile dihedrals. The glossaries are helpful, if not necessary, adjuncts to the de- tailed yet telegraphic species accounts that follow. One of the greatest contributions of the work is its detailed, precise, and largely ac- curate range maps (36 in the “eastern” vol- ume and 54 in the “western”). Although they are likely to fall out-of-date quickly as the ranges of several species fluctuate overall, and as the migratory behavior of many species changes the demarcation lines of summer and winter ranges, Wheeler’s range maps will con- tinue to serve as a useful historic record of the summer, winter, and year-round ranges of North American raptors for some time. A sug- gestion is in order here: a web site with 5-year updates of range maps would provide a useful addition to the work. Wheeler’s photographs, which are wonder- ful individually, and are spectacular when grouped together in species portfolios, are in some ways the heart and soul of the work. My only suggestion is that the location at which each photograph was taken be included in the legend, should the book be reprinted. Wheeler’s attempts to identify the full range of sub-species differences in field identifica- tion are both laudable and exhausting. The devil, as they say, is in the details, and, much to his credit, Wheeler spends considerable time dealing with this “hellaciously” compli- cated aspect of raptor field identification. The true mark of any great field guide is that it serves as a useful reference for workers in the field: something that both the beginning and experienced hawk watcher will use again and again, not only to teach themselves, but also to teach others about important field marks and identification cues. Wheeler clearly succeeds in meeting this goal. Although the work is not the easy read of a Peterson ( 1 934) or a Dunne, Sibley, and Sutton (1988) and, although it is not a work that one is likely to take into the field regularly, together the two volumes are essential reference for bird-of- prey field identification and distribution that should be in the possession of all raptor afi- cionados, professional and amateur. Even though the two volumes overlap extensively in both content and photographs, the range maps and parts of the text dealing with each species’ areas of occurrence differ, and having both will be important to those interested in each species’ North American range. Finally, I must say that I cannot imagine anyone writing and successfully publishing more detailed descriptions of raptor field marks in a stand-alone field guide. What needs to oc- cur next is for someone to boil down the com- plexity Wheeler has identified, so that those of us without encyclopedic memories can identify North America’s birds of prey in the field . . . at least most of the time. — KEITH L. BILD- STEIN, Acopian Center for Conservation Learning, Hawk Mountain Sanctuary, 410 wSummer Valley Road, Orwigsburg, PA 17961, UvSA; e-mail: Bildstein@hawkmtn.org THE WILSON JOURNAL OF ORNITHOLOGY Editor CLAIT E. BRAUN 5572 North Ventana Vista Road Tucson, AZ 85750-7204 E-mail: TWILSONJO@comcast.net Editorial Board RICHARD C. BANKS KATHY G. BEAL JACK CLINTON EITNIEAR SARA J. OYLER-McCANCE Editorial NANCY J. K. BRAUN Assistant Review Editor MARY GUSTAESON Texas Parks and Wildlife Department 2800 South Bentsen Palm Drive Mission, TX 78572, USA E-mail: WilsonBookReview@aolcom GUIDELINES FOR AUTHORS Please consult the detailed “Guidelines for Authors” found on the Wilson Ornithological Society web site (http://www.wilsonsociety.org). All manuscript submissions and revisions should be sent to Clait E. Braun, Editor, The Wilson Journal of Ornithology, 5572 North Ventana Vista Road, Tucson, AZ 85750-7204, USA. The Wilson Journal of Ornithology office and fax telephone number are (520) 529-0365. The e-mail address is TWilsonJO@comcast.net. NOTICE OF CHANGE OF ADDRESS Notify the Society immediately if your address changes. Send your complete new address to Ornithological Societies of North America, 5400 Bosque Boulevard, Suite 680, Waco, TX 76710. The permanent mailing address of the Wilson Ornithological Society is: %The Museum of Zoology, The University of Michigan, Ann Arbor, MI 48109, Persons having business with any of the officers may address them at their various addresses given on the inside of the front cover, and all matters pertaining to the journal should be sent directly to the Editor. MEMBERSHIP INQUIRIES Membership inquiries should be sent to Timothy J. O'Connell, Department of Natural Resource Ecology and Management, Oklahoma State University, 205 Life Sciences West, Stillwater, OK 74078; e-mail: oconnet® okstate.edu THE JOSSELYN VAN TYNE MEMORIAL LIBRARY The Josselyn Van Tyne Memorial Library of the Wilson Ornithological Society, housed in the University of Michigan Museum of Zoology, was established in concurrence with the University of Michigan m 1930. Until 1947 the Library was maintained entirely by gifts and bequests of books, reprints, and ornithological magazines from members and friends of the Society. Two members have generously established a fund for the purchase of new books; members and friends are invited to maintain the fund by regular contribution. The fund will be administered by the Library Committee. Robert Payne, University of Michigan, is Chairman of the Committee. The Library currently receives over 200 periodicals as gifts and in exchange for The Wilson Journal of Orni- thology. For information on the Library and our holdings, see the Society’s web page at http:// www.wilsonsociety.org. With the usual exception of rare books, any item in the Library may be borrowed by members of the Society and will be sent prepaid (by the University of Michigan) to any address m the United States, its possessions, or Canada. Return postage is paid by the borrower. Inquiries and requests by borrowers, as well as gifts of books, pamphlets, reprints, and magazines, should be addressed to: Josselyn Van Tyne Memorial Library, Museum of Zoology, The University of Michigan, 1109 Geddes Avenue, Ann Arbor, MI 48109-1079, USA. Contributions to the New Book Fund should be sent to the Treasurer. This issue of The Wilson Journal of Ornithology was published on 28 May 2008. 378 384 390 393 395 398 401 404 407 409 413 416 419 422 425 Continued from outside back cover Gender identification of Caspian Terns using external morphology and discriminant function analysis Joshua T. Ackerman, John Y. Takekawa, Jill D. Bluso, Julie L. Yee, and Collin A. Eagles-Smith Effects of traffic noise on auditory surveys of urban White-winged Doves Jeffrey B. Breeden, Fidel Hernandez, Ralph L. Bingham, Nova J. Silvy, and Gary L. Waggerman Short Communications Specimen shrinkage in Cinnamon Teal Robert E. Wilson and Kevin G. McCracken Breeding range extension of the Coastal Plain Swamp Sparrow Bryan D. Watts, Michael D. Wilson, Fletcher M. Smith, Barton J. Paxton, and J. Bill Williams Polyandry and sex ratio in the Song Sparrow Michael H. Janssen, Peter Arcese, Mark S. Sloan, and Kelly J. Jewell Novel courtship behavior in the Little Greenbul {Andropadus virens) Alexander N. G. Kirschel A recording of a Type B song of the Yellow-throated Warbler Bailey D. McKay Substrate and vegetation selection by nesting Piping Plovers Jonathan B. Cohen, Elizabeth H. Wunker, and James D. Fraser Nest site selection by a male Black-capped Vireo Andrew J. Campomizzi, Shannon E Farrell, and Jerrod A. Butcher Nests of Black-throated Green Warblers in tree cavities Douglas C. Tozer Nests and fledglings of the Red-ruffed Fruitcrow {Pyroderus scutatus) Mercival R. Francisco, Paulo R. R. Oliveira Jr., and Vitor O. Lunardi Palila {Loxioides bailleui) fledgling fed by Hawai‘i Amakihi {Hemignathus virens) Chris Farmer, Bridget A. Frederick, Paul C. Banko, Robert M. Stephens, and Carter W Snow Adoption: adaptation or reproductive error in Eastern Bluebirds? Daniel P Wetzel and C. Ray Chandler An intraspecific killing in adult Pacific Reef Egrets {Egretta sacra) Christa Beckmann Ornithological Literature Compiled by Mary Gustafson The Wilson Journal of Ornithology (formerly The Wilson Bulletin) Volume 120, Number 2 CONTENTS June 2008 Major Articles 239 New insight to old hypotheses: Ruffed Grouse population cycles Guthrie S. Zimmerman, Rick R. Horton, Daniel R. Dessecker, and R. J. Gutierrez 248 Communal calling and prospecting by Black-headed Trogons {Trogon melanocephalus) Ghristina Riehl 256 Song variation in Buff-breasted Flycatchers {Empidonax fulvifrons) M. Ross Lein 268 Phylogenetic relationship and song differences between closely related Bush Warblers {Cettia seebohmi and C. diphone) Shoji Hamao, Maria J. S. Veluz, Takema Saitoh, and Isao Nishiumi 111 Autumn stopover near the Gulf of Honduras by Nearctic-Neotropic migrants Andrew B. Johnson and Kevin Winker 286 Numbers of migratory birds stopping over in New Orleans, Louisiana, USA in relation to weather Peter H. Yaukey and Shawn C. Powell 296 Mass changes of migratory landbirds during stopovers in a New York City park Chad L. Seewagen and Eric J. Slayton 304 Migration of Florida sub-adult Bald Eagles Elizabeth K Mojica, J. Michael Meyers, Brian A. Millsap, and Katherin L Haley 311 Wetland features that influence occupancy by the endangered Hawaiian Duck Kimberly J. Uyehara, Andrew Engilis Jr., and Bruce D. Dugger 320 Habitat features associated with Barrow’s Goldeneye breeding in eastern Canada Michel Robert, Bruno Drolet, and Jean-Pierre L. Savard 331 Distribution, abundance, and nest-site characteristics of Black Swifts in the southern Rocky Mountains of Colorado and New Mexico Richard G. Levad, Kim M. Potter, Christopher W Shultz, Carolyn Gunn, and Joseph G. Doerr 339 Nest reuse by Vermilion Flycatchers in Texas Kevin S. Ellison 345 Natural history and breeding biology of the Rusty-breasted Antpitta {Grallaricula ferrugineipectus) Alina M. Niklison, Juan I. Areta, Roman A. Ruggera, Karie L. Decker, Carlos Bosque, and Thomas E. Martin 353 Foraging ecology of parrots in a modified landscape: seasonal trends and introduced species Greg D. Matuzak, M. Bernadette Bezy, and Donald J. Brightsmith 366 The signal function of a melanin-based plumage ornament in Golden-winged Warblers Emily Anne McKinnon and Raleigh J. Robertson 371 Factors influencing fidelity of House Finches to a feeding station Andrew K Davis Continued on inside back cover iiOil- Wilson Journal of Ornithology Volume 120, Number 3, September 2008 MC2 library MAY 19 2009 harvard UNIVERSITY Published by the Wilson Ornithological Society THE WILSON ORNITHOLOGICAL SOCIETY FOUNDED 3 DECEMBER 1888 Named after ALEXANDER WILSON, the first American ornithologist. President— James D. Rising, Department of Zoology, University of Toronto, Toronto, ON MSS 3G5, Canada; e-mail; rising@zoo.utoronto.ca First Vice-President— E. Dale Kennedy, Biology Department, Albion College, Albion, MI 49224, USA; e-mail: dkennedy@albion.edu Second Vice-President-Robert C. Beason, USDA, Wildlife Services, 6100 Columbus Avenue, Sandusky, OH 44870, USA; e-mail: beason@netzero.com Editor— Clait E. Braun, 5572 North Ventana Vista Road, Tucson, AZ 85750, USA; e-mail: TWILSONJO@ comcast.net Secretary— John A. Smallwood, Department of Biology and Molecular Biology, Montclair State University, Montclair, NJ 07043, USA; e-mail: smallwoodj@montclair.edu Treasurer-MelindaM.Clark,52684HighlandDrive,SouthBend,IN46635,USA;e-mail:MClark@tcservices.biz Elected Council Members— Carla J. Dove, Greg H. Farley, and Mia R. Revels (terns expire 2009), Rebecca Holberton, Robert S. Mulvihill, and Timothy O’Connell (terms expire 2010); Jameson F. Chace, Sara R. Morris, and Margaret A. Voss (terms expire 2011). Membership dues per calendar year are: Active, $21.00; Student, $15.00; Family, $25.00; Sustaining, $30.00; Life memberships, $1,000.00 (payable in four installments). The Wilson Journal of Ornithology is sent to all members not in arrears for dues. THE WILSON JOURNAL OF ORNITHOLOGY (formerly The Wilson Bulletin) THE WILSON JOURNAL OF ORNITHOLOGY (ISSN 1559-4491) .s pubhshed ^uarteriy m MarcK September, and December by the Wilson Ornithological Society, 810 East 10th Street, Lawrence, KS suLription price, both in the United States and elsewhere, is $40.00 per year. Penodie^s postage paid at Lawrence, KS. POSTMASTER; Send address changes to OSNA, 5400 Bosque Boulevard, Suite 680, Waco, TX 76710. All articles and communications for publications should be addressed to the Editor^ Exchanges should be addressed to The Josselyn Van Tyne Memorial Library, Museum of Zoology, Ann Arbor, MI 48109, USA. Subscriptions changes of address, and claims for undelivered copies should be sent to OSNA, 5400 Bosque BoulLrd, Suite 680, Waco, TX 76710, USA. Phone: (254) 399-9636; e-mail: single copies are available for $12.00 each. Most back issues of the journal are available and may be ordered from OSNA^ SpecL^pric^ L quoted for quantity orders. All issues of the journal published before 2000 are accessible on a free web site^at the University of New Mexico library (http://elibrary.unm.edu/sora/). The site is frilly searchable, and fril reproductions of all papers (including illustrations) are available as either PDF or DjVu files. © Copyright 2008 by the Wilson Ornithological Society Printed by Allen Press, Inc., Lawrence, KS 66044, U.S.A. COVER; Wilson’s Phalaropes {Phalaropus tricolor). Illustration by Robin Corcoran. 0 This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). MC2 LIBRARY MAY 19 2009 HARVARD UNIVERSITY FRONTISPIECE. Analysis of microsatellite DNA suggests that substantial mixing of breeding populations of Swainson’s Warbler {Limnothlypis swainsonii) may occur in wintering areas in Mexico and Jamaica. Painting by Daniel E Lane. Wilson Journal of Ornithology Published by the Wilson Ornithological Society VOL. 120, NO. 3 September 2008 PAGES 433-666 The Wilson Journal of Ornithology 120(3):433^45, 2008 GENETIC STRUCTURE OF BREEDING AND WINTERING POPULATIONS OF SWAINSON’S WARBLER KEVIN WINKER‘-2-3 AND GARY R. GRAVES' ABSTRACT. — Swainson’s Warbler (Limnothlypis swainsonii) is a species of conservation concern because of its small wintering range in the Caribbean Basin, relatively low population densities, and habitat fragmentation in its core breeding range in the southeastern United States. We investigated microsatellite DNA variation among 1 1 breeding populations from eastern Texas to Virginia and two populations from wintering areas in Jamaica and Mexico. Analyses of six polymorphic loci indicated a moderate level of gene flow among breeding popu- lations, relatively small effective population sizes (<200 individuals in each sampled population), and subtle population variation. We detected no evidence of population bottlenecks in breeding or wintering populations. Bayesian assignment tests suggested that substantial mixing of breeding populations may occur in wintering areas. Genetic differences between the Mexican and Jamaican populations indicate they may be drawn from different subsets of breeding populations. Patterns of genetic variation among breeding and wintering populations suggest a network of local and regional conservation programs may be necessary to maintain genetic diversity in Swainson’s Warbler. Received 4 May 2007. Accepted 17 December 2007. Genetic differentiation in migratory species is often associated with migratory divides where populations that winter in different re- gions meet during the breeding season in par- apatric contact zones (Salomonsen 1955). Studies of Sylvia warblers (Sylviidae) in Eu- rope have demonstrated that migratory behav- iors can have a strong genetic basis and can evolve rapidly (Berthold and Querner 1981; Helbig 1991, 1996; Berthold 2003; Bearhop ' Department of Vertebrate Zoology, MRC-1 16, and Laboratory of Molecular Systematics, National Mu- seum of Natural History, Smithsonian Institution. P. O. Box 37012, Washington, D.C. 20013, USA. ^ Current address: University of Alaska Museum. 907 Yukon Drive, Fairbanks, AK 99775, USA. -^Corresponding author; e-mail: ffksw@uaf.edu et al. 2005). Experimental crosses between populations that winter in different regions produced offspring that exhibited intermediate degrees of migratory orientation and restless- ness. Helbig (1991) suggested that hybrids be- tween populations of Eurasian Blackcap {Syl- via atricapilla) from opposite sides of the mi- gratory divide would be selected against, be- cause they would probably attempt to migrate across the Mediterranean and end in the un- inhabitable expanses of the Sahara Desert. This example suggests a plausible mechanism for maintenance of genetic structure in breed- ing populations of migratory species that win- ter in different areas (allohiemy). SwainsoiTs Warbler {Linmothlypi.s .swain - .sonii) is a sparsely-distributed wood warbler 433 434 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 (Parulidae) that breeds in the unglaciated southeastern United States and winters in the Caribbean Basin (Meanley 1971; Brown and Dickson 1994; Graves 2001, 2002). Geo- graphic variation in plumage color is subtle and the species is currently regarded as mono- typic (Brown and Dickson 1994). Although locally common, Swainson’s Warbler has been ranked as one of the most vulnerable breeding songbirds in the southeastern United States because of habitat fragmentation, relatively low population densities, and a small disjunct wintering range (Morse 1989, Terborgh 1989, Hunter et al. 1993, Rappole 1995). Most con- temporary breeding populations occur in sec- ond-growth forest on alluvial soils in the Mis- sissippi Valley and on the coastal plain from eastern Texas to southeastern Virginia (Graves 1998, 2001, 2002). The nonbreeding distri- bution is poorly known (Brown and Dickson 1994, Graves 1996), but the primary winter- ing areas appear to be in Cuba (Kirkconnell et al. 1996), Jamaica (Graves 1996), and in humid forests of mainland Middle America from Veracruz to Tabasco in Mexico (Winker et al. 1992, Winker et al. 1999b) and in Belize (K. Winker, unpubl. data). Winker et al. (2000) compared allozyme variation at 26 loci among five breeding pop- ulations of Swainson’s Warbler. Allele fre- quencies at five loci indicated modest popu- lation structure, chiefly between samples in the Mississippi Valley from Arkansas and populations on the coastal plain from Louisi- ana to Virginia; patterns of genetic variation were inconsistent with an isolation-by-dis- tance model. It was hypothesized the observed population structure was due to genetic drift due to the absence of barriers to gene flow in the contemporary landscape. Alternatively, genetic structuring could be influenced by the warbler’s divided wintering range (Greater Antilles vs. mainland Middle America). If mi- gratory behavior is under genetic control in this noctumally migrating species, interbreed- ing between populations that winter on the mainland and those that winter in the Greater Antilles might produce offspring that end their migrations over open water in the Caribbean or in the Gulf of Mexico. We investigated allelic variation at six poly- morphic microsatellite loci in II breeding populations of Swainson’s Warbler sampled from eastern Texas to Virginia and two pop- ulations from wintering areas in Jamaica and Mexico. We had three principal objectives: (1) characterize microsatellite diversity within breeding and wintering populations, (2) ex- amine the evidence for mixing of breeding populations on wintering areas, and (3) con- trast microsatellite and allozyme variation in the subset of breeding populations studied by Winker et al. (2000). METHODS Data Collection.—Sv^ainson's Warblers are territorial in breeding (Meanley 1971) and wintering areas (Graves 1996); they are also monogamous, although polygyny may occur (Graves 1992), and pairing occurs in breeding areas. We obtained tissue samples from 205 territorial males from breeding locations be- tween 27 April and 31 May (1986—1996) in the southeastern United States (Fig. 1) under state and national permits: Sulphur River, Tex- as (TXl: n = 17); Sam Houston National For- est, Texas (TX2: n = 20); Atchafalaya River, Louisiana (LAI: n = 22); Tensas River, Lou- isiana (LA2: n = 10); Homochitto River, Mis- sissippi (MSI: n = 20); Little Sunflower Riv- er, Mississippi (MS2: n = 10); White and Mississippi rivers, Arkansas (AR: n = 20); Apalachicola River, Florida (FL: n = 24); Oc- mulgee River, Georgia (GA: n = 20); Cooper and Santee rivers. South Carolina (SC: n = 21); and the Great Dismal Swamp, Virginia, and Chowan River, North Carolina (VA: n = 21). These samples nearly span the latitudinal and longitudinal distribution of the core breeding range (Graves 2002). Individuals from five population samples (LAI, AR, FL, SC, and VA) were previously assayed in the allozyme study of Winker et al. (2000). Ge- netic samples from the wintering range were obtained in Veracruz {n = 25) and Tabasco {n = 1) in southern Mexico (MEX) and in the Blue Mountains of Jamaica (JAM: n = \9\ Fig. 1). Voucher specimens (except for Ja- maican samples, for which only blood was ob- tained) are housed in the National Museum of Natural History (Smithsonian Institution), Bell Museum of Natural History (University of Minnesota), and the Coleccion Nacional de Aves (Universidad Nacional Autonoma de Mexico). Genomic DNA was extracted from body Winker and Graves • GENETIC STRUCTURE OE SWAINSON’S WARBLER 435 FIG. 1. Sampling localities within the principal breeding and wintering ranges of Swainson’s Warbler (after Graves 2002). tissues or from small pieces of museum skins using a diatomaceous earth/guanidine thiocy- anate extraction protocol (Carter and Milton 1993) and diluted to 20 ng/jaL for amplifica- tion of microsatellite loci developed for this species by Winker et al. (1999a). Extractions and amplifications were conducted in a PCR- free laboratory. All amplification batches in- cluded negative controls. Individuals were randomized across extractions, amplifications, and gel runs. Fragments were generated using dye-labeled primers and visualized on an ABI 373A automated sequencer. Fragment size was measured using an internal size standard (350 TAMRA) and GeneScan software (both from Applied Biosystems Inc., Foster City, CA, USA). All electropherograms were ex- amined manually, allele scoring was done without reference to source population, and results were not sorted to population until completion of the study. Analyses. — Basic statistics of allelic fre- quencies, genetic distance measures between populations, allelic diversity, and expected and observed heterozygosities were calculated using the computer programs BIOSYS-1 and GDA version 1 .0 (Swofford and Selander 1981, Lewis and Zaykin 1999). Tests for Har- dy-Weinberg (H-W) equilibrium, allelic het- erogeneity, levels of population structure, and differences between population pairs were performed using GENEPOP version 3. Id (Raymond and Rousset 1995). Levels of pop- ulation structure are given using the B of Weir and Cockerham (1984), which is an unbiased estimator of traditional We used CER- I'AlilJi I. Allele IVeciiieneies for six inierosatellile loei among I I hreecling and two wintering populations of Swainson’s Warhier (// - 250). I’opiilation (/i) 436 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 ir-. ir, ir, (N Cl C: ir. Cl ic, ic, Cl — Cl yr, ic, ic, c- ic, Cl Cl ic, yr, yr, c- Tf — — ^ r-, ^ oc ^ ^ IC, C IC, DC DClC.vCaCDC'^.Cl — X Cl ^ DC X C ^ ^1 X >C IC. fC r-. Cl X X ic, yr. yr, yr, yr. Cl Cl Cl Cl Cl O ^ Cl O yr, yr, yr. yr. c c c Cl O Cl Cl C — yr. yr, yr, yr. yr. yr, yr. yr. yr. yr. yr. yr. yr, c. Cl C yr. yr, yr. yr. Cl Cl c C — yr. c- yr. yr, yr. Cl c ir. Cl yr. yr, yr, yr, c- Cl c ic. yr. ci — Cl Cl c, fC CS X c, c- Cl C: — Cl C Cl c c. X X d Cl O — Cl C C: yr. r-. c, X C C. X c. Cl c d X C' C — TABLE I. Continued. Winker and Graves • GENETIC STRUCTURE OF SWAINSON'S WARBLER 437 CM sC ^ ir; o rj CM (M oc ^ Tt (M X DC X vC nC ^,^,0 — 0^5^ Tj- X X (M sC ''t — O — ir, o C CM O O O vC CM yr, Tt 'C CM X O c^, r^, c^- X — ir, sC c^- o o (M O' in. m CM t^ IT-, CM r' X — • irv IT-, CM X — X CM — X vC CM X c^, C^ O' r<^. Tt ir-, r;;; !£< — C^, CM iri O' X c<~, — yr', CM O' ^ vC — O' vC — CM CN CM O c^, X c^, O' vC — (^- O — CN O ir, yr, ir-, ir-, ic', yr, CM CM iri CM CM CM CM CM — CM C iTi X vC c^, c<", sC CM X ^ ^ ir, yr', ir; IT; CM C CM CM — CM — O ir, iTi CM ur, CM sC c^, O r-- r<-, CM ir, CM CM o ir, yr'. yQ ^ IT', Tt r-- CM -rf — in c^, O — ^X — O;^^ in, CM CM CM — CM xO'Tj-xr^xxO' !C' ^ ~ ” 1C' i— cMmrtinvcr^xo' Three individuals could not be scored at one locus each and n = 2.‘5 at loci 14, 18, and 3. 438 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 VUS version 1.0 (Marshall et al. 1998) to in- fer frequencies of null alleles. We used anal- ysis of covariance (ANCOVA) to evaluate population variability in allelic diversity. The relationship between genetic and geographic distances among populations was assessed with Mantel and permutation tests using NTSYS-pc version 1.50 (Rohlf 1988). We ex- amined geographic subsets of the population matrix with Spearman rank correlation coef- ficients (rj. Pairwise comparisons of popula- tion parameters are not independent, because each population is subjected to multiple con- trasts. Therefore, we emphasize the relative strength of the correlative relationship (rj be- tween genetic and geographic distances for these populations rather than P values. We used STRUCTURE, version 2 (Prit- chard et al. 2000, Falush et al. 2003) to ex- amine how well the predefined populations corresponded to genetic groups (K). Individ- ual genotypes in this Bayesian clustering ap- proach are assigned to clusters with Hardy- Weinberg equilibrium and linkage equilibrium achieved within each cluster. A Markov Chain Monte Carlo approach was used to identify the number of clusters (K) that are most likely given the observed genotypes. STRUCTURE was used twice for each defined (1-1 1) after a bum-in of 10^ iterations, followed by an ad- ditional 10^ iterations on the full breeding area data set (11 populations). No prior informa- tion (e.g., on the population of origin of each individual) was used. Admixture and correla- tions models were used in which individuals can have mixed ancestry, and where the allele frequencies of closely related populations may be correlated. We further refined our analysis of genetic groups (K) using the approach of Evanno et al. (2005), which examines the sec- ond order rate of change of the log probability of data with respect to the number of clusters. These analyses were based on 20 independent STRUCTURE runs for each K under the same conditions of a 10^ bum-in, plus another 10^ iterations. A bootstrapped (100 replicate), neighbor-joining tree was developed using SEQBOOT, GENDIST, NEIGHBOR, and CONSENSE in the software package PHYLIP (Felsenstein 1993). We tested for population bottlenecks using the computer program M (Garza and William- son 2001). This test calculates the ratio (M) of total alleles to the range of allele sizes with- in populations. Reduced populations are ex- pected to have a smaller M ratio than popu- lations in mutation-drift equilibrium (Garza and Williamson 2001). Simulations estimating the probability of the observed M ratio used Garza and Williamson’s (2001) suggested val- ues for the proportion of one-step mutations (90%) and the average size of non-one-step mutations (Ag = 3.5). We used 10,000 repli- cates for simulations and set 0 (4N^[ju) to 1, 10, and 25, values corresponding to equilib- rium effective population sizes {NJ before a bottleneck of 500, 5,000, and 12,500, respec- tively, at a mutation rate (|jl) of 5 X lO""^ (Goldstein and Schlotterer 1999). Population sizes of Swainson’s Warbler are unknown, but values of 0 well above 1 seem reasonable giv- en our field experience. Estimates of 4A^jjl were obtained using MI- GRATE (Beerli and Felsenstein 2001), where is effective population size and p, is the microsatellite mutation rate. Our estimates of long-term effective population sizes {NJ were made using a mutation rate (p) of 5 X lO "^ (Goldstein and Schlotterer 1999). Estimates of levels of gene flow among breeding popula- tions were made using the rare alleles method (Slatkin 1985, Barton and Slatkin 1986, Slat- kin and Barton 1989) as implemented in GENEPOP version 3. Id (Raymond and Rous- set 1995) and assignment tests; the latter (Cor- nuet et al. 1999) enable a more direct estimate of gene flow by calculating the likelihood that an individual genotype originated from the population where it was obtained. These tests remove an individual from the sample and compare it with the remainder (Rannala and Mountain 1997, Comuet et al. 1999) to iden- tify the most likely breeding population of or- igin for wintering individuals. We used an a of 0.01, a stringency level shown to have high levels of accuracy in assigning populations of origin for dispersing individuals (Berry et al. 2004). We used the Dunn-Sidak correction (Sokal and Rohlf 1995) where appropriate to adjust probabilities for simultaneous statistical tests. RESULTS Variability of Microsatellite DNA. — The six microsatellite loci exhibited 5—16 alleles per locus (Table 1). Sixty-seven alleles were de- Winker and Graves • GENETIC STRUCTURE OE SWAINSON’S WARBLER 439 TABLE 2. Sample sizes {n), average allelic diver- sity (A), and expected and observed heterozygosities and HJ for six microsatellite loci in breeding and wintering populations of Swainson’s Warbler. Population n A He Ho Arkansas (AR) 20 6.17 0.676 0.667 Florida (FL) 24 6.83 0.724 0.660 Georgia (GA) 20 6.00 0.687 0.658 Louisiana 1 (LAI) 22 7.50 0.712 0.652 Louisiana 2 (LA2) 10 5.17 0.754 0.750 Mississippi 1 (MSI) 20 6.83 0.734 0.700 Mississippi 2 (MS2) 10 5.33 0.698 0.767 South Carolina (SC) 21 6.17 0.674 0.690 Texas 1 (TXl) 17 6.33 0.733 0.667 Texas 2 (TX2) 20 6.50 0.687 0.725 Virginia (VA) 21 5.83 0.667 0.690 Mexico (MEX) 26 6.50 0.705 0.620 Jamaica (JAM) 19 6.33 0.672 0.684 tected across all loci, ranging from a mini- mum of 32 detected in the MS2 population (n = 10 individuals) to a maximum of 45 in the LAI population {n = 22 individuals). Ten al- leles were unique to a single population (FL = 3; JAM = 2; LA2 = 1; MEX = 2; TX2 = 2). The number of alleles detected per popu- lation was correlated (r^ = 0.54; P < 0.005) with sample size. Populations from localities that drain into the Atlantic Ocean (VA, SC, GA) had significantly lower allelic diversity (ANCOVA, = 9.57; P = 0.015) than pop- ulations from drainages that empty into the Gulf of Mexico after factoring out the effects of sample size (FL, MSI, MS2, LAI, LA2, AR, TXl, TX2; Table 2). An overall hetero- zygote deficiency (P = 0.01 1) was caused by loci LmpL5B and Ls\v'\l\9 (Table 3). Hetero- zygote deficiency can be caused by a number of factors, including inbreeding, the Wahlund effect, and null alleles. Inferred frequencies for the latter possibility suggested that null al- leles were not responsible for the H-W dis- equilibrium (Table 3). Population Structure. — Breeding popula- tions had significant but weak structure (0 = 0.0068, P < 0.00005; Table 4) where 0 is the unbiased estimator for this was driven by heterogeneous distributions of alleles in loci L5H’|jl5B (P < 0.00005) and L.vh|jl18 (P = 0.039; Table 4). The Bayesian assessment of population clusters reflected this weak struc- ture with the highest probabilities of K occur- ring for 1-7 populations (Fig. 2A). Further TABLE 3. Exact tests for Hardy-Weinberg equi- librium (P-values; GENEPOP, Raymond and Rousset 1995) and inferred frequencies of null alleles (CER- VUS, Marshall et al. 1998). Locus p Null alleles L5H’|jl14 0.293 0.006 L5H’fX 18 0.185 0.016 L5H’|Jl3 0.429 0.011 L5H’ )jl5B 0.029 0.045 Lsw\v9 0.616 0.004 Lsw\i 1 9 0.035 0.019 testing to identify the true number of genetic clusters (K), performed separately for both breeding and wintering populations, suggested that 2-4 populations are involved (2-4 among breeding populations and 2-3 in the wintering samples), although the peak height of delta K in both sets of analyses suggested a lack of strong signal in the data set (Fig. 2B, C). A bootstrapped distance tree (10,000 replicates) provided little support for any breeding pop- ulation associations; just two populations formed one supported clade (FL and MSI supported by 60% of bootstrap replicates; not shown). Mantel and permutation tests (NTSYS-pc) of the pooled breeding population data set in- dicated that genetic distance was at best only weakly associated with geographic distance (Z = 0.27, P = 0.11; 10,000 random permuta- tions). However, when population contrasts were partitioned by geographic region, a clearer view of population differentiation TABLE 4. Allelic heterogeneity and levels of pop- ulation structure (6 and associated f*-values; Weir and Cockerham 1984) among 11 breeding and between two wintering populations of Swainson's Warbler (GENEPOP, Raymond and Rous.set 1995). Breeding Wintering Locus 0 pa B P' L,vvv:14 0.003 0.122 0.014 0.098 L.vvv: 1 8 0.007 0.039 -0.012 0.585 Ls\v:3 -0.009 0.924 -0.014 0.455 Ls\v:5B 0.034 0.000 -0.012 0.338 Lsu:9 0.012 0.192 0.144 0.004 Lsvv:\9 -0.001 0.387 -0.005 0.577 Overall 0.007 <0.00005 0.0 1 1 0.044 “ Adjusted rt for series of tests on individual l(K'i is 0.(K)8.^. An a of O.O.S for exact tests (bottom row) is not adjusted. 440 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 A B K FIG. 2. A. Probabilities, In P(D), of population samples coming from K genetic groups using Bayesian analyses of the full breeding area data set (STRUC- TURE Version 2, Pritchard et al. 2000; Falush et al. 2003). The set of 1 1 breeding populations oversampled the number of actual genetic clusters present, a value that ranges from 1 to 7. The second and third figures show delta K, the second order rate of change of FIG. 3. Rogers’ (1972) genetic distance and geo- graphic distance between breeding populations of Swainson’s Warbler; (circles) pairwise comparisons between localities that drain into the Gulf of Mexico (TXl, TX2, LAI, LA2, MSI, MS2, AR, FL); (gray triangles) pairwise comparisons between localities that drain into the Atlantic Ocean (GA, SC, VA); (black triangles) pairwise comparisons between Gulf and At- lantic drainage localities. emerged (Figs. 1, 3). Pairwise comparisons between Gulf populations {n = 28 contrasts) indicated a slightly negative relationship be- tween genetic and geographic distance (r^ = -0.39). The relationship between genetic and geographic distance was substantially weaker (r^ = -0.23) when pairwise comparison was limited to the seven populations from the Mis- sissippi and Sabine river drainages (n = 21 contrasts) of the western Gulf. In contrast, pairwise contrasts (n = 24) between popula- tions from Atlantic drainages with those from Gulf drainages exhibited a markedly positive correlation (r, = 0.49), suggesting a weak lon- gitudinal trend toward genetic isolation-by- distance. The most geographically isolated population in our survey (VA) also exhibited the greatest average pairwise genetic distance <— In P(D) in relation to K (Evanno et al. 2005), of breed- ing (B) and wintering (C) populations suggesting that 2-4 genetic clusters are most probable among the 1 1 sampled breeding populations, and that just 2—3 are most likely represented in the wintering populations (although peak heights and signal in these data sets are low). Winker and Graves • GENETIC STRUCTURE OE SWAINSON’S WARBLER 441 from other populations. The most centrally lo- cated population (FL) also showed the lowest average genetic distance between populations. Wintering populations exhibited significant but weak structure (0 = 0.0111, P = 0.044; Table 4). This was driven largely by locus Lsw\x9 (P = 0.004; other loci P > 0.098; Ta- ble 4). The loci responsible for breeding and wintering population structure were different (Table 4). Allozyme and Microsatellite Compari- sons.— We compared genetic distance matri- ces derived from allozymes and the microsat- ellite loci examined in this study for five breeding populations (LAI, AR, FL, SC, VA). The null hypothesis that the two matrices were not associated could not be rejected with Man- tel and permutations tests (Z = 0.18, P = 0.30; 10,000 random permutations) indicating the two nuclear genetic marker systems ex- amined for this subset of breeding populations exhibited discordant patterns of differentia- tion. Gene Flow. — The rare alleles method, an indirect technique to estimate the number of migrants per generation, inferred Nm = 9.16 among breeding populations after correction for sample size (where mean sample size was 18.6 and mean frequency of private alleles P[l] = 0.026). Assignment tests, which are more reflective of contemporary gene flow, in- dicated that three to seven individuals (15- 70%) from each breeding population had ge- notypes that were significantly unlikely to have originated there (averaging 4.4 individ- uals or 25% per population; Table 5). These values ranged from two to five individuals per population (averaging 3.2 individuals or 18% per population; Table 5) when adjusted for type-I error (Dunn-Sidak test). These numbers were similar in the two wintering populations (averaging 4.5 individuals per population, or 3.5 with the Dunn-Sidak correction; Table 5). The genotypes of three individuals from Flor- ida and single individuals from Mississippi (MSI), South Carolina, and Texas (TX2) could not be assigned to any breeding popu- lation (all with P < 0.()()()()5). Similarly, two individuals from Jamaica and one from Mex- ico could not be assigned to any breeding or wintering population. Both the rare alleles method and assignment tests suggested a TABLE 5. Assignment test results comparing ge- notypes of individuals with genotypic characteristics of the breeding populations from which they were sampled (GeneClass, Cornuet et al. 1999). Population n Mean Bayesian probability (± SDf Genotypes not from sampled population'’ Arkansas (AR) 20 0.37 ± 0.38 5 Elorida (EL) 24 0.34 -1- 0.30 5 Georgia (GA) 20 0.32 ± 0.31 3 Louisiana 1 (LAI) 22 0.28 -H 0.29 5 Louisiana 2 (LA2) 10 0.11 H- 0.20 7 Mississippi 1 (MSI) 20 0.29 + 0.29 4 Mississippi 2 (MS2) 10 0.16 -H 0.27 3 South Carolina (SC) 21 0.31 -+- 0.32 3 Texas 1 (TXl) 17 0.23 -+- 0.25 4 Texas 2 (TX2) 20 0.32 ± 0.28 5 Virginia (VA) 21 0.38 + 0.30 4 Mexico (MEX) 26 0.35 + 0.34 4 Jamaica (JAM) 19 0.32 -+- 0.29 5 ^ Averaged Bayesian posterior probabilities of membership among a pop- ulation’s individuals (Rannala and Mountain 1997, Cornuet et al. 1999). Number of individuals with probabilities of membership from popula- tion of origin with P < 0.01. moderate amount of gene flow among breed- ing populations {Nm = 3. 2-9. 2). Long-term Population Size. — We obtained estimates of 4A^|jl, where is effective pop- ulation size and |ul is the microsatellite muta- tion rate. These analyses suggested uniformly small long-term effective population sizes {N^) ranging from a low of 44 (LA2) to a high of 174 (LAI; Table 6). There was no evidence for population genetic bottlenecks in breeding or wintering populations at any of the mod- eled values of 0 (1, 10, and 25; P > 0.1), despite these rather low estimates of long- term effective population size. Evidence of Winter Mixing. — We compared the allelic frequencies of wintering popula- tions from Mexico and Jamaica with those of each of the 1 1 breeding populations to inves- tigate the genetic relationships between breed- ing and wintering populations. No genetic dif- ferences were found between any pairwise combination of wintering and breeding pop- ulation when Gi was adjusted for multiple tests (Table 7). Evidence of geographic segregation of breeding populations in wintering areas re- mained weak under a less conservative ap- proach (no adjustment of a). Allelic frequen- cies of the Mexican wintering population were significantly different from those of breeding populations from Arkansas (AR) in the Mis- 442 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 TABLE 6. Estimates of and the 95% confidence interval (following Beerli and Felsenstein 2001) where N, is effective population size and ijl is the mutation rate of microsatellite loci. Subsequent estimates of long-term effective population sizes (N, est.) and the corresponding 95% confidence intervals are based on a mutation rate of 5 X 10“^ (Goldstein and Schlotterer 1999) and are well below contemporary census sizes of these populations. Population Arkansas (AR) Florida (EL) Georgia (GA) Louisiana 1 (LAI) Louisiana 2 (LA2) Mississippi 1 (MSI) Mississippi 2 (MS2) South Carolina (SC) Texas 1 (TXl) Texas 2 (TX2) Virginia (VA) Mexico (MEX) Jamaica (JAM) 95% CF 0.18 (0.16-0.21) 0.24 (0.21-0.26) 0.17 (0.15-0.20) 0.35 (0.30-0.41) 0.09 (0.08-0.10) 0.16 (0.14-0.18) 0.09 (0.08-0.11) 0.22 (0.19-0.25) 0.14 (0.13-0.16) 0.20 (0.18-0.23) 0.16 (0.14-0.18) 0.34 (0.30-0.38) 0.17 (0.15-0.19) Ne est. 95% Cl 92 (81-105) 118 (105-132) 86 (75-98) 174 (150-204) 44 (38-51) 80 (71-91) 45 (39-53) 108 (96-123) 71 (63-81) 102 (89-117) 78 (69-88) 170 (152-192) 84 (74-96) a Based on multiple runs after initial, random parameter estimates were used to seed additional runs and convergence on similar values was verified. sissippi Valley and Virginia (VA) and South Carolina (SC) on the Atlantic coastal plain (Table 7). Similarly, allelic frequencies of the Jamaican wintering population were signifi- cantly different from those of breeding pop- ulations from Louisiana (LA2), Arkansas (AR), Texas 1 (TXl), and Virginia (VA; all P < 0.02; Table 7). We assigned wintering individuals to their most likely breeding population by choosing the highest probability P-value from the array of breeding populations. All but two individ- uals from Mexico and two from Jamaica could be assigned to one of the 1 1 breeding popu- lations (P < 0.01). Of the 24 assignable in- dividuals from Mexico, only Arkansas (AR) and Georgia (GA) were eliminated as likely breeding populations; and among the 16 as- signable individuals from Jamaica, only Lou- isiana (LA2), Mississippi (MS2), and Virginia (VA) were unlikely breeding sources. This suggests that individuals from a substantial number of geographically distinct breeding populations mix in wintering areas in Mexico and Jamaica. DISCUSSION The breeding distribution of Swainson’s Warbler has decreased significantly during the past century because of deforestation of bot- TABLE 7. Pairwise comparisons showing genetic similarities and differences between two wintering and 1 1 breeding populations of Swainson’s Warbler (0 and associated /"-values; GENEPOP). Mexico (MEX) Jamaica (JAM) Arkansas (AR) 0.0211 Florida (FL) -0.0046 Georgia (GA) 0.0004 Louisiana (LAI) 0.0006 Louisiana (LA2) -0.0082 Mississippi (MSI) -0.0122 Mississippi (MS2) -0.0195 South Carolina (SC) 0.0217 Texas (TXl) 0.0015 Texas (TX2) 0.0125 Virginia (VA) 0.0207 0.0126 0.0200 0.0107 0.5485 0.0081 0.1444 0.3073 0.0112 0.1641 0.1796 -0.0007 0.2652 0.3890 0.0221 0.0070 0.9319 0.0087 0.1468 0.9580 0.0163 0.0831 0.0115 0.0117 0.1059 0.0645 0.0201 0.0062 0.1380 0.0042 0.3184 0.0169 0.0173 0.0178 ^ Experiment error held at a = 0.05, and adjusted alpha for each series of tests - 0.0045. Winker and Graves • GENETIC STRUCTURE OE SWAINSON’S WARBLER 443 tomlands, flood mitigation projects, and for- estry practices that minimize early succession- al habitats (Twedt and Loesch 1999; Graves 2001, 2002). Several populations near the northern periphery of the warbler’s breeding range have disappeared in recent decades and many of the remaining populations are isolat- ed. We detected no evidence of genetic bot- tlenecks in breeding or wintering populations despite relatively small effective population sizes and fragmentation of the warbler’s geo- graphic range. However, the lower allelic di- versity exhibited among Atlantic-drainage populations and the significant correlation be- tween genetic and geographic distances ob- served in pairwise comparisons of Gulf-drain- age populations with those sampled in Atlan- tic drainages suggests that subtle population structure exists among breeding populations. Clustering tests further suggested that breed- ing populations sampled in this study are composed of two to four genetic populations. Whether differentiation is caused by genetic drift, stochastic sampling effects, selection as- sociated with the warbler’s disjunct wintering range, or other factors is unknown. The moderate levels of gene flow among breeding populations revealed by microsatel- lite data {Nm = 3. 2-9. 2) were consistent with the results of an earlier survey of allozymes (Winker et al. 2000) in a subset of five pop- ulations {Nm = 1.5-11.7). However, spatial patterns of genetic variation revealed by mi- crosatellites and allozymes were discordant, a not uncommon event between molecular marker systems (Allendorf and Seeb 2000). Significant population structure was caused by two microsatellite loci in the present study and three of 16 loci examined in our earlier allo- zyme study. Mixing of breeding populations in winter- ing areas is believed to be a common phenom- enon because breeding ranges of most Nearc- tic-Neotropic migratory songbirds are consid- erably larger than their wintering ranges (Ter- borgh 1989). The signal of geographic structure in our microsatellite data is weak, but Bayesian assignment tests imply that at least six breeding populations (TXl, TX2, LAI, MSI, FL, and SC) could have contrib- uted wintering individuals to both the Mexi- can and Jamaican wintering populations. However, genetic differences between the Mexican and Jamaican populations suggest they are comprised of different subsets of breeding populations. CONSERVATION IMPLICATIONS Should breeding populations of Swainson’s Warbler be managed as a single conservation unit? Patterns of microsatellite variation among breeding and wintering populations suggest that multi-regional and international efforts will likely be required to maintain the current level of genetic diversity in the spe- cies. Implementation of a viable management plan for the species throughout its global range, however, is contingent upon the devel- opment of additional genetic markers that can more fully resolve the genetic structure of breeding populations and their distributions during the nonbreeding season. Future field efforts should focus first on obtaining genetic samples from breeding populations in the Ap- palachian and Ozark mountains and from win- tering populations in Cuba and Belize. ACKNOWLEDGMENTS The Legacy Resource Management Program of the U.S. Department of Defense (DAMD17-93-J-3073, DACA87-94-H-0012) and the Research Opportunities Fund of the Smithsonian Institution supported this study. Numerous agencies kindly granted permits for the research; Jamaican Conservation and Development Trust; National Resources Conservation Authority (Kingston); USDA Forest Service; U.S. Fish and Wild- life Service; U.S. Department of Defense; Arkansas Game and Fish Commission; Florida Fish and Wildlife Conservation Commission; Georgia Department of Natural Resources; Louisiana Department of Wildlife and Fisheries; Mississippi Department of Wildlife, Fisheries, and Parks; North Carolina Wildlife Resourc- es Commission; South Carolina Department of Natural Resources; Texas Parks and Wildlife Department; and Virginia Department of Game and Inland Fisheries. R. T. Brumheld and T. C. Glenn helped with marker de- velopment. R. M. Zink permitted us to sample study skins in the Bell Museum of Natural History. Patricia Escalante P provided additional samples from Mexico. B. K. Schmidt, C. M. Milensky, J. P. Angle, and J. P. Dean assisted with specimen preparation. Francois Roussel provided helpful correspondence on making multiple comparisons. M. J. Braun, R. T. Brumfield, P. O. Lewis, and C. L. Pruett provided helpful comments on an earlier draft. LITERATURE CITED Ali.i:nd()RF, F W. and L. W. Sehb. 2()()(). Concordance of genetic divergence among sockeye salmon pop- 444 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 ulations at allozyme, nuclear DNA, and mitochon- drial DNA markers. Evolution 54:640-651. Barton, N. H. and M. Slatkin. 1986. A quasi-equi- librium theory of the distribution of rare alleles in a subdivided population. Heredity 56:409—415. Bearhop, S., W. Fiedler, R. W. Furness, S. C. Votier, S. Waldron, J. Newton, G. J. Bowen, P. Bert- HOLD, AND K. Farnsworth. 2005. Assortative mating as a mechanism for rapid evolution of a migratory divide. Science 310:502—504. Beerli, P. and j. Felsenstein. 2001. Maximum like- lihood estimation of a migration matrix and effec- tive population sizes in n subpopulations by using a coalescent approach. Proceedings of the Nation- al Academy of Sciences USA 98:4563—4568. Berry, O., M. D. Tocher, and S. D. Sarre. 2004. Can assignment tests measure dispersal? Molec- ular Ecology 13:551-561. Berthold, P. 2003. Genetic basis and evolutionary as- pects of bird migration. Advances in the Study of Behavior 33:175-229. Berthold, P. and U. Querner. 1981. Genetic basis of migratory behavior in European warblers. Science 212:77-79. Brown, R. E. and J. G. Dickson. 1994. Swainson’s Warbler {Limnothlypis swainsonii). The birds of North America. Number 126. Carter, M. J. and I. D. Milton. 1993. An inexpensive and simple method for DNA purification on silica particles. Nucleic Acids Research 21:1044. CORNUET, J.-M., S. PiRY, G. LUIKART, A. ESTOUP, AND M. SOLIGNAC. 1999. New methods employing multilocus genotypes to select or exclude popu- lations as origins of individuals. Genetics 153: 1989-2000. Evanno, G., S. Regnaut, and J. Goudet. 2005. De- tecting the number of clusters of individuals using the software STRUCTURE: a simulation study. Molecular Ecology 14:2611-2620. Falush, D., M. Stephens, and J. K. Pritchard. 2003. Inference of population structure using multilocus genotype data: linked loci and correlated allele frequencies. Genetics 164:1567—1587. Felsenstein, J. 1993. PHYLIP (Phylogeny Inference Package) Version 3.5c. Distributed by the author. Department of Genetics, University of Washing- ton, Seattle, USA. Garza, J. C. and E. G. Williamson. 2001. Detection of reduction in population size using data from microsatellite loci. Molecular Ecology 10:305- 318. Goldstein, D. B. and C. Schlotterer. 1999. Micro- satellites: evolution and applications. Oxford Uni- versity Press, Oxford, United Kingdom. Graves, G. R. 1992. A case of aggregated nest place- ment and probable polygyny in Swainson’s War- bler. Wilson Bulletin 104:370-373. Graves, G. R. 1996. Censusing wintering populations of Swainson’s Warbler: surveys in the Blue Moun- tains of Jamaica. Wilson Bulletin 108:94-103. Graves, G. R. 1998. Stereotyped foraging behavior of the Swainson’s Warbler. Journal of Field Orni- thology 69:121-127. Graves, G. R. 2001. Factors governing the distribution of Swainson’s Warbler along a hydrological gra- dient in Great Dismal Swamp. Auk 1 18:650-664. Graves, G. R. 2002. Habitat characteristics in the core breeding range of the Swainson’s Warbler. Wilson Bulletin 114:210-220. Helbig, a. j. 1991. SE- and SW-migrating Blackcap (Sylvia atricapilla) populations in central Europe: orientation of birds in the contact zone. Journal of Evolutionary Biology 4:657-670. Helbig, A. J. 1996. Genetic basis, mode of inheritance and evolutionary changes of migratory directions in Palearctic warblers (Aves: Sylviidae). Journal of Experimental Biology 199:49-55. Hunter, W. C., D. N. Pashley, and R. E. F. Escano. 1993. Neotropical migratory landbird species and their habitats of special concern with the southeast region. Pages 159-171 in Status and management of neotropical migratory birds (D. M. Finch and P. W. Stangel, Editors). USD A, Forest Service General Technical Report RM-229. Fort Collins, Colorado, USA. Kirkconnell, a., G. E. Wallace, and O. H. Garrido. 1996. Notes on the status and behavior of the Swainson’s Warbler in Cuba. Wilson Bulletin 108: 175-178. Lewis, P. O. and D. Zaykin. 1999. Genetic data anal- ysis: computer program for the analysis of allelic data, (www.eeb.uconn.edu/people/plewis/software. php). Marshall, T. C., J. Slate, L. E. B. Kruuk, and J. M. Pemberton. 1998. Statistical confidence for like- lihood-based paternity inference in natural popu- lations. Molecular Ecology 7:639-655. Meanley, B. 1971. Natural history of the Swainson’s Warbler. North America Fauna 69:1-90. Morse, D. H. 1989. American warblers: an ecological and behavioral perspective. Harvard University Press, Cambridge, Massachusetts, USA. Pritchard, J. K., M. Stephens, and P. J. Donnelly. 2000. Inference of population structure using mul- tilocus genotype data. Genetics 155:945-959. Rannala, B. and j. L. Mountain. 1997. Detecting immigration by using multilocus genotypes. Pro- ceedings of the National Academy of Sciences USA 94:9197-9201. Rappole, j. H. 1995. The ecology of migrant birds. Smithsonian Institution Press, Washington, D.C., USA. Raymond, M. and E Rousset. 1995. GENEPOP (Ver- sion 1.2): a population genetics software for exact tests and ecumenicism. Journal of Heredity 86: 248-249. Rogers, J. S. 1972. Measures of genetic similarity and genetic distance. Studies in Genetics 7:145-153. Rohlf, F. j. 1988. NTSYS-pc: numerical taxonomy and multivariate analysis system. Exeter Software, Setauket, New York, USA. Salomonsen, F. 1955. The evolutionary significance of Winker and Graves • GENETIC STRUCTURE OF SWAINSON’S WARBLER 445 bird-migration. Det Kongelige Danske Videnska- bernes Selskab 22:1-62. Slatkin, M. 1985. Rare alleles as indicators of gene flow. Evolution 39:53-65. Slatkin, M. and N. H. Barton. 1989. A comparison of three indirect methods for estimating average levels of gene flow. Evolution 43:1349-1368. SOKAL, R. R. AND E J. Rohlf. 1995. Biometry. Third Edition. W. H. Freeman and Company, New York, USA. SwoFFORD, D. L. AND R. B. Selander. 1981. BIO- SYS-1: a FORTRAN program for the comprehen- sive analysis of electrophoretic data in population genetics and systematics. Journal of Heredity 72: 281-283. Terborgh, J. 1989. Where have all the birds gone? Princeton University Press, Princeton, New Jer- sey, USA. Twedt, D. j. and C. R. Loesch. 1999. Forest area and distribution in the Mississippi alluvial valley: im- plications for breeding bird conservation. Journal of Biogeography 26:1215-1224. Weir, B. S. and C. C. Cockerham. 1984. Estimating F-statistics for the analysis of population struc- ture. Evolution 38:1358-1370. Winker, K., T. C. Glenn, and G. R. Graves. 1999a. Dinucleotide microsatellite loci in a migratory wood warbler (Parulidae: Limnothlypis swainson- ii), and amplification among other songbirds. Mo- lecular Ecology 8:1553-1556. Winker, K., G. R. Graves, and M. J. Braun. 2000. Genetic differentiation among populations of a migratory songbird: Limnothlypis swainsonii. Journal of Avian Biology 31:319-328. Winker, K., S. Arriaga Weiss, P. J. Lourdes Trejo, AND P. Escalante 1999b. Notes on the avifauna of Tabasco. Wilson Bulletin 1 1 1:229-235. Winker, K., R. J. Oehlenschlager, M. A. Ramos, R. M. Zink, J. H. Rappole, and D. W. Warner. 1992. Avian distribution and abundance records for the Sierra de Los Tuxtlas, Veracruz, Mexico. Wilson Bulletin 104:699-718. The Wilson Journal of Ornithology 1 20(3);446^54, 2008 DOES AGE INELUENCE TERRITORY SIZE, HABITAT SELECTION, AND REPRODUCTIVE SUCCESS OF MALE CANADA WARBLERS IN CENTRAL NEW HAMPSHIRE? LEONARD R. REITSMA,* MICHAEL T. HALLWORTH,i ^ AND PHRED M. BENHAM2 ABSTRACT— The Canada Warbler (Wilsonia canadensis) is currently in decline in the northeastern United States and basic demographic parameters remain to be described. We studied marked populations (76 ASYs, 14 SYs, and 2 of unknown age) of Canada Warblers on two study sites from 2003 to 2006. We mapped 92 territories (including males returning in multiple years) of 71 males using handheld GPS and ArcMap. We compared the pairing and fledging success of older and younger males on both sites, a red maple {Acer rubrum) swamp and a young forest intensively harvested in 1985 with -10% residual standing trees used by males as song perch trees. Both sites had a high proportion of ASYs (84% ASY for all territorial males, 77.5% of all males including non-territorial individuals). Both pairing (91%) and fledging (78%) success was comparatively high suggesting these two sites were of high value to this species. A higher proportion of SYs were transients. Pairing success was lower for younger males which established territories, but paired SYs fledged at least one young at a rate comparable to older males. This study corroborates the benefits of age and experience to reproductive perfor- mance. The results suggest that both red maple swamps and post-harvest forests with thick subcanopy vegetation and emergent trees provide high quality habitat for breeding Canada Warblers. Received 11 July 2007. Accepted 12 November 2007. Demographic studies are important in un- derstanding the population ecology of a bird species, especially during the breeding season. A bird’s age can greatly affect its ability to hold high quality territories, pair, and success- fully raise young (e.g., Ficken and Ficken 1966, Holmes et al. 1996). Older age and greater experience of males are often associ- ated with higher reproductive performance. This has been shown for a diversity of bird species from several different families. Mates with older male Great Tits {Pants major) laid earlier, fledged more young, and their fledg- lings survived at higher rates, although age of the male did not affect clutch size (Perrins and McCleery 1985). Age of the European Pied Flycatcher (Ficedula hypoleuca) was similarly correlated with laying date and fledgling sur- vivorship, and older males also had greater clutch sizes (Harvey et al. 1985). Nol and Smith (1987) found that older Song Sparrows {Melospiza melodia) began breeding earlier and raised more young to independence than younger Song Sparrows. ' Plymouth State University, 17 Highland Avenue, Plymouth, NH 03264, USA. 2 New England Institute for Landscape Ecology, 266 Prospect Hill Road, Canaan, NH 03741, USA. ^ Corresponding author; e-mail: leonr@plymouth. edu The effects of age on territory quality and reproductive success have been frequently studied among Parulidae. Habitat features and/or food availability affect individual bird territory quality, which in turn influences ter- ritory size, site fidelity, and reproductive suc- cess; older males disproportionately occupy higher quality habitat (e.g., Rodenhouse and Holmes 1992). The age of Ovenbirds (Seiurus aurocapilla) positively correlates with habitat quality and pairing success. Older males ac- quire higher quality habitats and pair at higher rates than younger males (Bayne and Hobson 2001). Holmes et al. (1996) found that older Black- throated blue Warblers (Dendroica ca- erulescens) occurred in areas with higher shrub density and fledged significantly more young. Older males also returned at higher rates than yearlings and had significantly smaller territory sizes. Older American Red- start {Setophaga ruticUla) males arrive earlier, and pair and fledge young at higher rates than first year males. Older males also relegate first year males to less optimal breeding habitat (Sherry and Holmes 1989). These findings in- dicate that habitat variables and age may in- fluence breeding success and return rates of warblers. Understanding the relationship between age and reproductive performance and how age 446 Reitsma et al. • AGE DIFFERENCES IN CANADA WARBLERS 447 Structure differs among habitats provides the needed tools for management, especially for species with documented declines. Habitat type is not a reliable predictor of reproductive success (Van Horne 1983, Vickery et al. 1992), but demographic patterns among dif- ferent habitats within a species’ distribution indicate age-related effects. Younger birds are often displaced into lower quality habitat where they are incapable of attracting a mate or raising young (e.g.. Holmes et al. 1996, Za- nette 2001). This is consistent with age-spe- cific models of habitat selection that suggest older birds assert dominance over younger birds and secure the best habitat and, conse- quently, have higher fitness (e.g.. Holmes et al. 1996, Hunt 1996). This study focused on the Canada Warbler (Wilsonia canadensis), a small (10-12 g) Neotropic-Nearctic migrant. The Canada War- bler breeds in areas with high shrub density, such as red maple {Acer rubrum) swamps and young upland forest (Titterington et al. 1979, Hagan et al. 1997, Golet et al. 2001). It occurs in both early-successional forest stands and within the deciduous (DeGraaf et al. 1998) and mixed-species understories of mature for- ests in New Hampshire during the breeding season. The species’ population is in a 2.7- 4.6% per year decline over the last four de- cades in the region that includes this study, as estimated from North American Breeding Bird Survey data (Sauer et al. 2005). The rea- sons for this decline are poorly understood as few studies have been performed at the level of the individual or population. Most studies involving the Canada Warbler have been con- ducted at the community level, involving cen- suses in differing habitats. This study com- pares age-related habitat selection and breed- ing success in both naturally occurring red maple swamps and young upland forest, clear- cut in 1985 with residual tree retention. The objeetives of this study were to ex- amine differences in the habitat characteris- tics, breeding success, and territory sizes be- tween after-second-year (ASY) and second- year (SY) males within the same breeding population. We predicted these measures will show consistent patterns similar to those doc- umented with other species. Older birds were predicted to have smaller territories, and high- er pairing and fledging success consistent with age-based dominance-hierarchical models. METHODS Study Site. — This study was conducted in Canaan, New Hampshire (43° 40' N, 72° 03' W) on the Canaan Town Forest (40 ha, here- after refen'ed to as the lower plot) and the ad- jacent Bear Pond Natural Area (BPNA) (363 ha, hereafter referred to as the upper plot) (Fig. 1). The lower plot is a 40-ha red maple swamp dominated by balsam fir {Abies bal- samea), red spruce {Picea rubens), and red maple with interspersed mixed upland forest. The upper plot (all within BPNA) is a 43-ha mixed deciduous upland forest, which was heavily harvested in 1985. Young regrowth dominated the harvest zone; remnant non- merchantable trees emerge from the regener- ating layer as scattered individuals and in clumps. Trees (>8.0 cm dbh) of the upper plot were dominated by red maple, balsam fir, east- ern hemlock {Tsuga canadensis), red spruce, and big-toothed aspen {Populus grandidenta- ta). Balsam fir dominated the smallest size class of shrubs and saplings (<2.5 cm dbh) on the upper plot followed by Ilex and Viburnum spp., white birch {Betula papyrifera), and red maple. Characterizing Male Territories. — Seventy- one territorial males were captured from 2003 to 2006 using playback of songs and calls. All birds were uniquely color-banded. Birds were assigned to age classes following plumage characteristics, i.e., molt limits of the outer primary eoverts, and boldness and extent of the necklace (Rappole 1983, Pyle 1997). Male territories were mapped from 27 May to 10 July each year. Every territorial male was ob- served for six 30-min periods. The bird’s lo- cation was recorded during each observation period with a Global Positioning System (GPS) unit at 5-min intervals. Every territorial male was observed for six observation bouts resulting in a minimum of 42 data points. All birds were followed between dawn and 0930 hrs EDT We used Adaptive Kernel (ADK) to construct territory boundaries (95% ADK) and the “core” territory areas (50% ADK) from each individual's location data (Barg el al. 2005). Territory boundaries and sizes were calculated using CALHOME home range analysis program (Kie et al. 1994). 448 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 FIG. 1. Canada Warbler territories during 2003-2006 in central New Hampshire used to compare reproduc- tive performance. After-second-year (ASY) male territories are shown in black and second-year (SY) male territories are shown in white. Each territory is the result of an adaptive kernel (ADK) analysis using GPS locations taken during 30-min observation bouts. Measuring Reproductive Success. — Pairing and fledging success were documented for each male within observation bouts. A male was considered paired if a female was repeat- edly detected within his territory. Males sing- ing from high perches throughout the breeding season with no female detected were consid- ered unpaired. A nest on a male’s territory was considered successful if at least one fledgling was seen (Hewlett and Stutchbury 2003), or if males and females concurrently delivered food to multiple areas within the territory. Vegetation Sampling. — Minimum convex polygons (MCP) were calculated using CAL- HOME (Kie et al. 1994) to position up to four vegetation sampling locations on each male’s territory. We used MCPs to be consistent with data collected across all years. Only males with the full set of six complete observation bouts were used in vegetation analyses in 2005 and 2006 {n = 39 male territories). Ninety-two male territories had a complete set of six observation bouts from 2003 to 2006 and were used for comparing reproductive success. All vegetation characteristics were quantified using a modification of the Breed- ing Biology Research and Monitoring [BBIRD] protocol (Martin and Conway 1994). Thirty-nine sets of vegetation plots, each set consisting of 4 subplots, were placed within territories. The position of subplots fol- lowed the BBIRD protocol except on territo- ries where subplots had to be shifted to either remain more than 50% within territory bound- aries or include actual locations of birds with- in their territories. If a subplot contained fewer than two observation points or was more than 50% outside of the polygon, it was shifted the least distance possible to meet the criteria. Thirty-two sets of non-territory plots were created using a random number generator and randomly placed in habitat that was not oc- cupied. We characterized a song perch tree as any Reitsma et al. • AGE DIFFERENCES IN CANADA WARBLERS 449 tree that emerged at least 3 m above the sur- rounding canopy with a radius of >5 m that was not overlapping with the surrounding can- opy. These perch sites were presumed to pro- vide high visibility and maximize the distance during singing at which a song may be heard by conspecifics. The density of the shrub fo- liage was quantified using a pole (2.5 m long and 2.5 cm diameter) marked at half-meter in- tervals. We recorded the plant species and number of times a branch or leaf touched the pole within each half-meter interval. The pole was placed 1 m from the center of the plot in each of the four cardinal directions. All four subplots for each territory or non-territory an- alog were averaged for analysis to best ap- proximate the habitat in the area sampled. Statistical Analyses. — Chi-square analyses were used to compare pairing and fledging success between older (ASY) and younger (SY) males. Fisher’s exact test was used to compare age classes of males in 2005 and 2006 on both plots. Wilcoxon signed rank analysis was used, due to non-normal distri- butions, to compare the territory and core area sizes of older and younger males that had a complete set of observation bouts (42 location points), and to analyze differences in habitat characteristics between the two age classes. Habitat comparisons were made by combining plots to maximize sample size within each age class (Bonferroni adjustments were made for habitat characteristics that were not consid- ered independent, shrub stems: P < 0.01, shrub foliage density: P < 0.01, tree stems: P < 0.0083). We also analyzed age class differ- ences in habitat characteristics within each plot and made the same Bonferroni adjust- ments for multiple comparisons of variables that were not considered independent. RESULTS Ninety-two territories of 71 males over 4 years (2()03-2()06) were mapped with a com- plete set of six 30-min observation bouts in- cluding 17 males mapped in two consecutive years and four males mapped in three consec- utive years. Ninety-three percent of territorial males (66 of 71) were banded. Territories of these marked individuals surrounded unhand- ed territorial males, making it possible to ac- curately map the five unmarked individuals. Of the 92 male territories mapped over 4 B > S_ D. n > s: +1 5 -u ■§ >- - ? 00 W .c -g +1 >- CO 2d, = < 1) u ^ V xz .t: y c A.SY 1.12 (0.238) 0.249 (0.065) 0.67 (0.140) 0.107 (0.017) 1.11(0.179) 0.152 (0.036) 1.18 (0.113) 0.232 (0.024) SY 1.16 One SY 0.150 No SYs in 2004 1.36 (0.165) 0.286 (0.037) 1.27 (0.227) 0.203 (0.041) 450 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 TABLE 2. Number (%) of each Canada Warbler age class on each of two plots for each year in west-central New Hampshire during 2003 to 2006. All males Territorial males >30 observations 2005 2006 2005 2006 Age Upper Lower Upper Lower Upper Lower Upper Lower ASY SY 11 (52) 14 (82) 24 (83) 10(48) 3(18) 5(17) 17 (77) 5 (23) 9 (64) 5 (36) 13 (87) 2 (13) 21 (81) 5 (19) 17 (85) 3 (15) years, 76 were defended by ASYs (83%), 14 by SYs (15%), and two by individuals of un- known age (2%). These latter two males were excluded from all analyses (Fig. 1). The 4-year average (± SE) territory size of ASYs (1.07 ± 0.75 ha) did not differ from SYs (1.3 ± 0.23 ha; Wilcoxon signed-rank test: Z = -1.19, P = 0.23) for the territories for which we had a complete set of six 30- min observation periods on known-aged birds {n = 90, Table 1). Similarly, the average ASY core area for the 4 years (0.19 ± 0.14 ha) did not differ from that of SYs (0.23 ± 0.19 ha; Wilcoxon signed-rank test: Z = -0.89, P = 0.37). The study area had a clear preponderance of older males (Table 2). A total of 108 male territories had >30 location observations. Of these 108 males, 91 were ASYs (84%) and 17 were SYs (16%). There were 16 additional territories mapped for which a complete set of six 30-min observation periods was not ob- tained, and 19 birds that were caught or mapped on the two sites, but not observed again. Of this comprehensive total of 127 birds, 27 were SYs (21.2%). Including these additional captured birds, the proportion of SYs at the red maple swamp site was 15.5% compared to 30% at the early-mid succession site. Age ratios in 2005 and 2006, the years with adequate sample sizes for comparisons, did not differ for all males considered terri- torial (>30 location observations, Fisher’s ex- act test: P = 0.215 in 2005, and P = 1.00 in 2006, Table 2). Reproductive Success. — ASYs paired (96%, 73 of 76) at higher rates than SYs (64%, 9 of 14, ^2 = 14 73^ df = 1, P < 0.001; Table 3) for all 4 years. However, the proportion of successfully paired males that fledged at least one young did not differ between the two age classes (ASYs = 76.7% [56 of 73], SYs - 67% [6 of 9], x" = 0.438, df = 1, P > 0.05). Habitat Characteristics. — Twenty-one hab- itat characteristics were measured. One of the 21 variables was significantly different be- tween the two age classes when combining both plots and adjusting for multiple compar- isons. ASY territories had a significantly low- er canopy height than those of SY birds (Wil- coxon signed-rank test: Z = —2.13, P = 0.033, Table 4). Three habitat characteristics were significantly different between age clas- ses within each plot. On the upper plot, ASYs had significantly more song perch trees per 0.04 ha (average ± SD for ASYs = 1.98 ± 0.12 vs. 1.41 ± 0.16 for SYs, Wilcoxon signed-rank test: Z = —2.26, P = 0.024). The total number of shrub stems per 0.00785 ha (5-m radius) was significantly greater on ASY territories (average for ASYs = 42.05 ±3.41 vs. 26.66 ± 4.05 for SYs, Wilcoxon signed- rank test: Z = —2.61, P = 0.009). On the lower plot, SYs had significantly more decid- uous shrub stems >2.5 cm dbh (5-m radius) than ASYs (average for ASYs = 2.29 ± 0.32 vs. 10.13 ± 3.40 for SYs, Wilcoxon signed- rank test: Z = -3.12, P = 0.002). DISCUSSION The high proportion of older males in our study area, together with the high overall pair- ing and fledging success of males over the 4 years for all males combined clearly indicates this is a productive area for the Canada War- bler. ASY males paired at significantly higher rates than SYs, which is consistent with other warbler species, such as the Black-throated Blue Warbler (Holmes et al. 1996) and Ov- enbird (Bayne and Hobson 2001). We found that ASY males were more reliably observed throughout the breeding season and a greater proportion were presumed to be territorial based upon observations of males for which we did not have sufficient data to do adaptive kernel analysis. Proportionally fewer SYs per- TABLE 3. Number {%) of older and younger male Canada Warblers successfully pairing and fledging at least one young in each year at two sites in central New Hampshire during 2003 to 2006. Reitsma et al. • AGE DIFFERENCES IN CANADA WARBLERS 451 W) ' — ' ■o 00 E o o (N |g ■o ^ ■ — . g ^ o o -o 00 OD E m lo — 00 (N (N in 8 "O 'vO tS TD Wj E —I 00 « s ip) 00 "O ro c£ - T3 W) “O — O' ^ i (N 0 *o a CL O' ^ ' — . O' >- < ^ >- < C/!) sisted throughout the breeding season. How- ever, SYs demonstrated equal success in fledging > 1 young if they were able to main- tain a territory and attract a female. Certain SYs were successful in the same years that some ASYs were not. This finding demon- strates impressive variation in individual fit- ness across age classes. We do not have ade- quate data at present to analyze whether suc- cess as an SY male influences the probability of success in subsequent years. Differences in pairing success could occur because ASY males may secure territories in habitats that females find more attractive or the males themselves may be more attractive as they are brighter, with bolder necklaces (Rappole 1983), which may signal greater ex- perience. Bold markings have been shown to increase reproductive success of Yellow War- blers {Dendroica petechia) (Yezerinac and Weatherhead 1997). There was no significant difference in size of territories (95% ADK) or core areas (50% ADK) between ASYs and SYs with complete sets of observations. This comparison was based upon combined samples from both plots. In a separate analysis, Hallworth et al. (2008) found that males of both age classes used 50% more area in the upper second- growth plot compared to the lower red maple swamp, but neither reproductive performance nor food abundance differed between the two plots. Territory size is known to vary inverse- ly with food supply in some species (e.g., Stenger 1958). Rodenhouse and Holmes (1992) found that food-rich habitat increases reproductive success for Black-throated Blue Warblers. Paired ASY and SY Canada War- blers did not differ in the rate of fledging at least one young in our study suggesting sim- ilar food resources existed in the territories of both age classes, but we did not measure food abundance on territories for each age class of males. Few habitat characteristics differed between ASY and SY territories. Shrub foliage densi- ties up to 1 111 on SY territories were higher than on ASY territories, but not significantly different. Shrub foliage density within this ho- rizon on ASY territories may influence nest site selection in that dense shrubs in the lower stratum may prevent the formation of sphag- num hummocks, a nesting substrate that was 452 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 TABLE 4. Habitat characteristics sampled on territories of both ASY and SY male Canada Warblers. All habitat values denote diameter at breast height (dbh), except values under shrub foliage density, where values denote heights. Means ± SE are shown along with Z and P values with significant differences indicated by an asterisk. Canopy height was the only significantly different vegetative feature. Bonferroni adjustments for shrub stem and foliage densities were both P = 0.01, and P = 0.0083 for tree stems. Characteristic ASY SY z p Canopy height, m 7.45 ± 0.26 8.91 ± 0.56 -2.130 0.033* Song posts, 0.04 ha 1.92 ± 0.10 2.06 ± 0.25 -0.555 0.579 Sum shrubs 45.20 ± 2.02 44.55 ± 3.10 -0.151 0.880 Deciduous shrubs, >1 m <2.5 cm 30.46 ± 2.98 22.17 ± 3.58 -0.331 0.741 >2.5 cm 6.20 ± 0.96 5.39 ± 1.48 -0.238 0.815 Coniferous shrubs, >1 m <2.5 cm 10.40 ± 1.09 8.03 ± 1.12 -0.506 0.613 >2.5 cm 4.19 ± 0.33 4.00 ± 0.61 -0.302 0.763 Shrub foliage density 0-0.5 m 0.88 ± 0.20 1.87 ± 0.48 -2.269 0.023 0.5-1 m 1.13 ± 0.29 1.73 ± 0.50 -2.057 0.040 1-1.5 m 1.10 ± 0.21 1.49 ± 0.33 -1.793 0.073 1.5-2 m 0.91 ± 0.13 1.45 ± 0.28 -1.793 0.073 2-2.5 m 0.94 ±0.16 1.71 ± 0.42 -1.698 0.090 Trees Small, 8-22.9 cm 13.30 ± 0.82 16.43 ± 2.08 -1.515 0.130 Medium, 23-37.9 cm 2.34 ± 0.34 2.64 ± 0.55 -0.813 0.416 Coniferous trees Small, 8-22.9 cm 6.69 ± 0.62 7.33 ± 1.28 -0.597 0.550 Medium, 23-37.9 cm 0.90 ±0.14 1.08 ± 0.25 -1.165 0.244 Deciduous trees Small, 8-22.9 cm 6.51 ± 0.57 9.05 ± 1.24 -1.916 0.055 Medium, 23-37.9 cm 1.32 ± 0.16 1.56 ± 0.51 -0.80 0.936 Snags Sum of all size classes 1.75 ± 0.22 2.69 ± 0.86 -0.929 0.353 Small, 8-22.9 cm 1.59 ± 0.22 2.48 ± 0.85 -0.831 0.406 Medium, 23-37.9 cm 0.16 ± 0.04 0.20 ± 0.09 -0.845 0.398 common if not preferred (Conway 1999, this study). Twelve of 18 nests found in 2005 and 2006 were within sphagnum hummocks, and were among the most cryptic of nest locations. In contrast, greater shrub cover on SY terri- tories may inhibit the ability of predators from locating nests (Bowman and Harris 1980). Average canopy height was signihcantly lower within ASY territories compared to SY territories. The lower canopy height on ASY territories for both plots may factor into fe- male assessment of prospective mates. We found that ASYs had more song perch trees and total shrub stem densities on the upper, second-growth plot in the single-plot habitat analyses between age classes. The greater number of emergent perch sites and higher to- tal shrub stem densities in this second-growth may combine to attract females more fre- quently to ASY territories, which is consistent with the higher pairing success of ASYs in our study. Larger shrubs (>2.5 cm dbh) were more abundant on SY territories on the lower, red maple swamp plot. ASYs occupied most of the dense deciduous thickets typical of red maple swamp on the lower plot. These dense thickets tended to have stems <2.5 cm dbh, whereas peripheral areas had larger-stemmed shrubs. The latter was more typical of SY ter- ritories on the lower plot. The habitat complex in which this study was conducted is presumed to be of high qual- Reitsrna et al. • AGE DIFFERENCES IN CANADA WARBLERS 453 ity given the high ASY: SY ratio (Hunt 1996) and reproductive success (Holmes et al. 1996) of territory-holders. More work needs to be done to compare ASY and SY reproductive performance in areas with different demo- graphic ratios assuming that within-population differences are likely to be more pronounced in more marginal habitat. ACKNOWLEDGMENTS This research was supported by the New England Institute for Landscape Ecology, Plymouth State Uni- versity, the Vermont Institute of Natural Science, Sweet Water Trust, The A. V. Stout Fund, and the Charles E. and Edna T Brundage Scientific and Wild- life Conservation Foundation. We thank Jameson Chace for introducing us to key methods used in this study. Clinton Parrish, Robin Jahne, Amy Ueland, Eric An- derson, Nicholas Dalzell, Keith Doran, Christopher Bordonaro, Marissa Goodnow, and Kyle Parent assist- ed in data collection. J. A. Jones and C. E. Braun made valuable comments on the manuscript. We thank the Town of Canaan and the Mascoma Watershed Conser- vation Council for permission to use the study sites. LITERATURE CITED Barg, J. J., J. Jones, and R. J. Robertson. 2005. De- scribing breeding territories of migratory passer- ines; suggestions for sampling, choice of estima- tor, and delineation of core areas. Journal of An- imal Ecology 74:139-149. Bayne, E. M. and K. A. Hobson. 2001. Effects of habitat fragmentation on pairing success of Ov- enbirds: importance of male age and floater be- havior. Auk 1 18:380-388. Bowman, G. B. and L. D. Harris. 1980. Effect of spatial heterogeneity on ground-nest depredation. Journal of Wildlife Management 44:806-813. Conway, C. J. 1999. Canada Warbler {Wilsonia can- adensis). The birds of North America. Number 421. DeGraae, R. M., j. B. Hestbeck, and M. Yamasaki. 1998. Associations between breeding bird abun- dance and stand structure in the White Mountains, New Hampshire and Maine. Forest Ecology and Management 103:217-233. Ficken, M. and R. W. Ficken. 1966. Notes on mate and habitat selection in the Yellow Warbler. Wil- son Bulletin 78:232-233. Goee:t, F. C., Y. Wang, J. S. Merrow, and W. R. DeRagon. 2001. Relationships between habitat and landscape features and the avian community of red maple swamps in southern Rhode Island. Wifson Bulletin I 13:217-227. Hagan, J. M., P. S. McKinley, A. L. Meehan, and S. L. Grove. 1997. Diversity and abundance of laiul- birds in a northeastern industrial forest. Journal of Wildlife Management 61:718-735. Hai.lworiii, M. T, P. M. Beniiam. J. D. Lambi;ri. and L. R. Reitsma. 2008. Canada Warbler {Wilsonia canadensis) breeding ecology in young forest stands compared to a red maple {Acer ruhriim) swamp. Forest Ecology and Management 255; 1353-1358. Harvey, P. H., M. J. Stenning, and B. Campbell. 1985. Individual variation in seasonal breeding success of Pied Flycatchers {Ficedida hypoleiica). Journal of Animal Ecology 54:391-398. Holmes, R. T, P. P. Marra, and T. W. Sherry. 1996. Habitat specific demography of breeding Black- throated Blue Warblers {Dendroica caerulescens): implications for population dynamics. Journal of Animal Ecology 65:183-195. Howlett, j. S. and B. j. M. Stutchbury. 2003. De- terminants of between-season site, territory, and mate fidelity in Hooded Warblers {Wilsonia citri- na). Auk 120:457-465. Hunt, P. D. 1996. Habitat selection by American Red- starts along a successional gradient in northern hardwoods forest: evaluation of habitat quality. Auk 1 13:875-888. Kie, j. G., j. a. Baldwin, and C. J. Evans. 1994. CALHOME Home range analysis program. USDA, Forest Service, Pacific Southwest Re- search Station, Fresno, California, USA. Martin, T. E. and C. J. Conway. 1994. Breeding Bird (BBIRD) field protocol. Montana Cooperative Wildlife Research Unit, University of Montana, Missoula, USA. Nol, E. and j. N. M. Smith. 1987. Effects of age and breeding experience on seasonal reproductive suc- cess in the Song Sparrow. Journal of Animal Ecol- ogy 56:301-313. Perrins, C. M. and R. H. McCleery. 1985. The effect of age and pair bond on the breeding success of Great Tits Pams major. Ibis 127:306-315. Pyle, P. 1997. Identification guide to North American birds. Part I. Columbidae to Ploceidae. Slate Creek Press, Bolinas, California, USA. Rappole, j. H. 1983. Analysis of plumage variation in the Canada Warbler. Journal of Field Ornithology 54:152-159. Rodenhouse, N. L. and R. T. Holmes. 1992. Results of experimental and natural food reductions for breeding Black-throated Blue Warblers. Ecology 73:357-372. Sauer, J. R., J. E. Hines, and J. Fallon. 2005. The North American Breeding Bird Survey, results and analysis 1966-2004. Version 2005.2. USGS, Patuxent Wildlife Re.search Center, Laurel, Mary- land, USA. www.mbr-pwrc.usgs.gov/bbs/bbs. html (acees.sed 12 October 2006). Shi;rry, T. W. and R. T. Holmes. 1989. Age-specific social dominance affects habitat use by breeding American Redstarts {Setopha(>a rnticilla): a re- moval experiment. Behavioral Ideology and Socio- biology 25:327-333. Sti;ngi;r, .1. 1958. Food habits and available food of Ovenbirds in relation to territory si/e. Auk 75: 333-346. 454 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 3, September 2008 Titterington, R. W., H. S. Crawford, and B. N. Bur- GASON. 1979. Songbird response to commercial clear-cutting in Maine spruce-fir forests. Journal of Wildlife Management 43:602-609. Van Horne, B. 1983. Density as a misleading indi- cator of habitat quality. Journal of Wildlife Man- agement 47:893-901. Vickery, R D., M. L. Hunter, and J. V. Wells. 1992. Is density an indicator of breeding success? Auk 109:706-710. Yezerinac, S. M. and P. J. Weatherhead. 1997. Ex- tra-pair mating, male plumage coloration and sex- ual selection in Yellow Warblers (Dendroica pe- techia). Proceedings of the Royal Society of Lon- don Series B 264:527-532. Zanette, L. 2001. Indicators of habitat quality and the reproductive output of a forest songbird in small and large fragments. Journal of Avian Biology 32: 38-46. The Wilson Journal of Ornithology 120(3);455-459, 2008 SOLITARY WINTER ROOSTING OF OVENBIRDS IN CORE FORAGING AREA DAVID R. BROWN>-2-3 AND THOMAS W. SHERRY' ABSTRACT. — We used radio-telemetry to locate night roosts of 54 Ovenbirds {Seiurus aurocapilla) wintering in a coastal second-growth scrub habitat in Jamaica. All Ovenbirds roosted within their daytime home range and most individuals roosted within the core of their diurnal home range. Sixty-six percent of individuals roosted within the 30% core use distribution (i.e., most heavily used portion of their home range) and 35% of individuals roosted within the 10% core. The average distance between roost sites for 29 individuals located on more than one night was 34 m and at least three birds roosted solitarily in the same location on different nights. Roost location relative to the 10% core area of diurnal home range did not differ between males and females, adults and immatures, or between individuals studied in subsequent years. The wintering Ovenbird population we studied appeared to roost solitarily. This study is the first to provide quantitative evidence that individual migrant songbirds in a tropical wintering population consistently roost at night within their foraging home range. These results suggest that roosting behavior is correlated with daytime space use patterns of the winter social system. Received 29 May 2007. Accepted 7 January 2008. Our knowledge of roosting behavior of tropical wintering migrant songbirds is limit- ed. The only quantitative studies of winter roosting are for species that roost in separate habitats from their daytime activity; Protho- notary Warbler (Protonotaria citrea) (War- kentin and Morton 1995), and Northern Wa- terthrush {Seiurus noveboracensis) (Reitsma et al. 2002, Burson et al. 2005). The literature is limited to descriptive examples of com- munal roosting of Northern Parula {Parula cimericana) (Staicer 1992), Cape May Warbler (Dendroica tigrina) (Staicer 1992, Baltz and Latta 1998), Prairie Warbler (D. discolor) (Staicer 1992), and Prothonotary Warbler. Morton (1980) and Staicer (1992) observed individuals going to roost and vocalizing at dawn on diurnal home ranges, and it is gen- erally assumed that many winter-territorial mi- grant songbirds roost solitarily at night on their diurnal home range. However, quantita- tive evidence of these patterns is lacking. Roosting behaviors have conservation impli- cations because multiple habitats may require protection for species that roost in habitats separate from their diurnal activities. We describe ( 1 ) the roosting behavior of ' Department of Ecology and Evolutionary Biology, Tulane University, New Orleans, LA 701 IS, USA. ^ Current address: Department of Biological Scienc- es, Eastern Kentucky University, Richmond, KY 40475, USA. '’Corresponding author; e-mail: david.brown5@ eku.edu wintering Ovenbirds {Seiurus aurocapilla) in Jamaica based on radio-tracking data, and (2) consider how roosting behavior is associated with other aspects of the winter social system. Specifically, we predicted that Ovenbirds would roost solitarily within their home range because most wintering Ovenbirds maintain long-term, minimally overlapping home range cores. METHODS We located roosts of Ovenbirds in a second growth scrub habitat at Luana Point, 8 km west of Black River, Jamaica (18° 02' N, 77° 55' W) during the dry season of February and March 2003-2005. The scrub habitat is rela- tively homogeneous and dominated by Log- wood {Haematoxylum campechianum), a dry deciduous introduced tree that ranged in height from 3 to 8 m (Strong 2000). The study area includes patches of black mangrove {Av- icennia germinans) and red mangrove {Rhi- zophora mangle) within 50—200 m of scrub study plots. Ovenbirds were abundant within the scrub habitat leaving little, if any, unoc- cupied space (Brown 2006). We mist-netted individuals, and banded and classified them to age (adult or immature). We ascertained gen- der using molecular markers or morphological measurements (Brown and Sherry 2006). Ra- dios weighing 0.6-0. 9 g (Holohil wSystems Ltd., Carp, ON, Canada) were attached with leg-loop harnesses (Rappole and Tipton 1991 ). We began searching for roosts 1 hr past 455 456 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 sunset. Roosts were located either by direct searches until we isolated the site or by tri- angulation with Program LOAS (Ecological Software Solutions 2002). We also collected locations (x = 39, SE = 1 ) of each bird during the day, and modeled home range character- istics using fixed-kernel use distributions with Arcview 3.2 (Hooge and Eichenlaub 1997). A use distribution is a probability density func- tion of the frequency of use of a given area within the home range based on the density of all locations (Kernohan et al. 2001). We report the area contained within a fixed cate- gory (e.g., 10%, 30%, 95%) of the use distri- bution volume. Smoothing parameters for fixed-kernel functions were selected by least squares cross validation. We calculated average home range size and neighbor overlap for 95% and 30% use dis- tributions. We considered the 30% use distri- bution to be the core area but, for some anal- yses, used the 10% core area to delineate the center of activity more precisely. We identi- fied where birds roosted in relation to their home ranges using the distance from the roost site to the edge of the 10% core area, (distance = zero for roosts within this core). We also classified which kernel contour, in increments of 10%, contained each roost location and used a histogram to compare this graphically to the total home range area within each 10% increment. We had only a single roost location for many birds. We considered individual birds rather than roost sites as replicates by random- ly selecting one location for analyses of roost proximity to home range core for the subset of individuals for which we had multiple lo- cations. We examined if birds tended to roost close to their home range core using a paired r-test that compared distances to home range cores from actual roost sites versus the aver- age distance of 10 random locations within each home range. Random locations were se- lected uniformly across the home range and not with respect to areas of increased use. We arbitrarily selected 10 as the sample size of random locations because we believed it would produce an adequate average distance. We tested for differences between males and females, and ages in the distance of roosts from home range cores using a two-way AN- OVA. Test assumptions were not violated and no data transformations were conducted. We used Systat Version 10 (Wilkinson 2000) for all analyses and report means ± SE. RESULTS Mean (± SE) home range area of 90 sed- entary (non-floating) birds was 0.69 ± 0.05 ha (area of 95% fixed-kernel use distribution). Mean 30% use distribution core area was 0.04 ± 0.004 ha. Neighboring home ranges over- lapped considerably (overlap area of 95% use distribution with each neighbor = 22%), but core areas overlapped little (overlapping area of 30% core with each neighbor = 3%). We located at least one night roost of 54 individuals and multiple roost locations for 29 of these birds. No roost was outside the home range (Fig. 1), but most birds (n — 26) roosted in different locations when tracked on multi- ple nights (average distance from previous roost site = 34 ± 4 m). The other three Ov- enbirds roosted close to the previous sites, i.e., within the same tree. Roost locations were closer than random locations to home range cores (ts4 = 9.34, P < 0.001). The distance from the 10% contour of the home range to actual roosts was 7 ± 1 m compared with 36 ± 3 m at random locations. The distance of roost locations from home range cores did not differ by gender (Ei 4g = 2.04, P = 0.16; n = 25 males, 25 females), age class (E, 52 = 0.14, P = 0.71; n = 30 adults; n = 24 immatures), or year (F251 1.13, F = 0.33; 2003: n = 10, 2004: n = 24, 2005: n = 20). Sixty-six per- cent of roosts were within the 30% core area of the use distribution, which encompassed only 7% of the total area of the home range (Fig. 2). Territorial birds, regardless of age or gender roosted centrally in relation to where they foraged. Birds were at least 2 m above ground for several cases {n — 1) where we identified the roost tree. DISCUSSION Ovenbirds roosted exclusively within their home range and most individuals roosted within their home range core. We did not di- rectly observe solitary roosting (i.e., it is pos- sible that more than one individual roosted in close proximity), but our results support soli- tary roosting because Ovenbird home range cores overlapped minimally. Thus, roost lo- cations also must have been separated. This is Brown and Sherry • WINTER ROOSTING OF OVENBIRDS 457 30% (shaded areas) and 95% (outline) fixed-kernel use distributions. Individuals located on multiple nights are indicated by multiple small numbers. Numerous non radio-marked Ovenbirds were also in the study area. 1.0 0.8 0.6 0.4 0 2 0 0 FIG. 2. Frequency of roost locations of Ovenbirds by kernel increment. Bars indicate the number ol indi- viduals located once each at night roost sites for fixed percentages ot each bird's use distribution. Core areas ( 10-30%) contain a small proportion of the total home range area (line). Each category represents the percentage of the total volume of the use distribution, but may represent a smaller proportion of the hcnne range area. For example, the 50% use distribution encompas.ses 50% of a bird’s activity, but only 16% of the total home range area. 458 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 the first quantitative description of a neotrop- ical migrant songbird roosting solitarily within the core of its foraging home range. We can- not rule out the possibility that Ovenbirds roost communally or outside their home range elsewhere, but the social behavior at our Ja- maican study site is typical of that in other habitats (Strong and Sherry 2000); we suspect the roosting pattern is typical as well. Few comparable data are available for other migratory species. Radio-tracked Hermit Thrushes (Catharus guttatus) wintering in the temperate zone also roosted solitarily within their diurnal home range (Brown et al. 2000). Many species of migratory songbirds occupy small home ranges in winter and it is likely that some of these also roost within their home range. Evidence of large-scale move- ments for most species at dusk is lacking and Morton’s (1980) observations are supportive. However, other patterns of roosting have also been documented. Wintering Bicknell’s Thrushes (C. bicknelli), which occupy exclu- sive territories during the day, were radio- tracked to “loosely communal” night roosts in a separate forest habitat from their diurnal activities, presumably because of predation threat from rats (Rattus spp.) in their preferred habitat (Rimmer et al. 2001:13). At least some individuals of Cape May Warbler, Prairie War- bler, and Northern Parula roosted communal- ly, and most individuals of these species gen- erally maintain solitary home ranges (Staicer 1992). Diurnal space use or roosting patterns for these three species has not been measured with telemetry, which is the best technique to quantify secretive space use patterns. Individ- ual Northern Waterthrush forage solitarily, shifting habitat use to track seasonal changes in food supply, and roost in structurally com- plex stands of red mangrove some distance from foraging locations (Reitsma et al. 2002, Burson et al. 2005). Wintering Prothonotary Warblers forage in mono- and interspecific flocks, and roost communally in mangrove habitat (Petit 1999). Movement to red man- grove roost sites by both Northern Water- thrush and Prothonotary Warblers may be an anti-predation behavior, because these man- grove stands are isolated from dry land. How- ever, movement to a distinctive roosting hab- itat also corresponds with the absence of sta- ble, long-term (over winter) home ranges. We hypothesize, based on these patterns, that win- ter roost behavior is associated with diurnal social systems, as exemplified by both solitary foraging and roosting by Ovenbirds. However, considerable variation in social behavior ex- ists both within and among species of winter- ing neotropical migrants, and additional re- search is needed to advance understanding of the causation of roosting behavior. We do not understand why Ovenbirds tend- ed to roost near their core diurnal foraging area and not randomly in their winter home range or in separate habitats. The relatively sedentary behavior of Ovenbirds would tend to minimize energetic costs associated with commuting. It is unlikely these costs would be important considering the small size of Ov- enbird home ranges and the proximity of our study site to red mangrove habitat (<50 m for some birds). Roosting behavior of tropical wintering birds is likely not subject to ther- moregulatory constraints. We suggest it is more likely that roosting behavior is associ- ated with social system or other ecological factors including predation. Roosting within core foraging areas would tend to space in- dividuals throughout the habitat, which could help protect against particular predators. Roosting behavior is important beyond our understanding of habitat use and social sys- tems. Conservation planning necessitates in- tegrating day with night activity locations for several species that roost apart from where they forage. Conservation planning is consid- erably simplified for species that roost within their daytime home range, as do these Oven- birds. ACKNOWLEDGMENTS We are indebted to field assistants Adam Anderson, Lynn Duda, Andrea Flowers, Lucas Forester, and Greg Levandoski. S. L. Burson and 2 anonymous reviewers provided constructive comments on the manuscript. Bird banding and radio-marking were conducted with approval of the U.S. Geological Survey and the Ja- maican National Environment and Planning Agency. This work was funded by a National Science Foun- dation (NSF) grant to TWS (DEB-0089541) and an NSF Doctoral Dissertation Improvement Grant to DRB and TWS (DEB-0408117). LITERATURE CITED Baltz, M. E. and S. C. Latta. 1998. Cape May War- bler {Dendroica tigrina). The birds of North America. Number 332. Brown and Sherry • WINTER ROOSTING OF OVENBIRDS 459 Brown, D. R. 2006. Food supply and the dry-season ecology of a tropical resident bird community and an over-wintering migrant bird species. Disserta- tion. Tulane University, New Orleans, Louisiana, USA. Brown, D. R. and T. W. Sherry. 2006. Food supply controls the body condition of a migrant bird win- tering in the tropics. Oecologia 149:22-32. Brown, D. R., R C. Stouffer, and C. M. Strong. 2000. Movement and territoriality of wintering Hermit Thrushes in southeastern Louisiana. Wil- son Bulletin 112:347—353. Burson, S. L., L. R. Reitsma, and P. D. Hunt. 2005. Conservation implications of multiple habitat use by Northern Waterthrushes during the nonbreed- ing season. Journal of Caribbean Ornithology 18: 72-76. Ecological Software Solutions. 2002. LOAS. Eco- logical Software Solutions LLC, Schwagalpstrasse 2, 9107, Urnasch, Switzerland. Hooge, P. and B. Eichenlaub. 1997. Animal move- ment extension 2.0. Alaska Science Center, Bio- logical Science Office, USDI, Geological Survey, Anchorage, Alaska, USA. Kernohan, B. J., R. a. Gitzen, and J. J. Millspaugh. 2001. Analysis of animal space use and move- ments. Pages 126-166 in Radio tracking and an- imal populations (J. J. Millspaugh and J. M. Mar- zluff. Editors). Academic Press, San Diego, Cali- fornia, USA. Morton, E. S. 1980. Adaptations to seasonal changes by migrant land birds in the Panama Canal Zone. Pages 437-453 in Migrant birds in the Neotropics: ecology, behavior, distribution, and conservation (A. Keast and E. S. Morton, Editors). Smithsonian Press, Washington, D.C., USA. Petit, L. J. 1999. Prothonotary Warbler {Protonotaria citrea). The birds of North America. Number 408. Rappole, j. H. and a. R. Tipton. 1991. New harness design for attachment of radio transmitters to small passerines. Journal of Field Ornithology 62: 335-337. Reitsma, L., P. Hunt, S. L. Burson, and B. B. Steele. 2002. Site fidelity and ephemeral habitat occupan- cy: Northern Waterthrush use of Puerto Rican black mangroves during the nonbreeding season. Wilson Bulletin 114:99-105. Rimmer, C. C., K. P. McFarland, W. G. Ellison, and J. E. Goetz. 2001. Bicknell’s Thrush (Catharus bicknelli). The birds of North America. Number 592. Staicer, C. a. 1992. Social behavior of the Northern Parula, Cape May Warbler, and Prairie Warbler wintering in second-growth forest in southwestern Puerto Rico. Pages 308-320 in Ecology and con- servation of neotropical migrant landbirds (J. M. Hagan III and D. W. Johnston, Editors). Smith- sonian Institution Press, Washington, D.C., USA. Strong, A. M. 2000. Divergent foraging strategies of two neotropical migrant warblers: implications for winter habitat use. Auk 1 17:381-392. Strong, A. M. and T. W. Sherry. 2000. Habitat-spe- cific effects of food abundance on the condition of Ovenbirds wintering in Jamaica. Journal of An- imal Ecology 69:883-895. Warkentin, I. G. AND E. S. Morton. 1995. Roosting behavior of Prothonotary Warblers in the non- breeding season. Wilson Bulletin 107:374—376. Wilkinson, L. 2000. Systat 10.0 for Windows. SPSS Inc., Chicago, Illinois, USA. The Wilson Journal of Ornithology 1 20(3):460^66, 2008 REPRODUCTIVE SUCCESS OE THE PUERTO RICAN VIREO IN A MONTANE HABITAT ADRIANNE G. TOSSAS' 2 ABSTRACT. — I studied the reproductive success of the Puerto Rican Vireo {Vireo latimeri) from 1998 to 2000 in Maricao State Forest, a montane reserve in the southwestern part of Puerto Rico. No parasitism by the Shiny Cowbird {Molothrus honariensis) was found but 63% of 38 active nests were depredated with an overall daily nest survival of 0.932 ± 0.007 (± SE). Mean nest survival estimates following Hurricane Georges did not vary signihcantly before and after the hurricane. However, a return rate of only 39% was estimated for color- marked adults the year including the hurricane compared to 72% for the year without a hurricane. A 26% decline in density of territorial males was observed in the post-hurricane year. Received 26 May 2007. Accepted 11 November 2007. The nesting suceess of a bird speeies can vary geographically as a result of differences in abundance of brood parasites, particularly if habitat characteristics favor the parasitic species (Ward and Smith 2000, Purcell 2006). These effects can be exacerbated if the host has not evolved nest defense strategies follow- ing recent colonization by the parasitic species (Rothstein 1990). This may be the case of the Puerto Rican Vireo (PRVI; Vireo latimeri), which has experienced sharp decreases in population size in Guanica State Forest (henceforth Guanica) on the dry coastal plain in southern Puerto Rico (Faaborg et al. 1997). However, the species is still commonly found throughout the island (Raffaele 1989). The population decline has been mainly associated with high rates (73-83%; Woodworth 1997) of brood parasitism by the Shiny Cowbird {Molothrus honariensis), a species first re- ported in Guanica by 1969 (Kepler and Kepler 1970), while expanding its range into the Ca- ribbean region from its native distribution in South America (Cruz et al. 1985). The PRVI population in Guanica from 1973 to 1996 showed an average annual decrease of 5% (Faaborg et al. 1997). Mean annual survival estimates for the PRVI in Guanica decreased from 0.68 in 1973-1990 (Faaborg and Arendt 1995) to 0.61 for a longer time period includ- ing data through 1996 (Faaborg et al. 1997). Little is known about the reproductive suc- ' Department of Biology, University of Puerto Rico, Rio Piedras, Puerto Rico 00931, USA. 2 Current address: Villas del Rio, 1 100 Bambu, Ma- yagiiez, Puerto Rico 00681, USA; e-mail: agtossas@gmail.com cess of the PRVI and the effects of cowbird parasitism in habitats other than Guanica. Par- asitism rates may differ within the island, par- ticularly since the Shiny Cowbird is rare or absent in some habitats. Cowbird parasitism of vireo nests has been found to be less inten- sive in highly forested montane habitats than in the more fragmented coastal habitats (Cruz et al. 1985, Perez-Rivera 1986). Perez-Rivera (1986) reported that 36% of PRVI nests and 35% of Black-whiskered Vireo {Vireo altilo- quus) nests in the central mountain region were parasitized. In contrast, cowbirds para- sitized 87% of the Black- whiskered Vireo nests in coastal forests (Cruz et al. 1985). The objectives of my study were to exam- ine if the Puerto Rican Vireo was affected by Shiny Cowbird parasitism in Maricao as has been reported in Guanica (Woodworth 1997, 1999; Woodworth et al. 1998, 1999), and to compare the breeding biology between the two populations. The incidental passage of Hurricane Georges through the Maricao study area in September 1998 provided an oppor- tunity to evaluate how reproductive parame- ters were affected by the natural event. I hy- pothesized that higher nesting success should occur at Maricao due to the scarcity of cow- birds at this site, compared to Guanica, and that parameters related to the breeding biology of the Puerto Rican Vireo should differ in the years before and after the passage of the hur- ricane. METHODS Study Area. — Maricao (18°09'N, 66° 59' W) is in the western part of the central moun- tain range of Puerto Rico, 16 km northwest of 460 Tossas • REPRODUCTIVE SUCCESS OF THE PUERTO RICAN VIREO 461 Guanica. It comprises 4,150 ha with eleva- tions ranging from 150 to 875 m (Silander et al. 1986). Annual rainfall and temperature from 1961 to 1990 averaged 232.6 cm and 21.7° C, respectively. The area contains sub- tropical moist forest, subtropical wet forest, and lower montane wet forest (Ewel and Whitmore 1973). Dominant tree species in- clude Micropholis chrysophylloides, Tere- braria resinosa, Linociera dominguensis, Homalium racemosum, Tabebuia schumanni- ana, and Eugenia stahlii (Silander et al. 1986). Maricao was struck by Hurricane Georges on 21-22 September 1998. This hurricane (category 3 on Saffir-Simpson scale of 5) damaged the forest structure with sustained winds of 184 kph and gusts of 240 kph (Ben- nett and Mojica 1998). The strong winds opened the canopy by severe defoliation, tree falls, and stem and branch breakage, all of which affected the resident avifauna (Tossas 2006). Nest Searching and Monitoring. — Thirty- six adult PRVIs were color-banded from 1998 to 2000 to facilitate behavioral observations. Individuals were lured to mist nets by playing the male’s song on a tape recorder near or within their territories. Gender, age, and breeding condition were assigned for captured vireos when possible. Twenty adult vireos were confidently separated by gender, 1 5 were males and five were females. All captured in- dividuals were marked with a USGS alumi- num band and a unique combination of two plastic color bands. At least one member of all pairs studied was color-banded. Vireo pairs were monitored from March to July each year from 1998 to 2000 along three trails with similar floristic composition and vegetation structure. I followed singing males looking for behavioral cues that could lead to their nests, or pairs carrying nesting material up to —50 m into the forest on each side of the trails surveyed. Nest contents were ex- amined every 2-5 days for signs of predation, parasitism or nest abandonment. Nests were checked directly or with a mirror fixed to a 6-m pole. Nests were checked when possible from a distance of —10 m to avoid disturbing breeding pairs, altering the nesting habitat, or increasing the risk of predation. 1 looked for Shiny Cowbird eggs or chicks in PR VI nests as signs of brood parasitism. Predation was assumed if eggs disappeared before their hatching date. Chicks were considered pre- dated if they disappeared from the area around the nest before the expected fledging date. In- cubation and nestling stages are 15 and 12 days, respectively (Woodworth 1997). The dates marking the initiation or end of each stage were assigned by direct observation or by back-dating or forward-dating from other observed events. Nest abandonment was re- corded if adults or chicks were not seen in the nest or in the vicinity during three consecutive visits of 30 min each. I noted if abandoned nests included eggs or were empty. Nest searching in 1999 was limited because of changes in forest structure related to Hur- ricane Georges and management practices that restricted access to trails. However, I moni- tored pairs throughout the breeding season and was able to ascertain their productivity even when I did not find the nest structures. Breeding season length was calculated for each study year from the day when the first pair started building a new nest until no active nests were found. Nest Success. — Daily nest survival rates and standard errors were generated following Mayfield (1975) and Johnson (1979). Esti- mates were calculated for two nesting inter- vals separately, from egg laying to incubation, and for the nestling period. An overall esti- mate was calculated for both intervals com- bined. Index of Annual Reproductive Success. — An estimate of annual reproductive success was calculated for all nests built by a pair moni- tored throughout the entire breeding season. However, only single nests were found for most pairs in 1998 (17/22) and 2()()0 (11/22), and only three nests were monitored in 1999. Thus, identifying whether a pair had success- fully raised chicks was based on observations of juveniles attended by color-marked adults at the end of the breeding season. This method was possible because juvenile PRVIs remain within their natal territories and are ted by adults for as long as 2 months after fledging (pers. obs.). Territories of color-marked individuals were visited every 2-5 days throughout the breed- ing season in search of nests, but searches for juveniles were more intensive from June to 462 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 July, the final weeks of the breeding season. I played the song of the PRVI in a tape recorder 1-2 times from the center of each territory for 5-min intervals and looked for both members of the pair during visits that lasted from 30 min to 1 hr. Adults usually approached the source of the intruder’s song and responded with alarm calls, providing an opportunity to observe accompanying juveniles. Each terri- tory was visited 10-15 times, and 20-22 pairs were monitored each year with at least one member of the pair being observed in every visit. I calculated the proportion of juveniles at- tended by adults as an index of annual repro- ductive success by the end of the breeding season. The index was the mean number of young fledged/pair/year. This method may un- derestimate reproductive success if chicks were depredated after leaving the nest, if their presence was not detected by being concealed and quiet, or if the brood was split between the pair. However, even with these biases, the index allows comparisons among study years. Shiny Cowbird Surveys. — I performed 60 point counts with playbacks of the Shiny Cowbird chatter call along the two trails where I did most nest searching. Counts were conducted in June 1998 (n = 30) and 1999 (n = 30) when PRVIs and most resident bird species were breeding. Fifteen point counts of unlimited radius, 100 m apart, were conducted in each trail. I played the call in the center of each plot for 5 min and waited 5 min for a cowbird aural or visual response. I also care- fully looked for indications of cowbird pres- ence during field visits throughout the project. Return Rates and Territory Mapping. — Marked PRVIs were followed over time to ob- tain estimates of their return rates (finite sur- vival rate described by Krebs 1989). The an- nual return rate was estimated as the number of marked adults present in their territories in year t that were recaptured or reobserved in their territories in year r + 1. The probability of relocating banded PRVIs using this method is high (Woodworth et al. 1999), although it does not account for individuals that may be alive but not in the sampling area. Thus, adults were assumed to have died or dispersed outside the search area if not observed the fol- lowing year. The study included 26 adult vir- eos color-banded in 1998 (n = 13) and 1999 {n = 13). Males and females were pooled for analysis due to small sample size and inability to differentiate gender of all individuals. Re- turn rates were calculated for two time peri- ods, a year including the hurricane event (1998-1999) and a year without a hurricane (1999-2000). Confidence intervals for return rates were calculated assuming binomial sam- pling. Spot maps were prepared for PRVIs along three transects to measure size and density of male territories. I assumed territorial bound- aries occurred where males or pairs engaged in aggressive encounters with neighbors and intruders (Bibby et al. 2000). The location of singing males, nests, and movements of indi- viduals also helped delineate territories. The mating status of males was assigned from ob- servations during repeated visits to the terri- tories throughout the breeding season. Un- paired males were excluded from the total density of territorial males (Wenny et al. 1993). All results are presented as mean (± SD), except for nest survival rates (x ± SE). RESULTS Nest Success. — I found 38 active PRVI nests in Maricao during the breeding seasons of 1998 to 2000. Most were found during nest building or early in the incubation period. All cases of nesting failure (24 of 38 nests) were attributed to predation, resulting in loss of the entire clutch or brood. Average clutch size was 2.03 ±0.16 eggs (range = 2-3, n — 38). Mean nest height was 4.9 ± 1 .29 m (range = 2.4-7. 0 m, n — 19). Breeding seasons lasted 90, 107, and 88 days in 1998 (26 Apr to 24 Jul), 1999 (23 Mar to 15 Jul), and 2000 (14 Apr to 10 Jul), respectively. Nesting activity was not observed after 15 July of any year. None of the PRVI nests was parasitized by the Shiny Cowbird. Cowbirds were not de- tected on any of the point counts in the forest interior with or without playbacks. I only had incidental observations of individual cowbirds outside of the study sites, along edges or in disturbed areas such as around the forest head- quarters and near the communications anten- nae and vacation center. I observed single cowbirds on two occasions following Greater Antillean Orioles {Icterus dominicensis) at the forest edge and twice observed single cow- birds flying high over the forest canopy. Tossas • REPRODUCTIVE SUCCESS OE THE PUERTO RICAN VIREO 463 TABLE 1. Nest survival rates (x ± SE) of the Puerto Rican Vireo in Maricao State Eorest, Puerto Rico, 1998-2000. Daily survival rates^ Egg laying and Overall (egg laying Year Nests Observation days incubation Nestling through nestling) 1998 20 411 0.981 ± 0.008 0.953 ± 0.017 0.935 ± 0.029 1999 2 27 2000 16 246 0.951 ± 0.017 0.976 ± 0.017 0.928 ± 0.013 Totals 38 684 0.968 ± 0.008 0.963 ± 0.012 0.932 ± 0.007 a Incubation and nestling stages consist of 15 and 12 days, respectively (Woodworth 1997). Analysis not possible due to small sample size, data added to total. Fourteen nests were lost during 440 nest- days of incubation and nine nests were lost during 244 days of the nestling stage (Table 1). The probabilities (x ± SE) that a nest would survive 15 days of incubation and 12 days of the nestling stage were 0.62 ± 0.008 and 0.64 ± 0.012, respectively. The probabil- ity of survival from incubation to fledging was 40%. Daily survival rates did not differ be- tween the incubation and nestling periods (^2^ = 4.08, P = 0.25) or between study years (X^ = 0.05, P = 0.83). Annual Reproductive Success. — Puerto Ri- can Vireo pairs in Maricao renested as many as three times in a single season after losses to predation, but most pairs (17/22, based on 1998 data) had a single nest. The percent of successful pairs ranged from 36 in 1998 to 65 in 1999, and was 55% in 2000. None of the pairs had two successful clutches in one sea- son. A total of 40 fledglings was produced by 33 successful pairs {n = 64 total pairs) during the 3 years. Successful pairs produced an av- erage (± SD) of 1.21 ± 0.55 fledglings/year, ranging from 1.15 ± 0.55 fledglings/pair in 1999 to 1.25 ± 0.62 fledglings/pair in 2000. I was able to follow 16 PRVI pairs for more than one breeding season. Most pairs (81%) were successful in at least 1 of the 3 years. Four of the nine pairs I observed during 1998- 2000 were successful once, two were success- ful twice, and three pairs failed to produce fledglings in any year. None of the pairs was successful in all 3 years. Seven pairs were fol- lowed in two breeding seasons. Three suc- cessfully reared young in both years, while the rest were successful in 1 year. Fifty-one per- cent of all pairs (n = 64 pair years) produced offspring in at least 1 year. Return Rates and Territory Size. — The re- turn rate (T ± SD) of territorial adults from 1998 to 1999 was 0.39 ± 0.14 {n = 5/13; 95% Cl = 0.14-0.68). Thirteen individuals marked in 1999 were added to the sample group of five survivors from 1998. The return rate from 1999 to 2000 was 0.72 ± 0.11 {n - 13/18; 95% Cl = 0.47-0.90). None of the individuals marked in 1998 and missing in the 1999 sam- pling period was reobserved in 2000. Mean (± SD) territory size was 0.86 ± 0.20 ha and ranged from 0.56 to 1.08 ha {n = 9). I did not observe territory switching in marked individuals. Ninety-one percent of all territorial males (1998-2000) were paired {n = 69). This number ranged from 96% in 1998 {n = 23) to 87% in 1999 {n = 23) and 91% in 2000 {n = 23). Density of male territories was 0.57, 0.44, and 0.41/ha, in 1998, 1999, and 2000, respectively. DISCUSSION The main difference in nesting success be- tween Maricao and Guanica PRVI populations was related to parasitism rates. Brood parasit- ism was absent in the montane population, but 73-83% of the nests in the lowland were par- asitized by the Shiny Cowbird (Woodworth 1997). The dissimilarity is related to high cowbird abundance in the southern coastal plain where the species is favored by land uses such as agriculture and cattle ranching. Large pastures in the vicinity of the forest re- serve in Guanica (S. Molina-Coldn, pers. comm.) contrast to only 4% pastures in 19,382 ha surrounding Maricao (Tossas and Thomlin- son 2007). The lower and more open strucUire of the dry forest in Guanica may also facilitate nest searching by cowbirds. Differences in reproductive success be- tween PR Vis in Maricao and Guanica may be 464 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 affecting the amount of effort invested by pairs in nest building and egg-laying. Pairs in Maricao built new nests after a nesting failure but not after a successful nest. However, pairs in Guanica renested as many as six times after previous failures and also had second broods (Woodworth 1997). This behavior may com- pensate for high rates of nest loss and low reproductive success. Weather condition may be the ultimate factor affecting the number of nesting attempts and population dynamics overall, as extremely dry conditions have been shown to have a direct relationship with bird population declines in Guanica (Faaborg et al. 1984). Woodworth (1997) found the arrival of rains early in 1991 facilitated rapid initiation of breeding behavior by the PRVI in Guanica with increased opportunities for laying second clutches. That year, six pairs that had fledged young by the beginning of June were able to initiate second broods 1-2 weeks later. In ad- dition to higher chances of reproductive suc- cess, nests initiated early in the season were less likely to be parasitized (Woodworth 1997). Maricao and Guanica can be considered the extremes of a gradient differing in elevation above sea level and habitat characteristics, but aspects of the biology of PRVI were similar in both populations. For example, breeding season lengths at Maricao were similar to those observed in Guanica (106 days in 1991, 69 days in 1993; Woodworth 1997). High lev- els of nest predation were responsible for nest losses at both sites with 63% of nests at Mar- icao (this study) and 70% of nests at Guanica (Woodworth 1997) lost to predation. These numbers are typical for open-cup understory nests in the Neotropics, as 67% of nests in wet forests in Costa Rica (Skutch 1985), and 62 and 68% of nests in lowland Panama, in 1996 and 1997, respectively (Robinson et al. 2000) were lost to predation. The identity of predators could not be as- certained in most cases in the present study, but the lack of damage to the nests suggests birds or reptiles were responsible. Snakes (Al- sophis portoricensis, Epicrates inornatus) that could predate bird eggs or young are uncom- mon in Maricao, but potential avian predators were frequently observed. The latter include the Puerto Rican Sharp-shinned Hawk {Accip- iter striatus Venator), Pearly-eyed Thrasher {Margarops fuscatus), Puerto Rican Lizard Cuckoo (Coccyzus vieilloti), and Red-legged Thrush (Turdus plumbeus). I cannot dismiss the possibility that nest losses were caused by rodents due to the difficulty of predator iden- tification based on nest damage characteristics (Marini and Melo 1998). Woodworth (1999) indicated that Guanica is a population sink for the PRVI due to high levels of nest parasitism and depredation. A computer simulation model of the PRVI pop- ulation with the levels of reproductive success found at Maricao implies this population has a positive growth rate even when the produc- tivity index used in the model is probably an underestimate of the real number of fledglings produced (Tossas 2002). The differences in nest success between the two populations con- tribute to the idea of a source-sink metapop- ulation structure, but the short dispersal dis- tances reported for juvenile PRVIs (Wood- worth et al. 1998) suggest that Maricao indi- viduals may not migrate as far as Guanica. Alternatively, the surplus of individuals pro- duced at Maricao may disperse to forest frag- ments surrounding this forest reserve (Tossas and Thomlinson 2007). Hurricanes directly increase avian mortality rates due to collisions during the storm, or in- directly by destroying the resources on which they depend (Wiley and Wunderle 1993). A study of the impacts of Hurricane Georges in 1998 to the Maricao avian community showed that 16 of 21 species, including the PRVI, de- clined after the hurricane (Tossas 2006). The reasons for the decline of each species may vary since they have particular ecological re- quirements and evidence on how resources are affected is scant. Nesting success parameters of PRVI did not vary significantly before and after the hurricane, but a low return rate of color-banded adults was observed in 1999 suggesting that disappearance of individuals was related to the effects of the hurricane. This finding contrasts with the vireo’s high re- turn rate and territorial fidelity in a year with- out a hurricane. Thus, the 1999—2000 annual return rates of 72% was likely typical for the species, while the 39% return estimated from 1998 to 1999 was probably caused by either mortality or dispersal during or immediately after the hurricane. The estimated return in 1999-2000 is similar to the annual survival Tossas • REPRODUCTIVE SUCCESS OF THE PUERTO RICAN VIREO 465 probabilities (68-74%) of PRVIs at Guanica reported by Woodworth et al. (1999), Faaborg and Arendt (1995), and Faaborg et al. (1997). The decline in adult return rate resulted in 26% lower density of territorial males the year following the hurricane. Nest parasitism was not found to threaten the PRVI in this study, but the Shiny Cowbird population may increase at Maricao if hurri- canes facilitate their colonization by opening the canopy and altering microhabitat charac- teristics. Brood parasites and predators may become more abundant if the amount of dense forest in the area surrounding Maricao is af- fected by fragmentation and deforestation. Changes in land-use practices affecting habitat continuity include rural development by an in- creased human population and substitution of traditional shade-coffee plantations for sun- grown varieties or other crops with higher yields. ACKNOWLEDGMENTS This work was supported by the Department of Bi- ology at the University of Puerto Rico, Rio Piedras and grants from the National Science Foundation (Center for Research Excellence in Science and Tech- nology, HRD-9353549; Alliance for Graduate Educa- tion and the Professoriate, HRD-98 17642). The Puerto Rico Department of Natural and Environmental Re- sources provided permits to work in Maricao State Forest. Beatriz Hernandez and L. A. Muniz-Campos were helpful in the field. I am grateful to J. M. Wun- derle Jr. and R. B. Waide for guidance during the de- velopment of the research project, and to B. L. Wood- worth for thoughtful reviews of the manuscript. This paper also benefited from comments by J. D. Nichols, E. A. VanderWerf, and two anonymous reviewers. LITERATURE CITED Bennett, S. P. and R. Mojica. 1998. Hurricane Georg- es preliminary storm report: from the tropical At- lantic to the United States Virgin Islands and Puerto Rico. National Weather Service, San Juan, Puerto Rico. Bibby, C. j., N. D. Burge.ss, D. A. Hill, and S. Mus- TOE. 2()()0. Bird census techniques. Second Edi- tion. Academic Press, London, United Kingdom. Cruz, A., T. Manolis, and J. W. Wiley. 1985. The Shiny Cowbird: a brood parasite expanding its range in the Caribbean region. Ornithological Monographs 36:607-620. Ewel, J. J. AND J. L. Whitmori;. 1973. The ecological life zones of Puerto Rico and the U.S. Virgin Is- lands. Re.search Paper IT! -18. USDA, Forest Ser- vice, Institute of Tropical Forestry, Rfo Piedras, Puerto Rico. Faaborg, J. and W. J. Arendt. 1995. Survival rates of Puerto Rican birds: are islands really that dif- ferent? Auk 112:503-507. Faaborg, J., W. J. Arendt, and M. S. Kaiser. 1984. Rainfall correlates of bird population fluctuations in a Puerto Rican dry forest: a nine year study. Wilson Bulletin 96:575-593. Faaborg, J., K. M. Dugger, W. J. Arendt, B. L. Woodworth, and M. E. Baltz. 1997. Population declines of the Puerto Rican Vireo in Guanica Forest. Wilson Bulletin 109:195-202. Johnson, D. H. 1979. Estimating nest success: the Mayfield method and an alternative. Auk 96:651- 661. Kepler, C. B. and A. K. Kepler. 1970. Preliminary comparison of bird species diversity and density in Luquillo and Guanica forests. Pages El 83-1 91 in A tropical rain forest (H. T. Odum, Editor). U.S. Atomic Energy Commission, Oak Ridge, Tennes- see, USA. Krebs, C. J. 1989. Ecological methodology. Harper and Row, Publishers, New York, USA. Marini, M. A. and C. Melo. 1998. Predators of quail eggs, and the evidence of the remains: implica- tions for nest predation studies. Condor 100:395- 399. Mayfield, H. F. 1975. Suggestions for calculating nest success. Wilson Bulletin 87:456-466. Perez-Rivera, R. a. 1986. Parasitism by the Shiny Cowbird in the interior parts of Puerto Rico. Jour- nal of Field Ornithology 57:99-104. Purcell, K. L. 2006. Abundance and productivity of Warbling Vireos across an elevational gradient in the Sierra Nevada. Condor 108:315-325. Raffaele, H. a. 1989. A guide to the birds of Puerto Rico and the Virgin Islands. Princeton University Press, Princeton, New Jersey, USA. Robinson, W. D., T. R. Robinson, S. K. Robinson, and J. D. Brawn. 2000. Nesting success of understory forest birds in central Panama. Journal of Avian Biology 31:151-164. Rothstein, S. I. 1990. A model system for coevolu- tion: avian brood parasitism. Annual Review of Ecology and Systematics 21:481-508. SiLANDER, S., H. Gil de Rubio, M. Miranda, and M. Vazquez. 1986. Los bosques de Puerto Rico. Pag- es 210-236 in Compendio enciclopedico de los recursos naturales de Puerto Rico. Volumen 10. (J. L. Vivaldi, Editor). Departamento de Recursos Naturales, San Juan, Puerto Rico. Skutch, a. E 1985. Clutch size, nesting success, and predation on nests of neotropical birds, reviewed. Ornithological Monographs 36:575-594. Tossas, A. G. 2002. Puerto Rican Vireo demography in a montane habitat, with metapopulation impli- cations. Dissertation. University of Puerto Rico, Rio Piedras. Tossas, A. G. 2006. Fiffects of Hurricane Georges on the resident avifauna of Maricao State Forest, southwestern Puerto Rico. C'aribbean Journal of Science 42:81-87. 466 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 Tossas, a. G. and J. Thomlinson. 2007. Landscape characteristics of Puerto Rican Vireo (Vireo lati- rneri) nesting habitat, with source-sink implica- tions. Ornitologia Neotropical 18:233-242. Ward, D. and J. N. M. Smith. 2000. Brown-headed cowbird parasitism in a sink population in War- bling Vireos. Auk 117:337-344. Wenny, D. G., R. L. Clawson, J. Faaborg, and S. L. Sheriff. 1993. Population density, habitat selec- tion and minimum area requirements of three for- est-interior warblers in central Missouri. Condor 95:968-979. Wiley, J. W. and J. M. Wunderle Jr. 1993. The ef- fects of hurricanes on birds, with special reference to Caribbean islands. Bird Conservation Interna- tional 3:319-349. Woodworth, B. L. 1997. Brood parasitism, nest pre- dation, and season-long reproductive success of a tropical island endemic. Condor 99:605-621. Woodworth, B. L. 1999. Modeling population dy- namics of a songbird exposed to parasitism and predation and evaluating management options. Conservation Biology 13:67-76. Woodworth, B. L., J. Faaborg, and W. J. Arendt. 1998. Breeding and natal dispersal in the Puerto Rican Vireo. Journal of Field Ornithology 69:1-7. Woodworth, B. L., J. Faaborg, and W. J. Arendt. 1999. Survival and longevity of the Puerto Rican Vireo. Wilson Bulletin 111:376-380. The Wilson Journal of Ornithology 120(3):467^72, 2008 NEST DEFENSE BY CAROLINA WRENS KELLY A. D’ORAZIO'-2 AND DIANE L. H. NEUDORE" ^ ABSTRACT. — We examined nest defense behavior of the Carolina Wren (Thryothorus ludovicianus) in re- sponse to avian intruders at the nest. Freeze-dried mounts of a brood parasite (female Brown-headed Cowbird [Molothrus aterf), nest predator (Blue Jay [Cyanocitta cristatof), and a control (Swainson’s Thrush [Catharus ustulatus]) were presented at wren nests to examine if wrens were able to distinguish among these threats. The primary nest defense behavior of Carolina Wrens was alarm calls. Wrens spent more time alarm calling to the Blue Jay model than the control and Brown-headed Cowbird models suggesting cowbirds were not recognized as threats to the nest. Wrens were less likely to respond during laying stage trials, probably due to a lack of nest visitations at this time. Intensity of alarm calls did not increase from the laying to nestling stage for any of the models presented. Carolina Wrens are both socially and genetically monogamous, and males should invest heavily in care of young due to their high confidence of paternity. Males and females did not differ in the amount of time spent alarm calling to any of the models, and defended their nests from intruders with equal intensity as predicted by the confidence of paternity hypothesis. Received 21 October 2006. Accepted 4 October 2007. Nest defense is a common behavior used by birds to improve survival of their young in an attempt to increase the potential of passing their genes on to future generations. Nest de- fense can be costly and parents are expected to make an optimal compromise between their own safety and that of their young (Andersson et al. 1980). Responses of nest owners are ex- pected to be influenced by the reproductive value of their young and the stimulus value of the intruder (Patterson et al. 1980). Thus, par- ents should respond selectively to intruders that pose the most threat in conjunction with timing of the nesting cycle and the expected future benefits of survival of young. The two main causes of nesting failure in songbirds are predation (Wilcove 1985, Mar- tin 1992, Robinson 1992) and brood parasit- ism (Mayfield 1977, Brittingham and Temple 1983, Robinson et al. 1995). Predation is re- sponsible for loss of over 50% of eggs and nestlings in some passerine species (Ricklefs 1969). Hosts of the parasitic Brown-headed Cowbird {Molothrus citer) have reduced nest- ing success because of removal or damage to their eggs, and nestling or fledgling death from competition (Friedmann 1963, Rasmus- sen and Sealy 2006). Cowbirds also depredate nests by destroying eggs or killing nestlings ' Department of Biological Sciences, Sam Houston State University, Huntsville, TX 77341, USA. ^ Current address: 792 East Cowboy Cove. Queen Creek, AZ 85242, USA. ^Corresponding author; e-mail: neudorf@shsu.edu (e.g., Friedmann 1929, Du Bois 1956, Gran- fors et al. 2001, Smith et al. 2003). Carolina Wrens {Thryothorus ludovicianus) are a non-migratory songbird of the eastern United States and portions of Central America (Haggerty and Morton 1995). Brood parasit- ism of the Carolina Wren by the Brown-head- ed Cowbird has previously been documented (Bent 1948, Mengel 1965, Woodward 1983, Haggerty and Morton 1995). Carolina Wrens are recognized as an acceptor of cowbird eggs and can raise fledglings to independence (Woodward 1983). However, little is known about the wren’s nest defense behavior toward cowbirds during their breeding season. We quantified nest defense behavior of the Carolina Wren in response to models of a: (1) female Brown-headed Cowbird, (2) Blue Jay {Cyanocitta cristata) (an avian nest predator), and (3) Swainson’s Thrush {Catharus ustula- tus) (a non-threatening species), during the egg-laying and nestling stages. We predicted that wrens would increase their level of de- fense towards an avian nest predator over the nesting cycle as young increa.sed in value to the parents. Responses to the female cowbird model are predicted to be most intense at the laying stage when parasitism is a threat (e.g., Neudorf and Sealy 1992). Alternatively, if wrens view cowbirds as nest predators, re- sponses to the cowbird model should also be inten.se at the nestling stage. We compared the responses of males and females to each in- truder to examine if parents differed in their responses. Carolina Wrens are both socially 467 468 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 and genetically monogamous (Haggerty et al. 2001); males have a high confidence of pater- nity and should invest heavily in offspring care. Thus, we predicted that both males and females would defend their nests with equal intensity (Hobson et al. 1988, Montgomerie and Weatherhead 1988). METHODS Studx Site. — This study was conducted from March- August 2001 through 2003 on and around Sam Houston State University’s Cen- ter for Biological Field Studies, Huntsville, Texas. The 100-ha area (30° 74' N, 95° 48' W) provides abundant habitat for Carolina Wrens, which are year-round residents at the field sta- tion. Habitat includes mixed pine {Finns spp.)-hardwood forests, open prairie inclu- sions, old-field succession, and riparian areas along two creeks (Dent and Lutterschmidt 2001). Nest Monitoring and Banding. — Carolina Wrens nest in both natural and artificial sites (Haggerty and Morton 1995) that provide a suitable cavity. We constructed 120 nest boxes and placed them in appropriate habitats, and within known territories, to entice the birds to nest. Boxes were checked every 2-3 days for signs of nesting activity. Males and females are nearly identical in plumage, but males often have greater mass with longer bills, wings, and legs (Haggerty and Morton 1995). Most individuals were cap- tured with mist nets coupled with song or call playback, uniquely color banded, and re- leased. Gender was assigned by the presence or absence of a brood patch or cloacal protu- berance, and vocalizations. Model Presentation. — Freeze-dried mounts of a female Brown-headed Cowbird. Blue Jay, and Swainson’s Thrush were presented at wren nests to quantify nest defense. Mounts were prepared with glass eyes and made to appear life-like. We used only one model of each of the three species we presented. These were salvaged specimens and should represent a random individual from each species. The mounts remained in good condition over the course of the study because wrens did not physically contact them. A thrush was chosen as a control because it is similar in size, shape, and coloration to the female cowbird. Thrushes are migrants through the study area and pose no threat to wrens. The Blue Jay is larger than the other two model species, but is the smallest known avian predator of eggs and nestlings that is also abundant in the study area. Models were wired to an alligator clip at- tached to a moveable perch to allow ease of manipulation during testing. The perch was placed 1 m from the nest and positioned with the mount facing the nest opening. Presenta- tions of models were in random order and each nest was tested only once per breeding season. Obser\'ations. — Procedures for nest defense testing generally followed Sealy et al. (1998). All observations were conducted by a single individual (KAD) for consistency and nests were tested between 0700 and 1100 hrs (CST). Laying stage observations were con- ducted anytime after the first egg was pro- duced until the penultimate egg was laid. Car- olina Wren clutch size is typically four eggs that are laid on successive days within 1-2 hrs of sunrise (Haggerty and Morton 1995). Nest- ling stage observations were conducted on nests containing 5 to 1 1 day-old nestlings. Ob- servations were made from a blind 5—10 m from the nest as surrounding vegetation al- lowed. The blind and perch were positioned at least 1 day prior to testing to allow birds to habituate to these objects near their nests. The observer hid in the blind until the mated pair was away from the nest and then positioned models on the perch and returned to the blind. The 5 -min trial commenced when we ob- served either bird return. Each model was pre- sented for a 5-min trial with at least a 15-min rest period between presentations. The model was immediately removed after a trial and the observer returned to the blind. We found 15 min to be sufficient to allow for birds to stop responding and resume normal activity after a model was removed. We did not put the next model in place if the parents were still at the nest after 15 min until they left the area. Ob- servations during testing were recorded on a portable cassette recorder and later tran- scribed. Three commonly used alarm calls including female “rasp ”, female “dit ”, male ti-dink and male “rasp”” (Haggerty and Morton 1995) were observed in response to models. All calls were quantified as the number of 10-sec in- D'Orazio and Neiidorf • NEST DEreNSE BY CAROLINA WRENS 469 TABLE 1. Alarm calls of Carolina Wrens to models during egg-laying and nestling stages. Response‘s Model Cowbird Blue Jay Thrush Egg laying 4.9 ± 3.7 n = 5 15.2 ± 4.0 n = 9 3.6 ± 2.4 n = 6 Nestling 6.6 ± 1.8" n = 16 17.2 ± 2.L n = 16 6.0 ± 2.2" n = 16 P 0.45 0.44 0.62 P 0.66 0.67 0.54 ® Responses are given as the mean ± SE. Alarm calls were measured as the number of 10-sec intervals that wrens responded. Responses were averaged w'hen two wrens responded. *’• ‘^Results of Bonferroni multiple comparison test (nestling stage only) examining differences in responses between models. Means with different superscripts differed significantly. ^ Two-sample f-tests to compare responses between nest stages. All tests were two-tailed. tervals in which they occurred during model presentation (maximum of 30 for a 5-min tri- al). Any aggressive behaviors sueh as close passes (flying within 0.5 m above the model) or strikes at the models were also noted. A bird was given a zero for the alarm call cat- egory if it returned to the nest but did not vocalize. The next model was put in place af- ter a 1 5-min rest period if no birds were ob- served returning to the nest. The trial was con- sidered a “no response” after 1 hr and the model was removed. The visibility of return- ing adults to the observer was limited to 5-10 m around the nest depending on the sur- rounding vegetation. It is possible that if a bird returned and sat quietly at a distance it may not have been seen by the observer. Statistical Analyses. — Statistical compari- son of wren alarm calls among models within the same nesting stage was only possible for the nestling stage due to the number of no responses during several trials at the egg lay- ing stage. Different alarm calls for males and females were combined into one alarm call category to facilitate analysis. The results were averaged if two individuals responded to a given model. A repeated measures ANOVA was performed to detect differences in nest defense behavior to presented models at the nestling stage. Significant results iP < 0.05) were analyzed by Bonferroni’s multiple com- parison test to identify those models that elic- ited signifieantly different responses. We used a two-sample r-test to compare nest defense responses between nest stages for each model. A paired /-test was used to analyze differ- ences between responses of males and females for each model. We used trials only where both members of the pair responded to a par- ticular model. Gender responses were com- pared only at the nestling stage due to the low number of individuals responding at the lay- ing stage. Chi-square or Fisher’s exact tests were used to compare the number of males and females that responded among models. All values are reported as mean ± SE and all tests were two-tailed. Data were analyzed with GraphPad Prism 4.0 or StatView 4.5 software. RESULTS The primary nest defense response of Car- olina Wrens to model intruders at the nest was alarm ealls. Only one female made a close pass at each model during a nestling stage trial and no strikes were observed. It was not un- usual for pairs to feed nestlings before or after giving alarm calls at the nest. Discrimination Among Models. — Discrimi- nation among models at the laying stage was difficult to assess due to lack of parental re- sponse. However, the data suggest a stronger response to the Blue Jay model (Table 1 ). Time spent alarm ealling at the nestling stage was greatest toward the Blue Jay model (re- peated measures ANOVA, F2.30 - 11.9, P = 0.0002; Table 1). Time spent alarm calling did not increase significantly over the nesting cy- cle for any of the models (Table 1 ). Differences Between Male and Female Re- sponses.— The number of lO-sec intervals spent alarm calling did not differ at the nest- ling stage between males and females in re- sponse to the Blue Jay model (male: 17.0 ± 2.9, female: 16.0 ± 3.3; paired /-test = 0.21. df = 12, P = 0.83) or the thrush model (male: 11.1 ± 3.6, female: 8.5 ± 4.5; paired /-test, / = 0.58, df = 7, P = 0.58). Females spent more time alarm calling to the cowbird model than did males but this difference was not sig- 470 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 nificant (male: 5.2 ± 2.3, female: 13.2 ± 3.6; t = 1.91, df = 10, P = 0.09). Male (2, 5, and 1) and female (5, 9, and 4) Carolina Wrens (n = \5 nests tested during the egg-laying stage) responded to the cow- bird, Blue Jay, and thrush models, respective- ly. Females appeared to respond more than males at the laying stage but differences were not significant (Fisher exact test, NS; compar- ing responses to thrush and Blue Jay models). Sixteen nests were tested during the nestling stage and 12, 14, and 13 males and 12, 14, and 13 females responded to the cowbird. Blue Jay, and thrush models, respectively. Males and females responded in similar num- bers at the nestling stage to all three models (X^ = 0.321, df = 2, P = 0.85). DISCUSSION Carolina Wrens relied on alarm calls as their primary defense. Their use of vocaliza- tions as a defensive system occasionally alert- ed other species of birds in the area which congregated near the model resulting in a mobbing response. Mobbing may function to let a predator know it has been discovered and cause it to leave the area (Curio 1978). Car- olina Wrens in an adjacent territory responded to the alarm calls of the nest owners during only one observation. Carolina Wrens often did not respond to model trials at the laying stage. Models in sev- eral instances were left in place for 1 hr and no birds were observed anywhere near the nest area. This is likely due to a lack of nest visitations by wrens during the laying stage rather than birds hiding from the observer’s view. Nice and Thomas (1948) reported few nest visitations by a pair of Carolina Wrens during the laying stage. Most observed visits were early in the morning and associated with egg laying. We found no difference among models in the number of males and females responding at the laying stage. We would ex- pect lower numbers of males and females re- sponding to the Blue Jay model than the thrush model if birds were hiding out of sight to avoid attracting a predator to the nest. Birds may reduce visits to their nests to avoid at- tracting attention of predators or brood para- sites and leading them to the nest (McLean 1987, Eggers et al. 2005). This could account for the low numbers of individuals responding at the laying stage in our study. However, when wrens did respond at the laying stage their level of response was similar to the nest- ling stage (Table 1). Recognition of Threats. — Carolina Wrens recognized and defended their nests from the model of the Blue Jay, a common nest pred- ator on our study site (Table 1). The intensity of nest defense by wrens in response to the Blue Jay model did not increase as the nesting cycle progressed from egg-laying to nestling stage. Parents are expected to increase re- sponses to predators over the nesting cycle ac- cording to the reproductive value hypothesis (Andersson et al. 1980, Patterson et al. 1980). This hypothesis has been supported in several studies of songbirds (Greig-Smith 1980, West- neat 1989, Burhans 2001). However, other studies did not find the predicted increase in nest defense over the nesting cycle (Knight and Temple 1986) or did not find an increase in all species examined (Neudorf and Sealy 1992). The primary nest defense behavior of Carolina Wrens was alarm calls, which is pre- sumably a lower risk activity to parents than close passes or strikes. The risk of injury from the much larger Blue Jay may have out- weighed the potential benefits of increased ag- gression at the nestling stage. Level of alarm calling to the cowbird model was similar to that of the control suggesting wrens do not recognize cowbirds as a signif- icant threat. There are no records of Brown- headed Cowbird parasitism of Carolina Wrens in our study area {n = 118 nests over 6 years; D. L. H. Neudorf, unpubl. data) perhaps be- cause wrens use nest boxes that are too small and enclosed for cowbirds. Many birds rec- ognize cowbirds as threats to their nests (e.g.. Song Sparrows [Melospiza melodia]. Smith et al. 1984; Red-winged Blackbirds [Agelaius phoeniceus], Neudorf and Sealy 1992; Hood- ed Warblers [Wilsonia citrinaf Mark and Stutchbury 1994; Yellow Warblers [Dendroi- ca petechia]. Gill and Sealy 1996). However, these species are more common cowbird hosts and presumably have more opportunities to interact with cowbirds. Differences in Responses of Males and Fe- males.— Male and female Carolina Wrens de- fended their nests from the models with equal intensity at the nestling stage. Neither differed in propensity to respond nor amount of time D’Orazio and Neudorf • NEST DEFENSE BY CAROLINA WRENS 471 spent alarm calling to the models. Differences between males and females in nest defense are not uncommon in birds and can be attributed to many factors (Montgomerie and Weather- head 1988). Generally, females have a higher confidence of parenthood (Trivers 1972) and should take more risks in an effort to increase their fitness (Montgomerie and Weatherhead 1988). Carolina Wrens are socially monoga- mous and extra-pair copulations apparently do not occur in this species (Haggerty et al. 2001). Thus, it appears parental care in the form of nest defense is shared equally in Car- olina Wrens supporting the confidence of pa- ternity hypothesis. ACKNOWLEDGMENTS We are grateful to Jana Story and Anne Domonoske for assistance during the 2001 field season. T. M. Hag- gerty provided advice on nest box design. Everett Wil- son generously constructed nest boxes. Keith Arnold and William Armstrong donated bird specimens for models. Robert Rhodes provided assistance with freeze-drying. Colleen Barber and two anonymous re- viewers provided comments on the manuscript. Fund- ing was provided by Sam Houston State University (SHSU), Department of Biological Sciences Research Award (2001 and 2002 to KAD) and a SHSU Research Endowment Fund Grant (2001 to DLHN). LITERATURE CITED Andersson, M., C. G. Wiklund, and H. Rundgren. 1980. Parental defense of offspring: a model and an example. Animal Behaviour 28:536-542. Bent, A. C. 1948. Life histories of North American nuthatches, wrens, thrashers, and their allies. U.S. National Museum Bulletin 195. Brittingham, M. C. and S. A. Temple. 1983. Have cowbirds caused forest songbirds to decline? BioScience 33:31-35. Burhans, D. E. 2001. Enemy recognition by Field Sparrows. Wilson Bulletin 113:189-193. Curio, E. 1978. The adaptive significance of avian mobbing. 1. Teleonomic hypotheses and predic- tions. Zeitscrift Tierpsychologie 48:175-183. Dent, L. and W. 1. Lutterschmidt. 2001. The ich- thyofauna of Harmon and Wynne creeks sampled within the Center for Biological Field Studies, Walker County, Texas. Texas Journal of Science 53:139-146. Du Bois, A. D. 1956. A cowbird incident. Auk 73: 286. Eggers, S., M. Grie.sser, and J. Ekman. 2005. Pred- ator-induced plasticity in nest visitation rates in the Siberian Jay (Perisoreus infaustus). Behavior- al Ecology 16:309-315. Friedmann, H. 1929. The cowbirds: a study in the bi- ology of social parasitism. C. C. Thomas, Spring- field, Illinois, USA. Friedmann, H. 1963. Social parasitism in birds. Quar- terly Review of Biology 3:554-569. Gill, S. A. and S. G. Sealy. 1996. Nest defense by Yellow Warblers: recognition of a brood parasite and an avian nest predator. Behaviour 133:263- 282. Graneors, D. a., P. J. Pietz, and L. A. Joyal. 2001. Frequency of egg destruction by female Brown- headed Cowbirds at grassland nests. Auk 118: 765-769. Greig-Smith, P. W. 1980. Parental investment in nest defense by Stonechats {Saxicola torqiiata). Ani- mal Behaviour 28:604-619. Haggerty, T. M. and E. S. Morton. 1995. Carolina Wren {Thryothorus ludovicianus). The birds of North America. Number 188. Haggerty, T. M., E. S. Morton, and R. C. Fleischer. 2001. Genetic monogamy in Carolina Wrens {Thryothorus ludovicianus). Auk 118:215-219. Hobson, K. A., M. L. Bouchart, and S. G. Sealy. 1988. Reponses of naive Yellow Warblers to a novel nest predator. Animal Behaviour 36:1823- 1830. Knight, R. L. and S. A. Temple. 1986. Why does intensity of avian nest defense increase during the nesting cycle? Auk 103:318-327. Mark, D. and B. J. Stutchbury. 1994. Response of a forest-interior songbird to the threat of cowbird parasitism. Animal Behaviour 47:275-280. Martin, T. E. 1992. Breeding productivity consider- ations: what are the appropriate habitat features for management? Pages 455-473 in Ecology and conservation of neotropical migrant landbirds (J. M. Hagan III and D. W. Johnston, Editors). Smith- sonian Institute Press, Washington, D.C., USA. Mayeield, H. 1977. Brown-headed Cowbird: agent of extermination. American Birds 31:107-113. McLean, I. G. 1987. Response to a dangerous enemy: should a brood parasite be mobbed? Ethology 75: 235-245. Mengel, R. M. 1965. The birds of Kentucky. Ornitho- logical Monographs 3. Montgomerie, R. D. and P. J. Weatherhead. 1988. Risks and rewards of nest defense by parent birds. Quarterly Review of Biology 63:167-187. Neudorf, D. L. and S. G. Sealy. 1992. Reactions of four passerine species to threats of predation and cowbird parasitism: enemy recognition or gener- alized responses? Behaviour 123:85-105. Nice, M. M. and R. H. Thomas. 1948. A nesting of the Carolina Wren. Wilson Bulletin 60:139-158. Paiter.son, T. L., L. Peirinovich. and D. K. James. 1980. Reproductive value and appropriateness of response to predators by White-crowned Spar- rows. Behavioral Ecology and Sociobiology 7: 227-23 1 . Rasmlissen, j. L. and S. G. Sealy. 2006. Hosts feed- ing only Brown-headed Cowbird fledglings: 472 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 3, September 2008 where are the host fledglings? Journal of Lield Or- nithology 77:269-279. Ricklefs. R. E. 1969. An analysis of nesting mortality in birds. Smithsonian Contributions to Zoology 9: 1-48. Robinson, S. K. 1992. Population dynamics of breed- ing neotropical migrants in a fragmented Illinois landscape. Pages 408-418 in Ecology and con- servation of neotropical migrant landbirds (J. M. Hagan III and D. W. Johnston, Editors). Smith- sonian Institute Press. Washington, D.C., USA. Robinson, S. K., S. I. Rothstein, M. C. Brittingh.am, L. J. Petit, .\nd J. A. Grzybowski. 1995. Ecology and behavior of cowbirds and their impact on host populations. Pages 428-460 in Ecology and man- agement of neotropical migratory birds (T. E. Martin and D. Einch, Editors). Oxford University Press, New York. USA. Sealy, S. G., D. L. Neudorf, K. A. Hobson, .and S. A. Gill. 1998. Nest defense by potential hosts of the Brown-headed Cowbird: methodological ap- proaches, benefits of defense, and coevolution. Pages 194-211 in Parasitic birds and their hosts: studies in coevolution. (S. I. Rothstein and S. K. Robinson, Editors). Oxford University Press, New York. USA. Smith, J. N. M., P. Arcese, and I. G. McLean. 1984. Age, experience, and enemy recognition by wild Song Sparrows. Behavioral Ecology and Socio- biology 14:101-106. S.MiTH, J. N. M., M. J. Taitt, L. Zanette. a.nd I. H. Myers-S.mith. 2003. How do Brown-headed Cowbirds {Molothrus ater) cause nest failures in Song Sparrows (Melospiza melodia)? A removal experiment. Auk 120:772-783. Trivers. R. L. 1972. Parental investment and sexual selection. Pages 139-179 in Sexual selection and the descent of man (B. Campbell. Editor). Aldine. Chicago, Illinois, USA. Westne.at. D. E 1989. Intensity of nest defense in In- digo Buntings increases with stage and not num- ber of visits. Auk 106:747-749. WiLCOVE. D. S. 1985. Nest predation in forest tracts and the decline of migratory songbirds. Ecology 66:1211-1214. Woodw ard, P. W. 1983. Behavioral ecology of fledg- ling Brown-headed Cowbirds and their hosts. Condor 85:151-163. The Wilson Journal of Ornithology 120(3):473^77, 2008 NEST, EGGS, AND PARENTAL CARE OF THE PUNA TAPACULO {SCYTALOPUS SIMONSI) PETER A. HOSNER'-3 AND NOEMl' E. HUANCA^ ABSTRACT. — We describe the nest and eggs of the Puna Tapaculo (Scytalopus simonsi) from Bolivia, and include observations of nest building, incubation, and parental care. The nest is similar to several other described nests in the genus jn construction and placement: a domed cup nest of grasses in an excavated burrow in a vertical bank. Both male and female constructed the nest, brooded, and provisioned the young, typical of Scytalopus and tracheophone suboscines. This is only the second described Scytalopus nest constructed of grasses, probably an adaptation to its drier habitat near and above treeline. The growing body of Scytalopus nest descriptions suggests they do not exhibit generic level stereotyped nest structure and placement, unlike other tracheophone suboscines, which show strong phylogenetic signal in nest architecture. Received 1 September 2007. Accepted 26 January 2008. Scytalopus is among the most diverse and abundant groups of Andean birds, yet their breeding biology is poorly understood. Nests of 15 of the 41 (Gill and Wright 2006, Rem- sen et al. 2008) currently recognized species of Scytalopus have been described, all of which share similarities but differ somewhat in placement or structure. The nest site is usu- ally some form of natural crevice or cavity (Skutch 1972, Hilty and Brown 1986, De San- to et al. 2002, Acros-Torres and Solano-Ugal- de 2007), excavated burrow in an earthen bank (Johnson and Goodall 1965, Stiles 1979, Young and Zuchowski 2003, Decker et al. 2007, Pulgarin-R 2007), cavity in a rotten log (Greeney and Gelis 2005), or hidden among dense foliage, roots, trunks, fallen leaves, and moss (Rosenberg 1986, Sick 1993, Christian 2001). Nest architecture in Scytalopus ap- proaches the entire range of variation of all the rhinocryptids (Krabbe and Schulenberg 2003) and nests usually are domed or globular in shape with a top or side entrance; only nests of White-browed Tapaculo {S. superciliaris) (Stiles 1979) and Long-tailed Tapaculo {S. mi- cropterus) (Greeney et al. 2005) have been found with an open cup. Nests are constructed from a variety of plant material including mosses, lichens, rootlets, leaves, small sticks, and are at times lined with grasses, hair, or feathers (Krabbe and Schulenberg 2003). ' Department of Ecology and Evolutionary Biology, University of Kansas, Lawrence, KS 66045, USA. - Asociacfon Armonia, Santa Cruz de la Sierra, San- ta Cruz, Bolivia. ^Corresponding author; e-mail: hosner@ku.edu Field observations of Scytalopus during the breeding period are extremely limited. Males lack a brood patch (Krabbe and Schulenberg 2003), but incubation has been reported by both males and females (Young and Zuchows- ki 2003, Decker et al. 2007), as has brooding (Greeney et al. 2005). The incubation period has been reported to be from 15 (Sick 1993) to 19 days (DeSanto et al. 2002); the only recorded fledging period is 1 1 days (DeSanto et al. 2002). Both parents actively provision the chicks with small arthropods (Christian 2001, Young and Zuckowski 2003, Greeney and Gelis 2005, Greeney and Rombough 2005, Greeney et al. 2005). Nothing is re- corded on nest building or post fledging care (Krabbe and Schulenberg 2003). The Puna Tapaculo {Scytalopus simonsi) occurs at high elevations (2,000-4,500 m) in Puna grassland, ravines with shrubby vegeta- tion, Polylepis woodland, and elfin forest edge from Departamento Cuzco, Peru south to De- partamento Santa Cruz, Bolivia (Hennessey et al. 2003, Krabbe and Schulenberg 2003). Once considered conspecific with Magellanic Tapaculo (S. magellanicus) (i.e., Zimmer 1939), differences in song, distribution, and morphology have split this complex into 1 1 recognized species with further taxonomic re- visions likely (Remsen et al. 2008). The ob- jectives of this paper are to describe the nest, eggs, and parental care in S. sitnonsi, and dis- cuss trends in Scytalopus breeding behavior. OBSERVATIONS On 16 December 2006 near the community of Palcapampa (17"20'S, 66° 24' W; 3,550 473 474 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 FIG. 1. Schematic of nest burrow of Puna Tapa- culo showing the false and nest chambers. Photograph of entrance hole (inset). m), Departamento Cochabamba, Bolivia, we observed a Scytalopus simonsi hopping across a footpath carrying grasses. The bird (pre- sumed to be a female as it lacked a white post- ocular line) re-crossed the footpath after 30 sec and hopped up a steep bank, disappearing into one of about nine burrows. When we re- tuned to the site on 17 December 2006, a S. simonisi was peering from the entrance of the nest hole (presumed to be the male of the pair based on the distinctive short whitish post-oc- ular stripe). The bird exited, dropped to the ground and crept away. The bank was 2.5 m tall and was constructed as a terrace between two small potato fields. The bank was almost vertical, sloping gradually to horizontal as it met the lower field. The nest hole was in plain view, 20 cm below the top of the bank, which was covered with grasses and other herba- ceous vegetation. There was a small (15 m wide, 3 m deep) ravine 20 m from the nest site with Alnus acuminata, Polylepis besserii, and other native shrubs including Gynoxis, Baccharis, Ribes, and Berberis, and Festiica and Cortaderia grasses, a more typical habitat of the species. The male was heard singing there infrequently (a few songs per hour) dur- ing investigation of the nest. The entrance hole (Fig. 1) measured 9X4 cm, the only oval hole among many circular holes of various sizes in the bank. The burrow (Fig. 1) went straight into the bank 18 cm, tapering slightly before turning sharply 80° to the right, narrowing to 7 X 4.5 cm, and con- tinuing another 16 cm before opening into two chambers: one empty to the left, and one to the right which contained the nest. A few root- lets from herbaceous vegetation emerged from the ceiling of the burrow. The nest was in an FIG. 2. Nest cup of Puna Tapaculo after extraction from the burrow; the dome could not be extracted for photographs. The two eggs (inset). area replete with rodent burrows (9 other holes within a few meters of the nest). The entrance to the nest chamber was oval, mea- suring 10X8 cm, the nest chamber itself was almost spherical atl2X 12X 18 cm. The sides and ceiling of the chamber were lined with ~4 cm of loosely woven dried Cortad- eria grasses that formed a dome over the nest. The nest cup (Fig. 2) was 12 cm in diameter and 10 cm in height, filling the lower two- thirds of the nest chamber. The walls and bot- tom of the cup were lined with 3.5 cm of much more tightly woven grasses than the loose dome. Five blades of Cortaderia grass taken at random from the nest cup measured 5, 7, 8, 12, and 17 cm in length; all grasses were about 0.5 mm in width. The cup had an inside diameter of 5 cm and inside depth of 6.5 cm. The lining of the interior of the cup was of much smaller, finer grasses than the rest of the nest and included a few long black horse hairs, several species of bird feathers in- cluding domestic chicken {Callus gallus), and a single piece of moss. The nest was clean and devoid of droppings upon extraction. The nest contained two eggs that were cold on 17 December 2007, although the male was in attendance. The eggs were white, the egg- shells seemed thin and had a transparent qual- ity; the air sac at the large end of the egg was easily visible. Eight sets of transparent parallel lines 2 mm apart went half the length of the eggs, converging at the small tip. Egg #1 mea- sured 23.30 X 17.60 mm, weighing 3.9 g; egg #2 measured 23.36 X 17.26 mm and weighed 3.8 g. The male continued to bring stems of grass to the nest after the eggs were replaced. The eggs hatched by 4 January and the Hosner and Huanca • NESTING OF PUNA TAPACULO 475 chicks were estimated to be 2 days of age, weighing 4.5 and 4.9 g with nare to tip of bill measurements of 3.8 and 3.9 mm. Assuming incubation began on 18 December 2007, the day after the two eggs were found to be cold, the incubation period would have been 16 days. The chicks retained the egg tooth, had eyes closed, and were partially covered with long, thick mouse gray natal down on the crown, back, wings, and legs. After the chicks were returned to the nest, the female brooded for 27 min. The male arrived with food and the pair switched with the male staying to brood. The pair alternated brooding and for- aging 33% of the time observed. The brood- ing bird would stay in the nest until the other returned with food, then exited allowing the provisioning bird to enter, feed the chicks, and brood the nestlings. The female made six for- aging bouts and the male four from 1500 to 1630 hrs CST The food brought by the adults appeared to be a mix of —50% small (—10 mm) insect larvae and 50% (—5-10 mm) small adult insects with one small white moth (—10 mm). The female spent 34 min of the 1.5 hrs brooding and at dusk spent the night brooding in the nest. The male spent 16 min brooding during this period and roosted at night in nearby shrubs. The chicks weighed 5.7 and 6.2 g on 6 Jan- uary 2007, and provisioning behavior by the adults was similar although they spent little time brooding, and exited the nest directly af- ter feeding rather than waiting for the arrival of their mate. The female made seven foraging bouts, the male four in 1.15 hrs of observa- tion. Adults entered the nest hole three times but were not identified to gender. Begging calls of the chicks were recorded (ML #132524) by placing a microphone inside the nest for 30 min during active feeding. The chicks called (soft high pitch peeping) almost continuously in the presence and absence of adults. Provisioning behavior was similar in 30 min of observation on 7 January 2007 and both adults were observed removing fecal sacs from the nest for the first time. The chicks were covered in pin feathers on 14 January 2007, although their eyes were still closed and they retained some natal plumes on the crown and back. The majority of the feathers on the back had lost their pins and were brown with thick black barring typ- ical of all juvenile Scytalopus. The chicks weighed 18.1 and 19.8 g, had nare to tip of bill measurements of 6.2 and 6.3 mm, and tar- si of 19.7 and 20.4 mm. The nest was empty on 6 February and was removed from the cav- ity. The fledging period was >12 days and, given the stage of development at 12 days, a fledging period of 15-20 days seems reason- able. DISCUSSION Given the identification issues and taxo- nomic changes in this genus (i.e., Krabbe and Schulenberg 1997). a recording of the male at the nest was deposited at the Macaulay Li- brary at Cornell University (ML #132525). Young and Zuchowski (2003) erroneously cit- ed a nest of S. simonsi found by T. S. Schu- lenberg in Puno, Peru (Rosenberg 1986). This nest correctly pertains to Diademed Tapaculo (5. schulenbergi) (Krabbe and Schulenberg 2003; T. S. Schulenberg, pers. comm.). The domed nest was similar in structure to other Scytalopus nests; however, use of bunch grasses as construction material has been pre- viously reported only in S. superciliaris (Stiles 1979). Use of bunch grasses is probably an adaptation for living in areas above treeline where it is one of the only nesting materials available. Mosses were available at the site, although in much less quantity than the ex- tremely humid subtropical and temperate mossy forests where the majority of the Scy- talopus nests have been described. This pair of S. simonsi also used a longer tunnel than previously described for nests of the genus; however, the growing number of Scytalopus nest descriptions suggests the genus is ex- tremely variable and opportunistic in selecting a nest site. They apparently use a variety of structures for nest sites and, perhaps, vary nest architecture based upon the individual nest site (Greeney and Gelis 2005, Greeney 2008), rather than stereotyped nest architecture with a genetic basis (i.e., Zyskowski and Prum 1999, Zimmer and Isler 2003. Brumfield et al. 2007) seen in other tracheophone groups. It appears likely the nest burrow was modified from an old rodent excavation and that ro- dents. rather than Scytalopus, constructed the false chamber and nest chambers. Alternative- ly. the shape of the entrance resembled the burrow of Bar-winged Cinclodes (Cinclodes 476 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 fuscus) and it could have been a re-used bur- row of that species. Recently, Greeney (2008) described the first unambiguously re-used nest by a Scytalopus, the Long-tailed Tapaculo, which had taken over a nest built and used to raise three broods by a pair of Spotted Barb- tails, (Premnoplex hrunnescens). Parental care of the chicks was typical for the genus (Greeney and Gelis 2005, Greeney and Rombough 2005, Greeney et al. 2005) with both parents contributing to virtually all activities. The sexual dimorphism in S. simon- si allowed for our novel observations that males also contribute to nest building, and that while the male actively contributed to feeding and brooding, it provisioned slightly less fre- quently and spent less time brooding than the female. The used nest devoid of fecal sacs from the chicks suggests that, like in other nests (Greeney and Gelis 2005, Greeney and Rombough 2005), S. simonsi chicks may ex- crete outside of the nest and the parents keep the nest extraordinarily clean. The appearance of the chicks and unique natal down were typ- ical of Scytalopus (Greeney and Gelis 2005). However, the soft peeping calls of the chicks differed strongly from previous descriptions (Greeney and Gelis 2005) that Scytalopus chick vocalizations were loud and insect or woodpecker-like. Whether this is due to the young age of the S. simonsi chicks or inter- specific differences is unknown. Incubation in S. simonsi appears much longer than that of magellanicus (DeSanto et al. 2002), but exact dates of laying and hatch are unknown. Little is known about the post-fledgling care of Scytalopus and we were unable to ob- serve the birds post fledging. PAH frequently observed a pair of S. parkeri at Reserva Tap- ichalaca in Loja, Ecuador accompanying and provisioning a single fledgling in the same ter- ritory for 3 weeks from mid October through early November 2005. This suggests a long period of post-fledging care by adults in the group, much as in related tracheophone fam- ilies (Zimmer and Isler 2003). Many aspects of Scytalopus breeding be- havior (i.e., shared nest building, shared in- cubation although only the female incubates at night, shared provisioning of young, ex- tended parental care after fledging, and low rates of nest success) appear similar to the other tracheophones. We hypothesize that Scy- talopus are also similar to tracheophones in their monogamy, extended pair bonds, and other aspects of breeding biology yet to be studied or observed in this fascinating group. ACKNOWLEDGMENTS We thank the American Bird Conservancy for fi- nancing the study on Cochabamba Mountain Finch (Compospiza garleppi) during which we made these observations. We also thank Asociacion Civil Armon- la/BirdLife International and the communnities of Pal- capampa and Portrero for allowing access to the study areas and providing support. The Macaulay Library at Cornell University loaned recording equipment. The Herbario Nacional Martin Cardenas assisted in plant identification. A. B. Hennessey, S. K. Herzog, Luis Llanos, Felix Huanca, Israel Huanca, and Josias Huan- ca assisted with logistics and fieldwork. M. B. Robbins provided comments that improved the manuscript. LITERATURE CITED Acros-Torres, a. and A. Solano-Ugalde. 2007. First description of the nest, nest site, eggs, and nestlings of Narino Tapaculo {Scytalopus vici- nior). Ornitologia Neotropical 18:445-448. Brumfield, R. T, J. G. Tello, Z. A. Cheviron, M. D. Carling, N. Crochet, and K. V. Rosenberg. 2007. Phylogenetic conservation and antiquity of a tropical specialization: army-ant following in the typical antbirds (Thamnophilidae). Molecular Phylogenetics and Evolution 45:1 — 13. Christian, D. G. 2001. Nests and nesting behavior of some little known Panamanian birds. Ornitologia Neotropical 12:327-336. Decker, K. L., A. M. Niklison, and T. E. Martin. 2007. First description of the nest, eggs, and breeding behavior of the Merida Tapaculo (5cv- talopus meridaniis). Wilson Journal of Ornithol- ogy 119:121-124. De Santo, T. L., M. F. Willson, K. E. Sieving, and J. J. Armesto. 2002. Nesting biology of tapaculos (Rhinocryptidae) in fragmented south-temperate rainforests of Chile. Condor 104:482—495. Gill, F. and M. Wright. 2006. Birds of the world. Recommended English names. Princeton Univer- sity Press, Princeton, New Jersey, USA and Ox- ford, United Kingdom. Greeney, H. F. 2008. Additions to our understanding of Scytalopus tapaculos reproductive biology. Or- nitologia Neotropical 19: In press. Greeney, H. F. and R. A. Gelis. 2005. The nests and nestlings of the Long-tailed Tapaculo {Scytalopus micropterus) in Ecuador. Ornitologia Colombiana 3:88-91. Greeney, H. F and C. J. F. Rombough. 2005. First nest of the Chusquea Tapaculo {Scytalopus par- keri) in southern Ecuador. Ornitologia Neotropical 16:439-440. Greeney, H. F, A. D. L. Buecker, and N. Harbers. Hosner and Huanca • NESTING OF PUNA TAPACULO 477 2005. Parental care of the Blackish Tapaculo {Scy- talopus latrans) in northeastern Ecuador. Ornito- logia Neotropical 16:283-286. Hennessey, A. B., S. K. Herzog, and F. Sagot. 2003. Lista anotada de las Aves de Bolivia. Quinta Ed- icion. Asociacion Armonia/Birdlife International, Santa Cruz de la Sierra, Bolivia. Hilty, S. L. and W. L. Brown. 1986. A guide to the birds of Colombia. Princeton University Press, Princeton, New Jersey, USA. Johnson, A. W. and J. D. Goodall. 1965. The birds of Chile and adjacent regions of Argentina, Boliv- ia, and Peru. Volume. 2. Platt Establecimientos Graficos S.A., Buenos Aires, Argentina. Krabbe, N. and T. S. Schulenberg. 1997. Species limits and natural history of Scytalopus tapaculos (Rhinocryptidae), with descriptions of the Ecua- dorian taxa, including three new species. Ornitho- logical Monographs 48:47-88. Krabbe, N. and T. S. Schulenberg. 2003. Family Rhinocryptidae (tapaculos). Pages 748-787 in Handbook of the birds of the world. Volume 8. Broadbills to tapaculos (J. Del Hoyo, A. Elliot, and D. Christie, Editors). Lynx Edicions, Barce- lona, Spain. Pulgarin-R, P. C. 2007. The nest and eggs of Spill- mann’s Tapaculo {Scytalopus spillmanni). Ornito- logia Colombiana 5:91-93. Remsen Jr., j. V, A. Jaramillo, M. A. Nores, M. B. Robbins, T. S. Schulenberg, E G. Stiles, J. M. C. DA Silva, D. F. Stotz, and K. J. Zimmer. 2008. A classification of the bird species of South America. American Ornithologists’ Union. http://www. museum.lsu.edu/~Remsen/SACCBaseline.html. Rosenberg, G. 1986. The nest of the Rusty-belted Ta- paculo (Liosceles thoracicus). Condor 88:98. Sick, H. 1993. Birds in Brazil: a natural history. Princeton University Press, Princeton, New Jer- sey, USA. Skutch, a. 1972. Studies of tropical American birds. Publications of the Nuttall Ornithological Club Number 10. Stiles, E. 1979. The nest and eggs of the White- browed Tapaculo {Scytalopus superciliaris). Con- dor 81:208. Young, B. E. and W. Zuchowski. 2003. First descrip- tion of the nest of the Silvery-fronted Tapaculo {Scytalopus argentifrons). Wilson Bulletin 115: 91-93. Zimmer, J. T. 1939. Studies of Peruvian birds 32. The genus Scytalopus. American Museum Novitates. 1044:1-18. Zimmer, K. J. and M. J. Isler. 2003. Family Tham- nophilidae (typical antbirds). Pages 448-681 in Handbook of the birds of the world. Volume 8. Broadbills to tapaculos (J. Del Hoyo, A. Elliot, and D. Christie, Editors). Lynx Edicions, Barce- lona, Spain. Zyskowski, K. and R. O. Prum. 1999. Phylogenetic analysis of the nest architecture of neotropical ov- enbirds (Furnariidae). Auk 116:891-911. The Wilson Journal of Ornithology 120(3):478-493, 2008 PHYLOGEOGRAPHIC PATTERNS OF DIFFERENTIATION IN THE ACORN WOODPECKER MAGALI HONEY-ESCANDON,'^ BLANCA E. HERNANDEZ-BANOS.^ ADOLFO G. NAVARRO-SIGUENZA,! hESIQUIO BENITEZ-DIAZ,- AND A. TOWNSEND PETERSON^ ABSTRACT. Acorn Woodpecker {Melanerpes fonniciyoriis) populations were sampled to evaluate geo- graphic patterns of differentiation and connectivity across the species' range. We observed patterns of differ- entiation generally coincident with geographic patterns in plumage patterns with distinct subpopulations m Baja California Sur, northern Central America, southern Central America, and mainland Mexico north into the south- western United States. We confirmed the existence of geographic genetic structuring of populations of this species, although shared haplotvpes between Baja California Sur and mainland Mexico suggest that lineage sorting is not >it complete. The process of geographic differentiation and speciation is likely still underway m this group. Received 31 January 2007. Accepted 26 October 2007. The montane forests of North and Central America have had a complex history and geo- graphy over the past 100,000 years (Graham 1975, Wells 1983). During the Pleistocene, montane areas, particularly in the northern part of the region, appear to have been largely covered by ice or tundra and. thus, uninhab- itable for forest birds, whereas desert basins filled with what are presently 'montane’ con- iferous forests (Wells 1983). Pleistocene and Holocene climatic shifts must have had im- portant implications for avian biogeography in terms of population connectivity and isolation, and likely affected the species inhabiting these biomes profoundly. Studies ha\e now addressed the climatic and biogeographic implications of Pleisto- cene-Holocene climate shifts (Hugall et al. 2002. Martmez-Meyer et al. 2004. Martmez- Meyer and Peterson 2006, Ruegg et al. 2006), but surprisingly few detailed phylogeographic studies of birds have been conducted to illus- trate how climatic changes and habitat shifts influenced the evolution and differentiation of birds. Only Aphelocoma jays (Peterson 1992. * Museo de Zoologia. Facultad de Ciencias. Univ- ersidad Nacional Autonoma de Mexico. Apartado Postal 70-399. Mexico. D.F. 04510. Mexico. - Comision Nacional para el Uso y Conocimiento de la Biodiversidad. Liga Periferico - Insurgentes Sur 4903, C.P. 14010. D.F. 14010. Mexico. - Natural History Museum and Biodiversity Re- search Center. University of Kansas. Lawrence, KS 66045. USA. Corresponding author; e-mail; behb@hp.fciencias.unam.mx Rice et al. 2003), Sphyrapicus sapsuckers (Cicero and Johnson 1995), MacGillivray s Warbler {Oporornis tolmiei) (Mila et al. 2000), and Hutton's Vireos {Vireo huttoni) (Cicero and Johnson 1992) have been studied in North American pine-oak {Pinus-Qiierciis) woodlands and forests. Patterns of genetic dif- ferentiation and the extent to which they do — or do not — relate to Pleistocene patterns of connection and disjunction of habitats are only beginning to be understood. The objective of this paper is to present the results of molecular genetic studies of 98 in- dividuals from 15 populations of Acorn Woodpeckers {Melanerpes formicivorus) across North and Central America. An earlier contribution based on many of the same sam- ples as in this paper (Benftez-Diaz 1993) iden- tified a series of morphologically distinct pop- ulations with major units including popula- tions in California. Baja California Sur. main- land Mexico, Central America, and Colombia. Samples are lacking to represent the distinc- tive populations of northwestern South Amer- ica. but sampling of the remainder of the dis- tribution of the species is more or less inten- sive. This study, based on sequences of two mitochondrial genes, offers a first view of geographic patterns of genetic differentiation among populations of the Acorn Woodpecker. METHODS Samples and Sequencing. — Samples of muscle, heart, and liver collected from 98 in- dividual Acorn Woodpeckers across most of the species’ range (the distinct Colombian 478 Hoiiey-Escandon et al. • PHYLOGEOGRAPHY OF THE ACORN WOODPECKER 479 populations, and those of lowland areas in Be- lize and the remainder of the Peten region were not included for lack of access to sam- ples; Fig. 1). We included sequences from 10 individuals of seven related species, including Melanerpes lewis, M. aurifrons (3 individu- als), M. uropygialis, M. pygmaeus (2 individ- uals), M. pucherani, Sphyrapicus nuchalis, and (more distantly) Coracias spatulatus (Ap- pendix). These samples were obtained from field collections by several of the authors; full specimen voucher specimens are deposited in the Museo de Zoologia “Alfonso L. Herrera” of the Universidad Nacional Autonoma de Mexico (UNAM), Field Museum of Natural History, and the University of Kansas Natural History Museum, supplemented by tissue samples associated with specimens kindly provided by the Barrick Museum of Natural History (University of Nevada-Las Vegas) and the Museum of Vertebrate Zoology (Uni- versity of Califomia-Berkeley). Data were obtained from GenBank for two outgroup in- dividuals. Total tissue DNA was extracted via DNEa- sy Extraction Kits (Qiagen, Valencia, CA, USA). Specific fragments were amplified via polymerase chain reaction (PCR) using prim- ers spanning 334 bp of the mitochondrial gene ND2 segment (L5215 TAT CGG GCC CAT ACC CCG AAA AT; H5578 CCT TGA AGC ACT TCT GGG A AT CAG A) (Hacked 1996) and a 608 bp fragment of the cytochrome b gene (LI 54 13 CTG AC A AAA TTC CAT TTC ACC C; HI 6064 CTT CAG TTT TTG GTT TAC AAG ACC) (Kocher et al. 1989 and Sorenson et al. 1999, respectively). All numbers refer to the 3-prime end of the primer reference of the complete mtDNA sequence of the domestic chicken (Gallus gallus) (Desjar- dins and Morais 1990). A typical ND2 amplification involved 35 cycles of 95° C for 1 min, 48° C for 2 min, 72° C for 3 min, and a final 10 min extension period at 72° C. Cyih amplification involved 27 cycles of 94° C for 1 min, 50° C for 1 min, and 72° C for 2 min, followed by a 7 min extension period at 72° C. PCRs were con- ducted on a GeneAmp PCR System 9700 (Ap- plied Biosystems, Foster City, CA, USA). Products were verified on a 1% agarose gel with added ethidium bromide and cleaned us- ing a QiaQuick Kit (Qiagen, Valencia, CA, USA), obtaining a final volume of 15-30 p.L of PCR products. We purified PCR products using Gene- clean® (Qbiogene, BiolOl® Systems, Krack- eler Scientific Inc., Albany, NY, USA) and Millipore purification kits following manufac- turers’ protocols. Purified PCR products were sequenced on a Perkin-Elmer ABI 373 auto- matic sequencing machine. Sequences were cleaned using Chromas 1.45 (McCarthy 1996), and aligned using ClustalX (Thompson et al. 1997). We corroborated the origin of our sequences by combining at least two of the following: amplifying overlapping gene seg- ments, sequencing both DNA strands, and/or using multiple individuals of single popula- tions. Statistical Analyses. — We used MEGA 2.0 (Kumar et al. 2004) to derive basic statistics regarding sequences, and their variation and diversity. We used Arlequin (Schneider et al. 2000) to calculate Nei’s pairwise differences (raw distances corrected following Nei [1987]) among populations, as well as F^, val- ues. DnaSP Version 4.10 (Rozas et al. 2003) was used to calculate nucleotide diversity (it) and haplotype diversity (k). We used TCS Version 1.13 (Clement et al. 2000) to estimate networks summarizing mutational differences among haplotypes. We compared matrices of Nei’s corrected genetic distances with matri- ces of straight-line geographic distances sep- arating populations using a Mantel test; we plotted the ratio of genetic to geographic dis- tances on maps to visualize spatial patterns of genetic differentiation on a per kilometer ba- sis. Only informative characters and unique haplotypes were used for parsimony searches using Coracias as the only designated out- group to avoid problems of non-monophyly ot in-group taxa. Maximum parsimony trees were constructed for ND2 and cyih sequences both separately and combined, using heuristic search options in PAUP 4.0 (Swofford 1999) with TBR and ACCTRAN optimization op- tions. We used character-based bootstrap anal- ysis (100 replicates) to estimate support for each node in the resulting tree. ModelTest 3.0 (Posada and Crandall 1998) was used to identify appropriate models of se- quence evolution for haplotypes of Melaner- pes forniicivorus. Bayesian inference (BI) ap- 480 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 FIG. 1. Geographic distribution of Melanerpes formicivorus (shaded) showing localities where samples were obtained. proaches, as implemented in Mr.Bayes Ver- sion 3.1 (Huelsenbeck and Ronquist 2001) used the substitution model GTR (nset = 6) for the number of rate parameters and a gam- ma distribution for rates at each site. We ran four Markov Chains (random starting trees) for 10^ generations, each sampling every 250 generations and identifying stationarity visu- ally. We allowed an initial “bum-in” of 250 trees to avoid non-optimal solutions and com- puted a majority-mle consensus tree, as well as posterior probabilities for each node (Huel- senbeck et al. 2002). RESULTS Genetic Variation. — We obtained a total of 942 base pairs across the two genes. Of these sites, 586 were conserved, 356 were variable, and 231 were parsimony-informative. The transition/transversion ratio was 3.4 and nu- cleotide composition was T = 0.26, C = 0.36, A = 0.27, and G = 0.11. Nucleotide diversity was 0.00482 and haplotype diversity was 0.851 with lowest nucleotide diversity values in Baja California Norte and Oaxaca popula- tions. Overall, we found 44 haplotypes (Fig. 2) among the 98 sequences that were distin- guishable by 66 polymorphic sites. Almost all (41) haplotypes were restricted to single pop- ulations. Haplotype H4 was present in single individuals from population samples from Guerrero and Hidalgo, Mexico; haplotype H35 occurred in seven individuals from Baja California Norte, Mexico, and California, USA, and (most impressively) H29 was found in 33 individuals from 10 localities from Ar- izona south to Honduras. Pairwise average population differences (Table 1) ranged from 0 to 8.07 within the Arizona sample, and 0 to 8.35 in the Baja Cal- Honey-Escandon et al. • PHYLOGEOGRAPHY OF THE ACORN WOODPECKER 481 r > Chiapas Baja California Norte, California FIG. 2. Haplotype network of the 44 haplotypes of Melanerpes formicivorus. Mutational steps are indicated by the number of line segments connecting haplotypes. The size of the ovals represents the number of .samples with that haplotype; shaded ovals are haplotypes with more than one .sample. ifornia Norte versus Baja California Sur, Mex- ico samples. The overall F^, statistic for the species was 0.484. Pairwise values between population samples ranged from 0 (several population pairs) to >0.7, most related to the Baja California Sur and Baja California Norte populations and, to a lesser extent, with the Central American populations. The haplotype network (Fig. 2) had several features. One haplotype (H29) was common, occurring in about one-third of all individuals. Closely associated to this haplotype were 22 other haplotypes that differed by <3 muta- tions from H29; overall, this group of haplo- types generally corresponds to populations of mainland Mexico and Arizona (with one rep- resentative from as far south as Honduras). Closely associated to the mainland Mexico 482 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 'C ir-, O' 'C 2C IT', O' ir-, >c .w ^ tr-, <2 yz — 1 — c ' ' Y > K 'T- zc fN > O' O' 'V', ir^. t\ > r' 2C 0 m. y— — -Z 'T ir-, > > > '> t\ K 2c yz “ — " — > t\ K <2 'C 'C wm — ^ 'V^ K O' O' t\ s p 'T- 'T- N. ^ 5j :c ZC ZC 'C N. I\ O' ' 'C C ^ t\ 'C > (N X t\ fS rsi O^ ir-. yz >Z X t\ — ly^, 0 ■= " 2C :c i\ l\ rx ri 'C — 0 - = ' — ' ' n: u - iC «v^ 'C tr-, (N fS — 'C z O' 0 (N yz o — "l- > — ^ C; d 0 C: 0 d Iz t\ -y. t\ > ir, Cc > > > i ‘ . Zc O' 'T- (N ir, K fN 0 'r > > 2C ir, nC IT', > • — r ^ 22 t\ t\ t\ r\ t\ — O' 0 i 1 '5 y, 'C d O " N, rs) 3 3 :< c d X d c 't ".l| rj y6 r^. 5C X rJ O' Os O' IT', S C: 0 ” _ 'C c ^ 0 ^ 0 — y r<~, X — ■ (N X — . :c > , ^ ir# r- t\ K K — ^ v: z B' 5 T T ~ T T r< ■ ^ 1 - IT( _ (N ■ ^ ()!)(! X IT, X S s t\ d 5 '^, 500 equally parsimonious 532-step trees (Cl = 0.594, RI = 0.812; Fig. 4). These trees grouped all Melanerpes fonnicivorus populations as a monophyletic group with high bootstrap support (100% of bootstrap replicates). Subclades corresponding to indi- viduals from Baja California Sur (84% sup- port), northern Central America (Chiapas, Honduras; 80% support), and southern Central America (Costa Rica; 73% support) were found within this clade, although none had solid branch support in the bootstrap analyses. The remaining individuals in the study were grouped in one large, but poorly supported clade (51% bootstrap support) of individuals from mainland Mexico, Arizona, California, and Baja California Norte. Two individuals (from Guerrero and Zacatecas) were not con- nected with any of the subclades within the species, one Baja California Sur individual (haplotype H36) grouped with the mainland Mexico assemblage, and one Arizona individ- ual (haplotype Hll) grouped with the Chia- pas-Honduras clade. The BI analyses were based on the TVM + I + G model of substitution and showed a to- pology (Fig. 5) generally close to that of the MP tree. The clade corresponding to all Melanerpes fonnicivorus populations was well-defined and subclades with intriguing but inconclusive con- stimtion were encountered. In particular, we re- covered the Baja California Sur (0.97 posterior probability), northern Central America with the single Arizona sample (0.99 posterior probabil- ity), and southern Central America (0.78 pos- terior probability) nodes. We encountered a weakly supported node corresponding to the California and Baja California Norte samples (0.63 posterior probability); the mainland Mex- ico and Arizona and single Baja California Sur samples formed a large and poorly-defined as- semblage. DISCUSSION An earlier morphological analysis (Benitez- Diaz 1993), in many cases of precisely the same individuals as were analyzed in this study, found marked subdivision of the spe- cies into seven groups, two of which (Belize and Colombia) were not analyzed in this study. These groups were supported by the distribution of genetic variation found in our study, albeit not strongly or with marked ge- netic differentiation. Recalculating statis- tics hierarchically, we found that 73.7% of overall genetic variation was assorted among these five groups, as opposed to 26% within them, suggesting these groups have explana- tory power regarding population differentia- tion in the overall complex. The five groups included in this study, with one exception, were distinct from one another in terms of mutational steps in a haplotype network. The exception was that of the Cali- fornia/Baja California Norte populations, which, although they grouped together, were only one mutational step from the mainland Mexico haplotype group. Other groups were more distinct; each was >6 mutational steps removed from all other groups. Thus, the hap- lotypes of the plumage-based groups appear to differ markedly from group to group. Given the high number of unique haplotypes, addi- tional sampling may prove necessary for the details of the situation to be completely clear. 484 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 FIG. 3. Geographic patterns of genetic differentiation; (top) relationship between genetic distance and geo- graphic distance based on 15 collecting localities; (bottom) map of genetic distance/km illustrating patterns of genetic connectivity among Acorn Woodpecker population samples. Thick continuous lines indicate rates of < 1 genetic distance unit/km, thin continuous lines indicate rates of 1—5 genetic distance unit/km, and thin broken lines indicate rates of >5 genetic distance unit/km. Our results clearly indicate lack of full es- tablishment of reciprocal monophyly among the various populations in spite of the overall picture of differentiation. This muted differ- entiation is visible in both the relatively un- resolved and poorly supported trees that were recovered, and in the mixture of one Baja Cal- ifornia Sur haplotype among the “mainland Mexico” haplotypes and the presence of one (H29, the most common haplotype) in Hon- duras in both the phylogenetic analyses and the haplotype network. Honey-Escandon et al. • PHYLOGEOGRAPHY OF THE ACORN WOODPECKER 485 51 H44 H24 H26 H28 H3 H5 H43 H14 H15 H16 H40 H13+ H29* H33 H36 H34 H4+ H27 H35+ H25+ H39 H30+ Mexico, Arizona, California, Baja California Norte Costa Rica Chiapas, Honduras Baja California Sur Guerrero Zacatecas M. aurifrons M. aurifrons M. aurifrons M. uropygialis M. pygmaeus M. pygmaeus M. pucherani Sphyrapicus nuchalis Coracias spatulatus FIG. 4. Maximum parsimony tree (50% majority rule consensus) of the 44 haplotypes of Acorn Woodpeckers and lO outgroup samples. Numbers on branches indicate bootstrap support. + = haplotypes with more than one sample; * = the most common haplotype (which was represented in a single sample from Honduras). Note that some branches have relatively low bootstrap support and may not be robust hypotheses of relationships. 486 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 1.00 ii35+ _ H35+ _ H27 _ H26 _ H24 H28 H43 H16 HI 4 HI 5 HI 2 Arizona H41 Michoacan H17+ Jalisco California Norte California Michoacan Jalisco M. pucherani 0.99 [ 1.00 1 M. ova 0.99 -M. aurifrons -M. aurifrons 1.00 M. pygmaeus Coracias spatulatus FIG. 5. Bayesian inference tree of the 44 haplotypes and 10 outgroups. Numbers below the branches show the values of the posterior probability of each branch. + = haplotypes with more than one sample; * = the most common haplotype (which was represented in a single sample from Honduras). Note that some of the branches have relatively low probabilities associated and may not be robust hypotheses of relationships. Benftez-Dfaz (1993) documented the exis- tence of seven subgroups within Melanerpes formicivorus on the basis of external pheno- type. These groups should be considered for formal taxonomic recognition (Navarro and Peterson 2004), at least under the Evolution- ary Species Concept (Wiley 1978) and the Phylogenetic Species Concept (Zink and McKitrick 1995, Zink 1996). Benftez-Dfaz (1993) recommended recognition of Melaner- pes bairdi of California and Baja California Norte, M. angustifrons of Baja California Sur, M. formicivorus of the southwestern United States and mainland Mexico, M. lineatus of northern Central America, M. striatipectus of southern Central America, M. albeolus of Be- lize, and M. flavigula of Colombia. M. albeo- lus and M. flavigula were not available to us for molecular analysis and we did not find marked differentiation between populations in California and Mexico. Hence, we focus at- tention on M. formicivorus (including Califor- nia populations of the bairdi group), M. an- gustifrons, M. lineatus, and M. striatipectus in the rest of our discussion. Benftez-Dfaz’s (1993) general picture of dif- ferentiation of Acorn Woodpecker populations was supported, but decisions regarding species limits were less clear. From the perspective of the Biological Species Concept (AOU 1998), these populations can be interpreted either as (1) exchanging few genes after a relatively recent separation, or (2) still exchanging genes (which may cause the intermixing of haplotypes), which would probably point to caution in split- ting populations under this concept. The Phy- logenetic Species Concept would clearly rec- ognize these different forms as species in view of their distinctiveness in plumage, but would hold back from recognition using molecular Honey-Escandon et al. • PHYLOGEOGRAPHY OF THE ACORN WOODPECKER 487 characters on the basis of intermixing of hap- lotypes from Baja California Sur, mainland Mexico, and Central America. Finally, under the Evolutionary Species Concept, one would most likely accord them species status, given that not only are populations apparently in the process of diverging, but unique phenotypic characters are now fixed in at least some populations. The patterns of genetic variation and dif- ferentiation identified would appear to corre- spond closely to known Pleistocene geogra- phy of pine-oak woodlands at the Last Glacial Maximum (LGM, ca. 20,000 years ago). That is, at LGM, montane woodlands moved on large spatial scales, broadly invading the southwestern North American deserts (Banner and Van Devender 1981, Spaulding et al. 1983, Wells 1983). The major zones of ge- netic differentiation in Melanerpes formici- vorus are between the southern tip of Baja California and the Califomia/Mexico portion of the range, and across the Isthmus of Te- huantepec— the lack of differentiation across the Mohave Desert may reflect the Pleistocene connectivity of populations of this species. ACKNOWLEDGMENTS We thank our companions (particularly Noe Vargas- Barajas, Scott Baker, Patricia Escalante, and Laura and Fernando Villasenor-Gomez, among many others) for invaluable assistance and support during field collec- tions. Colleagues at the Museum of Vertebrate Zool- ogy and the Barrick Museum of Natural History kindly provided samples to add to our set of specimens; Ed- uardo Morales (Instituto de Historia Natural y Ecolo- gia) kindly provided samples from Chiapas. Funding was provided by the National Geographic Society and the U.S. National Science Foundation. The laboratory research was supported by DGAPA IN-208700 and lN-2 11407, SEMARNAT-CONACYT Sectorial Fund CO 1-0265, and a CONACYT scholarship to MHE. We thank Gabriela Garcia-Deras, Nandadevi Cortes-Rod- riguez, and Laura Marquez Valdelamar for technical support. LITERATURE CITED American Ornithologists LInion (AOU). 1998. Check-list of North American birds. Seventh Edi- tion. American Ornithologists’ Union, Washing- ton, D.C., USA. Benitez-Di'az, H. 1993. Geographic variation in col- oration and morphology of the Acorn Woodpeck- er. Condor 95:63-71. Cicero, C. A. and N. K. Johnson. 1992. Genetic dif- ferentiation between populations of Hutton’s Vir- eo (Aves: Vireonidae) in disjunct allopatry. South- western Naturalist 37:344-348. Cicero, C. and N. K. Johnson. 1995. Speciation in sapsuckers {Sphyrapicus). III. Mitochondrial- DNA sequence divergence at the cytochrome-b locus. Auk 112:147-163. Clement, M., D. Posada, and K. A. Crandall. 2000. TCS: a computer program to estimate gene ge- nealogies. Molecular Ecology 9:1657-1659. Desjardins, P. and R. Morais. 1990. Sequence and gene organization of the chicken mitochondrial genome. Journal of Molecular Biology 212:599- 634. Graham, A. 1975. Late Cenozoic evolution of tropical lowland vegetation in Veracruz, Mexico. Evolu- tion 29:723-735. Hackett, S. J. 1996. Molecular phylogenetics and bio- geography of tanagers in the genus Ramphocelus (Aves). Molecular Phylogenetics and Evolution 5: 368-382. Huelsenbeck, J. P. and F. Ronquist. 2001. MRB AY- ES: bayesian inference of phylogenetic trees. Bioinformatics 17:754-755. Huelsenbeck, J. P, B. Larget, R. E. Miller, and F. Ronquist. 2002. Potential applications and pitfalls of Bayesian inference of phylogeny. Systematic Biology 51:673-688. Hugall, a., C. Moritz, A. Moussalli, and J. Stan- isic. 2002. Reconciling paleodistribution models and comparative phylogeography in the wet trop- ics rainforest land snail Gnarosophia bellendenk- erensis (Brazier 1875). Proceedings of the Nation- al Academy of Sciences USA 99:61 12—61 17. Kocher, T. D., W. K. Thomas, A. Meyer, S. V. Ed- wards, S. PaABO, E X. ViLLABLANCA, AND A. C. Wilson. 1989. Dynamics of mitochondrial DNA evolution in animals: amplification and sequenc- ing with conserved primers. Proceedings of the National Academy of Sciences USA 86:6196- 6200. Kumar, S., K. Tamura, and M. Nei. 2004. MEGA3: integrated software for molecular evolutionary ge- netics analysis and sequence alignment. Briefings in Bioinformatics 5:150-163. Lanner, R. M. and T R. Van Devender. 1981. Late Pleistocene pinon pines in the Chihuahuan Desert. Quaternary Biology 15:278-290. MartInez-Meyer, E. and A. T. Peterson. 2006. Con- servatism of ecological niche characteristics in North American plant species over the Pleisto- cene-to-recent transition. Journal of Biogeography 33:1779-1789. Martinez-Meyer, E.. A. T. Peter.son. and W. W. Hargrove. 2004. Ecological niches as stable dis- tributional constraints on mammal species, with implications for Pleistocene extinctions and cli- mate change projections for biodiversity. Global Ecology and Biogeography 13:305—314. McCarthy, C. 1996. Chromas 1.45. Griffith Univer- sity, Southport, Queensland, Australia. Mila, B.. D. J. Girman, M. Kimura, and T. B. Smith. 488 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 2000. Genetic evidence for the effect of a post- glacial population expansion on the phylogeog- raphy of a North American songbird. Proceedings of the Royal Society of London Series B 267: 1033-1040. Navarro, A. G. and A. T. Peterson. 2004. An alterna- tive species taxonomy of the birds of Mexico. Biota Neotropica 4 (2). http://www.biotaneotropica. org.br/v4n2/pt/abstract?taxonomic-review-(- BN03504022004. Nei, M. 1987. Molecular evolutionary genetics. Co- lumbia University Press, New York, USA. Peterson, A. T. 1992. Phylogeny and rates of molec- ular evolution in the jays of the genus Apheloco- ma (Corvidae). Auk 109:134-148. Posada, D. and K. A. Crandall. 1998. Modeltest: testing the model of DNA substitution. Bioinfor- matics 14:817-818. Rice, N. H., E. Martinez-Meyer, and A. T. Peterson. 2003. Ecological niche differentiation in the Aphelocoma jays: a phylogenetic perspective. Bi- ological Journal of the Linnaean Society 80:369- 383. Rozas, J. J., C. Sanchez-DelBarrio, X. Messeguer, AND R. Rozas. 2003. DnaSP, DNA polymorphism analyses by the coalescent and other methods. Bioinformatics 19:2496-2497. Ruegg, K. C., R. j. Humans, and C. Moritz. 2006. Climate change and the origin of migratory path- ways in the Swainson’s Thrush, Catharus ustula- tus. Journal of Biogeography 33:1172-1182. Schneider, S., D. Roessle, and L. Excoffier. 2000. Arlequin Version 2.000: a software for population genetics data analysis. University of Geneva, Ge- neva, Switzerland. Sorenson, M. D., J. C. Ast, D. E. Dimcheff, T. Yuri, AND D. P. Mindell. 1999. Primers for a PCR- based approach to mitochondrial genome se- quencing birds and other vertebrates. Molecular Phylogenetics and Evolution 12:105-114. Spaulding, W. G., E. B. Leopold, and T R. Van De- vender. 1983. Late Wisconsin paleoecology of the American Southwest. Pages 259-293 in Late Quaternary environments of the United States. The late Pleistocene (S. C. Porter, Editor). Volume 1. University of Minnesota Press, Minneapolis, USA. SwoFFORD, D. L. 1999. PAUP* Phylogenetic analysis using parsimony. Version 4.0. Sinauer Associates, Sunderland, Massachusetts, USA. Thompson, J. D., T J. Gibson, E Plewniak, F. Jean- MOUGiN, AND D. J. HiGGiNS. 1997. The ClustalX windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Research 24:4876—4882. Wells, P. V. 1983. Paleobiogeography of montane is- lands in the Great Basin since the last glacioplu- vial. Ecological Monographs 53:341-382. Wiley, E. O. 1978. The evolutionary species concept reconsidered. Systematic Zoology 27:17-26. Zink, R. M. 1996. Species concepts, speciation and sexual selection. Journal of Avian Biology 27:1-6. Zink, R. M. and M. C. McKitrick. 1995. The debate about species concepts and its implications for or- nithology. Auk 112:701-719. Honey-Escandon et al. • PHYLOGEOGRAPHY OF THE ACORN WOODPECKER 489 (N bb T3 C o a c/2 (U fc o o (U a "a X 3 1 E , g zz i s e s ■X O C/T c« .-H 3 c3 3 3 33 C C 3 3 0^0 W Z Z g « 6 S ^ ^ ^ t ^ (U 2 « . t w S ^ ^ (U OO >-7 flJ lU T}- O 04 — ' rn r- O' 04 04 04 04 CO Z >< X Z X s s s s s Os fN IT) 10, 103 >n O' 04 — r/'i fO m ro ro X X XX XXX vC — 04 — Tt 3- sD 00 00 sO O', O', 10, ON O' O', O', O', XXX z z z s s s U- U, U- 04 00 O' sD 10, sD ON 10, O', O', X X Z Z u. u. — 04 O', 3- „ I I I I I Z Z Z Z C>5 u u o u u CQ CD 02 03 OQ >-, C, c« 3 C O 00 3 Td 3 rn o ■J”*- X 3 CD 3 00 3 3 CD ■J” hJ 3 fc Cc03t3 003Td 003-0 003T3 3 CD cd 3 CD rP 3 PQ 3 pg O P J ^ F J 01 r^t 1 3- 10, 1 nC 1 c/5 1 CO GO 1 GO C/5 U u u u U CD CD CD CD 02 00 ON ro ■3- in r-' 00 00 00 00 00 00 3 S O' r- r- t'' o> NO in in m in. in in m. On ON O' On On ri-. ro O', r*-. < < (U a Q/ nc W) o 3 to X 'X CO X ^ — S 2 ^ a c 3 o o Z (U ^ X CO U W) O O O a ^ £ OJ ^ CO CO 3 3 CO CO 3 3 u u 3 > CO (U g 3 . X X a >.z = £ ^ oo ' 2 3 3 3 3 3 3 3 3 (U CO '£ ’£ ’£ '£ '£ '£ ’£ ’£ CO CO CO 1 1 1 1 ,o X ^4-H 1 a a X 3 3 CL 3 O I—) a a a a a a a a X X X 3 u u u u u u u u u u u GO O O O Oi Qt, at, C2, Di .£ .£ .£ 3 3 3 3 3 3 3 3 3 ^ ^ ^ C/5 CO to to CO C/5 to to ^ OJ OJ Q 0 0 O O O o 0 O S S S u u U u U u u u u m X 00 T+ X 00 On o (N ro O O O X (N (N (N (N E2 ^2 E2 Q 5 5 5 5 Q Q Q Q X X X m CQ CQ m M CQ PQ CQ CQ U U U X X X X X X X X X X (N DC in (N X in m DC in (N X (N (N DC X 00 X ^ DC X K X 00 00 X X li (N O in (N 00 —I On 00 r- N N > > s s X 00 ON o (N m o o o fO m ro m ro m ro 'Ct 3- 3- 3- C\ ON ON ON ON ON ON ON cn m m m CO ro m m > > > > z h C C c3 ^ c/; > ^ 2 S 5 e2 Tj- vC oc O' sO sC O' fO (O r^, r- rsi rsi sC oc oc OC vC sC oc 'C 'O r". r<". (^, VO, VO, sC sC r<~, VO, VO, 3s O' O' O' O' VO, VO, O' r«~. r<~. r*~. r^, r<~. r<~. r«~. r*', ro I I I I I I I I I I z z z z z z z z z z X X X X X X X X X X 'C C/3 S w £££££££-S ir-, IT) (n to VO lo _r _r _r — - — - — — d ococooot; C C C C C _C _C cd '.£ '.£ '.£ ■.£ !£ 1£ OOCJUUOCJO yccccccct ^C3_rtc3c^?3c«c35 hhPFhhhu UJ £ £ £ o lo ^ — — ^ C ' 'C^ 1 ° E U (U . X PJ t/: tLl ^ cti 2 g z ^ fTl c 'C3 D- o U OJ o c K3 pQ pj £22 fc fc fc « (U (U 3 D 3 o a o o o c _u o _o 'S ’>< (U (U (U s s ^ O' >o (N O' o _ ■3' >0 O r*^ 'Vt X >< X S S ^ 2 CD CQ S ± O' Csl 3" rv". O' (N r', rr', O' (N O' (N O' (N •3- O' (N O' (N O' O' rsi (N (N I I I I I I I I I I III lO, (N OS O' o r- Tt rf lo O' O' O' (O r^-, XII z z z 2 s lJ.Ll.uU (N r*-. 1 VO, vC r-- _ (N 1 r^j ■O' vr-, sC 1 ) 1 6 o 6 1 o 6 6 6 6 1 O 6 o o o o a o a o o o CiL a: c^ DC DC DC G Q Q G G G G O O a o a o oooooooo ■t3-0'T3-OT3'OT3t3 I I I I I I I I coococoo ^ ^ .£ .£ .£ .£ ‘>< ’>< '>< ’>< (Uo in 5d •C H S ^ S c (U D- I = D cu § c y o 5 ^ ^ c2 N On \o in m i:q (N O in o o o o o ro O o in (n "Ct m 00 'Ct m o in m cn o 'Ct m m m 'Ct m 'Ct o \D m P3 Q CQ Q CQ Q CQ Q CQ Q CQ Q CQ Q CQ Q PQ Q OQ >< z 1— s Z 1— i X X X X 1—5 X i— i X Q Q X X X X X X X X 2 . 496 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 TABLE 2. Characteristics of granaries 2004-2005. (jc ± SD, /7 = 96). of Acorn Woodpeckers at La Primavera Forest, Jalisco, Mexico, Species used as granaries Number of granaries Tree condition (live/dead) Average DBH (cm) Acorn species stored Pinus oocarpa 95 91/4 39.23 ± 13.25 Quercus viminea Q. castanea Q. laeta Quercus castanea 1 1/0 32 Q. castanea Q. laeta mature acorns of Q. magnoliifolia and Q. re- sinosa were consumed most frequently but their consumption diminished after start of the rainy season. Acorns of these oaks are large (Table 1), and woodpeckers pierced and ate the central part of the seed. We did not find acorns of these species in granaries. We also observed individual Acorn Woodpeckers for- aging on immature acorns and, possibly, ob- taining insects and grit on the ground. Acorn Woodpeckers stored acorns from late August to early March and of primarily three species; Q. viminea, Q. castanea, and Q. laeta which are characterized by acorns of small size (Table 1). Qiiercus gentry i also had a small acorn but was not found in granaries, probably because of its lower density in the forest and low availability of acorns. Acorn Woodpeckers increased their consumption of stored acorns from November to February, which continued into April (Fig. 1). Wood- peckers consumed butterflies (Lepidoptera) (captured by hawking) and beetle (Coleoptera) larvae (from the bark of trunks of P. oocarpa) (20-33% of total time), fed on sap (25-32%), and consumed stored acorns (10-25%) during the hot dry season (Mar to May) (Fig. 1). We found nests of Acorn Woodpeckers with young before fledging from mid June to late July at La Primavera Forest. We observed me- dium size groups (6—8 individuals) and one group of 10 individuals after early July. These groups regularly contained three or four young. We observed several smaller groups (2-3 individuals) in the area from late Feb- ruary to early March. These medium size groups maintained granaries, mainly in live trunks of pine trees (Table 2). Granaries in the southwestern portion (oak-pine forest) of the forest were filled with acorns of Q. castanea and Q. laeta (n ^ 32) while granaries in the remainder of the forest were filled with acorns of Q. viminea (n = 64). Granaries with acorns of Q. castanea and Q. laeta had a small num- ber of holes, and acorns of Q. viminea were stored more often in tree cracks than in gra- naries. The average (± SD) number of holes per granary was 1,267.11 ± 1,427 with 853.71 ± 1,068.81 containing acorns {n = 96). DISCUSSION Acorn Woodpeckers in La Primavera Forest had a diverse diet similar to those reported for Acorn Woodpeckers in North (MacRoberts 1970, Stacey 1981) and Central America (Kat- tan 1988). Unlike what has been reported in Colombia (Kattan 1988), we did not observe Acorn Woodpeckers consuming fruits other than acorns, or feeding on flower nectar. These results suggest that Acorn Woodpecker diets were greatly affected by the abundant sources of acorns. Seasonal availability and use of resources by Acorn Woodpeckers de- creases from temperate to tropical regions. Sap is greatly consumed in spring, insects in summer, and acorns during fall and winter in North America (MacRoberts 1970). Acorn Woodpeckers in La Primavera Forest simul- taneously use several food sources seasonally while in Central America, the diet is more di- verse because more food sources are available (Kattan 1988). The number of oaks (5 species) used by Acorn Woodpeckers in La Primavera Forest is larger than reported in North and South Amer- ica (MacRoberts 1970, Stacey and Jansma 1977, Stacey 1981, Kattan 1988, Stanback 1989, Koenig and Benedict 2002, Arsenault 2004). Our results revealed that immature and mature acorns are an important component of the diet of this woodpecker. This may allow Rosas-Espinoza et al. • ACORN WOODPECKER DIET IN MEXICO 497 the Acom Woodpecker to be resident in La Primavera Forest as a large number of oak species increases the probability of finding sufficient food resources, even if a crop failure occurs in a particular species of oak (Bock and Bock 1974, Koenig and Haydock 1999). There is decreasing use of acorns in the diet (Kattan 1988, Wong 1989) of Acom Wood- peckers from North to South America, be- cause this food item may be replaced by other available resources including fleshy fmit, nec- tar, and flower catkins as in Colombia (Kattan 1988). This may be related to the decreasing number of filled holes in granaries from North to South America. Trees contain thousands of holes in temperate zones (MacRoberts 1970), but this number decreases in tropical areas such as La Primavera Forest (x — 1,267 holes) to 16 to 20 in Belize (Stacey 1981) with no storage of acorns in the Colombian Andes (Kattan 1988). The long period during which acorns are available for harvesting in La Pri- mavera Forest suggests that importance and necessity of acom storage is decreased com- pared to more temperate areas. Piercing and eating of the central part of large acorns with- out storage has also been reported by Stan- back (1989) in Costa Rica with Q. costaricen- sis. This increases the foraging flexibility of Acom Woodpeckers (Stanback 1989). MacRoberts and MacRoberts (1976) and Stacey and Bock (1978) suggested that group size of Acom Woodpeckers is large in areas with high numbers of species of oaks. How- ever, we observed medium size groups of Acom Woodpeckers in La Primavera Forest. We expected groups at least of medium size given the importance of acorns in their diet, and given that group size is related to main- tenance of granaries (MacRoberts and MacRoberts 1976, Stacey 1979). Acorn Woodpeckers at La Primavera Forest used live pines {P. oocarpa) for granaries, mainly in the surfaces of trunks or in pre-ex- isting tree cracks. This may be the result of the availability of live pines and their life ex- pectancy compared with snags. Use of pine trees as granaries has also been reported in Durango, Mexico where granaries were con- structed in P. engelmannii with a similar DBH (37.4 ± 10.6 cm) (Koenig and Williams 1979) as in our study area. Koenig and Benedict (2002) reported the three species of oaks present at the Hastings Reservation in central coastal California had differential use as granaries. In contrast, not all oaks ( 1 1 total species) were used as gra- naries in La Primavera Forest. Acom Wood- peckers appear to differentially and selectively use five oak species. This could be a conse- quence of tree density, acom production by species, size and nutrient quality, and content of secondary metabolites in acorns (Koenig and Benedict 2002). ACKNOWLEDGMENTS We thank Manuel Anguiano-Santana, Luz Maria Hi- nojosa-Medina, and Aleyda Barraza for field assis- tance, and the Executive Committee of La Primavera Forest for assistance with logistics. We thank Cecilia Neri-Luna for improving the manuscript. LITERATURE CITED Arsenault, D. P. 2004. Differentiating nest-sites of primary and secondary cavity-nesting birds in New Mexico. Journal of Field Ornithology 75: 257-265. Benitez-Diaz, H. 1993. Geographic variation in col- oration and morphology of the Acom Woodpeck- er. Condor 95:63-71. Bock, C. E. and J. H. Bock. 1974. Geographical ecol- ogy of the Acom Woodpecker: abundance vs. di- versity of resources. American Naturalist 108: 694-698. Henshaw', H. W. 1921. The storage of acorns by the California Woodpecker. Condor 23:109-118. Kattan, G. 1988. Food habits and social organization of Acom Woodpeckers in Colombia. Condor 90: 100-106. Koenig, W. D. and L. S. Benedict. 2002. Size, insect parasitism, and energetic value of acorns stored by Acorn Woodpeckers. Condor 104:539-547. Koenig, W. D. and J. Haydock. 1999. Oak, acorns, and the geographical ecology of Acorn Wood- peckers. Journal of Biogeography 26:159-165. Koenig, W. D. and R. L. Mumme. 1987. Population ecology of the cooperatively breeding Acorn Woodpecker. Princeton University Press, Prince- ton, New Jersey, USA. Koenig, W. D. and P. L. Williams. 1979. Notes of the status of the Acorn Woodpeckers in central Mex- ico. Condor 81:317-318. Koenig, W. D., P. B. Stacey, M. T. Stanback, and R. L. Mumme. 1995. Acorn Woodpecker {Melaner- pes formicivorus). The birds of North America. Number 194. Ligon, j. D. and P. B. Stacey. 1989. On the signifi- cance of helping behavior in birds. Auk 1()6:7(H)- 705. MacRoberts, M. H. 1970. Notes on the food habits and food defense of the Acorn Woodpecker. Con- dor 72:196-204. 498 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 MacRoberts, M. H. and B. R. MacRoberts. 1976. Social organization and behavior of the Acorn Woodpecker in central coastal California. Ornitho- logical Monographs 21:1-115. Olmo, G. B. 2005. Criterios e indicadores para la res- tauracion del area de proteccion de flora y fauna La Primavera, Jalisco, Mexico. Thesis. Universi- dad de Guadalajara, Mexico. Reyna, B. O. F. 2004. Arboles y arbustos del bosque La Primavera, guia ilustrada. Universidad de Gua- dalajara, CONABIO, Mexico. Ritter, W. E. 1921. Acorn-storing by the California Woodpecker. Condor 23:3-14. SEMARNAT (Secretaria de Medio Ambiente y Re- CURSOS Naturales). 2000. Programa de manejo area de proteccion de flora y fauna La Primavera. Subdireccion. General de Conservacion y Manejo de Areas Naturales Protegidas. CONANP, D. E, Mexico. Skutch, a. 1969. Life-histories of Central American birds. III. Pacific Coast Avifauna 35. Stacey, P. B. 1979. Habitat saturation and communal breeding in the Acorn Woodpecker. Animal Be- havior 27:1153-1166. Stacey, P. B. 1981. Foraging behavior of the Acorn Woodpecker in Belize, Central America. Condor 83:336-339. Stacey, P. B. and C. E. Bock. 1978. Social plasticity in the Acorn Woodpecker. Science 202:1298- 1300. Stacey, P. B. and R. Jansma. 1977. Storage of pinon nuts by the Acorn Woodpecker in New Mexico. Wilson Bulletin 89:150-151. Stanback, M. T. 1989. Observations on food habits and social organization of Acorn Woodpeckers in Costa Rica. Condor 91:1005-1007. Winkler, H., D. A. Christie, and D. Nurney. 1995. Woodpeckers, an identification guide to the wood- peckers of the world. Houghton Mifflin Company, Boston, Massachusetts and New York, USA. Wong, M. 1989. The implications of germinating acorns in the granaries of Acorn Woodpeckers in Panama. Condor 91:724-726. The Wilson Journal of Ornithology 120(3):499-504, 2008 SEASONAL VARIATION IN ACOUSTIC SIGNALS OF PILEATED WOODPECKERS SARAH B. TREMAIN,‘-2 KYLE A. SWISTON,' AND DANIEL J. MENNILL' ABSTRACT. — We used remote recorders to document temporal variation in acoustic signals of a population of Pileated Woodpeckers {Dryocopus pileatus) along the Choctawhatchee River in the Florida Panhandle, sam- pling seven different locations for 24 hrs once a week between mid January and mid April 2006. We found significant seasonal variation in the drumming behavior and vocal behavior of Pileated Woodpeckers. Drumming behavior peaked in mid March, just prior to onset of breeding activities. The three primary long-distance Pileated Woodpecker vocalizations, the cackle call, the wuk series call, and the wok call had similar patterns of seasonal variation. All four acoustic signals declined to low levels by early April when birds were nesting. The seasonal pattern of variation for all four Pileated Woodpecker acoustic signals had a similar pattern to that observed for song in many temperate passerines, and support the hypothesis that woodpecker calls and drumming displays are the functional counterparts to passerine song. Received 17 September 2007. Accepted 27 January 2008. The long-range acoustic signals of many territorial birds function in both territory de- fense and mate attraction (Catchpole and Slat- er 1995). Comparatively little research has been conducted on acoustic communication in woodpeckers (Piciformes) (Stark et al. 1998). The functions of the majority of acoustic sig- nals used by woodpeckers are not fully un- derstood, and seasonal patterns of variation in their acoustic behavior have yet to be inves- tigated. A clear understanding of seasonal var- iation in woodpecker acoustic signals may in- crease our understanding of the function of their calling and drumming behavior. Woodpecker vocalizations remain virtually unstudied, although woodpecker mechanical sounds have received modest attention (Stark et al. 1998). All woodpeckers produce non- vocal acoustic signals referred to as “drum- ming” and these far-carrying sounds are thought to function in long distance commu- nication (Dodenhoff et al. 2001). A wood- pecker drum consists of a repetitive series of strikes by the bird’s bill against an external substrate, and is distinguished as not being as- sociated with foraging or cavity excavation (Bent 1939, Pynnonen 1939, Short 1974). Woodpecker drumming is postulated to be the evolutionary counterpart to song (Pynnonen 1939, Lawrence 1967), replacing the complex vocalizations used by passerines for long-dis- tance communication (Dodenhoff et al. 2001). ' Department of Biological Science.s, University of Wind.sor. Windsor. ON N9B 3P4, Canada. ^Corresponding author; e-mail: tremai5@uwindsor.ca Several hypotheses for the function of drum- ming have been proposed; the best supported include territory announcement and mainte- nance, mate attraction, pair bond maintenance, and individual localization (Short 1982, Wil- kins and Ritchison 1999). Pileated Woodpeckers {Dryocopus pileatus) are sexually dimorphic, monogamous, non- migratory woodpeckers that inhabit deciduous and coniferous forests from southern Canada through the western, midwest, and eastern United States (Bent 1939, Bull and Jackson 1995). The ecology of Pileated Woodpeckers has been studied in greater detail than many other woodpecker species, perhaps owing to their majestic appearance, large size, and wide distribution. The basic characteristics of Pile- ated Woodpecker drums have been described (Bull and Jackson 1995, Stark et al. 1998), although a detailed study of variation in drum- ming behavior over time has not been previ- ously conducted. Additionally, there are dis- crepancies regarding vocalizations, especially concerning terminology and function (Hum- phrey 1946, Kilham 1959). We recorded the acoustic signals ot a pop- ulation of Pileated Woodpeckers along the Choctawhatchee River in the Florida panhan- dle over a 4-month period. Our objective was to examine seasonal patterns in the acoustic behavior of this species to better understand the possible functions of both vocal and non- vocal acoustic communication in woodpeck- ers. We predicted that acoustic signals of Pi- leated Woodpeckers would show a peak in ac- tivity at the start of the breeding season, sim- 499 500 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 3, September 2008 ilar to the documented pattern of seasonal variation in singing behavior of many song- birds (e.g., Catchpole 1973, Amrhein et al. 2004). METHODS Field Techniques. — We studied Pileated Woodpeckers in a plot of mature bottomland forest along the Choctawhatchee River in the Florida Panhandle (30° 3' N, 85° 5' W), dur- ing a bioacoustic search for Ivory-billed Woodpeckers (Campephilus principalis) (Hill et al. 2006). We placed seven automated “lis- tening stations” throughout our study site in January 2006. Each listening station consisted of a Sennheiser ME-62 omni-directional mi- crophone and K6 power module connected to a Marantz PMD-670 solid-state digital record- er powered by a sealed lead-acid battery. The microphones were shielded in protective rain guards made from polyvinyl chloride (PVC) tubing and were attached to the top of a 3-m wooden stake via a 30-cm shelf bracket. Wooden stakes were attached to small trees in an effort to capture sound from all directions and minimize any sound shadow resulting from the tree trunk. All visible materials were camouflaged with spray paint to minimize the influence of their presence on the behavior of resident birds. Each of the seven listening sta- tions recorded 24-hr MP3 files at 44.1 kHz, 16 bit, 160 kbps onto Hitachi 3GB microdrive cards. Memory cards and batteries were changed daily by quietly paddling or hiking to the site. Full details of the recording ap- paratus and sampling regime are given in Hill et al. (2006). We selected recordings from 14 dates be- tween 20 January and 20 April 2006 (specific recording dates: 20, 27 Jan; 3, 10, 16, 22, 28 Feb; 7, 14, 22, 28 Mar; and 5, 14, 20 Apr). We used recordings from seven distinct areas on each of the 14 dates which yielded 98 24- hr recordings. The recordings analyzed were selected from days with no precipitation and little to no wind to ensure consistency in re- cording conditions. Files were scanned using Syrinx-PC sound analysis software (John Burt, Seattle, WA, USA) and acoustic signals of interest were annotated using Syrinx-PC ’s time and frequency cursors. The seven recording units were separated by a distance of 766.4 ± 83.2 m (T ± SE) between the seven microphones. The territory size of Pileated Woodpeckers has not been carefully described, but reported breeding densities in bottomland forests of the southern United States are one pair/14-43 ha (Tanner 1942, Dennis 1951, Carter 1967). Conse- quently, we assume that each one of the seven recording units was recording different indi- viduals. We collected direct observations of birds on the Choctawhatchee River to assess the timing of Pileated Woodpecker breeding behavior. A team of field technicians searching for Ivory- billed Woodpeckers from January to May 2006 and 2007 documented any Pileated Woodpecker breeding behavior observed. We also referred to published records of the tim- ing of Pileated Woodpecker breeding. We col- lected average daily temperature data through- out the recording period from www. accuweather.com. Definition of Drums and Vocalizations. — The Pileated Woodpecker drum pattern is a tattoo of 1 1-30 beats lasting ~3 sec and trail- ing off in amplitude towards the end (Kilham 1959; Fig. lA). Birds also produce a quieter “drum-tapping” which is a slow, close-range communication that can be heard during times of feeding, courtship, and upon agreement re- garding the location of a nest hole (Kilham 1959). Many other woodpecker species pre- sent at our study site engaged in similar forms of drum-tapping, and we omitted drum-tap- ping given the difficultly in accurately distin- guishing between Pileated Woodpecker and heterospecific tapping without visual identifi- cation. Pileated Woodpecker vocalizations have been described by several authors, notably Kilham (1959) and Short (1982), although ter- minology is inconsistent across authors. Kil- ham (1959) described five basic calls while Short described three. However, across all de- scriptions, Pileated Woodpecker vocalizations are simple acoustic signals, consisting of 1—2 syllables given individually or in series (Bull and Jackson 1995). We examined temporal patterns in the three most common long-dis- tance Pileated Woodpecker vocalizations: (1) the cackle (also referred to as a fast wuk series call), (2) the wuk series call, and (3) the wok call (also referred to as the waa call). The cackle call (Fig. IB) consists of rapid series Tremain et al. • PILEATED WOODPECKER ACOUSTIC SIGNAL VARIATION 501 FIG. 1 . Sound spectrograms of four primary long- distance acoustic signals of Pileated Woodpeckers sampled with automated recorders in the Choctaw- hatchee River bottomland forests in Florida: (A) drum- ming, (B) the cackle call, (C) the wuk series call, and (D) the wok call. of notes that rise and then fall in pitch. Ca- pable of transmitting over long distances, it is thought to be used in territorial assertion, dis- tant interactions, and as a copulation readiness signal (Short 1982). The wuk series call (Fig. 1C) described by Short (1982), is a much lon- ger series of spaced-out notes. Some authors categorize the cackle call as a fast version of the wuk series call, but we distinguish the wuk as the substantially slower, drawn-out calls. The wuk series is used in situations of agita- tion or alarm (Bull and Jackson 1995). The wok call (Fig. ID) described by wShort (1982), is a lower amplitude vocalization generally given in a series of eight notes, at a rate of ~3 notes/sec. It is used during interactions be- tween birds and appears to be synonymous with Kilham’s (1959) “high call”. Statistical Analysis. — We applied repeated measures ANOVA using SPSS 14 to analyze temporal patterns in each of the four Pileated Woodpecker acoustic signals (SPSS Inc., Chi- cago, IL, USA). The within-subjects factor for each acoustic signal was the week number (14 levels), and the between-subjects factor was the recording location (7 levels). Values for each acoustic signal were calculated as the mean frequency of occurrence per hour across all daylight hours. Results are given as means ± SE and all tests are two-tailed. RESULTS Pileated Woodpeckers in the Choctawhat- chee River bottomlands demonstrated signifi- cant seasonal variation in all four of this spe- cies’ long-range acoustic signals. Drum pro- duction remained low throughout January and Eebruary, increased in early March, peaked in late March, and decreased to the January and February level by the beginning of April (AN- OVA: F = 2.49, P = 0.007; Fig. 2A). The frequency of occurrence of the cackle call in- creased from the end of January to a peak in mid March, followed by a decrease which continued into mid April (ANOVA: F — 4.50, P < 0.001; Fig. 2B). The wuk series call ex- hibited substantial variation from week to week, but it followed an oscillating pattern which peaked in late March, before decreasing substantially in April (ANOVA: F — 2.58, P = 0.005; Fig. 2C). The wok call at all times of year was less frequent than all other call types (Fig. 2). There was an increase in the frequency of occurrence for the wok call from January to mid March; after this date the wok call was not detected again until early April when levels increased once again, albeit at a low level (ANOVA: F = 2.70, P = 0.003; Fig. 2D). Pileated Woodpeckers were abundant at the study site and were encountered daily throughout the period from January to April. Pileated Woodpecker breeding activities were observed on four occasions; ( 1 ) one pair of birds excavated a cavity between 20 and 23 March (cavity excavation is believed to last Temp (-C) Acoustic signals/hr (mean ± SE) 502 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 Jan Feb Feb Mar Mar Mar Apr 20 3 17 3 17 31 14 Date FIG. 2. Seasonal variation in frequency of record- ing the four primary long-distance acoustic signals of Pileated Woodpeckers in the Choctawhatchee River bottomland forests in Florida: (A) drumming, (B) the cackle call, (C) the wuk series call, and (D) the wok call. Seasonal variation in temperature (E) shows a gradually increasing mean temperature across the re- cording period. —3-6 weeks); (2 and 3) two different nests were found, both with incubating parents, on 3 and 5 May (incubation is believed to last 18 days, suggesting eggs in these nests were laid in Apr); and (4) for a fourth pair fledglings were observed in early May (suggesting eggs were laid in late Mar or early Apr). This pe- riod of nesting and egg laying activity in late March and April corresponds to a period when the temperature on the Choctawhatchee River was beginning to increase (Fig. 2E). DISCUSSION Our recordings from the bottomland forests along the Choctawhatchee River in the Florida Panhandle revealed significant seasonal vari- ation in the acoustic behavior of Pileated Woodpeckers. All four of the Pileated Wood- pecker’s long-distance acoustic signals reached a peak in mid March. Pileated Wood- peckers in Florida commenced breeding in mid March; other studies have found that egg laying occurs between 17 March and 25 May in Florida (Bent 1939, Stevenson and Ander- son 1994). Our observations confirmed that mid March and early April mark the begin- ning of the breeding period for the Choctaw- hatchee Pileated Woodpecker population. Our results support the prediction that acoustic sig- nals peak in activity at the start of the breed- ing season. The pattern we observed for all four Pileated Woodpecker acoustic signals corresponds to the pattern of seasonal varia- tion documented for the singing behavior of many songbirds (e.g., Catchpole 1973, Am- rhein et al. 2004). All four of the long-range signals of Pile- ated Woodpeckers that we studied were given at low levels throughout late January and Feb- ruary. Pileated Woodpecker pairs are territo- rial at all times of the year (Hoyt 1957, Bull and Jackson 1995) and this early winter period corresponds with pre-breeding activities when males and females have not yet started exca- vating nest cavities. The wuk series call fol- lowed a non-linear pattern during this period of low vocal activity, which may be related to the proposed function of this call. This acous- tic signal is thought to be given as an alarm call or when a bird is agitated (Bull and Jack- son 1995). The stochastic pattern of variation we observed for the wuk series call likely re- lates to the stochastic nature of threats that may have elicited alarm. Both the cackle call and wok call increased in frequency in early March, reaching their peak in mid March. Both the drum and the wuk series call reached their maximum the following week. Drum- ming is believed to function in territorial an- nouncement (Lawrence 1967); this maximum Tremain et al. • PILEATED WOODPECKER ACOUSTIC SIGNAL VARIATION 503 presumably occurred because adults were ac- tively defending their territory and nest site at the time of intensified courtship and egg lay- ing. A decrease occurred in the production of all four acoustic signals in late March. Egg- laying has been documented to occur in Flor- ida between 17 March and 25 May (Bent 1939), and this minimum in acoustic signal production may have occurred because fe- males were incubating their clutches; high levels of calling may increase birds’ conspic- uousness and reveal the location of their nest hole to predators. The drum display, cackle call, and the wok call had a similar pattern of seasonal variation, suggesting that all three signals have roles of increased importance during the breeding sea- son. The drum and cackle call of the Pileated Woodpecker have both been proposed to func- tion in territorial advertisement; both males and females have been observed to respond to territorial intrusion by drumming and calling (Mellen et al. 1992). Pileated Woodpecker pairs defend territories all year, but do not de- fend their territories as vigorously during the non-breeding season in the fall and winter; during this time territorial intrusion by floaters is generally tolerated (Bull and Jackson 1995). The seasonal patterns we observed in both the drum and cackle calls of Pileated Woodpeck- ers in northern Florida reflect this increase in territorial aggression with hourly rates starting low during the pre-breeding season in January and mid February before peaking during the onset of breeding in mid March. The wok call is usually given in interactions between indi- viduals, which should similarly reach a max- imum during the period of affiliation associ- ated with the start of the breeding season. Un- derstanding how these vocalizations vary across the breeding season has important im- plications for survey techniques (Selmi and Boulinier 2003); our data show that Pileated Woodpeckers should be most readily detect- able early in the breeding season. Woodpecker drumming has long been pos- tulated to be the functional counterpart to pas- serine song (Pynnonen 1939, Lawrence 1967). Our analyses reveal a pattern of seasonal var- iation in drumming behavior, as well as call- ing behavior, that supports this hypothesis. Drumming has received modest attention in the literature (Stark et al. 1998, Dodenhoff et al. 2001), but more quantitative work is need- ed to clarify possible functions. ACKNOWLEDGMENTS We thank G. E. Hill, B. W. Rolek, and the many members of the Ivory-billed Woodpecker search team for assistance in the field. We thank R. D. Stark and an anonymous reviewer for comments on the manu- script. We thank the Natural Sciences and Engineering Research Council of Canada (NSERC), Canada Foun- dation for Innovation, Ontario government, and the University of Windsor for supporting DJM’s research program. We also thank NSERC, Pelee Island Winery, Nokuse Plantation Inc. and M. C. Davis, Northwest Florida Water Management District, Florida Fish and Wildlife Conservation Commission, U.S. Fish and Wildlife Service, and Marantz Canada for supporting the Florida search for Ivory-billed Woodpeckers, which gave rise to the data in this study. LITERATURE CITED Amrhein, V., H. P. Kunc, and M. Naguib. 2004. Sea- sonal patterns of singing activity vary with time of day in the Nightingale {Luscinia megarhyn- chos). Auk 121:110-117. Bent, A. C. 1939. Life histories of North American woodpeckers. U.S. National Museum Bulletin Number 174. Bull, E. L. and J. A. Jackson. 1995. Pileated Wood- pecker {Dryocopus pileatiis). The birds of North America. Number 148. Carter, W. A. 1967. Ecology of the nesting birds of the McCurtain Game Preserve, Oklahoma. Wilson Bulletin 79:259-272. Catchpole, C. K. 1973. The functions of advertising song in the Sedge Warbler {Acrocephalus schoe- nobaenus) and Reed Warbler {Acrocephalus scir- paceus). Behaviour 46:300—320. Catchpole, C. K. and P. J. B. Slater. 1995. Bird song: biological themes and variations. Cam- bridge University Press, New York, USA. Dennis, J. V. 1951. A comparative study of Florida woodpeckers in the nonbreeding season. Thesis. University of Florida, Gainesville, USA. Dodenhoff, D. J., R. D. Stark, and E. V. Johnson. 2001. Do woodpecker drums encode information for species recognition? Condor 103:143-150. Hill, G. E., D. J. Mennill, B. W. Rolek, T. J. Hicks, AND K. A. Swiston. 2006. Evidence suggesting that Ivory-billed Woodpeckers {Campephilus principalis) exist in Florida. Avian Conservation and Ecology 1(3):2. Hoyt, S. F. 1957. The ecology of the Pileated Wood- pecker. Ecology 39:246-256. Humphrey, P. S. 1946. Observations at the nest of a Pileated Woodpecker. Migrant 17:43 — 16. Kilham, L. 1959. Behavior and methods of commu- nication of Pileated Woodpeckers. Condor 61: 377-387. Lawrence, L. di-: K. 1967. A comparative life-history 504 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 3, September 2008 study of four species of woodpeckers. Ornitholog- ical Monographs Number 5. Mellen, T. K., E. C. Meslow, and R. W. Mannan. 1992. Summertime home range and habitat use of Pileated Woodpeckers in western Oregon. Journal of Wildlife Management 56:96-103. Pynnonen, a. 1939. Beitrage zur Kenntnis der Biol- ogiefinnisher Spechte. Annales Botanici Societatis Zoologicae Botanicae Lennicae Vanamo 7:1-166. Selmi, S. and T. Boulinier. 2003. Does time of season influence bird species number determined from point-count data? A capture-recapture approach. Journal of Lield Ornithology 74:349-356. Short, L. L. 1974. Habits and interactions of North American three-toed woodpeckers {Picoides arc- ticus and Picoides tridactylus). American Muse- um Novitates 547(2): 1-42. Short, L. L. 1982. Woodpeckers of the world. Dela- ware Museum of Natural History Monographs, Series Number 4. Stark, R. D., D. J. Dodenhoff, and E. V. Johnson. 1998. A quantitative analysis of woodpecker drumming. Condor 100:350-356. Stevenson, H. M. and B. H. Anderson. 1994. The birdlife of Llorida. University of Llorida Press, Gainesville, USA. Tanner, J. T. 1942. The Ivory-billed Woodpecker. Re- search Report Number 1. National Audubon So- ciety, New York, USA. Wilkins, H. D. and G. Ritchison. 1999. Drumming and tapping by Red-bellied Woodpeckers: descrip- tion and possible causation. Journal of Lield Or- nithology 70:578-586. The Wilson Journal of Ornithology 120(3):505— 512, 2008 COMMON POORWILL ACTIVITY AND CALLING BEHAVIOR IN RELATION TO MOONLIGHT AND PREDATION CHRISTOPHER P. WOODS' ^ AND R. MARK BRIGHAM'-^ ABSTRACT. — We investigated the influence of lunar and environmental factors on behavior of Common Poorwills {Phalaenoptilus nuttallii) in southern Arizona under a diverse set of natural and artificial conditions. Radio-marked poorwills were most active shortly after sunset during the new moon. Movements declined as evening progressed. Activity remained high for several hours after sunset when the moon was full. Poorwills were heard calling from March through October, but most calling occurred between early May and September. Only ambient light was correlated with number of poorwills heard calling. More poorwills responded to play- backs of conspecifics when the moon was full than when it was new. Poorwills did not change their response to conspecifics during full moon when playback of poorwill calls followed playback of Great Homed Owl {Bubo virginianus) calls but, during the new moon, fewer birds responded following the owl call. Poorwill behavior is strongly influenced by lunar conditions; their ability to detect and evade predators is important when calling advertises their location. Received 22 May 2006. Accepted 1 September 2006. The behavioral influence of sunlight on birds is evident, but the influence of moon- light on diurnal and nocturnal birds is less ap- parent. For example, many nocturnal birds call most actively at dusk in contrast to the well known singing or calling at dawn by di- urnal birds (Brauner 1952, Cooper 1981, Ga- ney 1990). Frequencies of calling, as well as overall activity patterns, are less well known during true night, when moonlight is the pri- mary source of illumination. Some animals that are eaten by owls limit activity during periods when the moon is full (Price et al. 1984, Brown et al. 1988, Kotler et al. 1994, Brigham et al. 1999, Beier 2005, Lang et al. 2006). Bright moonlight coincides with the highest rates of nocturnal predation on Black- vented Shearwaters {Pujfinus opisthomelas) and Cassin’s Auklets {Ptychoramphus aleuti- cus) by Western Gulls {Lcirus occidentalis) (Nelson 1989, Keitt et al. 2004). Predation by Great Black-backed Gulls (Larus marinus) on Manx Shearwaters (Puffinus puffinus) is also strongly influenced by lunar condition (Brooke 1990). The behavioral effects of different lu- nar conditions are not universal, however, and other nocturnal animals increase activity when the moon is full and the sky is relatively bright, presumably because visual constraints ' Department of Biology. University of Regina, Re- gina, SK S4S 0A2, Canada. ^Current address; 11181 West Seneca Drive, Boise, ID 83709, USA. Corresponding author; e-mail; mark.brigham@iiregina.ca are reduced during those periods (Cooper 1981, Mills 1986, Brigham and Barclay 1992, Jetz et al. 2003). The Caprimulgidae is a circumglobal fam- ily of visually acute, crepuscular or nocturnal insectivores for which some aspects of lunar influence have been evaluated. The most com- monly studied aspect of caprimulgid biology in the context of lunar illumination is the pos- sible synchrony of breeding with lunar cycle (Jackson 1985, Mills 1986, Brigham and Bar- clay 1992, Perrins and Crick 1996). The hy- pothesis that light levels constrain foraging activity has also been proposed (Csada et al. 1992, Bayne and Brigham 1995, Jetz et al. 2003). Common Poorwills {Phalaenoptilus nuttallii-, hereafter poorwill) are the smallest North American caprimulgid (45-50 g), and are one of the least understood of all North American birds, principally because of their cryptic coloration and nocturnal habits (Csada and Brigham 1992). We studied Common Poorwills under a diverse set of natural and artificial conditions to assess the influence of environmental and lunar factors on their be- havior. Specifically, we compared the likeli- hood of movement by radio-marked birds be- tween the full and new moons, and measured the rate of vocalization under a range of en- vironmental, lunar, and seasonal situations. We also used playbacks of poorwill and Great Horned Owl {Bubo virginianus) vocalizations to examine whether the proximity to a poten- tial predator influences calling responses by poorwills. 505 506 THE WILSON JOURNAL OF ORNITHOLOGY • Vol 120, No. 3, September 2008 METHODS We collected data during 1996-1999 at three sites within 100 km of Tucson, Arizona, USA: the National Audubon Society’s Apple- ton-Whittell Research Ranch (31° 36' N, 110° 30' W), central portions of the Buenos Aires National Wildlife Refuge (31° 41 ' N, 1 1 1° 26' W), and the east side of the Tortolita Moun- tains (32° 32' N, 111° 00' W). All three sites are within the Sonoran desert ecosystem and share environmental and ecological attributes, including an elevation range of 1,000 to 1,550 m, hot days and cool nights, limited rainfall that typically occurs during the summer mon- soon, and an abundance of cacti. Poorwills oc- curred year-round at the Tortolita and Buenos Aires sites; some over- wintered at the Re- search Ranch site, although most were absent during the coldest months. Poorwills were captured at night either us- ing mist nets in conjunction with playbacks of territorial calls or with a spotlight and long- handled net (Swenson and Swenson 1977, Jackson 1984, Brigham 1992). A few birds were captured or recaptured during daylight by flushing them into mist nets placed near roosts. Captured birds were banded with a uses metal band and fitted with a tempera- ture-sensitive radio transmitter (2 models that differed in mass and range were used: Model PD-2T, 2.8 g and 1-4 km range, and Model BD-2GT, 1.7 g and 0.5-2 km range; Holohil Systems Inc., Carp, ON, Canada). Transmit- ters were affixed using an elastic harness slipped over the wings (Brigham 1992, Hill et al. 1999). Likelihood of Activity. — Radio-marked poorwills were monitored remotely using a Lotek SRX 400 Telemetry Receiver (Lotek Engineering Inc., Newmarket, ON, Canada) to assess activity patterns by quantifying varia- tion in signal strength. Activity can be in- ferred because signal strength of successive radio pulses varies widely when radio-marked animals are active due to rapid changes in ori- entation of the transmitting antenna (Sutter et al. 1996, Brigham et al. 1999). We evaluated the relationship between time of night, lunar illumination, and movement by poorwills using telemetry data from 5-min pe- riods exactly 1, 3, and 5 hrs past sunset on nights for which there was either a full moon (>95% of the moon’s face illuminated; 82 bird-nights from 6 males, 5 females, and 2 unknown gender juveniles) or no moon (<5% of the moon’s face illuminated; 92 bird-nights from 9 males, 4 females, and 2 unknown gen- der juveniles). We only used data from birds that were euthermic (i.e., not torpid based on body temperature above 30° C) and for which there was either movement or inactivity in at least one of the 5-min periods. We inferred that movement had occurred when the stan- dard deviation of the intensity of three signals recorded by the Lotek receiver was >10. This value was derived from comparisons of vari- ation in signal strength when movement or in- activity by birds could be observed directly (cf. Sutter et al. 1996, Brigham et al. 1999). We used Chi-square tests to evaluate whether there were differences in the number of times that movement occurred during each 5-min period, for the two different lunar conditions. Environmental Influences on Vocaliza- tion.— We examined the influence of climatic, lunar, and seasonal variables on poorwill vo- calization by conducting night-time counts of calling poorwills at established points in the three study sites. Counts were made on 89 sur- vey nights between June 1996 and January 1999, although effort varied seasonally and between sites. Most surveys, especially during spring and summer, were at 7 to 10 day in- tervals, but logistics meant that five were sep- arated from a prior survey by only 2 to 4 days and 10 were separated by 2 weeks to 10 months. Two counts were done on 55 of 89 survey nights with 1-3 hrs separating each count making for 144 total counts. Poorwills were counted at the Research Ranch and Buenos Aires sites at nine points: three, each separated by 1 km, along each of three narrow and lightly-used dirt roads. Counts at the Tor- tolita Mountains site consisted of either five individual points or two sets of three points; in both cases points were separated by 1 km or more (logistics at Tortolita required two separate routes). Calling poorwills were counted at each point for 3 min during which time CPW, who conducted all counts, remained in darkness and silent (playbacks of poorwill calls were not used). The number of calling birds was assigned based on differing direction and dis- tance of calls and temporal overlap in calling Woods and Brigham • POORWILL ACTIVITY IN RELATION TO MOONLIGHT 507 (i.e., 2 or more birds calling at the same time). Counts were started at least 1 hr after sunset and completed in 1-2 hrs. No counts were conducted during civil twilight (when the sun is 0° to 6° below the horizon), even though poorwills called frequently during this time, so that solar light did not complicate our in- terpretation of the influence of lunar illumi- nation. No counts were conducted when it rained. We recorded environmental variables twice (Tortolita Mountains) or three (Research Ranch and Buenos Aires) times during each survey, and averaged them to establish overall count conditions. We measured ambient light levels, temperature, relative humidity, average wind speed, and percent cloud cover. Wind speed and/or relative humidity were not mea- sured during 49 counts. A Beseler PM2L Col- or Analyzer (Charles Beseler Co., Linden, NJ, USA) was used to measure ambient light lev- els (Hecker and Brigham 1999). This device generates a unit-less measure of light intensity to facilitate setting exposure times for film printing. It works at light intensities at which many light meters are ineffective. We used a Cole-Parmer thermistor thermometer (Model 8402-00; Cole-Parmer Instrument Co., Vernon Hills, IL, USA) to measure ambient tempera- ture and a General Eastern thermo-hygrometer (Model 880, General Eastern Instruments, Woburn, MA, USA) to measure relative hu- midity. We measured wind speed (±3 km/hr) over a 3-4 min period with a Kestrel wind meter (NK Electronics, Chester, PA, USA), which averaged measurements taken at 1-sec intervals. Percent cloud cover was visually es- timated. Most birds called during the warmer months and we included for our analysis the 92 counts between 5 May and 3 September, during which time 80% of all calling oc- curred. Forty-seven of the 92 counts included in our analysis were at the Research Ranch, 42 at Buenos Aires, and 3 at Tortolita Moun- tains. We assumed there was no site specific lunar effect. The maximum number of birds heard at each point along each route within any year was used as an estimate of year- and site-specific bird maximums. The number of calls detected during any count were stan- dardized against the maximums and our anal- ysis was based on the percentage of the max- imum number of birds calling, not the number of birds specifically. We used two separate backward stepwise multiple regressions to as- sess the importance of environmental vari- ables. First, for all 92 counts with light, tem- perature, and percent cloud cover as variables, and second, for the smaller sample of 43 counts for which data on relative humidity and wind speed were also available. Influence of Owls on Vocalization. — We as- sessed the influence of owls on calling behav- ior by poorwills by measuring the response to playbacks of poorwill calls both with and without the implied presence of a Great Homed Owl (these owls are common noctur- nal predators and were heard in all 3 study areas). We surveyed for calling poorwills (in- dependently from the counts described) using two playback treatments to discern the effect of potential owl presence on poorwill calling behavior under differing light levels. Play- backs were conducted at 1.6-km intervals on dirt road systems at the Research Ranch and Tortolita Mountains sites, as well as on lightly traveled dirt roads on the east side of the Rin- con Mountains, 50 km east of Tucson (32° 07' N, 110° 28' W), on desert flats near Black Mountain, 60 km north of Tucson (32° 49' N, 110° 57' W), and on the west side of the Sil- verbell Mountains, 55 km west of Tucson (32°26'N, 111°30'W). At each stop, we waited in silence and darkness for 30 sec and then played a treatment tape. The test was conducted between 6 June and 11 July 1998 on calm or nearly calm nights with little or no cloud cover, beginning at least 90 min after sunset. The test usually took several hours and was not repeated on the same evening. Three routes were surveyed twice, once during the full moon and again during no moon, and two were surveyed three times each, twice during a full moon (separated temporally by 1 month) and once when there was no moon. We randomly assigned a treatment to each stop prior to surveys, but with the condition that each set of two stops included each treat- ment. Counts were based on response to two playbacks treatments: poorwill only (15 sec poorwill, 60 sec silence, 15 sec poorwill) and poorwill. Great Horned Owl, poorwill (15 sec poorwill, 60 sec silence, 15 sec Great Horned Owl, 15 sec poorwill). The treatment variation thus occurred in the second call sequence. Re- 508 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 cordings used in treatments were taken from a standard collection of bird songs (Peterson Field Guides: Western Bird Songs, Houghton Mifflin Co., Boston, MA, USA). Complete 90-120 sec treatment tapes were created so that switching of tapes during counts was un- necessary. We used a portable cassette player, set at 3/4 maximum volume and judged that poorwills within —300 m responded. We measured the number of responding birds and calling intensity during the 60-sec interval following playback of the initial call and again during the 60-sec interval following playback of the treatment call(s). The number of birds calling was based on the direction and distance of calls, and overlap in calling. To characterize intensity, we counted the total number of calls, regardless of the number of birds calling, and assigned a categorical value based on six categories: (0) no calls, (1) 1-5 calls, (2) 6-10 calls, (3) 11-20 calls, (4) 21- 30 calls, and (5) 31+ calls. Categorical values were used to minimize potential errors in counting when calling was particularly in- tense. We completed 12 surveys with a total of 168 individual stops, 84 for each treatment. Stops at which no poorwills were heard (29 in total), were eliminated from our analysis. We initially compared the full versus new moon results prior to the treatment playback (i.e., the numerical and categorical response to the first poorwill call only). A subsequent analysis to assess treatment effects was based on the per point change in number of birds or calls from the 60-sec period after the first playback to the 60-sec period after the second playback. Preliminary analysis indicated that data for both survey treatments were not nor- mally distributed (Shapiro- Wilks’ W with as- sociated P < 0.01); therefore, we used a non- parametric Mann-Whitney U-test for all com- parisons. RESULTS Likelihood of Activity. — We used data from 135 activity records (59 full moon, 76 new moon) for the period 1 hr past sunset, 1 12 (56 full moon, 56 new moon) for 3 hrs past sunset, and 97 (48 full moon, 49 new moon) for 5 hrs past sunset to examine poorwill movements. The likelihood that a bird was active during full moon was similar in the three time peri- ods with movement occurring in 70% (1 hr). FIG. 1. Likelihood of activity by Common Poor- wills during 5-min periods 1, 3, and 5 hrs after sunset in southern Arizona. There were between 97 and 135 sampling periods within each time period, and activity differs within each (x^ tests, all P < 0.05). 61% (3 hrs), and 69% (5 hrs) of each. The likelihood of activity decreased over time dur- ing the new moon with movements occurring in 86, 39, and 25% of the three periods, re- spectively (Fig. 1). Birds were more likely to be active 3 and 5 hrs past sunset when the moon was full versus when it was new (x^i = 5.1, P = 0.023 and = 19.1, P < 0.001, respectively). The situation was reversed, however, 1 hr past sunset, when activity was more common during the new than the full moon (x^i = 5.1, P = 0.024). Environmental Influences on Vocaliza- tions.— Poorwills in southern Arizona called throughout an extended period from March through October (Fig. 2). Most calling oc- curred during the summer months. The central 50 and 80% of all calls were recorded between 27 May and 29 July, and 5 May and 9 Sep- tember, respectively. Ambient light levels sig- nificantly predicted poorwill calling, whether based on the data from 92 counts for which only light, temperature, and cloud cover were analyzed (Pi ^q - 92.2, = 0.509, P < 0.001, Pught = Fig- 3), or based on the data from 43 counts that also included relative hu- midity and wind speed (Pi^i = 49.1, — 0.545, P < 0.001, (3L.gh, = 0.74). No other en- vironmental variable measured was signifi- cantly related to the number of calling birds, regardless of the data set used. Influence of Owls on Vocalizations. — The effect of moonlight on calling was also evi- dent from playback surveys whose purpose Woods and Brigham • POORWILL ACTIVITY IN RELATION TO MOONLIGHT 509 Date Light FIG. 2. Seasonality of calling by Common Poor- wills in southern Arizona from point count surveys without playback of a poorwill call. Solid lines encom- pass the central 50% of all calls, and stippled lines the central 80%. was to evaluate the effect of owl calls on poorwill vocalizing. Poorwills responded in significantly greater numbers and more in- tensely during full moons versus new moons after playback of the first poorwill call. The number of birds responding during the new moon averaged only 38% of the number call- ing during the full moon (mean of 0.59 birds at each point vs. 1.55 birds; Z5485 = 5.66, P < 0.001). The categorical number of calls de- tected during the new moon averaged 52% of the number heard during full moon (1.57 vs. 3.02; Z5485 = 4.51, P < 0.001). There was no significant difference in the change in calling response following playback of the second treatment call(s) during the full moon either in number of birds responding or intensity of calling. The number of responding poorwills increased 16% (from 1.64 to 1.90 birds) per point following the poorwill only treatment during the full moon from the pe- riod after the first call to the period after the second. Calling increased 14% (from 1.47 to 1.67 birds) after the Great Horned Owl then poorwill treatment (Z4243 ~ 0.14, P > 0.1). The number of calls increased 21% (from 3.05 to 3.69) after the poorwill only treatment, and 1 1% (from 3.00 to 3.33) following the Great Horned Owl then poorwill treatment (Z4243 = 0.85, P > 0.1). In contrast, there was a sig- nificant difference in the change in response by poorwills when the moon was new, both in the number of birds responding and the cat- egorical intensity of calling. The number of FIG. 3. Relationship between ambient light at night and calling by Common Poorwills during spring and summer in southern Arizona without playback of a poorwill call. Light levels are unit-less measurements from a photographic light meter, with “10” represen- tative of no moon and little light, and “50” of a full moon and bright conditions. responding poorwills following the poorwill only treatment during the new moon increased 59% (from 0.56 to 0.89 birds) after the second call compared to after the first call, but de- creased 11% (from 0.63 to 0.56 birds) after the Great Homed Owl then poorwill treatment (Z27,27 = 2.42, P = 0.016). Categorically, the number of calls increased 35% (from 1.59 to 2.15) after the poorwill only treatment, but de- creased 4% (from 1.56 to 1.48) following the Great Homed Owl then poorwill treatment (Z2727 = 2.19, = 0.029). DISCUSSION The intensity of night-time lunar illumina- tion varies by more than a hundred-fold de- pending on phase (Austin et al. 1976). Con- sequently, it is not surprising that lunar phase should influence behavior of nocturnal ani- mals, especially those that orient visually, al- though the direction of this influence may not be obvious. In the absence of extenuating fac- tors, visually-oriented nocturnal predators like poorwills should concentrate foraging activity during bright moonlight but vocalize more frequently when there is little moonlight, when foraging should be least efficient (Brigham and Barclay 1992, Csada et al. 1992, Bayne and Brigham 1995). Chuck- wilFs-widows {Caprimulgus carolinensis) and Whip-poor-wills (C. vociferus) call with great- er frequency during moonlit conditions, how- 510 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 ever, and lunar illumination is the only envi- ronmental factor correlated with calling dur- ing the breeding season (Cooper 1981, Mills 1986, Ganey 1990). Poorwills in our study vo- calized most actively during the full moon and lunar illumination was the only environmental variable we measured with which the number of birds heard calling was correlated. In ad- dition, poorwills tended to be more active (which we interpret as foraging) during full moon, although the number of flying insect prey depended principally on temperature and was independent of ambient light (Woods 2002). Consequently, increased activity by poorwills during the brightest periods was not related to increased prey density, and vice ver- sa. Jetz et al. (2003) found a strong relationship between lunar cycle and timing of foraging activity by Standard-winged (Macrodipteryx longipennis) and Long-tailed {Caprimulgus climacurus) nightjars in equatorial West Af- rica. They noted that foraging was concen- trated during the crepuscular period during all phases of the lunar cycle but increased signif- icantly around new moon. They also detected a small but significant increase in abundance of some larger insects during full moon peri- ods which correlated with increased foraging during periods of the night with lunar light. In cooler more seasonal subtropical environ- ments, temperature as well as light levels strongly affects activity patterns of Freckled Nightjars {Caprimulgus tristigma', R. A. M. Ashdown and A. E. McKechnie, unpubl. data). Unfortunately, calling behavior was not evaluated in either of these studies. Peaks in calling occur for some nocturnal birds throughout the breeding season, and are often attributed to different breeding stages (e.g., Clark and Anderson 1997). We found no seasonal peaks in calling, possibly because of the cyclic influence of lunar phase on calling frequency. Poorwills are known to call persis- tently (Bent 1940, Gabrielson and Jewett 1940, Bailey and Niedrach 1965, this study), and we heard birds call in every month but January. These observations support the hy- pothesis that individuals are inclined to call frequently where they occur as year-round res- idents, regardless of breeding stage, but in lo- cations where they are migratory, calling tends to cease in late summer (Kalcounis et al. 1992). These results suggest a trade-off ex- ists between frequent calling and efficient for- aging, assuming that calling reduces foraging efficiency. The extent to which vocalizing may restrict foraging by caprimulgids is un- certain, but males deplete fat reserves during the breeding season whereas females, which probably call infrequently, do not (Csada and Brigham 1992, Thomas et al. 1996). Energetic shortfalls resulting from forgoing feeding when prey are most visible are probably min- imized since full moons are above the horizon for the most time. Consequently, more forag- ing time is available when the moon is rela- tively full. During new moon, activity was highest shortly after sunset and diminished as evening progressed, whereas during full moon, activity levels were consistent through- out late evening (cf. Brauner 1952). These contrasting patterns suggest that during new moons, poorwills must forage with greater in- tensity at and just past twilight, since low light levels will reduce foraging efficiency and may increase predation risk later in the night. Why are poorwills more vocal during pe- riods of high lunar illumination? In addition to the offsetting increase in time available for foraging, the ease with which males can move within and defend their territories is likely en- hanced by relatively bright conditions. Our re- sults suggest the ability to detect and evade predators is also important when calling ad- vertises location. This explanation is support- ed by the higher intensity of calling in re- sponse to conspecific calls following owl playbacks during the full moon, but an appar- ent reluctance to respond under similar cir- cumstances during the new moon. Other rel- atively small nocturnal birds that are espe- cially vocal when the moon is bright also have acute vision and rely heavily on vision to for- age (e.g.. Whip-poor-will). We surmise that visual acuity influences timing or extent to which smaller nocturnal animals vocalize at night. Eor nocturnal animals that may less eas- ily detect and out-maneuver predators, the in- creased risk of predation during the full moon selects for less activity. Gerbils (Gerbillus spp.) and other desert rodents, as well as Aus- tralian Owlet-Nightjars {Aegotheles cristatus), adjust foraging patterns to avoid bright moon- light and are all preyed upon by owls (Price et al. 1984, Brown et al. 1988, Kotler et al. Woods and Brigham • POORWILL ACTIVITY IN RELATION TO MOONLIGHT 511 1994, Brigham et al. 1999). Notably, one male radio-marked poorwill was killed by an owl {Otus sp.) in our study. The bird was taken under relatively dim conditions when the moon was in its first quarter and setting. Whether specific predators and or predator foraging strategies are the major selective pressures which have shaped the differences in response to lunar conditions between ca- primulgids and some rodents, remains to be understood. Future research is needed with an emphasis on explaining the role of lunar illumination on behavior of other birds that are active at night, including those that are generally considered diurnal (e.g., Johnson et al. 2003). The lunar cycle is relatively short in comparison with the overall breeding season for most birds, and consideration of this variable could pro- vide insight into foraging strategies and tim- ing of breeding. Knowledge of the lunar phase and its effect on both foraging by and preda- tion risk on nocturnal birds may also be an important consideration for surveying noctur- nal species (Ganey 1990, Downs 1998). Ad- ditionally, our study sites were relatively dis- tant from large urban centers and the light pol- lution associated with them. Light pollution could relieve some dependence on moonlight for efficient foraging by predators but, as a consequence, enhance predation pressure on their prey (Buchanan 2005, Frank 2005, Lloyd 2005). Perhaps more importantly, the loss of cyclic variation in nocturnal illumination may disrupt behaviors that evolved in association with regular fluctuation between bright and dark periods (Rich and Longcore 2005). Ad- ditional research into the impact of light pol- lution on the behavior of nocturnal animals is essential, considering the extent to which it has altered the night sky in many developed regions. ACKNOWLEDGMENTS R A. Bradshaw, R. K. Sechler, and B. A. Woods assisted in trapping and radio-marking poorwills. The National Audubon Society and W. V. Branan provided generous access to the Appleton-Whittell Research Ranch and logistic support. Comments by C. E. Braun and two anonymous reviewers greatly improved the manuscript. Funding tor this research was provided in part by awards to CPW from the Faculty of Graduate Studies and Research at the University of Regina and an NSERC research grant to RMB. LITERATURE CITED Austin, R. H., B. F. Phillips, and D. J. Webb. 1976. A method for calculating moonlight illuminance at the earth’s surface. Journal of Applied Ecology 13:741-748. Bailey, A. M. and R. J. Niedrach. 1965. Birds of Colorado. Denver Museum of Natural History, Denver, Colorado, USA. Bayne, E. M. and R. M. Brigham. 1995. Prey selec- tion and foraging constraints in Common Poor- wills (Phalaenoptilus nuttallii: Aves: Caprimul- gidae). Journal of Zoology 235:1-8. Beier, P. 2005. Effects of artificial night lighting on terrestrial mammals. Pages 19-42 in Ecological consequences of artificial night lighting (C. Rich and T. Longcore, Editors). Island Press, Washing- ton, D.C., USA. Bent, A. C. 1940. Life histories of North American cuckoos, goatsuckers, hummingbirds, and their al- lies. U.S. National Museum Bulletin 176. Brauner, j. 1952. Reactions of poor-wills to light and temperature. Condor 54:152-159. Brigham, R. M. 1992. Daily torpor in a free-ranging goatsucker, the Common Poorwill {Phalaenopti- lus nuttallii). Physiological Zoology 65:457-472. Brigham, R. M. and R. M. R. Barclay. 1992. Lunar influence on foraging and nesting activity of Com- mon Poorwills {Phalaenoptilus nuttallii). Auk 109:315-320. Brigham, R. M., R. C. A. Gutsell, R. S. Wiacek, and F. Geiser. 1999. Foraging behaviour in relation to the lunar cycle by Australian Owlet-Nightjars Ae- gotheles cristatus. Emu 99:253-261. Brooke, M. 1990. The Manx Shearwater. T. & A. D. Poyser, London, United Kingdom. Brown, J. S., B. P. Kotler, R. J. Smith, and W. O. WiRTZ II. 1988. The effects of owl predation on the foraging behavior of heteromyid rodents. Oec- ologia 76:408-415. Buchanan, B. W. 2005. Observed and potential effects of artificial night lighting on anuran amphibians. Pages 192-220 in Ecological consequences of ar- tificial night lighting (C. Rich and T. Longcore, Editors). Island Press, Washington, D.C., USA. Clark, K. A. and S. H. Ander.son. 1997. Temporal, climatic and lunar factors affecting owl vocaliza- tions of western Wyoming. Journal of Raptor Re- search 3 1 :358-363. Cooper, R. J. 1981. Relative abundance of Georgia caprimulgids based on call-counts. Wilson Bulle- tin 93:363-371. CsADA, R. D. AND R. M. Brigham. 1992. Common Poorwill {Phalaenoptilus nuttallii). The birds of North America. Number 32. CsADA, R. D., R. M. Brigham, and B. R. Phtendrigh. 1992. Prey selection in relation to in.scct avail- ability by the Common Poorwill {Phalaenoptilus nuttallii). Canadian Journal of Zoology 70:1299- 1303. Downs, K. D. 1998. Common Snipe surveys, habitat 512 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 3, September 2008 and evaluation of the Breeding Bird Survey. The- sis. University of Wyoming, Laramie, USA. Lrank. K. D. 2005. Effects of artificial night lighting on moths. Pages 305-344 in Ecological conse- quences of artificial night lighting (C. Rich and T. Longcore, Editors). Island Press, Washington, D.C., USA. Gabrielson, I. N. AND S. G. Jewett. 1940. Birds of Oregon. Oregon State Monographs 2. Oregon State College. Corvallis, USA. Ganey, J. L. 1990. Calling behavior of Spotted Owls in northern Arizona. Condor 92:485-490. Hecker, K. R. and R. M. Brigham. 1999. Does moon- light change vertical stratification of activity by forest dwelling insectivorous bats? Journal of Mammalogy 80:1196-1201. Hill, I. L, B. H. Cresswell, and R. E. Kenward. 1999. Field-testing the suitability of a new back- pack harness for radio-tagging passerines. Journal of Avian Biology 30:135-142. Jackson, H. D. 1984. Finding and trapping nightjars. Bokmakierie 36:86-89. Jackson, H. D. 1985. Aspects of the breeding biology of the Fierynecked Nightjar. Ostrich 56:263-276. Jetz, W, J. Steefen, and K. E. Linsenmair. 2003. Ef- fects of light and prey availability on nocturnal, lunar and seasonal activity of tropical nightjars. Oikos 103:627-639. Johnson, M., J. P. Beckmann, and L. W. Oring. 2003. Diurnal and nocturnal behavior of breeding Amer- ican Avocets. Wilson Bulletin 115:176—185. Kalcounts, M. C., R. D. Csada, and R. M. Brigham. 1992. The status and distribution of the Common Poorwill in the Cypress Hills, Saskatchewan. Blue Jay 50:38-44. Keitt, B. S., B. R. Tershy, and D. A. Croll. 2004. Nocturnal behavior reduces predation pressure on Black- vented Shearwaters Piiffimis opisthomelas. Marine Ornithology 32:173-178. Kotler, B. P, Y. Ayal, and A. Subach. 1994. Effects of predatory risk and resource renewal on the tim- ing of foraging activity in a gerbil community. Oecologia 100:391-396. •Lang, A. B., E. K. V. Kalko, H. Romer, C. Bock- HOLDT, AND D. K. N. Dechmann. 2006. Activity levels of bats and katydids in relation to the lunar cycle. Oecologia 146:659-666. Lloyd, J. E. 2005. Stray light, fireflies, and fireflyers. Pages 345-364 in Ecological consequences of ar- tificial night lighting (C. Rich and T. Longcore, Editors). Island Press, Washington, D.C., USA. Mills, A. M. 1986. The influence of moonlight on the behavior of goatsuckers (Caprimulgidae). Auk 103:370-378. Nelson, D. A. 1989. Gull predation on Cassin’s Auklet varies with lunar cycle. Auk 106:495-497. Perrins, C. M. and H. Q. P. Crick. 1996. Influence of lunar cycle on laying dates of European Nightjars (Caprimulgus europaeiis). Auk 113:705-708. Price, M. J., N. W. Waser, and T. A. Bass. 1984. Effects of moonlight on microhabitat use by desert rodents. Journal of Mammalogy 65:211-219. Rich, C. and T. Longcore (Editors). 2005. Ecological consequences of artificial night lighting. Island Press, Washington, D.C., USA. Sutter, G. C., D. J. Sawatzky, D. M. Cooper, and R. M. Brigham. 1996. Renesting intervals in Sprague’s Pipit, Anthus spragueii. Canadian Field- Naturalist 110:694-697. Swenson, J. E. and S. Swenson. 1977. Nightlighting as a method for capturing Common Nighthawks and other caprimulgids. Bird Banding 48:279- 280. Thomas, D. W., R. M. Brigham, and H. Lapierre. 1996. Field metabolic rates and body mass chang- es in Common Poorwills {Phalaenoptilus mittalliv. Caprimulgidae). Ecoscience 3:70-74. Woods, C. P. 2002. Ecological aspects of torpor use and inactivity during winter by Common Poor- wills. Dissertation. University of Regina, Regina, Saskatchewan, Canada. The Wilson Journal of Ornithology 1 20(3);5 1 3— 5 1 8, 2008 MAXIMIZING DETECTION PROBABILITY OF WETLAND-DEPENDENT BIRDS DURING POINT-COUNT SURVEYS IN NORTHWESTERN FLORIDA CHRISTOPHER P. NADEAU,'-’ COURTNEY J. CONWAY,' BRADLEY S. SMITH,' AND THOMAS E. LEWIS’ ABSTRACT. — We conducted 262 call-broadcast point-count surveys (1-6 replicate surveys on each of 62 points) using standardized North American Marsh Bird Monitoring Protocols between 31 May and 7 July 2006 on St. Vincent National Wildlife Refuge, an island off the northwest coast of Florida. We conducted double- blind multiple-observer surveys, paired morning and evening surveys, and paired morning and night surveys to examine the influence of call-broadcast and time of day on detection probability. Observer detection probability for all species pooled was 75% and was similar between passive (69%) and call-broadcast (65%) periods. Detection probability was higher on morning than evening (r = 3.0, P = 0.030) or night {t = 3.4, P = 0.042) surveys when we pooled all species. Detection probability was higher (but not significant for all species) on morning compared to evening or night surveys for all five focal species detected on surveys: Least Bittern {Ixobrychus exilis). Clapper Rail {Rallus longirostris). Purple Gallinule {Porphyrula martinica). Common Moor- hen (Gallinula chloropus), and American Coot {Fulica americana). We detected more Least Bitterns (r = 2.4, P = 0.064) and Common Moorhens {t = 2.8, P = 0.026) on morning than evening surveys, and more Clapper Rails (r = 5.1, P = 0.014) on morning than night surveys. Received 2 March 2007. Accepted 6 October 2007. Maximizing detection probability of rare or inconspicuous birds during point-count sur- veys is essential so that sufficient individuals are detected to reliably estimate population trends (Lynch 1995). Wetland-dependent birds (i.e., rails and bitterns) are among the most inconspicuous groups of birds in North Amer- ica. They vocalize infrequently and often oc- cur in isolated wetlands making them difficult to monitor (Bystrak 1981, Gibbs and Melvin 1993). A marsh bird monitoring protocol was developed in 1999 for conducting standard- ized surveys for wetland-dependent birds across North America (Conway 2007). The protocol instructs surveyors to use call-broad- cast after an initial 5-min passive period to increase vocalization probability of birds pre- sent during the survey period. However, call- broadcast might interfere with the observer’s ability to hear vocalizing birds during the sur- vey (Conway and Nadeau 2006). A decrease in observer detection probability could poten- tially negate the benefits of increased vocali- ' uses, Arizona Cooperative Fish and Wildlife Re- search Unit, 325 Biological Sciences P^ast, University of Arizona, Tucson, AZ 85721, USA. ^ St. Vincent National Wildlife Refuge. U.S. F^'ish and Wildlife Service, 479 Market Street, P. O. Ffox 447, Apalachicola, FL 32329, USA. ^ Corresponding author; e-mail: cnadeau@email.arizona.edu zation probability. Few studies have examined the effects of call-broadcast on observer de- tection probability and participants using this protocol are not obligated to use methods (i.e., multiple-observer surveys) to account for var- iation among observers. Vocalization probability of wetland-depen- dent birds varies with time of day and diurnal patterns may vary regionally and among spe- cies (Conway and Gibbs 2001). Thus, survey- ors should identify the optimal time of day to conduct surveys in their region to maximize detection probability. The Standardized North American Marsh Bird Monitoring Protocol (Conway 2007) instructs participants to con- duct surveys in the morning or evening when birds are most vocal. Participants using the protocol are not obligated to define which of the two daily time periods is optimal in their region and few participants have attempted to do so. Many species of wetland-dependent birds are known to vocalize at night (e.g.. Black Rail \Laterallns Jamaicensi.s]. Clapper Rail \Rallus longirosfri.s], Virginia Rail \R. littiicola]. Yellow Rail \Cotumicops novehor- acen.sis], and American Bittern \Rotauru.s len- tiginosus\\ Reynard 1974, Meanley 1985, Johnson and Dinsmore 1986, Gibbs et al. 1992, Bookhout 1995), yet few studies have attempted to examine the efficacy of night sur- veys. We ccmipared morning to evening sur- 514 THE WILSON JOURNAL OL ORNITHOLOGY • Vol 120, No. 3, September 2008 veys and morning to night surveys to ascertain the optimal time of day to conduct surveys for wetland-dependent birds in the southeastern United States. Our specific objectives were to: ( 1 ) compare observer detection probability be- tween passive and call-broadcast surveys in the context of the North American Marsh Bird Monitoring Protocol, and (2) ascertain the op- timal time of day to conduct surveys for wet- land-dependent birds in the southeastern Unit- ed States. METHODS Study Area. — All surveys were conducted on the St. Vincent Island (SVI) portion of St. Vincent National Wildlife Refuge (29° 40' N, 85° 05' W). SVI is a 4,968 ha forested barrier island in the northeastern Gulf of Mexico bounded by St. Vincent Sound, Apalachicola Bay, and the Gulf of Mexico. Thirty-seven percent of the island is considered suitable habitat for wetland-dependent birds: 308 ha of managed marsh, 67 ha of managed open wa- ter, 1,090 ha of estuarine marsh, and 383 ha of estuarine open water (Grace 2000). Palus- trine marsh (270 ha) and palustrine scrub (43 ha) wetlands exist on the island, but were not surveyed since few wetland-dependent birds have been detected in these wetland types. We established five survey routes on SVI: three were in managed marshes and two were in estuarine marshes. Vegetation on three survey routes was dominated by Cladiiim. The other two routes were co-dominated by Typha and Juncus or Spartina and Cladium. Salinity in the managed marshes increased throughout the study and ranged from 3.1 to 18.5 ppt. Salinity measurements were not available for the estuarine marshes. We established 62 sur- vey points, placing 12-14 points on each sur- vey route. Adjacent points were spaced 200 m apart (400 m on one route to conform to past surveys). We chose an interval of 200 m on the four newly established routes to increase our probability of detecting Black Rails, which are rarely detected beyond 100 m (Con- way et al. 2004). Survey points were marked in the field with a portable GPS receiver and with rebar or surveyor tape. Surveys. — Survey methods followed the Standardized North American Marsh Bird Monitoring Protocol (Conway 2007). We re- corded all aural and visual detections of 10 fo- cal species during each minute of both a 5-min passive listening period and a 5-min call-broadcast period at each point. The 10 fo- cal species thought to occur in the area includ- ed Pied-billed Grebe {PodUymbus podiceps). Least Bittern {Ixobrychus exilis), American Bittern, Black Rail, Clapper Rail, King Rail (Rallus elegans), Sora {Porzana Carolina), Pur- ple Gallinule (Porphyrula martinica). Common Moorhen {Gallinula chloropus), and American Coot (Fulica americana). The call-broadcast period was composed of 30 sec of broadcast calls followed by 30 sec of silence for each of five species in the following sequence: Black Rail, Least Bittern, Clapper Rail, Common Moorhen, and Purple Gallinule. We used the standardized call-broadcast recordings for the North American Marsh Bird Monitoring Pro- tocol (Conway 2007) that we obtained from the coordinators of the program. We did not broad- cast for all (5 of 10) focal species to limit the duration of the survey at each point. We ex- cluded species of lesser management concern or species commonly detected without the use of call-broadcast. We excluded King Rail from our call-broadcast due to the similarity between King Rail and Clapper Rail calls and because they commonly respond to each others’ calls (Conway and Nadeau 2006). Surveys were conducted on days without rain and when winds were <10 km/hr. All calls were broad- cast using a Memorex CD player (Model #MD6443SIL) and a Sony SRS-A27 Active Speaker System placed on the ground or bow of the canoe pointed perpendicular to the edge of the marsh. All broadcasts were —90 dB measured 1 m from the speaker. We used a Kestrel 3000 weather anemometer to record temperature and wind speed at the beginning and end of each survey route. We also esti- mated percent cloud cover at the beginning and end of each survey route. We measured tem- perature, wind speed, and cloud cover to ex- amine whether differences in detection proba- bility during different periods (morning, even- ing, and night) were potentially due to differ- ences in weather conditions. Double-blind Multiple-observer Surveys. — We conducted double-blind multiple-observer surveys at 26 points on two survey routes on 25 June and 5 July 2006. Double-observer surveys require two trained observers. We only had two trained observers and chose each Nadeau et al. • DETECTION PROBABILITY OF WETLAND BIRDS 515 route based on the route observer #1 planned to survey (for the second part of our study) on the day observer #2 was available. Both observers estimated the distance to each bird detected and recorded the call given by each bird. We were able to use these auxiliary data to easily identify (after the survey) which birds had and had not been detected by each observer. The two observers stood side-by- side or sat on opposite ends of a canoe during the surveys. They recorded their data incon- spicuously during the survey, by recording their data discreetly and by shielding their data sheets with their clipboard, to not alert the other observer when they detected a bird. They did not discuss their detections until the survey was complete. We followed Nichols et al. (2000) to esti- mate observer detection probability for each of two observers (Pi and P2) during each of three time periods; (1) the entire survey (pas- sive and call-broadcast combined), (2) the passive period only, and (3) the call-broadcast period only. We averaged Pj and P2 to obtain an overall estimate of observer detection prob- ability for each of the three time periods. Detection Probability During Different Times of Day. — We conducted 144 paired point-count surveys on three survey routes (38 points with 2-6 replicates/point) during both the morning and evening. Paired surveys were conducted on the same day or consecutive days between 31 May and 30 June 2006. Morning surveys were conducted 0.5 hrs be- fore sunrise until no later than 1000 hrs EDT, and evening surveys were conducted 3 hrs be- fore sunset until dark. We also conducted 104 paired point-count surveys on two different routes (26 points with 4 replicates per point) during both the morning and night. Paired sur- veys were conducted on the same day or con- secutive days between 7 June and 7 July 2006. Morning surveys were conducted 0.5 hrs be- fore sunrise until no later than 1000 hrs, and night surveys were conducted between 0100 and 0400 hrs. Paired morning versus evening surveys were conducted on different survey routes than paired morning versus night sur- veys with the exception of one pair of sur- veys. We used paired /-tests to compare the mean number of wetland-dependent birds detected (all species pooled) and the mean number of each species detected on each route between morning and evening surveys, and between morning and night surveys. We also used paired /-tests to compare temperature, wind speed, and percent cloud cover between morn- ing and evening surveys, and between morn- ing and night surveys. RESULTS Double-blind Multiple-observer Surveys. — We detected 5 of 10 focal species during dou- ble-blind multiple-observer surveys: Least Bittern, American Bittern, Clapper Rail, King Rail, and Common Moorhen. Observer detec- tion probability for all species pooled across the entire survey (passive and call-broadcast periods combined) was 75%. The observer de- tection probability was similar for both the passive period (69%) and call-broadcast peri- od (65%). Detection Probability During Different Times of Day. — We detected 5 of 10 focal spe- cies on at least one of the 12 paired morning and evening surveys; Least Bittern, Clapper Rail, Purple Gallinule, Common Moorhen, and American Coot. The mean (± SE) number of birds detected was higher (/ 3.0, P = 0.030) on morning {x = 19.8 ± 6.5 birds) than on evening surveys (T = 7.8 ± 3.3 birds) when we pooled all species. We detected 106% more Least Bitterns (/ = 2.4, P = 0.064) and 170% more Common Moorhens (/ = 2.8, P = 0.026) on morning than on even- ing surveys (Fig. 1 A). We did not observe sig- nificant differences in the number of individ- uals detected between morning and evening surveys for Clapper Rail, Purple Gallinule, and American Coot. We did not observe a dif- ference in percent cloud cover (/ = 0.8, P = 0.31) or wind speed (/ = 0, P = 1.00) between morning (.Cioud = 19 ± 7%, = 1 .5 ± 0.1 km/hr) and evening = 22 ± 8%, = 1 .5 ± 0.3 km/hr) surveys, but temperature was 8% higher on evening (.?„., „p = 28.6 ± 0.5° C) than on morning (.v,^.„ip = 26.6 ± 0.8° C) sur- veys (/ = 3.5, P = 0.017). We delected 3 of 10 focal species on al least one of the eight paired morning and night sur- veys; Least Bittern, Clapper Rail, and Com- mon Moorhen. The mean number of birds de- lected was higher (/ = 3.4, P = 0.042) on morning (.v = 5.5 ± 1.7 birds) than on night surveys (.v = 1 .2 ± 0.6 birds) when we pooled 516 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 3, September 2008 Least Clapper Purple Common American Bittern Rail Gallinule Moorhen Coot LIG. I. Mean (± SE) number of wetland-dependent birds detected during (A) six morning (gray) and six evening (black) surveys, and (B) four morning (gray) and four night (black) surveys on St. Vincent National Wildlife Refuge, northwestern Llorida, 31 May-7 July 2006. Data in (A) and (B) were from different survey routes and can not be appropriately compared. Sample sizes (n) refer to the total number of birds detected. all species. We detected 433% more Clapper Rails (r = 5.1, P = 0.014) on morning than on night surveys (Fig. IB). We did not ob- serve significant differences in number of in- dividuals detected between morning and night surveys for Least Bittern or Common Moor- hen. We did not observe a difference in tem- perature it = 0.3, P = 0.76) between morning (.Vj,^p = 25.4 ± 0.7° C) and night (T^emp = 25.1 ± 1.3° C) surveys. Percent cloud cover (r = 2.3, P = 0.10) and wind speed (t = 2.8, P = 0.066) were higher on morning (xcoud = 39 ± 17%, = 2.4 ± 0.3 km/hr) than on night = 23 ± 21%, = 0.8 ± 0.5 km/hr) surveys, however, the differences were only marginally signihcant. DISCUSSION Our results suggest that call-broadcast has little or no negative effect on observer detec- tion probability. These results support use of call-broadcast to maximize detection proba- bility during wetland-dependent bird surveys. Other studies have shown that observer detec- tion probability is higher during the call- broadcast period for Clapper Rails and similar between passive and call-broadcast periods for all other species (Conway and Nadeau 2006). We observed similar patterns although our sample sizes for individual species were small. Our estimate of observer detection probability (75%) is similar to previous esti- Nadeau et al. • DETECTION PROBABILITY OF WETLAND BIRDS 517 mates (50-75%) for wetland-dependent birds (Conway et al. 2004, Conway and Nadeau 2006). Our results suggest that trained observ- ers are missing 25% of the birds vocalizing during a survey that combines both passive and call-broadcast methodology. Thus, ac- counting for detection probability during sur- veys is essential if data will be used to esti- mate a measure of true abundance. Moreover, our individual estimates of observer detection probability (68 and 82%) differed between our two trained observers. Hence, estimating ob- server detection probability for each observer is important if data from multiple observers will be used to estimate population trends over time. We detected more birds on morning surveys compared to both night and evening surveys for all species detected. Higher temperatures in the evening may explain the decrease in the number of wetland-dependent birds detected between morning and evening surveys. Past studies examining the effects of temperature on the detection probability of wetland-depen- dent birds have reported conflicting results. Previous studies have reported a positive cor- relation between temperature and detection probability (e.g.. Mangold 1974, Tacha 1975, Spear et al. 1999) and others have shown a negative correlation (e.g.. Tango et al. 1997). However, authors of these studies failed to ac- count for the correlation between temperature and time of day or time of year (Conway and Gibbs 2001). Future studies that examine how weather affects detection probability need to first control for time of day. Robbins (1981) suggested that extreme heat reduces bird ac- tivity in other groups of birds. Higher wind speeds and percent cloud cover in the morning did not make morning less effective than night surveys. Robbins (1981) also reported two peak singing periods for birds (one in the morning and one in the evening) but the peak singing period in the morning was substan- tially longer. Thus, we may have detected more birds in the morning because we were able to survey more points on a route during the peak singing period. Our results suggest that surveying in the morning will maximize detection probability of wetland-dependent birds in the southeastern United States. We recommend that further studies be completed in other regions of North America to identify the optimal time of day to conduct surveys. Additional research is needed to examine the efficacy of night sur- veys for wetland-dependent birds that are rare or not present on SVI (e.g.. Black Rail, Yel- low Rail, and Virginia Rail). ACKNOWLEDGMENTS Monica Harris, Dale Shiver, and Charlotte Chumney provided logistical assistance on SVI. Robert Watt as- sisted with fieldwork. Wendy Gierhart provided GIS and logistical support. The manuscript benefited from comments provided by M. S. Woodrey and an anon- ymous reviewer. LITERATURE CITED Bookhout, T. a. 1995. Yellow Rail {Coturnicops nov- eboracensis). The birds of North America. Num- ber 139. Bystrak, D. 1981. The North American Breeding Bird Survey. Studies in Avian Biology 6:34-41. Conway, C. J. 2007. Standardized North American Marsh Bird Monitoring Protocols. Wildlife Re- search Report Number 2007-04. USGS, Arizona Cooperative Fish and Wildlife Research Unit, Tucson, USA. Conway, C. J. and J. P. Gibbs. 2001. Factors influ- encing detection probabilities and the benefits of call-broadcast surveys for monitoring marsh birds. Final Report. USGS, Patuxent Wildlife Research Center, Laurel, Maryland, USA. Conway, C. J. and C. P. Nadeau. 2006. Development and field testing of survey methods for a conti- nental marsh bird monitoring program in North America. Wildlife Research Report Number 2005- 1 1. USGS, Arizona Cooperative Fish and Wildlife Research Unit, Tucson, USA. Conway, C. J., C. Sulzman, and B. A. Raulston. 2004. Factors affecting detection probability of California Black Rails. Journal of Wildlife Man- agement 68:360-370. Gibbs, J. P. and S. M. Melvin. 1993. Call-response surveys for monitoring breeding waterbirds. Jour- nal of Wildlife Management 57:27-34. Gibbs, J. P, S. M. Melvin, and F A. Reid. 1992. American Bittern {Botauraus lenti^inosus). The birds of North America. Number 18. Grace, S. L. 2000. Final report of the vegetation sur- vey and map report project, St. Vincent National Wildlife Refuge. Apalachicola. Florida. USGS- USFWS Research Partnership Program Project #1 448-4 1650-97-N093. USGS. National Wetlands Research Center, Lafayette, Louisiana, USA. Johnson, R. R. and J. J. Dinsmore. 1986. The use of tape-recorded calls tt) count Virginia Rails and So- ras. Wilson Bulletin 98:303-306. Lynch, J. 1995. Effects of point-count duration, time- of-day, and aural stimuli on detectability of mi- gratory and resident bird species in Quintana Roo. 518 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 Mexico. Pages 1-6 in Monitoring bird populations by point counts (C. J. Ralph, J. R. Sauer, and S. Droege, Editors). USDA, Forest Service General Technical Report PSW-GTR-149. Pacific South- west Research Station, Berkeley, California, USA. Mangold, R. E. 1974. Research on shore and upland migratory birds in New Jersey: Clapper Rail stud- ies, 1974 Report. Accelerated Research Program, Contract Number 14-16-0008-937. USDI, Fish and Wildlife Service, Trenton, New Jersey, USA. Meanley, B. 1985. The marsh hen: a natural history of the Clapper Rail of the Atlantic coast salt marsh. Tidewater Publishing, Centreville, Mary- land, USA. Nichols, J. D., J. E. Hines, J. R. Sauer, F. W. Fallon, J. E. Fallon, and P. J. Heglund. 2000. A double- observer approach for estimating detection prob- ability and abundance from point counts. Auk 117:393-408. Reynard, G. B. 1974. Some vocalizations of the Black, Yellow and Virginia rails. Auk 91:747- 756. Robbins, C. S. 1981. Bird activity levels related to weather. Studies in Avian Biology 6:301-310. Spear, L. B., S. B. Terrill, C. Lenihan, and P. De- LEVORYAS. 1999. Effects of temporal and environ- mental factors on the probability of detecting Cal- ifornia Black Rails. Journal of Field Ornithology 70:465-480. Tacha, R. W. 1975. A survey of rail populations in Kansas with emphasis on Cheyenne Bottoms. Thesis. Fort Hays Kansas State College, Hays, USA. Tango, P. J., G. D. Therres, D. F. Brinker, M. O’Brien, E. T. Blom, and H. L. Wierenga. 1997. Breeding distribution and relative abundance of marshbirds in Maryland: evaluation of a tape playback survey method. U.S. Fish and Wildlife Service Grant #14-48-009-95-1280 Final Report. USDI, Fish and Wildlife Service, Office of Mi- gratory Bird Management, Denver, Colorado, USA. The Wilson Journal of Ornithology 120(3):5 19-524, 2008 MALE SONG VARIATION OF GREEN VIOLETEAR (COLIBRI THALASSINUS) IN THE TALAMANCA MOUNTAIN RANGE, COSTA RICA GILBERT BARRANTES,'-'' CESAR SANCHEZ,' BRANKO HILJE,^ AND RODOLFO JAFFE3 ABSTRACT. — We studied variation in acoustic and temporal characteristics of the static male song of the Green Violetear (Colibri thalassinus) in a single population in Costa Rica. The static song of 19 males was extremely variable. The song has two elements: the first was delivered exclusively at the beginning of each song while the second was present once, twice, or three times in the song of different males. Low frequency (LF), song duration (AT), and high frequency (HF) varied significantly among most individuals. The male population of Green Violetear has four song types that differ in acoustic and temporal characteristics. The great inter-male song variation suggests this type of vocalization may be under sexual selection. Received 2 February' 2007. Accepted 17 January 2008. Song variation among individual birds is well known for oscine and psittacid species (Farabaugh and Dooling 1996, Kroodsma 1996). Learning in these birds has an impor- tant role in syntaxes and structure of the song. This variation may have evolved to facilitate social interactions and/or by intra- or inter- sexual selection (Kroodsma 2004). Male re- productive success in birds is often associated with striking displays, such as complex songs, resulting from sexual selection (Catchpole 1982, Searcy and Yasukawa 1983, Johnsgard 1994, Kroodsma 2004). Individual song variation is relatively un- explored in hummingbirds, another song- leaming bird group (Baptista and Schuchmann 1990, Gaunt et al. 1994, Jarvis et al. 2000). Gaunt et al. (1994) showed that male Green Violetear {Colibri thalassinus) in neighboring populations share song types and similarity decreases with geographical distance. The ex- tent of intra-population variation in male song has not previously been reported for this hum- mingbird. Male Green Violetear show little (if any) ' Escuela de Biologi'a, Universidad de Co.sta Rica, Ciudad Universitaria, Co.sta Rica. ^ Institute Tecnoldgico de Costa Rica, R O. Box 159-7050, Cartage, Costa Rica. ^ Apartado 68941, Altamira, Caracas 1062, Vene- zuela. ■* Corresponding author; e-mail: gilbert.barrantes@gmail.com aggressive physical interactions with other males during the breeding season (e.g., darting chases) and visual displays are apparently ab- sent. Territorial males sing nearly continuous- ly during courtship from before dawn until sunset (Slud 1964, Feinsinger 1977). Males begin to sing in September and some continue until the end of March, investing up to 84% of daily time to this activity (Skutch 1967, Wolf 1976). The objective of our study was to describe the variation in male song features within a population of Green Violetear. METHODS We conducted fieldwork at the Estacion Biologica Cuerief, Talamanca Mountains, Costa Rica (09° 33' N, 83°40'W; elevation 2,600 m) during the dry season in January 2004. The area is dominated by oak {Quercus spp.) forest intermixed with several succes- sional growth stages with abundant flowering plants: Fuchsia paniculata. Bomarea costari- censis, Lamourouxia lanceolata, and Ceutro- pogon spp. We recorded the static songs (song uttered by perched birds) from 19 males sing- ing from exposed perches, ranging from 8 to 25 m in height, on a 1.5-km transect along the primary road (4 m wide); 12 males were re- corded one morning and seven the next morn- ing. wSinging males were separated by 20-100 m and perched at most 10 m into the forest (n = 17), facing the main road, or in a forest gap (// = 2). Each bird was recorded once for 519 520 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 at least 2 min or until the bird became silent. We recorded the hummingbirds sequentially as we walked along the road to avoid record- ing the same individual more than once. Be- cause of their high density, we could, in most cases, listen to the hummingbird previously recorded when we began to record the next focal individual. We considered singing males as territorial individuals following Skutch (1967); however, identifying territory limits was difficult, due to few aggressive interac- tions by males of this species. We define a male territory as a circular area of 20 m in diameter around the perch the singing male was most frequently using. We recorded songs using a Telinga Pro II Parabola and a Marantz PMD-222 tape re- corder, and analyzed the songs using the pro- gram CANARY Version 1.2.1 (Charif et al. 1993). Tapes were deposited at the Laborato- rio de Bioacustica, Escuela de Biologfa, Univ- ersidad de Costa Rica. Spectrograms were ob- tained with a filter bandwidth of 349.70 kHz, frame length of 256 points, and a sampling rate of 44,100 Hz. We identified the elements of the song, defining a song as a string of notes or elements, temporally separated from other such strings; an element was the shortest consistently recognizable trace on a sound spectrogram (Marler and Peters 1982). We randomly selected five songs separated by at least 10 sec to analyze the among-individual song variation from the total record of each hummingbird. We measured eight variables for each complete song: high frequency (HF), low frequency (LF), frequency range or band- width (AF) (HF-LF), peak frequency (PF, fre- quency with highest amplitude), peak time (PT, time with maximum amplitude), duration in seconds (AT), time elapsed between two consecutive songs (Tl), and time between the first and the second element (T2). We also quantified the number of songs and elements delivered in a random selected minute for each individual. We used a multiple analysis of variance (MANOVA) on six song variables (AT, LF, HF, PF, Tl, and T2) to examine the difference of these variables among individuals (based on a Discriminant Function Analysis (DFA), AF and PT were excluded because they did not differ among individuals). One-way anal- yses of variance and posteriori tests (Duncan) were conducted to examine which particular variables differ among individuals. We clas- sified the song of the hummingbirds into four types based on sonogram images, and used stepwise Discriminant Function Analysis (DFA) to test whether hummingbirds could be separated into four groups by their song fea- tures. We also recorded the following infor- mation during 6 days: non-systematic obser- vations of aggressive interactions and pres- ence of flower patches near perches of singing males. Means ± SD are presented. RESULTS The 19 male Green Violetears apparently held the same territories for at least 6 days as the same perches were occupied by singing males day after day and we assumed the same males were present. Territories included 2-3 perches from which males sang; they spent most of their time on one of them. The num- ber of males singing appeared to decrease from the forest border to the interior, except when a gap was present, in which case the number of males singing increased near the gap edge. Territories of the 19 recorded males varied by presence of flower patches used as nectar sources for hummingbirds. Twelve territories (main song perches) were within 10 m of large patches of Fuchsia paniculata, which were frequently visited by Green Violetears and Volcano Hummingbirds (Selasphorus flammula). Attacks between individual Green Violetears were rarely seen although the pres- ence of an individual in a flower patch occa- sionally triggered an attack by a singing male. The absence of obvious sexual dimorphism prevented us from knowing whether these few attacks were directed preferentially at males. Males sang from before dawn until dusk, only abandoning their perches to briefly visit nearby flower patches for feeding or to cap- ture insects on the wing with short sallies. The song of all males included two different ele- ments (Fig. 1). The first element (a) was pro- duced only once in the song of all individuals. Element (b) occurred once in the song of nine males (song type 1), twice in the song of sev- en individuals (type 2), three times in two oth- er males (type 3), and one or two times in songs produced by another male (type 4). On average, males produced 64.2 (± 8.8, range: Barrantes et al. • GREEN VIOLETEAR MALE SONG VARIATION 521 Unw (sec) EIG. 1. Spectrograms of the variation in number of elements, and structural characteristics of elements and songs of three male C. thalassinus. Element (a) was present only once, at the beginning of each male’s song, while presence of element (b) varied from one to three in songs of different males. 42-76.5) songs and 163.6 (± 33.7, range: 126-231) elements per minute. Structurally, the frequency range and high- est frequency were features of the static song that had the largest range (max-min values) for the 19 males: I 1.06 and 9.08 kHz respec- tively (Table 1 ). The time between the first and the second element (T2), followed by fre- quency range (AF), had the largest variation (CV) among individuals. Number of songs de- livered by individual per minute decreased with song duration (r = —0.41, n = \9, P = 0.08). However, the number of elements (r = 0.17, n = 19, P = 0.49) produced per minute was not related to AT. The large variation in AF was primarily caused by individual varia- tion in both HF and LF. We compared six characteristics of the complete song (AT, LF, HF, PF, Tl, and T2) among individuals using a MANOVA. The comparison of all song features showed sig- nificant differences among individuals (^108,414 ^ 21.66, P < 0.001). Significant dif- ferences were also found among individuals when variables were analyzed separately (one- way ANOVA and Duncan tests). Variables that differed among most individuals were LF ( 1 6 males: 76 = 37.64, P < 0.00 1 ), AT ( 1 4 males: = Al .21 , P < O.OOl ), and HF ( 13 males: F|«7^, = 77.56, P < O.OOl). Variation among individuals was lower for PF (8 males: = 6.32, P < O.OOl), Tl (4 males: = 2.59, P = ().0()2), and T2 (4 males: = 26.62, P < O.OOl). Male Green Violetears were divided in sub- groups based on structural and temporal char- acteristics of their song. DFA separated the males of this population into four groups 522 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 3, September 2008 TABLE 1. Characteristics of high frequency (HE), low frequency (LL), bandwidth (AL), peak frequency (PL), time duration (AT), time between two consecutive songs (TI), and time between the first and the second element (T2) for the complete song of 19 male Green Violetears in Costa Rica. HF (kHz) LF (kHz) AF (kHz) PF (kHz) AT (sec) Tl (sec) T2 (sec) Mean SD CV Range 10.98 2.19 19.90 9.04-18.12 3.16 0.57 17.98 2.08-4.82 7.82 2.36 30.17 4.28-15.34 6.11 0.81 13.31 4.05-8.53 0.58 0.10 17.73 0.42-0.95 0.37 0.03 8.86 0.30-0.51 0.33 0.15 44.41 0.069-0.84 (^15 240 ^ 11.98, P < 0.0001) based on six variables (AT, LF, HF, PF, Tl, and T2), and classified 82% of the songs correctly. DISCUSSION The difference in characteristics of the stat- ic song among male Green Violetears was high. This difference individualizes most sing- ing males in the population, suggesting a pos- sible role of inter- and intra-sexual selection (Morton 1986, Searcy and Andersson 1986); however, other factors such as age may also affect song characteristics. Frequency range in the song of C thalassinus is possibly the fea- ture from which the receiver (e.g., a female) obtains greater information, particularly when the song is composed by short, broad band- width elements or notes (Fig. 1, Table 1). This characteristic permits more precise location of singing males (Smith et al. 1978, Richards and Wiley 1980). Additionally, males producing songs with a wide frequency range may con- ceal the songs of other males, if these males’ songs have a narrower bandwidth that fit in part of the range of frequencies of other males. Despite the high variation in this population of male Green Violetears, groups of males produced similar song types. This suggests the presence of “different dialects” in a single population. However, causes of inter-group song differentiation are difficult to explain. Dialects usually evolve through geographic or microgeographic isolation (Kroodsma 1996) but, in this case, geographic isolation seems not to be the cause, since this species breeds in the area and then migrates to lower eleva- tion during the non-breeding season (Skutch 1967). In addition, young hear songs of adult males as males sing well beyond when fledg- lings abandon their nests (GB, pers. obs.). A possible explanation is the “song relearn hy- pothesis” proposed by Kroodsma (2004) to explain temporal song variation in dialects of the Three-wattled Bellbird (Procnias tricarun- culatus). This hypothesis proposes that adult birds “relearn” their songs throughout their life, and modify their songs through compe- tition (by imitation) with familiar rivals. For this hypothesis to be possible, groups of males have to hear and recognize neighboring males (Aoki 1989); this is the case with Green Vi- oletear. Consequently, at a particular time lag, different songs could be present in a single population. However, spatial distribution of song types is needed to begin testing this hy- pothesis. The information available allows us to compare variation of song traits of C. thal- assinus with other Trochilinae hummingbirds. The number of elements in the song of C. thalassinus is small compared to songs of Anna’s Hummingbird (Calypte anna) (Stiles 1982), Blue-throated Hummingbird (Lampor- nis clemenciae) (Ficken et al. 2000), Ame- thyst-throated Hummingbird (L. amethystin- us), and Green-throated Mountaingem (L. vir- idipallens). The number of elements is similar to Magnificent Hummingbird {Eugenes ful- gens) (Ornelas et al. 2002) and more complex (e.g., more elements and frequency modula- tion) than the static song of its congener Spar- kling Violetear {Colibri coruscans) (Gaunt et al. 1994). A characteristic present in all songs of this small sample of hummingbirds is the wide frequency range (bandwidth), although both species of the genus Colibri present the narrowest bandwidth. Song complexity is not higher in those species lacking visual displays, such as C. thalassinus, when compared to spe- cies with complex visual displays, such as C. anna and L. amethystinus. These results should be viewed with caution because num- ber of elements and acoustic structure of hum- Barrantes et al. • GREEN VIOLETEAR MALE SONG VARIATION 523 mingbird songs may be strongly influenced by environmental and phylogenetic features (Ir- win 1988, McCracken and Sheldon 1997). These aspects can be more closely analyzed within the monophyletic genus Colibri. When features of the song are compared between C thalassinus and C coruscans, the two high- land species of the genus (Gaunt et al. 1994), the static song of C. thalassinus has a higher number of elements. However, C. coruscans produces a “dynamic song” that males sing during a diving flight as part of the courtship display, which is absent in C. thalassinus. This supports Wagner’s (1954) suggestion that species lacking elaborate dynamic songs have complex static songs, and suggests the two highland species of the genus have evolved different courtship strategies. The other two species in the genus, C. delphinae and C. ser- rirostris are mid-elevation and apparently ter- ritorial (Stiles and Skutch 1989, Schuchmann 1999). Little is known about the vocalizations of these two species. ACKNOWLEDGMENTS We thank J. R. Eberhard, D. E. Kroodsma, C. E. Braun, and two anonymous reviewers for helpful and critical comments on the manuscript, and Luis San- doval for helping with the statistical analysis. We also thank the Universidad de Costa Rica for financial sup- port and The Organization for Tropical Studies and Estacion Biologica Cuerici for logistical support. LITERATURE CITED Aoki, K. 1989. A sexual-selection model for the evo- lution of imitative learning of song in polygynous birds. American Naturalist 134:599-612. Baptista, L. E and K. L. Schuchmann. 1990. Song learning in the Anna Hummingbird (Calypte anna). Ethology 84:15-26. Catchpole, C. K. 1982. The evolution of bird sounds in relation to mating and spacing behaviour. Pages 297-319 in Acoustic communication in birds. Pro- duction, perception, and design features of sounds (D. E. Kroodsma and E. H. Miller, Editors). Ac- ademic Press, New York, USA. Charif, R. a., S. Mitchell, and C. W. Clark. 1993. Canary 1.1 u.ser’s manual. Cornell Laboratory of Ornithology, Ithaca, New York, USA. Farabaugh, S. M. and R. J. Dooling. 1996. Acoustic communication in parrots: laboratory and field ex- periments of Budgerigars, Melopsittacus uniilatns. Pages 97-1 17 in Ecology and evolution of acous- tic communication in birds (D. E. Kroodsma and E. H. Miller, Editors). Cornell University Press, New York, USA. Eeinsinger, P. 1977. Notes on the hummingbirds of Monterverde, Cordillera de Tilaran, Costa Rica. Wilson Bulletin 89:159-164. Eicken, M. S., K. M. Rush, S. J. Taylor, and D. R. Powers. 2000. Blue-throated Hummingbird song: pinnacle on nonoscine vocalizations. Auk 117: 120-128. Gaunt, S. L. L., L. E Baptista, J. E. Sanchez, and D. Hernandez. 1994. Song learning as evidenced from song sharing in two hummingbird species {Colibri coruscans and C. thalassinus). Auk 111: 87-103. Irwin, R. E. 1988. The evolutionary importance of be- havioral development: the ontogeny and phylog- eny of bird song. Animal Behaviour 36:814-824. Jarvis, E. D., S. Ribeiro, M. L. da Silva, D. Ventura, J. ViELLiARD, AND C. V. Mello. 2000. Behaviour- ally driven gene expression reveals song nuclei in hummingbird brain. Nature 406:628-632. JoHNSGARD, P. A. 1994. Arena birds: sexual selection and behavior. Smithsonian Institution Press, Washington, D.C., USA. Kroodsma, D. 1996. Ecology of passerine song de- velopment. Pages 3-19 in Ecology and evolution of acoustic communication in birds (D. E. Kroods- ma and E. H. Miller, Editors). Cornell University Press, Ithaca, New York, USA. Kroodsma, D. 2004. The diversity and plasticity of birdsong. Pages 108-131 in Nature’s music. The science of birdsong (P. Marler and H. Slabbe- koorn. Editors). Academic Press, New York, USA. Marler, P. and S. Peters. 1982. Subsong and plastic song: their role in the vocal learning process. Pag- es 25-50 in Acoustic communication in birds. Production, perception, and design features of sounds (D. E. Kroodsma and E. H. Miller, Edi- tors). Academic Press, New York, USA. McCracken, K. G. and E H. Sheldon. 1997. Avian vocalizations and phylogenetic signal. Proceed- ings of the National Academy of Science USA 94: 3833-3836. Morton, E. S. 1986. Prediction from the ranging hy- pothesis for the evolution of long distance signals in birds. Behaviour 99:65-86. Ornelas, J. E, C. Gonzales, and J. Uribe. 2002. Complex vocalizations and aerial displays of the Amethyst-throated Hummingbird {Lampornis ametliystinus). Auk 1 19:1 141 — 1 149. Richards, D. G. and R. H. Wiley. 1980. Reverbera- tion and amplitude fluctuations in the propagation of sound in a forest: implications for animal com- munication. American Naturalist 115:381—399. Schuchmann, K. L. 1999. Family Trochilidae (hum- mingbirds). Pages 468-680 in Handbook of birds of the world. Volume 5. Barn Owls to humming- birds (J. Del Hoyo. A. Elliott, and J. Sargatal, Ed- itors). Lynx Edicions, Barcelona. Spain. Searcy, W. A. and M. Andf;rs.son. 1986. Sexual se- lection and the evolution of .song. Annual Review of Fxology and Systematics 17:507-533. Searcy, W. A. and K. Yasukawa. 1983. Sexual se- 524 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 lection and Red-winged Blackbirds. American Scientist 71:166-174. Skutch, a. F. 1967. Life histories of Central American highland birds. Nuttall Ornithological Club Num- ber 7. Cambridge, Massachusetts, USA. Slud, R 1964. The birds of Costa Rica. Distribution and ecology. Bulletin of the American Museum of Natural History 128:1-430. Smith, W. J., J. Pawlukiewicz, and S. T. Smith. 1978. Kinds of activities correlated with singing patterns of the Yellow-throated Vireo. Animal Behaviour 26:862-884. Stiles, F. G. 1982. Aggressive and courtship displays of the male Anna’s Hummingbird. Condor 84: 208-225. Stiles, F. G. and A. F. Skutch. 1989. A guide to the birds of Costa Rica. Cornell University Press, Ith- aca, New York, USA. Wagner, H. O. 1954. Versuche einer analyse der ko- libribalz. Zeitschrift fur Tierpsychologie 11:182- 212. Wole, L. L. 1976. Avifauna of the Cerro de la Muerte region Costa Rica. American Museum Novitates 2606:1-37. The Wilson Journal of Ornithology 120(3):525— 530, 2008 METABOLIZABLE ENERGY IN CHINESE TALLOW ERUIT EOR YELLOW-RUMPED WARBLERS, NORTHERN CARDINALS, AND AMERICAN ROBINS MICHAEL J. BALDWIN,'-^ WYLIE C. BARROW Jr.,‘ CLINTON JESKE,> AND FRANK C. ROHWER2 ABSTRACT. — The invasive exotic Chinese tallow tree (Triadica sebifera) produces an abundant fruit crop, which is primarily bird-dispersed. The fruit pulp of tallow is lipid-rich, high in saturated fatty acids, and con- sumed by many bird species. Long-chained fatty acids can be difficult for many birds to digest and we inves- tigated the ability of tallow consumers to assimilate energy in the pulp. We used the total collection method and compared apparent metabolizable energy (AME) of tallow fruit for three species of birds with differing fruit composition in their natural diets. All birds exhibited nitrogen deficits and lost body mass during the trials. Northern Cardinals {Cardinalis cardinalis) lost more mass (8.73%/day) than Yellow-rumped Warblers {Den- droica coronata) (5.29%/day) and American Robins {Turdus migratorius) (5.48%/day), and had larger nitrogen deficits (-120.1 mg N/g diet) than both species as well (-36.4 mg N/g diet and -68.9 mg N/g diet, respectively). Food intake relative to metabolic body mass was highest in Yellow-rumped Warblers (0.70 g-dry/g°’5-day). Northern Cardinal and American Robin food intake was lower and did not differ from each other (both species: 0.13 g-dry/g"^5-day). Nitrogen corrected values of AME were used to make species comparisons. Yellow- rumped- Warblers exhibited the highest values of AME (30.00 kJ/g), followed by American Robins (23.90 kJ/g), and Northern Cardinals (14.34 kJ/g). We suggest tallow may be an important winter food source for Yellow-rumped Warblers where their ranges overlap. Received 5 July 2006. Accepted 14 October 2007. The Chinese tallow tree {Triadica sebifera) is an invasive, non-native tree from Southeast Asia. It is now common in many areas along the northern Gulf of Mexico and south Atlan- tic coast (Scheld and Cowles 1981, Harcombe et al. 1993, Renne et al. 2000) and often exists in near monospecific stands (Baldwin 2005). It is primarily a bird-dispersed plant produc- ing an abundant fruit crop in the fall that can persist until early spring. The fruit contains a hard seed surrounded by waxy pulp, which is largely comprised of saturated fatty acids (42.5-75.1% of total fatty acids) (Khan et al. 1973, Raie et al. 1983, Xu et al. 1991). Many animals exhibit poor assimilation of high- melting point fatty acids (Scott et al. 1976) and their digestion may require specialized physiological traits (Place and Stiles 1992). Consequently, tallow may provide poor qual- ity food to most birds despite its high ener- getic value (33.5 kJ/g, range of other common plant species = 14.91-30.23 kJ/g; W. C. Bar- ' U.S. Geological Survey, National Wetlands Re- search Center, 700 Cajundome Boulevard, Lafayette, LA 70506, USA. ^ School of Renewable Natural Resources, Louisiana State University, Baton Rouge, LA 70803, USA. ’ Corresponding author; e-mail: michael_baldwin@usgs.gov row Jr. and C. Jeske, unpubl. data) and con- sumption by many species (24 species: Con- way et al. 2002; 64 species: W. C. Barrow Jr. and W. R. Fontenot, unpubl. data). This issue has become more significant with the increas- ing expansion of tallow in the region. We conducted feeding trials on three spe- cies of birds with varying diets all known to consume tallow fruit. Yellow-rumped War- blers (Dendroica coronata) were selected be- cause they were the most frequent tallow con- sumer in coastal Texas (Conway et al. 2002) and have a specialized digestive system that allows them to assimilate waxy foods such as wax myrtle {Morelia cerifera) and northern bayberry (M. pensylvanica) fruit (Place and Stiles 1992). They were also the most abun- dant bird observed in tallow-dominated wood- lands during winter in southwest Louisiana (Baldwin 2005). Yellow-rumped Warblers can assimilate high-melting point fatty acids, but they prefer unsaturated to saturated fatty acids (McWilliams et al. 2002). American Robins {Turdus migratorius) and Northern Cardinals {Cardinalis cardinalis) were selected based on their relatively high consumption of tallow fruit (Renne et al. 2000, 2002). These two spe- cies were the second and third most abundant birds detected in tallow woodlands during 525 526 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 winter (Baldwin 2005). Neither species is known to have specialized digestive traits similar to the Yellow-rumped Warbler. Amer- ican Robins have exhibited relatively high di- gestive efficiencies of lipids (Witmer and Van Soest 1998, Lepczyk et al. 2000) and Zurov- chak et al. (1999) found no change in absorp- tion efficiency with increasing lipid levels in a test diet. These studies did not test for dif- ferences in fatty acid composition and the di- ets examined were comprised primarily of un- saturated fatty acids. Zurovchak (1997) mea- sured food preferences of American Robins offered synthetic fruits high in either saturated or unsaturated fatty acids and observed pref- erences for the latter by most individuals test- ed; none preferred saturated fatty acids. Zu- rovchak (1997) suggested the few individuals that exhibited no preference were likely se- lecting fruit based on color. Willson and Har- meson (1973) studied Northern Cardinal seed preferences and found no correlation between seed preference and lipid concentration. No information is available on Cardinal fruit pref- erences based on lipid content or fatty acid composition, but they are known to consume high-lipid fruit (Fontenot 1998, Borgmann et al. 2004). Our objective was to examine how well Yellow-rumped Warblers, Northern Cardinals, and American Robins could metabolize the available energy in tallow fruit during con- trolled feeding trials. We hypothesized that Yellow-rumped Warblers would assimilate a greater amount of the available energy than Northern Cardinals and American Robins be- cause of their diet and specialized digestive traits. METHODS Capture and Care of Birds. — Yellow- rumped Warblers {n = 9), Northern Cardinals {n = 10), and American Robins {n = 5) were captured near Lafayette, Louisiana in the win- ters of 2001-2002 and 2002-2003, and trans- ported to an indoor environmental chamber where they were kept under controlled tem- peratures (20° C) and photoperiod (10.5L: 13.5D). Birds were housed individually in 30 X 36 X 36 cm metal cages. Water and main- tenance diet were provided ad libitum. The diet consisted of a commercial mash designed for captive birds (Avian-Fare and/or Soft- billed-Fare) mixed in a plastic food cup with tallow fruit and mealworms. In addition, Yel- low-rumped Warblers were supplemented with fruit from wax myrtle and poison ivy {Toxicodendron radicans) while American Robins and Northern Cardinals were given fruit from hackberry {Celtis laevigata) and de- ciduous holly {Ilex decidua). Northern Cardi- nals were also provided sunflower {Helianthus spp.) and safflower {Carthamus tinctorius) seeds. The total collection method (Servello et al. 2005) was used to measure apparent metabo- lizable energy (AME) in tallow fruit. Birds were allowed a minimum of 12-14 days to acclimate to captivity prior to onset of AME trials and no individual was used more than once. Birds were only offered tallow fruit dur- ing feeding trials. Whole fruits (seed included) were fed to Northern Cardinals and American Robins, while Yellow-rumped Warblers were presented with pulp only. Yellow-rumped Warblers feed on tallow by pecking or scrap- ing off the waxy, outer layer. Pulp was re- moved from tallow fruit using a dissecting needle or by rubbing fruit over a 2-mm sieve. Northern Cardinals and American Robins did not digest the seed, and all AME estimates were calculated for pulp only. Fruit was placed into a plastic food cup and weighed prior to each feeding. The first day of the AME trial was used to remove nonex- perimental foods from the digestive tract and no fecal material was collected. Fecal material was collected daily after day one; the remain- ing food was removed and weighed, including spilled food, and fresh food was weighed and provided to the birds. Food and fecal samples were placed in scintillation vials, vacuum- dried, and stored at -80° C. Gross energy in food and fecal samples was measured with a Parr 1261 Isoperibol Bomb Calorimeter. Values for AME may be influenced by ni- trogen loss/gain flux during feeding trials caused by protein tissue growth/catabolism (Scott et al. 1976), particularly for low-protein foods such as fruit. This affects endogenous urinary nitrogen levels in an individual’s feces and zero nitrogen balance AME calculations were needed. Percent total nitrogen in food and fecal samples was measured with a Ther- mo Finnigan® FlashEA 1112 Elemental Ana- lyzer. Previous studies have used the Kjeldahl Baldwin et al. • METABOLIZABLE ENERGY IN TALLOW FRUIT 527 method to measure nitrogen composition (Scott et al. 1976, Carl and Brown 1985, Ser- vello et al. 2005). Our method was chosen be- cause it is as accurate as the Kjeldahl method, allows for smaller sample size, and requires less preparation and run time (Matejovic 1995). Amount of nitrogen lost/retained was measured as the difference in nitrogen con- sumed via tallow pulp and the amount of ni- trogen in the feces. Nitrogen corrected AME values were calculated by: (AME (kJ/g) = 4.1841 -{(F-FE) - [(X-XE + 8.22-(NF - NE)]}/F), where F and X equal the dry weights (g) of ingested food and excrement, FE and XE equal the energy (kcal/g) of in- gested food and excrement, 8.22 is the energy constant (kcal/g) of nitrogen in uric acid, and NF and NE equal the grams of nitrogen in the diet and excrement, respectively (Scott et al. 1976). Statistical Analyses. — Dry food intake rel- ative to metabolic body mass (g/g°^^-day), body mass change, and nitrogen corrected AME values (kJ/g) for tallow pulp were com- pared using nonparametric Kruskal-Wallis tests (Conover 1980) in SAS (Proc NPARIWAY, SAS Institute Inc. 1999). Spe- cies differences in nitrogen flux (mg N/g) were compared with ANOVA. Contrasts were performed on ranks with Tukey’s pairwise comparisons (Proc MIXED, SAS Institute Inc. 1999) when the null hypothesis of no differ- ence among species was rejected (a = 0.05). One Northern Cardinal exhibited a negative AME; it was treated as an outlier and exclud- ed in the analysis. RESULTS Mean (± SD) gross energy and percent to- tal nitrogen in tallow pulp were 34.43 ±0.12 kJ/g and 0.32 ± 0.004%, respectively. Food intake (g/g^'^^-day) varied among species (x^ = 1 1.14, P = 0.004) and was highest in Yel- low-rumped Warblers, but did not differ be- tween Northern Cardinals and American Rob- ins (Table 1). All individuals lost body mass during the trials and exhibited nitrogen defi- cits. Mass loss differed among species (x^ = 1 1.71, P = 0.003) and was highest in North- ern Cardinals (Table 1 ). The amount of nitro- gen lost per gram of tallow consumed was not equal between species (P2.20 “ 29.4, P < 0.001), and was highest in Northern Cardinals £ C/3 ^ ^ g 1 (U -o a. c UJ F. (U C« a i: .S (U c ■“ !- O (U c fc o — U (U = I M) ^ O ^ - fc = ‘I g yi O tzi C ^ c m 15 ^9 X) c O II D X) O (U ^ i: Vi c 10 ha) patches, and vary in habitat quality (Lloyd and Marsden in press). These patches continue to be eroded due to ongoing firewood collection and burn- ing (Fjeldsa and Kessler 1996). Consequently, some Polylepis bird species are considered globally-threatened (BirdLife International 2004) and are presumed to have extremely small populations in the tiny forest patches that remain (Fjeldsa 1993). If factors that af- fect resource partitions in remnant forest patches can be identified, this information could enable biologists to predict with greater accuracy the responses of Polylepis birds to further habitat loss and degradation, and pro- vide a link to the mechanisms that could be used for active habitat restoration strategies. The objectives of this paper are to describe (1) foraging patterns, and (2) guild composi- tion of the insectivorous Polylepis bird com- munity in remnant forest patches in the Cor- dillera Vilcanota, southern Peru. METHODS Study Sites. — I recorded the foraging ecol- ogy of Polylepis bird species at three sites in the Cordillera Vilcanota mountain range (Fig. 1). Mantanay (13° 12' S, 72° 09' W) at 3,400- 4,500 m elevation, above the village of Yan- ahuara, is one of the largest areas of Polylepis racemosa woodland in the Cordillera Vilcan- ota. The site was surveyed during 67 field days in July 2003, October 2004, and Septem- ber 2005. Yanacocha (13° 17' S, 72° 02' W), -3,700-4,500 m elevation, above the village of Huaocari, was surveyed for 28 days in Oc- tober 2003 and June 2004. Laguna Queuna- cocha (13° 12' S, 72° 10' W) is a small area of P. pepei woodland, -4,200-4,500 m eleva- tion above the village of Huilloc (hereafter re- ferred to as Huilloc), and was surveyed during 22 field days in December 2003 and July 2004. Foraging Observations. — Polylepis birds were located visually and vocally foraging with or independently of mixed-species flocks in both large (>10 ha) and small (<1.0 ha) forest patches. Individual birds of all species were followed opportunistically between 0530 and 1730 hrs for as long as they could be kept in view. Sequential foraging observations for each individual were made using 8 X 42 bin- oculars, recorded on micro-cassette, and later Lloyd • FORAGING ECOLOGY OF POLYLEPIS BIRDS 533 transcribed to data sheets. Many individuals of each species were sampled reducing prob- lems associated with pseudoreplication. No more than three consecutive sequences of for- aging events were taken of each individual to further limit bias (Morrison 1984, Kratter 1995). The point of each foraging sequence was marked with colored tape. Variables estimated and recorded at the point of each foraging se- quence were species, height above ground, and height below the canopy. Vegetation cov- er at each marked foraging point was esti- mated using a sighting tube (James and Shu- gart 1970) once the bird had moved from view. Terminology and strategy for recording prey-attack maneuvers (Table 1 ) follows Rem- sen and Robinson (1990). This scheme rec- ognizes the importance of links between the prey-attack maneuver and substrate used by the bird, and places importance on the posi- tion from which the prey-attack-maneuver is launched (Remsen and Robinson 1990, Mac Nally 2000). Bird Counts. — I examined the influence of bird abundance on foraging ecology by sur- veying birds at census plots along transects within large and small remnant patches using a point count distance sampling method (Reynolds et al. 1980 as adapted by Jones et al. 1995). Each transect was randomly placed with census stations positioned 150 m from each other (Blake and Karr 1987). I used en- counter rates of birds recorded within 50 m of the census stations’ central point as a measure of relative abundance. Bird surveys were con- ducted between 0530 and 1630 hrs and only during hours of suitable weather (i.e., in the absence of snow, rain or strong wind). I was the only recorder for the bird surveys and was familiar with the vocalizations of the region’s bird species. I sat quietly for a 5-min ‘settling down’ period (Bibby et al. 2000) following arrival at each plot, before spending 20 min recording all birds seen and heard. Two rep- licates of each station were made in each patch size category making a total of 180 point counts (100 in large patches and 80 in small patches). The direction of surveys along transects was rotated in an attempt to mini- mize any bias associated with variable bird activity at different times of day. Data Analyses: Guild Composition. — Se- 2 ^3 Dl, P- Q, Cu 3 C -is C 3 D 3 C3 § S (U jz i c c T3 cd CO ^ (U (U CL 3 I (L) bX) bXX bX) OJO J- Cu CC X ^ 3 0 c3 ^ CC CO 3 C (U 3 i I 3 .S O "3 .3 t B %c C 3 .3 Li. 3 c -3 3 3 3 S £ cz o o J < (U 3: u 3 3 -3 > > y y -i;’ y y i; uj ^ o o 0^ 534 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 quential observations rather than single point observations are the preferred method for sampling foraging behavior of birds from a large sample of individuals because they have a greater tendency to reveal rare behaviors (Hertz et al. 1976. Eckhardt 1979, Holmes et al. 1979, Moermond 1979, Morrison 1984). I used 30 sequential foraging events as the min- imum sample size following Morrison (1984). Twenty-three high Andean bird species were recorded during the surveys for which suffi- cient sample sizes were collected for 12 spe- cies. Two of these 12 species were omitted from further analyses. The Royal Cinclodes {Cinclodes aricomae) was observed using the ground or boulders as a perch type and moss or bark-ground litter as substrate and was, therefore, separated from the other bird spe- cies by perch type, substrate, and niche posi- tion. The Thick-billed Siskin (Carduelis cras- sirosths) was observed feeding on either flowers or seeds of Polylepis trees and was not considered to be an insectivore. I selected 100 random point (individual) foraging maneuvers and 100 sequential for- aging maneuvers for each species to examine any possible bias of temporal dependence in sequential observations (e.g.. Bell et al. 1990) and tested for significant differences using Mann-Whitney f/-tests. Only two of 12 spe- cies examined had significant differences in foraging height above ground; White-browed Tit-spinetail {Leptasthetnira xenothorax) (U = 3.576, P < 0.05) and Cinereous Conebill {Conirostriim cinereiim) (U = 2,373, P < 0.001). and only the Giant Conebill (Oreo- manes fraseri) had a significant difference {U = 3,621, P < 0.01) in foraging height below the canopy. I attributed these results to chance alone, since only 5% of tests were significant, and included all sequential foraging observa- tions in all subsequent analyses. I used a hierarchical cluster analysis using X" measures of between-group linkage to group species into distinctive guilds based on frequency use of perch type, substrates, and prey-attack maneuvers for each of the 10 bird species. 1 defined guild as a group of species (>1) that exploit the same class of environ- mental resources in a similar way (Root 1967, Morrison and Hall 2003). Guilds were as- signed by visual inspection of the dendrogram since currently no statistical precedent exists for quantitative separation of species into guilds (Holmes et al. 1979). Differences in foraging (niche) position between species in each guild were examined using Mann-Whit- ney U and Kruskal-Wallis tests. Differences in frequency of use of each perch type, substrate, and prey-attack maneuver between species in each guild were examined with tests. Standardized residuals in contingency ta- bles were analysed to learn whether birds for- aged disproportionately at certain perch types or substrates, or disproportionately used cer- tain prey-attack maneuvers (Kratter 1995, Gabbe et al. 2002, Hartung and Brawn 2005). Sequential Bonferroni adjustments of P values for overall X“ values were not made because they are considered over-conservative with lit- tle consensus as to their application for eco- logical studies (Cabin and Mitchell 2000. Moran 2003, Driscoll and Wier 2005). Cotnparison of Foraging Regimes in Dif- ferent Patch Size Categories. — Eight bird spe- cies were selected for analysis based on a minimum sample size of 50 foraging sequenc- es in each patch size category. Twenty random samples of niche position, substrate use, and prey-attack manoeuvres for each of the eight species in both large and small patches were selected. 1 defined niche position as mean ± SD foraging height range (foraging height above ground and foraging height below can- opy). 1 used J' evenness scores to examine whether a significant shift in foraging regimes occurs for each species (e.g.. Morse 1971. Yeaton 1974) within each guild between large and small patches. I calculated J' evenness scores of total number and frequency of use of perch types, substrates, and prey-attack ma- neuvers as measurements of niche breadth, substrate breadth and breadth of prey-attack maneuvers, and tested for significant differ- ences using Mann-Whitney U-tests. Influence of Bird Abundance on Foraging Regimes. — Spearman’s rank correlation anal- yses were used to examine the influence of bird abundance on foraging regimes. I exam- ined the relationship between density shift of each species between large and small patches (abundance in large patches minus abundance in small patches) with the difference in rich- ness and breadth of both substrates and prey- attack maneuvers. I then examined the rela- tionship of species tolerance of small patches Lloyd • FORAGING ECOLOGY OF POLYLEPIS BIRDS 535 0 5 10 15 20 25 A. alpinus Ar parvlus Arboreal sally-gleaners 1 Arboreal foliage-gleaners C, cinereum C, fenvgineivantra L. yanacensis X parina L xenothorax L. and/cofa Arboreal bark-foliage gleaners C. a/bicaplUa O ft^asen ^^nd^tory bark-subsurface gleaners FIG. 2. Hierarchical cluster analysis with dendrogram (using measures of between-group linkage) of guild composition of 10 species of insectivorous bird species in Polylepis forests based on percent (0-25) use of perch types, substrate, and attack maneuvers. with differences in richness and breadth of substrates and prey-attack maneuvers. Species tolerance was defined as species abundance in small patches. RESULTS Guild Composition. — Individuals (1,122) of 10 species followed during 110 days in the field corresponded to 8,694 foraging events. Guild 1 (Fig. 2) arboreal sally- gleaners, in- cluded Ash-breasted Tit-tyrant {Anairetes al- pinus) and Tufted Tit-tyrant {A. parulus) that fed primarily on the inner- and outer-most leaves of live Polylepis trees using principally sally-strike and perch glean prey-attack ma- neuvers. Guild 2, arboreal foliage gleaners, included both Cinereous Conebill and White- browed Conebill (Conirostrum ferrugineiven- tre), which fed principally on the outer-most leaves of Polylepis trees using mainly hang- up and hang-down prey-attack maneuvers. Guild 3, arboreal bark-foliage gleaners, in- cluded three Leptasthenura Tit-spinetail spe- cies and Tit-like Dacnis (Xenodacnis parina) that fed on the inner- and outer most branches and foliage using hang-up glean as the dom- inant prey-attack maneuvers. Guild 4, under- story bark-subsurface gleaners, included the Creamy-crested Spinetail {Cranioleuca albi- capilla) and Giant Conebill that fed mainly on the surface and subsurface of trunks and branches in the understory of live Polylepis trees using similar surface prey-attack maneu- vers, but different subsurface prey-attack ma- neuvers. Within-guild Niche Position and Niche Breadth: Arboreal Sally-gleaners. — I found a difference {P < 0.001) in niche position and mean foraging vegetation cover between the two Anairetes species in this guild (Table 2). The Tufted Tit-tyrant foraged higher above ground and nearer the canopy than the Ash- breasted Tit-tyrant, but had a narrower niche position and niche breadth. The Ash-breasted Tit-tyrant foraged in a narrow range of dense Polylepis vegetation. Arboreal Foliage Gleaners. — There were differences in mean height below the canopy {P < 0.001) but not in mean height above ground or mean vegetation cover at foraging sites for both arboreal foliage gleaners (Table 2). The White-browed Conebill had the more narrow niche position in foliage closer to the ground, a more narrow niche breadth, and for- aged within a narrow range of dense vegeta- tion cover. The Cinereous Conebill had the more narrow niche position for mean foraging height below the canopy. A rb o re a I B a rk -fo I iage Gleaners. — There was a difference (P < ().()() 1) in the niche po- 536 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 o ^ c 0) t. > ,; ~ St) >> ^ ^ ^ -F f^, IT) Tf (y, O' sc O' — CM ir, — Tt IT) sc os sc X O' II - a. +1 +1 r-; q o rn r- V a, vC q q q — (N sC c^ lE ^ +1 +1 +1 +1 ~ sC in r<~. '' tn ri — 50 m from the forest’s edge were considered to be in the interior of the forest fragment (Cahill and Matthysen 2007). OBSERVATIONS Breeding Season and Behavior. — We found one nest with the parents incubating on 30 Oc- tober 2003 in Fragment A. We found six ad- ditional nests in mid November (Fragment A, 2 nests; Fragment C, 1 nest) and early Decem- ber (Fragment J, 1 nest; Fragment B, 2 nests) during the breeding period. No active nests were found from January to September 2004. Both parents took turns incubating the eggs. Two young hatched by 5 November and both parents provisioned them. Observations of four of six nests showed that one parent, act- ing as a sentinel, perched on a high branch of a Polylepis tree until the other returned with food, approaching the nest by short flights and jumps between Polylepis branches. The par- ents did not emit songs while approaching the nest or during the provisioning process. Chicks made loud chirps when a parent ap- proached and, when close to fledging, they moved to the edge of the nest rim to be fed. Both parents removed fecal sacs. Nests. — All nests were open cups (Fig. 1). The external portion of each was composed of twigs and sticks interwoven tightly with soft plant material. Three collected nests had a mean (± SD) mass of 201 ± 44 g with an FIG. 1. Schematic drawing of the nest of the Giant Conebill. average external height of 9.5 ± 0.7 cm and an average width of 20 ± 1 cm. The internal egg cup had a diameter of 9 cm and a depth of 5 cm. The bodies of hatchlings were 100% covered within this space. All nests were supported by 3-5 branches of a Polylepis tree. Nests were placed within the space surrounded by branches and secured by twigs and grasses. The tree branches of three nests were interwoven tightly with the nest materials and the branches were used as part of the nest structure. The nest material used consisted mostly of Polylepis parts. Additionally, 21 species of plants were used in smaller amounts, 76% of which were mosses (Table 1). The nest had an external layer with a structural and a decora- tive part. The lower, structural part was com- posed of twigs, sticks, and grasses. The upper decorative part, which was most visible, was comprised of many types of mosses (densely interwoven), flakes of thin Polylepis bark, and Polylepis leaves. This layer gave the nest a dark greenish color. The internal structural layer had Polylepis bark strips with finer twigs, some of which were partly covered with small amounts of mud. The nest lining consisted of two parts. The outermost part had mosses, stems of Ca- jophora sp. (Loasacaeae), thin flakes of Po- lylepis bark, and parts of Cyperaceae and Po- aceae (roots, heads, and stems). The innermost part, which came in contact with the chicks, contained bird feathers (most likely from tin- amou such as Nothura darwini) and sheep wool (Table 1). Cahill et al. • GIANT CONEBILL NESTING BIOLOGY 547 TABLE 1. Plant materials in the nest of the Giant Conebill in a Polylepis besseri forest in Sacha Loma and Cuturi (Cochabamba, Bolivia). Material Layer Species Twigs and sticks External and inter- nal structural Polylepis besseri, Ribes sp. Leaves External decorative Polylepis besseri Roots External decorative Cyperacea Mosses External decorative Neckera complanata, Bartramia potosica, Papillaria c.f. nigrences, Drepanocladus revolvens, Neckera andina c.f., Racomitrium crispipilum, Hedwigidium integrifolium, Syntrichia princeps, Bryum andicola, Macromitrium sp., Lekea sp., Breutelia sp., Cyclodietyon sp., Leptodontium sp. Lining Metzgeria sp., Frulania sp. Grasses External structural and lining Festuca sp., Deyeuxia sp. Bark External decorative and internal structural Polylepis besseri Stems Lining Ribes sp., Cajophora horridus Nestlings. — One nest found in the incuba- tion stage contained two white eggs with tiny brown specks with a mean (± SD) weight of 3.0 ± 0.2 g. Three of the six nests found in the nesting period had two chicks, two had one chick, and one had three chicks. We found a dead chick shortly after the start of nest monitoring in two of the three nests with two hatchlings. The average mass and tarsus length of chicks at an average age of 13 ± 2 days {n = 11) were 21.01 ± 1.18 g and 21.95 ± 0.94 mm, respectively. Chicks at this stage were close to fledging since most pin feathers had become feathered. The young {n = 11) fledged 14-16 days after hatching. Fledglings had a conspicuous large white cheek patch, white supercilium, and retained the yellow bill flanges. They were easily identifiable up to 7 weeks, even in mixed flocks. We observed the supercilium of juveniles began to turn russet after this period. Juveniles remained with both parents during the first 3 weeks post fledging. We observed them on six occasions after 5 weeks in mono-specific flocks of 4-7 individ- uals; in one case, two juveniles were with both parents and one unknown adult. Nest Site. — The seven nests were in the in- terior (>5() m from the edge) and relatively dense parts of the forest fragments where some boulders with abundant lichens and mosses were also present. The average (± SD) height of trees in which nests were placed was 3.16 ± 0.56 m. The height of nests above ground was 2.43 ± 0.25 m and nests were well camouflaged inside dense foliage. DISCUSSION The breeding season of the Giant Conebill extended from October to December, which coincided with the beginning of the rainy sea- son (Nov-Mar) with 10° C mean monthly temperatures (Fernandez et al. 2001). This timing may be related to an increase in the amount of food resources (insects living in Polylepis trees) during and especially after the nestling period. This also coincides with re- cords of Giant Conebill fledglings in Puno (Nov) and Cusco (Dec), Peru (Fjeldsa and Krabbe 1990). Herzog et al. (2003) similarly observed that most Polylepis-'mh'dh'ximg bird species had peak breeding activity in October and November. However, Gonzalez and Tor- res-Mura (2000) report one active nest of the Giant Conebill on 2 March near Belen (Pari- nacota Province in Chile), which would be at the end of the rainy season. The shape of the nest (open cup) follows the pattern found in most Thraupidae species for which nests are known (Fjeldsa and Krab- be 1990, Isler and Isler 1999). This character- istic of the Giant Conebill is shared with the phylogenetically related species from the Con- i rostrum genus ((’. speciosunu C. hicolor. C. cinerenm, C. tamarugense) (Estades and L6- pez-Calleja 1995, Ribon and Simon 1997). However, size and rim thickness (5.5 cm) re- 548 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 corded for the nest of Giant Conebill, are much larger than those reported for nests of the Tamarugo Conebill (Coni rostrum tamci- rugense) (width 6.5-9 cm, rim 1.5-2 cm) (Es- tades and Lopez-Calleja 1995) and the Chest- nut-vented Conebill (C. speciosum) (width 6.2 cm, rim 0.85 cm) (Ribon and Simon 1997). The large size of the nest of the Giant Conebill is primarily the result of the rim thickness (55% of total width). This could be to assure support and, perhaps more insulation as nest microclimate is important for chick de- velopment (Dawson et al. 2005). The densely interwoven mosses in the upper part of the external layer filled the spaces among the twigs and sticks, probably to improve the ex- ternal structure and increase protection against wind. The presence and thickness of the lining layers of mosses, bark, roots, stems, feathers, and sheep wool in the interior of the nest seemed to additionally insulate the chamber. Temperatures in the high Andes and within Polylepis forest can drop drastically at night or after rainfall (from 12° to 1°C; J. R. A. Cahill, pers. obs.). A nest that is well insulated would not be subject to drastic heat loss. The depth of the nest chamber (5 cm) could also protect chicks from heat loss. The depth of this chamber in the nest of the Giant Conebill in Polylepis rugulosa is also mentioned by Gonzalez and Torres-Mura (2000), but no di- mensions are given. Use of Polylepis twigs and bark as nesting materials by the Giant Conebill was also re- ported by Gonzalez and Torres-Mura (2000) underlining its strong relation with the Poly- lepis habitat and its designation as a Polylepis specialist (Fjeldsa and Krabbe 1990, Fjeldsa and Kessler 1996, Fjeldsa 2002). The use of mosses suggests search for a material that is less abundant than that available from Poly- lepis, but which is most probably present in the interior of forest fragments, which are more humid and the canopy is more dense than at edges (Fernandez et al. in prep.). Co- incidentally, nests were only found in the in- terior of forest fragments. Apparently, in frag- ments of P. besseri, these areas are selected and more intensively used by Giant Conebills (Cahill and Matthysen 2007). The Giant Conebill uses mostly plant material for nest construction, which agrees with the phylogenetically related species: Tamarugo Conebill and Chestnut-vented Conebill (Estades and Lopez-Calleja 1995, Ribon and Simon 1997). Use of large amounts of twigs and leaves of a specific tree for the nest of the Giant Cone- bill is similar to the specific use of the rachis of leaves of Tamarugo (Prosopis tamarugo) trees in the nest of the Tamarugo Conebill (Estades and Lopez-Calleja 1995). Moss material in nests of the Giant Conebill has also been recorded for nests of the Moss-backed Tanager (Bangsia ed- wardsi) (Robbins and Glenn 1988) and the Par- adise Tanager (Tangara chilensis) (Wood et al. 1992). The nests we found were well camouflaged in P. besseri trees. This has also been reported by Gonzalez and Torres-Mura (2000), who even mention that observations of the parents at the nest were difficult due to the branches and leaves obstructing vision. Nest camou- flage could be a strategy to avoid predators that hunt visually. Predation by avian preda- tors, such as the Yungas Pygmy-owl (Glau- cidium bolirianum), seems to be an imminent risk for the Giant Conebill since it forms and leads flocks (Herzog et al. 2002). Substrate trees (height 3.16 m) selected by the Giant Conebill were least abundant in three of the four fragments where nests were found. It is possible the largest trees in these fragments, which usually have few branches near the ground, were chosen for terrestrial predator avoidance, although no signs of pre- dation were observed at nests during this study. This apparent selection for tall substrate trees by the Giant Conebill is not found in the phylogenetically related Tamarugo Conebill (Estades and Lopez-Calleja 1995). The clutch size of this species of two coincides with most Thraupidae (Isler and Isler 1987) and with birds at this latitude (Skutch 1985). The nest- ling period of 14-16 days is similar (within 1-2 days) to the Tangara species for which similar data exist (Isler and Isler 1987, Long and Heath 1994). Human induced fires and logging are the most prevalent threats for Polylepis forests at present in the High Andes (Kessler and Driesch 1993, Hensen 2002). Bark, thin twigs, and mosses are likely to be damaged during fires. Pruning also induces open spaces in the canopy and probably a reduction of mosses. Probable nest sites and abundance of nest ma- terial could be negatively affected as a con- Cahill et al. • GIANT CONEBILL NESTING BIOLOGY 549 sequence. Protection of the habitat of the Gi- ant Conebill is necessary because it is a near threatened species of a monotypic genus. ACKNOWLEDGMENTS We thank the people of Sacha Loma and Cuturi for allowing us to work in their forest, and Greissy Arriaran, Claudia Salazar, Lenny Terceros, and Dania Jarro for as- sistance during field work, and the guidance of Magaly Mercado for plant identification. This study was funded by the Flemish Interuniversity Cooperation (VLIR-IUC), Belgium and the Centro de Biodiversidad y Genetica, Universidad Mayor de San Simon, Bolivia. LITERATURE CITED BirdLife International. 2006. Species fact sheet: Or- eomanes fraseri. http://www.birdlife.org (accessed 15 September 2006). Burns, K. J., S. J. Hackett, and N. K. Klein. 2003. Phylogenetic relationships of neotropical honey- creepers and the evolution of feeding morphology. Journal of Avian Biology 34:360-370. Cahill, J. R. A. and E. Matthysen. 2007. Habitat use by two specialist birds in High-Andean Polylepis forests. Biological Conservation 140:62-69. Dawson, R. D., C. C. Lawrie, and E. L. O’Brian. 2005. The importance of microclimate variation in determining size, growth and survival of avian offspring: experimental evidence from a cavity nesting passerine. Oecologia 144:499-507. Estades, C. E and M. V. Lopez Calleja. 1995. Eirst nesting record of the Tamarugo Conebill {Coni- rostrum tamanigense). Auk 112:797—800. Fernandez, M., J. R. A. Cahill, E. Martinez, and J. M. Lazcano. In Prep. Disturbance and edge ef- fects in High-Andean Polylepis besseri forests. In prep. Fernandez, M., M. Mercado, S. Arrazola, and E. Martinez. 2001. Estructura y composicion flor- istica de un fragmento boscoso de Polylepis bes- seri Hieron subsp. besseri en Sacha Loma (Co- chabamba). Revista Boliviana de Ecologia y Con- servacion Ambiental 9:15-17. Fjeldsa, j. 2002. Key areas for conserving the avifau- na of Polylepis forests. Ecotropica 8:125-131. Fjeldsa, J. and M. Kessler. 1996. Conserving the bi- ological diversity of Polylepis woodlands of the highlands of Peru and Bolivia — a contribution to sustainable natural resource management. Norde- co, Copenhagen, Denmark. Fjeldsa, J. and N. Krabbe. 1990. Birds of the High Andes. Zoological Museum, University of Copen- hagen and Apollo Books, Svendborg, Denmark. Gonzalez, G. E. and J. C. Torres-Mura. 2()()0. Ni- dificacidn de Oreomanes fraseri (Passeriformes, Fringillidae) en Tarapaca, Andes de Chile. Boletfn Chileno de Ornitologfa 7:19-23. Greeney, H. E, M. Juina, and A. E Sornoza M. 2006. Nest descriptions for Conotliraupis speculigera and Thlypopsis ornate in Ecuador. Boletfn SAO Volume 16 (Number 1). Hansell, M. 2000. Bird nests and construction behav- ior. Cambridge University Press, Cambridge, United Kingdom. Hensen, I. 2002. Impacts of anthropogenic activity on the vegetation of Polylepis woodlands in the region of Cochabamba, Bolivia. Ecotropica 8:183-204. Herzog, S. K., R. Soria A., and E. Matthysen. 2003. Seasonal variation in avian community composi- tion in a High-Andean Polylepis (Rosaceae) forest fragment. Wilson Bulletin 115:438-445. Herzog, S. K., R. Soria A., A. Troncoso J., and E. Matthysen. 2002. Composition and structure of avian mixed-species flocks in a High-Andean Po- lylepis forest in Bolivia. Ecotropica 8:133-143. ISLER, M. L. AND P. R. ISLER. 1987. The tanagers: natural history, distribution, and identification. Smithsonian Institution Press, Washington, D.C., USA. ISLER, M. L. AND P. R. ISLER. 1999. The tanagers: natural history, distribution, and identification. Smithsonian Institution Press, Washington, D.C., USA. Kessler, M. 2002. The Polylepis problem”: where do we stand? Ecotropica 8:97-1 10. Kessler, M. and P. Driesch. 1993. Causas e historia de la destruccion de bosques altoandinos en Bo- livia. Ecologia en Bolivia 21:1-18. Long, A. J. and M. F. Heath. 1994. Nesting ecology and helping behavior in the Azure-rumped Tana- ger in Mexico. Condor 96:1095-1099. Mazar Barnett, J., G. Pugnali, and M. Della Seta. 1998. Primer registro del Sai grande Oreomanes fraseri (Passeriformes: Coerebidae? Thraupidae?) en la Argentina. El Hornero 15:44-46. Ribon, R. and j. E. Simon. 1997. The nest and eggs of the Chestnut-vented Conebill Conirostrum spe- ciosum (Temmick, 1824). Ornitologia Neotropical 8:71-72. Ridgely, R. S. and G. Tudor. 1989. The birds of South America. Volume I. The oscine passerines. Oxford University Press, Oxford, United Kingdom. Robbins, M. B. and G. S. Glenn. 1988. Eirst descrip- tion of the nest and eggs of the Moss-backed Tan- ager {Buthraupis [Bangsia] edwardsi). Condor 90: 947-948. Salaberry, M., j. Aguirre, and J. Yanez. 1992. Ad- iciones a la lista de aves de Chile: descripcion de aves nuevas para el pais y otros datos ornitologi- cos. Noticiario Mensual Museo Nacional Historia Natural 321:3-10. Skutch, a. F. 1985. Clutch size, nesting success, and predation on nests of neotropical birds, reviewed. Ornithological Monographs 36:575-594. Stotz, D. E, j. W. Fitzpatrick, T. A. Parker III, and D. K. Moskovits. 1996. Neotropical birds, ecol- ogy and conservation. University of Chicago Press, Chicago, Illinois, USA. VuiLLEUMiER, E 1984. Patchy distribution and system- atics of Oreomanes fraseri (Aves, Coerebidae) of Andean Polylepis woodlands. American Museum Novitates Number 2777:1 — 17. Wood, T. M.. E Gallo, and P. K. Donahue. 1992. Observations at a Paradise Tanager nest. Wilson Bulletin 104:360-361. The Wilson Journal of Ornithology 120(3):550— 564, 2008 BIRD DENSITY AND MORTALITY AT WINDOWS STEPHEN B. HAGER,'^ HEIDI TRUDELL,^ KELLY J. MCKAY,^ STEPHANIE M. CRANDALL, AND LANCE MAYER“^ ABSTRACT. — Little is known about impacts to birds from collisions with windows at commercial buildings. We monitored bird mortality from striking windows at five commercial buildings on two college campuses in northwestern and southwestern Illinois. Bird mortality at Augustana College (northwestern), which was evaluated from 2002 to 2006, totaled 215 individuals in 48 species for an average rate of 55 birds/building/year. We calculated a mortality rate of 24 birds/building/year for 2004-2005 from 142 individuals within 37 species at Principia College (southwestern). Mortality of North American migrant (NAM) and neotropical migrant (NTM) birds was higher during migration than during summer or winter. We tested the hypothesis at Augustana that density of birds at a given location will be positively correlated with numbers of birds that die due to strikes with windows. Bird density only partially explained strikes with windows since mortality was also a function of landscaped habitat that attracted birds. Annual bird mortality at commercial buildings may be about five times higher than previous estimates. These buildings may place bird populations at high risk of strikes at windows. Received 10 May 2007. Accepted 7 September 2007. Annual bird mortality from collisions with windows in North America could be as high as 1 billion (Klem 1990). Windows may be the most significant cause of mortality second only to habitat loss (Klem 2006). In an evo- lutionary sense, both the fit and unfit are at risk wherever birds exist in close proximity to windows (Klem 1990). Experiments and systematic monitoring, mainly at residential structures (e.g., houses), suggest that mortality to birds from collisions with windows is highest in winter (Klem et al. 2004, Klem 2006) or migration and winter (Klem 1989), and disproportionately affects species that frequent bird feeders (Dunn 1993, Klem et al. 2004). It is thought that landscap- ing features in the vicinity of houses, which maintain bird feeders, provide habitat (shelter and fruiting trees and shrubs) for birds and, thus, increase their vulnerability to window collisions (Klem 1989). It is also known that mortality at skyscrapers in large cities, such ' Department of Biology, Augustana College, Rock Island, IL 61201, USA. 2 Principia College, Elsah, IL 62028, USA. ^ BioEco Research and Monitoring Center, Hamp- ton, IL 61256, USA. “^2218 South 23rd Street, Abilene, TX 79605, USA. 5 University of Illinois Extension, Rock Island County, 4550 Kennedy Drive, Suite 3, East Moline, IL 61244, USA. ^ 1006 Oakcrest Street, Apartment 102, Iowa City, lA 52246, USA. ^ Corresponding author; e-mail: stevehager@augustana.edu as Toronto, Chicago, and New York is signif- icant, but little has been published about the details of negative impacts. Internet reporting by monitoring programs has identified thou- sands of birds that die annually during spring and fall migration periods; the species most vulnerable are neotropical and North Ameri- can migrants, and species of conservation concern (New York City Audubon Society 2004, Hunsinger 2005, Fatal Light Awareness Program 2006). Less attention has been given to studying the impacts to birds at commercial buildings, which may be broadly defined as buildings (—600 m^) used in service, office, education, and healthcare (Swenson 1997). Recent de- velopments related to commercial building construction may be placing birds at a higher risk of collisions with windows than current estimates of annual bird mortality. We esti- mated that 5.58 million commercial buildings were in the United States in 2006 which is up 1.63 million since 1986 (Klem 1990, Environ- mental Protection Agency 2004). Blem and Willis (1998) suggested that in the last 10-20 years, commercial building construc- tion had increased in suburban areas and these buildings are surrounded by landscaping fea- tures that provide food and shelter for birds. We found support for this contention by examining city ordinances which mandate establishment of landscaping around commercial buildings (City of Rock Island, Illinois 2004; City of Wheaton, Illinois 2007). Furthermore, the American So- 550 Hager et al. • BIRD MORTALITY AT WINDOWS 551 ciety of Landscape Architects expected that in 2007 the demand would be higher than ever for environmentally-friendly landscaping, which in- cludes adding native plants and water resources in close proximity to commercial buildings (Owens 2006). One published study detailing the system- atic monitoring in suburban Virginia of bird mortality at commercial buildings found: (1) bird mortality was about three times higher than the estimated 1-10 dead birds/building/ year, (2) North American and neotropical mi- grants died at higher proportions compared to permanent residents, and (3) mortality was highest during spring and fall migration and lowest in winter and summer (O’Connell 2001). Similar results were reported by Blem and Willis (1998), whose data came partly from office buildings in Virginia, and by John- son and Hudson (1976) in an analysis of bird impacts related to a glassed-in walkway con- necting two commercial buildings in Wash- ington State. It is critical that details of mortality due to collisions with windows at commercial build- ings be identified given that relatively little is known about the interactions between birds and commercial buildings, and that more of these buildings are being constructed in areas that contain “bird-friendly” habitat. More- over, little is known about the intrinsic factors affecting whether or not birds fly into win- dows aside from the contention that birds do not perceive clear and tinted glass as barriers (Klem 1989, Klem et al. 2004). Extrinsic fac- tors thought to affect collision frequency in- clude behavior, window characteristics, and the environment (season, time of day, and weather) (Klem 1989, Klem et al. 2004). In addition, Klem (1989, 2006) hypothesized the best predictor of collision rate at any one site is the density of birds in the vicinity of glass. This hypothesis — hereafter referred to as the bird density hypothesis — was generally sup- ported for bird mortality observed at a house in southern Illinois (Klem 1989) and by Dunn (1993), who analyzed data from Project Feed- er Watch (Cornell Laboratory of Ornithology 2006) for the winter months at houses where participants provided estimates of bird abun- dance only for species observed at feeders. We present the results of two studies. In Study 1 we systematically monitored avian mortality from strikes at windows at five com- mercial buildings in northwestern and south- western Illinois: Augustana College, Rock Is- land, and Principia College, Elsah, respective- ly. Our objectives were to: (1) document the abundance and richness of birds killed by buildings, and (2) assess the relationship be- tween season and migratory class of birds killed by windows, and between window area and mortality within sections of a building and among buildings. We tested the bird density hypothesis in Study 2 for a commercial build- ing at Augustana College in spring (Apr-May 2006) and winter (Dec 2006-Jan 2007). Our objective for this study was to evaluate the relationship between estimates of bird density using point counts and birds killed by strikes with windows. METHODS Study 1. — We monitored bird mortalities at two geographic locations in Illinois: Augus- tana College in Rock Island and Principia Col- lege in Elsah. Monitoring at Augustana Col- lege was conducted at the Science Building (90° 33' W, 41° 30' N) from November 2002 to November 2006 (Fig. 1). The college is within the Dissected Hill Plains Physiographic Area (Ruth 2006) and at the edge of a terrace overlooking the Mississippi River (National Cooperative Soil Survey 1998). The slopes of the bluffs are composed of upland hardwood forest, whereas the flat areas are landscaped with a variety of deciduous and evergreen shrubs and trees. The Science Building is at the western edge of campus. The western and southern edges of the building are in close proximity to wooded terrace slopes, while the northern edge faces a parking lot and paved roadway. The eastern side faces the “quad”, which is a mosaic of landscaped grasses and woody vegetation including ginkgo (Ginkgo hiloha), sycamore (Platanus spp.), oaks (Quercus spp.), maples (Acer spp.), and crab apple (Mains spp.). The Science Building is 74 m long. 25 m wide, and four stories in height. The main en- try ways (2 along the eastern edge and 1 at the western edge), and the walls in which the entry ways are found, are almost completely glass. Much of the remainder of the eastern and western walls is ccnnposed of glass. We identified sections of the building that cone- 552 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 Top view of Principia College (rotated 180° relative to north) identifying the buildings monitored: 2 (leftmost upper wing) = Dining Room, 3 = School of Government, 4 = School of Nations, and 5 = Library. Telescope Top view of Science Building (rotated 90° clockwise relative to north), Augustana College, depicting sampling sections, A-P. Section P is a patio on the 4th floor and is above Section A. All other sections are four levels in height. FIG. 1 . Locations and building characteristics for studies at Augustana and Principia colleges, Illinois. sponded to variation in wall characteristics and window size and shape. Each section was given a unique label ranging from “A” through “P” (Fig. 1) and we calculated win- dow area. Little vegetation occurs at the im- mediate edge of the building and this made visual surveys for dead birds relatively un- obstructed. Some ivy ground cover (Hedera spp.) occurs at the northeast corner of the building, but the remainder is either paved walkway or decorative stone ground cover. We monitored avian mortality by completing about two surveys/week around the building. We focused on a 2-m wide transect surrounding the entire building that extended from the build- ing’s edge. An analysis of survey effort sug- gested that two surveys/week, rather than daily surveys, were sufficient to accurately estimate mortality rate. Bodies reported to us by others that were within 2 m of the building were also included in the data set. We assumed that all dead birds adjacent to the building resulted from window collisions. We salvaged all bodies (including inciden- tal findings exclusive of surveys), identified them to species whenever possible, and placed them in the Vertebrate Museum of Augustana College. Common and scientific names fol- Hager et al. • BIRD MORTALITY AT WINDOWS 553 lowed the American Ornithologists’ Union (1998). We also recorded the date and section of the building at which each body was found. Scavengers, grounds workers, custodians, and students may have impacted detection proba- bility. However, we believe that any effect was minimal since many of the faculty, staff, and students at the Science Building were in- formed of this project and reported bird bodies to us. Moreover, scavengers may have had lit- tle impact since some of the bodies found were fairly well decomposed (estimated to be 2-3 days old) and recent work found only 13% of experimental baits were taken or moved slightly from their location by scav- engers (Klem et al. 2004). We tested for differences in average weekly mortality by bird season using Kruskal-Wallis tests. Seasons were defined according to the seasonal movements of passerine birds in the region of northwest and southwest Illinois: (1) North American Migrant (NAM): Fall Migra- tion = 2nd half of September-November, Winter = December— February, Spring Migra- tion = March- 1st half of May, Summer Breeding = 2nd half of May- 1st half of Sep- tember; (2) neotropical migrant (NTM): Fall Migration 2nd half of August- 1st half of October, Winter = 2nd half of October- March, Spring Migration = April-May, Sum- mer Breeding = June- 1st half of August; and (3) permanent resident (PER): Non-Breeding Season = October-April, Breeding Season = May-September. Differentiating each migra- tory class in this manner allowed us to more precisely assess the timing of mortality rela- tive to seasonal movements of individuals. We used the log likelihood ratio and examined whether the proportion of birds separated by migratory class was different relative to known proportions (total species 99; NAM - 45.5%; NTM = 36.4%; PER = 18.1%; S. B. Hager, unpubl. data) at the campus. We conducted simple linear regression to assess the relationship between window area (i.e., building section) and mortality. Normality was examined using the Shapiro Wilk Test (Zar 1984) and all statistical tests were com- pleted using JMP 6 (SAS Institute Inc. 2006). We monitored bird collisions at Principia College at windows at four commercial build- ings (90°21'W, 38°56'N) between January 2004 and November 2005 (Fig. 1); however. in 2004 we could not monitor from 4 June to 4 September and 20 November to 29 Decem- ber, nor for July 2005. The college is within the Prairie Peninsula Physiographic Area (Ruth 2006) and is on the bluffs of the Mis- sissippi River (National Cooperative Soil Sur- vey 1999). The general character of campus vegetation is similar in landscaping design and plant species to Augustana. Standardized surveys for mortality were conducted at the Library, School of Nations, School of Government, and Dining Room (Fig. 1). The Library (Lib) is roughly square in shape and three stories in height. Relatively more vegetation surrounded this building than the others, although this did not impact visual surveys. School of Nations (SN) is two stories in height, roughly rectangular, and contains an inegularly shaped main entrance at the north- east edge of the building. This building was the most difficult at which to conduct surveys due to the presence of creeping vegetation (e.g., Hedera spp.) that surrounded most win- dows. This may have prevented us from re- trieving all bodies. School of Government (SG), three stories in height, is “L” shaped from top view. The Dining Room (DR), two stories tall, comprises the eastern section of Howard Center and contains only three sides. No measurements were taken for each build- ing. Qualitatively speaking, we ranked the fol- lowing for relative size of buildings: SG > Lib > SN > DR; and for window area: Lib > SN > SG > DR. Survey methods and documentation for sal- vaging bird bodies followed those used for Augustana, except that daily surveys were completed around each building and all bodies were deposited in the Museum of Natural His- tory, University of California, Santa Cruz. In- cidental mortality is reported as the number of bodies opportunistically encountered on cam- pus exclusive of Lib, SN, SG, and DR; these are not included in any statistical analysis. Kruskal/Wallis Ranked Sums test was used to evaluate differences in mortality for migra- tory class and bird seasons. We used the log likelihood ratio to examine whether the pro- portion of birds separated by migratory class was different relative to known proportions (total species = 147; NAM = 36.7%; NTM = 49.0%; PER = 14.3%r; M. J. Hoff, unpubl. data) at the campus. Differences in mortality 554 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 by building (i.e., estimated window area) were examined using the log likelihood ratio. Study 2. — We tested the bird density hy- pothesis at the Augustana College Science Building during spring migration (Apr-May 2006) and winter (Dec 2006-Jan 2007). We chose spring migration since species detection via singing males, including non-breeding mi- grants, could be maximized relative to fall, when male singing is rare for most species. A winter analysis was used so that comparisons could be made with spring since other studies speculated that window-caused mortality is low in winter as fewer birds may be present at a site (Johnson and Hudson 1976, Klem 1989, Blem and Willis 1998). We assessed avian richness and abundance in spring with fixed radius (50 m) point counts, which were 10 min in duration (Bibby et al. 2000). The Science Building is dispro- portionately long along a north-south axis and we established one point count station each on the west facing (Augie Point #1) and east fac- ing (Augie Point #2) sides of the building. One point count survey was completed on Saturdays of each week between 0700—0800 hrs CDT (surveys on 28 Apr and 5 May were completed on Friday). Data from both point count stations were combined for each survey date. We followed the weather protocol for the North American Breeding Bird Survey (Par- dieck 2001). Weekly point count data were compared to the number of dead birds discov- ered 3 days prior to and 3 days following the point count (surveys completed on a Friday were compared to mortality 2 days prior and 4 days after this day). We assumed for this comparison that (1) migration for a species at a specific location may be several weeks in duration (Devore et al. 2004), and (2) within individuals, average stopover times would be ~1 week in duration given data for White- crowned Sparrow (Zonotrichia leucophrys) and Wood Thrush {Hylocichla mustelina) (Chilton et al. 1995, Wang and Moore 1997), and recent mathematical models (Schaub et al. 2001, Efford 2005). We used the same point count survey meth- odology for winter as in spring to estimate species abundance and richness except that surveys were completed between 0700 and 0800 hrs CST We believe these methods were appropriate for winter since bird populations are generally stable during this season, as was found for winter home ranges in the Northern Cardinal (Cardinalis cardinalis) (Halkin and Linville 1999). Bird mortality due to collisions with win- dows reported by others was highest in the spring and fall at commercial buildings (John- son and Hudson 1976, Blem and Willis 1998, O’Connell 2001). Thus, based on the bird den- sity hypothesis, we predicted the following at the Science Building: (1) in spring, a time of high mortality, bird abundance would be rel- atively high; (2) in winter, a time of low mor- tality, bird abundance would be relatively low; and (3) the abundance and richness of species killed at windows will be proportional to the abundance and richness of birds living in the vicinity of the building. We used simple linear regression to analyze the relationship between the abundance of mortalities and living birds, as well as the richness of mortalities and living birds within season. The Kruskal-Wallis Ranked Sums test was used to evaluate the abundance of mor- talities to the abundance of living birds for each species. Differences in abundance and richness of living birds for spring and winter were assessed using ANOVA. RESULTS Study 1. — We documented 215 window- killed birds within 48 species at Augustana College (Fig. 2, Appendix 1). Average mor- tality was 54.8 birds/building/year. Species with >10 dead individuals (this level ap- peared to be a natural break in the data) in- cluded White-throated Sparrow {Zonotrichia albicollis), Ovenbird (Seiurus aurocapilla), American Robin {Turdus migratorius). Swain- son’s Thrush {Catharus ustulatus). Dark-eyed Junco {Junco hyemalis). Ruby-throated Hum- mingbird {Archilochus colubris), and North- ern Cardinal. These represented 44.2% of the total mortality at the Science Building. Nine to 12 new species were added each year after an initial species total of 23 in 2003 (including 8 weeks in 2002). The mean weekly rate of mortality for each migratory class (Fig. 3) differed among bird- defined seasons (NAM: H = 35.9, P < 0.0001; NTM: H = 72.9, P < 0.0001; PER: H = 6.33, P = 0.012). The proportions of window-killed NAMs (39.9%), NTMs (54.2%) Hager et al. • BIRD MORTALITY AT WINDOWS 555 14 12 10 S 6 *5) 4 c ? 2 :: =9 0 7 6 5 4 3 2 1 0 C/) ■O ■O 03 0) TD C 03 CD A , ill 1=1 - -T 1=: , ' --r- 1 1 1 1 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec FIG. 2. Detections of dead birds/building/year, (A) 2002-2006 at Augustana College and (B) 2004-2005 at Principia College, Illinois. and PERs (6.3%) differed from proportions known to occur on campus (G = 9.06, df = 2, P - 0.01 1). There was a significant positive relationship between window area and mor- tality for each building section {n = 16, = 0.54, P = 0.0012). We documented 142 window-killed indi- viduals of 37 species at Principia College (Fig. 2, Appendix 1). Average mortality was 24.0 birds/building/year. Species killed most often (>8 individuals; natural break in the data) included Ruby-throated Hummingbird, American Robin, White-throated Sparrow, and Ovenbird. These species represented 56.3% of the total mortality. Twelve new spe- cies were found in 2005 relative to those found in 2004. Incidental mortality totaled 58 individuals, which were dominated by the same species recorded during standardized surveys, except for Ovenbirds (Appendix 1). The mean weekly rate of mortality (Fig. 3) differed among seasons for NTMs {H = 34.3, P < 0.0001), but not for NAMs (// = 3.24, P = 0.36) or PERs {H = 0.69, P = 0.41). The proportions of window-killed NAMs (29.7%), NTMs (54.1%), and PERs (16.2%) was not different from proportions known to occur on campus (G = 0.81, df == 2, P = 0.67). Sig- nificantly more birds died from collisions with windows at Lib (71.1%) than at SG (11.3%), SN (1 1.3%), and DR (6.30%) (G = 135.5, df = 3, P < 0.0001); this corresponded to qual- itative estimates of window area by building. This may also be explained by the finding that all dead Ruby-throated Hummingbirds, which had the highest mortality at Principia, were at Lib; however, significantly more birds died at Lib than at the other buildings (G = 56.1, df = 3, P < 0.0001) even if hummingbirds are excluded from the data set. Study 2. — We documented 55 and 22 live species during spring and winter, respectively (Table 1, Appendix 2). Simple linear regres- sion revealed no relationship between dead Mean dead birds/week (±1SE) 556 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 3, September 2008 Augustana College P< 0.0001 J l_ ^ ^ ! Fall Winter Spring Summer Principia College NAM Seasons 2 1.5 1 0.5 0 Fall Winter Spring Summer P< 0.0001 T 4 3.5 3 2.5 2 1.5 1 0.5 0 Fall Winter Spring Summer NTM Seasons 0.14 r — 0.12 0.1 0.08 0.06 0.04 - 0.02 0 Breeding P = 0.012 Non-Breeding PER Seasons FIG. 3. Detections of dead birds/week for bird-defined seasons separated by campus and migratory class, Augustana and Principia colleges, Illinois. and living birds for abundance {n = 8, P = 0.014, P = 0.79) and richness {n = 8, P = 0.18, P = 0.29) in spring. We found no rela- tionship in spring between species of living birds observed around the Science Building and those that were killed, all surveys com- bined (// = 23.1, P = 0.57). In spring, only 8 of 55 (14.5%) species observed during point count surveys were recorded as window- killed. We found no window-killed birds in winter despite relatively abundant species (de- fined as >5 individuals/survey), such as American Crow {Corvus hrachyrhynchos), American Robin, European Starling {Sturnus vulgaris). Cedar Waxwing (Bombycilla ced- rorum). Dark-eyed Junco, Northern Cardinal, and House Sparrow {Passer domesticus). We found no differences in the abundances of liv- ing birds between spring and winter {F = 2.81, P = 0.12); however, richness in living Hager et al. • BIRD MORTALITY AT WINDOWS 557 TABLE 1 . Abundance of birds at point counts and birds found dead near windows at Augustana College, Illinois in Study 2. Season Observation date“ # Live birds # Birds found dead Abundance Richness Abundance Richness Spring 8 Apr 2006 113 25 2 1 15 Apr 2006 98 25 0 0 22 Apr 2006 101 20 0 0 28 Apr 2006 107 17 1 1 5 May 2006 70 22 4 3 13 May 2006 76 22 1 1 20 May 2006 127 32 3 3 27 May 2006 85 22 2 2 Mean ± SE 97.1 ± 6.80 23.13 ± 1.56 1.63 ± 0.50 1.38 ± 0.42 Winter 6 Dec 2006 128 16 0 0 12 Dec 2006 86 14 0 0 20 Dec 2006 87 12 0 0 27 Dec 2006 68 14 0 0 3 Jan 2007 68 16 0 0 10 Jan 2007 71 13 0 0 19 Jan 2007 73 14 0 0 25 Jan 2007 Mean ± SE 59 80.0 ± 7.62 10 13.6 ± 0.71 0 0 ^ Weekly point count data were compared to bird kills discovered ± 3 days relative to the date on which a point count survey was conducted. birds differed between these seasons (F = 30.7, P < 0.0001). Relatively abundant spe- cies observed during point counts in both spring and winter experienced no mortality from window collisions, including American Robin, Northern Cardinal, and House Spar- row. The monitoring data revealed no differ- ences among years for mortality in spring (H - 1.81, P = 0.61) and winter (H = 4.59, P = 0.33). Thus, the mortality we documented in spring 2006 and winter 2006-2007 was not different than the mortality observed for these same seasons in previous years. DISCUSSION The results of systematic monitoring of window strikes at commercial buildings in northwestern and southwestern Illinois (Study 1) demonstrate that bird mortality was high. We calculated a mortality rate of almost 55 dead birds/building/year at Augustana (north- western Illinois); this is about twice as high as Principia in southwestern Illinois (24 birds/ building/year) and relative to an ofhce park (29 birds/building/year) in Richmond, Virgin- ia (O’Connell 2001). Bird mortality was high- est among sections of the Science Building with the most window area and corresponded to buildings at Principia with the highest es- timated window area. Species richness of bird mortality was higher at Augustana (n = 48) than at Principia (n = 37) and in Virginia {n = 40; O’Connell 2001). Differences among sites in numbers and species of birds dying may be a consequence of factors related to behavior, environment, and window area. We found that 9 to 12 new species died each year at both Augustana and Principia. Thus, multi- year studies may observe a larger range of species dying from collisions with windows than monitoring projects of relatively short duration (O’Connell 2()01). Mortality for Study 1 at Augustana and Principia was high for particular species: Ruby-throated Hummingbird, American Rob- in, White-throated Sparrow, and Ovenbird. Swainson’s Thrush, Northern Cardinal, and Dark-eyed Junco also died at high rates at Au- gustana. Deaths of Ruby-throated Humming- birds at Principia were more than twice as high as other species; about half of these deaths occurred in September, which includes the peak of fall migration in this region (Rob- inson et al. 1996). Robinson et al. ( 1996) iden- 558 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 tilled Ruby-throated Hummingbirds as being possibly vulnerable to window strikes and Graham (1997) suggested that traplining (for- aging for nectar over great distances of un- defended plants) of hummingbirds may make them vulnerable to collisions. High abundance of this species in southwestern Illinois (K. J. McKay, pers. obs.) coupled with fall flower- ing honeysuckle (Lonicera spp.) on the Prin- cipia campus (M. J. Hoff, pers. obs.) may at- tract high numbers during fall migration. White-throated Sparrows at Augustana were killed more often than other species. Most in- dividuals (72%) died in the fall, a time during which migration for this sparrow is more pro- longed than during spring (Borror 1948). We found individuals from September through November, whereas spring mortality was re- stricted to May. Our work contributes toward a better un- derstanding of those species commonly killed at windows of different building structures. Regular fatalities at commercial buildings in- clude Ruby-throated Hummingbird, Yellow- bellied Sapsucker (Sphyrapicus varius). Brown Creeper {Certhia americana), thrush- es, waxwings, and wood-warblers (Klem 1989, Blem and Willis 1998, O’Connell 2001), whereas at houses they are grosbeaks, chickadees, and woodpeckers (Klem 1989, Dunn 1993). Some species are common at both structures: kinglets, American Robin, Northern Cardinal, White-throated Sparrow, Dark-eyed Junco, finches, and House Sparrow (Klem 1989, Dunn 1993, Blem and Willis 1998, O’Connell 2001). To our knowledge, this is the first study to examine mortality of migratory classes against proportions known from a site and across specific bird-defined seasons for each migratory class. These analyses allowed us to more precisely understand bird mortality re- lated to seasons. Generally, migrating species tended to die relatively more during spring and fall migration, although patterns of sea- sonal mortality were different between cam- puses. We believe the lack of monitoring sur- veys conducted in summer and December at Principia resulting in less than a full year of monitoring data, explains part of these differ- ences. In addition, differences may be attri- buted to variation in the composition of bird populations among seasons at each site. We tested the bird density hypothesis in Study 2 at the Science Building of Augustana. We predicted the following at this building: (1) in spring, a time of high mortality, bird abundance would be relatively high, (2) in winter, a time of low mortality, bird abun- dance would be relatively low, and (3) the abundance and richness of species killed at windows will be proportional to the abun- dance and richness of birds living in the vi- cinity of the building. Only the first of these predictions was supported by the data. Thus, our work does not support the bird density hypothesis per se. Our data suggest that in addition to bird density, window-related fac- tors and bird behavior (Klem 1989) explain the patterns of collisions at windows observed at the Science Building. The hypothesis of window-related factors indicates that habitat variables attract particular avian species to the vicinity of windows which results in these species being more vulnerable to dying at windows than birds not found near windows. Factors in this explanation include: size and location of windows in a building relative to ground level; suburban and urban habitats; and habitat surrounding buildings that con- tains bird feeding stations, fruiting trees, water supplies, and nesting and perching sites (Klem 1989). The bird behavior hypothesis suggests that collisions with windows occur due to in- tra- and interspecific interactions (e.g., male- male chasing and escape flights due to sudden presence of a potential predator) and physio- logical effects to the body during migration, e.g., migratory restlessness and aggression (Klem 1989, Berthold 2001). The results of Study 2 are consistent with these hypotheses because: (1) mortality was documented in spring and not in winter despite the observation during point counts of no sig- nificant differences in abundance between these two time periods; (2) there was no relationship for species richness between birds found living near the building and those that died; (3) for birds that did not die from collisions with win- dows, abundances were similar between spring and winter (e.g.. Mourning Dove [Zenaida ma- croura]. Downy Woodpecker [Picoides pubes- cens\ and House Finch [Carpodacus mexican- us]) or the species was more abundant in one season or another (e.g., American Crow and Ce- dar Waxwing); (4) abundance was higher in Hager et al. • BIRD MORTALITY AT WINDOWS 559 winter for some birds that died in spring (e.g., European Starling); and (5) relatively abundant birds did not die at windows during Study 2 (e.g.. Northern Cardinal and House Sparrow) (Table 1, Appendix 2). Only one season each for spring and winter was examined for this work, but the mortality data were not statisti- cally different among years for these seasons; several bird surveys completed on the Augus- tana campus in previous years in winter and spring suggest that abundance and diversity were similar among years (S. B. Hager, unpubl. data). The bird density hypothesis was previously supported for houses during winter, and land- scaping features and bird behavior (related to anti-predation) were found to contribute to collisions at windows (Dunn 1993). Our re- sults for a commercial building suggest that birds are vulnerable to collisions with win- dows for different reasons. The presence of bird feeding stations at houses may best ex- plain these differences. Dunn (1993) noted that houses maintaining feeders attracted a minimum of 84 birds during a count period, which tends to confine high densities to a rel- atively small area near a house. Mortality at houses is known to occur from the sudden presence of a potential predator that forced birds to abruptly take flight, but fail to rec- ognize windows of a house as a barrier (Dunn 1993). Klem et al. (2004) found a significant positive relationship between increasing dis- tance of feeders from houses and the rate of bird-window collisions. Thus, it seems that window strikes by birds occur when densities per unit area are high and birds are congre- gated some distance from the structure. We observed similar abundances at the Science Building as those reported by Dunn (1993) and suggest the density of birds was relatively lower when compared to the area available to birds at a house. In addition, birds known to frequent feeders in winter, such as Dark-eyed Junco, Northern Cardinal, and House Spar- row, were common at the Science Building, but experienced no mortality during Study 2. Houses with feeder stations may increase bird density beyond some threshold value that sig- nificantly increases their vulnerability to win- dows strikes. Conversely, at commercial buildings without feeding stations, bird den- sity per unit area and vulnerability to colliding with windows are relatively low. We suggest that birds in the United States and Canada are vulnerable to window collisions be- cause: (1) increasing numbers of commercial buildings each year, zoning laws for and societal interest in increased naturalized habitat around houses and businesses, and that each of these is more common in suburban contexts (Blem and Willis 1998; Owens 2006; C. G. Mahaffey, pers. comm.); (2) bird mortality due to window col- lisions at commercial buildings disproportion- ately affects neotropical and North American migrants mostly during spring and fall migra- tion; and (3) the reported annual mortality at commercial buildings is at least 3-5 times high- er than estimated by Klem (1990). We recom- mend implementing measures directed at dis- rupting factors which place birds at high risk of striking windows. The most practical in the con- text of heightened interest by the public in nat- uralized landscaping are external window screening that covers windows and angled win- dow mounting in buildings (Klem 2006). These significantly reduce mortality by migratory and non-migratory species (Klem et al. 2004, Klem 2006). Proper placement of bird feeders, if pre- sent in the vicinity of a building, appears to be effective in reducing mortality to species that frequent feeding stations (Klem et al. 2004). ACKNOWLEDGMENTS All window-killed birds at Augustana College were salvaged under Federal Migratory Bird Scientific Col- lecting Permit MB096406-0 and Illinois DNR Scien- tific Collecting Permit NH07.313. We are grateful to S. A. Hallstrom, who helped with monitoring bird mortality throughout the study at Augustana. A. B. Francois assisted in fall 2005. M. J. Holt kindly pro- vided a species list for Principia College. J. J. Williams offered insightful comments on an early dralt ot the paper. We thank Daniel Klem Jr. and one anonymous reviewer for helpful suggestions on the manuscript. LITERATURE CITED American Oknithoi.ogksts' Union. 1998. Check-list of North American birds. Seventh Edition. Amer- ican Ornithologists' Union, Washington. D.C.. USA. Bhktiioi.d, P. 2001. Bird migration: a general survey. Second Edition. Oxford University Press. New York. USA. Bimby. C. j.. N. D. Burgess. D. A. Hu e. and S. Mus- TOE. 2000. Bird census techniques. Second F2di- tion. Academic Press. London. United Kingdom. 560 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 3, September 2008 Blem, C. R. and B. a. Willis. 1998. Seasonal varia- tion of human-caused mortality of birds in the Richmond area. The Raven 69:3-8. Borror, D. J. 1948. Analysis of repeat records of banded White-throated Sparrows. Ecological Monographs 18:411-430. Chilton, G., M. C. Baker, C. D. Barrentine, and M. A. Cunningham. 1995. White-crowned Sparrow {Zonothchia leucophrys). The birds of North America. Number 183. City of Rock Island, Illinois. 2004. Rock Island Il- linois Zoning Ordinance, www2.rigov.org/pdf/ ordinancesZzoningordinance.pdf (accessed 16 Jan- uary 2007). City of Wheaton, Illinois. 2007. Wheaton Zoning Or- dinance. www.wheaton.il.us/Service/Planning-D/ index.htm (accessed 27 January 2007). Cornell Laboratory of Ornithology. 2006. Project feeder watch. Cornell Laboratory of Ornithology, Ithaca, New York, USA. www.birds.cornell.edu/ pfw (accessed 2 January 2007). Devore, S., S. D. Bailey, and G. Kennedy. 2004. Birds of Illinois. Lone Pine Publishing, Edmon- ton, Alberta, Canada. Dunn, E. H. 1993. Bird mortality from striking resi- dential windows in winter. Journal of Eield Orni- thology 64:302-309. Efford, M. G. 2005. Migrating birds stop over longer than usually thought: comment. Ecology 86:3415- 3418. Environmental Protection Agency. 2004. Buildings and the environment: a statistical summary com- piled by: U.S. Environmental Protection Agency Green Building Workgroup. Washington, D.C., USA. www.epa.gov/greenbuilding/pubs/gbstats. pdf (accessed 20 January 2007). Eatal Light Awareness Program. 2006. Collision data, www.flap.org/new/nestegg_3.htm. Eatal Light Awareness Program, Toronto, Ontario, Can- ada (accessed 3 January 2007). Graham, D. L. 1997. Spider webs and windows as potentially important sources of hummingbird mortality. Journal of Eield Ornithology 68:98- 101. Halkin, S. L. and S. U. Linville. 1999. Northern Car- dinal {Cardmalis cardinalis). The birds of North America. Number 440. Hunsinger, R. L. 2005. Chicago bird collision moni- tors: prospects for a Great Lakes network. Passen- ger Pigeon 67 : 1 1 3— 1 1 6. Johnson, R. E. and G. E. Hudson. 1976. Bird mor- tality at a glassed-in walkway in Washington State. Western Birds 7:99-107. Klem Jr., D. 1989. Bird-window collisions. Wilson Bulletin 101:606-620. Klem Jr., D. 1990. Collisions between birds and win- dows: mortality and prevention. Journal of Field Ornithology 61:120-128. Klem Jr., D. 2006. Glass: a deadly conservation issue for birds. Bird Observer 34:73-81. Klem Jr., D., D. C. Keck, K. L. Marty, A. J. Miller Ball, E. E. Niciu, and C. T. Platt. 2004. Effects of window angling, feeder placement, and scav- engers on avian mortality at plate glass. Wilson Bulletin 116:69-73. National Cooperative Soil Survey. 1998. Soil survey of Rock Island County, Illinois. USDA, Natural Re- sources Conservation Service, Washington, D.C., USA. soildatamart.nrcs.usda.gov/Manuscripts/ IL161/0/RockIslancLIL.pdf (accessed 12 January 2007). National Cooperative Soil Survey. 1999. Soil survey of Madison County, Illinois. USDA, Natural Re- sources Conservation Service, Washington, D.C., USA. soildatamart.nrcs.usda.gov/Manuscripts/ ILl 19/0/madison_IL.pdf (accessed 14 January 2007). New York City Audubon Society. 2004. List of birds found dead and injured at selected build- ings in New York City, 1990 to spring 2001. New York City Audubon Society, New York, USA. www.birdbash.org/NYCASBirdWatch/ BirdKillsCumulativebyStatusonly.htm (ac- cessed 29 March 2004). O’Connell, T. J. 2001. Avian window strike mortality at a suburban office park. The Raven 72:141-149. Owens, A. 2006. Landscape architects identify trends for 2007. American Society of Landscape Archi- tects, Washington, D.C., USA. www.asla.org/ press/2006/release 122 106.htm (accessed 27 Janu- ary 2007). Pardieck, K. 2001. North American Breeding Bird Survey. Training: main menu. USGS, Patuxent Wildlife Research Center, Laurel, Maryland, USA. www.pwrc.usgs.gov/bbs/participate/training (ac- cessed 15 January 2007). Robinson, T. R., R. R. Sargent, and M. B. Sargent. 1996. Ruby-throated Hummingbird {Archilochus colubris). The birds of North America. Number 204. Ruth, J. M. 2006. Partners in Flight, Physiographic Area Map. USGS, Patuxent Wildlife Research Center, Laurel, Maryland, USA. www. partnersinflight.org (accessed 14 January 2007). SAS Institute. 2006. JMP 6. SAS Institute Inc., Cary, North Carolina, USA. ScHAUB, M., R. Pradel, L. Jenni, and J.-D. Lebreton. 2001. Migrating birds stop over longer than usu- ally thought: an improved capture-recapture anal- ysis. Ecology 82:852-859. Swenson, A. 1997. Profile of commercial buildings in 1995. Energy Information Administration, Wash- ington, D.C., USA. www.eia.doe.gov/emeu/ cbecs/char95/profile.html (accessed 20 January 2007). Wang, Y. and F. R. Moore. 1997. Spring stopover of intercontinental migratory thrushes along the northern coast of the Gulf of Mexico. Auk 1 14: 263-278. Zar, j. H. 1984. Biostatistical analysis. Second Edi- tion. Prentice Hall, Upper Saddle River, New Jer- sey, USA. Hager et al. • BIRD MORTALITY AT WINDOWS 561 CL DO C/3 3 & DO £ c o 2 X) -o (U TD cd 1 2 u. O <£ c/, c -o o dJ x c 1° 3 D O "S 0^ (L) X > ■ ~ c .£ ^ 60 O C3 •2 s >1 ^ §. |l DC 2 2 -2 L: '3 2 5U L Q "2 2 C3 ^ ^ § s ^ s .§ •§ K 2 .s 5 2 -I s 3 I . g II ^ 5 2 g £ g g ■2 fl -5; 2 ■2 L: 2 L 2 L ^ ^ 2 ^ - -2 2 -2 o C3 . . . ^ s ^ L 2 h:Uc 2 sj ^ -c 2 2 ^1 C3 > ■5^ L . Q Q L/J 7i X ^ I ^ 3 Q ii •3 X C -V III I CO 2 >- 3 ^ E E 3 X II a s 1 f I ^ Ll 3 2 c 3 Y ^ c S 3 0> O w U QC >■ Q Z 3 CL ■§ E ^11 uu w > u - § o X X I- s ^ .;; i 2 ■I 11 ? ■I " ■E -c c -o ^ a: .3 2 r3 C3 H •C F 5j >% E 2 E < O CD P -3 3 dJ uj U _u DO X ^ 3 X $ i ^ ■§ I I g » I » V := 1 X 2 !S o z X) Urn J V -a II ^ I O y. X V D X >■ u 562 THE WILSON JOURNAL OF ORNITHOLOGY • Vol 120, No. 3, September 2008 o < — r^i— — — '^'OC^O'n'^0(N — (NCSONr^OCiTi 04 ^ - 04 - H r3 ?3 ■ S-c C "O fv. — — — 04 OC^IO, — 0404r-00 ^ ^ l/~) ^ 3 m high, not branched at the base (arboreal stratum). Flock Composition. — We measured coexis- tence between Worthen’s Sparrows and other species in the flocks. Coexistence of species was calculated using Ochiai’s Similarity Co- efficient = A/[VA + B'S/A + C], where A is the number of places where species one and species two were found, B is the number of places where species one has been found but not species two, and C is the number of places where species two has been found but not species one. This coefficient provides in- formation about co-occurrence of two species in the same area or locality in the range 0 to 1.0 (Janson and Vegelius 1981, Jackson et al. 1989). Assuming that is approximately nor- mally distributed with the mean, asymptotic standard error was calculated as: = (a^c + a^b + 4abc)/[4{a + bf-{a + cY-n] d(r^), where a, b, and c are proportions ob- tained from division of A, B, and C by n that, in this case, is the total number of flocks. We quantified foraging and vigilance time of Worthen’s Sparrows in both heterospecific and monospecific flocks. We randomly select- ed a focal individual (following Austin and Linwood 1972) and observed it for 3 min, re- cording foraging and vigilance activity. For- aging (food-search) and food-capture behav- iors were treated as feeding activity while vig- ilance was measured as time spent looking around from the ground or perches. We used r-tests to compare mean group size, and mean foraging and vigilance times in both types of flocks; mean values are reported with ± stan- dard error. We used a one-way ANOVA to compare size of monospecific and heterospe- cific flocks between breeding and non-breed- ing seasons using Statistica Kernel release 5.1 (StatSoft Inc. 1998). RESULTS We observed 245 bird flocks along 49 tran- sects including 64 mixed-species flocks and 182 monospecific flocks. Worthen’s Sparrows occurred in 27 (43%) mixed flocks and 19 (10%) monospecific flocks. The average num- ber of species in mixed flocks containing Wor- then’s Sparrows was 2.7 ± 1.19 per flock and mean flock size was 12.9 ± 8.97 birds (range 3-31). Mean group size in monospecific flocks of Worthen’s Sparrows was 10.7 ± 8.81 individuals per flock (range 2-30). We found no difference between mean sizes of mono- specific and heterospecific flocks = —0.81, P = 0.42). Worthen’s Sparrows associated with 16 oth- er species in heterospecific flocks (Table 1). Based on Ochiai’s Similarity Coefficient, the highest values of coexistence between Wor- then’s Sparrows and species in the mixed flocks were Vesper Sparrows (Pooecetes gra- mineus) (0.60), Black-throated Sparrows {Am- phispiza bilineata) (0.33), Homed Lark {Ere- mophila alpestris) (0.27), and Western Blue- bird (Sialia mexicana) (0.22). The breeding season in our study area oc- curred from mid-April to late August. Mono- specific and mixed-species flocks were ob- served in both breeding and non-breeding sea- sons. We recorded five monospecific and sev- en heterospecific flocks during the breeding period, and 14 monospecific and 20 hetero- specific flocks during the non-breeding period. A one-way ANOVA indicated that mean flock size in the two types of flocks did not differ (^3,42 = 2.331, P = 0.0879) between breeding and non-breeding seasons (Fig. 1). Flocking incidence was higher during the non-breeding season than during either the breeding season Canales-DelgadiUo et al. • FLOCKING IN WORTHEN’S SPARROW 571 TABLE 1. Coexistence coefficient species flocks. ± asymptotic SE for Worthen’s Sparrow and associated species in mixed- Species Scientific name Ochiai’s Similarity Coefficient Black-throated Sparrow Amphispiza bilineata 0.33 -+- 0.23 American Pipit Anthus rubescens 0.19 -H 0.24 House Finch Carpodacus mexicanus 0.21 -F 0.22 Lark Sparrow Chondestes grammacus 0.20 -h 0.25 Yellow-rumped Warbler Dendroica coronata 0.17 -h 0.25 Horned Lark Eremophila alpestris 0.28 -h 0.25 Northern Mockingbird Mimus polyglottos 0.10 -F 0.22 Canyon Towhee Pipilo fuscus 0.20 ± 0.25 Vesper Sparrow Pooecetes gramineus 0.60 -F 0.22 Rock Wren Salpinctes obsoletus 0.27 -F 0.24 Say’s Phoebe Sayornis saya 0.09 4- 0.20 Western Bluebird Sialia mexicana 0.23 -F 0.22 Clay-colored Sparrow Spizella pallida 0.19 -F 0.24 Chipping Sparrow S. passerina 0.15 -F 0.23 Field Sparrow S. pusilla 0.29 -F 0.25 Curve-billed Thrasher Toxostoma curvirostre 0.09 0.20 or the entire season for both monospecific and mixed-species flocks (Fig. 2). Flocks of both types were virtually absent from June to Au- gust when reproductive activity of Worthen’s Sparrows was high (Fig. 3) and single males or pairs were prevalent. During September we found only three pairs of Worthen Sparrows. Foraging occurred most frequently on the ground (76%) and the herbaceous stratum re- ceived less attention by birds (4%). Worthen’s Mono Hetero SEASON: Breeding Mono Hetero SEASON: Non-breeding d] □ ±SD *SE Mean Flock type FIG. 1. Flock size of Worthen’s Sparrow (monospecific flocks = mono) and mixed-species (heterospecific flocks = hetero) in the breeding and non-breeding seasons in Coahiiila. Mexico. 572 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 Flock type □ Breeding ■ Non-breeding FIG. 2. Flocking incidence of Worthen’s Sparrow during the breeding and non-breeding season by flock type. Numbers above the bars are the frequency values of Worthen’s Sparrows in each type of flock. Sparrows also searched shrub stems, branches, and foliage from perches, and at times hawked for flying insects from these perches. Most perches were in the shrub stratum (87%), but fences also were used (12%). We observed foraging behavior in 41 and 39 focal individuals of Worthen’s Sparrow in monospecific and heterospecific flocks, re- spectively. Sparrows in monospecific and mixed species flocks differed strongly in for- aging time, spending an average of 27.0 ± 27.73 sec (n = 33) and 69.2 ± 46.11 sec (n = 38), respectively (Gg — —4.98, P < 0.000). However, we found no difference (Gg — P = 0.06) in vigilance times with an average vigilance time of 32.5 ± 39.60 sec by indi- viduals in both monospecific and mixed flocks. DISCUSSION Our observations confirm monospecific flocking by Worthen’s Sparrows (Wege et al. 18 16 14 12 - |10 - o 8 - o E 6 ^ 4 2 0 - Jan Feb Mar Apr May Jun Month Jul Aug Sep Oct Nov m MBrS m HBrS m MNBrS B HNBrS FIG. 3. Flocking incidence of Worthen’s Sparrow during the breeding (MBrS = monospecific flocks; HBrS = heterospecific flocks) and non-breeding seasons (MNBrS = monospecific flocks; HNBrS = heterospecific flocks). Ccmales-Delgadillo et al. • FLOCKING IN WORTHEN’S SPARROW 573 1993, Behrstock et al. 1997), but also reveal the species commonly occurs in mixed species flocks in the non-breeding season. We found Worthen’s Sparrows occurred more often in heterospecific than in monospecific flocks. Flocking incidence of Worthen’s Sparrows decreased in late winter and early spring, as reported by Powell (1979), who found the largest flock sizes among neotropical birds oc- curred from November to February. The de- crease in flocking incidence during the breed- ing season and flocking associations with 16 species, some of which are migratory species, suggest that flocking in Worthen’s Sparrows was seasonal and that flock composition be- comes more heterogeneous with inclusion of migrants. Our data showed a flocking inci- dence of 35% in mixed-species flocks during the breeding season, but most of these obser- vations were recorded early in the breeding season when flocks were starting to disperse. Most species that coexist with Worthen’s Sparrows in heterospecific flocks had similar food habits, feeding mainly on insects and seeds (Austin and Linwood 1972). Moreover, some species used the same foraging tech- nique, searching systematically for food across the study area. Powell (1979) and Hut- to (1994) reported that some species in het- erospecific bird flocks obtain food from the activities of others. Waite and Grubb (1988) reported similar behavior in mixed-species bird flocks of temperate-deciduous wood- lands. Worthen’s Sparrows closely associated with Vesper Sparrows, which is a migratory species that winters in our study area. Gram (1998) suggested the association between resident and migrant neotropical species in northeast- ern Mexico has an important winter survival function. Thus, migrant species in our study area could be joining opportunistically with resident species to derive survival benefits in winter. This may be one reason for the strong association with Vesper Sparrows. We frequently witnessed cases of alarm sig- naling by flock associates (e.g.. Western Blue- birds, Vesper Sparrows, Yellow-rumped War- blers [Denclroica coronata]) feeding with Worlhen’s Sparrows in mixed flocks. Others have emphasized the importance for ground- foragers of collective vigilance provided by many individuals and several species foraging together in mixed groups (e.g., Powell 1979, Goldman 1980, Hutto 1994, Latta and Wun- derele 1996, Benkman 1997, Gram 1998, Dol- by and Grubb 1999). ACKNOWLEDGMENTS The research was funded by CONACYT (COl- 0700), the University of Nuevo Leon (CN 926-04), and The Nature Conservancy (MBP/PW 010703). We thank Nick Reid, Celina Garza, and Regina Perez for helping with revision and comments on this paper, and Claudia Doria Trevino, Leonel Resendiz, and Rogelio Hernandez for helping process data and with field work. LITERATURE CITED Austin, G. T. and E. Linwood. 1972. Winter foraging ecology of mixed insectivorous bird flocks in oak woodland in southern Arizona. Condor 74:17-24. Behrstock, R. A., C. W. Sexton, G. W. Lasley, T. L. Eubanks, and J. P. Gee. 1997. First nesting re- cords of Worthen’s Sparrow (Spizella wortheni) for Nuevo Leon, Mexico, with habitat character- ization of the nest site and notes on ecology, voice, additional sightings and leg coloration. Co- tinga 8:27-33. Benkman, C. W. 1997. Feeding behavior, flock size- dynamics, and variation in sexual selection in crossbills. Auk 114:163-178. BirdLiee International. 2000. Threatened birds of the world. Lynx Editions and BirdLife Interna- tional, Barcelona, Spain and Cambridge, United Kingdom. Dolby, A. S. and T. C. Grubb Jr. 1999. Functional roles in mixed-species foraging flocks: a field ma- nipulation. Auk 116:557-559. Goldman, P. 1980. Flocking as possible predator de- fense in Dark-eyed Juncos. Wifson Bulletin 92: 88-95. Gram, W. K. 1998. Winter participation by neotropical migrant and resident birds in mixed-species flocks in northeastern Mexico. Condor 100:44-53. Hutto, R. L. 1994. The composition and social orga- nization of mixed-species flocks in a tropical de- ciduous forest in western Mexico. Condor 96: 105-118. Jackson, D. A., K. M. Somers, and H. H. Harvey. 1989. Similarity coefficients of co-occurrence and association or simply measures of occurrence? American Naturalist 133:436-453. Jan.son, S. and J. Vegelius. 1981. Measures of eco- logical association. Oecologia 49:371-376. Laita, S. C. and j. M. Wunderele. 1996. The com- position and foraging ecology of mixed-species flocks in pine forests of Hispaniola. Condor 98: 595-607. Moreno, P. N. 1984. Glosario botanico ilustrado. In- stituto Nacional de Investigaciones sobre Recur- sos Bidticos. Xalapa, Veracruz, Mexico. 574 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 3, September 2008 Orta, D. M. 1988. Influencia del Perrito de las prad- eras {Cynomys mexicamis) en la vegetacion del suelo y pastizal mediano abierto en Coahuila. Thesis. Universidad Autonoma Agraria Antonio Narro, Saltillo, Coahuila, Mexico. Powell, G. V. N. 1979. Structure and dynamics of interspecific flocks in a neotropical mid-elevation forest. Auk 96:375-390. Ralph, C. J., G. R. Geupel, M. Pyle, T. E. Martin, D. L. Desante, and B. Mila. 1996. Manual de Metodos de Campo para el Monitoreo de Aves Terrestres. USDA, Lorest Service, General Tech- nical Report PSW-GTR- 159-Web. Rioja, P. T. M. 2003. Comportamiento reproductivo del perrito llanero (Cynomys mexicanus) en el Alti- plano Mexicano. Thesis. Lacultad de Ciencias Lo- restales, Universidad Autonoma de Nuevo Leon, Mexico. Sada, a. M. 1987. Locations for finding Worthen’s Sparrow (Spizella wortheni) in Nuevo Leon [Coa- huila]. MBA Bulletin Board l(87-3):2. Scott-Morales, L., E. Estrada, L Chavez-Ramirez, AND M. CoTERA. 2004. Continued decline in geo- graphic distribution of the Mexican prairie dog (Cynomys mexicanus). Journal of Mammalogy 85: 1095-1101. StatSoft Inc. 1998. STATISTIC A for Windows. Tul- sa, Oklahoma, USA. Villasenor, G. J. L and E. Santana C. 2003. El mon- itoreo de poblaciones: herramienta necesaria para la conservacion de aves en Mexico. Pages 224- 262 in Conservacion de Aves, Experiencias en Mexico (H. Gomez de Silva and I. A. Oliveras, Editors). CONABIO, D.E, Mexico. Waite, T. A. and T. C. Grubb Jr. 1988. Copying of foraging locations in mixed-species flocks in tem- perate-deciduous woodland birds: an experimental study. Condor 90:132-140. Wege, D. C., S. N. G. Howell, and A. M. Sada. 1993. The distribution and status of Worthen’s Span'ow Spizella wortheni: a review. Bird Conservation In- ternational 3:211-220. The Wilson Journal of Ornithology 120(3):575-581, 2008 STATUS OF CRESTED PENGUIN {EUDYPTES SPP.) POPULATIONS ON THREE ISLANDS IN SOUTHERN CHILE DAVID A. OEHLER,' 8 STEVE PELIKAN,^ W. ROGER FRY,^ LEONARD WEAKLEY Jr.,-* ALEJANDRO KUSCH,^ AND MANUEL MARIN*-’ ABSTRACT. — We surveyed the coast of Chile between Isla Terhalten (54° 20' S) and Islote Leonard (55° 44' S) by boat during November-December 2005 to document breeding locations of Rockhopper (Eudyptes chrysocome chrysocome) and Macaroni {E. chrysolophus) penguins, and to count the number of nests during the early incubation period. A yacht-based observation recorded 1,000 Rockhopper Penguin nests at Isla Ter- halten. Ground-based counts at Isla Noir estimated a total of 158,200 Rockhopper and 3,470 Macaroni penguin nests in 17 mixed colonies and one Macaroni Penguin breeding colony. The Rockhopper population on Isla Noir increased significantly from the previous estimate of 70,000 pairs, making this island the most important breeding site for the southern subspecies of Rockhopper Penguin. The new estimate for Rockhopper Penguins suggests a population increase over the last 20 years, consistent with recent data for the main colony in Argentina. Conservation status of the species and the factors potentially responsible for population declines elsewhere may need to be re-evaluated. Received 11 July 2007. Accepted 9 December 2007. The Southern Rockhopper Penguin {Eudyp- tes chrysocome chrysocome) and Macaroni Penguin {E. chrysolophus) are the most nu- merous of the Eudyptes penguins (Woehler 1993, Reilly 1994). Both species have been classified as Vulnerable, based on declining populations of at least 30% over the last 30 years and continued anthropogenic pressures (e.g., fisheries activities and changes in the marine environment), by the International Union for the Conservation of Nature (lUCN)/ BirdLife International Red List (BirdLife In- ternational 2007a). The world population of the Southern Rockhopper Penguin subspecies has been es- timated at 475,000 breeding pairs at 51 sites in Argentina, Chile, and the Falkland Islands (Bingham and Majias 1999). Macaroni Pen- guins occur in at least 216 colonies at 50 breeding sites on many islands throughout subantarctic and Antarctic waters north of the ' Feather Link Incorporated, Cincinnati, OH 45244, USA. 2 University of Cincinnati, Math Department, MLO 025, Cincinnati, OH 45221, USA. ' 7920 Brill Road, Cincinnati, OH 45243, USA. ‘*6613 Wyndwatch Drive, Cincinnati, OH 45230, USA. ^Casilla 19, Punta Arenas, Chile. ^Section of Ornithology, Natural History Museum of Los Angeles County, Los Angeles, 900 Exposition Boulevard, Los Angeles, CA 9()()07, USA. ’Current address: Casilla 15 Melipilla, Chile. ^ Corresponding author; e-mail: david.oehler@fu.se.net pack ice. The world population estimate for the Macaroni Penguin is 9.0 to 11.7 million breeding pairs (Reilly 1994, Bingham and Majias 1999, BirdLife International 2007b). Long-term monitoring of these breeding sites shows a substantial decline in these popula- tions over the last 50 years (Woehler and Croxall 1997; Barlow et al. 2002; Clausen and Huin 2003; Crawford et al. 2003, 2006; Piitz et al. 2003). Recent surveys in Argentina, in- volving the inventory of 180,000 pairs of Rockhopper Penguins on Staten Island, sug- gest the Bahia Franklin colony has dramati- cally increased from a few thousand pairs to 167.000 pairs and may indicate a geographical shift of birds from the Falkland Islands (Schiavini 2000). The number of Southern Rockhoppers in Chile was estimated at 175.000 pairs with the largest colonies of 70.000 and 13,000 pairs on Isla Noir and Di- ego Ramirez islands, respectively, but with no comprehensive monitoring to identify long- term trends (Venegas 1984, 1991; Woehler 1993). Recent studies indicate the Diego Ra- mirez population comprises 132,721 Rock- hopper Penguin pairs (Kirkwood el al. 2007). The colony on Isla Noir of 70,000 pairs con- tains an estimated 35% of the total number along the coast of Chile (Venegas 1998). The available data on total numbers and population trends of Southern Rockhopper Penguins in the Falkland Islands and along the coast of Argentina are more complete than data for Chilean populations. The Falkland Is- 575 576 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 land Seabird Monitoring Programme, initiated in 1986/1987, has provided annual documen- tation of numbers and breeding success of penguins in the Falkland Islands (Thompson 1993, Ptitz et al. 2003). These estimates have been based on quadrat sampling to equate oc- cupied nest densities with measurements of the total area of the colonies (Schiavini 2000). Inventories of Southern Rockhopper colo- nies in Chile have been sporadic or yacht- based (Woehler 1993). More in-depth quadrat sampling may be hampered by limited access to outlying islands. Nesting habitats used by penguin colonies, often on steep or heavily vegetated areas, makes direct observations of nests difficult without disruption of the colo- nies (Downes et al. 1959, Piitz et al. 2003). Woehler’s (1993) reports of 100,000 pairs on the Ildefonso Group, 10,800 pairs on Bame- velt Island, and 3,000 pairs on Terhalten Is- land are estimates of adult birds accurate only to the nearest order of magnitude and were derived from Gerry Clark’s expedition (Clark 1988). The lUCN established a high priority recommendation for the Southern Rockhopper Penguin, which included identification of the status and trends in Argentina, Chile, Falkland Islands/Malvinas (Ellis et al. 1998). Macaroni Penguin populations along South America may be stable although sufficient data are lacking to identify trends (BirdLife International 2007b). Small colonies of Mac- aroni Penguins have been reported on Cabo Pilar, Isla Desolacion, Isla Diego Ramirez, Isla Deceit, Isla Terhalten, and Isla Recalada, Chile (Araya and Millie 1986, Wallace 1991). The population on Isla Recalada declined from 559 in 1989 to 421 in 1990 with no Mac- aroni Penguins observed in 1991 and 2005 (Soto 1990, Venegas 1991, Venegas and Soto 1992, Oehler et al. 2007). The total number of pairs on Isla Noir has been estimated at 12,500 to 25,000 (Venegas 1984, Woehler 1993). The objectives of our survey were to (1) estimate the numbers of breeding Rockhopper and Macaroni penguins within the study areas (Fig. 1), and (2) establish a more accurate baseline population estimate to track future demographic trends, using direct counts in open areas and stratified random sampling in areas of heavy vegetation. METHODS We surveyed known locations of Rockhop- per and Macaroni penguin colonies at Isla Noir (54° 20' S, 73° 01 ' W) from 8 through 25 November 2005 using maps and Global Po- sitioning System (GPS) coordinates from pre- vious studies. We conducted coastal searches, based on historic data for reference, by boat to expand the survey to adjacent islands. The goal was to collect data from each colony site to estimate total penguin populations. Study sites on Isla Noir and Islote Leonard (53° 23' S, 74° 04' W) demonstrated the timing of our survey coincided with peak colony attendance for the two species. Observations on penguins at Isla Terhalten (55° 26' S, 67° 04' W) on 22 December 2005 were limited to yacht-based counts. Investigators on Isla Noir and Islote Leonard conducted land-based systematic sampling of all geographic areas known to be used by penguins as described by Venegas (1984, 1991, 1998), Soto (1990), and Venegas and Soto (1992). Direct ground counts were used only in ar- eas where all birds within a colony could be observed completely, on Islote Leonard on 14 November 2005 and in the main Macaroni Penguin colony at Isla Noir on 9 November 2005. Three observers conducted simulta- neous but independent counts of each area. The counts were repeated if any of the three counts for any given colony varied by more than 10% from the others. Each count was recorded separately. We used quadrat sampling to estimate abun- dance (Krebs 1999) in areas of irregular to- pography and dense vegetation that precluded methodology developed for estimating breed- ing population size through use of ground counts (CCAMLR 1999). We conducted quad- rat sampling on Isla Noir from 8 to 12 No- vember 2005 in seven colonies selected at ran- dom from the 17 Rockhopper colonies. We delineated the outer perimeter of quadrats with yellow twine to facilitate plotting of nests. We recorded the total number of occu- pied penguin nests and nests with eggs in 45 randomly located 10 X 20-m quadrats within the Rockhopper colonies and 10 randomly lo- cated quadrats outside the colonies. We de- scribed habitat features within the quadrats, including vegetation type, soil, topography, el- Oehler et al. • CRESTED PENGUINS IN SOUTHERN CHILE 577 20° S 25° S Pacific Ocean 30° s 35° S 40° S Islote Leonard O 0 2.5 KM 0 500 Miles 1 1 I 0 500 KM FIG. I. Islote Leonard, Isla Noir, and Isla Terhalten along the coast of Chile. 80° W Isla Terhalten 45° S Atlantic Ocean 55° S 50° s Falkland Islands 578 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 evation, and slope. We manually recorded GPS waypoints at ~20-m intervals along the outer perimeter of all colonies to calculate the area of the colonies. The area of the 17 accessible Southern Rockhopper Penguin breeding sites on Isla Noir was calculated from GPS coordinates and elevations using the GIS program AutoCAD. The variability in the estimate of the total density of nests can be attributed al- most entirely to the variation due to sampling based on a conservative estimate of the co- efficient of variation for the measurement of area (Bogaert et al. 2005). We choose to ig- nore the variability in the GPS measurement of the colony areas. The density of nests was calculated following Krebs (1999, 2002; C. J. Krebs, pers. comm.). We used confidence in- tervals for quantifying uncertainty of our den- sity estimates (Krebs 1999). RESULTS Isla Terhalten — Rockhopper Penguin. — A yacht-based census estimated —1,000 Rock- hopper Penguin nests within the colony on Isla Terhalten. Islote Leonard — Macaroni Penguin. — The survey documented a previously unrecorded colony of Macaroni Penguins on Islote Leo- nard (6.4 nautical km southeast of Isla Reca- lada). Direct counts recorded 132 (± 4 SD) occupied nests and 96% of Macaroni Pen- guins on nests were incubating with the re- maining pairs attending empty nests. Isla Noir — Rockhopper Penguin. — The total area of all colonies was 31.61 ha producing a population estimate of 158,200 (95% Cl = 139,716-176,700) Rockhopper Penguin pairs for all colonies combined (Table 1). Pairs at occupied nests in the random quadrats within the colonies were incubating eggs (98%) with the remainder attending empty nests. The number of nests per quadrat did not differ between colonies (Fig. 2; one-way AN- OVA: F = 0.99, df = 6, P = 0.45) and we pooled data from all censused quadrats to es- timate the number of penguins in the colonies. The variance-mean ratio test (x^ = 668.8, df = 44; Krebs 1999:119) indicated that nests were more clustered (P < 0.0001) than they would be if spatially random. The counts pro- duced parameter estimates p. = 100.11 and k = 4.257 by fitting a negative binomial distri- TABLE 1. Rockhopper Penguin colonies on Isla Noir, November 2005. Areas of individual colonies and estimated total nests based on 95% confidence in- tervals for density of nests. Rockhopper colony Area (ha) Estimated nests (95% Cl) 1 0.39 1,953 (1,724-2,180) 2 1.35 6,757 (5,967-7,547) 3 1.34 6,706 (5,923-7,491) 4 3.45 17,267 (15,249-19,286) 5 3.23 16,166 (14,277-18,056) 6 0.25 1,251 (1,105-1,398) 7 1.41 7,057 (6,232-7,882) 8 0.35 1,752 (1,547-1,957) 9 5.63 28,177 (24,885-31,472) 10 1.06 5,305 (4,685-5,925) 11 2.80 14,013 (12,376-15,651) 12 4.29 21,471 (18,962-23,981) 13 0.58 2,903 (2,564-3,242) 14 0.75 3,754 (3,315-4,193) 15 1.15 5,756 (5,083-6,429) 16 1.36 6,807 (6,011-7,602) 17 2.22 11,111 (9,812-12,410) Totals 31.61 158,200 ( 139,716-176,700) bution (following Krebs 1999: equation 4.14). A log(x + k!2) transformation generated a 95% confidence interval of 88.4-1 1 1.8 for the number of nests per 200-m^ quadrat. A Chi- square goodness-of-fit test (x^ = 18.334, df = 12, P = 0.11) provided additional support for the negative binomial distribution. Rockhopper Penguin nests were almost ex- clusively in tall grasses and most nests were associated with Anthoxanthum redolens. Nests were rarely found in wet areas containing oth- er dominant vegetation such as locenes acan- thifolius or in areas containing trees (e.g., Nothofagus antarctica). Isla Noir — Macaroni Penguin. — We used the quadrat data from the Rockhopper Pen- guin census to estimate a total of 1,649 (95% Cl = 386-2,913) Macaroni Penguin nests within the Rockhopper colonies. Macaroni Penguin nests were not spatially distributed randomly based on a variance-to-mean ratio test (x^ = 312.33, df - 45, P < 0.00001). A negative binomial distribution best described the census data (x^ 3.49, df = 3, P = 0.32) with [X = 1.0435 and k = 0.193. These pop- ulation parameter estimates produce a 95% confidence interval for the number of nests/ 200-m^ quadrat of 0.244-1.843. Direct counts in the main Macaroni Pen- Oehler et al. • CRESTED PENGUINS IN SOUTHERN CHILE 579 Quadrat counts vs. Colony Colony FIG. 2. Rockhopper Penguin nests per quadrat in seven randomly selected colonies on Isla Noir. Median counts per quadrat are indicated by thick solid lines, boxes indicate interquartile range, and thin solid lines represent minimum and maximum values. guin colony on Isla Noir, recorded 1,816 (± 55 SD) occupied nests. We documented 3,470 (2,157-4,784) Macaroni Penguin nests on Isla Noir including those within the Rockdiopper colonies. All Macaroni Penguins on nests within the mixed colonies were incubating eggs. Penguins (96%) within the main Maca- roni colony were incubating eggs on nests and the remaining pairs were attending empty nests. DISCUSSION We describe a new colony of Macaroni Penguins on Islote Leonard with a relatively small population of 132 breeding pairs. The absence of historical data precludes clear in- terpretation of the findings, although with dis- appearance of the former colony on Isla Re- calada (Oehler et al. 2007), these baseline data may suggest a possible shift from Isla Reca- lada to Islote Leonard. The yacht-based survey of Isla Terhalten documented fewer Rockhopper Penguins than in past surveys. However, it is premature to draw conclusions from the current yacht- based observations in reference to historical accounts. Further land-based counts are re- quired to more accurately describe population trends of RockLopper and Macaroni penguins at this site. This study represents the first complete in- ventory of Rockhopper Penguin numbers on Isla Noir through direct observation of each individual colony using standardized meth- odology. Population trends of the Southern Rockhopper Penguin on Isla Noir appear to mirror documented increases along the coast of Argentina (Schiavini 2000). Previously published estimates of 70,000 pairs of South- ern Rockhopper Penguins on Isla Noir (Ve- negas 1984) are significantly lower than the 158,200 (139,716-176,700) pairs of Rockhop- pers present in November 2005. However, site inventories by Venegas ( 1984) did not encom- pass the entire island and occurred during the later portion of the breeding season (Dec). In- ability to access all colonies and attrition due to nest failure could partially account for the lower counts in 1983. The initial survey by Venegas (1984) described a signilicant popu- lation of Southern Rockhopper Penguins on Isla Noir, but those results may have under- estimated population size and cannot reliably be used to demonstrate an increase in the pop- ulation since 1983. We cannot definitively conclude, based on 580 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 past estimates, the Rockhopper Penguin pop- ulation on Isla Noir has increased in numbers in the last 20 years. However, evidence exists of the physical expansion of colonies through modifications in vegetation. The Rockhopper colonies along the coast of Isla Noir were sim- ilar in structural characteristics. Colony ex- pansion into boundary areas containing trees, such as Nothogagus antarctica, may cause a reduction of N. antarctica, creating suitable conditions for succession of grass taxa. Cor- relation of active nests and weak or dead N. antarctica could demonstrate an expansion of the Rockhopper colonies and the related veg- etation within the study sites, due to associ- ated edaphic disturbances similar to those documented in seabird colonies (Ellis 2005). Evaluation of the remaining Southern Rockhopper Penguin colonies should continue where past estimates were yacht-based and lack accuracy. Continued evaluation of the numbers of penguins on Isla Noir should also be conducted to better refine the estimated numbers and to document trends. Past esti- mates did not use appropriate methodologies and provided incomplete estimates of the pen- guin populations. Our census results provide suitable baseline data for future studies and could be used to review the conservation sta- tus of the subspecies, and factors potentially responsible for population declines elsewhere. These data support designation of Isla Noir as an important nesting area for Southern Rockhopper Penguins. Sustainable manage- ment of Isla Noir for nesting Rockhoppers would require additional legal protection for the island. Developing a management plan will require detailed information on the pop- ulation and further study of the biology of the penguins, including foraging areas and popu- lation recruitment. Isla Noir is presently not within Chile’s park system, it lacks statutory protection, and is not monitored as closely as other areas containing penguin colonies. Man- agement of Isla Noir would offer future pro- tection for the existing population of Rock- hoppers and provide a favorable breeding site that may attract birds from regions where pop- ulations are under adverse pressures. ACKNOWEEDGMENTS We thank the reviewers and editor for helpful com- ments on this manuscript. This work was supported through Feather Link, Inc. by grants-in-aid for scien- tific research from local and international foundations. We thank Brian and Holly Hunt of African Safari Wildlife Park, Indianapolis Chapter of the American Association of Zoo Keepers, Faunia, Cincinnati Zoo and Botanical Garden, and the Wave Foundation of the Newport Aquarium for financial support of this and past expeditions. We also thank Bob Brinkman at Woolpert LLP for assistance with the AutoCAD sys- tem, and Claudio Venegas for support. We express spe- cial appreciation to our contacts in Chile, the crew of the Chonos, staff members of Fantastico Sur Birding and Nature, and government officials of the Ministerio de Economia, Fomento y reconstruccion, Republica de Chile for granting permission to work on Isla Noir. We thank the crew and staff of the M/S Nordnorge for close approaches to Terhalten. Our gratitude is also expressed to Clinton Nagy, Edward Maruska, and Greg Hanson for their continued support and participation in this project. LITERATURE CITED Araya, B. and G. Millie. 1986. Guia de campo de las aves de Chile. Editorial Universitaria, Santiago de Chile. Barlow, K. E., I. L. Boyd, J. P. Croxall, K. Reid, I. J. Staniland, and a. S. Brierley. 2002. Are pen- guins and seals in competition for Antarctic krill at South Georgia? Marine Biology 140:205-213. Bingham, M. and E. Mejias. 1999. Penguins of the Magellan Region, Scientia Marina 63: Supple- ment 1:485-493. BirdLife International 2007a. Species factsheet: Eu- dyptes chrysocome. http://www.birdlife.org (ac- cessed 7 June 2007). BirdLife International 2007b. Species factsheet: Eu- dyptes chrysolophus. http://www.birdlife.org (ac- cessed 7 June 2007). Bogaert, P, j. Delince, and S. Kay. 2005. Assessing the error of polygonal area measurements: a gen- eral formulation with applications to agriculture. Measurement Science and Technology 16:1170- 1 178. CCAMLR. 1999. Commission for the Conservation of Antarctic Marine Living Resources. CCAMLR Ecosystem Monitoring Program. Standard Method Publication, Tasmania, Australia. Clark, G. 1988. The Totorore voyage. Century Hutch- inson Publisher, Aukland, New Zealand. Clausen, A. P. and N. Huin. 2003. Status and numer- ical trends of King, Gentoo, and Rockhopper pen- guins breeding in the Falkland Islands. Waterbirds 25:389-402. Crawford, R. J. M., J. Cooper, and B. M. Dyer. 2003. Population of the Macaroni Penguin Eudyp- tes chrysolophus at Marion Island, 1994/95-2002/ 03, with information on breeding and diet. African Journal of Marine Science 23:475—586. Crawford, R. J. M., B. M. Dyer, J. Cooper, and L. G. Underhill. 2006. Breeding numbers and suc- cess of Eudyptes penguins on Marion Island, and Oehler et al. • CRESTED PENGUINS IN SOUTHERN CHILE 581 the influence of mass and time of arrival of adults. CCAMLR Science 13:175-190. Downes, M. C., E. H. M. Ealey, A. M. Gwynn, and P. S. Young. 1959. The birds of Heard Island. Australian National Antarctic Resource Expedi- tions Report Series Bl. Ellis, C. E. 2005. Marine birds on land: a review of plant biomass, species richness, and community composition in seabird colonies. Plant Ecology 181:227-241. Ellis, S., J. P. Croxall, and J. Cooper. 1998. Penguin conservation and assessment plan. Apple Valley: lUCN/SSC Conservation Breeding Specialist Group (1996 CAMP Report). Cape Town, South Africa. Kirkwood, R., K. Lawton, C. Moreno, J. Valencia, R. Schlatter, and G. Robertson. 2007. Esti- mates of Southern Rockhopper and Macaroni pen- guin numbers at the Ildefonso and Diego Ramirex Archipelagos, Chile, using quadrat and distance- sampling techniques. Waterbirds 30:259-267. Krebs, C. J. 1999. Ecological methodology. Second Edition. Addison-Wesley Educational Publication, Menlo Park, California, USA. Krebs, C. J. 2002. Ecological methodology. Second Edition Software. Department of Zoology, Uni- versity of British Columbia, Vancouver, Canada. Oehler, D. A., W. R. Fry, L. A. Weakley, and M. Marin. 2007. Rockhopper and Macaroni penguin colonies absent from Isla Recalada, Chile. Wilson Journal of Ornithology 119:502-506. PuTZ, K., A. P. Clausen, N. Huin, and J. P. Croxall. 2003. Re-evaluation of historical Rockhopper Penguin population data in the Falkland Islands. Waterbirds 26:169-175. Reilly, P. 1994. Penguins of the world. Oxford Uni- versity Press, South Melbourne, Australia. SCHIAVINI, A. C. M. 2000. Staten Island, Tierra del Fuego: the largest breeding ground for Southern Rockhopper Penguins? Waterbirds 23:286-291. Soto, N. 1990. Proyecto de proteccion y manejo de las colonias de pingUinos presentes en isla Rupert e isla Recalada, Reserva Nacional Alacalufes. In- forme de temporada 1989-1990. CONAF-XII Re- gion, Punta Arenas, Chile. Thompson, K. R. 1993. Falkland Islands Seabird Mon- itoring Programme summary of results. Falklands Conservation Report SMP/3. Stanley, Falkland Is- lands. Venegas, C. 1984. Estado de las poblaciones de Pin- guino de Penacho Amarillo y Macaroni en la Isla Noir, Chile. Informe Instituto de la Patagonia 33. Venegas, C. 1991. Estudio de cuantificacion pobla- cional de pinguinos crestados en Isla Recalada. Informe Instituto de la Patagonia 55. Venegas, C. 1998. Pinguinos crestados {Eudyptes chrysocome Forster 1781, E. chrysolophus Brant 1837) y de Magallanes {Spheniscus rnegellanicus Forster 1781) en isla Noir, Chile. Anales del In- stituto de la Patagonia 26:59-67. Venegas, C. and N. Soto. 1992. Estudio de pinguinos eudyptidos en isla Recalada, R. N. Alacalufes, Chile. Informe Convernio UMAG-CONAF, Punta Arenas, Chile. Wallace, G. E. 1991. Noteworthy bird records from southernmost Chile. Condor 93:175-176. WoEHLER, E. J. 1993. The distribution and abundance of Antarctic and subantarctic penguins. SCAR, Cambridge, United Kingdom. WoEHLER, E. J. AND J. P. Croxall. 1997. The status and trends of Antarctic and subantarctic seabirds. Marine Ornithology 25:43-66. The Wilson Journal of Ornithology 120(3):582— 593, 2008 NEST HABITAT SELECTION OE WHITE- WINGED SCOTERS ON YUKON ELATS, ALASKA DAVID E. SAFINE>-23 AND MARK S. LINDBERG' ABSTRACT. — Breeding bird surveys indicate a long-term decline in numbers of scoters (Melanitta spp.) breeding in North America. Little is known about the breeding habitat and reproductive life history of White- winged Scoters (M. fusca) in their primary breeding areas in the boreal forest of Alaska and northern Canada. We characterized selection of nest habitats and attributes within those habitats by measuring variables at nests and random sites on the Yukon Flats National Wildlife Refuge, Alaska. White-winged Scoters avoided nesting in meadows, but nested in scrub or forested habitat types in proportion to their availability (x^s = 9.7, P = 0.08). Nests of radio-marked females were farther from water and edge ( + 210 ± 43 and +10 ± 4 m, respec- tively), and in slightly thicker cover ( + 6 ± 4%) than nests located without aid of radio transmitters. Females selected sites with more variable and abundant overhead and lateral cover, and sites closer to edge and water than random sites. The results imply nearly random use of scrub and forested habitat types within the study area, but selective use of attributes within those habitat types. This generalist approach to nest site selection at a larger scale may be an adaptive response to reduce detection by nest predators. Nests located without use of radio-marked females may not be representative of the population of nests at a study site. White-winged Scoters often selected nest sites with dense cover far from water, which may increase nest survival. However, concealed sites are difficult for heavy-bodied birds to escape and females may be trading productivity against their own mortality. Received 3 November 2006. Accepted 26 December 2007. White-winged Scoters {Melanitta fusca) breed from the Canadian prairies north and west through the boreal forest of Canada into interior Alaska (Bellrose 1980). The majority of the 884,000 scoters surveyed in North America breed in the northern boreal forest of Canada and Alaska (Canadian Wildlife Ser- vice Waterfowl Committee 2006). The Yukon Flats National Wildlife Refuge (hereafter Yu- kon Flats) in eastern interior Alaska has one of the densest populations of breeding White- winged Scoters in North America (Bellrose 1980). North American surveys of scoters (Black [Melanitta nigra]. Surf [M. perspicil- lata], and White-winged scoters) indicate breeding populations have declined 1.1% per year in areas surveyed since 1961 (Canadian Wildlife Service Waterfowl Committee 2006). The Alaska population of breeding scoters has been stable or gradually declining (-0.4%/ year, P > 0.05), whereas scoter populations in the western Boreal Canada and Canadian Prai- rie strata have been declining more rapidly (-1.3 and -4.6%/year, respectively) since ' University of Alaska Fairbanks, Department of Bi- ology and Wildlife, and Institute of Arctic Biology, Fairbanks, AK 99775, USA. 2 Current address; Alaska SeaLife Center, R O. Box 1329, Seward, AK 99664, USA. ^ Corresponding author; e-mail: david_safine@alaskasealife.org 1961 (Canadian Wildlife Service Waterfowl Committee 2006). Studies of breeding White-winged Scoters (hereafter scoter) in North America are almost exclusively limited to island nesting popula- tions in the prairie-parkland of Saskatchewan and Alberta (Brown and Brown 1981, Kehoe 1989, Traylor et al. 2004). However, the land- scape and plant communities of prairie park- land, grasslands and agricultural fields inter- spersed with groves of deciduous trees are quite different from the northern boreal forest, which is dominated by coniferous trees and includes a much lesser extent of grasslands (Johnson et al. 1995). Nest sites in the boreal forest may differ from those in the southern portion of the breeding range where less for- ested area is available. Characterizing the breeding habitat of White-winged Scoters in Alaska and Canada is an information need identified by the Sea Duck Joint Venture Man- agement Board (2001). Oil and gas develop- ment has been proposed on both the Macken- zie Delta (Haszard 2001) and Yukon Flats (USDI 2005), both important scoter breeding areas. Describing patterns of nest habitat use in the northern boreal forest will provide base- line information important to managers de- veloping future conservation plans (Haszard 2001, 2004). Nest site selection is also important in un- 582 Safine and Lindberg • NEST HABITAT SELECTION OE SCOTERS 583 derstanding population dynamics because nest location can affect nest (Martin 1993b, Fillia- ter et al. 1994, Gloutney and Clark 1997) and female survival; thus, individual females may have trade-offs between selecting sites that maximize nest survival while minimizing their own mortality risk (Hoekman et al. 2002a). This trade-off may have population- level effects because both nest and female sur- vival are relatively important factors affecting population growth (Hoekman et al. 2002b). Nest site characteristics that can affect sur- vival include habitat type and vegetation lay- ers (Crabtree et al. 1989; Martin 1993a, 1995); nests on islands often have higher survival than those on the mainland (Lokemoen and Woodward 1992, Walker et al. 2005). Dis- tance of nests from water and edge (Clark and Shutler 1999, Traylor et al. 2004), vegetative heterogeneity (Crabtree et al. 1989), and cover at nest sites (Badyaev 1995, Clark and Shutler 1999, Traylor et al. 2004) can also affect nest survival. White-winged Scoters in previous studies have been observed nesting far from water in dense and often thorny shrubs, and on islands (Brown and Brown 1981, Brown and Fredrickson 1989, Traylor et al. 2004); following the general patterns of nest site se- lection in waterfowl. We believed that scoters in the boreal forest would select similar sites to those in prairie-parkland and predicted we would observe greater vegetative cover and variability, greater distances to water and edge, and more scrub plant communities at scoter nests than at random sites. Quantifying habitat differences between nests and random sites has revealed patterns of habitat use that have improved survival of nests and females over evolutionary time (Clark and Shutler 1999). Additionally, be- cause selection can be quantified hierarchical- ly (Johnson 1980), we believed it would be useful to investigate differences between nests and random sites at multiple scales. The ob- jectives of this study were to examine patterns of site use for nesting White-winged Scoters in the northern boreal forest at two spatial scales; (1) comparison of used and available habitat types, “third-order” .selection or the selection of specific habitat components with- in a home range (Johnson 1980); and (2) com- parison of the site attributes of nests and ran- dom points, “fourth-order” selection or a more specific level of use within that habitat type (Johnson 1980). METHODS Study Area. — We collected data during the breeding season (May-Aug) from 2002 to 2004 on the Yukon Flats, —170 km north of Fairbanks, Alaska (Fig. 1). The Yukon Flats includes —3.5 million ha along the Yukon River floodplain in east-central Alaska and en- compasses the largest interior wetland basin in Alaska (Heglund 1988). This basin is an area of major importance under the North American Waterfowl Management Plan (USDI 1986). We studied breeding ecology at the Scoter Lake complex (66° 14' N, 146° 23' W) in south central Yukon Flats. This area includes a series of relatively large (—1.5 km long) inter-connected lakes and boreal forest (taiga) habitat covering —4,400 ha. The forest habitats are dominated by: white and black spruce (Picea glauca and P. mariana, respec- tively), paper birch (Betula papyrifera), and trembling aspen (Populus tremuloides). Wil- low (Salix spp.), shrub birch {Betula glandu- losa), alder (Alnus spp.), and immature or stunted tree species dominate the scrub habi- tats. Grasses (e.g., Calamagrostis spp. and Hordeum spp.), sedges {Carex spp.), and emergent plants (e.g., Typha spp., Scirpiis spp., and Nuphar spp.) predominate in her- baceous habitats. Nest Searching. — White-winged Scoters of- ten nest far from water or in thick cover (Brown and Brown 1981) and we used two different methods to locate nests: foot search- es with the aid of a dog (Kehoe 1989) and tracking of females marked with radio trans- mitters prior to nesting. We captured scoters by driving them into floating mist nets (Kai.ser et al. 1995), modified for duck capture, from 31 May to 13 June 2002— 2004. We outfitted females with prong and suture radio transmit- ters (Model A4430, 9 g. Advanced Telemetry Systems, Isanti, MN, USA; Mauser and Jarvis 1991, Rotella et al. 1993) modified with glue. Each female was tracked daily from the ground and once or twice weekly from an air- plane until we cither found her nest or con- firmed her as a failed or non-breeder. We at- tempted to ascertain the status of females lo- cated on water (paired or not paired) without 584 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 FIG. 1. Scoter Lake complex, south central Yukon Flats National Wildlife Refuge, Alaska. Darkened areas indicate water in the study area, 2002-2004. flushing them. We attempted to find nests of females on land without flushing them. We searched lakeshores, islands, peninsu- las, bog perimeters, and areas within 600 m of water for nests on foot from 0800 to 1600 hrs ADT from 21 June to 17 July each year. We defined a scoter nest as a depression with either down, eggs, or contour feathers identi- fied as White-winged or Surf scoter. We found nests initiated in the current or previous year and included active, destroyed, and hatched scoter nests in our sample. We were able to include nests initiated in the previous year be- cause typically sufficient feathers and/or egg shells remained in the depression to positively identify the nest. We recorded latitude and longitude data for all nests with a compact Global Positioning System (GPS) unit (± 6 m accuracy). Nest Habitat. — We entered GPS coordinates of all nests from 2002 to 2004 into a data base and plotted them on Arc View 3.3 (Environ- mental Systems Research Institute, Redlands, CA, USA) geographic information system program. We used the Animal Movement ex- tension (Hooge and Eichenlaub 2000) to draw a minimum convex polygon for the entire sample of nest sites and generated 80 random locations within this polygon. Random sites were spaced at least 200 m apart with no dis- tance to polygon border restrictions. We ex- cluded random sites in lakes, but visited all sites within 50 m of the mapped lakeshore, as lake levels have changed since U.S. Geolog- ical Survey maps were developed in 1956. We visited all nest and random sites from 28 July to 14 August in 2003 and 2004 to record site characteristics. We recorded (1) habitat type, (2) edge type, (3) distance to edge, (4) distance to water, (5) overhead cover. Safine and Lindherg • NEST HABITAT SELECTION OF SCOTERS 585 and (6) lateral cover at each site. We measured this suite of variables because we predicted they would affect female, nest, and duckling survival and, potentially, the process of site selection. We classified habitat type using the level II categories in the Alaska Vegetation Classification (Heglund 1992, Viereck et al. 1992) defined as the proportion of cover types in a 10-m radius circle centered on the nest or random point. We defined distance to edge as the distance to the nearest change in habitat type (Clark and Shutler 1999, Clark et al. 1999) and edge type as the habitat type pre- sent beyond that change or nearest different habitat type (Clark et al. 1999). We measured distance to water (Clark and Shutler 1999, Traylor et al. 2004) as the minimum distance to a body of water sufficiently large to appear on infrared photographs of the area. We marked additional points 5 m from the nest or random location to better characterize each site. In 2003, we visited nests found in 2002 and 2003, and marked four additional points in the cardinal directions around these nests. In 2004, we visited nests located that year and all random points marking two ad- ditional points at random bearings around each site. We reduced additional points (from 4 to 2) in 2004 because of logistical con- straints associated with the four-fold increase in number of sites to visit that year. We recorded overhead and lateral cover only at each of the two (four) outside points (Fig. 2.) We measured overhead cover (Clark and Shutler 1999, Traylor et al. 2004) with a spherical convex crown densiometer placed on the ground and averaged from the four car- dinal directions. We measured lateral cover as the average percent obstruction of five white 6.5-cm^ blocks on a black cardboard square (Clark and Shutler 1999) viewed 2 m from the side at a height of 60 cm taken from the four cardinal directions. Each site was character- ized by the average value of overhead and lat- eral cover measured at the center and outside points. We defined overhead and lateral cover variation as the standard deviation of the three or five measurements of overhead and lateral cover at each site. We sampled additional ran- dom sites in the dwarf tree and tall scrub hab- itats after visiting all random sites, because they were rare. We needed to increase the sample of random sites in the two rare habitat FIG. 2. Sampling protocol for nests and random points. 1 = nest site or GPS coordinates of the random point, 2 and 3 are additional sampling points 5 m from the center, and 4 is the habitat type of the 10-m circle around the center point. X and Y are random bearings for additional points. types to be greater than or equal to the number of nests in rare habitat types. Use versus Availability. — We performed all statistical analyses using SAS software (SAS Institute 1999). We used a Chi-square test of homogeneity (PROC FREQ) to test for equal proportions of nests and random sites in each habitat type because availability was estimat- ed from random sites (Marcum and Lofts- gaarden 1980, Thomas and Taylor 1990). We performed the analysis with and without hab- itat types that were commonly available but rarely used (Thomas and Taylor 1990). We also performed a test of homogeneity on the proportion of edge habitat types at both nest and random sites in habitat types used for nesting. We verified that expected frequencies were greater than one and the average ex- pected frequencies were greater than six to as- sure appropriateness of the Chi-square test (Zar 1999). Effects of Sampling Design and Nest Searching Method. — We examined differences in mean cover and variation values at nests in 2003 and 2004 (5 and 3 sampling points per site, respectively; PROC TTEST) to ascertain if both protocols provided sufficiently similar results to warrant their combination in the 586 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 TABLE 1. Proportion of White-winged Scoter nests and random points in each habitat and edge type with associated cell Chi-square values from the test of homogeneity; Scoter Lake complex, Yukon Flats National Wildlife Refuge, Alaska, 2002—2004. Nests Random sites Overall n % Cell n % Cell Total x^ P value Habitat type Coniferous forest 16 40 0.0 26 43 0.0 Deciduous forest 2 5 0.1 4 7 0.0 Mixed forest 13 33 0.1 17 28 0.1 Dwarf tree scrub 5 13 2.9 1 2 1.9 Tall scrub 4 10 0.0 6 10 0.0 Graminoid herbaceous 0 0 2.8 7 12 1.8 Totals 40 100 5.9 61 100 3.9 9.7 0.08 Edge type Coniferous forest 7 18 0.2 7 13 0.2 Deciduous forest 3 8 0.2 6 11 0.1 Mixed forest 7 18 0.0 10 19 0.0 Dwarf tree scrub 4 10 0.9 2 4 0.6 Tall scrub 4 10 2.0 15 28 1.4 Graminoid herbaceous 5 13 0.4 4 7 0.3 Water 9 23 0.1 10 19 0.1 Totals 39 100 3.8 54 100 2.7 6.5 0.37 analysis. We also investigated differences in nest site characteristics attributable to the search method (radiotelemetry or ground searches with aid of dogs; PROC TTEST) to understand how method used may affect sub- sequent analyses. Site Attributes. — We used logistic regres- sion (PROC LOGISTIC) to characterize which habitat features affected selection (Alldredge et al. 1998) within habitat types. Our sampling protocol was consistent with a use-availability study as an approximation to a case-control design, which requires the as- sumption that use of available sites is rare (Keating and Cherry 2004). We interpreted the results of logistic regression as odds ratios and not resource selection functions as we were making the rare use assumption (Keating and Cherry 2004). The explanatory variables used in the mod- els were habitat type, distances to edge and water, lateral and overhead cover, and varia- tion of lateral and overhead cover. We includ- ed 12 additional random points to increase sample size in rare habitat types to achieve a sample of available sites in approximate pro- portion to use. We used habitat type to explain variation in the logistic regression models be- cause we sampled in proportion to use, but not to infer selection of habitat types themselves. We investigated correlations among explana- tory variables with correlation analysis (PROC CORR) and scatter plots. We chose an a priori model set of 41 biologically relevant models and all models with more than two parameters included only additive effects. We used Akaike’s Information Criteria (Akaike 1973) adjusted for small sample size (AIC^; Burnham and Anderson 1998) for model se- lection. We tested goodness-of-fit to the logis- tic model with the Hosmer and Lemeshow (1989) statistic. Values reported are means ± SE. RESULTS Use versus Availability. — We visited ran- dom sites that were on land {n = 61) in all six terrestrial habitat types: coniferous, decid- uous, and mixed forest; dwarf tree and tall scrub; and graminoid herbaceous. We located scoter nests {n = 3, 17, and 20 [one of which was a Surf Scoter] in 2002, 2003, and 2004, respectively) in five of the six terrestrial hab- itat types on the study area; only graminoid herbaceous habitat was unused (Table 1). The edge habitat at nests and random points in- cluded all six terrestrial habitats plus water (Table 1). Nests {n = 40) and random sites {n Safine and Lind berg • NEST HABITAT SELECTION OE SCOTERS 587 TABLE 2. Site attributes of White-winged Scoter nests and random points; Scoter Lake complex, Yukon Flats National Wildlife Refuge, Alaska, 2002-2004. Nests (n = 39) Random points (« = 62) Parameter® Mean SE Mean SE Difference'’ SE Distance to edge 12.3 2.0 29.2 5.3 *-16.9 6.9 Distance to water 144.3 26.1 240.6 23.3 *-96.3 35.9 Overhead cover 78.9 1.8 73.8 1.9 5.1 2.8 Overhead cover variation 12.1 1.4 11.0 1.0 1.1 1.7 Lateral cover 55.7 2.4 45.1 2.5 *10.6 3.7 Lateral cover variation 16.3 1.2 13.9 1.0 2.4 1.6 a Distance in meters, cover as percent obstruction, and variation in cover as standard deviation of percent obstruction, b * p-value <0.01. = 54 without graminoid sites, « = 61 with graminoid points) were present in the same proportions among habitat types (Table 1) whether or not we included the graminoid her- baceous habitat type that was available but un- used (without graminoid; x^4 - 4.7, P = 0.32). The proportion of edge type at nests {n = 39) and random sites (n = 54) was equal among the seven edge habitat types (Table 1). Effects of Sampling Design and Nest Searching Method. — We located 40 nests of which six were initiated the year prior to dis- covery; 15 nests were found by monitoring radio-marked females and 25 nests were found while conducting ground searches with a dog. There were minimal differences in mean overhead cover (2 ± 4%) and lateral cover (7 ± 5%) between nests with three sam- pling points {n = 19) and five sampling points {n = 20). Mean variation in overhead (0 ± 3%) and variation in lateral cover ( — 2 ± 2%) did not differ between nests with three sam- pling points {n = 19) and five sampling points (/I = 20). Sampling design differences be- tween years did not affect parameter estimates in the regression model and combining sam- pling designs was warranted. However, mean distances to water and edge were greater at nests found using telemetry ( + 210 ± 43 and + 10 ± 4 m, respectively) than at nests found with ground searches; there was some evi- dence that overhead cover was also greater ( + 6 ± 4%). Site Attributes. — We included random points (n = 50) in the five habitat types that scoters used for nesting: coniferous, decidu- ous, and mixed forest, dwarf tree and tall scrub, and additional points in dwarf tree and tall scrub {n = 9 and 3, respectively). Four random points had water within 5 m and were not included in the analysis because water re- stricted the directions available to place the additional points. Nests were closer to edge and water, and had denser and more variable cover than random points (Table 2). The best approximating model in the logis- tic regression was that site use depends on the additive relationship of all measured variables (distance to edge and water, overhead and lat- eral cover, variation in both overhead and lat- eral cover, and habitat type simplified into 2 categories, forest or scrub). Five other models were within seven AIQ units, but none was within two AIC^ units of the most parsimo- nious model (Table 3). The top three models were similar and all included distance effects (edge and water) and cover effects (overhead and lateral). The top model included habitat and variation in cover effects, whereas the next best model added only variation in cover effects. The Hosmer and Lemeshow good- ness-of-fit test indicated the most parameter- ized model fit the logistic model (x^s ~ 4.6, P = 0.80). Probabilities predicted by the top model were 85% concordant and 15% discor- dant with the observed data. Coefhcient and odds ratio estimates from the top model (Ta- bles 4, 5) indicated all variables explain var- iation in the data with the exception of vari- ation in lateral cover. The odds of site use de- creased at sites farther from water and edge; however, odds of site use increased with great- er lateral and overhead cover as well as vari- ation in lateral and overhead cover. In the top model, distance to water changed the odds ra- tio of Li.se much slower than distance to edge (Fig. 3); the same relationship was true for the odds ratios of lateral and overhead cover (Fig. 588 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 TABLE 3. Model selection from logistic regression of White-winged Scoter nest site attributes and location (nest or random point); Scoter Lake complex, Yukon Flats National Wildlife Refuge, Alaska, 2002-2004. Models are shown with sources of variation in location; plus symbols indicate additive relationships among parameters. Number of parameters (k), -2*log likelihood (-2log(l)), the difference in Akaike’s information criterion adjusted for small sample size from the best approximating model (AAIQ), and the coefficient of determination {R^) are included with results from models with AAIQ. < 7. ModeF A -2 log120 m; a negative effect. A negative distance to edge effect was also reported for Ring-necked Pheasants {Phasianus colchicus) (Clark et al. 1999), and White-winged Scoters nested closer to edge than random sites at Redberry Lake, Saskatch- ewan (Traylor et al. 2004). In contrast, dis- tance to edge did not differ between nests and unused sites for Wild Turkeys (Meleagris gal- lopavo) (Badyaev 1995), and in five species of dabbling ducks (Clark and Shutler 1999). Distance to edge for scoters has important implications for nest and female survival. Be- ing farther from edge may improve nest sur- vival (Clark and Shutler 1999) but decrease female survival as they are farther from suit- able escape cover often present at edges. If edge habitat is lower or more open than nest- ing habitat, it may form an opening sufficient for these heavy-bodied birds with “relatively low and slow take-offs” (Brown and Fred- rickson 1989: 245) to fly safely from ap- proaching nest predators. Nesting near open- ings may be extremely important for female survival. Most nests were within 10 m of an opening suitable for escape, but often this opening was too small to be recorded as a unique edge at the scale used in this analysis. The odds of use slowly decreased as dis- tance to water increased. Nests were on av- erage closer to water than random points (142.7 ± 25.5 m and 231.3 ± 22.8 m, re- spectively), but sufficient nests were farther from water than random points (18% of the sample) to produce a gradually declining odds ratio. This pattern is similar to that of White- winged Scoters in Saskatchewan, which se- lected nest sites approximately the same dis- tance from water as random points (—107 m; Traylor et al. 2004). Scoters are known to nest long distances from water (Bellrose 1980, Brown and Brown 1981, this study), but what advantage do scoters receive for nesting far from the safety of lakes? Nesting farther from water may improve nest survival sufficiently to offset potentially negative impacts on fe- males and ducklings during long distance movements to brood rearing habitats. Nesting farther from travel routes of mammalian nest predators (e.g., shorelines and habitat edges) may improve female and nest survival (Brown and Fredrickson 1989, Paton 1994). Nesting scoters appear to be generalists; individual fe- males place their nests varying distances to water and edge in most habitat types, and then seek thick cover at those sites. Scoters nested in sites with more overhead and lateral cover than random sites, similar to Safine and Lindberg • NEST HABITAT SELECTION OF SCOTERS 591 other waterfowl species (Lokemoen et al. 1984, Clark and Shutler 1999, Traylor et al. 2004). High levels of cover were likely se- lected by females because they improved nest survival (Badyaev 1995, Clark and Shutler 1999, Traylor et al. 2004). This strategy may reduce detection of the nest, but well-con- cealed sites are more difficult for these heavy- bodied ducks to exit. Relatively high female mortality (survival probability of 0.80) was observed during the nesting period for White- winged Scoters on this study area as females often nested in heavy cover (Safine 2005). Therefore, females need dense cover to con- ceal nests and less dense cover nearby for es- cape, which could be edge habitat or a small opening. The use of escape cover is likely the reason overhead cover variation was an im- portant variable in the model, higher levels of which increased the odds of use. On average, the three points sampled at random sites tend- ed to be more similar to each other, indicating more uniform cover at random sites. Scoters not only selected for high overhead cover, but they chose to place their nests at sites with more heterogeneity in cover. Scoters selected nest sites from a continu- um of available cover densities and distances to water and edge at the Scoter Lake complex. Placement of nests at various levels along this continuum constitutes different solutions to trade-offs in female, nest, and brood survival. Scoters represent waterfowl at one extreme of the cover and distance continuum, often nest- ing at sites with dense cover far from water (D. E. Safine, pers. obs.). Thus, we would ex- pect scoters to have the highest nest survival and lowest female survival during nesting. Dabbling ducks (Tribe Anatini) are in the cen- ter of the continuum and pochards (Tribe Ay- thyini) are on the other extreme, typically nesting in open sites with floating nests or near the water. We expect pochards to have the lowest nest survival because of the open habitat they select, but high female survival during nesting, as females may easily escape from nests. Despite their poor take-off capa- bilities and longevity. White-winged ^Scoters at the Scoter Lake complex often nested at the extreme of the cover and distance continuum. Over the long-term they must experience rel- atively high nest survival compared to dab- bling ducks and pochards; otherwise, this strategy would not persist. Nesting White-winged Scoters on the Yu- kon Flats had little selectivity at a larger scale (for specific habitat types) other than avoid- ance of graminoid meadows. However, scoters appear to be selective at a smaller scale (spe- cific sites within those habitats). This lack of selectivity at a larger scale may improve nest survival over other more selective duck spe- cies as predators cannot focus search efforts on scoter nests. ACKNOWLEDGMENTS This project was made possible through financial support from the U.S. Fish and Wildlife Service, Yu- kon Flats National Wildlife Refuge; University of Alaska Fairbanks, Department of Biology and Wildlife and Institute of Arctic Biology; the Jay Hammond Scholarship Fund; and the Sea Duck Joint Venture. We extend our gratitude to our field assistants for their hard work despite the difficult field conditions: K. H. Martin, H. F Knudsen, J. E. Minerva, S. J. Dufford, James Akaran, C. M. Harwood, Jessica Mensik, A. J. Leppold, K. A. Sesser, J. G. Carlson, Johann Walker, Casey Quitmeyer, S. L. Freeman, L. L. Kennedy, C. S. Van Stratt, M. R. Bertram, B. D. Whitehill, E. T. Heuer, S. W. Wang, Emma, and Minto. We thank the staff of Yukon Flats National Wildlife Refuge for their enthusiastic project support and providing extensive field equipment, and extend a special thanks to M. R. Bertram for assuring we had the equipment, personnel, and logistical support to achieve our goals, and M. T Vivion for countless hours of aircraft logistical support and radio-telemetry flights. We thank D. H. Rosenberg and M. J. Petrula, Alaska Department of Fish and Game, and P. L. Flint and J. L. Schamber, U.S. Geo- logical Survey, for use of bird capture equipment. We appreciate the GIS assistance provided by S. J. Dufford and N. J. Pamperin. J. W. Hupp, C. P. H. Mulder, E. C. Murphy, and two anonymous referees provided valuable reviews of previous versions of this manu- script. LITERATURE CITED Akaike, H. 1973. Information theory and an extension of the maximum likelihood principle. Pages 267- 281 in International symposium on information theory (B. N. Petran and F. Csaki, Editors). Aka- demiai Kiado, Budapest, Hungary. Ai.ldredgh, j. R., D. L. Thomas, and L. L. Mc- Donald. 1998. Survey and comparison of meth- ods for study of resource selection. Journal of Ag- ricultural, Biological, and Environmental Statis- tics 3:237-253. Badyai-v, a. V. 1995. Nesting habitat and nesting suc- cess of eastern Wild furkeys in the Arkansas O/ark Highlands. Condor 97:221-232. 592 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 Bellrose, F. C. 1980. Ducks, geese and swans of North America. Third Edition. Stackpole Books, Harrisburg. Pennsylvania. USA. Brown, P. W. and M. A. Brown. 1981. Nesting biol- ogy of the White-winged Scoter. Journal of Wild- life Management 45:38 — 15. Brown, P. W. and L. H. Fredrickson. 1989. White- winged Scoter populations and nesting at Redber- ry Lake. Saskatchewan. Canadian Field-Naturalist 103:240-247. Burnham, K. P. .and D. R. Anderson. 1998. Model selection and inference: a practical information- theoretic approach. Springer- Verlag, New York. USA. C.ANADLAN WILDLIFE SERVICE W.ATERFOWL COMMITTEE. 2006. Population status of migratory game birds in Canada: November 2006. Migratory Birds Reg- ulatory Report 19. Canadian Wildlife Service, Ot- tawa. Ontario. Canada. Clark. R. G. .and D. Shltler. 1999. Avian habitat selection: pattern from process in nest-site use by ducks. Ecology 80:272-287. Clark. W. R., R. A. Schmitz, .and T. R. Bogenschltz. 1999. Site selection and nest success of Ring- necked Pheasants as a function of location in Iowa landscapes. Journal of Wildlife Management 63: 976-989. Cr-abtree. R. L.. L. S. Broome, .and M. L. Wolfe. 1989. Effects of habitat characteristics on Gadwall nest predation and nest-site selection. Journal of Wildlife Management 53:129-137. Filllater. T. S., R. Breitwisch. .and P. M. Ne.alen. 1994. Predation on Northern Cardinal nests: does choice of nest site matter? Condor 96:761-768. Glol-tney, M. L. .and R. G. Cl.ark. 1997. Nest-site selection by Mallards and Blue-winged Teal in re- lation to microclimate. Auk 114:381—395. H.ASZ.ARD, S. L. 2001. Habitat requirements of White- winged and Surf scoters in the Mackenzie Delta region. Northwest Territories. Arctic 54:472-473. H.ASZ.ARD, S. L. 2004. Habitat use by White- winged Scoters (Melanitia fiisca) and Surf Scoters {Me- lanitta perspicillata) in the Mackenzie Delta re- gion, Northwest Territories. Thesis. University of Saskatchewan, Saskatoon, Canada. Heglund, P. J. 1988. Relationship between waterbird use and limnological characteristics of wetlands on the Yukon Flats National Wildlife Refuge, Alaska. Thesis. University of Missouri, Columbia. USA. Hegll'ND, P. J. 1992. Patterns of wetland use among aquatic birds in the interior boreal forest region of Alaska. Dissertation. University of Missouri, Co- lumbia. USA. Hoekman, S. T, I. J. B.all, .and T. F. Fondell. 2002a. Grassland birds orient nests relative to nearby vegetation. Wilson Bulletin 1 14:450 — J56. HOEKM.AN, S. T, L. S. Mills. D. W. Howtrter. J. H. Devries, .and I. J. Ball. 2002b. Sensitivity anal- yses of the life cycle of midcontinental Mallards. Journal of Wildlife Management 66:883-900. Hooge, P. N. and B. Eichenlaub. 2000. Animal move- ment extension to Arcview. Version 2.0. USGS, Alaska Science Center-Biological Science Office, Anchorage. USA. Hos.mer. D. W. and S. Lemeshow . 1989. Applied lo- gistic regression. John Wiley and Sons. New York. USA. Johnson. D. H. 1980. The comparison of usage and availability measurements for evaluating resource preference. Ecology 61:65-71. Johnson. J. D., L. Kershaw , A. M.acKinnon. a.nd J. Poj.AR. 1995. Plants of the western boreal forest and aspen parkland. First Edition. Lone Pine Pub- lishing, Renton. Washington, USA. Kaiser. G. W.. A. E. Derocher. S. Cr.awford, M. J. Gill, .and I. A. M.anley. 1995. A capture tech- nique for Marbled Murrelets in coastal inlets. Journal of Field Ornithology 66:321—333. Keatlng. K. a. .and S. Cherry. 2004. Use and inter- pretation of logistic regression in habitat-selection studies. Journal of Wildlife Management 68:774- 789. Kehoe, F. P. 1989. The adaptive significance of crech- ing behaviour in the White-w inged Scoter Melan- itta fiisca deglandi. Canadian Journal of Zoology 67:406-411. Lokemoen. j. T. .and R. O. Woodw ard. 1992. Nesting w aterfowl and w aterbirds on natural islands in the Dakotas and Montana. Wildlife Society Bulletin 20:163-171. Lokemoen, J. T, H. F. Duebbert, .and D. E. Sharp. 1984. Nest spacing, habitat selection, and behav- ior of w aterfow 1 on Miller Lake Island. North Da- kota. Journal of Wildlife Management 48:309— 321. Marcum. C. L. and D. O. Loftsg.a.arden. 1980. A nonmapping technique for studying habitat pref- erences. Journal of Wildlife Management 44:963- 968. Martln, T. E. 1993a. Nest predation among vegetation layers and habitat types: revisiting the dogmas. American Naturalist 141:897-913. Martin. T. E. 1993b. Nest predation and nest sites. Bioscience 43:523-532. Martin, T. E. 1995. Avian life history evolution in relation to nest sites, nest predation, and food. Ecological Monographs 65:101-127. M auser. D. M. .and R. L. J arvis. 1991. Attaching ra- dio transmitters to 1 -day-old Mallard ducklings. Journal of Wildlife Management 55:488-491. P.ATON, P. W. C. 1994. The effects of edge on avian nest success: how strong is the evidence? Conser- vation Biology 8:17-26. Rotella. j. j., D. W. Howerter. T. P. S.ankow ski. .and J. H. Devries. 1993. Nesting effort by wild Mal- lards w ith 3 types of transmitters. Journal of Wild- life Management 57:690-695. S.AFTNE, D. E. 2005. Breeding ecology of White- winged Scoters on the Yukon Flats, Alaska. The- sis. University of Alaska. Fairbanks. USA. Safine and Lindberg • NEST HABITAT SELECTION OE SCOTERS 593 SAS Institute. 1999. SAS. Version 8.0. SAS Institute Inc., Cary, North Carolina, USA. Sea Duck Joint Venture Management Board. 2001. Sea Duck Joint Venture Strategic Plan: 2001- 2006. Unpublished Report. Sea Duck Joint Ven- ture Continental Technical Team, USDI, Fish and Wildlife Service, Anchorage, Alaska. USA; Ca- nadian Wildlife Service, Sackville, New Bruns- wick, Canada. Thomas, D. L. and E. J. Taylor. 1990. Study designs and tests for comparing resource use and avail- ability. Journal of Wildlife Management 54:322- 330. Traylor, J. J., R. T. Alisauskas, and F. P. Kehoe. 2004. Nesting ecology of White-winged Scoters Melanitta fusca deglandi at Redberry Lake, Sas- katchewan. Auk 121:950-962. U.S. Dep.artment of Interior (USDI). 1986. North American waterfowl management plan: a strategy for cooperation. USDI, Fish and Wildlife Service, Washington, D.C., USA. U.S. Dep.artment of Interior (USDI). 2005. Evalua- tion and review of a proposed land exchange and acquisition of native lands within the Yukon Flats National Wildlife Refuge, Alaska. USDI, Fish and Wildlife Service, Anchorage, Alaska, USA. ViERECK, L. A., C. T. Dyrness, a. R. Batten, and K. J. Wenzlick. 1992. The Alaska vegetation clas- sification. USDA, Forest Service, General Tech- nical Report PNW-GTR-286. Pacific Northwest Research Station, Portland, Oregon, USA. Walker, J., M. S. Lindberg, M. C. M.acCluskie, M. J. Petrula, and j. S. Sedinger. 2005. Nest sur- vival of scaup and other ducks in the boreal forest of Alaska. Journal of Wildlife Management 69: 582-591. Z.AR, J. H. 1999. Biostatistical analysis. Fourth Edition. Prentice Hall, Upper Saddle River, New Jersey, USA. Short Communications The Wilson Journal of Ornithology 120(3):594-599, 2008 Use of Legs and Feet for Control by Scoters during Aerial Courtship William J. Wilson^ ABSTRACT. — Scoters (Melanitta spp.) exhibit ex- traordinary maneuvers during courtship flight, attitudes which are not commonly seen in flight. Scoters drop their legs and spread webbed feet during these maneu- vers. There appears to be a correlation between how the feet of scoters meet the airflow and the maneuver in progress. Received 25 April 2007. Accepted 5 Au- gust 2007. Block Island Sound, New York/Rhode Is- land, USA supports large populations of sco- ters {Melanitta spp.) in late winter and early spring. Groups of each species of scoter form courting parties at this time, comprised of a single hen with an entourage of drakes (Gunn 1927, McKinney 1959). Airborne courting parties in late winter exhibit uncommon flight patterns which include steep climbs (Wilson 1980), abrupt turns, and slow fluttering flight. Often during these maneuvers the birds drop their legs so the spread webbed feet can en- gage the airflow. I spent weekends through the 1970s pho- tographing scoters during late winter and early spring near Montauk, New York (41°03'N, 72° 00' W) at the eastern end of Long Island, where the Montauk Lighthouse faces the southern terminus of Block Island Sound and the adjoining Atlantic Ocean. The objective of this paper is to describe how the feet of sco- ters are used in aerial maneuvers during this early courtship period. Six figures are dis- cussed that support these descriptions. OBSERVATIONS A Surf Scoter {Melanitta perspicillata) courting party transitions from straight ahead flight to an abrupt turn beeause the lone hen (Fig. 1, arrow 1) departs the formation. In this sequence the highest drake brakes with both feet prior to turning. Just below him a drake (Fig. 1, arrow 2) rotates his extended right leg ' P. O. Box 2285, Montauk. NY 11954, USA; e-mail: jwwilson@optonline.net in a counter-clockwise direction about the roll axis to a position that, rather than below, is beside his body (Fig. 2). His open (braking) foot, being perpendicular to the ineident air- flow, is subject to only the force of drag. This force is proportional to the square of the air- flow velocity (Videler 2005:7) and, acting in the same direction as airflow and coupling to the bird’s body by the extended leg, produces clockwise rotational force about the yaw axis that is part of the mechanics of the turn. Similarly, the feet of at least two of the de- scending Black Scoter (M. nigra) (Fig. 3B, ar- rows 1 and 2) are presented to the headwind to produce only drag. The descending courting party is depicted (Fig. 3A) in varied transient attitudes, all with deployed webbed feet. The formation flutters to the water with a yawing motion resembling a falling leaf. Supporting this maneuver is a headwind that, judging from the white-caps on Block Island Sound, is a sea state that appears more rough than moderate, indicating a fresh breeze (Beaufort Scale #5) and wind velocity approximated at 30-39 km/hr. The wings of scoters 1 and 2 are configured in a negative dihedral that min- imizes control, while their feet are rotated about their roll axis. They shift centers of force that produce the oscillating yawing mo- tion and the strong headwind enhances this control. The birds’ large angle of attack is needed to counteract the resulting rotational force the feet produce about the pitch axis. In contrast, the feet of a Black Scoter eourting party (Fig. 4A) are swept back in execution of an abmpt turn. The hen (Fig. 4A, arrow 1) is initially seen in a large angle of attack and brak- ing prior to following a drake (Fig. 4A, arrow 2) out of formation. This drake is executing a banking turn with his swept back feet astride the spread tail. The feet are held at roughly right angles to each other, rather than parallel, and the surface of the right foot is held vertically. Sim- ilarly, the hen (Fig. 4B) is making her turn and also holds her right foot vertically (Fig. 5). In 594 SHORT COMMUNICATIONS 595 2 FIG. 1. Members of a Surf Scoter courting party execute an abrupt turn to follow the departing hen (arrow 1). A rarely seen turning technique is exhibited by the drake (arrow 2) with the extended foot. aviation terminology, i.s this right foot function- ing as a vertical stabilizer or a mdder? It is note- worthy the right foot is outboard of the turn, giving it a greater distance from the center of the turn and, thus, a greater angular velocity. The side-slipping Black Scoter (Fig. 3B, ar- row 3) appears almost totally devoid of pitch control due to the vertical attitude of the wings and tail. Control of pitch normally re- sults from the horizontal movement of the 596 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 1. Rotational Force 2. Drag FIG. 2. As the stationary paddle turns a canoe, so too does the force of drag acting on this outstretched foot create rotational force. wing (Brown 1961:301, Pennycuick 1975:52). The tail's primary use is for pitch (Brown 1961:302). However, the legs and swept back, spread-webbed feet must have significance, if only by default. A courting party of White-winged Scoter (M. fiisca. Fig. 6). when heading into the wind, executes a steep climb. A “slight sea state (Fig. 6) allows approximating the wind speed at 7-19 km/hr. a light to gentle breeze on the Beaufort Scale (#s 2-3). Typically, dur- ing this maneuver, the legs deploy the spread webs of the feet astride the tail. The lead drake is at the peak of the climb. The near vertical attitude of the feet in this terminal attitude eliminates increased lift as a reason for their deployment. Their exact function, however, remains unclear. DISCUSSION The literature is gradually reflecting that legs and webbed feet of water birds are used for attitude control and maneuver. Brown (1951) suggested it is generally accepted that birds' legs do not contribute to flight, while Barlee (r964:303) believed the feet can pro- duce unbalanced forces. Aymar (1935:32) and Pennycuick (1975:55) both suggest the auk's {Alcidae) small tail necessitates control and lift augmentation by the feet. Videler (2005: 27) advocates that, in many species, every as- pect of the body is related to flight and. in addition, that legs and feet can alter the center of gravity (Videler 2005:47). 1 FIG. 3. A descending Black Scoter courting party uses the strong headwind to perform this less than syn- chronized maneuver. The enlargement (B) better reveals foot positioning and use. SHORT COMMUNICATIONS 597 FIG. 4. A two-frame sequence of ~1 sec duration depicts a Black Scoter courting party executing an abrupt turn. Other than the hen in frame A (arrow 1) using her feet to brake, the entire formation has feet swept back. It is reasonable to assume there is an aero- dynamic reason why scoters deploy their legs and spread webbed feet in flight. An extended webbed foot clearly can exert a rotational force 1. Rotational Force 2. Reactive Force FIG. 5. Is the swept back right foot a vertical sta- bilizer, or is it a rudder? Any deflection of the airflow by the foot will result in a reactive force that pivots on the center of gravity, resulting in rotational force. about the bird’s center of gravity depending on how it interfaces with airflow. This rotational force is exerted upon the airborne body and is either balanced for a trim attitude, or the bird uses the imbalance to change attitude. The rea- son for the extended foot is elusive in some instances. This is not so for some scoters (Fig. 1, arrow 2; Fig. 3B, arrows 1 and 2). The force of drag acting on the foot with its resulting ro- tational force is obviously part of an ongoing maneuver; the use of the foot as a control sur- face is unambiguous. It is only when the foot, from the extended leg, is swept back that conclusions cannot be absolute. The vertically-held right foot (Fig. 4) is part of the maneuver but its function is open to speculation. The higher airflow veloc- ity over the right foot allows it to generate a greater reactive force, suggesting its use as a control surface. The side-slipping Black ,Scoter (Fig. 3B, ar- row 3) is using his swept back feet. At that instant his only maneuver is descent. That the surfaces of the feet are oriented perpendicular to one another suggests attitude control. 598 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 FIG. 6. Gravity surmounts thrust at the peak of a steep climb by White-winged Scoters. SHORT COMMUNICATIONS 599 If the lead White-winged Scoter (Fig. 6) was a fixed-wing aircraft, its thrust spent and in a large angle of attack, it would be devoid of control and prone to falling into a spin with possible catastrophic results. However, birds do not spin out. The drake (Fig. 6) maintains a delicate balance prior to the dive that must follow. In pondering how this is done, it is applicable to recall Videler’s (2005:27) pre- viously referenced observation, “Not only the wings and the tail but also the head, neck, the body and the hind limbs have features directly related to flight in many species.” ACKNOWLEDGMENTS I thank R A. Johnsgard (University of Nebraska- Lincoln), V. A. Tucker (Duke University), and two anonymous reviewers for commenting some thirty years ago on early drafts of this subject. I am grateful for the helpful comments made by the reviewers of this current manuscript. I am indebted to Lindsey Rohrbach (South Fork Natural History Museum) for organizing the digital imagery. Finally, I must ac- knowledge the literate comments, research, and editing by my wife Janice on whom my dependency is abso- lute. The original photographs are contained on 35 mm Kodachrome transparencies. LITERATURE CITED Aymar, G. C. 1935. Bird flight. Dodd, Mead and Company, London, United Kingdom. Barlee, J. 1964. Flight. Pages 299-307 in The new dictionary of birds (A. L. Thomson, Editor). Mc- Graw-Hill, New York, USA. Brown, R. H. J. 1951. Flapping flight. Ibis 93:333- 359. Brown, R. H. J. 1961. Flight. Pages 289-304 in Bi- ology and comparative physiology of birds. Vol- ume 2 (A. J. Marshall, Editor). Academic Press, New York, USA. Gunn, D. 1927. The courtship of the Common Scoter. British Birds 20:193-197. McKinney, E 1959. Waterfowl at Cold Bay, Alaska, with notes on the display of the Black Scoter. Wildfowl Trust Annual Report 10:133-140. Pennycuick, C. j. 1975. Mechanics of flight. Pages 1-75 in Avian biology. Volume 5 (D. S. Earner and J. R. King, Editors). Academic Press, New York, USA. ViDELER, J. J. 2005. Avian flight. Oxford University Press, Oxford, United Kingdom. Wilson, W. J. 1980. Inverted flight of White-winged Scoters during courtship flight. American Birds 34:746-747. The Wilson Journal of Ornithology 120(3):599-602, 2008 Bill Entanglement in Subcutaneously-anchored Radio Transmitters on Harlequin Ducks Jeanine BoncF’^"^ and Daniel EsleU ABSTRACT. — We report two incidences of Harle- quin Ducks (Histrionicus histrionicus) entangling their bills in subcutaneously-attached anchor transmitters. We suggest caution should be exercised when using these transmitters. Received 19 May 2007. Accepted I November 2007. ' Centre for Wildlife Ecology, Department of Bio- logical Sciences, Simon Fraser University, 8888 Uni- versity Drive, Burnaby, BC V5A 1S6, Canada. 2 Centre for Wildlife Ecology, Simon Fraser Uni- versity, 5421 Robertson Road, Delta, BC V4K 3N2, Canada. Current address: 2843 Woodland Drive, Vancou- ver, BC V5N 3P9, Canada. Corresponding author; e-mail: jbond@alumni.sfu.ca Radio telemetry is an important tool for wildlife research that can reveal valuable in- formation about wild birds that would other- wise be difficult to obtain, such as move- ments, home ranges, survival, and nest site at- tributes. However, methods of radio transmit- ter attachment vary and researchers must choose an attachment that provides adequate retention time, and minimizes detrimental ef- fects. Attachment methods on waterfowl have included backpack harnesses (Dwyer 1972), glue (Perry et al. 1981), tail-mount (Giroux et al. 1990), suture and glue (Wheeler 1991). an- chor-suture (Mauser and Jarvis 1991, Pietz et al. 1995), interscapular implants (Korschgen et al. 1996a), and intra-abdominal implants 600 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 3, September 2008 (Korschgen et al. 1996b). Some evaluations of radio transmitters have revealed deleterious effects such as feather wear, lower survival, reduced reproductive effort, behavioral chang- es, and mass loss (Paquette et al. 1997, Guyn and Clark 1999, Ackerman et al. 2004), whereas other studies have found no signifi- cant consequences of transmitters (Houston and Greenwood 1993, Esler et al. 2000, Hepp et al. 2002, Hupp et al. 2006, Iverson et al. 2006). Entanglement is an issue, with poten- tial direct effects on survival, that has been reported for harness transmitters (Keedwell 2001, Duriez et al. 2005), but less frequently with other attachment methods (Conway and Garcia 2005). We report two incidences in which the bills of Harlequin Ducks {Histrion- icus histriofiiciis) became entangled with sub- cutaneously-anchored radio transmitters. This information is important from both ethical and scientific perspectives and informs researchers of potential complications when using these transmitters. METHODS We investigated Harlequin Duck breeding and nesting attributes in southern British Co- lumbia during summer in 2003 and 2004. We used transmitters attached by a subcutaneous anchor (Pietz et al. 1995) to adult females to follow them through the breeding season. The transmitters, made by Holohil Systems Ltd. (model RI-2B, Carp, ON, Canada), were a 6-g model with a mortality switch and a life span from 3 to 9 months. They were circular, measured 2 cm in diameter, and had a maxi- mum height of 0.8 cm. We glued stainless steel anchors (Advanced Telemetry Systems, Isanti, MN, USA) to the transmitters opposite the antennae with marine-grade epoxy (Pig. 1). Each transmitter was attached in the small depression between the scapulae, dorsal to the approximate junction of the cervical and tho- racic vertebrae. The feathers were parted us- ing isopropyl alcohol and the site was pre- pared with Betadine solution (Purdue Freder- ick Co., Stamford, CT, USA). This attachment site was naturally void of contour feathers and only downy feathers needed to be removed. The transmitter anchor also was cleaned with Betadine solution. The skin was pinched and lifted at the attachment site, and a sterile 18- gauge hypodermic needle was used to punc- ture a small hole in the skin without contact- ing the underlying muscle, nerves or blood vessels. No local anesthetic was used because the incision was small and there is believed to be little enervation in this area (Altman 1981, Pietz et al. 1995). The anchor of the trans- mitter was inserted into the hole, curled end first, and manipulated until the anchor had the head of the arrow pointing forward and rest- ing flat on the bird’s back. A quick-drying, veterinary-grade adhesive (Vetbond, 3M, St. Paul, MN, USA) was used around the inser- tion site to seal the opening, and to glue the top and bottom of the transmitter to the feath- ers. The females were released immediately after the adhesive dried. These transmitters are normally shed after a period of weeks to months (Iverson et al. 2006). OBSERVATIONS We radio-marked 34 female Harlequin Ducks and monitored them from their arrival in breeding areas (late Apr) until the majority of females had departed for wintering areas (early Aug). We had two females (one each year) hook their bills on part of the anchor of the radio transmitter as it began to extrude during the shedding process. The females did not have their bills caught on the transmitter attachment in the same way. The bill of the first female was caught on the curled end of the anchor (Fig. 1), which was slightly open and pierced through the lower mandible of her bill while the looped side of the anchor was still attached subcutaneously to her back. The bill of the second female was caught when her lower mandible became wedged into the loop side of the anchor while the curled end was still attached to her back. The result for both birds was that their heads were held to their backs and, although they were still able to move about, they were likely unable to ac- quire food during this time. We recaptured both of these females and removed the trans- mitters. However, if this complication had oc- curred after we had completed monitoring, these females may have died from starvation or been highly vulnerable to predation. DISCUSSION No other study has reported this type of complication from using anchor transmitters. SHORT COMMUNICATIONS 601 ANCHOR I ■ I 0 '2 cm FIG. 1. Illustration of subcutaneously-anchored radio transmitter. We surmise that Harlequin Ducks are partic- ularly susceptible to this problem because of their small bill size. Pietz et al. (1995) found this type of attachment was retained on some birds for over a year and, in these cases, could lead to feather wear and some abrasion. How- ever, they reported no incidences of the trans- mitter anchors catching the bills of Mallards {Anas platyrhynchos) or Gadwalls {A. stre- pera). Paquette et al. (1997) reported negative effects of anchor transmitters on reproduction and survival of Mallards and suggested this resulted from additional energetic costs asso- ciated with increased mass and drag, as well as increased preening. Overall there have been few reports of negative effects from this type of transmitter attachment and many research- ers have advocated their use, especially for shorter-term studies (Hepp et al. 2002, Fleskes 2003, Iverson et al. 2006). This attachment method is less surgically invasive than inter- scapular subcutaneous or intra-abdominal transmitters, and retention times are often ap- propriate for the battery life of the transmit- ters. We do not conclude that anchor attach- ments should be avoided when it is appropri- ate for the study design, but suggest that po- tential problems should be considered. Precautionary measures such as using a small- er anchor size, ensuring the curled end is tightly closed, and narrowing the loop end could prevent complications with using this radio attachment method. We advocate that re- searchers involved in radio telemetry studies make efforts to evaluate, and subsequently re- port in the primary literature, the effects of 602 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 3, September 2008 radio transmitters, given their widespread use and potential risks to individual animals and results of studies. ACKNOWLEDGMENTS Lunding was provided by BC Hydro Bridge Coastal Pish and Wildlife Restoration Program. The Centre for Wildlife Ecology at Simon Eraser University provided logistical support. Chester Alec and Sam Copeland were valuable field technicians, and Sam Iverson, Sun- ny LeBourdais, Ken Wright, Sean Boyd, and Bobby Vinnie assisted with captures. LITERATURE CITED Ackerman, J. T, J. Adams, J. Y. Takekawa, H. R. Carter, D. L. Whitworth, S. H. Newman, R. T. Golightly, and D. L. Orthmeyer. 2004. Effects of radiotransmitters on the reproductive perfor- mance of Cassin’s Auklets. Wildlife Society Bul- letin 32:1229-1241. Altman, R. B. 1981. General principles of avian sur- gery. Compendium on Continuing Education for the Practicing Veterinarian 3:177-183. Conway, C. J. and V. Garcia. 2005. Effects of radi- otransmitters on natal recruitment of Burrowing Owls. Journal of Wildlife Management 69:404- 408. Duriez, O., Y. Perrand, P. Binet, E. Corda, P. Goss- MANN, AND H. pRiTZ. 2005. Habitat selection of the Eurasian Woodcock in winter in relation to earthworm availability. Biological Conservation 122:479-490. Dwyer, T. J. 1972. An adjustable radio package for ducks. Bird-Banding 43:282-284. Esler, D., D. M. Mulcahy, and R. L. Jarvis. 2000. Testing assumptions for unbiased estimation of survival of radio-marked Harlequin Ducks. Jour- nal of Wildlife Management 64:591-598. Pleskes, j. P. 2003. Effects of backpack radiotags on female Northern Pintails wintering in California. Wildlife Society Bulletin 31:212-219. Giroux, J.-E, D. V. Bell, S. Percival, and R. W. Sum- mers. 1990. Tail-mounted radio transmitters for waterfowl. Journal of Pield Ornithology 61:303- 309. Guyn, K. L. and R. G. Clark. 1999. Decoy trap bias and effects of markers on reproduction of North- ern Pintails. Journal of Pield Ornithology 70:504- 513. Hepp, G. R., T. H. Polk, and K. M. Hartke. 2002. Effects of subcutaneous transmitters on reproduc- tion, incubation behavior, and annual return rates of female Wood Ducks. Wildlife Society Bulletin 30:1208-1214. Houston, R. A. and R. J. Greenwood. 1993. Effects of radio transmitters on nesting captive Mallards. Journal of Wildlife Management 57:703-709. Hupp, J. W., J. M. Pearce, D. M. Mulcahy, and D. A. Miller. 2006. Effects of abdominally implant- ed radiotransmitters with percutaneous antennas on migration, reproduction, and survival of Can- ada Geese. Journal of Wildlife Management 70: 812-822. Iverson, S. A., W. S. Boyd, D. Esler, D. M. Mul- cahy, AND T. D. Bowman. 2006. Comparison of the effects and performance of four types of ra- diotransmitters for use with scoters. Wildlife So- ciety Bulletin 34:656-663. Keedwell, R. 2001. Evaluation of radio transmitters for measuring chick mortality in the Banded Dot- terel. Waterbirds 24:217-223. Korschgen, C. E., K. P. Kenow, A. Gendron-Pitz- PATRiCK, W. L. Green, and P. J. Dein. 1996a. Im- planting intra-abdominal radio transmitters with external whip antennas in ducks. Journal of Wild- life Management 60:132-137. Korschgen, C. E., K. P. Kenow, W. L. Green, M. D. Samuel, and L. Sileo. 1996b. Technique for im- planting radio transmitters subcutaneously in day- old ducklings. Journal of Pield Ornithology 67: 392-397. Mauser, D. M. and R. L. Jarvis. 1991. Attaching ra- dio transmitters to 1 -day-old Mallard ducklings. Journal of Wildlife Management 55:488-491. Paquette, G. A., J. H. Devries, R. B. Emery, D. W. Howerter, B. L. Joynt, and T. P. Sankowski. 1997. Effects of transmitters on reproduction and survival of wild Mallards. Journal of Wildlife Management 61:953-961. Perry, M. C., G. H. Haas, and J. W. Carpenter. 1981. Radio transmitters for Mourning Doves: a comparison of attachment techniques. Journal of Wildlife Management 45:524-527. PiETZ, P. J., D. A. Brandt, G. L. Krapu, and D. A. Buhl. 1995. Modified transmitter attachment method for adult ducks. Journal of Field Orni- thology 66:408-417. Wheeler, W. E. 1991. Suture and glue attachment of radio transmitters on ducks. Journal of Field Or- nithology 62:271-278. SHORT COMMUNICATIONS 603 The Wilson Journal of Ornithology 120(3):603-606, 2008 Estimate of Trichomonas gallinae-mducQ.d Mortality in Band-tailed Pigeons, Upper Carmel Valley, California, Winter 2006-2007 Mark R. Stromberg,* Walter D. Koenig,^ Eric L. Walters,'’^ and John Schweisinger- ABSTRACT — Band-tailed Pigeons {Patagioenas fasciata) wintering at Hastings Reservation in central coastal California during winter 2006-2007 died in large numbers between January and March 2007. Lab- oratory analysis of carcasses indicated that Tricho- monas gallinae was responsible for the die-off. During the height of the die-off, a survey of 2.5 km of suitable riparian habitat resulted in 373 pigeon carcasses being found. Based on a subsample of carcasses, mean turn- over rate was 2.8 days with a 95% confidence interval of 2-10 days. Extrapolating to suitable habitat over the 52.7-km^ study area resulted in a conservative estimate of 43,059 dead pigeons, assuming a conservative car- cass turnover rate of 10 days. This estimate of mor- tality is nearly three times the largest trichomoniasis mortality event previously recorded for Band-tailed Pi- geons and at least twice the number harvested annually in the United States. Local mortality of pigeons in Monterey County, California may have been several times this estimate based on the presence of consid- erable similar habitat in the nearby Ventana Wilder- ness. Received 27 July 2007. Accepted 14 December 2007. Band-tailed Pigeons {Patagioenas fasciata) range from Alaska into South America and are seasonally migratory on the Pacific Coast (Keppie and Braun 2000). Most large flocks wintering in central coastal California depart in late April, arriving in British Columbia in May (Keppie and Braun 2000). At Hastings Reservation in the upper Carmel Valley, Mon- terey County in central coastal California, pi- geons are most numerous in winter and fluc- tuate among years (Davis et al. 1980). From October 2006 to March 2007 large flocks were present feeding largely on coast live oaks {Quercus agrifolia) acorns, many of which ' Ha.stings Natural History Re.servation and Museum of Vertebrate Zoology, University of California, Berke- ley, 38601 East Carmel Valley Road, Carmel Valley, CA 93924, USA. “Turf Images Inc., 177 Webster Street, Monterey. CA 93940, USA. Corresponding author; e-mail: ewalters@berkeley.edu were still present on a small number of trees throughout the winter. Although a few dead birds are found in most years when pigeons are abundant, unusually high mortality was noted starting in January 2007 with many chronically sick birds observed that were un- able to fly and dying. Band-tailed Pigeons in California and else- where are known to have significant mortality due to infections of Trichomonas gallinae (Stabler and Braun 1975, 1979; Cole 1999). Our objectives in this paper are to report on the causes of the observed pigeon mortality and present an estimate of the number of pi- geons involved. METHODS We documented mortality of Band-tailed Pigeons in the vicinity (36°23'N, 121° 33' W) of Hastings Reservation in upper Carmel Valley, California where elevation ranges from 310 to 1,100 m. Pigeons moved widely during the day over a larger area, but roosting at night was usually along riparian corridors. Dead and dying birds were regularly found in these riparian corridors from early January to late March 2007. Estimating the number of dead pigeons re- quired several steps. First, three observers walked a 2.5-km census transect on Hastings Reservation along Robertson Creek on 7 Feb- ruary 2007 searching the ground on both sides of the creek for fresh pigeon carcasses. Patch- es of fresh, dry feathers from birds recently depredated, dead birds, and still alive but non- volant individuals were mapped using hand- held GPS units and coordinates were taken for each bird. These data were transformed to dis- tances from a central axis for use with the program DISTANCE® (J'homas et al. 2006) to calculate sample estimates and the effective strip width of the sampling tran.sect. .Second, we estimated the carcass decom- position (turnover) rate by surveying a 1.4-km 604 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 Stretch of dirt road paralleling Robertson Creek at the Hastings Reservation over 18 days (8-26 Feb 2007). Thirteen marked birds, visible from the road, were checked each day until a carcass was removed by scavengers or the feather pile became indiscernible because it was matted and reduced by rains. We cal- culated the mean decomposition time (time until they would no longer be counted as pre- sent) of these 13 birds using Kaplan-Meier survival analysis (SPSS 2000) and used the upper bound of the 95% confidence interval of mean survival time to calculate a conser- vative estimate of carcass turnover time. Third, an estimate of the total riparian areas used by pigeons for roosting at night and at a similar elevation to the Robertson Creek site examined in the first step was obtained using CIS layers in ArcGIS (ESRI 2007). All con- tiguous stream courses fitting our criteria within the larger study area, supported with observations of dead or dying pigeons in their vicinity, were identified and the area along them was buffered with a 56.4-m strip (based on the effective strip width from program DISTANCE®, as detailed below) to calculate area. We did not include the high, dry, south- facing streambeds of the watershed as these areas are dominated by relatively short shrubs, lack tall oak trees, and are rarely used by Band-tailed Pigeons. Informal telephone in- terviews with land managers of the nearby U.S. Forest Service and private lands con- firmed the presence of many dead and dying Band-tailed Pigeons within the study area. We conservatively estimated the period during which mortality occurred to be 60 days (15 Jan to 15 Mar). The total number of Band- tailed Pigeons carcasses occurring during this time period within the study area was esti- mated using the formula below. Total pigeon carcasses = D-RA-(1/MST)-P, where D = the density of pigeons (carcasses/ ha) in the Robertson Creek census transect, RA = the total estimated riparian area within the study area, 1/MST is the turnover rate of dead carcasses (the inverse of the upper bound of the 95% confidence interval of mean sur- vival time [MST]), and P = the length of the period during which pigeon mortality was sig- nificant. The California Animal Health and Food Safety Laboratory System (CAHES) and R. W. Gerhold, Wildlife Disease Section, Uni- versity of Georgia, Athens, Georgia, evaluated the dead pigeons; the latter used an Inpouch TriTrichomonas Test pouch kit (Cover et al. 1994). Samples were collected at several times during the study and pigeons consis- tently showed diagnostic symptoms of infec- tion by T. gallinae (caseous, obstructive le- sions within the upper areas of the digestive tract; Cole 1999). Ereshly dead pigeon car- casses (25) were collected, frozen, and depos- ited in the Museum of Vertebrate Zoology, University of California, Berkeley. RESULTS All Band-tailed Pigeons found dead or dy- ing had clinical signs of trichomoniasis in- cluding low body weight, listlessness, caseous or cheesy, yellowish lesions in the mouth, throat, and around the beak, and tendency to fall over when forced to move (Stabler and Braun 1975, 1979; Cover et al. 1994). Labo- ratory test results were positive for Tricho- monas gallinae and negative for West Nile vi- rus (R. W. Gerhold and K. D. Hanni, pers. comm.). We counted 373 dead pigeons over the 2.5- km transect surveyed along Robertson Creek on 7 February 2007. Based on the GPS co- ordinates of the birds, DISTANCE® calculated an effective (one-half) strip width of 28.2 m (fx in Buckland et al. 1993). We therefore used 56.4 m as the ArcGIS buffer function to cal- culate the area of the survey transect along the winding riparian corridor, producing an esti- mated density of dead birds of 26.45 birds/ha. We estimated 271.3 ha of comparable riparian habitat within the larger 52.7-km2 ^rea. Assuming densities of dead birds comparable to that found on the transect conducted at Has- tings, this yields a total of 7,177 pigeon car- casses detectable at any one point in time along the riparian corridors of the study area. We observed relatively high densities of feral pigs (Sus scrofa), coyotes {Canis latrans), and raptors including Bam Owls {Tyto alba). Red- tailed Hawks (Buteo jamaicensis), and Cooper’s Hawks (Accipter cooperii) in the census tran- sect. Unsurprisingly, this resulted in relatively rapid estimated carcass turnover times. The mean time to carcass decomposition was 2.8 days with a 95% confidence interval of 2-10 SHORT COMMUNICATIONS 605 days based on Kaplan-Meier survival analysis of 1 3 marked carcasses checked over an 1 8-day period in February 2007. We assume the death rate was constant over the peak of the mortality event. The number of dead individuals was conservatively esti- mated to have turned over at least six times yielding a total estimate of 43,059 dead pi- geons within the total study area during the 2-month period of peak mortality. DISCUSSION Mortality attributable to T. gallinae during winter 2006-2007 was high relative to other major causes of mortality of Band-tailed Pi- geons. The estimated numbers of this species harvested over their entire range in the United States for 1999, 2000, 2004, and 2005 aver- aged 20,550 per year (USDI 2006a, b). Epi- zootic trichomoniasis in Band-tailed Pigeons has been reported previously (Cole 1999) with the largest prior estimate of mortality being 16,000 pigeons. A series of reports of tricho- moniasis in Band-tailed Pigeons from north- ern California (USGS 1995, 2004, 2006) es- timated 2,000, 2,000, and 300 deaths, respec- ; tively. Our survival analysis assumes that mortal- “ ity was constant over the period considered. It is possible, however, that we missed the pe- riod of peak mortality, as we did not imme- diately recognize the severity and extent of the j die-off, and were spurred to conduct quanti- tative sampling some time after peak mortality may have occurred. It is also likely the mortality extended much ^ farther than we were able to conhrm. U.S. Forest Service employees were not available ! to conduct searches for pigeons along the ri- parian areas in the nearby Ventana Wilder- ness, a large (~3,5()0 km^) mostly inaccessi- I ble area that is similar to the habitat at Has- I tings Reservation. Mortality may also have I extended well into the upper Carmel and Ar- royo Seco watersheds of the Ventana Wilder- j ness where large flocks of Band-tailed Pi- I geons are common in winter (Roberson and Tenny 2002). Thus, the total number of birds affected by the event within Monterey County ; may have been several times our estimate of ! -43,000 birds. I The mortality event reported here is far in j excess of what has been previously observed and is far more signiheant than hunting mor- tality over the species’ North American range. Despite the die-off, flocks of pigeons re- mained common through late March within the study area. ACKNOWLEDGMENTS We thank B. J. Piculell and R. D. Drobek for assis- tance with surveying and R. W. Gerhold for confirming trichomoniasis. Testing was arranged by J. D. Cann (California Department of Fish and Game) and K. D. Hanni (Monterey County Health Department). ELW was supported by NSF grant IOB-05 16851 to WDK. LITERATURE CITED Buckland, S. T, D. R. Anderson, K. R Burnham, AND J. L. Laake. 1993. Distance sampling: esti- mating abundance of biological populations. Chapman and Hall, London, United Kingdom. Cole, R. A. 1999. Trichomoniasis. Pages 201-206 in Field manual of wildlife diseases: general field procedures and diseases of birds (M. Friend and J. C. Franson, Editors). USDI, Geological Survey- BRD, Washington, D.C., USA. Cover, A. J., W. M. Harmon, and M. W. Thomas. 1994. A new method for the diagnosis of Trich- omonas gallinae infection by culture. Journal of Wildlife Diseases 30:457-459. Davis, J., W. D. Koenig, and P. L. Williams. 1980. Birds of Hastings Reservation, Monterey County, California. Western Birds 11:113-128. ESRI, Inc. 2007. ArcGIS. Version 9.2.2. Earth Sys- tems Research Institute, Inc., Redlands, Califor- nia, USA. Keppie, D. M. and C. E. Braun. 2000. Band-tailed Pigeon (Columha fasciata). The birds of North America. Number 530. Roberson, D. and C. Tenny. 2002. Monterey Birds. Second Edition. Monterey Peninsula Audubon So- ciety, Monterey, California, USA. SPSS Institute, Inc. 2000. SPSS for Windows. Ver- sion 10.0. SPSS Institute, Inc., Chicago, Illinois, USA. Stabler, R. M. and C. E. Braun. 1975. Effect of vir- ulent Trichomonas gallinae on the Band-tailed Pi- geon. Journal of Wildlife Diseases 1 1:482-483. Stabler, R. M. and C. E. Braun. 1979. Effects of a California derived strain of Trichomonas gallinae on Colorado USA Band-tailed Pigeons. California Fish and Game 65:56-58. Thomas, L., J. L. Laaki;, S. Strindberg, H E C. Marques, S. T. Buckland, D. L. Borchers, D. R. Anderson, K. P. Burnham, S. L. Hedlf:y, J. H. F’oli.ard, j. R. B. Bishop, and T. A. Marqiie:s. 2006. DISTANCE 5.0. Release 2.0. University of St. Andrews, United Kingdom. http://www. ruwpa.st-and.ac.uk/distance (accessed 15 June 2007). U.S. Di;pARrME:NT oi the: Interior (USDI). 2006a. Mi- 606 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 gratory bird hunting activity and harvest during the 1999 and 2000 hunting seasons — Final Re- port. USDI, Fish and Wildlife Service, Washing- ton, D.C., USA. U.S. Department oe the Interior (USDI). 2006b. Mi- gratory bird hunting activity and harvest during the 2004 and 2005 hunting seasons: preliminary estimates. USDI, Washington, D.C., USA. U.S. Geological Survey (USGS). 1995. Quarterly Wildlife Mortality Report April 1995 to June 1995. USGS, National Wildlife Health Center, Madison, Wisconsin, USA. U.S. Geological Survey (USGS). 2004. Quarterly Wildlife Mortality Report April 2004 to June 2004. USGS, National Wildlife Health Center, Madison, Wisconsin, USA. U.S. Geological Survey (USGS). 2006. Quarterly Wildlife Mortality Report April 2006 to June 2006. USGS, National Wildlife Health Center, Madison, Wisconsin, USA. The Wilson Journal of Ornithology 120(3):606-610, 2008 Winter Ecology of Yellow Rails Based on South Carolina Specimens William Post^ ABSTRACT— Arthur T. Wayne collected 58 Yel- low Rails (Coturnicops noveboracensis) during seven winters between 1903 and 1918 at one locality on the Atlantic coast in Charleston County, South Carolina. The collection represents the largest known series of Yellow Rails from a single wintering site and provides information about the winter ecology of this species. There was no evidence that Yellow Rail numbers var- ied between winters. The sex ratio was significantly biased toward females suggesting the occurrence of differential wintering. Yellow Rails were collected mainly in wet (freshwater) fields with short dense grass, the same features of Yellow Rail habitats in coastal Texas. Yellow Rails were consistently located in the same habitats as LeConte’s Sparrow (Ammodra- mus leconteii). Two other grassland species, Henslow’s Sparrows (A. henslowii) and Sedge Wrens {Cistothorus palustris), had habitat occupancy patterns significantly different from that of Yellow Rails. Received 6 June 2007. Accepted 21 November 2007. The Yellow Rail {Coturnicops novebora- censis) has been studied in breeding areas (Peabody 1922, Terrill 1943, Stalheim 1974, Anderson 1977, Bookhout and Stenzel 1987, Gibbs et al. 1991, Robert and Laporte 1999, Popper and Stem 2000, Robert et al. 2000) and in captivity (Stalheim 1975). Historically, migratory populations of the Yellow Rail oc- cupied a discontinuous breeding range from the Northwestern Territories to New Bruns- wick, Canada, south to the latitudes of Con- necticut and Oregon, USA (Bookhout 1995), ' Charleston Museum, 360 Meeting Street, Charles- ton, SC 29403, USA; e-mail: grackler@aol.com and wintered primarily on the coastal plain from North Carolina to Texas. Most infor- mation about the species in winter is from studies conducted on the Gulf coast of Texas (Grace et al. 2005), and little is known about Yellow Rails overwintering on the Atlantic coast. A series of 58 specimens collected at one locality in South Carolina between 1903 and 1918 provides information about patterns of occurrence, habitat use, and sex ratios in winter. The objective of this paper is to com- pare these data with that available from other areas of its wintering range. METHODS This paper is based on the field work of Arthur T. Wayne who, in 1903-1913, collect- ed 56 Yellow Rails in Charleston County, South Carolina. I examined 36 of the skins and confirmed the information on their labels accurately reflected entries in his specimen ledgers, which are archived at the Charleston Museum. Additional data were contained in Wayne’s correspondence and published pa- pers. These sources provided the gender, col- lection date, and collection location of each specimen. The skins are not accompanied by habitat information, but Wayne recorded the name of the field where each bird was col- lected, and his letters and articles contained descriptions of the fields. I assume that each of Wayne’s collecting trips involved about the same amount of time. He usually worked at one location each day and took 3-4 speci- SHORT COMMUNICATIONS 607 mens. The primary or secondary sex ratios of Yellow Rail are unknown. The species is be- lieved to be monogamous (Bookhout 1995) and I assume the sex ratio is approximately * equal at the end of the breeding season RESULTS Wayne first discovered Yellow Rails win- tering in South Carolina on 4 February 1898, when he obtained a cat (Felis catus)-captmQd bird, too mangled to preserve (Wayne 1910). The next Yellow Rail that Wayne encountered, also captured by a cat (19 Dec 1903), was a male (Wayne 1905; specimen #ATW 4529; Museum of Comparative Zoology #309649). Wayne then began searching appropriate hab- itats using hunting dogs (pointers) and from ’ 1903 to 1913, collected 56 more Yellow Rails, all within 10 km of his house at Porcher’s Bluff, Mt. Pleasant, Charleston County, South Carolina (32° 5 1 ' 22" N, 79° 46' 49" W). Wayne collected and reported the gender of 56 Yellow Rails during seven winters in I 1903-1913 (Table 1). Of these, 77% were fe- males, representing a significant departure from a balanced sex ratio (y^ = 8.5, df == 1, j P < 0.005). His peak period was winter 1909-1910, when he collected 35 Yellow Rails, 83% of which were females, also rep- resentative of an unbalanced winter sex ratio (X^ = 8.5, df = 1, P < 0.005). Sex ratios are based on birds that were removed from the same population. Forty-one of the 50 trips that Wayne made in 1909-1910 were to one field. Too few birds were collected in other years to examine the significance of any departure from a 1:1 ratio (Table 1). Wayne noted that four females he collected from 1 1 November , to 10 December were molting. The notes and papers of Wayne provide an I idea of the habitats used by Yellow Rails. A main collecting site was “a low wet place of open land with a dense growth of short dead I grass” (Wayne 1905:396-397). Wayne collected ' three other species in the same habitats as the ' rails: LeConte’s Sparrows {Ammodrcumis lecon- I teii). Sedge Wrens (Cistothonis platens is), and Henslow’s Sparrows (A. henslowii). Based on Wayne’s notes, most Yellow Rails (32/41; 78%) occurred in the wettest of three sites (Table 2), as did LeConte’s Sparrows (35/ ' 44; 80%). Henslow’s Sparrows were collected most often in the driest of the fields (12/17; •* Number of birds/trip. ^Significantly more t'cmalc^ than males were collected {X~ = 8.7, df = \, F < ().0()5). • .Number of different sites over total period. "'Significantly more females than males were collected (X- = 8.5, df = I, /^ < ().0()5). 608 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 TABLE 2. Habitats used by wintering Yellow Rails and other species at Mt. Pleasant, South Carolina. Species Porcher’s Rough Ground^ Porcher’s Myrtle 1909-1913 (57)^* Pasture'^ 1906-1918 (36) Cat island'-' 1903-1911 (27) Yellow Rail LeConte’s Sparrow Henslow’s Sparrrow Sedge Wren a Wet: two-layered vegetation (a dense layer of short graminoids and forbs interspersed with broom sedge). Moist to wet: three-layered vegetation (short grasses and sedges, broom sedge and wax myrtle). Moist to dry: primarily one-layered (short grasses and sedges and scattered clumps of broom sedge). Number of trips. 32 2 7 35 5 4 5 0 12 10 34 4 71%), where few Yellow Rails were collecteci (7/41; 17%). Most Sedge Wrens (34/48; 71%) were collected in sites with a mid-range of moisture, and with greater cover of shrubs. Only 5% (2/41) of the Yellow Rails occurred in habitat with shrubs. LeConte’s Sparrows were collected on the same days at the same sites in each of the seven winters in which Yellow Rails occurred. Rails co-occurred with Henslow’s Sparrows and Sedge Wrens in four and three winters, respectively. The habitat use pattern of Yellow Rails and LeConte’s Sparrows did not differ from that expected un- der a null hypothesis (G = 11.8, df = 6, P > 0.05). However, I reject the null hypothesis that Yellow Rails had the same pattern of hab- itat use as Henslow’s Sparrows (G = 20.8, df = 6, P < 0.001) and Sedge Wrens (G = 13.0, df = 6, P < 0.05). DISCUSSION Loomis (1891) collected four Yellow Rails (1 male, 1 female, 2 unknown gender) in 1887 in the Piedmont of South Carolina (Chester County), but Wayne (1905) did not find the species on the coast until 1898. Wayne col- lected Yellow Rails in seven winters over 11 years (1903-1913). Two apparent peaks oc- curred; one in winter 1909-1910 (35 Yellow Rails), and another in 1903-1904 (7). There were few differences between years when considering numbers of rails collected per trip (Table 1). For example, in winter 1909-1910, when Wayne collected 35 rails in 50 trips, the success rate (0.70 rails/trip) was not much greater than that obtained during 1907-1908, when only five Yellow Rails were collected in eight trips (0.63/trip; Table 1). Besides un- equal effort between years, Wayne concen- trated his work in fields where he frequently encountered Yellow Rails, curtailing visits to other sites. For example, in 1909-1910, 41 of 50 trips were to one field (Porcher’s Rough Ground). Unequal effort between years and localities is also a source of bias when con- sidering annual variation in LeConte’s Spar- row populations. McNair and Post (2000) hy- pothesized that numbers of the latter species varied significantly between years, and they considered winters such as 1909-1910, when Wayne collected 38 LeConte’s Sparrows, to be “incursion” years. Wayne’s sampling was not random. If he obtained an unusual specimen, he intensified his search effort at the collection site. As in the case of Yellow Rails, when numbers collected are considered with effort, there was little difference between years. Wayne obtained four LeConte’s Sparrows in winter 1904-1905 (0.67/trip), almost the same rate as in 1909-1910 when he collected 35 (0.70/trip; Table 1). Annual variation in the numbers of Yellow Rails and LeConte’s Spar- rows can be most parsimoniously explained by uneven collecting effort between years and among collection sites. Few workers on the Atlantic coast have mentioned the habitats in which they found wintering Yellow Rails, but the sites for which descriptions are available appear to be similar to fields visited by Wayne. For example, the Beaufort-Morehead (North Carolina) airport, where >28 Yellow Rails were counted in 1 day during a controlled bum (Chapman 1969), was covered primarily by shorter vegetation such as panicgrass {Panicum spp.), inter- spersed with clumps of broom sedge {Andro- pogon spp.). Anderson (1977) suggested that Yellow Rails prefer drier portions of Spartina marshes (salinity not indicated) in winter. Wayne encountered Yellow Rails only in freshwater or low salinity sites, although he made numerous trips to salt marshes, as dem- SHORT COMMUNICATIONS 609 onstrated by the 381 Seaside Sparrows {Am- I modramus maritimus) he collected. Yellow Rails and LeConte’s Sparrows con- I sistently occurred in the same habitats, old- I fields dominated by short vegetation such as panicgrass and by rushes {Juncus spp.). These fields appeared to have a secondary vegetative layer composed of broom sedge, which prob- ably was more frequent at higher elevations. Some sites also had a third layer, composed of shrubs such as wax myrtle {Myrica ceri- fera) and groundsel {Baccharis halimifolia). Landowners periodically cattle-grazed and burned the fields, and Wayne’s notes imply that shrubs were low and scattered. Yellow Rails and LeConte’s Sparrows both occurred most often in wetter, more open fields (Table 2). Henslow’s Sparrows and Sedge Wrens were most common in drier sites. Sedge Wrens were frequently collected in fields that j had shrubs, but few Yellow Rails or Ammo- i dramus sparrows were collected there (Table j 2). Based on Wayne’s studies, the winter hab- I itats of Yellow Rails in South Carolina cor- i respond to those in coastal Texas, where the most constant features of all Yellow Rail lo- j cations are dense vegetative cover and low ! plant height (Grace et al. 2005). I Wayne collected a significantly larger num- I ber of female than male Yellow Rails, a dis- i parity which suggests that males and females overwinter at different latitudes or use differ- ent habitats. Disproportionate numbers of fe- males of another rallid, the American Coot {Fidica americana), have been found winter- ing farther south than males (Eddleman 1983, ^ Brisbin and Mowbray 2002). Wayne’s work demonstrates sexual segregation in one region I and in different habitats. Limited data suggest j sex ratios favor male Yellow Rails in areas I north of South Carolina. In a series of 17 win- , ter specimens collected north of Chesapeake j Bay (39° 32' 55" N), the sex ratio was 9 j males:2 females; the comparable ratio for I birds collected south of Chesapeake Bay (ex- I eluding South Carolina) was 3 males:3 fe- I males. These data are too few to examine sta- I tistically, but they suggest that in comparison ! to sex ratios found by Wayne, male Yellow Rails do not migrate as far south as females. I If differential wintering is widespread, it could j have implications for management practices. Female-biased dispersal may lead to male-bi- ased sex ratios in breeding areas, increasing vulnerability to extinction of small or isolated populations, or populations at the periphery of a species’ range (Dale 2001). This suggests that one objective of future research might be to examine how Yellow Rail sex ratios are af- fected by prescribed fires and grazing in dif- ferent overwintering areas (Mizell 1998, Grace et al. 2005, Baldwin et al. 2007). ACKNOWLEDGMENTS The following museums provided access to their collections: American Museum of Natural History, British Museum, Charleston Museum, Field Museum, Museum of Comparative Zoology (Harvard), Phila- delphia Academy of Science, and Smithsonian Insti- tution. The American Museum of Natural History and Museum of Comparative Zoology kindly provided copies of Wayne’s letters. This paper benefited from discussions with D. B. McNair. D. R. C. Prescott and an anonymous reviewer provided useful comments en- abling me to improve the manuscript. LITERATURE CITED Anderson, J. M. 1977. Yellow Rail (Coturnicops nov- eboracensis). Pages 66-70 in Management of mi- gratory shore and upland game birds in North America (G. C. Sanderson, Editor). International Association of Fish and Wildlife Agencies, Wash- ington, D.C., USA. Baldwin, H. Q., J. B. Grace, W. C. Barrow Jr., and F. C. Rowher. 2007. Habitat relationships of birds overwintering in a managed coastal prairie. Wil- son Journal of Ornithology 119:189-197. Bookhout, T. a. 1995. Yellow Rail {Coturnicops nov- eboracensis). The birds of North America. Num- ber 139. Bookhout, T. A. and J. R. Stenzel. 1987. Habitat and movements of breeding Yellow Rails. Wilson Bul- letin 99:441-447. Brisbin Jr., I. L. and T. B. Mowbray. 2002. American Coot (Fulicci americana). The birds of North America. Number 697. Chapman, F. L. 1969. Yellow Rails at Beaufort, N.C. Chat 33:103. Dale, S. 2001. Female-biased dispersal, low female recruitment, unpaired males, and the extinction of small and isolated bird populations. Oikos 92: 344-356. Eddleman, W. R. 1983. A study of migratory Ameri- can Coots, Fulica americana. Dissertation. Oklahoma State University, Stillwater, USA. Gibbs, J. P, W. G. Shriver, and S. M. Melvin. 1991. Spring and summer records of the Yellow Rail in Maine. Journal of Field Ornithology 62:509-516. Graci;, j. B.. L. K. Allain, H. Q. Baldwin, A. G. Bil- L(X'K, W. R. Eddli:man, A. M. Givi:n, C. W. JiiSKi;, and R. M. Moss. 2(X)5. Effects of prescribed fire in the coastal prairies of Texas. USDI, Geological Sur- 610 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 vey and Fish and Wildlife Service, Region 2. Open- File Report 2005-1287. htpp://www.nwrc.usgs.gov/ facshts/2005- 1 287-fire-in-coastal-texas-repoit.pdf. Loomis, L. M. 1891. A further review of the avian fauna of Chester County, South Carolina. Auk 8: 49-59,167-173. McNair, D. B. and W. Post. 2000. Historical winter sta- tus of three upland Ammodramiis sparrows in South Carolina. Studies in Avian Biology 21:32-38. Mizell, K. L. 1998. Effects of fire and grazing on Yellow Rail habitat in a Texas coastal marsh. Dis- sertation. Texas A&M University, College Station, USA. Peabody, P. B. 1922. Haunts and breeding habits of the Yellow Rail Cotiirnicops noveboracensis. Journal of the Museum of Comparative Oology 2: 33-44. Popper, K. J. and M. A. Stern. 2000. Nesting ecology of Yellow Rails in southcentral Oregon. Journal of Field Ornithology 71:460-466. The Wilson Journal of Ornithology 120(3):610— 613, 2008 Robert, M. and P. Laporte. 1999. Numbers and movements of Yellow Rails along the St. Lawr- ence River. Condor 101:667—671. Robert, M., P. Laporte, and R. Benoit. 2000. Sum- mer habitat of Yellow Rails, Cotiirnicops nove- boracensis, along the St. Lawrence River, Quebec. Canadian Field-Naturalist 114:628-635. Stalheim, P. S. 1974. Behavior and ecology of the Yellow Rail {Cotiirnicops noveboracensis). The- sis. University of Minnesota, Minneapolis, USA. Stalheim, P. S. 1975. Breeding and behavior of captive Yellow Rails. Aviculture Magazine 81:133-141. Terrill, L. M. 1943. Nesting habits of the Yellow Rail in Gaspe County, Quebec. Auk 60:171-180. Wayne, A. T. 1905. Notes on certain birds taken or seen near Charleston, South Carolina. Auk 22: 395-400. Wayne, A. T. 1910. Birds of South Carolina. Contri- butions from the Charleston Museum I. Daggett Printing Company, Charleston, South Carolina, USA. Nest Raising by Red-crowned Cranes in Response to Human-mediated Flooding at Zhalong Nature Reserve, China Qiang Wang*’^’^ and Feng ABSTRACT — Water was released from a reservoir in May 2005 to restore the wetland of international importance at Zhalong, China. As a result, the water level on the floodplain rose rapidly. A pair of Red- crowned Cranes {Grus japonensis) was observed rais- ing their nest to avoid submersion and loss of eggs. The nest and two eggs were tended by both adults, but the eggs did not hatch. Hatching failure may have re- sulted from low egg temperatures due to addition of wet and new nesting materials, and reduction of in- cubation time during the flooding event. Reservoir dis- charge should avoid the breeding period of waterbirds and discharge rates should be reduced. Received 29 May 2007. Accepted 14 November 2007. Water level is an important factor affecting habitat selection by waterfowl (Kingsford and Johnson 1998, Timoney 1999, Adamo et al. 2004, Bradter et al. 2005), and can have both ' Northeast Institute of Geography and Agricultural Ecology, CAS, Changchun, 130012, China. 2 Graduate School, CAS, Beijing, 1300039, China. ^ Department of Zoology, Northeast Forestry Uni- versity, Harbin, 150040, China. Corresponding author; e-mail: lifeng_1956@ 126.com positive and negative effects on breeding and survival of waterbirds (Poiani 2006). Floods can increase food resources for some water- birds and subsequently increase their chances of successful breeding (Nielsen and Gates 2007). However, floods can also lead to nest- ing failure (Borad et al. 2002, Mukherjee et al. 2002, Sanders and Maloney 2002, Gilbert and Servello 2005). The Red-crowned Crane {Grus japonensis) is classified as a globally endangered species with a total population of —2,000 birds (Meine and Archibald 1996). There is little informa- tion regarding the impact of floods on breed- ing success of this species but, floods were a major cause for breeding failure of the Indian Sams Crane {Grus antigone antigone) with 12.2% of egg losses attributed to high water (Borad et al. 2002). Mukherjee et al. (2002) reported that a pair of Indian Sams Cranes attempted to raise the nest platform by adding nest materials in response to flooding. This nest eventually became submerged and the eggs were lost. Red-crowned Cranes nest within Zhalong SHORT COMMUNICATIONS 611 FIG. 1. Red-crowned Crane nest before flooding, Zhalong Nature Reserve, China. Nature Reserve in China, a Ramsar wetland area of international importance. Increased de- mand for irrigation water for agriculture up- river from Zhalong Nature Reserves can lead to lack of water in the wetland, endangering local waterbirds (Lemly et al. 2000). The local government in 2005 decided to discharge wa- ter from a reservoir to recharge the wetland prior to the main period of irrigation. There is little recorded information regarding the im- pact of flooding on nesting behavior of Red- crowned Cranes. The objective of this paper is to report on observations of a pair of cranes raising their nest when flooding occurred. OBSERVATIONS We observed a pair of Red-crowned Cranes raising their nest as flood waters approached in the core zone of Zhalong Nature Reserve (123° 47'-124° 37' E, 46° 52'-47° 32' N) in western Heilongjiang Province of northeastern China in May 2005. We discovered and mea- sured the nest on 9 May 2005 when the nest height was 23 cm with an inner diameter of 46 cm and an outer diameter of 130 cm; there were 2 eggs (Fig. 1 ). The nest was in a reed (Phragmites australis) floodplain marsh; the marsh vegetation consisted of reed, cattail (Typha angustifolia), and common bulrush (Scirpus planiculmis). The vegetation in the vicinity had been harvested the previous win- ter, leaving only stubble —20 cm high (Fig. 1). The nest was built of dry, long reed stems on stubble in 15 cm of water. There was a shallow water pond 6 m northwest of the nest; the water depth in the pond was —35 cm and its surface area was — 1 2 iiF. The nest was FIG. 2. Red-crowned Crane nest during flooding, Zhalong Nature Reserve, China. — 80 m from a river, where fishermen in row boats frequently passed. The water level at the nest site increased rapidly beginning on 10 May as water from an up stream reservoir was released. We ob- served the Red-crowned Crane pair from an elevated location 1.2 km from the nest using a monocular telescope. The cranes continu- ously submerged their heads to collect below water material, flicking the material to elimi- nate water and mud before using it to raise the nest. We revisited the site at 1630 hrs and measured the raised nest; the nest height was 62 cm with an inner diameter of 41 cm and an outer diameter of 66 cm. The water depth adjacent to the nest was —56 cm. We moved some dry reeds to the nest site, before resum- ing observation on 1 1 May. The pair contin- ued to add material to the nest. We ap- proached the nest at 0740 hrs and found the cranes had not used the dry, long reeds which we had placed there earlier. The new nest ma- terials were primarily wet cattail stubble (with attached roots), and short stems of reed and common bulrush. We again measured the nest; the nest height had increased to 71 cm with an inner diameter of 44 cm and an outer di- ameter of 79 cm (Fig. 2). The water depth next to the nest was —63 cm, and the inner and outer diameters were larger tlian when measured on 10 May (Fig. 3). We checked the nest on 22 May. 8 June, and 15 June, but the two eggs did not hatcli. Fiight- to 1 0-day-old dead embryos were found in the eggs on 15 June. 612 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 3, September 2008 9 May 2005 10 May 2005 11 May 2005 LIG. 3. Nest measurements (cm) by date, Zhalong Nature Reserve, China. DISCUSSION The average air temperature in Zhalong Na- ture Reserve was 14.4° C and average water temperature was 1 1.6° C in May from 1981 to 1990 (Wu 1999). We hypothesize that internal egg temperatures were too low, especially at night, as a result of the added wet nest ma- terials, and that incubation time was shortened during flooding, resulting in embryo death. Selection of nesting materials changed during the flooding. Red-crowned Cranes typically choose dry, long reed stems to build their nests (Wu 1999). We do not know why this pair of Red-crowned Cranes chose wet vege- tation growing next to the nest instead of dry vegetation placed there by humans. It is pos- sible that water levels could have affected the birds’ choice of nesting materials as reported by Kuyt (1995) for Whooping Cranes {Grus americana). Nesting failure by the pair of Red-crowned Cranes observed following human-mediated flooding suggests better water management practices are needed at Zhalong Nature Re- serve. Discharging large amounts of water at a high rate during incubation or when chicks are small can drastically affect hatching suc- cess and chick survival. Releasing excess wa- ter in spring prior to egg-laying allows Red- crowned Cranes and other birds time to adjust to higher water levels, and should still provide protection from predators (Pieman 1988, Pie- man et al. 1993). Similar results would likely also occur if water was released at a slower rate. However the amount of water needed for agriculture is large in spring. Rainfall is plen- tiful in Zhalong Nature Reserve in autumn (Wu 1999) and there is little need for water for agriculture. Releasing water following the nesting season when chicks are larger is likely to reduce negative impacts, and mitigate the conflict for water between wetland restoration and agriculture. The objective of the water re- lease was to re-flood the wetland following use for agriculture. However, the needs of wa- terbirds and other wildlife using the wetland should be considered regarding the amount and speed of water discharge. ACKNOWLEDGMENTS Our research was supported by National Natural Science Loundation of China (Warrant No. 30370221) and the International Crane Loundation. We thank L S. Li for helpful comments. We also thank Y. M. Zhang, Y. J. Cai, J. G. Lu, and Y. B. Ba for help in the investigation. We are grateful to Zhalong National Nature Reserve for permission and access to conduct this study. M. A. Hayes and C. D. Littlefield provided helpful comments which improved the clarity of the original manuscript. LITERATURE CITED Adamo, M. C., L. Puglisi, and N. E. Baldaccini. 2004. Lactors affecting Bittern Botaurus stellaris distribution in a Mediterranean wetland. Bird Conservation International 14:153-164. Borad, C. K., a. Mukherjee, S. B. Patel, and B. M. Parasharya. 2002. Breeding performance of In- dian Sams Crane Grus antigone antigone in the paddy crop agroecosystem. Biodiversity and Con- servation 11:795-805. Bradter, U., S. Gombobaatar, C. Uuganbayar, T. E. Grazia, and K. M. Exo. 2005. Reproductive per- formance and nest-site selection of White-naped Cranes Grus vipio in the Ulz River Valley, north- eastern Mongolia. Bird Conservation International 15:313-326. Gilbert, A. T. and L Servello. 2005. Water level dynamics in wetlands and nesting success of Black Terns in Maine. Waterbirds 28:181—187. Kingsford, R. T. and W. Johnson. 1998. Impact of water diversions on colonially-nesting waterbirds in the Macquarie marshes of arid Australia. Co- lonial Waterbirds 21:159-170. Kuyt, E. 1995. The nest and eggs of the Whooping Crane, Grus americana. Canadian Lield-Naturalist 109:1-5. Lemly, a. D., R. T. Kingsford, and J. R. Thompson. 2000. Irrigated agriculture and wildlife conserva- tion: conflict on a global scale. Environmental Management 25:485-512. Meine, C. D. and G. W. Archibald. 1996. The cranes: status survey and conservation action plan. lUCN, Gland. Switzerland, and Cambridge, United King- dom. SHORT COMMUNICATIONS 613 Mukherjee, a., C. K. Borad, and B. M. Parasharya. 2002. Breeding performance of the Indian Sams Crane in the agricultural landscape of western In- F dia. Biological Conservation 105:263-269. Nielsen, C. L. R. and R. J. Gates. 2007. Reduced nest C predation of cavity-nesting Wood Ducks during flooding in a bottomland hardwood forest. Condor 109:210-215. PiCMAN, J. M. 1988. Experimental study of predation on eggs of ground-nesting birds: effects of water -j depth and distance from edge. Condor 90:124- 131. y PiCMAN, J. M., M. L. Milks, and M. Leptich. 1993. Patterns of predation on passerine nests in marsh- The Wilson Journal of Ornithology 120(3):613-617, 2008 es: effects of water depth and distance from edge. Auk 110:89-94. PoiANi, A. 2006. Effects of floods on distribution and reproduction of aquatic birds. Advances in Eco- logical Research 39:63-83. Sanders, M. D. and R. E Maloney. 2002. Causes of mortality at nests of ground-nesting birds in the Upper Waitaki Basin, South Island, New Zealand: a 5-year video study. Biological Conservation 106:225-236. Timoney, K. 1999. The habitat of nesting Whooping Cranes. Biological Conservation 89:189-197. Wu, C. S. 1999. Study and management on resources in Zhalong National Nature Reserve. Northeast Forestry University Press, Harbin, China. White-winged Diuca Finch (Diuca speculifera) Nesting on Quelccaya Ice Cap, Peru Douglas R. Hardy and Spencer R Hardy^ ABSTRACT. — We found evidence of birds nesting directly on glacier ice of the Quelccaya Ice Cap in the Cordillera Vilcanota, Peru at elevations up to 5,300 m. Observations during June and July over several years consisted of numerous nests not in situ having obvi- ously fallen from the steep and dynamic, retreating glacier margin. A typical nest was a bulky stmcture of grass and twigs with a dry mass of 160 g. The inner cup was nicely formed and lined with fine grass, mea- suring 6-7 cm in diameter and 4-5 cm deep. Feathers and entire wings of White-winged Diuca Finch {Diuca speculifera) were observed in association with the nests; this was the passerine species most commonly seen in the area. The evidence indicates the glacier nests were built and used by White-winged Diuca Finch, probably during the Austral autumn when on- site automated measurements indicate the wet season ends and air temperatures have not yet decreased. This is the first well-documented case of high-elevation avi- an nesting on glacier ice. Received 30 November 2006. Accepted 26 December 2007. Some birds are well adapted to environments which are seasonally dominated by snow or sea ice, but birds are not generally associated with ' Climate System Research Center, Geosciences De- partment, University of Mas.sachusetts, Amherst, MA 01 003, USA. 2 Marion W. Cross School, Norwich, VT 05055, USA. Corresponding author; e-mail: dhardy@geo.umass.edu glaciers. Only the Emperor Penguin (Aptenod- ytes forsteri) is known to routinely nest on ice, typically frozen sea-ice (i.e., fast-ice) but at times in association with ice shelves derived from glaciers (Kooyman 1993). Transient birds have been observed passing over mountain gla- ciers at high elevations outside the polar regions or discovered after succumbing to harsh envi- ronmental conditions (Krajick 2002; L. G. Thompson, pers. comm.). However, glacier sur- faces are usually cold, actively changing through accumulation and ablation, and at times wet; conditions that are poorly suited for nesting and raising young birds. The ornithological literature contains only one detailed account of nesting on a glacier. This was the unusual circumstance where gla- ciers advancing into Alaska’s Prince William Sound overran a Black-legged Kittiwake {Ris- sa tridactyla) colony (Irons 1988). Previously- used nest sites were unavailable and 77 kitti- wake nests were constructed on the glacier face. All of these nests failed due to ablation and/or meltwater runoff which either dis- lodged the nests or caused them to disintegrate (Irons 1988; D. B. Irons, pers. comm.). The objective of this paper is to present ev- idence of nesting by White-winged Diuca Finch (Diuca speculifera) directly on glacier ice of the Quelccaya Ice Cap in Cuzco De- 614 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 3, September 2008 LIG. 1. (A) White-winged Diuca Linch on Quelccaya Ice Cap at —5,200 m on 23 June 2007, (B) the same glacier in June 2005 with arrow indicating the approximate location of the first nest observation in 2005 and the nest described in text. Circled area of margin just above the proglacial lake is enlarged in (C) looking under the ice. The circle on (C) encloses a nest which is likely in situ, 3 m above rock surface with enlarged view m (D); note silt from meltwater runoff. partment, Peru (14° S, 71° W). Observations were made over several years during the Aus- tral winter (Jun-Jul) in the course of con- ducting glacier and climate research on the ice cap. This is the first well-documented case of high-elevation avian nesting on a glacier, cor- roborating a second-hand report of “ice cave” nesting by White-winged Diuca Finch (John- son 1967). OBSERVATIONS During June 2005, while exploring the re- treating Quelccaya Ice Cap margin, we came upon a nest which appeared to have fallen re- cently from the glacier (within weeks- months). Several other older nests were ob- served nearby including one in a cave under the ice margin (Fig. 1). More extensive nest searches were conducted in 2006 and 2007 along —1,500 m of glacier margin, resulting in location of numerous nests. Most were along two sections with respective lengths of 350 and 530 m, elevation ranges of 50 and 100 m, and upper elevations of 5,200 and 5,300 m. Searches were restricted to this 1,500 m section; the extent to which this is representative of the Quelccaya Ice Cap is un- known. However, nesting on the ice cap has likely not been limited to the last 3 years as L. G. Thompson has occasionally observed nests over the past —30 years (pers. comm.). Nest remains were most often found on rocks at the base of near-vertical sections of ice margin. At least 14 nests were found in 2006 and at least 16 in 2007. These varied in apparent age from weeks (i.e., previous breed- ing season) to several years and the 2007 count almost certainly includes some found the previous year. Two of the freshest nests in 2006 were only 3 m apart. Typically, sections of glacier margin with nest remains were nearly vertical, somewhat grooved or fluted, and -5—10 m high. Over- hanging icicles and steeply-sloping rock be- low the margin made access to some nest re- mains difficult, and nests still attached to the glacier could have been overlooked. One nest was found on the glacier 22 m from the mar- SHORT COMMUNICATIONS 615 FIG. 2. (A) Dislodged nest along Quelccaya Ice Cap margin found on 15 July 2006 in inverted position within 1 m of vertical ice wall at 5,190 m. The nest was not present the previous year (26 Jun 2005), although a different residual nest was found —10 m distant. The dark bar on the field book cover is 1.25 X 9.5 cm. (B) Another dislodged nest in position found, 23 Jun 2007 at 5,230 m; note feathers and 13-cm pen for scale. gin, below a steep, step-like seetion of ice (i.e., not in situ). Some nests were entirely intact when found, while others were disintegrating or par- tially buried by sediment; however, almost all nests were not in situ (cf. Fig, ID). Net retreat of the ice margin is roughly 1 m/year or more along this part of the Quelccaya Ice Cap and nests constructed on the steeply-sloping mar- gin could only be observed in situ within a brief interval following construction (i.e., breeding season) prior to falling. Evidence for this interpretation is that several nests were found in inverted position and, in almost ev- ery case, a vertical trace of nest material was observed above the nest remains, frozen to the ice. Remains of varying ages indicate multi-year occupation of favorable sites along the ice mar- gin suggesting that reproductive efforts on the ice cap are successful. In addition, we observed fecal sacs in one of the fresh nests, presumably from nestlings just prior to fledging. No off-gla- cier nesting evidence was found despite search- ing areas adjacent to the glacier. Several nests appeared entirely intact and the following is based upon one of the fresh- est-appearing nests observed in 2006. Nests were bulky structures of grass and twigs with a deep, well-made inner cup (Fig. 2). It ap- pears that a rough platform is initially con- structed (32 X 18 cm), which roughly tapers upward and becomes increasingly well-woven towards the top (13 cm outer diameter). Only this better-woven upper portion was found in some cases. The inner cup measured —6.5 cm in diameter and was 4.6 cm deep. Overall dry mass of the nest (Fig. 2A) was 160 g. Nests consisted of woven grass (—80-90%), partic- ularly the locally abundant Calaniagrostis chrysantha. The inner cup was lined with finer grass and feathers were observed in or adja- cent to some nest remains (Fig. 2B). Analysis by Carla Dove at the Smithsonian Institution revealed feathers of Rufous-bellied Seedsnipe {Attagis gayi), Andean Goose {Chloephaga melanoptera), and tail feathers of White- winged Diuca Finch. This diverse assemblage suggests that nests are lined with feathers re- covered from the surrounding landscape. No evidence of camelid fleece was seen in any nest (cf. Johnson 1967). DISCUSSION Although indirect, all evidence indicates that Quelccaya glacier nests were built and used by White-winged Diuca Finch. Little information exists on nesting habits of this species from any- where in the Andes although Johnson (1967: 368) repoiled that in the Parinacota area of 616 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 Chile, a White-winged Diuca Finch nest was found “on the ground beneath some loose stones on a hillside.” In lieu of breeding-season observations of White- winged Diuca Finches at Quelccaya, our deduction is based upon their local presence, feathers, and the species known association with glaciers. Small groups of White-winged Diuca Finches of unknown age class were frequently observed among rocky moraine surfaces and bogs in the area. This species is known to not typically retreat northward or to lower eleva- tions during winter (Johnson 1967, 1972). On one occasion in June 2007, while we were in- vestigating an apparent roosting site within a crack of the glacier, a flock of —20 White- winged Diuca Finches began gathering late in the afternoon at the glacier margin, both on and off the glacier (Fig. 1). The birds acted disturbed by our presence in contrast to their behavior during diurnal feeding. Several feathers observed in proximity to the nests were White-winged Diuca Finch wing or tail feathers. In addition, two nearly- entire wings (chord = 105 mm) of White- winged Diuca Finch were found on the ice at a 2007 nest site. No feathers were found from other passerine species (e.g., ground tyrants [Muscisaxiola spp.]) observed in close prox- imity to the glacier. White-winged Diuca Finch is the only spe- cies to be repeatedly associated in the orni- thological literature with Western Hemisphere glaciers. Niethammer (1953) observed an es- timated 100 White- winged Diuca Finches gathering for the night inside a glacier cre- vasse at Chacaltaya (5,200 m) Bolivia in mid- summer. He also collected one male specimen on 20 December in breeding condition with enlarged testes (6X4 mm vs. —3 mm for another specimen collected in late August). A second association between White-winged Diuca Finch and glaciers was in the mid- 1960s when P. R. Parker of the Chacaltaya Astrophysical Observatory found a nest “in an ice cave” at 5,300 m, leading Johnson (1967:368) to postulate the species “quite possibly nests at a higher altitude than any other passerine form.” Subsequently, several sources mention that White-winged Diuca Finch has been recorded roosting in glacier fissures (Meyer de Schauensee 1970), glacier crevasses (Ridgely and Tudor 1989) or glacier cracks (Fjeldsa and Krabbe 1990). Relatively little is known about the timing of White-winged Diuca Finch nesting in Peru or elsewhere. Johnson (1967) suggested this species nested after the summer rains in the Parinacota region (Chile— Bolivia) where White-winged Diuca Finches were present all year. Immature White-winged Diuca Finches in Bolivia were noted by Fjeldsa and Krabbe (1990) during July and August (La Paz), and August (Cochabamba). These findings are consistent with the Quelccaya situation, where climatic conditions present multiple difficul- ties for breeding birds, especially snow and low temperature (cf. Hendricks and Norment 1992, Martin and Wiebe 2004). We began operating an automated weather station (AWS) in 2003 at the ice cap summit, —3 km east of the nest sites and —500 m high- er in elevation. These data permit a close ap- proximation of the climate in which nest building, incubation, and rearing of nestlings occurs on the glacier. Pronounced seasonality of precipitation at Quelccaya typically features considerable and frequent snowfall from late September to early- mid April. During this wet season, 2 m or more of snow accumulates at the ice cap summit but, at the slightly lower elevation margin, ablation predominates; meltwater flows off during the day and freezes at night. Some precipitation at these nest sites (5,150-5,300 m) may be in the form of rain, as they are close to a rising at- mospheric freezing level (Thompson et al. 1993, Bradley et al. 2006). This would impact the bird’s exposure to moisture, thermal regulation, food availability, and other factors, and the vi- ability of glacier nesting in the area. Diurnal fluctuations in air temperature at Qu- elccaya are greater than the annual variation. By assuming a constant environmental lapse rate, summit AWS measurements can be adjusted to the elevation of nesting site, revealing average daily minima of —3.1° and — 6.3° C during the wet and dry seasons, respectively. Thus, air tem- peramre is low during the night throughout the year with extreme radiational cooling whenever cloud cover is low (especially in the dry sea- son). Maxima for the same periods reach 2.9° and 0.9° C at the nesting sites. Successful nesting in this dynamic environ- ment and extreme climate requires not only SHORT COMMUNICATIONS 617 thermal and mechanical adaptations to the ice substrate, but also careful timing within the seasonal cycle of climate. The AWS data sug- gest that Quelccaya glacier nesting by White- winged Diuca Finch most likely occurs as the wet season concludes in April when nesting sites become exposed and drier, as the tran- sient snow line elevation rises. Daily mean temperatures decrease after March as decreas- ing cloud cover results in colder nights, but the decrease in daily maximum is consider- ably more gradual until June (DRH, unpubl. data). Young finches had apparently fledged in each of the past 3 years by this date. Note Added in Proof. — Quelccaya field- work during June 2008 provided additional observations. The first nests clearly in situ were observed on the glacier, in locations and orientation as hypothesized. One contained two abandoned eggs consistent in size and color for White-winged Diuca Finch (Johnson 1967). Also, a single off-glacier nest, similar to that described above, was found in situ un- der boulders —500 m from the glacier. A late- aftemoon gathering of 20-30 White-winged Diuca Finch was again observed, 30-60 min prior to sundown, at the same section of frac- tured glacier margin where this behavior was observed in 2007. Several roosts were located nearby, all entirely within (vs. beneath) the glacier. One expedition member observed and photographed White-winged Diuca Finch roosting at the site prior to sunrise (J. A. Cas- taneda Gil, pers. comm.). Supplemental ma- terial on glacier-nesting White-winged Diuca Finch at Quelccaya Ice Cap is available at http://www.geo.umass.edu/climate/quelccaya/ diuca.html. ACKNOWLEDGMENTS This material is based upon work supported by the National Science Foundation and NOAA Office of Global Programs, Climate Change Data and Detection Program (Grant No. 0402557, awarded to DRH). The able field assistance of Mathias Vuille, Carsten Braun, and D. R. Dockstader, along with Mountain Guide Fe- lix Vicencio and staff, is gratefully acknowledged. Ini- tial discussion of these observations with Jon Fjeldsa, Alvaro Jaramillo, and M. A. Plenge was helpful. Carla Dove at the Smithsonian Institution (Washington, D.C.) undertook feather identification in her laboratory and provided considerable enthusiasm. We also thank G. A. Clark, who encouraged preparation of the man- uscript and provided insightful comments on an earlier draft. Two anonymous reviewers considerably im- proved the manuscript. LITERATURE CITED Bradley, R. S., M. Vuille, H. F. Diaz, and W. Ver- gara. 2006. Threats to water supplies in the trop- ical Andes. Science 312:1755-1756. Fjeldsa, J. and N. Krabbe. 1990. Birds of the High Andes. Zoological Museum, University of Copen- hagen, and Apollo Books, Svendborg, Denmark. Hendricks, P. and C. J. Norment. 1992. Effects of a severe snowstorm on subalpine and alpine popu- lations of nesting American Pipits. Journal of Field Ornithology 63:331-338. Irons, D. B. 1988. Black-legged Kittiwakes nest on advancing glacier. Wilson Bulletin 100:324-325. Johnson, A. W. 1967. The birds of Chile and adjacent regions of Argentina, Bolivia and Peru. Volume 2. Platt Establecimientos Graficos, Buenos Aires, Argentina. Johnson, A. W. 1972. The birds of Chile and adjacent regions of Argentina, Bolivia and Peru, Supple- ment. Platt Establecimientos Graficos, Buenos Ai- res, Argentina. Kooyman, G. L. 1993. Breeding habitats of Emperor Penguins in the western Ross Sea. Antarctic Sci- ence 5:143-148. Krajick, K. 2002. Melting glaciers release ancient rel- ics. Science 296:454-456. Martin, K. and K. L. Wiebe. 2004. Coping mecha- nisms of alpine and arctic breeding birds: extreme weather and limitations to reproductive resilience. Integrative and Comparative Biology 44:177-185. Meyer de Schauensee, R. 1970. A guide to the birds of South America. Academy of Natural Sciences of Philadelphia and Livingston Publishing Com- pany, Wynnewood, Pennsylvania, USA. Niethammer, G. 1953. Zur Vogelwelt Boliviens. Bon- ner Zoologische Beitriige 4:195-303. Ridgely, R. S. and G. Tudor. 1989. The birds of South America. Volume I. The o.scine passerines. University of Texas Press, Austin, USA. Thomp.son, L. G., E. Mosley-Thomp.son. M. Davis. P- N. Lin, T. Yao, M. Dyurgerov, and J. Dai. 1993. “Recent warming”: ice core evidence from trop- ical ice cores with emphasis on Central Asia. Global and Planetary Change 7:145-156. 618 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 The Wilson Journal of Ornithology I20(3);618— 619, 2008 Nest Success and Nest Predation of the Endangered Rota White-eye (Zosterops rotensis) Lainie Berry^-^ and Estanislao Taisacan^ ABSTRACT — The Rota White-eye {Zosterops ro- tensis) is an endangered species endemic to the island of Rota in western Micronesia. We monitored eight nests from 2003 to 2005, four of which produced at least one fledgling. Clutch size was two in each of five nests that were counted; the average number of fledg- lings from successful nests was 1.5 {n = 4). We filmed six nests and captured two nest predation events on video. A Mariana Crow {Corxnis kubaryi), which is also an endangered species, was filmed taking nest- lings from one white-eye nest. Another nest containing eggs was depredated by a rat {Rattiis spp.); however, this may have occurred after the nest was abandoned due to the presence of the camera. Received 24 Sep- tember 2007. Accepted 16 January 2008. The Rota White-eye {Zosterops rotensis) is endemic to the island of Rota, Northern Mar- iana Islands, in western Micronesia (14° 09' N, 145° 12' E) and is listed as federally en- dangered (USDI 2004). The population is cur- rently limited to —250 ha of mature limestone forest above 150 m elevation (Fancy and Snet- singer 2001). Island-wide surveys yielded a population estimate of -10,000 white-eyes in 1982 (Engbring et al. 1986); surveys in 1996- 1999 indicate the population has since de- creased to — 1 ,000 individuals ( Amidon 2000, Fancy and Snetsinger 2001). Predation by the introduced Black Drongo (Die rums macro- ceriis) or rats {Rattus spp.) is a possible cause of the decline of the Rota White-eye (Craig and Taisacan 1994, Fancy and Snetsinger 2001). Direct evidence on causes of the de- cline is lacking (Amidon 2000, Fancy and Snetsinger 2001), partly because of the treach- erous terrain, thick vegetation, and difficulty of working on Rota. ' Rota Avian Behavioral Ecology Program, Depart- ment of Psychology, University of Washington, P. O. Box 1298, Rota MP 96951, USA. - Isa Consulting Services, P. O. Box 1381, Rota MP 96951, USA. ^ Corresponding author; e-mail: lainieb@u. washington.edu We searched the As Akoddo and Uyulan Hulo regions of Rota between 2003 and 2005. These regions occur on north-facing slopes with an elevation range of 150-400 m and are heavily forested with a dense cover of Mer- rilliodendron megacarpwn. Eight active nests were found and monitored; six of these were filmed continuously until either the nestlings fledged or the nest failed. The nests were filmed using a black and white CCD infra-red camera placed 5—7 m from the nest, and re- corded on a Sanyo SRT-2400DC real-time video cassette recorder. Our objectives were to examine nest fate and identify nest preda- tors for possible future control measures. OBSERVATIONS All eight white-eye nests were in Merril- liodendron megacarpum trees. We observed nesting activity from March to July each year, although we also observed nest building in September. Four of the eight nests were suc- cessful in producing at least one fledgling. One nest was depredated by a Mariana Crow {Corviis kubaryi) during the nestling stage. One was depredated by a rat during the in- cubation stage, but this might have occurred after the nest was abandoned due to placement of the video camera too close to the nest (Tha- lia Sachtleben, pers. comm.). One nest failed because the eggs were cracked. The fate of the eighth nest was unknown. Clutch size was two eggs in all five nests for which clutch size was confirmed. The average number of fledglings from successful nests was 1.5 (n = 4). DISCUSSION All eight nests were in Merrilliodendron megacarpum, most likely because this was the predominant tree species in the area searched. Amidon et al. (2004) found Rota White-eye nests in four tree species (Elaeocarpus joga, Hernandia labyrinthica, Merrilliodendron megacarpum, and the introduced Acacia con- SHORT COMMUNICATIONS 619 fusa), but their study included searches in re- gions that were not dominated by Merrillio- dendron megacarpum. Amidon et al. (2004) observed active nests in December, March, July, and August, and re- ported indirect evidence of nesting in January, April, and June. Lusk and Taisacan (1997) re- ported a nesting season of at least March to June. In combination with our finding of nests between March and July and one instance of nest building in September, these observations suggest the Rota White-eye nesting season may extend as long as 10 months from December to September, and breeding may even occur year- round (Amidon et al. 2004). Four of the eight nests monitored were suc- cessful in producing at least one fledgling. In comparison, seven of 10 nests monitored by Amidon et al. (2004) were successful in pro- ducing at least one fledgling; one was de- stroyed by a typhoon, one was depredated, and one was either depredated or abandoned, possibly due to human interference. Sample sizes are small, but nest predation does not appear to be a significant factor af- fecting the reproductive success of Rota White- eyes in either the current study or in Amidon et al. (2004). The only species that was positively identified as a nest predator was the Mariana Crow, which is also endangered (USDI 1984); thus, control of this species is not possible. In- troduced rats have caused population declines of birds on other Pacific islands (Atkinson 1977, King 1985), but nest predation by rats did not appear to be a significant cause of nest failure in this limited study. Sachtleben (2005) filmed nests of the Bridled White-eye (Zosterops con- spicillatus). Golden White-eye {Cleptomis mar- chei), and Rufous Fantail (Rhipidura mfifrons) on Saipan, Northern Mariana Islands, and cap- tured four predation events. Three nests were depredated by Micronesian Starlings (Aplonis opaca) and one by a Collared Kingfisher {To- diramphus chloris). Both species are native to Saipan and Rota. No nests were found to be depredated by rats, contrary to expectations (Sachtleben 2005). There was no evidence for nest predation by Black Drongos on Rota. Ad- ditional evidence on nest predator identity and the effects of nest predation on the breeding suc- cess of the Rota White-eye is required before predator control is initiated. ACKNOWLEDGMENTS We thank Shelly Kremer for encouragement and support of this work, Thalia Sachtleben for her insight- ful review of nest footage, Renee Ha for transferring video footage to DVD, and Monica Awasthy, Fred Amidon, and an anonymous reviewer for comments on the manuscript. This study was funded by a U.S. Fish and Wildlife Service Traditional Section 6 endangered species grant. LITERATURE CITED Amidon, F A. 2000. Habitat relationships and life his- tory of the Rota Bridled White-eye {Zosterops ro- tensis). Thesis. Virginia Polytechnic Institute and State University, Blacksburg, USA. Amidon, F. A., C. A. Haas, and J. M. Morton. 2004. Breeding biology of the endangered Rota Bridled White-eye. Wilson Bulletin 116:342-346. Atkinson, I. A. E. 1977. A reassessment of factors, particularly Rattus rattus L., that influenced the decline of endemic forest birds in the Hawaiian Islands. Pacific Science 31:109-133. Craig, R. J. and E. Taisacan. 1994. Notes on the ecology and population decline of the Rota Bri- dled White-eye. Wilson Bulletin 106:165-169. Engbring, j., E L. Ramsey, and V. J. Wildman. 1986. Micronesian forest bird survey, 1982: Saipan, Ti- nian, Agiguan, and Rota. U.S. Eish and Wildlife Service, Honolulu, Hawaii, USA. Fancy, S. G. and T. J. Snetsinger. 2001. What caused the decline of the Bridled White-eye on Rota, Mar- iana Islands? Studies in Avian Biology 22:274-280. King, W. B. 1985. Island birds: will the future repeat the past? Pages 3-15 in Conservation of island birds (P. J. Moors, Editor). International Council for Bird Preservation Technical Publication Num- ber 3. International Council for Bird Preservation, Cambridge, United Kingdom. Lusk, M. R. and E. Taisacan. 1997. Description of a Bridled White-eye nest from Rota, Mariana Is- lands. Micronesica 30:183-185. Sachtleben, T. 2005. Predation and nest success of forest birds in native and non-native habitat on Saipan, Mariana Islands. Thesis. Colorado State University, Fort Collins, USA. U.S. Department of Interior (USDI). 1984. Endan- gered and threatened wildlife and plants: deter- mination of endangered species status for seven birds and two bats of Guam and the Northern Mariana Islands; final rule. Federal Register 49: 33881-33885. U.S. Debar IMENT of Interior (USDI). 2004. Endan- gered and threatened wildlife and plants: endan- gered status for the Rota Bridled White-eye {Zos- terops rotensis) from the Commonwealth of the Northern Mariana Islands: final rule. P'ederal Reg- ister 69:3022-3029. 620 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 The Wilson Journal of Ornithology 120(3):620-624. 2008 Predators at Nests of the Western Slaty Antshrike {Thamnophihis atrimicha) Corey E. Tarwater^-^ ABSTRACT — Predation is the primary cause of nest loss in most passerine species. While nest preda- tion is important, the predator community and behav- iors of parents and predators during predation events are poorly documented. I witnessed one and video- taped four predation events at nests of Western Slaty Antshrikes {Thamnophihis atrimicha) in central Pana- ma. Predators included a snake {Pseustes poecilono- tiis), a monkey (white-faced capuchin [Cebiis capiici- niis]), and three species of birds. Predators spent little time at the nest, yet some returned repeatedly to the same nest. Parents also returned several times to nests after the predation event and parental behaviors varied depending on the predator. Nest disturbance was not an accurate indicator of predator type. Received 27 January 2007. Accepted 11 November 2007. Nest predation is the primary source of nest loss in most passerine species (Ricklefs 1969) and is likely a strong selective pressure in avi- an life histories (Martin 1996). Nest predation in tropical regions may be especially high and has been hypothesized to influence life history traits such as small clutch size and parental activity at the nest (Skutch 1985, Martin 1996). Most evidence of the identity of nest predators in the Neotropics is based on anec- dotal or indirect evidence (Skutch 1985, Ro- per 1992, Sieving 1992, Robinson and Rob- inson 2001). A wide diversity of predators is typical of these forests (Ricklefs 1969, Skutch 1985, Sieving 1992) and studies suggest the importance of mammals (Roper 1992), snakes (Skutch 1985, Robinson et al. 2005a), and raptors (Skutch 1985, Robinson and Robinson 2001) in predation of bird nests. Identifying the predator community is important for un- > Department of Natural Resources and Environ- mental Sciences, University of Illinois at Urbana- Champaign, 606 East Healey Street. Champaign, IL 61820, USA. 2 Current address; Program in Ecology, Evolution, and Conservation Biology, University of Illinois at Ur- bana-Champaign, 606 East Healey Street, Champaign, IL 61820, USA; e-mail: tarwater@uiuc.edu derstanding predator-prey interactions (Weath- erhead and Bloun-Demers 2004), nest site se- lection (Weatherhead and Bloun-Demers 2004, Remes 2005), and evolution of parental care behaviors (Skutch 1949). Understanding parental behavior during predation events is necessary for examining whether parents ac- cept risks to themselves by engaging in nest defense (Montgomerie and Weatherhead 1988, Pietz and Granfors 2005) and whether predator-specific responses occur (Montgo- merie and Weatherhead 1988). Direct observations of predation events are rare since the length of time predators spend at nests is often short. This results in low probability of the observer and the predator being at the nest at the same time (Robinson et al. 2005a). Identification of predators based on disturbance to the nest is also unreliable (Thompson et al. 1999, McCallum and Han- non 2001, Robinson et al. 2005a). Use of vid- eo cameras has become a popular and reliable tool for observing predation events (Thomp- son et al. 1999, Pietz and Granfors 2000, Rob- inson et al. 2005a) because of the challenges in identifying the predator community. The objectives of this paper are to describe five predation events of nests of a neotropical passerine, the Western Slaty Antshrike {Tham- nophilus atrimicha), and to evaluate the suc- cess of recording predation events by video- taping nests. METHODS I videotaped and monitored nests of West- ern Slaty Antshrikes in 2004-2006 in Parque Nacional Soberama (mainland site) and Barro Colorado Island. Both sites are in lowland, tropical moist forests in the Republic of Pan- ama (Karr 1971, Robinson 1999, Robinson et al. 2000). Most nests were found during build- ing and were monitored every 2—3 days de- pending on stage. The incubation period is 16 days and the nestling period is 10 days SHORT COMMUNICATIONS 621 (Skutch 1969, Oniki 1975, Robinson et al. 2000). The open cup nests of Slaty Antshrikes are in forks of trees and range from <1 m above ground to 6 m in height (Roper and Goldstein 1997). Thirty-eight nests were videotaped of which 25 were filmed during the nestling period (24 nests in 2004). Video cameras (model: Sony CCD TRV 328) were placed 2-4 m from the nest (depending on the vegetation) in 2004 and nests were videotaped continuously from 0700 to 1400 hrs (EST), except during heavy rain. Nests were filmed for 2-4 hr periods in 2006 {n = 1) and for shorter periods (<30 min) in 2005 {n = 7). Cameras were placed 10 m from the nest in 2005 and 2006 and nests were filmed throughout the day. I could not see nest defense by parents outside of a 1 -m area around the nest, but vocalizations by parents could be heard at farther distances (—20 m). Camouflage and the surrounding vegetation were used in all years to minimize disturbance to the nest. Parents resumed ap- parently normal behavior after cameras were installed suggesting that disturbance to the nest was minimal. RESULTS Four predation events were filmed and one was observed. Three predation events filmed at the mainland site were by Double-toothed Kite {Harpagus bidentatus). Keel-billed Tou- can (Ramphastos sulfuratus), and white-faced capuchin (Cebus capucinus). I also observed a Fasciated Antshrike {Cymbilaimus lineatus) prey on a nest at the mainland site. One pre- dation event by a snake (Pseustes poecilono- tus) was filmed on Barro Colorado Island. Double-toothed Kite. — The antshrike nest was 2 m from the edge of the vegetation and 2 m above ground. The predation event oc- curred on day 7 of the nestling period. Vid- eotaping of the nest (with 1 nestling) began at 0719 hrs. A Double-toothed Kite arrived at the nest at 0738 hrs. The kite pulled at the nestling and the nestling vocalized an appar- ent alarm sound. The kite bit at the nestling once again and departed after 1 min at the nest. An hour and 46 min later, another or the same Double-toothed Kite returned to the nest. The kite pecked at the nest for 1 min and then departed. A kite returned after 3 min, looked into the nest, and then flew away. Adult antshrikes were not heard vocalizing during predation or subsequent visits by the kite. Each parent returned twice with prey fol- lowing the predation event. The male arrived 24 min after predation, followed 30 min later by the female, and then the male 27 min later, and lastly the female after another 53 min (for a total of 140 min after predation). The par- ents did not return during the remaining 226 min of filming. The return intervals of parents to the nest were not unusually long as Slaty Antshrike parents provision young only a few times per hour. The adults called continuously each time they arrived with food. The parents, clearly disturbed, looked into and hopped in and out of the nest, and sang faintly. They then consumed the food item brought, and de- parted. Parents spent an average of 1.5 min at the nest. In addition to vocalizations at the nest, parents also called and sang faintly out- side of the range of the video camera, but near the nest. Parents vocalized for about 3 min before returning to the nest and up to 5 min after leaving the nest. Keel-billed Toucan. — The antshrike nest (with 2 nestlings) was far from any edge and 1 m in height above ground. The predation event occurred on day 7 of the nestling period. Filming of the nest began at 0719 hrs. The first indication of a disturbance was at 1330 hrs when one of the adults began alarm calling every 15 sec. Sounds of toucans flying over- head were heard at 1331 hrs and became loud- er as toucans approached the nest. The adult antshrikes were silent at 1333 hrs and move- ment was heard near the nest. A Keel-billed Toucan flew near the nest at 1334 hrs and hopped to a branch above the nest. The toucan took one of the nestlings without disturbing the nest. Both nestlings gave alarm calls dur- ing the event and one of the adults responded with alarm calls every 5 sec. One nestling fled the nest and fell to the ground while emitting alarm calls when the toucan took the first nest- ling. The toucan ate the nestling and left after spending <30 sec at the nest. Another or the same toucan returned to the nest 3 min later. Sounds of both the toucan and the remaining nestling were heard and then suddenly both were silent suggesting the toucan found the nestling. The adults continued to emit alarm calls for 18 min after the predation event at 622 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 which time the toucans were no longer heard in the video. The parents did not return to the nest during the remaining 38 min the nest was filmed after the predation event. The pair of antshrikes renested soon after the predation event confirming the loss of the second off- spring. White-faced Capuchin.— antshrike nest (with 2 eggs) was on the edge of a small gap and 1.5 m above ground; the predation event occurred in the early stage of the incubation period. Filming of the nest began at 0830 hrs. The female antshrike was incubating for 2.5 hrs when white-faced capuchins were first heard on the tape (1101 hrs). Sounds of the monkeys became louder during the next 4 min. The incubating female began looking around more actively and left the nest at 1105 hrs. Soon (<30 sec later), the nest tree shook, followed 3 sec later with a monkey in view. The monkey lowered the nest branch from the ground and reached into the nest to take the eggs, breaking the branch in the process. The monkey preyed on the nest and left the video within 10 sec of arrival. Monkeys were heard in the video for an additional 8 min. The par- ents began calling at 3-5 sec intervals 9 min (1123 hrs) after the monkeys were no longer heard in the video. Their alarm calls got loud- er in the video as they approached the nest. The male arrived at the nest at 1136 hrs, looked into the nest, and called three times. He departed after 1 min, while calling. Alarm calls continued becoming quieter as the pair moved away from the nest until 1140 hrs, when no further vocalizations were heard. Nothing else noteworthy happened during the remaining 57 min the nest was filmed. Fasciated Antshrike. — This predation event occurred in the early stage of the incubation period. The nest had two eggs and was in the middle of a small gap at a height of 3.5 m above ground. I was nearby at 1400 hrs when I heard the Slaty Antshrike alarm call. I ar- rived at the nest and saw a male Fasciated Antshrike pecking at eggs in the nest. The male and female Slaty Antshrikes were calling and flying back and forth, 0.5-2 m from the nest. The male Slaty Antshrike attacked the male Fasciated Antshrike, striking him three times. The Fasciated Antshrike would leave the nest after each attack, but would return within a few seconds. Once the male Fasciated Antshrike finished at the nest, the Fasciated Antshrike pair left the vicinity of the Slaty Antshrike nest and vocalized for the next mi- nute. The female Fasciated Antshrike stayed within the vicinity of the nest while the male preyed upon the eggs, but was not part of the predation event. The Slaty Antshrikes called while the Fasciated Antshrikes were in the area. The male Slaty Antshrike returned to the nest <1 min after the predation event oc- curred, pecked into the nest 3-4 times, and then flew away. I did not continue observing the nest after the male departed. Pseiistes poecilanotiis. — This nest (with 1 nestling) was far from an edge and 1.25 m in height above ground. The predation event oc- curred on day 2 of the nestling period. I flushed the female from the nest just prior to videotaping the nest (filming began at 0804 hrs). Both parents subsequently vocalized, but when filming began, they were >20 m from the nest and silent. The snake came into view' at 0811 hrs. The snake was at the nest for 2 min while swallowing the nestling. The par- ents were not heard vocalizing during preda- tion of the nestling. The female returned 25 min after the snake had departed and sat on the empty nest for 61 min. The male arrived with food a few seconds after the female de- parted. The male vocalized (faint songs and contact calls), hopped around the nest, ate the food item, and then departed after 3 min. The nest was filmed for an additional 158 min dur- ing which time neither birds nor snake re- turned. Success of Recording Predation Events by Videotaping Nests. — Four hundred and sev- enteen hrs of observations were filmed {n = 38 nests) of which 397 hrs were from 24 nests during the nestling period. Twenty-three of 38 nests that were filmed were depredated (60.5%). Of 354 nests that were not filmed, 281 (79.4%) were lost due to predation. The difference in percent nests lost was likely be- cause most nests were filmed during the nest- ling period. A greater number of nests were depredated during incubation (252 of 304 total predation events occurred during the incuba- tion period). The number of predation events recorded (of the total number of hours nests were filmed) was 0.01 predation events/hr filmed. SHORT COMMUNICATIONS 623 DISCUSSION Keel-billed Toucans, Pseustes, Double- toothed Kites, and white-faced capuchins have been observed preying on bird nests (Willis 1972, Skutch 1985, Robinson and Robinson 2001, Robinson et al. 2005a). This is the first observation of a Fasciated Antshrike as a nest predator. Slaty and Fasciated antshrikes are commonly sympatric, often attending under- story antwren flocks together. However, when Slaty Antshrikes have a nest or young fledg- lings, the parents call and displace Fasciated Antshrikes when they approach their off- spring. This suggests that Fasciated Ant- shrikes are perceived by Slaty Antshrikes as a threat. The behaviors of the predators during the predation events were similar to other obser- vations of predators at nests. Predators may return repeatedly to the nest (Marzluff 1985, Stake and Cimprich 2003, Stake et al. 2004) and spend little time at nests (Robinson et al. 2005a). Predators at antshrike nests did not spend more than 2 min at a nest. Additionally, features of nest disturbance do not provide re- liable information to identify predators (Thompson et al. 1999; McCallum and Han- non 2001; Robinson et al. 2005a, b). Predation due to mammals is often associated with nest disturbance and predation by birds is not. However, both the Double-toothed Kite and the white-faced capuchin damaged the nest in this study. The toucan, Fasciated Antshrike, and snake left no evidence that might identify the predator. Parental defense behavior can vary depend- ing upon the risk the predator poses and like- lihood of successfully defending the nest (Montgomerie and Weatherhead 1988, Ghal- ambor and Martin 2001). Toucans and Fasci- ated Antshrikes are probably not threats to an adult Slaty Antshrike, whereas a Double- toothed Kite might be (Willis 1972). The Slaty Antshrike pair vocalized during and after the predation event by the toucan, and adults were aggressive in their nest defense with the Fas- ciated Antshrike. Fasciated and Slaty ant- shrikes are similar in size. The parents may have been willing to attack the predator be- cause of the small threat to themselves and greater probability of successfiilly defending the nest. However, in the case of the Double- toothed Kite(s), the parents did not vocalize when this species was at the nest. The parents potentially remained silent when the kite(s) was present to reduce their risk. The parents were silent while monkeys were close to their nest. It is unlikely that a capuchin could prey on an adult off the nest, but vocalizations by parents are also unlikely to deter monkeys. Adults returned to nests after predation events. Although attendance of empty nests may occur, (Pietz and Granfors 2000; M. M. Libsch and GET, pers. obs.), the phenomena is rarely reported. Despite high predation of Slaty Antshrike nests, few predation events were observed on videotape. More filming during the highest mortality period of the nest- ing cycle is required to record a greater num- ber of predation events. A diversity of pred- ators exists in the lowland forests of Panama and parental behavior varies depending on species of predator. More work is needed to understand the importance of the different predators on nest success and to identify the predator community. ACKNOWLEDGMENTS I thank S. K. and R. D. Bassar, C. G. Batista, B. M. Hancock, B. G. Lascelles, M. G. Meadows, and I. R. Ochoa for help collecting the field data that made this study possible. I also thank J. R Kelley and M. M. Libsch for support and the Smithsonian Tropical Re- search Institute for logistical support. I thank J. D. Brawn, R J. Weatherhead, and A.V. Suarez for helpful comments on the manuscript. This work was funded by National Science Foundation grant IBN-02 12587 and grants from the University of Illinois. LITERATURE CITED Ghalambor, C. K. and T E. Martin. 2001. Fecun- dity-survival trade-offs and parental risk-taking in birds. Science 292:494-497. Karr, J. R. 1971. Structure of avian communities in selected Ranama and Illinois habitats. Ecological Monographs 41:207-233. Martin, T. E. 1996. Life-history evolution in tropical and south temperate birds: what do we really know? Journal of Avian Biology 27:263-272. Marzluff, J. M. 1985. Behavior at a Rinyon Jay nest in response to predation. Condor 87:559-561. McCallum, C. A. and S. J. Hannon. 2001. Accipiter predation of American Redstart nestlings. Condor 103:192-194. Montgomfrih, R. D. and R. J. Wfathfrmfad. 1988. Risks and rewards of nest defense by parent birds. Quarterly Review of Biology 63:167-187. Oniki, Y. 1975. The behavior and ecology of Slaty 624 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 Antshrikes {Thamnophilus punctatus) on Barro Colorado Island, Panama Canal Zone. Annals de Academia Brasil Ciencias 47:471-515. PiETZ, P. J. AND D. A. Granfors. 2000. Identifying predators and fates of grassland passerine nests using miniature video cameras. Journal of Wildlife Management 64:71-87. PiETZ, P. J. .AND D. A. Granfors. 2005. Parental nest defense on videotape: more reality than “myth.” Auk 122:701-705. Remes, V. 2005. Birds and rodents destroy different nests: a study of Blackcap Sylvia atricapilla using the removal of nest concealment. Ibis 147:213- 216. Ricklefs, R. 1969. An analysis of nesting mortality in birds. Smithsonian Contributions to Zoology 9:1- 48. Robinson, W. D. 1999. Long-term changes in the avi- fauna of Barro Colorado Island, Panama, a tropi- cal forest isolate. Conservation Biology 13:85-97. Robinson, W. D. and T. R. Robinson. 2001. Obser- vations of predation events at bird nests in central Panama. Journal of Field Ornithology 72:43-48. Robinson, W. D., G. Rompre, and T. R. Robinson. 2005a. Videography of Panama bird nests shows snakes are principal predators. Omithologia Neo- tropical 16:187-195. Robinson, W. D., J. N. Styrsky, and J. D. Brawn. 2005b. Are artificial bird nests effective surrogates for estimating predation on real bird nests? A test with tropical birds. Auk 122:843—852. Robinson, W., T. Robinson, S. Robinson, and J. Br.aw?^. 2000. Nesting success of understory for- est birds in central Panama. Journal of Avian Bi- ology 31:151-164. Roper, J. J. 1992. Nest predation experiments with quail eggs — too much to swallow. Oikos 65:528- 530. Roper, J. J. and R. R. Goldstein. 1997. A test of the Skutch hypothesis: does activity at nests increase nest predation risk? Journal of Avian Biology 28: 111-116. Sieving, K. E. 1992. Nest predation and differential insular extinction among selected forest birds of central Panama. Ecology 73:2310-2328. Skutch, A. F. 1949. Do tropical birds rear as many young as they can nourish? Ibis 91:430-455. Skutch, A. F. 1969. Life histories of Central American birds. III. Pacific Coast Avifauna 35:172-179. Skutch, A. F. 1985. Clutch size, nesting success, and predation on nests of neotropical birds, reviewed. Ornithological Monographs 36:575-594. Stake, M. M. and D. A. Cimprich. 2003. Using video to monitor predation at Black-capped Vireo nests. Condor 105:348-357. Stake, M. M., J. Faaborg, and F. R. Thompson. 2004. Video identification of predators at Golden- cheeked Warbler nests. Journal of Field Ornithol- ogy 75:337-344. Thompson, F. R., W. Duak, and D. E. Burhans. 1999. Video identification of predators at songbird nests in old fields. Auk 116:259-264. Weatherhead, P. j. and G. Bloun-Demers. 2004. Un- derstanding avian nest predation: why ornitholo- gists should study snakes. Journal of Avian Biol- ogy 35:185-190. Willis, E. O. 1972. The behavior of Spotted Antbirds. Ornithological Monographs 10. SHORT COMMUNICATIONS 625 The Wilson Journal of Ornithology 120(3):625-627, 2008 Evidence for Bachman’s Sparrow Raising Brown-headed Cowbirds to Fledging Matthew J. Reetz,^-^ Elizabeth Farley, ^ and Thomas A. Contreras^ ABSTRACT — We report the first records of Brown- headed Cowbirds {Molothrus ater) raised to fledging by Bachman’s Sparrow {Aimophila aestivalis). Re- cords are based on field observations of parasitized sparrow nests monitored during two separate avian re- productive studies. One record is of a Bachman’s Spar- row nest in southwestern Florida in 2002 and four re- cords are of unpublished data from sparrow nests in central Arkansas during 1983-1985. These observa- tions suggest that Bachman’s Sparrow can successfully raise cowbird young. Ours is also the first record of a parasitized Bachman’s Sparrow nest in Florida. Re- ceived 5 October 2007. Accepted 16 January 2008. Bachman’s Sparrow {Aimophila aestivalis) is a resident of pine {Finns spp.) flatwoods, grasslands, and dry prairies of the southeast- ern United States. It ranges from southern Florida to southern Virginia and west to east- ern Texas (Dunning 1993). The range of Bachman’s Sparrow receded after an initial in- crease in distribution and abundance around the turn of the 20th century (Bent 1968), and populations have shown a >50% decrease during 1966-2006 (Sauer et al. 2006). Habitat loss and degradation are the most significant threats to Bachman’s Sparrow populations, but the expansion of the Brown-headed Cowbird {Molothrus ater) into the southeastern United States may negatively impact host populations that lack historical exposure to parasitism. Nests of Bachman’s Sparrow are infre- quently found to be parasitized by the Brown- headed Cowbird (Dunning 1993). Brown- headed Cowbirds are known to parasitize ' Department of Wildlife Ecology and Conservation, University of Florida, P. O. Box 1 10430, Gainesville, FL 32611, USA. - Department of Biology, University of Texas at Ty- ler, 3900 University Boulevard, Tyler, TX 75799, USA. Department of Biology, Washington and Jefferson College, 60 South Lincoln Street, Washington, PA 15301, USA. ‘‘Corresponding author; e-mail; aves@ut1.edu >220 species, yet more than a third of para- sitized species are not known to raise cowbird young (reviewed in Ortega 1998, Rasmussen and Sealy 2006). We found no instances in the literature of a Bachman’s Sparrow raising a cowbird to fledging or feeding a cowbird fledgling outside of the nest and only one where a Bachman’s Sparrow nest contained a cowbird nestling (Haggerty 1988). We present evidence for the first known instances of Brown-headed Cowbirds successfully fledging from nests of Bachman’s Sparrow. METHODS We monitored avian productivity at Myak- ka River State Park (MRSP) in southwest Florida (27° 12' N, 82° 15' W) from April to August 2000-2004. MRSP is 11,800 ha of primarily dry prairie and flatwood habitats characterized by an understory of herbaceous ground cover and saw palmetto {Serenoa re- pens) with a canopy of scattered longleaf pine {Finns palnstris), slash pine {F. elliottii), live oak {Qnercns virginiana), and sabal palm {Sa- bal pahnetto). Prescribed fire programs at MRSP were initiated in the mid-1970s to re- store natural characteristics of fire-dependent habitats, including those occupied by Bach- man’s Sparrows, which are year-round resi- dents. OBSERVATIONS We monitored 285 nests of 25 breeding bird species in MRSP from 2000 to 2004. Ten nests were from at least nine Bachman Spar- row pairs. One nest was discovered on 9 June 2002 when the female Hushed from the nest which was on the ground and hidden beneath two saw palmetto fronds. The nest contained three pinkish-white sparrow eggs and one speckled cowbird egg. The nest contents were photographed following discovery to confirm identification of the cowbird egg. We assume the egg was of a Brown-headed Cowbird and 626 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 3, September 2008 not a Shiny Cowbird {Molothrus bonariensis) which has not been confirmed breeding in North America. The nest was monitored every 2-4 days and on 21 June was found to contain three nest- lings (i.e., 1 cowbird, 2 sparrows) and one cracked sparrow egg. The cowbird nestling was slightly larger than the two sparrow nest- lings indicating that it likely hatched on 19- 20 June while the sparrow nestlings hatched on 20-21 June. The three nestlings were in apparent good health and begged on subse- quent nest-checks. All nestlings were large and well-feathered on 29 June with the cow- bird nestling being slightly larger than the sparrows. All birds appeared to be within 1-3 days of fledging from the nest. The nest was last checked on 2 July and was empty and showed no signs of disturbance. Two adult sparrows were chipping nearby and at least one sparrow young was heard calling from nearby vegetation. We estimate that all nest- lings fledged between 30 June and 1 July but the cowbird likely fledged earlier than the sparrow nestlings. We were not able to locate the cowbird fledgling outside of the nest, pos- sibly due to the presence of dense vegetation where fledglings could be hidden. We assume the cowbird fledged from the nest because there was no evidence that it or any of the sparrow young did not fledge. Haggerty (1988) monitored 66 Bachman’s Sparrow nests in Hot Spring County, Arkan- sas (34° 11' N, 92° 48' W) from May to Au- gust 1983-1985. Sites primarily consisted of shortleaf (Pinus echinada) and loblolly pine {P. taeda) plantations. With the author’s per- mission, we present observations of all four parasitized Bachman’s Sparrow nests, two of which likely fledged cowbird young. Nest #6 was found with one cowbird egg and three Bachman’s Sparrow eggs. All eggs hatched but the nest was depredated on day 5 of the nestling period. Nest #61 was found with three sparrow eggs and one cowbird egg. The cowbird egg and two sparrow eggs hatched. The nest was depredated on about day 7 of the nestling period. Nest #35 was found with a single cowbird egg and four Bachman’s Sparrow eggs. Three sparrow eggs disap- peared during the laying period and the first few days of incubation, one of which was found pierced on the ground near the nest. Both the remaining sparrow egg and cowbird egg hatched and both nestlings fledged on day 9 of the nestling period. They were not ob- served again in the adult territory. Nest #60 was found on day 4 or 5 of the nestling period based on the size of a recently dead sparrow nestling. The two nestlings remaining in the nest were both cowbirds of similar age. The fate of this nest is not known. DISCUSSION Reports of cowbird parasitism of Bach- man’s Sparrow nests are rare. Friedmann (1963) notes only three records of parasitism, one each from Kentucky, Missouri, and West Virginia. No cases of parasitism were reported in 34 Bachman’s Sparrow nests in Alabama (Tucker et al. 2006) and 56 nests monitored during three studies in Georgia and South Car- olina (reviewed by Kilgo and Moorman 2003). Perkins et al. (2003) did not report cowbird parasitism in any of the 40 nests of Bachman’s Sparrow they monitored in central Florida. Congeneric Botteri’s (Aimophila bot- terii). Rufous-crowned (A. ruficeps), and Cas- sin’s sparrows (A. cassinii) are also infre- quently parasitized (Webb and Bock 1993, Collins 1999, Dunning et al. 1999), while Five-striped (A. quinquestriata) and Rufous- winged sparrows (A. carpalis) have low to in- termediate parasitism rates (Groschupf 1992, Lowther et al. 1999). Rufous-winged, Rufous- crowned, and Five-striped sparrows each have been reported to successfully raise a cowbird or have been seen feeding a cowbird fledgling (Ohmart 1969, Mills et al. 1980, Miles 1986). Our observations represent the first evidence of Bachman’s Sparrow successfully fledging cowbird young. Furthermore, our records are the first to document successful hatching of cowbird eggs or nests found with cowbird nestlings. To our knowledge, this is also the first reported case of cowbird parasitism of Bachman’s Sparrow in Florida. ACKNOWLEDGMENTS We thank T. M. Haggerty for access to nest records from central Arkansas and permission to publish ob- servations of cowbird parasitism. We thank the man- agers and staff at Myakka River State Park and K. W. Outcalt (USDA Southern Research Station) for logis- tical support and access to field sites. Invaluable field assistance was provided by M. D. Alderman, G. S. Kaufmann, and E. S. Owens. We thank K. E. Sieving SHORT COMMUNICATIONS 627 for support and help with field protocols. This is Con- tribution Number 175 of the National Fire and Fire Surrogate Project (FFS) funded by the U.S. Joint Fire Science Program. Protocols were approved under Uni- versity of Florida lACUC #A530. LITERATURE CITED Bent, A. C. 1968. Life Histories of North American cardinals, grosbeaks, buntings, towhees, finches, sparrows, and allies. U.S. National Museum Bul- letin 237, Part 2. Collins, P. W. 1999. Rufous-crowned Sparrow {Aim- ophila ruficeps). The birds of North America. Number 472. Dunning, J. B. 1993. Bachman’s Sparrow (Aimophila aestivalis). The birds of North America. Number 38. Dunning Jr., J. B., R. K. Bowers Jr., S. J. Suter, and C. E. Bock. 1999. Cassin’s Sparrow (Aimophila cassinii). The birds of North America. Number 471. Friedmann, H. 1963. Host relations of the parasitic cowbirds. U.S. National Museum Bulletin 233. Groschupf, K. 1992. Five-striped Sparrow (Aimophila quinque striata). The birds of North America. Number 21. Haggerty, T. M. 1988. Aspects of the breeding biol- ogy and productivity of Bachman’s Sparrow in central Arkansas. Wilson Bulletin 100:247-255. Kilgo, j. C. and C. E. Moorman. 2003. Patterns of cowbird parasitism in the southern Atlantic Coast- al Plain and Piedmont. Wilson Bulletin 115:277- 284. Lowther, P. E., K. D. Groschupf, and S. M. Russell. 1999. Rufous-winged Sparrow (Aimophila car- palis). The birds of North America. Number 422. Miles, D. B. 1986. A record of Brown-headed Cow- bird (Molothrus ater) nest parasitism of Rufous- crowned Sparrows (Aimophila ruficeps). South- western Naturalist 51:253-254. Mills, S., J. Silliman, K. Groschupf, and S. Speich. 1980. Life history of the Five-striped Sparrow. Living Bird 18:95-110. Ohmart, R. D. 1969. Physiological and ethological ad- aptations of the Rufous-winged Sparrow (Aimo- phila carpalis) to a desert environment. Disserta- tion. University of Arizona, Tucson, USA. Ortega, C. P. 1998. Cowbirds and other brood para- sites. University of Arizona Press, Tucson, USA. Perkins, D. W, P. D. Vickery, and W. G. Shriver. 2003. Spatial dynamics of source-sink habitats: ef- fects on rare grassland birds. Journal of Wildlife Management 67:588-599. Rasmussen, J. L. and S. G. Sealy. 2006. Hosts feed- ing only Brown-headed Cowbird fledglings: where are the host fledglings? Journal of Field Or- nithology 77:269-279. Sauer, J. R., J. E. Hines, and J. Fallon. 2006. The North American Breeding Bird Survey, results and analysis 1966-2005. Version 6.2.2006. USGS, Patuxent Wildlife Research Center, Laurel, Mary- land, USA. http://www.mbr-pwrc.usgs.gov/bbs/ bbs.html (accessed 1 1 September 2007). Tucker Jr., J. W, W. D. Robinson, and J. B. Grand. 2006. Breeding productivity of Bachman’s Spar- rows in fire-managed longleaf pine forests. Wilson Journal of Ornithology 118:131-137. Webb, E. A. and C. E. Bock. 1993. Botteri’s Sparrow (Aimophila botterii). The birds of North America. Number 216. 628 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 The Wilson Journal of Ornithology 120(3):628-630, 2008 Simultaneous Incubation by Two Females and Nestling Provisioning by Four Adults at a Savannah Sparrow Nest Nathan J. Zalik^ and Noah G. Perlut^’^’^ ABSTRACT — We present the first observations of misdirected parental care by Savannah Sparrows {Pas- serculus sandwichensis) including a rare occurrence of simultaneous incubation. Two females simultaneously incubated eggs, brooded, and fed nestlings, and two males fed nestlings in one nest. These behaviors may have been prompted by strong parental instincts in combination with a stressful breeding environment me- diated by hayfield management, as any genetic benefits were unlikely. Received 1 September 2007. Accepted 24 October 2007. Cooperative breeding, in which care for young is provided by individuals other than the breeding pair, is estimated to occur in only 9% of avian species (Cockburn 2006). Co- operative breeding is thought to evolve when the additional help allows for greater produc- tion of young or increased survival of adults through reduced breeding effort (Brown 1978, Crick 1992). Helpers may receive resource benefits from being a member of a group or may enhance future breeding opportunities. Furthermore, helpers that are closely related to the breeders they assist gain indirect genetic benefits from raising kin (Hamilton 1963, 1964). Alternatively, Jamieson (1986, 1991) hypothesized that helping is a behavioral re- sponse to the presence of begging young. This “misdirected parental care” (Price et al. 1983: 192) may explain rare cases of feeding or in- cubation by individuals which do not appear to receive resource or genetic benefits. Avian incubation is usually performed by only one individual at any given time regard- less of the parental care system. Incubation and brooding by more than one bird simulta- ' The Rubenstein School of Environment and Nat- ural Resources, 81 Carrigan Drive, University of Ver- mont, Burlington, VT 05405, USA. 2 Vermont Cooperative Fish and Wildlife Research Unit, 81 Carrigan Drive, University of Vermont, Bur- lington, VT 05405, USA. 2 Corresponding author; e-mail: nperlut@uvm.edu neously has rarely been documented for tem- perate species (Forbush 1929, Howell 1942, Bellrose 1943, Hawksley and McCormack 1951, Brackbill 1952, Fuller and Bolen 1963). We present observations of misdirected paren- tal care by the Savannah Sparrow {Passercu- lus sandwichensis), a ground-nesting songbird with a mixed-mating strategy and biparental care (Wheelwright and Rising 1993). We de- scribe simultaneous incubation, brooding, and provisioning of young by two females, and provisioning young by two males. METHODS We captured and uniquely banded (U.S. Geological Survey [USGS] metal band and three colored leg bands) all breeding adult Sa- vannah Sparrows in 2002—2007 on a 10.5-ha managed hayfield (44° 39' N, 73° 27' W) in Shelburne, Vermont, USA. We located Savan- nah Sparrow nests {n = 515) and monitored them every 1-3 days until they fledged or failed, and banded nestlings (USGS band only) on post-hatch day 6-7 (Perlut et al. 2006). OBSERVATIONS On 12 July 2007 we found a nest containing four eggs being incubated by a female Savan- nah Sparrow with two orange bands on the left leg and a yellow and metal band on the right leg (hereafter referred to as the primary female). On the same day, we discovered an- other nest containing four eggs being incu- bated by a female with metal and blue bands on the left leg and orange and red bands on the right leg (hereafter referred to as the sec- ondary female). These nests, 73 m apart, were on adjacent territories. Both females were still incubating four eggs on 13 July. By 16 July, the secondary female’s nest had failed, while the primary female was still incubating four eggs. On 17 July, two birds flushed —0.5 m from each other from the area of the primary female’s nest. One of these birds was identi- SHORT COMMUNICATIONS 629 fied as the secondary female. The secondary female’s nest had recently failed and no fur- ther attempts to follow this bird were made as it was too soon for her to have renested. On 20 July, two birds flushed directly from the primary female’s nest. The rim of the nest was flattened on one side, making the nest cup atypically wide. Within minutes, both primary and secondary females returned to the nest. We flushed the birds two more times to re- confirm these observations, flushing from within 1 m of each other and directly from the nest together. We searched the area extensive- ly for a second nest, but none was located. On 20 July, at 1000 hrs EDT the nest had four eggs, one of which appeared damaged with a small indentation. We videotaped the nest us- ing a small (9X3X3 cm) wide angle lens at 1325 hrs, at which time the nest contained one nestling and three unhatched eggs. We made two recordings, one of 30 min and one of 17 min; both females were observed incubating si- multaneously as well as singly. They sat togeth- er on the nest for 1 1 min during one period with the primary female in a normal incubating pos- ture and the secondary female sitting partially on the back of the primary female and partially on the rim of the nest. Both birds were facing roughly the same direction. This simultaneous incubation ended when we approached the nest to remove the camera. On six occasions a bird forced itself underneath the other already incu- bating bird to begin incubating the nest itself. During one 22-sec period, this forceful ex- change occurred three times. We recorded (4.5 hrs) the nestling stage at post-hatch 4-5 days. Additional aggressive be- havior between the two females beyond the forceful incubation exchanges occurred. When a female was on the nest and the other female approached, it would often do so with one or both wings raised and give aggressive squealing and chattering calls. During one interaction, while the primary female was brooding, the sec- ondary female arrived at the nest with both wings raised and made a squealing vocalization. The two birds faced each other with bills agape lor 1 min as the primary female remained on the nest. The interaction ended when the pri- mary female pecked at the secondary female, causing the secondary female to retreat. These vocalizations and displays were notably differ- ent from those observed when one female and one male interacted at the nest. Despite these aggressive interactions, the primary female tol- erated the secondary female, as evidenced by simultaneous incubation, brooding, and provi- sioning of young. Two of the four eggs hatched and the two damaged or infertile eggs were removed by 23 July. Both females continued to brood and provision the nestlings. Two males also were observed provisioning nestlings. The male which frequently brought food was the social male associated with the primary female’s two previous nesting attempts. The second male was the social male associated with the sec- ondary female’s previous nesting attempts. This second male visited the nest and fed nest- lings only once during the recordings. The nestlings fledged on 29 July. DISCUSSION Our observations represent the first docu- mented case of misdirected parental care oc- curring naturally at a Savannah Sparrow nest. Weatherhead and Robertson (1978, 1980) de- scribe a case of helping induced artificially and a second case where clutches were laid simultaneously by two females in the same nest. Why did the secondary female (and her mate) help at the nest rather than initiate an- other nesting attempt? We have no reason to suspect intraspecific brood parasitism, as both females had active nests simultaneously, and the number of eggs in the primary female’s nest remained at four throughout the incuba- tion period. Paternity assignment of 109 broods in a related study revealed no evidence of egg dumping (N. G. Perlut, unpubl. data); Savannah Sparrows on Kent Island, New Brunswick show no evidence of egg dumping (C. R. Freeman-Gallant, pers. comm.). We do not know whether the two birds were related, as both were first banded as adults on 15 May 2007. It seems improbable the secondary fe- male had any direct genetic investment in the nest despite this uncertainty. We suspect the secondary female discovered the primary female’s nest while foraging and her maternal instinct was to incubate the eggs. Fe- males likely monitor the breeding status of their neighbors, as the field is densely populated (2 females/lia in 2(K)7) and individuals commonly interact. Other females at this site initiated re- nests around the date the nest of the secondary 630 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 female failed; some females have laid up to six clutches in a year, fledging as late as 24 August. Thus, there were sufficient resources and time with which to renest. The sight of eggs in a nest may have stimulated hormones that led to an incubation response. This effect was perhaps similar to the response of adults to nestlings of brood parasites, in which parental instincts and mistaken identity cause a maladaptive response to feed nestlings (Price et al. 1983). Similarly, the secondary male likely provisioned young simply because his mate was caring for the nest. Stress, mediated by the effects of hayfield management, likely contributed to this unusu- al behavior. The secondary female had at- tempted three previous nests. The field was cut twice during the nesting season, each cut- ting followed by manure application; all ac- tive nests failed as a result of haying. These sources of nest failure, in addition to failure from natural causes, may have placed these birds under high stress. It is possible these fac- tors contributed to the secondary female’s “decision” to attend this nest rather than ini- tiate another clutch. The secondary female and her mate assisted in incubation and in provisioning nestlings. However, whether these birds helped rather than hindered development of the young is de- batable. For instance, only two young reached the fledgling stage, whereas the average num- ber of fledged young from successful post- second harvest nests was 3.0 (n = 10) in 2007. Of the two young that survived, one nestling’s leg was splayed to the side. This could have been a genetic defect, but was more likely an injury sustained in the nest, possibly during an interaction between the two females. Our observations of simultaneous incuba- tion and provisioning of young by Savannah Sparrows are highly unusual and may have been prompted by misguided parental instincts in combination with a stressful breeding en- vironment. ACKNOWLEDGMENTS This project was supported by The Rubenstein School of Environment and Natural Resources, Uni- versity of Vermont, as well as the Initiative for Future Agricultural and Food Systems, and the National Re- search Initiative of the USD A, Cooperative State Re- search, Education and Extension Service, grant num- bers 2001-52103-11351 and 03-35101-13817, respec- tively. We thank Shelburne Farms and Sam Dixon for access to their land, and Allan Strong, James Rising, and Patrick Weatherhead for helpful comments on the manuscript. LITERATURE CITED Bellrose Jr., E 1943. Two Wood Ducks incubating in the same nesting box. Auk 60:446-447. Brackbill, H. 1952. A joint nesting of Cardinals and Song Sparrows. Auk 69:302-307. Brown, J. L. 1978. Avian communal breeding sys- tems. Annual Review of Ecology and Systematics 9:123-155. CocKBURN, A. 2006. Prevalence of different modes of parental care in birds. Proceedings of the Royal Society B-Biological Sciences 273:1375-1383. Crick, H. P. Q. 1992. Load-lightening in cooperatively breeding birds and the cost of reproduction. Ibis 134:56-61. Forbush, E. H. 1929. Birds of Massachusetts and other New England states. Volume 3. Berwick and Smith Company, Norwood, Massachusetts, USA. Fuller, R. W. and E. Bolen. 1963. Dual Wood Duck occupancy of a nest box. Wilson Bulletin 75:94-95. Hamilton, W. D. 1963. The evolution of altruistic be- havior. American Naturalist 97:354-356. Hamilton, W. D. 1964. The genetical evolution of so- cial behavior, I and II. Journal of Theoretical Bi- ology 7:1-52. Hawksley, O. and a. P. McCormack. 1951. Doubly occupied nests of the Eastern Cardinal Richmon- dena cardinalis. Auk 68:515—516. Howell, J. C. 1942. Notes on the nesting habits of the American Robin (Turdus migratorius L.). Amer- ican Midland Naturalist 28:529-603. Jamieson, I. G. 1986. The functional approach to be- havior: is it useful? American Naturalist 127:195- 208. Jamieson, I. G. 1991. The unselected hypothesis for the evolution of helping behavior: too much or too little emphasis on natural selection? American Naturalist 138:271-282. Perlut, N. G., a. M. Strong, T. M. Donovan, and N. J. Buckley. 2006. Grassland songbirds in a dynamic management landscape: behavioral re- sponses and management strategies. Ecological Applications 16:2235-2247. Price, T, S. Millington, and P. Grant. 1983. Helping at the nest in Darwin’s Finches as misdirected pa- rental care. Auk 100:192—194. Weatherhead, P. J. and R. J. Robertson. 1978. Intra- specific nest parasitism in the Savannah Sparrow. Auk 95:744-745. Weatherhead, P. J. and R. J. Robertson. 1980. Al- truism in the Savannah Sparrow? Behavioral Ecology and Sociobiology 6:185-186. Wheelwright, N. T. and J. D. Rising. 1993. Savannah Sparrow (Passerculus sandwichensis). The birds of North America. Number 45. SHORT COMMUNICATIONS 631 The Wilson Journal of Ornithology 120(3):631-632, 2008 Grey Heron (Ardea cine re a) Predation on the Aldabra White-throated Rail (Dryolimnas cuvieri aldabranus) Pierre A. Pistorius^ ABSTRACT — Grey Herons {Ardea cinerea) typi- cally prey upon fish and other aquatic organisms with occasional reports of predating small birds. I report observations of one or more Grey Herons predating on Aldabra White-throated Rails {Dryolimnas cuvieri al- dabranus) on Picard Island (Aldabra Atoll, Sey- chelles). Grey Herons were observed on two occasions having caught rails and, on another occasion, a heron was observed ambushing and pursuing rails with no success. Heron(s) were not observed ingesting the rails, as they flew or ran off with their prey, but it is likely the rails were consumed. These are the first ob- servations of any form of predation on the Aldabra White-throated Rail, which is one of the largest re- corded avian prey of the Grey Heron. Received 25 June 2007. Accepted 1 November 2007. The Grey Heron {Ardea cinerea), wide- spread throughout the Seychelles, is one of the few waders that breed on Aldabra Atoll (9° 24' S, 46° 20' E) (Feare and Watson 1984). The Aldabra White-throated Rail {Dryolimnas cuvieri aldabranus) is endemic to Aldabra and represents the last remaining flightless bird in the Indian Ocean. It is believed the White- throated Rail lost the ability to fly on Aldabra due to lack of predatory pressure on the atoll (Wanless 2003). There are wild cats {Felis sil- vestris) on the largest of the four islands (Grand Terre) making up the atoll; rails occur only on the three islands which are cat-free. Rails were reintroduced to Picard Island in 1999 (Wanless 2002, Wanless et al. 2002). Rails tend to flee on sight of a large bird (e.g., heron or crow), possibly an instinct retained from their distant past, and it has been noted there are no birds on Aldabra sufficiently large to kill adult Aldabra Rails (Wanless 2002). OBSERVATIONS Grey Herons were observed predating on White-throated Rails on 14 June 2007 al 1430 ' Seychelles Island Foundation, Mont Fleuri, P. O. Box 853, Victoria, Mahe, Seychelles; e-mail: ppistorius@zoology.up.ac.za hrs (GMT + 0400) and again on 21 June at 1615 hrs. The first event occurred on the beach and the second in an open area close to shrub vegetation on Picard Island when atten- tion was drawn to the events by loud vocali- zation by the rails. Observations were made with the naked eye from distances between 20 and 30 m. On the first occasion, an adult Grey Heron was seen with a rail of unknown age in its beak. When approached by the observer, it rushed into the shrub vegetation carrying the rail with it. The rail was still alive at last sight and it is not known how events pro- ceeded. On the second occasion, a Grey Heron was seen holding an adult rail with one foot and pecking vigorously at the still alive rail. It then flew with the rail in its beak. Further to these observations, on 25 June at 1030 hrs an adult Grey Heron was seen ambushing and pursuing a juvenile rail with no success. All three incidents reported occurred in the same area and it is likely the same individual heron was involved each time. DISCUSSION Grey Herons generally target aquatic prey, mostly fish, but have been reported to feed on small mammals, invertebrates, and occasion- ally on other birds (Marquiss and Leitch 1990, McCanch 2003, Jakubas and Mioduszewska 2005, Gwiazda and Amirowicz 2006, Wil- liams and Ward 2006). Avian prey of Grey Heron is usually relatively small, e.g.. Com- mon Moorhen {Gallinula chlorojnis) chicks, young Mallard {Anas platyrliynclios) and Tuft- ed Duck {Aythya fuligula) ducklings, and Cape Cormorant {Leucocarho capensis) chicks (Milstein et al. 1970, Marquiss and Leitch 1990, Williams and Ward 2006). At- tempts of hunting for larger birds have been reported occasionally. McCanch (2003) de- scribed a Grey Heron that choked to death while trying to ingest a Little Grebe {Tachy- haptus ruficollis). Adult Little Cirebes weigh 632 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 ~140 g (Ogilvie 2003) compared to Aldabra White-throated Rails that weigh ~180 g (Wanless 2003). During the time of the re- ported observations, juvenile Aldabra rails would have been about 6 months of age and of similar weight to the adults. The fate of the two rails that were caught (nor the heron) could not be ascertained. A rail with a broken wing was last seen in the study area during the period when the predation was observed, and may have been hunted by the heron. That a heron was observed spending several hours hunting for rails suggests rails have success- fully been captured and consumed, and form a rewarding target species. Population growth of Aldabra White-throat- ed Rails during the first 2 years after the in- troduction of rails to Picard Island was ex- ponential and it was estimated the population would reach carrying capacity between 2007 and 2009 (Wanless 2002). Recent observa- tions of large numbers of rails injured in ter- ritorial disputes and juveniles remaining in adult territories (pers. obs.) suggest that car- rying capacity may have been reached. Indi- vidual rails are also often observed feeding in the inter-tidal zone, behavior which is gener- ally uncommon in the Aldabra White-throated Rail (Penny and Diamond 1971). This in- crease in numbers and shift in habitat use sug- gest that rails are now more exposed to Grey Herons than in the past and could explain why the reported predation has not been observed before. The Grey Heron is the first recorded natural predator of the Aldabra White-throated Rail. As the last remaining flightless bird in the In- dian Ocean, the conservation status of these rails could necessitate intervention if the pop- ulation becomes threatened through increased Grey Heron predatory behavior. ACKNOWLEDGMENTS I am grateful to Adrian Skerret and two anonymous reviewers for comments that improved this manuscript. Seychelles Island Foundation provided logistical and financial support during the study. LITERATURE CITED Feare, C. J. and J. Watson. 1984. Occurrence of mi- grant birds in the Seychelles. Pages 559—574 in Biogeography and ecology of the Seychelles Is- lands (D. R. Stoddart, Editor). W. Junk Publishers, The Hague, Netherlands. Gwiazda, R. and a. Amirowicz. 2006. Selective for- aging of Grey Heron {Ardea cinerea) in relation to density and composition of the littoral fish com- munity in a submontane dam reservoir. Waterbirds 29:226-232. Jakubas, D. and a. Mioduszewska. 2005. Diet com- position and food consumption of the Grey Heron {Ardea cinerea) from breeding colonies in north- ern Poland. European Journal of Wildlife Re- search 51:191-198. Marquiss, M. and a. F. Leitch. 1990. The diet of Grey Herons Ardea cinerea breeding at Loch Le- ven, Scotland, and the importance of their preda- tion on ducklings. Ibis 132:535-549. McCanch, N. 2003. Grey Heron choking on Little Grebe. British Birds 96:86. Milstein, P. L., I. Prestt, and A. A. Bell. 1970. The breeding cycle of the Grey Heron. Ardea 58:171- 257. Ogilvie, M. 2003. Grebes of the world. Bruce Cole- man Books, Uxbridge, United Kingdom. Penny, M. J. and A. W. Diamond. 1971. The White- throated Rail, Dryolimnas cuvieri on Aldabra. Philosophical Transactions of the Royal Society of London, Series B, Biological Sciences 260:529- 548. Wanless, R. M. 2002. The reintroduction of the Al- dabra Rail Dryolimnas cuvieri aldabranus to Pi- card Island, Aldabra Atoll. Thesis. University of Cape Town, South Africa. Wanless, R. M. 2003. Can the Aldabra White-throated Rail {Dryolimnas cuvieri aldabranus) fly? Atoll Research Bulletin 508:1-9. Wanless, R. M., J. Cunningham, P. H. R. Hockey, J. Wanless, R. W. White, and R. Wisemen. 2002. The success of a soft-release reintroduction of the flightless Aldabra Rail {Dryolimnas cuvieri alda- branus) on Aldabra Atoll, Seychelles. Biological Conservation 107:203-210. Williams, A. J. and V. L. Ward. 2006. Sacred Ibis and Grey Heron predation of Cape Cormorant eggs and chicks; and a review of ciconiiform birds as seabird predators. Waterbirds 29:321-327. SHORT COMMUNICATIONS 633 The Wilson Journal of Ornithology 120(3):633-635, 2008 Purple Swamphens (Porphyrio porphyrio) Attempting to Prey upon Black Swan {Cygnus atratus) Eggs and Preying upon a Cygnet on an Urban Lake in Melbourne, Australia Shandiya Balasubramaniam^ and Patrick- Jean Guay^’^ ABSTRACT — We report the predation of Black Swan {Cygnus atratus) eggs and cygnets by adult Pur- ple Swamphens {Porphyrio porphyrio), a mostly-her- bivorous species of rail, in Melbourne, Victoria, Aus- tralia. Received 9 October 2007. Accepted 28 Novem- ber 2007. The Purple Swamphen {Porphyrio porphy- rio) is a large rail (males: 1,050 g; females: 850 g) whose range encompasses southern Af- rica, southern Asia, Papua New Guinea, Aus- tralia, and New Zealand (Marchant and Hig- gins 1993). Quantitative investigations into di- ets of Purple Swamphens in Australia and New Zealand demonstrated this species is al- most completely herbivorous (Carroll 1966, Norman and Mumford 1985). Animal material constitutes only a minor part of the diet and is mainly represented by insects and arachnids (Carroll 1966, Norman and Mumford 1985). However, they have been reported to eat fish, lizards, and birds (Wheeler 1949, Oliver 1974). They have also been observed preying on nestlings and adults of different passerine species (McKenzie 1967, Egan 1992, Fitzsi- mons 2003), as well as eggs (Binns 1953, Fitzgerald 1966, Brown 1997) and ducklings from different waterfowl species including Chestnut Teal {Anas castanea) (van Tets 1965) , Pacific Black Duck {A. superciliosa) (Nixon 1983, Bonser 1984), and domestic Muscovy Duck {Cairina moschata) (Lowe 1966) . We present the first record of Purple Swamphens taking Black Swan {Cygnus atra- tus) eggs and cygnets. OBSERVATIONS Observations were conducted during an in- vestigation of incubation behavior of Black ' Department of Zoology, University of Melbourne, Parkville, VIC 3010, Australia. ^Corresponding author; e-mail: p.guay@zoology.unimelb.edu.au Swans at Albert Park Lake (37° 50' S, 144° 58' E), an urban lake in the center of Mel- bourne, Victoria, Australia. We observed a Purple Swamphen’s opportunistic attempt to prey on a Black Swan egg on 7 September 2007 at 1230 hrs EST during an agonistic en- counter between an incubating male and an- other adult swan. The male swan was incu- bating the eggs at the nest when another adult swan came within close range (~1 m) of the nest. The incubating male swan stood up in a threat display, raising its wings. It then pro- ceeded to swim beside the intruding swan while displaying its curled feathers. This es- calated into a chase as the incubating male bit the other adult’s tail and pursued it for a dis- tance of —100 m. As soon as the incubating male swan had left, three swamphens gathered around the nest. One of them perched on the rim of the bowl and started pecking at one of the four eggs in the nest. This continued for 30 sec after which time the male swan returned to the nest and the swamphens scattered. The swan stood over the eggs and preened its feathers for 10 min, and then rearranged some of the nesting material within the nest. This partic- ular incident was unsuccessful. However, we have observed Purple Swamphens feeding on swan eggs on three occasions and 38% of Black Swan nests in this population lost at least one egg to swamphens or other predators during the 2007 breeding season. We witnessed the predation of a newly- hatched Black Swan cygnet by a Purple Swamphen on 4 September 2007. At 1500 hrs, a pair of breeding swans at their nest started hissing and flapping their wings; moments lat- er we observed a Purple Swamphen running from the nest with a cygnet in its beak. Nei- ther adult swan left the nest to pursue the swamphen, which ran under a nearby swamp 634 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 paperbark {Melaleuca ericifolia) shrub. The swamphen shook the cygnet vigorously by the neck and then proceeded to tear at the cygnet’s bill while holding its body down with one foot. The swamphen may then have been alarmed by our presence because it stopped shaking the cygnet and stood still for 1 min. During this time the cygnet was trying feebly to move away but the swamphen did not ap- pear to take any notice of it. The swamphen then grabbed the cygnet by the neck and ran from us through the vegetation. It remained out of our sight for 3 min and then returned to the swamp paperbark shrub. It started pick- ing at the dead cygnet’s neck and breast area. After 5 min the swamphen again ran away, this time leaving the carcass behind. Upon ex- amining the cygnet carcass, we noted that it had been decapitated. A cursory search of the nearby area did not reveal the location of the head. A few moments later, after we had moved away, we noticed the swamphen run- ning back and forth carrying small pieces of cygnet between the carcass and a grass tus- sock. This continued for —15 min, after which we lost sight of the swamphen. We presumed, but did not confirm, the tussock was a Purple Swamphen nest. DISCUSSION These observations constitute the first re- port of predation by Purple Swamphens on Black Swan eggs and cygnets. This cygnet predation event is similar to the descriptions of swamphen predation on Pacific Black Duck ducklings by Nixon (1983) and Bonser (1984). Gizzard content analyses suggest that Purple Swamphens are herbivorous (Carroll 1966, Norman and Mumford 1985), but giz- zard analyses tend to underestimate the animal component (Swanson and Bartonek 1970, Briggs et al. 1985) due to the relatively rapid degradation of soft animal tissues in compar- ison to fibrous plant material and seeds (e.g., van Koersveld 1950). Egg yolk is unlikely to be detected in gizzard contents (Fitzgerald 1966). The methodology used by Carroll (1966) and Norman and Mumford (1985) may have underestimated the animal content in the diet of Purple Swamphens, but animal mate- rial is unlikely to represent a large part of the adult diet as it constituted less than 1 % of the total food content in the gizzard analyses. Consumption of the prey was not observed (e.g., van Tets 1965) in most cases of preda- tion on vertebrates by adult swamphens. Pur- ple Swamphens are known to feed their chicks (Craig 1980) and adults have been observed feeding young pieces of flesh from frogs (Brown and Brown 1977) and European Star- lings (Sturnus vulgaris) (Egan 1992). We did not observe the swamphen feeding chicks, but its behavior suggests that it was bringing food to another individual. Juvenile Purple Swam- phens, similar to juvenile waterfowl (e.g., Sugden 1973), may be more carnivorous than adults due to higher protein requirements to meet the demands of growth and develop- ment. While adult Purple Swamphens are largely herbivorous, it appears they can op- portunistically take advantage of new animal food sources as they become available to feed their young. Underestimation of the propor- tion of animal material in swamphen diets is likely due to a lack of juvenile samples and poor representation of animal material in giz- zard contents. ACKNOWLEDGMENTS This study was supported by an ARC Discovery Grant DP0558607 awarded to R. A. Mulder and M. A. Elgar, and the David Hay Scholarship to Patrick- Jean Guay. LITERATURE CITED Binns, G. 1953. Birds of Terang, south-western Vic- toria. Emu 53:211-221. Bonser, C. M. 1984. Purple Swamphen vs duck. Bird Observer 630:70. Briggs, S. V., M. T. Maher, and R. P. Plamer. 1985. Bias in food habits of Australian waterfowl. Aus- tralian Wildlife Research 12:507-514. Brown, M. 1997. Birds of the wetlands: Purple Swam- phen. Bird Observer 774:9-10. Brown, R. J. and M. N. Brown. 1977. Observations on swamphens breeding near Manjimup, W.A. Corella 1:82-83. Carroll, A. I. K. 1966. Food habits of the Pukeko {Porphyrio melanotus Temminck). Notornis 13: 133-144. Craig, J. L. 1980. Pair and group breeding behaviour of a communal gallinule, the Pukeko, Porphyriop. melanotus. Animal Behaviour 28:593-603. Egan, K. 1992. Purple Swamphen’s predation on Eu- ropean Starling. Australian Birds 26:84. Fitzgerald, M. F. 1966. Letter. Notornis 13:222. Fitzsimons, j. a. 2003. Purple Swamphen Porphyrio porphyrio killing a Noisy Miner Manorina melan- ocephala nestling. Corella 27:90. SHORT COMMUNICATIONS 635 Lowe, V. T 1966. Notes on the Musk Duck. Emu 65: 279-290. Marchant, S. and R J. Higgins. 1993. Handbook of Australian, New Zealand and Antarctic birds. Vol- ume 2. Raptors to Lapwings. Oxford University Press, Melbourne, Australia. McKenzie, H. R. 1967. Foods of the Pukeko. Notornis 14:41-42. Nixon, C. 1983. Ducked, by a Purple Swamphen! Bird Observer 623:1 14. Norman, F. I. and L. Mumford. 1985. Studies on the Purple Swamphen, Porphyria porphyrio, in Vic- toria. Australian Wildlife Research 12:263-278. Oliver, W. R. B. 1974. New Zealand birds. Reed, Wel- lington, New Zealand. SuGDEN, L. G. 1973. Feeding ecology of Pintail, Gad- wall, American Wigeon and Lesser Scaup duck- lings in Southern Alberta. Canadian Wildlife Ser- vice Report Series 24:1-43. Swanson, G. A. and J. C. Bartonek. 1970. Bias as- sociated with food analysis in gizzards of Blue- winged Teal. Journal of Wildlife Management 34: 739-746. VAN Koersveld, E. 1950. Difficulties in stomach anal- ysis. Proceedings of the International Ornitholog- ical Congress 10:592-594. VAN Tets, G. F. 1965. Eastern Swamphen takes a downy from a pair of Chestnut Teal. Emu 64:100. Wheeler, R. 1949. Nature notes. Bird Observer’s Club Monthly Notes 212(July 1949):!. The Wilson Journal of Ornithology 120(3):635-636, 2008 Northern Fulmar Predation of Common Murre Stephan Lorenz^-^ and Sampath Seneveratne' ABSTRACT. — We observed a Northern Fulmar (Fulmarus glacialis) prey upon a live Common Murre {Uria aalge) off Cape Spear, Newfoundland, Canada on 4 November 2007. Active predation of Northern Fulmars on other seabirds has not previously been re- ported. Received 19 November 2007. Accepted 16 Jan- uary 2008. Northern Fulmars {Fulmarus glacialis) are widespread throughout the northern Atlantic and Pacific oceans and populations have in- creased in Atlantic Canada (Stenhouse and Montevecchi 1999). The diet of Northern Ful- mars includes a wide range of fish, squid, co- pepods, amphipods and, to a lesser extent, polychaetes, pteropods, and cnidarians (Cam- phuysen and van Franeker 1996, Hatch and Nettleship 1998, Phillips et al. 1999). In ad- dition, Northern Fulmars also feed on offal and carrion, mainly fish refuse and remains of marine mammals (Fisher 1952, Hobson and Welch 1992, Camphuysen and Garthe 1997). They forage by dipping, surface seizing, sur- face plunging (Hatch and Nettleship 1998, Brooke 2004), and pursuit diving (Hobson and Welch 1992, Garthe and Furness 2001, ' Department of Biology, Memorial University of Newfoundland, St. John’s, NL AIB 3X9, Canada. ^Corresponding author; e-mail: slorenz@mail.com Brooke 2004). Fulmars have been observed scavenging carcasses of birds, including can- nibalism (Fisher 1952). Fisher (1952) also cites indirect evidence of fulmars eating wounded birds, direct observations of attack- ing gulls (Laridae), and one captive fulmar killing, but not consuming a European Green- finch (Carduelis chloris). OBSERVATIONS We report a previously undocumented for- aging behavior and prey item of the Northern Fulmar. On 4 November 2007 while watching seabird congregations off Cape Spear, New- foundland (47° 31' 25" N, 52° 37' 13" W) we observed a fulmar kill and feed on a Common Murre {Uria aalge). At 0930 hrs (NST) after ~2 hrs of continued observation of seabirds passing close to land, consisting of large num- bers of Dovekies {Alle a He) and smaller num- bers of Common and Thick-billed (U. lotuvia) murres. Razorbills {Alca tarda), and Black- legged Kittiwakes (Rissa tridactyla), we ob- served a light phase fulmar actively pursuing a winter plumage Common Murre. Both birds were ~50 m offshore and were observed through binoculars and a spotting scope. Ini- tially the fulmar was .seen biting the murre's neck while both birds were sitting on the wa- 636 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 ter. As the struggle continued the fulmar per- sistently clasped the murre by the bill and neck, apparently in an attempt to drown it. The murre freed itself several times only to dive for a few seconds and being attacked on resurfacing. The entire struggle lasted 15 min until the murre had succumbed, floating on the surface. The fulmar proceeded to feed on the dead bird, plucking at the vent area. Great Black-backed Gulls {Lams marinus) harassed the feeding fulmar, which actively and suc- cessfully defended its prey using aggressive postures. We continued to observe the fulmar for another 30 min and it was still feeding on and defending the murre carcass when we left. DISCUSSION We found no previous reports of Northern Fulmar actively preying on seabirds or pro- cellariid species preying upon alcids. Only the Southern (Macronectes giganteus) and North- ern Giant (M. halli) petrels within the family Procellariidae regularly prey on seabirds, mainly penguins (Spheniscidae) at breeding colonies (Brooke 2004). Fisher (1952) reports Northern Fulmars scavenging Rhinoceros Auklet {Cerorhinca monocerata) and Puffin (Fratercula spp.). Documented predators of adult Common and Thick-billed murres, and the related Razorbill include a wide range of large raptors and mammals, especially foxes {Alopex lagopus and Vulpes vulpes) and, at times, polar bears {Ursus maritimus) (Gaston and Hipfner 2000, Ainley et al. 2002, Hipfner and Chapdelaine 2002). Common Murres ap- pear to be high-risk prey items for fulmars (body mass 600-700 g) due to their large body size (>800 g) and powerful beak. We conclude that fulmar predation of sympatric seabirds is a rare event, in this case possibly triggered by unfavorable weather leading to decreased feeding or a purely opportunistic event due to a sick or exhausted bird. We were unable to judge the condition of the murre, but believe that the bird was relatively healthy due to the vigorous defense and prolonged strug- gle. Common Murres are one of the largest Atlantic alcids; thus. Northern Fulmars could potentially prey on other seabirds, especially many of the smaller alcids. ACKNOWLEDGMENTS We thank C. E. Braun and two anonymous review- ers who provided useful comments on this manuscript. LITERATURE CITED Ainley, D. G., D. N. Nettleship, H. R. Carter, and A. E. Storey. 2002. Common Murre {Uria aal- ge). The birds of North America. Number 666. Brooke, M. 2004. Albatrosses and petrels across the world. Oxford University Press, New York, USA. Camphuysen, C. J. and S. Garthe. 1997. Distribution and scavenging habits of Northern Fulmars in the North Sea. Journal of Marine Science 54:654- 683. Camphuysen, C. J. and J. A. van Franeker. 1996. Jellyfish and fishery waste as food sources of Northern Fulmars Fulmariis glacialis feeding around St. Kilda. Sula 10:143—150. Fisher, J. 1952. The fulmar. Collins, London, England. Garthe, S. and R. W. Furness. 2001. Frequent shal- low diving by a Northern Fulmar at Shetland. Wa- terbirds 24:287-289. Gaston, A. J. and J. M. Hipfner. 2000. Thick-billed Murre (Uria lomvia). The birds of North America. Number 497. Hatch, S. A. and D. N. Nettleship. 1998. Northern Fulmar {Fulmariis glacialis). The birds of North America. Number 361. Hipfner, J. M. and G. Chapdelaine. 2002. Razorbill {Alca torda). The birds of North America. Num- ber 635. Hobson, K. A. and H. E. Welch. 1992. Observations of foraging Northern Fulmars {Fulmariis glaci- alis) in the Canadian high arctic. Arctic 45:150- 153. Phillips, R. A., M. K. Petersen, K. Lilliendahl, K. SOLMUNDSSON, K. C. HaMER, C. J. CAMPHUYSEN, AND B. ZONFRILLO. 1999. Diet of the Northern Ful- mar Fulmariis glacialis'. reliance on commercial fisheries? Marine Biology 135:159—170. Stenhouse, I. J. AND W. A. Montevecchi. 1999. In- creasing and expanding populations of breeding Northern Fulmars in Atlantic Canada. Waterbirds 22:382-391. SHORT COMMUNICATIONS 637 The Wilson Journal of Ornithology 1 20(3):637— 640, 2008 Diet of Nestling Black-crowned Night-herons in a Mixed Species Colony: Implications for Tern Conservation C. Scott and Stephen W. Kress^-^ ABSTRACT — Boluses were collected from Black- crowned Night-heron (Nycticorax nycticorax) nestlings in 1992 to examine the impact of night-heron preda- tion on a restored tern colony. Boluses (n = 101) were collected from 18 nests. Fish remains occurred in 89% of nests, sand shrimp (Crangon septemspinosa) in 50%, birds in 28%, and amphibians in 16% of nests sampled; mammalian, eel, squid, and marine inverte- brate remains were also noted. Regurgitated bird re- mains were found in five nests and included four spe- cies, Common Terns {Sterna hirundo). Common Ei- ders (Somateria mollissima). Gulls (Larus sp.), and the legs of an unknown wading bird. Nestling night-herons from three nests were fed tern chicks, but 92% of tern chicks known to have been eaten were fed to nestling Black-crowned Night-herons in one nest. No tern chicks fledged in 1992 and night-herons were observed in the tern colony on multiple occasions. This study suggests that individual night-herons will specialize on waterbird prey. The subsequent removal of a specialist night-heron predator resulted in improved tern produc- tivity. Received 16 February 2007. Accepted 16 Oc- tober 2007. Black-crowned Night-herons {Nycticorax nycticorax) are a generalist predator with a varied diet, which includes fish, mollusks, crustaceans, mammals, amphibians, reptiles, invertebrates, vegetation, and bird eggs and chicks (Marshall 1942, Palmer 1962, Collins 1970, Wolford and Boag 1971, Szlivka 1985, Yen 1991, Davis 1993). They are reported to feed by sight (Watmough 1978) and sound (Hunter and Morris 1976), and will take ad- vantage of temporary abundances of food (Davis 1993). Food preference is reported to ‘ National Audubon Society, Seabird Restoration Program, 12 Audubon Road, Bremen, ME 04551, USA. ^Current address: National Audubon Society, Sea- bird Restoration Program, 159 Sapsucker Woods Road, Ithaca, NY 14850, USA. Current address: National Audubon Society, Sea- bird Restoration Program, 118 High Street, Suite 2, #26, Belfast, ME 04915, USA. Corresponding author; e-mail: shall («huidubon.org vary between colonies and years (Collins 1970). Night-heron predation in colonial wa- terbird colonies has been widely noted (Beck- ett 1964, Collins 1970, Hunter and Morris 1976, Nisbet and Welton 1984) and negative impacts have been documented on nesting co- lonial seabirds (Hunter and Morris 1976, Shealer and Kress 1991). Stratton Island is a diverse waterbird colony in Maine that provides habitat for both a re- stored nesting colony of terns (Arctic [Sterna paradisaea]. Common [S. hirundo]. Least [5. antillarum], and Roseate [S. dougallii]) (Kress [1983] provides a discussion of tern restora- tion techniques) and nesting wading birds in- cluding Black-crowned Night-herons. Noctur- nal losses of tern eggs and chicks were first observed in 1989, 3 years after tern restoration was initiated. Tern reproductive success was poor in 1989 (0.07 chicks/pair; Shealer 1989) and 1991 (0.0 chicks/pair; Hedges and Whi- taker 1991), and good in 1990 (1.5 chicks/ pair; Skinner 1990) when productivity moni- toring only occurred in an area where startle lights were used to deter night-heron preda- tion. Overall, recruitment to the tern colony slowed during this period and birds were abandoning the colony at night and between seasons. Observations of terns mobbing adult and juvenile night-herons in or adjacent to the tern colony were recorded from 1 989 to 1 992, but night-herons were only observed taking tern chicks and eggs in 1992 (Benz and D’Angelo 1992). The objectives of our study were to: ( 1 ) quantify Black-crowned Night-heron nestling diet, (2) identify the number of night-heron pairs that were feeding terns to their young, and (3) examine the impact of night-heron predation on a tern colony. METHODS Study Area. — The study was conducted in 1992 on Stratton Island (43°31'N, 70° 19' 638 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 TABLE 1. Black-crowned Night-heron nestling diet, Stratton Island, Maine, 1992. Frequency equals the number of nests containing the designated prey type and not the total number of individuals recovered for each prey type. Nests (n = 18) Frequency Prey item of nests (%) Bird 5 (27.7) Sterna sp. 3 (16.7) Larus sp. 3 (16.7) Somateria moUissima 3 (16.7) Unk - ardeid 2 (11.1) Unk - bird 1 (5.6) Fish 16 (88.9) Unknown fish 16 (88.9) Alosa pseudoharangus 1 (5.6) Catostomus sp. 1 (5.6) Funduliis majalis 1 (5.6) Unk - Catfish 1 (5.6) Sand Shrimp (Crangon septemspinosa) 9 (50.0) Amphibian 3 (16.7) Rana clamitans 2 (11.1) Rana sp. 1 (5.6) Eel 1 (5.6) Mammal 1 (5.6) Tarnais striatus 1 (5.6) Invertebrate 2 (11.1) Squid 1 (5.6) W), 3 km east of Prout’s Neck, York County, Maine. The National Audubon Society’s Sea- bird Restoration Program monitors and man- ages this mixed species, 12.2-ha inshore is- land. Stratton Island has been Maine’s most diverse waterbird colony with 21 nesting spe- cies; in 1992, the wading bird colony included —250 pairs of Snowy Egret {Egretta thula). Little Blue Heron (E. caerulea). Tri-colored Heron {E. tricolor). Cattle Egret {Bubulcus ibis). Glossy Ibis {Plegadis falcinellus), and Black-crowned Night-heron. Bolus Collection and Processing. — Night- heron nests were flagged during the annual wading bird census for subsequent visits. We collected spontaneous regurgitations from captured nestlings and obtained additional bo- luses by gently massaging the neck of cap- tured nestlings. Samples were labeled, placed in plastic bags and preserved in a 70% ethanol solution for future processing. Samples were not collected from every nest on each visit to reduce disturbance. The colony contained 53 night-heron pairs and 1 8 of the nests were included in the study. Nest inclusion was dependent on hatching fate, chick survival, and our ability to sample nestlings (nests at the top or on outer branches of a nesting tree could not be sampled). Nests were visited 5-14 times between 11 June and 7 August while tern chicks were present on the island. One to 33 samples were collected from each nest (mode = 4 samples/nest). Samples were identified in January 1993 to the most precise taxa possible by initially sort- ing bolus contents by size and taxa. The con- tents were screened by passing water over the boluses and through a hardware cloth sieve. Voucher specimens were preserved in glass jars with 70% ethanol. All bird remains and fish scales (mounted on slides) were washed (feathers were dried) before examination un- der a stereo dissecting scope. Unidentified fish samples were submitted to the Maine Depart- ment of Marine Resources for identification. Tern Monitoring. — An annual nest census was conducted by walking through the tern colony and marking all nests following the procedures of Kress and Hall (2004). The count was timed to coincide with the maxi- mum number of nests prior to hatch. Two hun- dred and twenty-one pairs of Common Terns, seven pairs of Roseate Terns, and five pairs of Arctic Terns nested in 1992. Productivity (fledglings produced/pair) was measured from a sample of nests {n = 25) in five enclosures, which were enclosed to permit recapture of chicks. Enclosures measured 7.62 X 3.66 m and were constructed from chicken wire wrapped with landscape fabric. Enclosure lo- cations were random and constructed in tern nesting habitat prior to clutch initiation. All chicks were banded within 1 day of hatching with a uses band and followed through fledging. RESULTS One hundred and one boluses were collect- ed from Black-crowned Night-heron nest- lings; 58% contained bird remains, 33% fish, and 16% sand shrimp. However, sixty-two bo- luses were collected from just three nests. Ad- ditional prey items (small proportions) includ- ed eel, squid, amphibians, mammals, and in- sects. Fish were found at 89% of the sampled nests, sand shrimp at 50%, and bird remains at 28% {n = 5). Fifty percent (39 total sam- SHORT COMMUNICATIONS 639 pies) of the bird remains were tern chicks (most were S. hirundo), 24% (19 total sam- ples) were Common Eider ducklings, and 22% (18 total samples) were gull chicks (either L. argentatus or L. marinus). Most bird remains {91%) were found at three nests and 92% of the tern chick remains were in one nest in- cluding 21 of 22 recovered tern bands from a total available sample of 66 banded Common Tern chicks. Remains of gull chicks and eider ducklings were found in three nests, although 70% of gull chicks were found at one nest (different than the nest with tern chicks) and 53% of eider ducklings were found in another nest. No tern chicks fledged from enclosures in 1992. Sixty-five eggs were laid in 25 enclo- sure nests, 26 disappeared (6 were pipping), 12 were abandoned, 2 were depredated, and 25 hatched. All hatched chicks disappeared before they reached 8 days of age; 18 disap- peared by the age of 3 days. We assumed all chicks that disappeared were predated as no remains were found within or outside enclo- sures and night-herons were observed feeding in the colony. DISCUSSION Black crowned Night-heron nestlings on Stratton Island were fed a diverse diet con- sisting of marine, tidal/estuary, freshwater, and upland-derived vertebrate and inverte- brate prey. Fish (4+ species), shrimp, and birds (4+ species) were the dominant prey by frequency in boluses and at nests. Unlike oth- er studies (Szlivka 1985, Yen 1991), amphib- ians occurred in less than 5% of boluses, sug- gesting that night-herons nesting on Stratton Island forage primarily in marine or estuarine habitats, rather than freshwater habitat. Most food habit studies or summaries of Black- crowned Night-heron diets have reported that fish were the principal food (Collins 1970, Hoffman 1978, Szlivka 1985, Yen 1991, Da- vis 1993, Riehl 2006) but, in this study, while fish occurred in 89% of nests, shrimp and bird remains occurred in greater numbers and fre- quency in boluses. Bird remains have been re- ported in many other food studies of night- herons (Collins 1970, Wolford and Boag 1971, Hunter and Morris 1976, Yen 1991). The presence of wading bird remains in night- heron boluses is also consistent with other studies (Beckett 1964, Wolford and Boag 1971, Riehl 2006). Although bird remains were prevalent, nest- lings at only five of 1 8 nests were provisioned with birds and, of these, 97% of bird remains were at three nests. Parents at only one of these three nests provisioned their young ex- clusively with birds and this pair concentrated on gulls, but also delivered eider ducklings and tern chicks. Tern chicks were observed at all three nests with bird remains, but 92% of tern remains (36 of 39 total tern remains) were concentrated at one nest. However, 49% of the diet at this nest was comprised of sand shrimp (and 1 green frog [Rana sp.]); 51% were bird remains (11% gulls, 13% eiders, 76% terns) suggesting that each adult in this pair had spe- cific food or foraging habitat preferences. Thus, night-heron predation in this tern colony appeared to be a specialized feeding behavior with just one adult consistently using this food resource. Despite the low frequency of this special- ized feeding behavior by night-herons, the consequences for nesting terns on Stratton Is- land were pronounced. No tern chicks fledged from study nests in 1992. Direct predation (in- ferred from predated eggs, missing young chicks, and from direct observation of preda- tion with night vision binoculars) by night- herons was the principal cause of mortality in 1992. Several other studies have noted signif- icant impacts of night-heron predation on tern reproductive success (Collins 1970, Hunter and Morris 1976, Nisbet and Welton 1984, Shealer and Kress 1991). A single adult Black-crowned Night-heron was shot on 5 July 1994 while feeding on tern chicks. No additional predation by night-her- ons was noted in the tern colony in 1994. Night-heron predation was not observed again on Stratton Island until 2002. In response to the removal of one night-heron, tern produc- tivity increased to 1.9 chicks/nesting pair in 1994. The average tern productivity during years of night-heron predation (1989-1993, 2002) was 0.42 chicks/pair and, with no ob- served night-heron predation, it was 1.6 chicks/pair from 1994 to 2001. This suggests that tiight-heron predation was a major factor in affecting tern nesting success on Stratton Island. Night-heron predation can have significant 640 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 impacts on nesting waterbirds and the exploi- tation of tern chicks is likely a specialized feeding behavior. The subsequent removal of an individual night-heron and corresponding positive breeding response by terns suggests that removal of specialist night-heron preda- tors can reduce impacts to nesting terns. ACKNOWLEDGMENTS Funding was provided by grants and donations to the National Audubon Society Seabird Restoration Program and by the Front’s Neck Chapter of the Na- tional Audubon Society. We thank Stephanie Schmidt of the Manomet Center for Conservation Sciences for comments on the manuscript. We are grateful to Char- lie and Sally Lee for their generous hospitality. We thank the field teams for vigilantly collecting regurgi- tated samples. We are especially indebted to Eric Kershner for processing samples and preparing a report of the bolus collection effort. We also thank Jean Chenoweth from the Maine Department of Marine Re- sources Laboratory in West Boothbay for assistance with fish identification and Charles Dardia of the Cor- nell Vertebrate Collections for assistance and labora- tory space. LITERATURE CITED Beckett, T. A. 1964. Black-crowned Night-heron feeding behavior. Chat 29:93-94. Benz, S. and P. D’Angelo. 1992. Stratton and Bluff islands. The National Audubon Society’s Phineas W. Sprague Memorial Sanctuary, Saco Bay, Maine. 1992 Season Report. National Audubon Society Seabird Restoration Program, Ithaca, New York, USA. Collins, C. T. 1970. The Black-crowned Night-heron as a predator of tern chicks. Auk 87:584—586. Davis, W. E. 1993. Black-crowned Night-heron. The birds of North America. Number 74. Hedges, S. and H. Whitaker. 1991. Stratton and Bluff islands. The National Audubon Society’s Phineas W. Sprague Memorial Sanctuary, Saco Bay, Maine. 1991 Season Report. National Audubon Society Seabird Restoration Program, Ithaca, New York, USA. Hoffman, R. D. 1978. The diets of herons and egrets in southwestern Lake Erie. Pages 365-369 in Wading birds (A. Sprunt IV, J. C. Ogden, and S. Winkler, Editors). Research Report Number 7. Na- tional Audubon Society, New York, USA. Hunter, R. A. and R. Morris. 1976. Nocturnal pre- dation by a Black-crowned Night-heron at a Com- mon Tern colony. Auk 93:629-633. Kress, S. W. 1983. The use of decoys, sound record- ings, and gull control for re-establishing a tern colony in Maine. Colonial Waterbirds 6:185-196. Kress, S. W. and C. S. Hall. 2004. Tern management handbook: coastal northeastern United States and Atlantic Canada. USDI, Fish and Wildlife Service, Hadley, Massachusetts, USA. Marshall, N. 1942. Night desertion by nesting Com- mon Terns. Wilson Bulletin 54:25-31. Nisbet, I. C. T. AND M. J. Welton. 1984. Seasonal variations in breeding success of Common Terns: consequences of predation. Condor 86:53-60. Palmer, R. S. 1962. Handbook of North American birds. Volume 1. Loons through flamingos. Yale University Press, New Haven, Connecticut, USA. Riehl, C. 2006. Widespread cannibalism by fledglings in a nesting colony of Black-crowned Night-her- ons. Wilson Journal of Ornithology 1 18:101-104. Shealer, D. a. 1989. Stratton and Bluff islands. The National Audubon Society’s Phineas W. Sprague Memorial Sanctuary, Saco Bay, Maine. 1989 Sea- son Report. National Audubon Society Seabird Restoration Program, Ithaca, New York, USA. Shealer, D. A. and S. W. Kress. 1991. Nocturnal abandonment response to Black-crowned Night- heron disturbance in a Common Tern colony. Co- lonial Waterbirds 14:51-56. Skinner, J. M. 1990. Stratton and Bluff islands. The National Audubon Society’s Phineas W. Sprague Memorial Sanctuary, Saco Bay, Maine. 1990 Sea- son Report. National Audubon Society Seabird Restoration Program, Ithaca, New York, USA. SzLiVKA, L. 1985. Data on the food of the Purple {Ar- dea purpurea). Night (Nycticorax nycticorax) and Squacco (Ardeola ralloides) herons on Lake Lu- das. Larus 36-37:175-182. Watmough, B. R. 1978. Observations on nocturnal feeding by night-herons {Nycticorax nycticorax). Ibis 120:356-358. Wolford, J. W. and D. A. Boag. 1971. Food habits of Black-crowned Night-herons in southern Al- berta. Auk 88:435-437. Yen, Chung-Wei. 1991. Food of nestling egrets and night herons in the western lowlands of central Taiwan. Journal of the Taiwan Museum. 44:309- 320. SHORT COMMUNICATIONS 641 The Wilson Journal of Ornithology 120(3):641-645, 2008 Diet of the Long-eared Owl in the Northern and Central Negev Desert, Israel Zohar Leader, ‘ Yoram Yom-Tov,^’^ and Uzi Motro' ABSTRACT — This is the first report of the diet composition of Long-eared Owl {Asio otus) in the northern and central Negev desert, Israel. The diet con- sisted of 71.3% small mammals, 26.5% birds, 2.0% invertebrates, and 0. 1 % reptiles. There were no signif- icant differences among the seven localities studied or among seasons in percent rodents or invertebrates in the diet. However, the proportion of psammophilious rodents within the diet was larger in settlements where the soil was sand or sandy-loess and smaller where the soil was loess or rocky. Percent birds in the diet did not differ among localities, but differed among sea- sons. Migratory birds formed a significantly larger part of the total birds consumed during migration than dur- ing the non-migrafory months. Received 15 August 2007. Accepted 24 December 2007. The Long-eared Owl {Asio otus) is a Hol- arctic, nocturnal bird of prey whose diet has been extensively studied in North America and Europe. The literature on its diet indicates that it feeds mainly on rodents, especially voles, and other small mammals, comple- mented by other animals, including birds and invertebrates (Cramp and Simmons 1985). It forages in the open, but also hunts near and below trees (Cramp and Simmons 1985) where it feeds on birds that roost in trees. In Israel there are resident, migratory, and win- tering populations of Long-eared Owls (Paz 1987, Shirihai 1996). The species inhabits most low-lying areas in Israel, chiefly in the Mediterranean region, but also in agricultural settlements in the desert during the last three decades. This species prefers semi-open areas in Israel such as agricultural settlements, plan- ' Department of Evolution, Systematics and Pxolo- gy. The Hebrew University of Jerusalem, Jerusalem 91904, Israel. ^ Department of Zook)gy, 4'el Aviv University, 'Pel Aviv 69978, Israel. ’ Corresponding author; e-mail: yomtov@post.tau.ac.il tations, and patches or lines of trees (Shirihai 1996). The diet of the Long-eared Owl has not been studied in Israel. We collected data on the diet of the Long-eared Owl and report on the composition of the diet in the northern and central Negev desert, Israel. METHODS We studied Long-eared Owls in the north- ern and central Negev, Israel, a relatively arid region where rain occurs only during winter (Nov-Apr). Annual precipitation ranges be- tween 300 mm in the north to 100 mm in the south and varies greatly from year to year. Mean monthly temperature ranges between 26° C in July and 11°C in January (Jaffe 1986). We located communal winter roosts of Long-eared Owls and nesting sites during the breeding season in or near agricultural settle- ments; these were visited once or twice every month between May 2002 and December 2003. Long-eared Owls roost in dense vege- tation which, in our study area, occurs only in settlements. All pellets were collected during each visit and the area was cleaned of pellets and remains of prey; each collection was com- posed of “fresh” pellets accumulated since the last visit. Each pellet was placed in a sep- arate bag and its date and locality were re- corded. All pellets were allocated to t:>ne of four seasons: winter (Dec-Feb), spring (Mar- May), summer (Jun-Aug), and autumn (wSep- Nov). Pellets were identified by their gray or light-black color and relatively (in comparison with those of the sympatric Barn Owl \7\lo alhci\) narrower width, and with duller color. Pellets where species identilication was not clear were not used. Each pellet was either separated to its components in the laboratory using tweezers or soaked in water until the remains, and cranial and post-cranial elements 642 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 could be separated. Species identification was by comparison with identified specimens in the collection of the Zoological Museum of Tel Aviv University and the collection of The Hebrew University of Jerusalem. We treated each single pellet as a unit con- taining the remains of the complete portion of food eaten. Bones of one prey item under field conditions usually appear in one pellet and only rarely are they discarded in two or more pellets (Raczynski and Ruprecht 1974). How- ever, there is the possibility that owls will cache large prey items and return to them lat- er. In these cases, over estimation of the large prey items would occur using our pellet anal- ysis. Our data suggests that over estimation could have occurred for only one prey species {Rattus rattus) in one locality (Nirim). The main identifiable elements were crania, mandibles, and femura for mammals, skulls and humeri for birds, mandibles for reptiles, and exoskeleton pieces for invertebrates. The minimum number of individuals (MNI) was calculated from the most common element for every species. All remains in the pellets were identified to the lowest possible taxon. Differ- ences between species were often small and most remains were identified to genus. A small proportion (2-11%) of post cranial re- mains could not be identified to genus. These remains belonged to either Meriones or Ger- billus (Gerbilidae) and we divided them be- tween the two genera by their relative amount actually found in the identifiable remains. Identified bird species were categorized by their status in Israel (residents or migrants). However, some bird genera and even species have both resident and migratory populations in Israel. The remains in these cases were cat- egorized to the status of the most common species in the genus in the study area. The seven settlements from which we col- lected pellets were (number of pellets collect- ed in parentheses): Sde Boqer 30° 52' N, 34° 48' E (280 pellets), Revivim 31°03'N, 34° 44' E (380), Gevulot 31° 12' N, 34° 28' E (340), Tse’elim 31° 12' N, 34° 32' E (97), Tzohar settlements (a group of 6 settlements within 1-2 km) 31°14'N, 34°25'E (480), Omer 31° 17' N, 34° 50' E (157), and Nirim 31° 19' N, 34° 22' E (1,300). All but one (Omer) of the sites were agri- cultural settlements, and all had ornamental trees, bushes, and grass lawns. The surround- ing fields contained agricultural crops (pota- toes, peanuts, carrots, etc.) as well as planta- tions (olives, avocado) and had lines of or scattered ornamental trees (various genera of palms. Tamarisk ssp.. Acacia ssp.). The nat- ural small mammalian fauna consists of sev- eral species of rodents and shrews with a strong psammophilous element where loess soil was mixed with sand. Commensal rodents (mice and rats) were also present. We used two-way ANOVA tests for each of the three food categories (rodents, inverte- brates, and birds) to test if there were differ- ences in diet composition among localities and seasons. We used the Shannon- Wiener In- dex to calculate a value for diet diversity for each locality and two-way ANOVAs to test differences in diet diversity among localities and seasons. RESULTS We collected 3,034 pellets of Long-eared Owls, comprising 4,668 prey items with a composition of 71.3% small mammals, 26.5% birds, 2.0% invertebrates, and 0.1% reptiles (Table 1). There were no significant differenc- es among localities or among seasons in per- cent rodents (f*-values were 0.331 and 0.076, respectively) nor in percent invertebrates (P = 0.513 and P = 0.473, respectively) in the diet. Percent birds in the diet did not differ among localities (P = 0.144), but differed among sea- sons (P = 0.037). There were no significant differences in diet diversity (Shannon- Wiener Index) among localities {P = 0.290) or among seasons (P = 0.803). There were no differences in percent ro- dents (as a group) in the diet among localities, but rodent genera composition differed among localities: percent Gerbillus among rodents changed from 63—70% in Gevulot, Revivim, and Tse’elim to 43% in Tzohar, and 16% in Sde Boqer, and only 1-2% in Nirim and Omer, while the proportion of Meriones was highest (42-46%) in Nirim and Omer, 17- 20% in Revivim, Tse’elim, Gevulot, and Tzo- har and 10% in Sde Boqer. Jaculus composed 9, 6, and 1% in Revivim, Sde Boqer, and Ge- vulot, respectively. Commensal species com- prised 67% {Mus only) of the rodents in Sde Boqer, 56% (30% Mus and 26% Rattus) in Nirim (with a possible risk of over estimation SHORT COMMUNICATIONS 643 TABLE 1. Genera and species identified in Long- eared Owl pellets, (n = number of individuals in each taxonomic class; % = percent of these individuals among the total number of individuals in the pellets). Mammalia (n = 3,327; 71.3%) Rodentia Acomys spp. Gerbillus spp. Jaculus jaculus Meriones spp. Microtus guentheri Mas musculus Rattus rattus Spalax ehrenbergi Soricomorpha Crocidura spp. Chiroptera Otonycteri hemprichii Pipistrellus spp. Aves {n = 1,236; 26.5%) Passeriformes Alaudidae sp. Carduelis carduelis C. chloris Cercomela melanura Emberiza spp. Fringilla coelebs Hippolais spp. Hirundo spp. Lanius spp. Motacilla spp. Nectarinia osea Oenanthe spp. Passer domesticus Coraciiformes Merops apiaster Galliformes Coturnix coturnix Reptilia {n = 3; 0.1%) Squamata Gekkonidae sp. Scincidae sp. Insecta {n = 93; 2.0%) Blattodea Blattidae sp. Coleoptera Scarabaeus spp. Curculionidae sp. Hymenoptera Doryliis fulvus Mantodea sp. Orthoptera Gryllidae sp. Gryllotalpa gryllotalpa Arachnida (n = 2; 0.0%) Solifugae sp. Unidentified (n = 7; 0.1%) of Rattus in the pellets), 52% (42% Mus and 10% Rattus) in Omer, and 37, 18, and 11% in Tzohar, Tse’elim, and Gevulot, respectively. We found a total of 1,236 bird remains (Ta- ble 1) and identified 881 of them. These be- longed to at least 23 species, and between 31 and 85% of which (depending upon location) were House Sparrows {Passer domesticus). Percent birds in the diet differed significantly among seasons. The number of remains of resident birds was 156, 92, 203, and 295 in autumn, winter, spring, and summer, respec- tively, while the respective numbers for mi- grants were 33, 11, 73, and 18. The number of resident bird remains was similar during the migratory (spring and autumn) and non-mi- gratory (winter and summer) seasons (359 and 387, respectively), but the number of migrant remains, which consisted mainly of small pas- serines, was 3.7 times larger during the mi- gratory seasons than during the non-migratory period (106 and 29, respectively). Thus, mi- gratory birds comprised a significantly larger part (of the total birds consumed) during mi- gration than during the non-migratory months (Fisher’s exact test, P = 2 X 10~^0- DISCUSSION The diet of the Long-eared Owl in North America, Europe, and Japan is composed mainly of voles (Glue and Hammond 1974; Marti 1974, 1976; Nilsson 1981; Village 1981; Marks 1984; Bosakowski and Smith 1991; Tome 1994; Capizzi et al. 1998; Alivi- zatos and Goutner 1999; Navarro et al. 2003; Chiba et al. 2005). The one species of vole in the Mediterranean region in Israel {Microtus guentheri) is not found in the desert and oc- curs only in the northern part of our study area. We found only three voles in the diet of the Long-eared Owl (at Nirim, the most north- ern of our localities). The main species of prey in our study area were gerbils, jirds, mice, and rats. We believe the proportions of rodent spe- cies in the diet of the Long-eared Owl reflects the composition of the rodent communities where the pellets were collected. The soil in our study area is either sand or loess (wind deposited) or a mixture of different propor- tions of these two types. There are four spe- cies of psammophile Gerbillus (G. pyrami- dum, G. (dlenhyi, G. gerbillus. and G. henleyi) and one non-psammophile (G. dasyurus) that 644 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 occurs mainly in rocky habits. The diet of the Long-eared Owl in areas where the soil was sandy or had a large proportion of sand (Re- vivim, Gevulot, and Tse’elim) was mainly Gerhillus (63-70% of the rodents in the diet), while in loess-dominated areas (Nirim, Omer, Sde Boqer) it was below 17%. The proportion of Gerhillus constituted 43% of the rodents in Tzohar, where the soil is sandy-loess. Gerbil- lus was replaced by other rodents, mainly Meriones, where the soil was not sandy. Com- mensal rodents (house mouse [Mus musculus] and black rat [Rattus rattus]) comprised >50% of the diet in Sde Boqer, Nirim, and Omer. The high proportion of house mice in the diet in Sde Boqer is explained by a mice plague that occurred there in 2003, while the high proportion of commensal speeies in Omer is probably the result of this being an urban suburb of the city of Beer Sheva. Jer- boas {Jaculus jaculus) occur on both types of soil, and occurred in a low proportion in Re- vivim (sand) and Sde Boqer (loess). The numbers of resident bird remains were similar during the migratory (spring and au- tumn) and non-migratory (winter and sum- mer) seasons (359 and 387, respectively). The number of remains of migrant birds was 3.7 times larger during the migratory season than during the non-migratory period (106 and 29, respectively). The Long-eared Owl has been characterized as having a more specialized diet than other sympatric owls (Marti 1976, Andrews 1990). Other studies (Marks 1984, Bertolino et al. 2001) demonstrated that Long-eared Owls feed opportunistically. Our study indicates the diet of the Long-eared Owl in the northern and central Negev desert eonsists mainly of rodents (71.3%), but also of birds (26.5%). The proportion of psammophilious rodents within the diet was large in settlements where the soil was sand or sandy-loess, and small where the soil was loess or rocky, and the pro- portion of migratory birds was 3.7 times larg- er during migratory seasons than during non- migratory periods. These variations reflect the availability of the different prey types, and suggest a noticeable plasticity and an oppor- tunistic feeding behavior of the Long-eared Owl. ACKNOWLEDGMENTS We are grateful to Yoav Motro, Miriam Belmaker, Noam Leader, Mali Tores, Eran Levin, Leonid Fried- man, Asaf Tsoar, Boaz Shacham, Igor Gavrilov, and three reviewers for their help, advice, and comments. LITERATURE CITED Alivizatos, H. and V. Goutner. 1999. Winter diet of the Barn Owl {Tyto alba) and Long-eared Owl {Asia otus) in northeastern Greece: a comparison. Journal of Raptor Research 33:160-163. Andrews, R 1990. Owls, caves and fossils. Natural History Museum Publications, London, United Kingdom. Bertolino, S., E. Ghiberti, and A. Perrone. 2001. Feeding ecology of the Long-eared Owl {Asio otus) in northern Italy: is it a dietary specialist? Canadian Journal of Zoology 79:2192-2198. Bosakowski, T. and D. G. Smith. 1991. Comparative diets of sympatric nesting raptors in the eastern deciduous forest biome. Canadian Journal of Zo- ology 70:984-992. Capizzi, D., L. Caroli, and P. Varuzza. 1998. Feeding habits of sympatric Long-eared Owl Asio otus. Tawny Owl Strix aluco and Barn Owl Tyto alba in a Mediterranean coastal woodland. Acta Orni- thologica 33:85-92. Chiba, A., M. Onojima, and T. Kinoshita. 2005. Prey of the Long-eared Owl {Asio otus) in the suburbs of Niijata City, central Japan, as revealed by pellet analysis. Ornithological Science 4:169-172. Cramp, S. and K. E. L. Simmons. 1985. Handbook of the birds of Europe, the Middle East, and North Africa. The birds of the Western Palearctic. Vol- ume IV. Terns to woodpeckers. Oxford University Press, Oxford, United Kingdom. Glue, D. E. and G. J. Hammond. 1974. Feeding ecol- ogy of the Long-eared Owl in Britain and Ireland. British Birds 67:361-369. Jafee, S. 1986. Climate of Israel. Pages 79-94 in The zoogeography of Israel (Y. Yom-Tov and E. Tcher- nov. Editors). W. Junk, Dordrecht, The Nether- lands. Marks, J. S. 1984. Feeding ecology of breeding Long- eared Owls in southwestern Idaho. Canadian Jour- nal of Zoology 62:1528-1533. Marti, C. D. 1974. Feeding ecology of four sympatric owls. Condor 76:45-61. Marti, C. D. 1976. A review of prey selection by the Long-eared Owl. Condor 78:331-336. Navarro, J., J. A. Sanchez-Zapata, M. Carrete, F. Botella, a. Gavrilov, S. Sklyarenoko, J. A. Donazar, O. Ceballos, and E Hiraldo. 2003. Diet of three sympatric owls in steppe habitats of eastern Kazakhstan. Journal of Raptor Research 37:256-258. Nilsson, I. N. 1981. Seasonal changes in food of the Long-eared Owl in southern Sweden. Ornis Scan- dinavica 12:216-223. SHORT COMMUNICATIONS 645 Paz, U. 1987. Birds of Israel. S. Steimatzky, Tel Aviv, Israel. Raczynski, J. and a. L. Ruprecht. 1974. The effect of digestion on the osteological composition of owl pellets. Acta Ornithologica 14:25-37. Shirihai, H. 1996. The birds of Israel. Academic Press, London, United Kingdom. Tome, D. 1994. Diet composition of the Long-eared Owl in central Slovenia: seasonal variation in prey use. Journal of Raptor Research 28:253-258. Village, A. 1981. The diet and breeding of Long- eared Owls in relation to vole numbers. Bird Study 28:215-224. The Wilson Journal of Ornithology 120(3):645-648, 2008 First Observed Instance of Polygyny in Flammulated Owls Brian D. Linkhart,'-^ Erin M. Evers, Julie D. Megler,' "* Eric C. Palm,* ^ Catherine M. Salipante,^ '’ and Scott W. Yanco'’^ ABSTRACT. — We document the first observed in- stance of polygyny in Flammulated Owls {Otus flam- meolus) and the first among insectivorous raptors. Chronologies of the male’s two nests, which were 510 m apart, were separated by nearly 2 weeks. Each brood initially consisted of three owlets, similar to the mean brood size in monogamous pairs. The male delivered considerably fewer prey to the secondary nest, com- pared with prey-delivery rates at nests of monogamous males during the nestling period. Evidence suggested that all owlets fledged from the primary brood, but only one fledged from the secondary brood. We were uncertain of the cause of polygyny, but a possible ex- planation is the Hayman Fire shifted the operational sex ratio of the owls in favor of females. The extent of polygyny in Flammulated Owls may be limited by costs to the reproductive success of secondary females. Received JO February 2007. Accepted 22 December 2007. Facultative polygyny has been reported in nine species of strigiforms (Korpimaki 1983; Solheim 1983; Marti 1990; Marks et al. 1989, 1999, 2002), most of which feed primarily on ' Department of Biology, Colorado College, 14 East Cache La Poudre Street, Colorado Springs, CO 80903, USA. 2 416 South Leyden Street, Denver, CO 80224, USA. 1 Alger Place, Gros.se Pointe, MI 48230, USA. U.S. Geological Survey, Western Ecological Re- .search Center, San Francisco Bay PNtuary Field Sta- tion, 505 Azuar Drive, Vallejo, CA 94592. USA. ^94 Thornton Street, Hamden, CT 06517, USA. ^ 154 Holliston Street, Medway, MA 02053, USA. ’Corresponding author; e-mail: blinkhart@coloradocollege.edu small mammals. We document the first ob- served instance of polygyny in Flammulated Owls {Otus flammeolus), insectivores that feed primarily on moths (Reynolds and Link- hart 1987). This owl is a neotropical migrant, breeding in western North America and win- tering as far south as Central America (McCallum 1994). Flammulated owls are cav- ity nesters and, in Colorado, breed in mature conifer forests dominated by ponderosa pine {Pinus ponderosa) (Linkhart 2001). The ob- jective of this paper is to describe the polyg- ynous event, and compare provisioning rates and fledging success at the male’s two nests. OBSERVATIONS We initiated a study in 2003 of recoloni- zation by Flammulated Owls of the area af- fected in the Flayman Fire, which burned the largest area (560 km^) in recorded Colorado history in 2002 (Graham 2003). We located four territorial males in 2004, each of which occurred within a small parcel (20-30 ha) of unburned forest, and each separated by 7-20 km. One territorial male exhibited polygynous behavior, defined as a male feeding two fe- males and their young at different nests at which no other male was detected. The two nests were 510 m apart, a distance somewhat greater than the mean (± ,SE) diameter of ter- ritories of monogamous males (428 ± 29 m; Linkhart 2001). We discovered the nests, both of which were in natural cavities in aspen {Po- pulus trettiuloides) trees, on 22 June (CS2 nest) and 12 July (CSl nest) after hearing 646 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 adult feeding vocalizations near each nest. We captured and banded the CS2 female (USGS aluminum leg band #418) on 12 July, and the CSl female (#381) on 13 July while each was day-roosting in her nest cavity. We captured the male with a small mist net (Reynolds and Linkhart 1984) on 15 July as he exited the CSl nest at 1949 hrs MST and banded him (#419). We presumed he had delivered prey to the brooding CSl female because his flight to the nest was preceded by food location calls, a distinctive sequence of behaviors shown by nesting males (Reynolds and Link- hart 1984). We captured the same male as he exited the CS2 nest at 2150 hrs following a presumed prey delivery. The male was not re- captured thereafter, but observations indicated he was the sole male tending the nests over the remainder of the nestling period. First, we observed only a banded male visiting the nests and the distance (7 km) to the nest of the clos- est banded male was sufficiently great (~6 times the maximum observed flight distance of radio-marked males; BDL, unpubl. data) that visits by other banded males were unlike- ly. Second, the male gave food location calls from the same particular perches near nests and we noted no deviations from this pattern that might suggest a second male. It also was unlikely a second male tended either nest prior to its discovery, since cooperative breeding is not known in this species (McCallum 1994). Each brood initially consisted of three owl- ets, similar to the mean brood size in monog- amous pairs (2.5 ± 0.1 owlets) (Linkhart and Reynolds 2006). We estimated, based on pat- terns of plumage development and mass gain in broods of monogamous pairs (BDL, un- publ. data), the CSl owlets were 13, 17, and 18 days of age on 13 July, and the CS2 owlets were 1, 2, and 5 days of age on 12 July. Thus, the chronologies of the nests differed by near- ly 2 weeks. The male delivered considerably fewer prey to the secondary nest (CS2) during the nest- ling period compared with prey-delivery rates at nests of monogamous males. The male de- livered 5.3 prey/hr (8 prey in 1.5 hrs) on one night (13 Jul) during the mid/latter portion of the nestling period at the CSl nest, a some- what lower rate than that by monogamous males at this time (8 prey/hr; Reynolds and Linkhart 1987). The male delivered 1.1 prey/ hr (6 prey in 5 hrs) over three nights (12, 13, and 15 Jul) at the CS2 nest during the first week of the nestling period, which is much lower than by monogamous males at this time (7 prey/hr; Reynolds and Linkhart 1987). The male failed to deliver any prey to the CS2 nest in 2 hrs of observation on the last night of the nestling period (31 Jul). Each female deliv- ered <2 prey/hr to her nest on the above nights, a similar rate to that of monogamous females (Reynolds and Linkhart 1987). The evidence suggested that all owlets fledged from the CSl nest, whereas only one fledged from the CS2 nest. The youngest owl- et in the CSl nest fledged on 20 July at —20 days of age. We presume its two older siblings had fledged by this date, given that on 18 July the eldest owlet would have been —23 days of age, which is the mean (± SD) duration (± 1 day) of the nestling period in Flammulated Owls (Reynolds and Linkhart 1987). It was unlikely the older siblings perished due to poor health, given that mass and feather de- velopment of each owlet on our previous visit (13 Jul) had been comparable to other owlets of similar age. In contrast, only the eldest owl- et (#420) fledged from the CS2 nest at 23 days of age (30 Jul) when it had 24% less mass but similar primary development to other owlets at this age (n = 10; BDL, unpubl. data). The youngest owlet was missing when we visited the CS2 nest on 19 July and the second-eldest owlet was in poor condition, as indicated by having 42% less mass and a 70% shorter 5th primary feather compared to other 9-day old owlets (n = 18; BDL, unpubl. data). Two days later (21 Jul), we found the second-eldest owl- et dead in the nest (specimen saved; BDL). DISCUSSION Our observations constitute the first report- ed instance of polygyny among insectivorous raptors. We were uncertain of its cause. One possible explanation is the Hayman Fire, which likely caused a significant loss of owl habitat (few trees, if any, survived in 51% of the forested area; Graham 2003), reduced the density of breeding males more than females, which are not territorial (Linkhart and Reyn- olds 2007). Polygyny occurred in passerines when the operational sex ratio shifted in favor of females (Greenlaw and Post 1985, Kem- penaers 1994). SHORT COMMUNICATIONS 647 Polygyny has obvious potential benefits to the reproductive success of males, but the ex- tent of polygyny among birds is believed to be most limited by costs to females (Orians 1969). Polygynous males often have reduced rates of provisioning at secondary nests com- pared to primary nests (Sejberg et al. 2000) with the possible consequence of reducing the reproductive success of secondary females (Johnson et al. 1993). Reduced provisioning by the male at the CS2 nest during the nestling period likely caused the mortality of two of the three owlets. Reproductive success of female Flammu- lated Owls is particularly reliant on provision- ing by males, not only because males provide most of the prey to nests during the nestling period, as is true with most owls (Marks et al. 1999), but also because most prey are small (—0.1 g) and are delivered singly to nests (Linkhart et al. 1998). Documenting mortality of nestling Flammulated Owls is uncommon in Colorado, as 95% (144 of 151) of banding- age owlets of monogamous males fledged from 1981 to 1999 (Linkhart 2001). Even more unusual is observing partial mortality within a brood, given there was just one other occasion when a portion of a brood died while the rest survived to fledge {n = 79; Linkhart 2001). Polygyny in species with biparental care is more likely when chronologies of nesting cy- cles are staggered, according to the ‘asynchro- nous-settlement model’ (Leonard 1990), be- cause of the difficulty of provisioning multiple nests when chronologies are similar (Johnson et al. 1993). A potential tradeoff of staggered nesting chronologies is that nestlings in sec- ondary broods fledge at later dates, which may reduce survival rates of young in species such as Flammulated Owls that migrate long dis- tances (Rappole 1995). Fledglings cannot for- age independently of parents until 4-5 weeks after fledging (Linkhart and Reynolds 1987). Thus, owlet #420 could not have departed for migration before early September when ar- thropod prey becomes increasingly scarce in Colorado (BDL, unpubl. data). Owlets from only three broods (/? = 132 broods) fledged at later dates than owlet #420 since 1981 (BDL, unpubl. data). This suggests there is strong se- lection against late fledging and, indirectly, against polygyny in this species. ACKNOWLEDGMENTS The authors thank C. E. Braun, J. S. Marks, and S. O. Williams III for helpful comments on the manu- script. We thank Colorado College, the Hulbert Center for Southwest Studies, the Hughes Undergraduate Re- search Program, the Lois Webster Fund (Audubon So- ciety), and the USDA Forest Service for funding. The Rocky Mountain Research Station provided lodging. LITERATURE CITED Graham, R. T. (Editor). 2003. Hayman Fire case study. USDA, Forest Service, General Technical Report RM-114. Rocky Mountain Research Station, Fort Collins, Colorado, USA. Greenlaw, J. S. and W. Post. 1985. Evolution of mo- nogamy in Seaside Sparrows, Arnmodramiis rnar- itimus: tests of hypotheses. Animal Behaviour 33: 373-383. Johnson, L. C., L. H. Kermott, and M. R. Lein. 1993. The cost of polygyny in the House Wren Trog- lodytes aedon. Journal of Animal Ecology 62: 669-682. Kempenaers, B. 1994. Polygyny in the Blue Tit: un- balanced sex ratio and female aggression restrict male choice. Animal Behavior 47:943-957. Korpimaki, E. 1983. Polygamy in Tengmalm’s Owl Aegolius funereus. Ornis Fennica 60:86—87. Leonard, M. L. 1990. Polygyny in Marsh Wrens: asynchronous settlement as an alternative to the polygyny-threshold model. American Naturalist 136:446-458. Linkhart, B. D. 2001. Life history characteristics and habitat quality of Flammulated Owls (Otus flam- meolus) in Colorado. Dissertation. University of Colorado, Boulder, USA. Linkhart, B. D. and R. T. Reynolds. 1987. Brood division and postnesting behavior of Flammulated Owls. Wilson Bulletin 99:240-243. Linkhart, B. D. and R. T. Reynolds. 2006. Lifetime reproduction of Flammulated Owls in Colorado. Journal of Raptor Research 40:29-37. Linkhart, B. D. and R. T. Reynolds. 2007. Rate of return, fidelity, and dispersal in a breeding popu- lation of Flammulated Owls. Auk 124:264-275. Linkhart, B. D., R. T. Reynolds, and R. A. Ryder. 1998. Home range and habitat of breeding Flam- mulated Owls in Colorado. Wilson Bulletin 1 10: 342-35 1 . Marks, J. S., R. J. Cannings, and H. Mikkola. 1999. Family Strigidae. Pages 76-151 in Handbook of the birds of the world. Volume 5 (J. del Hoyo, A. Elliott, and J. Sargatal, Editors). Lynx Editions, Barcelona, Spain. Marks, J. S., J. L. Dickin.son, and J. Haydogk. 2002. Serial polyandry and alloparenting in Long-eared Owls. Condor 104:202-204. Marks, J. S., J. H. Doremlis, and R. J. Cannings. 1989. Polygyny in the Northern Saw-whet Owl. Auk 106:732-734. 648 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 Marti, C. D. 1990. Same-nest polygyny in the Barn Owl. Condor 92:261-263. McCallum, a. 1994. Flammulated Owl (Otus flam- meolus). The birds of North America. Number 93. Orians, G. H. 1969. On the evolution of mating sys- tems in birds and mammals. American Naturalist 103:589-603. Rappole, J. H. 1995. The ecology of migrant birds: a neotropical perspective. Smithsonian Institutional Press, Washington, D.C., USA. Reynolds, R. T. and B. D. Linkhart. 1984. Methods and materials for capturing and monitoring Flam- mulated Owls. Great Basin Naturalist 44:49-5 1 . Reynolds, R. T. and B. D. Linkhart. 1987. The nest- ing biology of Flammulated Owls in Colorado. Pages 239-248 in Biology and conservation of northern forest owls (R. W. Nero, R. J. Clark, R. J. Knapton, and R. H. Hamre, Editors). USDA, Forest Service, General Technical Report RM- 142, Rocky Mountain Research Station, Fort Col- lins, Colorado, USA. Sejberg, D., S. Bensch, and D. Hasselquist. 2000. Nestling provisioning in polygynous Great Reed Warblers {Acrocephalus arundinaceus): do males bring larger prey to compensate for fewer nest vis- its? Behavioral Ecology and Sociobiology 47: 213-219. SoLHEiM, R. 1983. Bigyny and biandry in the Teng- malm’s Owl (Aegolius funereus). Ornis Scandi- navica 14:51-57. The Wilson Journal of Ornithology 120(3):648-65 1, 2008 The Giant Hummingbird (Patagona gigas) in the Mountains of Central Argentina and a Climatic Envelope Model for its Distribution Henrik von Wehrden*’^ ABSTRACT — I present the first published obser- vations of the Giant Hummingbird {Patagona gigas) in the mountains of central Argentina. This species was recorded in early and late summer 2006 and 2007. This new range resembles other summer habitats of the species, which are in the Andes >500 km distant. A climatic envelope model configured with known lo- cations obtained from the literature predicts a high probability of occurrence in its “new” range. Received 11 July 2007. Accepted 18 October 2007. The Giant Hummingbird {Patagona gigas) is the largest hummingbird in the world. Its distribution includes Ecuador (Ortiz-Crespo 1974, King and Holloway 1990), Colombia (Woods et al. 1998), Peru and Bolivia (Kok- shaisky 2001, Wester and Classen-Bockhoff 2006), Chile (Vasquez and Simonetti 1999), and Argentina (Acreche et al. 1998, Oses 2003). This species typically inhabits high mountain habitats (Fjeldsa and Krabbe 1990, Barnett and Pearmann 2001) and most records ' Research Institute of Wildlife Ecology, Savoyen Strasse 1, 1160 Vienna, Austria. 2 Institute of Biology, Geobotany and Botanical Gar- den, Martin-Luther-University Halle-Wittenberg, 06099 Halle, Germany; e-mail: HenrikvonWehrden@ web.de originate from the Andes. The altitudinal dis- tribution of the species ranges to 4,600 m (Oses 2003) and is energy-driven (Fernandez and Bozinovic 2003). Narosky and Yzurieta (2003) indicate this species overwinters with- in the eastern lowlands of Argentina. How- ever, all summer records originate from west- ern Argentina in the vicinity of the Andes with important over-wintering habitats in north-western Argentina (Fjeldsa and Krabbe 1990, Schuchmann 1999). A recent phyloge- netic analysis of hummingbirds confirmed the species as being relatively isolated in relation to other hummingbird taxa (Altshuler et al. 2004). Oses (2003) classifies the species as comparatively primitive compared to other Trochilidae, which may be the reason for the variety of plants used by the taxon (Sahley 1996, Kokshaisky 2001, Schlumpberger and Badano 2005, Wester and Classen-Bockhoff 2006). OBSERVATIONS The Giant Hummingbird was encountered in the lower ranges of the Sierras Grandes de Cordoba, in central Argentina on 1 1 February 2006 at an altitude of —1,430 m elevation (31° 40' S, 64°40'W). The large size, char- SHORT COMMUNICATIONS 649 85°W J 0°N- 10°S- 20°S- 30°S- 40°S- 80°W L 75°W 70°W 65°W 60°W 55°W no low medium high very high excellent FIG. I. Climatic envelope predicting distribution of the Giant Hummingbird in South America based on 99 known records. The arrow shows the area of the new records presented in this paper. 650 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 acteristic eye-ring, straight bill, and dull col- oration made the identification definite (De La Pena and Rumboll 1998, Oses 2003). Further, the observed gliding flight is rather unusual within the Trochilidae. All other humming- birds in the region are much smaller; the Red- tailed Comet {Sappho sparganura) was pre- sent, while the Blue-tufted Starthroat {Heliom- aster furcifer) and the Glittering-bellied Em- erald {Chlorostilbon aureoventris) were observed at lower altitudes and are easily dis- tinguishable. The Giant Hummingbird was encountered in the same location on 3 days during summer 2007 (18 and 19 Feb, 3 Apr) with multiple observations each day. The species was seen nine times by the author on the first day and confirmed by other scientists. The habitat was heterogeneous grassland, which contained some shrubs and a few trees (mainly Mayten- us boaria). Several flowering plants repre- sented potential feeding sources for the hum- mingbird, yet it was most often seen feeding on Siphocampylus foliosus (Campanulaceae). The hummingbird was observed at a distance of <3 m, hovering close to the observers on several occasions. The species was also seen at another location closer to the mountain range at a distance of 4 km (1,630 m eleva- tion) from the first site. The birds were ob- served resting frequently in trees at both lo- cations. Weather conditions were variable and observations were made on clear sunny and misty rainy days. DISCUSSION A spatial data base was created to construct a climatic envelope model for the species based on data obtained from Oses (2003). The new observations presented here were not in- corporated into this model. Thus, a presence only model was created, which is a standard tool within ecological sciences (Peterson 2001, Pearce and Boyce 2006). Annual mean temperature, annual precipitation, mean tem- perature of the warmest quarter of the year, and annual temperature range were used as predictors and a public domain climate model (Hijmans et al. 2005) was used to derive the spatial model. The model predicted the species with a high probability for the mountains of central Ar- gentina, where I observed the species (arrow Fig. 1). I suggest the presence of the species in the area will be persistent since the species was present in both late and early summer over 2 years. The nearest known observations are from Salta (Acreche et al. 1998), San Juan (Contreras 1978), and Mendoza (Oses 2003) provinces, all of which are at a distance >500 km. The mountainous habitat where the ob- servations were made resembles the known preference of the species, which is regarded to be selective for its habitat (Vasquez and Si- monetti 1999). The observations support the unique island-like ecology of mountains of central Argentina (Nores 1995, Cabido et al. 1998). Thus, a permanent presence during the summer seems highly probable, which is also indicated by the climatic envelope model. The climatic distribution obtained from this model has a higher spatial resolution when compared to the standard literature. It demonstrates that climatic envelope models represent a valuable approach to identify the potential distribution of a species and helps researchers locate new distribution ranges. Model construction for other species might enhance our understand- ing of the distribution of these taxa as well. ACKNOWLEDGMENTS The author expresses gratitude to Heike Zimmer- mann and Heidi Hirsch for confirming his observation. Maurice Rumboll verified the record as new for the region. Danny McCluskey read earlier versions of this manuscript. Daniel Renison, C. E. Braun, and one anonymous referee provided helpful comments on the manuscript. The author thanks the Austrian Science Fund (Project PI 8624). LITERATURE CITED Acreche, N., H. A. Nunez, and M. V. Albeza. 1998. Vulnerabilidad de la avifauna en el Parque Na- cional Los Cardones, Salta, Argentina. Revista de Biologia Tropical 46:81 1-816. Altshuler, D. A., R. Dudley, and J. A. McGuire. 2004. Resolution of a paradox: hummingbird flight at high elevation does not come without a cost. Proceedings of the National Academy of Sci- ences of the United States of America-Biological Sciences 101:17731-17736. Barnett, J. M. and M. Pearmann. 2001. Annotated checklist of the birds of Argentina. Lynx Edicions, Barcelona, Spain. Cabido, M., G. Funes, E. Pucheta, F. Vendramini, AND S. Diaz. 1998. A chorological analysis of the mountains from central Argentina. Is all what we call Sierra Chaco really Chaco? Contribution to SHORT COMMUNICATIONS 651 the study of the flora and vegetation of the Chaco. XII. Candollea 53:321-331. Contreras, J. 1978. Some comments on the races of the Giant Hummingbird, Patagona gigas, in the provinces of Mendoza and San Juan (Aves: Tro- chilidae). Neotropica 24:47-49. De La Pena, M. R. and M. Rumboll. 1998. Birds of southern South America and Antarctica. Princeton University Press, Princeton, New Jersey, USA. Fernandez, M. J. and F. B. Bozinovic. 2003. Flight efficiency and energetic costs along an altitudinal gradient in the world’s largest hummingbird: Pa- tagona gigas. Integrative and Comparative Biol- ogy 43:820. Fjeldsa, j. and N. Krabbe. 1990. Birds of the high Andes. Apollo books, Copenhagen, Denmark. Humans, R. J., S. E. Cameron, J. L. Parra, P. G. Jones, and A. Jarvis. 2005. Very high resolution interpolated climate surfaces for global land areas. International Journal of Climatology 25:1965- 1978. King, J. R. and S. J. Holloway. 1990. Notes on the Giant Hummingbird Patagona gigas in southern Ecuador. Bulletin of the British Ornithologists Club 110:79-80. Kokshaisky, N. V. 2001. Species composition, phe- nology and trophic relations in hummingbirds from Central Valley, Bolivia. Zoologichesky Zhurnal 80:210-221. Narosky, T. and D. Yzurieta. 2003. Birds of Argen- tina and Uruguay. Vazquez Mazzini editores, Buenos Aires, Argentina. Nores, M. 1995. Insular biogeography of birds on mountain-tops in north western Argentina. Journal of Biogeography 22:61-70. Ortiz-Crespo, F. 1974. The Giant Hummingbird Pa- tagona gigas in Ecuador. Ibis 116:347-359. OsES, C. S. 2003. Taxonomy, phylogeny and bioge- ography of the Andean hummingbird Genera Coe- ligena LESSON, 1832; Pterophanes GOULD, 1849; Ensifera LESSON, 1843; and Patagona GRAY, 1840 (Aves: Trochiliformes). Dissertation. University of Bonn, Germany. Pearce, J. L. and M. S. Boyce. 2006. Modeling dis- tribution and abundance with presence-only data. Journal of Applied Ecology 43:405-412. Peterson, A. T. 2001. Predicting species’ geographic distributions based on ecological niche modeling. Condor 103:599-605. Sahley, C. T. 1996. Bat and hummingbird pollination of an autotetraploid columnar cactus, Weber- bauerocereus weberbaueri (Cactaceae). American Journal of Botany 83:1329-1336. ScHLUMPBERGER, B. O. AND E. I. Badano. 2005. Di- versity of floral visitors to Echinopsis atacamensis ssp. pasacana (Cactaceae). Haseltonia 11:18-26. SCHUCHMANN, K.-L. 1999. Order Apodiformes. Eamily Trochilidae (Hummingbirds). Pages 468-535 in Handbook of the birds of the world. Volume 5. Barn-owls to hummingbirds (J. del Hoyo, A. El- liott, and J. Sargata, Editors). Lynx Edicions, Bar- celona, Spain. Vasquez, R. a. and j. a. Simonetti. 1999. Life his- tory traits and sensitivity to landscape change: the case of birds and mammals of Mediterranean Chile. Revista Chilena de Historia Naturalista 72: 517-525. Wester, P. and R. Classen-Bockhoff. 2006. Hum- mingbird pollination in Salvia haenkei (Lami- aceae) lacking the typical lever mechanism. Plant Systematics and Evolution 257:133-146. Woods, S., E Ortiz-Crespo, and P. M. Ramsay. 1998. Presence of Giant Hummingbird Patagona gigas and Ecuadorian Hillstar Oreotrochilus chimbora- zo jamesoni at the Ecuador-Colombia border. Co- tinga 10:37-40. The Wilson Journal of Ornithology 120(3):65 1-653, 2008 Giant Hummingbirds {Patagona gigas) Ingest Calcium-rich Minerals Cristian F. Estades,''^ M. Angelica Vukasovic,' and Jorge A. Tomasevic' ABSTRACT. — We report Giant Hummingbirds {Pa- tagona gigas), regularly and deliberately, ingesting wood ashes and slaked lime in central Chile. These two minerals have high concentrations of calcium, which may be a scarce element in the nectar-based diet ' Laboratorio de Ecologfa de Vida Silvestre, Depar- tamento de Manejo de Recursos Forestales, Universi- dad de Chile, Casilla 9206, Santiago, Chile. - Corresponding author; e-mail: cestades@uchile.cl of the species. Both observations occurred during the post-breeding period suggesting the birds were females ingesting calcium-rich compounds to replace minerals lost during eggshell production. Received Jl March 2007. Accepted 10 Novendjcr 2007. The nectar-based diet of hummingbirds poses nutritional restrictions such as a limited 652 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 supply of proteins and some minerals. These nutrients are usually obtained from insects that hummingbirds include in their diet (Rem- sen et al. 1986). However, minerals may be lost at a much higher rate through the highly diluted urine of hummingbirds, or during egg production (Adam and des Lauriers 1998). The latter may cause hummingbirds to sup- plement their diet with mineral-rich com- pounds (Adam and des Lauriers 1998). We re- port two different observations of the Giant Hummingbird (Patagona gigas) deliberately ingesting calcium-rich minerals in central Chile. OBSERVATIONS We observed an adult Giant Hummingbird on several occasions during February 2005 (austral summer) visit a site where ashes from a barbecue had accumulated on the ground at the Pantanillos Forest Research Station (35° 32' S, 72° 17' W) near Constitucion in central Chile. Our first impression was the bird was looking for insects but we soon realized it was apparently ingesting ashes. Each time the bird would sit on the ground and place its beak into the soil or at a short distance above the ground while licking the ashes with its tongue. The bird came to the site regularly, making 2-3 short visits (<2 min) during the morning (0800-1000 hrs) and afternoon (1700-1800 hrs). MAV observed an adult Giant Humming- bird during most of February 2007 coming regularly to a wall of an old country house painted with slaked lime (calcium hydroxide) near the town of Quirihue (36° 20' S, 72° 38' W) in central Chile. The bird flew directly to two different spots in the wall where lime dust accumulated inside small crevices. At the spot closer to the ground (0.3 m) the bird would hover while licking the wall, whereas at the higher location (1.7 m) the bird would usually sit while reaching inside the crevice with its beak. Close inspection of these crevices failed to detect the presence of insects, spider webs or other nest building materials, which could be attracting the birds. The bird visited the site in the morning and in the afternoon, but made only one visit at each time of the day. DISCUSSION Calcium is one of the main components of wood ashes (des Lauriers 1994) and the slaked lime dust, although probably mixed with dirt, was likely rich in calcium as well. We believe the observed birds were regularly and delib- erately ingesting calcium-rich compounds. We could not identify the gender of the birds because the Giant Hummingbird is sex- ually monochromatic with males and females differing only slightly in size (Araya et al. 1986). This is important because most records of hummingbirds ingesting calcium are of fe- males (Verbeek 1971, des Lauriers 1994, Adam and des Lauriers 1998) that are likely trying to meet their requirements of calcium for eggshell production. Our observations likely corresponded to birds in the post-breeding period as nesting for this species in this region normally occurs be- tween October and January (pers. obs.). We cannot completely eliminate the possibility the birds were late breeders because, unfortunate- ly, we did not look for nests. However, we doubt the latter was the case because the spe- cies migrates from the region in March. It is believed that hummingbirds, being the smallest birds alive, do not have sufficient medullary bone for storing calcium for egg- shell production (Adam and des Lauriers 1998). Ingestion of extraneous calcium by birds that cannot store calcium in medullary bones should peak near the period of maxi- mum breeding activity (Dhondt and Hochach- ka 2001). Our observations suggest that Giant Hummingbirds were ingesting minerals to re- place the calcium lost during eggshell produc- tion that had likely occurred 2-3 months ear- lier. It is possible the Giant Hummingbird, be- ing four times heavier (18-22 g) than most hummingbird species (3-5 g), has a higher calcium storage capacity. Ours are not the only observations of hum- mingbirds ingesting calcium after the breeding period. Verbeek (1971) reported a female Anna’s Hummingbird (Calypte anna) feeding on calcium-rich sand more than a month after the last reported date for a clutch of the spe- cies. Most records of female hummingbirds apparently taking calcium are from the nesting period (des Lauriers 1994, Adam and des Lau- riers 1998), including several Anna’s Hum- mingbirds (Adam and des Lauriers 1998). We did not observe consumption of calcium by Giant Hummingbirds when nests are active SHORT COMMUNICATIONS 653 (Oct-Nov) during several years of work in both field sites. Hummingbirds have been reported eating calcium-rich sand and soil (Verbeek 1971, Adam and des Lauriers 1998). Ash-eating be- havior had been previously described for some North American hummingbirds (des Lauriers 1994) but, to our knowledge, there are no reports of this behavior for the Giant Hummingbird or any other hummingbird in the Southern Hemisphere. We believe the re- cord of the consumption of slaked lime is the first for a hummingbird, and probably the sec- ond for a bird, after the observation reported by Richmond (1953) of Purple Martins {Prog- ne subis) ingesting slaked lime from the ground in Oregon. ACKNOWLEDGMENTS We appreciate the comments of two anonymous ref- erees. LITERATURE CITED Adam, M. A. and J. R. des Lauriers. 1998. Obser- vations of hummingbirds ingesting mineral-rich compounds. Journal of Field Ornithology 69:257- 261. Araya, B., G. Millie, and M. Bernal. 1986. Guia de Campo de las Aves de Chile. Editorial Universi- taria, Santiago, Chile. DES Lauriers, J. R. 1994. Hummingbirds eating ashes. Auk 111:755-756. Dhondt, a. a. and W. M. Hochachka. 2001. Varia- tions in calcium use by birds during the breeding season. Condor 103:592-598. Remsen Jr., j. V., F. G. Stiles, and P. E. Scott. 1986. Frequency of arthropods in stomachs of tropical hummingbirds. Auk 103:436-441. Richmond, S. M. 1953. The attraction of Purple Mar- tins to an urban location in western Oregon. Con- dor 55:225-249. Verbeek, N. A. M. 1971. Hummingbirds feeding on sand. Condor 73:112-113. The Wilson Journal of Ornithology 120(3):653-654, 2008 Nocturnal Foraging Observations of the Blue-crowned Motmot (Momotus momota) in San Jose, Costa Rica Alejandro Solano-Ugalde^’^ and Agustina Arcos-Torres' ABSTRACT — We provide documentation on the first observations of nocturnal foraging by the Blue- crowned Motmot {Momotus momota). The motmot we observed mainly fed on sphinx (Sphingidae) moths; the capture rate seemed low for this fairly large prey. Received 27 April 2007. Accepted 26 October 2007. Nocturnal foraging has only been reported once (Thurber and Komar 2002) for the neo- tropical Momotidae: Coraciformes, despite many hours of field observations (Skutch 1945, 1947, 1964, 1971; Orejuela 1980, Rem- sen et al. 1993, Chacon-Madrigal and Barran- tes 2004, Garcia-C and Zahawi 2006). The Thurber and Komar (2002) report involved a ' Fundacion Imaymana, Lincoln 199 y San Ignacio, Quito, Ecuador. ^Corresponding author; e-mail: jhale/.ion@ gmail.com single Turquoise-browed Motmot (Eumomota superciliosa) in the Republic of El Salvador, which for 2 consecutive years (1975-1976) fed on insects attracted to an artificial light source in a residential yard. We provide doc- umentation in this note of the first known ob- servations of nocturnal foraging by the Blue- crowned Motmot {Momotus momota). OBSERVATIONS We report opportunistic observations of a single Blue-crowned Motmot nocturnally for- aging on aerial insects in the vicinity of an electric street lamp at Coopecabahas, San Jose, Costa Rica (09° 55' N, 84° 12' W; 910 m elevation). Our first observation was on 5 De- cember 2004 at 2 1 28 hrs when we observed a motmot on the lawn of a residence feeding on a fairly large sphinx moth (Sphingidae). The bird perched 2 m above the ground, after 654 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 consuming the moth, on an almost vertical branch of a fallen tree in the hilly terrain next to the artificial light source. This perch seemed to provide the bird with cover from direct light while allowing an ample vantage point. The observed bird performed five suc- cessful foraging sallies in 30 min of obser- vation; all captured prey appeared to be sphinx moths. On two other occasions, we found presumably the same bird feeding near the same street light; 18 December 2004 and 1 January 2005. The bird used the same for- aging method each time, usually sallying up- wards or horizontally from its perch to attempt prey capture. Only on two occasions did the motmot succeed in direct aerial capture. In the remaining foraging attempts (40), it had prob- lems capturing prey in the air and usually would have a vertical downward chase to the lawn, where it occasionally succeeded and proceeded with feeding. We checked the area during six nights in August 2006 but did not find any motmots feeding in the area. The inferred capture efficiency was quite low. However, the large size of prey may com- pensate, as we observed the motmot did not need many feeding events to be apparently sa- tiated as it left the scene. The capture rate could reflect the lack of adaptative response, which a more specialized bird would have to couple with its new foraging behavior. Artificial lights and motmots are common in suburbs around San Jose and other nearby Costa Rican provinces. However, diurnal in- sectivorous birds attracted to artificial light sources during the night have not previously been recorded for Costa Rica. ACKNOWLEDGMENTS We thank Julio Sanchez and Gilbert Barrantes for encouraging preparation of this note. We also thank Christopher Canaday, Oliver Komar, and an anony- mous reviewer for helpful comments which greatly im- proved this manuscript. LITERATURE CITED Chacon-Madrigal, E. and G. B.arr.antes. 2004. Blue-crowned Motmot {Momotus momotd) pre- dation on a long-tongued bat (Glossophaginae). Wilson Bulletin 116:108-110. Garcl\-C, J. M. and R. a. Zahawi. 2006. Predation by a Blue-crowned Motmot {Momotus momota) on a hummingbird. Wilson Journal of Ornithology 118:261-263r Orejuela, J. 1980. Niche relationships between Tur- quoise-browed and Blue-crowned motmots in the Yucatan Peninsula. Mexico. Wilson Bulletin 92: 229-244. Remsen. j. V., M. Hyde, .and A. Ch.apm.an. 1993. The diets of neotropical trogons. motmots. barbets and toucans. Condor 95:178-192. Sketch, A. 1945. Life history of the Blue-throated Green Motmot. Auk 62:489-517. Sketch, A. 1947. Life history of the Turquoise-browed Motmot. Auk 64:201-217. Sketch, A. 1964. Life history of the Blue-diademed Motmot Momotus momota. Ibis 106:321-332. Sketch, A. 1971. Life history of the Broad-billed Mot- mot. with notes on the Rufous Motmot. Wilson Bulletin 83:74-94. Thurber, W. a. .and O. Ko.aear. 2002. Turquoise- browed Motmot {Eumomota superciliosa) feeds by artificial light. Wilson Bulletin 114:525—526. SHORT COMMUNICATIONS 655 The Wilson Journal of Ornithology 120(3):655-657, 2008 Kleptoparasitism by Grey Kingbirds (Tyrannus dominicensis) in Barbados Sarah E. Overington,i’2 paure Cauchard,^ and Kimberly- Ann Cote^ ABSTRACT — We observed Grey Kingbirds {Tyr- annus dominicensis) from late February to May 2007 stealing food items from the bills of Carib Crackles {Quiscalus lugubris). This behavior occurred at two baited walk-in bird traps on the grounds of Bellairs Research Institute of McGill University in St James, Barbados. Grey Kingbirds were not seen entering traps, but were regularly observed in tree branches near traps, often chasing Carib Crackles and Zenaida Doves {Zenaida aurita) as they exited the traps with food. We describe six instances of kleptoparasitism by Grey Kingbirds from Carib Crackles. To our knowl- edge, this is the first report of kleptoparasitism for this species. Received 18 July 2007. Accepted 30 Decem- ber 2007. The Grey Kingbird {Tyrannus dominicen- sis) is a mid-sized tyrant flycatcher found throughout the Caribbean, northern South America, and into southern North America (Smith and Jackson 2002). Its diet consists of insects, lizards, berries, fruits, and seeds (Smith and Jackson 2002), but it has also been observed at food scrap assemblages in Bar- bados (Lefebvre et al. 2001). Grey Kingbirds hunt using a typical flycatcher technique, de- scending from high perches to capture prey in mid-air (Smith and Jackson 2002). They may also take insects from the surface of water (Sprunt 1942, Lack 1976), capture fish in shal- low water and sand bars (Lefebvre and Spahn 1987), and have occasionally been observed feeding at night near lights and street lamps (Reader et al. 2002). Grey Kingbirds are noted for their aggression toward mammals and oth- er birds, particularly during the breeding sea- son as they defend their nests (Sprunt 1942). However, no cases of aggressive behavior in- volving food theft (kleptoparasitism; Brock- mann and Barnard 1979) have been described for this species and only a single case of klep- ' Department of Biology, McGill University 1205 Avenue Docteur Penfield, Montreal, QC H3A IBl, Canada. ^Corresponding author; e-mail: Sarah. overington @ mai 1 .mcgi 1 1 .ca toparasitism has been reported for the genus (i.e., Cassin’s Kingbird [T. vociferans] steal- ing worms from “Blackbirds”; Merriam 1896). No reports of kleptoparasitism by Grey Kingbirds were found in a recent exhaustive survey covering 856 field reports of klepto- parasitism in birds published since 1969 (Morand-Ferron et al. 2007). OBSERVATIONS We observed six acts of interspecific klep- toparasitism by Grey Kingbirds from late Feb- ruary to May 2007 at the Bellairs Research Institute of McGill University in St James, Barbados. All kleptoparasitic attacks occurred near two walk-in traps (IX 0.55 X 0.55 m) on the grounds of the research institute (13° 10' 60 N, 59° 38' 60 W). These traps were made of wire affixed to a wooden frame, and we baited them regularly with moistened dog food pellets and cooked rice. We placed food near the back of the traps so that birds had to move at least 30 cm into a trap to reach the bait, because this increased our success in capturing our target species, Carib Grackles {Quiscalus lugubris). The two traps had the same dimensions, but were ~200 m apart, separated by a large building. One was closer to the front entrance of the research institute and we refer to this as the “front trap” with the other being the “back trap”. Each trap was within 5 m of trees where Grey Kingbirds were seen perching. On 19 February 2007 at 1723 hrs AST, we saw a Grey Kingbird descend from a tree branch above the front trap and, in mid-flight, steal a moistened dog food pellet from the bill of an adult Carib Crackle that was walking out of the trap. On 17 March 2007, a Grey Kingbird descended from a tree above the back trap and stole food from an adult Carib Crackle without landing on the ground. On 21 March at 850 hrs, a Grey Kingbird stole a dog food pellet from the bill of an adult female Carib Crackle at the front trap. Between 1240 656 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 3, September 2008 and 1250 hrs, we observed two acts of klep- toparasitism at the back trap. First, a Grey Kingbird stole a dog food pellet from a Carib Crackle without landing on the ground. Sec- ond, a Carib Crackle released a dog food pel- let from its bill as a Grey Kingbird descended from the tree and the Grey Kingbird took the pellet from the ground. On 23 April at 1446 hrs, we observed a different kleptoparasitic technique at the back trap where a Grey King- bird chased a Carib Crackle and stole a dog food pellet while both birds were in flight. Kleptoparasitism was observed several times by other researchers at both traps until they were removed from the research grounds on 21 May 2007, but the details of these obser- vations were not recorded. Two Grey Kingbirds were often observed together around the research institute and near the traps, although they were not observed stealing food at the same time. All birds we observed stealing food had adult plumage. We were unable to identify males from females by sight, although Grey Kingbirds show some sexual size dimorphism (Haberman et al. 1991). None of the kleptoparasitizing individ- uals had any obvious features that would al- low us to distinguish between them and we cannot be sure of the number of individuals engaging in kleptoparasitism. Grey Kingbirds were regularly seen in trees near the traps throughout the trapping period (Feb-May 2007), but were not observed entering either trap and were rarely seen on the ground. Grey Kingbirds often arrived shortly after the traps had been baited and were observed chasing both Zenaida Doves (Zenaida aurita) and Car- ib Crackles from the trap area on at least four occasions in what appeared to be failed at- tempts to steal food. DISCUSSION Animals may steal food to gain access to items that would otherwise be difficult to ac- quire due to morphological or behavioral con- straints. For example, juvenile Eurasian Oys- tercatchers (Haematopus ostralegus) reduce their rates of stealing mussels (Mytilus edulis) from adult conspecifics as they become more efficient at opening the mussel shells (Goss- Custard et al. 1998). Even if a species is ca- pable of performing difficult foraging tasks, kleptoparasitism may allow it to access food items at a lower cost. Food stealing by gulls (Laridae) allows them to obtain foods such as fish (Duffy 1980), mussels (Khatchikian et al. 2002), and refuse (Burger and Gochfeld 1976) without performing the diverse foraging tech- niques required to access these items (e.g., diving, cracking open hard shells, digging). This may also be the case for the Grey King- birds we observed. The design and dimen- sions of our walk-in traps made it impossible for individuals to retrieve dog food pellets us- ing their usual in-flight foraging strategy. Consequently, kleptoparasitism represented an alternative foraging strategy for these birds, allowing them to access a valuable (protein- rich) food source. The Grey Kingbird is a territorial species that aggressively chases other birds and mam- mals from nesting areas. The food stealing and chasing we observed may be a modifi- cation of existing foraging behavior as well as agonistic behavior toward other birds within the breeding territory. There was a Grey King- bird nest within 50 m of one of the two traps where we observed kleptoparasitic behavior, although we could not confirm this nest be- longed to the individuals engaged in food stealing. No studies have been conducted on the population of Grey Kingbirds at this lo- cation and we cannot speculate on territory size or whether individuals stealing food were territory owners. Our research group has been trapping birds at Bellairs Research Institute for several de- cades, but neither stealing nor eating of dog food pellets by Grey Kingbirds have been ob- served before (Louis Lefebvre, pers. comm.). Our repeated observations suggest that food stealing is an opportunistic foraging technique used by at least one individual at our field site. Opportunistic feeding behavior may contrib- ute to the success of the Grey Kingbird in Bar- bados, an island dominated by human-modi- fied habitats. ACKNOWLEDGMENTS We thank the staff at Bellairs Research Institute for their support and encouragement. Comments from R A. Buckley and an anonymous reviewer improved the manuscript greatly. We thank Julie Morand-Lerron, Louis Lefebvre, and Neeltje Boogert for helpful com- ments on a previous version of this work. This research was funded by NSERC through a graduate fellowship to SEO and a Discovery grant to Louis Lefebvre. SHORT COMMUNICATIONS 657 LITERATURE CITED Brockmann, H. J. and C. J. Barnard. 1979. Klepto- parasitism in birds. Animal Behaviour 27:487- 514. Burger, J. and M. Gochfeld. 1976. Age differences in Ring-billed Gull kleptoparasitism on starlings. Auk 96:806-808. Dueey, D. C. 1980. Patterns of piracy by Peruvian sea- birds. Ibis 122:521-525. Goss-Custard, j. D., j. T. Cayeord, and S. E. G. Lea. 1998. The changing trade-off between food find- ing and food stealing in juvenile oystercatchers. Animal Behaviour 55:745-760. Haberman, K., D. I. Mackenzie, and J. D. Rising. 1991. Geographic variation in the Gray Kingbird. Journal of Field Ornithology 62:117-131. Khatchikian, C. E., M. Favero, and A. I. Vassallo. 2002. Kleptoparasitism by Brown-hooded Gulls and Grey-hooded Gulls on American Oystercatch- ers. Waterbirds 25:137-141. Lack, D. 1976. Island biology: illustrated by the land birds of Jamaica. University of California Press, Berkeley, USA. Leeebvre, L. and D. Spahn. 1987. Gray Kingbird pre- dation on small fish {Poecilia sp.) crossing a sand- bar. Wilson Bulletin 99:291-292. Leeebvre, L., S. Reader, and S. J. Webster. 2001. Gray Kingbirds join a bread scrap assemblage in Barbados. Bulletin of Ornithology 1211:247-249. Merriam, F. a. 1896. Notes of some of the birds of Southern California. Auk 13:115-124. Morand-Ferron, j., D. Sol, and L. Leeebvre. 2007. Food-stealing in birds: brain or brawn? Animal Behaviour 74:1725-1734. Reader, S. M., J. Morand-Ferron, I. Cote, and L. Leeebvre. 2002. Unusual feeding behaviors in five species of Barbadian birds. El Pitirre 15:1 17- 123. Smith, G. A. and J. A. Jackson. 2002. Gray Kingbird (Tyrannus dominicensis). The birds of North America. Number 668. Sprunt Jr., a. 1942. Tyrannus dominicensis domini- censis (Gmelin) Gray Kingbird. Pages 29-50 in Life histories of North American flycatchers, larks, swallows, and their allies (A. C. Bent, Ed- itor). U.S. National Museum Bulletin Number 179. The Wilson Journal of Ornithology 120(3):657-659, 2008 Double-scratching by Yellow-headed Blackbirds Alan de Queiroz’ ABSTRACT — Some ground-foraging birds, includ- ing most New World sparrows in the tribe Emberizini, uncover food items in litter by double-scratching, a backward hop that displaces the litter posteriorly. I re- port double-scratching by Yellow-headed Blackbirds (Xanthocephalus xanthocephalus), a species not pre- viously known to exhibit this behavior. Double- scratching by one individual that was observed in de- tail occurred with the bill pointed downward with its tip near or touching the ground, as in other icterines but unlike the behavior in Emberizini. This individual, between double-scratches, also used bill-sweeping to displace litter. Double-scratching has a similar form in at least three of the four icterine species reported to show the behavior, suggesting the trait is homologous among the.se species. However, phylogenetic relation- ships imply that double-scratching evolved conver- gently in these taxa. Received 6 June 2007. Accepted 5 January 200H. ' 826 Delmar Way, Reno, NV 89509. USA; e-mail: alandqz@yahoo.com A number of ground-foraging birds use a distinctive double-scratching behavior to un- cover food items under litter or other material. Double-scratching, or bilateral scratching (Greenlaw 1976, 1977), refers specifically to synchronous movement of both legs to scratch away surface material, as opposed to using one leg at a time. This behavior typically in- volves a forward hop immediately followed by a backward hop (the scratching stroke) that displaces litter posteriorly (Hanison 1967; Greenlaw 1976, 1977). Double-scratching is especially widespread in New World sparrows of the tribe Emberi- zini and seems to occur in few species outside of this group (Davis 1957; Harrison 1967; Greenlaw 1976, 1977; Hailman 1984; Whalen and Watts 2()()(); | taxonomy follows Klicka et al. 2()()()1). Among icterines, double-scratching has been reported for only three species — Red-winged Blackbird (Agelains phoeniceiis). 658 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 Brown-headed Cowbird {Molothrus ater), and Shiny Cowbird (M. bonariensis) (Greenlaw 1976). Red-winged Blackbirds and Brown- headed Cowbirds apparently use the behavior somewhat commonly (Greenlaw 1976). I report double-scratching by Yellow-head- ed Blackbirds (Xanthocephalus xanthocephal- us), a species distantly related to other icter- ines known to show this behavior (Lanyon and Omland 1999, Jpnsson and Fjeldsa 2006), and briefly discuss the phylogenetic history of the behavior within Icterini. I also report a Yellow-headed Blackbird alternately using double-scratching and bill-sweeping on the same substrate, a specific combination of be- haviors that apparently has not been reported previously in any bird species. OBSERVATIONS Birds were observed in a yard in Ely, White Pine County, Nevada, USA (39° 13' 30.5" N, 1 14° 52' 19.5" W). On the afternoon of 6 May 2007, a large flock of Brown-headed Cow- birds and at least four Yellow-headed Black- birds flew into the yard and fed on bird seed (a mix of millet, cracked com, and sunflower seeds) that had been cast onto the lawn and into a shrub bed covered with bark (mostly pieces <1 cm wide by 2 cm long). Many of the cowbirds and at least two of the Yellow- headed Blackbirds double-scratched in the shrub bed to displace the bark and uncover the seeds. I watched an immature male Yellow-headed Blackbird from 1610 to 1615 hrs PST on 6 May perform at least a dozen double-scratches in the shrub bed. In double-scratching, the bird made a forward hop immediately fol- lowed by a backward hop. The legs of the bird were usually offset at the beginning of the double-scratch (i.e., one leg was farther for- ward than the other) and this offset position at times remained at the end of the backward hop. The movements of the two legs did not seem to be perfectly synchronous in some in- stances, although evaluating synchrony was difficult. The blackbird’s double-scratching movements were clearly slower than those of cowbirds, a Green-tailed Towhee {Pipilo chlo- rurus), and several White-crowned Sparrows (Zonotrichia leucophrys) that were also feed- ing in the shrub bed on the same afternoon. All of these species also fed on the lawn, but only the Green-tailed Towhee was observed double-scratching on that substrate. I watched an adult male Yellow-headed Blackbird feeding in the same shrub bed from -1310 to -1340 hrs PST on 8 May 2007. I focused my observations of 8 May on the po- sitions of the bill and legs. This adult male kept its head pointed downward with the tip of the bill touching or nearly touching the ground while double-scratching. The legs of the bird were at times obviously splayed at the end of a double-scratch. The backward hop propelled bark or other litter behind the bird and the wings of the bird were at times held slightly away from the body. As on 6 May, I had the impression the bird did not always move its legs synchronously while double- scratching. I performed a timed 5 -min watch of this bird from 1334 to 1339 hrs to obtain a sample of the frequency of double-scratching while the bird was feeding continuously. The bird double-scratched four times during this period and also, from a position with the bill tip touching or nearly touching the substrate, swept its bill upwards many times with both a forward and a lateral component. These bill- sweeping movements also displaced litter. DISCUSSION This report, to my knowledge, represents the first published record of double-scratching by Yellow-headed Blackbirds. This behavior in this species appears to be similar to that reported by Greenlaw (1976) for Red- winged Blackbirds and Brown-headed Cowbirds. These three icterine species hold the bill pointing downward with the tip near or touch- ing the substrate and they often end the back- ward hop with legs splayed. All three may also move the legs asynchronously (although this needs to be verified for Yellow-headed Blackbirds). The position of the bill during double-scratching is clearly different from that in New World sparrows (Emberizini), which typically double-scratch with the bill held well above the substrate (Greenlaw 1976). The splayed legs and asynchronous leg move- ments may also distinguish icterine double- scratching from that of most members of the Emberizini, although some species in the lat- ter group show these characteristics in some circumstances (Greenlaw 1976). SHORT COMMUNICATIONS 659 The positions of the head and legs suggest the trait is homologous in the three double- scratching icterine species (Red-winged Blackbird, Brown-headed Cowbird, Yellow- headed Blackbird) for which the behavior has been described in detail. However, estimates of phylogenetic relationships within the icter- ines (Lanyon and Omland 1999, Jpnsson and Fjeldsa 2006) suggest double-scratching has evolved independently in these three taxa. One explanation for this apparent contradic- tion is that double-scratching may be present but unreported in other icterine species. The assumption is that, if the taxonomic distribu- tion of double-scratching was more accurately known, phylogenetic analyses would show the trait has arisen only once within the group. This explanation seems plausible considering that, until the 1970s, double-scratching had been reported in only one icterine species, the Shiny Cowbird. It is now known to occur in three other common and conspicuous species within this group (Greenlaw 1976, this study). Observations of other icterine species feeding in conditions that might elicit double-scratch- ing would help clarify this issue. Red-winged Blackbirds, Brown-headed Cowbirds, and Yellow-headed Blackbirds have all been ob- served double-scratching in bark litter (Green- law 1976, this study) suggesting one context in which the behavior might be observed in other icterines. The alternating use of double-scratching and bill-sweeping on the same substrate ap- parently has not been reported previously for any bird species. Other species are known to use both double-scratching and bill-sweeping to move litter or other material, but not in the manner described in this paper. Egyptian Plo- vers {Pluvianus aegyptius) use both behaviors, but in different feeding contexts (Howell 1979). Some species of brush finches (Atla- petes) also double-scratch and bill-sweep, but it is unclear whether individual birds use both behaviors (Paynter 1978). Three thrush {Tar- dus) species are known to double-scratch si- multaneously with bill-sweeping (Clark 1983). An obvious question concerning Yel- low-headed Blackbirds is whether switching from one behavior to the other is tied to some subtle difference in context, such as small dif- ferences in the size of pieces of litter. ACKNOWLEDGMENTS I thank M. C. Witmer and T A. Forbis for helpful comments on the manuscript and M. C. Witmer for references on Yellow-headed Blackbirds. LITERATURE CITED Clark Jr., G. A. 1983. An additional method of for- aging in litter by species of Turdus thrushes. Wil- son Bulletin 95:155-157. Davis, J. 1957. Comparative foraging behavior of the Spotted and Brown towhees. Auk 74:129-166. Greenlaw, J. S. 1976. Use of bilateral scratching be- havior by emberizines and icterids. Condor 78: 94-97. Greenlaw, J. S. 1977. Taxonomic distribution, origin, and evolution of bilateral scratching in ground- feeding birds. Condor 79:426-439. Hailman, j. P. 1984. Effect of litter on leaf-scratching emberizines. Wilson Bulletin 96:121-125. Harrison, C. J. O. 1967. The double-scratch as a tax- onomic character in the Holarctic Emberizinae. Wilson Bulletin 79:22-27. Howell, T. R. 1979. Breeding biology of the Egyptian Plover, Pluvianus aegyptius. University of Cali- fornia Publications in Zoology 113:1-76. J0NSSON, K. A. AND J. Fjeldsa. 2006. A phylogenetic supertree of oscine passerine birds (Aves: Passed). Zoologica Scripta 35:149-186. Klicka, j., K. P. Johnson, and S. M. Lanyon. 2000. New World nine-primaried oscine relationships: constructing a mitochondrial DNA framework. Auk 117:321-336. Lanyon, S. M. and K. E. Omland. 1999. A molecular phylogeny of the blackbirds (Icteridae): five line- ages revealed by cytochrome-/? sequence data. Auk 116:629-639. Paynter Jr., R. A. 1978. Biology and evolution of the genus Atlapetes (Emberizinae). Bulletin of the Museum of Comparative Zoology 148:323-369. Whalen, D. M. and B. D. Watts. 2000. Interspecific variation in extraction of buried seeds within an assemblage of sparrows. Oikos 88:574-584. The Wilson Journal of Ornithology 1 20(3 ):660— 665, 2008 Ornithological Literature Compiled by Mary Gustafson and Clait E. Braun THE ORNITHOLOGIST’S DICTIONARY. By Johannes Erritzoe, Kaj Kampp, Kevin Winker, and Clifford B. Frith. Lynx Edicions, Barcelona, Spain. 2007: 284 pages. ISBN: 9788496553439. $25.00 (paper).— Lynx Edi- cions, based in Barcelona, Catalonia, are best- known to birdwatchers and ornithologists as publishers of the monumental ''Handbook of the Birds of the World”, but their range in bird-related works is large and growing with numerous new titles in English, Spanish, and (I am glad to say) Catalan appearing every year. This latest work is another useful addi- tion to their catalogue. As the name suggests, this is specifically a dictionary. Aside from a six-page appendix listing bird families with the number of spe- cies in each, the entire book is devoted to an alphabetic treatment of some 5,000 words and terms. Each entry is terse and to-the-point; this is not a collection of essays, but a collec- tion of definitions and facts, expressed as con- cisely as possible. The authors seem to have gone out of their way to make their selection varied and encompassing, including such dis- parate items as non-obvious abbreviations, such as per se and incertae sedis, ornitholog- ical societies and publications, behavior pat- terns, bird diseases, a few of the more out- standing luminaries of biology such as Dar- win and Linnaeus, anatomical terms, Malay- sian mythology, advances in nucleic acid science, important fossils, some obscure words to describe habitats, etc. The net is cast very widely indeed. The similarity between the titles invites an obvious comparison between the present work and Sir Arthur Landsborough Thomson’s clas- sical "A New Dictionary of Birds” published more than 40 years ago, but this is scarcely a valid comparison. The “New Dictionary” was in fact more in the nature of an encyclopedia rather than what most people would under- stand as a dictionary with essays, rather than entries, contributed by some 200 distinguished scientists, and with numerous illustrations, photographs, and color plates, in more than 900 large-format pages. “The Ornithologist’s Dictionary” is precisely that, a far more con- cise venue for laconically-expressed facts and definitions, with the artwork confined to one little vignette at the head of each letter of the alphabet. How well does “The Ornithologist’s Dic- tionary” fulfill its objectives? In short, I think very well. Obviously, a determined reviewer with an obsession for nit-picking could com- ment on omissions — among the publications, ''Birds of the Western Palearctic” and "Co- tinga”-, and would doubtless dwell on minor errors — “owlet” is defined as a “young owl”, but it is also used for four genera of small strigids, and Canadians may be a little mor- tified to see the Christmas Bird Count de- scribed as “an annual bird count in USA”; last year, notwithstanding temperatures of minus 40° in some locations, there were 371 Christ- mas counts in Canada, per capita (a term which is defined in this dictionary) twice as many as were conducted by our more delicate cousins to the south! “The Ornithologist’s Dictionary” is a thor- oughly workmanlike and useful publication. It will not become preferred bedtime reading, but as a daily tool it deserves a place on the desk of any serious ornithologist, whether profes- sional or amateur. — DAVID BREWER, RR 1, Puslinch, ON NOB 2J0, Canada; e-mail: mbrewer@albedo.net HANDBOOK OF BIRDS OF THE WORLD, VOLUME 12. PICATHARTES TO TITS AND CHICKADEES. Edited by Josep del Hoyo, Andrew Elliott, and David Christie. Lynx Edicions, Barcelona, Spain. 2007: 815 pp., 56 color plates, numerous color photo- graphs and range maps. ISBN: 8496553426. $265 (cloth). — This large format (24 X 31 cm; 4.6 kg) 12th volume in this remarkable series begins with an essay entitled Fossil Birds by Kevin Caley. It is 45 pages in length, has a Glossary, the References section con- 660 ORNITHOLOGICAL LITERATURE 661 tains 335 citations, and the essay is an excel- lent summary of what we know (and perhaps more importantly, what we do not know) about fossil birds. Included is a discussion of the evolution of birds (including the “tree- down” “ground-up” controversy) and the au- thor concludes: “. . . [the] general opinion is that the evidence for birds being descendents of theropod dinosaurs is overwhelming.” The text is supplemented by line-drawings by the author. The bulk of the text consists of family and species accounts of the Picathartidae (Pitha- cartes), Timaliidae (babblers), Paradoxornithi- dae (parrotbills), Pomatostomidae (Australa- sian babblers), Orthonychidae (logrunners), Eupetidae (jewel-babblers and allies), Pachy- cephalidae (whistlers), Petroicidae (Australa- sian robins), Maluridae (fairy-wrens), Dasyor- nithidae (bristlebirds), Acanthizidae (thorn- bills), Epthianuridae (Australian chats), Neo- sittidae (sittellas), Climacteridae (Australasian treecreepers), and Paridae (tits and chicka- dees). The family accounts are extensive, of- ten exceeding the length of the combined spe- cies accounts, and are lavishly illustrated with stunning color photographs. The family ac- counts vary in length, with the longest, the Timaliidae (with 309 species), occupying 64 pages including 95 photographs (some full- page), and several accounts (e.g., the Orthon- ychidae) with eight pages and less than a doz- en photographs. The family accounts typically begin with a summary box that details Class, Order, Suborder and Eamily, a small map with world distribution in red, and a brief descrip- tion of the characteristics of the family, hab- itat, number of genera, species, and subspe- cies, and conservation status. The sections of each family accounts generally are: System- atics. Morphological Aspects, Habitat, Gen- eral Habits, Voice, Food and Feeding, Breed- ing, Movements, Relationship with Man, and Status and Conservation. There are no in-text references, but a General Bibliography at the end of each family account gives abbreviated references (corresponding full citations are in the References at the end of the book). The family accounts have a wealth of information including such things as debate over vernac- ular names. For example, in the Acantizidae account, the shift (which 1 have long opposed) from the traditional “Warbler” to “Gery- gone” (the genus name) of a group of acan- thizides is explained and the meaning of “Gerygone” given: “bom of sound” or “bom of song.” The species accounts (623 in this volume) typically begin with the common name, fol- lowed by the scientific name, the common name in French, German, and Spanish, other common names, and sections on Taxonomy, Distribution, Descriptive Notes (that include voice). Habitat, Food and Feeding, Breeding, Movements, and Status and Conservation. The account concludes with a bibliography with shortened citations as in the family ac- counts. The accounts of species I have worked with and know best (e.g., Tasmanian Scrub- wren [Sericornis humilis]. Slender-billed Thornbill [Acanthiza iredalei], Scrubtit [Acan- thornis magna]. Rufous Treecreeper [Climac- terus rufus^) were excellent with accurate and pertinent information and citing the major ref- erences. Each species account includes a range map in up to three colors: yellow for breeding distribution, blue for non-breeding, and green for year-round. The 56 color plates by Norman Arlott, Hi- lary Burn, John Cox, Ren Hathway, Ian Lew- ington, Chris Rose, Jan Wilczur, and Tim Worfolk, which accompany the species ac- counts, are uniformly excellent. As in previ- ous volumes, the plates are on even-numbered pages and many have text or photographs printed on the previous page (back-side of the sheet). Thus, there is a faint image of print lines showing through onto the plate, but I found this not to be a distraction. The plates are interspersed in the corresponding text so the image is not more than a few pages from the text. Where appropriate, subspecies are il- lustrated as well as females of dimorphic spe- cies. For example, in the Varied Sittella {Da- phoenositta chrysoptera), all hve Australian subspecies are illustrated and two of hve have female as well as male images; two of hve New Guinea subspecies are illustrated, one with a female image. The References section is divided into References of Scientific De- scriptions (1,020 entries) and General List of References (>3,000 full-citation entries). Both are gold mines of information. The index has individual species double-entered, for exam- ple: Varied Sittella and Sittella, Varied — a user-friendly feature. 662 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 This volume — and the entire series — rep- resent a remarkably bold initiative. To pro- duce comprehensive accounts of all families of birds and their nearly 10,000 species is a project of enormous proportions. That Lynx Edicions and their editorial staff have been able to bring together a product of such high quality is remarkable. The necessarily high price may limit the number of individuals who acquire the series, but every library ought to have one.— WILLIAM E. DAVIS JR., Profes- sor Emeritus, Boston University, 23 Knoll- wood Drive, East Falmouth, MA 02536, USA; e-mail; wedavis@bu.edu FALCONRY AND HUNTING IN ARA- BIA. By Paris Al-Timimi. Hancock House Publishers, Blaine, Washington, USA. 2007: 320 pages, 442 photographs. ISBN 978-0- 97388-530-9. $120.00 (cloth).— Al-Timimi became a veterinarian at the University of Baghdad and for many years operated a clinic for trained falcons in Qatar, and more recently in Saudi Arabia. His purpose, as a non-fal- coner, is to reveal traditional attitudes and cus- toms of Arab falconers, and the nature of their hunting falcons and their quarry. The narrative often rambles, and topics are only loosely as- sorted by chapter, and sometimes not at all by paragraph. This unusual style might be ex- cused because English is not the author’s pri- mary language. Chapters include a historical account of hunt- ing in Arabia (but with few dates), the character of Arab hunters (really a description of how they go about falconry), an exhaustive descrip- tion of their falcons (mainly Saker [Falco cher- rug] and Peregrine [F. peregrinus] falcons), and a graphically illustrated chapter on diseases (of- ten worse-case examples) of captive falcons. Plastic “astroturf”, of the type commonly used as doormats, was introduced to Arabia (in Du- bai) as a covering on falcon perches. It greatly reduced the incidence of disastrous infections on the bottom of the feet, not so much because it allows air flow under the feet, as the author as- serts, but because the many plastic fingers cush- ion the feet and distribute the weight of the sed- entary falcon more evenly, reducing pressure on prominent areas, and enhancing vascular cir- culation. Another interesting but seemingly unrelated chapter describes Arab expeditions to Somalia in 1982 and Sudan in 1983. The main purpose of the latter was apparently to obtain antelope such as oryx {Oryx spp.), addax (Addax naso- maculatus), and gazelle (Gazella spp.) for pri- vate zoos in Qatar. The expeditions are illus- trated by 44 color photos, most of which are of antelope. There is a chapter on taming and ini- tial training of falcons. In pattern, these methods are not unlike those used in the West, except for the miserable practice of using a single suture, through both lower eyelids, passed over the top of head and tied to hold them up, preventing the falcon from seeing its new surroundings. West- ern falconers use a hood (probably introduced to the West by Arabs in the Crusades) exten- sively in the first days of training to achieve the same purpose without risk of tearing and scar- ring the eyelids. Arab falconry involves only falcons; hawks are rarely used. Sakers are preferred to chase Houbara Bustards {Chlamydotis undulata) in direct flight from the falconer’s glove. Female Peregrines are used less frequently but have become more popular. First-year migrant fal- cons are more prevalent, but adults (haggards) of both species are trapped and trained. Per- egrines are apparently more fragile than Sa- kers and their monetary value plummets at the end of the hunting season because they do not hold up well through the long molt (Mar- Oct). Sakers are more durable in what the au- thor calls “harsh” treatment by Arab falcon- ers. Outstanding hunters are retained from season to season. Unwanted falcons are sold at greatly reduced rates or set free, the latter to uncertain fates. In Arab falconry there is great emphasis on the appearance of each falcon, especially Sa- kers. As in the similar Gyrfalcon {Falco rus- ticolus), Sakers have immense plumage vari- ation. Pale individuals are highly prized, dark birds a bit less so, and lastly there are the average individuals. The author maintains that plumage color and pattern are unrelated to performance, contrary to the widely held su- perstitions of Arab falconers. There are more than 178 color photographs of captive falcons in the chapter “The Arabs and Their Hunting Falcons”, many of them full-page. Some of the birds are hooded. Included are photos of a few dozen Gyrfalcon and Saker study skins ORNITHOLOGICAL LITERATURE 663 held by the British Museum. The assumption is that Saker color and pattern variants come from different regions and are different “forms”. Unfortunately, the living birds and most of the skins were from wintering areas and could not reveal regional patterns in nest- ing areas. The author, Al-Timimi, maintains that fal- conry in Arabia is now the “sport” of a large proportion of the population, commoner and royalty alike. Falconry was certainly practiced among the nomads (Bedouin) and, after WW II, grew in favor among the royal families whose wealth created huge collections of trained falcons and their caretakers. In the last two decades, no doubt several thousands of falcons were in captivity in the Gulf Region in any year. Because of high turnover in the captive population, a few thousand wild Saker and Peregrine falcons were supplied by deal- ers each year to replace the losses. Dealers in Pakistan, Syria, Iraq, and Iran smuggled fal- cons to the Arabian Peninsula, sometimes with poor results for the birds. Dealers have expanded their quest for falcons across the whole of Central Asia, east to China where Sakers are numerous. The author does not cover up the conservation concerns, but he provides only his impressions, and few data. Overall, the situation is changing. Captive- bred Gyrfalcons and gyr-hybrids with Pere- grine and Saker falcons have grown greatly in popularity, not so much because of increased effectiveness in hunting, but because many are phenomenally handsome, especially in the eyes of sheikhs who pay great prices just to have them in their collection. These captive- bred falcons are imported legally from the West by the hundreds annually. But at the same time, many of those who can afford the costs have become distracted by involvement in their exploding economies. Falconry ap- pears to be on the decline among the entitled families. And then there are the “poor Houbara”. They now migrate into the Peninsula so rarely that a sheikh may pay a huge reward for in- formation on the whereabouts of one to hunt. Further, as photos in the book show, Pakastani trappers net hundreds, intending to smuggle them to the Peninsula where they are used as live lures to train falcons. Apparently, many die in transit. Houbara populations have crashed, and there are few alternate species the Arabs wish to hunt. Arabs have been forced to mount, at great cost, large hunting expeditions for Houbara to regions from Lib- ya to Pakistan. This book will not likely serve as a reference, but it provides considerable insight into the way falconry works in Arabia. These workings are markedly different from those in the West. Many North American falconers do not use fal- cons, but use hawks instead. On reading this account, falconers in the West who actually hunt with their birds, will come to more deeply ap- preciate their situation. As a life-long falconer, I am happy for the variety of raptors Western falconers use, the abundances of quarry that match the raptors and the abilities of falconers, the commitments of the falconers to perfection of the art, the fine veterinary care available, and the many regulations that effectively govern uses of our raptor and prey resources. I thank J. D. Remple for comments on this review. — JAMES H. ENDERSON, Professor Emeritus, Colorado College, 3215 Austin Drive, Colorado Springs, CO 80909, USA; e-mail: jenderson@ coloradocollege.edu SWAN. By Peter Young. Reaktion Books, Ltd., London, UK. 2008: 200 pages, 45 black and white plates, and 55 color plates, ISBN 978-1-86189349-9. $19.95 (paper).— This book is part of a series entitled ANIMAL ed- ited by Jonathan Burt. Although the book in- cludes biological and conservation informa- tion, there are many more pages dedicated to the documentation of the influence of swans on and interactions with humans. The first chapter, “Clamorous Wings”, in- cludes a brief review of swan biology and nat- ural history. Although mostly accurate, the bi- ological treatment is superficial. It is stated in one place that eight species survive, but notes that ornithologists disagree on seven or nine species. No attempt is made to specify the “species” to which the disagreements refer. Chapter 2, “Grace and Favour” is based on human observations of swans and begins with some of the early taxonomic works with some emphasis on illustrations of swans that accom- panied the texts. Much of this chapter dis- 664 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 cusses swans in literature and art, as well as human impressions of swans. Swan transformations (Chapter 3) are the stuff of fairy tales (The Ugly Duckling), my- thology, and legend. Swan transformations ap- pear in numerous cultures on almost every continent. Most are familiar with Leda and the Swan from Greek mythology where Zeus, in the form of a swan, seduces Leda. This story has been the subject of many works of art and is referred to in all kinds of literary works from poetry to novels. Numerous other swan/ human transformations are discussed. History (Chapter 4) discusses chronological representations of swans including using swans in place names, architecture, reference to objects, and influence on historical or sci- entific discoveries. For example, the “swan- necked” flask was used by Louis Pasteur to demonstrate that germs exist in the atmo- sphere. In 1727, Thomas Young observed the intersecting ripples set up by swans in a pond in Cambridge. He believed that light traveled in waves and interacted much the same as the waves generated by swans on the surface of the water. Young subsequently demonstrated interference patterns through light experi- ments supporting this wave theory of light. Chapter 5 (Hazards) could have combined with conservation, as it concerns a combina- tion of exploitation of swans by humans, and injuries and deaths caused by human related products or activities. Included in the discus- sion are shooting by vandals, the hazards of fish hooks, fishing line, oil and chemical spills, lead poisoning by ingestion of shotgun pellets and fishing sinkers, and collisions with structures such as power lines, wind turbines, and communications towers. This chapter also discusses uses of swans and swan parts from the 1770s to the early 1990s that are strongly implicated in the decimation of some species like the Trumpeter Swan (Cygnus buccinator). Swan feathers made the best quill pens, down was used to stuff pillows and mattresses, and flight feathers were used for hats, fans, and muffs. The Hudson’s Bay Company shipped over 17,000 swan skins from Canada between 1853 and 1877. Also, discussed in this chapter are the traditional and cultural uses of swan feathers or parts by aboriginal or tribal groups for food, healing, and religious purposes or for practical applications like tying fishing flies. Swans have been used for food (Chapter 6) for thousands of years. In the Old Testament, swans are classified as “unclean” along with other animals with dark meat. Included in this chapter are instructions for raising, preparing, cooking, and serving swans. In England, swans were often prepared for banquets and Christmas dinners until the arrival of the tur- key from the New World in the sixteenth cen- tury. Chapter 7 discusses swan conservation. The earliest legal protection of swans appears to have occurred in England. There is evidence that some Mute Swans {Cygnus olor) were owned in England prior to the year 1000 and that they had developed a way of marking them. Mute Swans had royal protection under a 1483 act of King Edward IV. The Crown’s right over swans continues to this day under the Wild Creatures and Forest Laws Act of 1971. Protection in Canada and the United States began with the signing of the Migratory Bird Treaty in 1916. The author makes no mention of the subsequent treaties between the U.S. and Mexico, Japan, and the Soviet Union. This chapter discusses refuges and re- serves established for swan protection in var- ious countries, such as Red Rock Lakes Na- tional Wildlife Refuge in the U.S. One im- portant reserve not mentioned, is the Xinjiang Bayinbuluke Swan Lake Nature Reserve in China. The recent controversy in the U.S. over Mute Swan control is mentioned in this chap- ter. A court ruled that Mute Swans, a species introduced by man into North America, were protected under the Migratory Bird Treaty. The author does not report on a recent federal law that removed introduced migratory bird species from federal protection. He also char- acterized the court decision as “In the end the hunters lost”. It was never really a hunting issue. The final chapter (Promotion) concerns the use of swans in promoting businesses, prod- ucts, and organizations. Because of their pos- itive image of elegance and grace, swans im- ply quality. It is not surprising that swans are used as a logo for advertising. For some reason, the author omitted refer- ence to Banko’s classic monograph. The Trumpeter Swan (1960, North American Fau- na, Number 63). Also missing is A. Lindsey Price’s Swans of the World, in Nature, His- ORNITHOLOGICAL LITERATURE 665 tory, Myth and Art (1995, Council Oak Books). A reader looking for a biological summary of swans of the world will be dis- appointed. If, however, one is interested in the influence on and interaction with human civ- ilization, you will enjoy this interesting little book. — JOHN E. CORNELY, The Trumpeter Swan Society, 3 Cliffrose, Littleton, CO 80127, USA; e-mail: johncomely@msn.com BIRDS: THE ART OF ORNITHOLOGY. By Jonathan Elphick. Rizzoli International Publications Inc., 300 Park Avenue South, New York, USA. 2008: 336 pages and 300 illustrations. ISBN: 978-0-8478-3 1 34-0. $19.95 (hardcover). — This volume is indeed broad in scope. It spans the millennia of bird art from Paleolithic cave wall paintings in southern France and northern Spain, and im- ages left as frescos in Egyptian tombs through the present generation of bird painters. It is a formidable undertaking, and the author is to be commended for completing an admirable task. Text and illustrations trace the chrono- logical development of this genre from early, stylized depictions to the more naturalistic im- ages of the late 19th and 20th centuries. As the author alludes to in the preface, with a volume of such broad scope, the most dif- ficult task certainly would be who to include and who to exclude. In the end, any volume tracing the development of an art form de- pends on the desire of the author to include artists who seem to have made a significant contribution to the evolution of the subject matter. I feel Mr. Elphick has made a sensible selection of bird painters both past and pre- sent. The first impression one receives upon pe- rusing the volume is its rather small size, 140 X 160 mm (SVi by 6V4 in.). Apparently, di- mensions of the original, much larger edition were reduced to make the publication more affordable. This is a laudable decision; how- ever, the smaller format might be difficult for some to read. The illustrations are excellent in quality and are not overly affected by the re- duced size. Many of the artists discussed, some at length, are represented with reproductions of their work; others are not. For example, Bod- ani’s painting Two Icelandic Falcons, vividly described by the author, is not shown. In a visual world we are too often left with the written word, and word descriptions can not take the place of the painted image. Under- standably, the practical nature of a volume of this breadth limits the number of illustrations that can be included. The artist’s approach to bird illustration ver- sus the painting of the subject will be debated for a long while. Suffice it to say there is a difference. Generally speaking, illustrations are more documentary in nature and do not attempt to exploit the medium to the same ex- tent as paintings. It might have been interest- ing to show a reproduction of a Liljefors’ oil or Carel Fabritius’ image of the European Goldfinch (Carduelis carduelis) as examples of the more painterly qualities of bird art. The scope and visual imagery of this vol- ume remind readers that bird art has reflected world cultures since prehistoric times. Those who appreciate the beauty as well as the iden- tity of birds will want to add Birds: The Art of Ornithology to their library. — DON RA- DOVICH, Professor Emeritus, Western State College of Colorado, 17232 Woodgate Road, Montrose, CO 81403, USA; e-mail: canyon. creek@earthlink.net THE WILSON JOURNAL OF ORNITHOLOGY Editor CLAIT E. BRAUN Editorial Board RICHARD C. BANKS 5572 North Ventana Vista Road Tucson, AZ 85750-7204 E-mail; TWILSONJO@comcast.net KATHY G. BEAL JACK CLINTON EITNIEAR SARA J. OYLER-McCANCE Editorial NANCY J. K. BRAUN Review Editor MARY GUSTAFSON Assistant Texas Parks and Wildlife Department 2800 South Bentsen Palm Drive Mission, TX 78572, USA E-mail; WilsonBookRevievv@aolcom GUIDELINES FOR AUTHORS Please consult the detailed “Guidelines for Authors” found on the Wilson Ornithological Society web site (http;//www. wilsonsociety.org). All manuscript submissions and revisions should be sent to Clait E. Braun, Editor, The Wilson Journal of Ornithology, 5572 North Ventana Vista Road, Tucson, AZ 85750-7204, USA. The Wilson Journal of Ornithology office and fax telephone number are (520) 529-0365. The e-mail address is TWilsonJO@comcast.net. Notify the Society immediately if your address changes. Send your complete new address to Ornithological Societies of North America. 5400 Bosque Boulevard, Suite 680, Waco, TX 76710. The permanent mailing address of the Wilson Ornithological Society is; %The Museum of Zoology, The University of Michigan, Ann Arbor, MI 48109. Persons having business with any of the officers may address them at their various addresses given on the inside of the front cover, and all matters pertaining to the journal should be sent directly to the Editor. Membership inquiries should be sent to Timothy J. O’Connell, Department of Natural Resource Ecology and Management, Oklahoma State University, 205 Life Sciences West, Stillwater, OK 74078; e-mail; oconnet@ The Josselyn Van Tyne Memorial Library of the Wilson Ornithological Society, housed in the University of Michigan Museum of Zoology, was established in concurrence with the University of Michigan m 1930. Until 1947 the Library was maintained entirely by gifts and bequests of books, reprints, and ornithological magazines from members and friends of the Society. Two members have generously established a fund for the purchase of new books; members and friends are invited to maintain the fund by regular contribution. The fund will be administered by the Library Committee. Robert Payne, University of Michigan, is Chairman of the Committee. The Library currently receives over 200 periodicals as gifts and in exchange for The Wilson Journal of Orni- thology. For information on the Library and our holdings, see the Society’s web page at http;// www.wilsonsociety.org. With the usual exception of rare books, any item in the Library may be borrowed by members of the Society and will be sent prepaid (by the University of Michigan) to any address in the United States, its possessions, or Canada. Return postage is paid by the borrower. Inquiries and requests by borrowers, as well as gifts of books, pamphlets, reprints, and magazines, should be addressed to; Josselyn Van Tyne Memorial Library, Museum of Zoology, The University of Michigan, 1109 Geddes Avenue, Ann Arbor, MI 48109-1079, USA. Contributions to the New Book Fund should be sent to the Treasurer. NOTICE OF CHANGE OF ADDRESS MEMBERSHIP INQUIRIES okstate.edu THE JOSSELYN VAN TYNE MEMORIAL LIBRARY This issue of The Wilson Journal of Ornithology was published on 5 September 2008. Continued from outside back cover 582 Nest habitat selection of White-winged Scoters on Yukon Flats, Alaska David E. Safine and Mark S. Lindberg Short Communications 594 Use of legs and feet for control by scoters during aerial courtship William J. Wilson 599 Bill entanglement in subcutaneously- anchored radio transmitters on Harlequin Ducks Jeanine Bond and Daniel Esler 603 Estimate of Trichomonas gallinae-md\xcG:d mortality in Band-tailed Pigeons, upper Carmel Valley, California, winter 2006-2007 Mark R. Stromberg, Walter D. Koenig, Eric L. Walters, and John Schweisinger 606 Winter ecology of Yellow Rails based on South Carolina specimens William Post 610 Nest raising by Red-crowned Cranes in response to human-mediated flooding at Zhalong Nature Reserve, China (Jiang Wang and Eeng Li 613 White-winged Diuca Finch {Diuca speculifera) nesting on Quelccaya Ice Cap, Peru Douglas R. Hardy and Spencer R Hardy 618 Nest success and nest predation of the endangered Rota White-eye {Zosterops rotensis) Lainie Berry and Estanislao Taisacan 620 Predators at nests of the Western Slaty Antshrike ( Thamnophilus atrinucha) Corey E. Tarwater 625 Evidence for Bachman’s Sparrow raising Brown-headed Cowbirds to fledging Matthew J. Reetz, Elizabeth Earley, and Thomas A. Contreras 628 Simultaneous incubation by two females and nestling provisioning by four adults at a Savannah Sparrow nest Nathan J. Zalik and Noah G. Perlut 631 Grey Heron {Ardea cinerea) predation on the Aldabra White- throated Rail {Dryolimnas cuvieri aldabranus) Pierre A. Pistorius 633 Purple Swamphens {Porphyrio porphyrio) attempting to prey upon Black Swan {Cygnus atratus) eggs and preying upon a cygnet on an urban lake in Melbourne, Australia Shandiya Balasubramaniam and Patrick-Jean Guay 635 Northern Fulmar predation of Common Murre Stephan Lorenz and Sampath Seneveratne 637 Diet of nestling Black-crowned Night-herons in a mixed species colony: implications for Tern conservation C. Scott Hall and Stephen W. Kress 641 Diet of the Long-eared Owl in the northern and central Negev Desert, Israel Zohar Leader, Yoram Yom-Tov, and Uzi Motro 645 First observed instance of polygyny in Flammulated Owls Brian D. Linkhart, Erin M. Evers, Julie D. Megler, Eric C. Palm, Catherine M. Salipante, and Scott W Yanco 648 The Giant Hummingbird {Patagona gigas) in the mountains of central Argentina and a climatic envelope model for its distribution Henrik von Wehrden 651 Giant Hummingbirds {Patagona gigas) ingest calcium-rich minerals Cristidn E Estades, M. Angelica Vukasovic, and Jorge A. Tomasevic 654 Nocturnal foraging observations of the Blue-crowned Motmot {Momotus momota) in San Jose, Costa Rica Alejandro SoUno-Ugalde and Agustina Arcos-Torres 655 Kleptoparasitism by Grey Kingbirds {Tyrannus dominicensis) in Barbados Sarah E. Overington, Laure Cauchard, and Kimberly-Ann Cote 657 Double-scratching by Yellow-headed Blackbirds Alan de Queiroz 660 Ornithological Literature Compiled by Mary Gustafson and Clait E. Braun The Wilson Journal of Ornithology (formerly The Wilson Bulletin) Volume 120, Number 3 CONTENTS September 2008 Major Articles 433 Genetic structure of breeding and wintering populations of Swainsons Warbler Kevin Winker and Gary R. Graves 446 Does age influence territory size, habitat selection, and reproductive success of male Canada Warblers in central New Hampshire? Leonard R. Reitsma, Michael T Hallworth, and Phred M. Benham 455 Solitary winter roosting of Ovenbirds in core foraging area David R. Brown and Thomas W Sherry 460 Reproductive success of the Puerto Rican Vireo in a montane habitat Adrianne G. Tossas 467 Nest defense by Carolina Wrens Kelly A. D’Orazio and Diane L. H. Neudorf 473 Nests, eggs, and parental care of the Puna Tapaculo {Scytalopus simonsi) Peter A. Hosner and Noemt E. Huanca 478 Phylogeographic patterns of differentiation in the Acorn Woodpecker Magali Honey-Escandon, Blanca E. Herndndez-Banos, Adolfo G. Navarro-Siguenza, Hesiquio Benitez- Diaz, and A. Townsend Peterson 494 Diet of Acorn Woodpeckers at La Primavera Forest, Jalisco, Mexico Veronica Carolina Rosas-Espinoza, Elisa Maya-Elizarraras, Oscar Erancisco Reyna Bustos, and Erancisco Martin Huerta-Martinez 499 Seasonal variation in acoustic signals of Pileated Woodpeckers Sarah B. Tremain, Kyle A. Swiston, and Daniel J. Mennill 505 Common Poorwill activity and calling behavior in relation to moonlight and predation Christopher P Woods and R. Mark Brigham 513 Maximizing detection probability of wetland-dependent birds during point-count surveys in northwestern Florida Christopher P Nadeau, Courtney J. Conway, Bradley S. Smith, and Thomas E. Lewis 519 Male song variation of Green Violetear {Colibri thalassinus) in the Talamanca Mountain Range, Costa Rica Gilbert Barrantes, Cesar Sdnchez, Branko Hilje, and Rodolfo Jaffe 525 Metabolizable energy in Chinese Tallow fruit for Yellow-rumped Warblers, Northern Cardinals, and American Robins Michael J. Baldwin, Wylie C. Barrow Jr., Clinton Jeske, and Erank C. Rohwer 531 Foraging ecology of High Andean insectivorous birds in remnant Polylepis forest patches Huw Lloyd 545 Nesting biology of the Giant Conebill {Oreomanes fraseri) in the High Andes of Bolivia J. R. A. Cahill, E. Matthysen, and N. E. Huanca 550 Bird density and mortality at windows Stephen B. Hagar, Heidi Trudell, Kelly J. McKay, Stephanie M. Crandall, and Lance Mayer 565 Ectoparasites affect hemoglobin and percentages of immature erythrocytes but not hematocrit in nestling Eastern Bluebirds Renee E. Carleton 569 Observations on flocking behavior of ^Morthens Sparrow {^Spizella wortheni) and occurrence in mixed- species flocks Julio C. Canales-Delgadillo, Laura M. Scott-Morales, Mauricio Cotera Correa, and Marisela Pando Moreno 575 Status of Crested Penguin {Eudyptes spp.) populations on three islands in southern Chile David A. Oehler, Steve Pelikan, W Roger Pry, Leonard Weakley Jr., Alejandro Kusch, and Manuel Marin Continued on inside back cover Wilson Journal of Ornithology Volume 120, Number 4, December 2008 ubbab^ Published by the Wilson Ornithological Society THE WILSON ORNITHOLOGICAL SOCIETY FOUNDED 3 DECEMBER 1888 Named after ALEXANDER WILSON, the first American ornithologist. President James D. Rising, Department of Zoology, University of Toronto, Toronto, ON MSS 3G5, Canada; e-mail: rising@zoo.utoronto.ca First Vice-President — E. Dale Kennedy, Biology Department, Albion College, Albion, MI 49224, USA; e-mail: dkennedy@albion.edu Second Vice-President— Robert C. Beason, USDA, Wildlife Services, 6100 Columbus Avenue, Sandusky, OH 44870, USA; e-mail: beason@netzero.com Editor — Clait E. Braun, 5572 North Ventana Vista Road, Tucson, AZ 85750, USA; e-mail: TWILSONJO@ comcast.net Secretary John A. Smallwood, Department of Biology and Molecular Biology, Montclair State University, Montclair, NJ 07043, USA; e-mail: smallwoodj@montclair.edu Treasurer— MelindaM. Clark, 52684 Highland Drive, SouthBend, IN 46635, USA; e-mail: MClark@tcservices.biz Elected Council Members — Carla J. Dove, Greg H. Farley, Daniel Klemm Jr., and Mia R. Revels (terms expire 2009); Robert S. Mulvihill, and Timothy O’Connell (terms expire 2010); Jameson F. Chace, Sara R. Morris, and Margaret A. Voss (terms expire 2011). Membership dues per calendar year are: Active, $21.00; Student, $15.00; Family, $25.00, Sustaining, $30.00; Life memberships, $1,000.00 (payable in four installments). The Wilson Journal of Ornithology is sent to all members not in arrears for dues. the WILSON JOURNAL OF ORNITHOLOGY (formerly The Wilson Bulletin) THE WILSON JOURNAL OF ORNITHOLOGY (ISSN 1559-4491) is published quarterly in March, June, September, and December by the Wilson Ornithological Society, 810 East 10th Street, Lawrence, KS 66044-8897. The subscription price, both in the United States and elsewhere, is $40.00 per year. Periodicals postage paid at Lawrence, KS. POSTMASTER: Send address changes to OSNA, 5400 Bosque Boulevard, Suite 680, Waco, TX 76710. All articles and communications for publications should be addressed to the Editor. Exchanges should be addressed to The Josselyn Van Tyne Memorial Library, Museum of Zoology, Ann Arbor, MI 48109, USA. Subscriptions, changes of address, and claims for undelivered copies should be sent to OSNA, 5400 Bosque Boulevard, Suite 680, Waco, TX 76710, USA. Phone: (254) 399-9636; e-mail: business@osnabirds.org. Back issues or single copies are available for $12.00 each. Most back issues of the journal are available and may be ordered from OSNA. Special prices will be quoted for quantity orders. All issues of the journal published before 2000 are accessible on a free web site at the University of New Mexico library (http://elibrary. unm.edu/sora/). The site is fully searchable, and full-text reproductions of all papers (including illustrations) are available as either PDF or DjVu files. © Copyright 2008 by the Wilson Ornithological Society Printed by Allen Press, Inc., Lawrence, KS 66044, U.S.A. COVER: Wilson’s Plover (Charadrius wilsonia). Illustration by Don Radovich. 0 This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). MCZ LIBRARY MAY 19 2009 harvard UNIVERSITY Species Typical eggs Vesper Sparrow Savannah Sparrow Baird’s Sparrow Chestnut-collared Longspur Western Meadowlark Brown-headed Cowbird FRONTISPIECE. Examples of eggs of Brown-headed Cowbird {Molothrus ater) and six common grassland passerines in south-central Saskatchewan. Top row to second row from the bottom are host species: Sprague’s Pipit (Anthus spragueii). Vesper Sparrow (Pooecetes gramineus). Savannah Sparrow (Passerculus sandwichen- sis), Baird’s Sparrow (Ammodramus bairdii). Chestnut-collared Longspur (Calcarius ornatus), and Western Meadowlark (Sturnella neglecta). Bottom row: parasitic Brown-headed Cowbird. Experimentally parasitizing nests of these species with cowbird-like eggs and non-mimetic (blue) eggs, Klippenstine and Sealy (pages 667- 673) recorded more blue eggs ejected than cowbird eggs by each species, which suggests similarity in appearance of cowbird and host eggs impeded discrimination and may represent a form of cowbird egg mimicry. Wilson Journal of Ornithology Published by the Wilson Ornithological Society VOL. 120, NO. 4 December 2008 PAGES 667-962 The Wilson Journal of Ornithology 120(4):667-673, 2008 DIFFERENTIAL EJECTION OF COWBIRD EGGS AND NON-MIMETIC EGGS BY GRASSLAND PASSERINES DWIGHT R. KLIPPENSTINEi 2 AND SPENCER G. SEALY' ^ ABSTRACT. — Grassland passerines are purported to tolerate parasitism by Brown-headed Cowbirds {Mol- othrus ater) because of adaptations by cowbirds that constrain egg discrimination and removal by their hosts, i.e., evolutionary equilibrium, rather than because of an absence of these defensive behaviors, i.e., evolutionary lag. We tested these hypotheses by experimentally parasitizing six grassland species with cowbird-like eggs and non-mimetic (blue) eggs in south-central Saskatchewan. Sprague’s Pipits (Anthus spragiieii). Vesper (Pooecetes gramineus). Savannah (Passerculus sandwichensis), and Baird’s {Ammodramus bairdii) sparrows, and Chestnut- collared Longspurs (Calcarius ornatus) accepted all or nearly all cowbird eggs, but ejected or attempted to eject between 9 and 20% of blue eggs with 54% of rejected eggs not removed from the nests, i.e., failed ejection attempts. Western Meadowlarks (Sturnella neglecta) ejected an intermediate number of cowbird eggs (67%) and most blue eggs (92%), none of which was recovered. Ejection of non-mimetic eggs versus cowbird eggs by each species suggests that similarity in appearance of cowbird eggs and hosts’ eggs impeded discrimination and may represent a form of cowbird egg mimicry. The low number of blue eggs ejected by five of the six species, their failure to remove these eggs from the nest sites, and the damage wrought on some hosts’ eggs during ejection suggest the morphology of cowbird eggs also constrains ejection behavior. These results support the evolutionary equilibrium hypothesis as the better explanation for acceptance of cowbird parasitism observed in these grassland passerines. Received 4 May 2007. Accepted 15 February 2008. Hosts of the parasitic Brown-headed Cow- bird {Molothrus ater) generally accept or re- ject cowbird eggs, depending on their ability to discriminate and eject foreign eggs from their nests (Rothstein 1975a, Peer and Sealy 2004). Birds learn the appearance of their own eggs as they are laid, which permits detection ' Department of Biological Sciences, University of Manitoba, Winnipeg, MB R3T 2N2, Canada. ^Current address: 896 Garfield Street North, Win- nipeg, MB R3G 2M7, Canada. ^ Corresponding author; e-mail: sgsealy@cc.umanitoba.ca of foreign eggs that differ visually from the hosts’ eggs (reviewed in Kilner 2006). Ejec- tion is the removal of a targeted egg by grasp- ing its girth, i.e., grasp-ejection, or puncturing the shell and lifting the egg out of the nest or removing it piecemeal, i.e., puncture-ejection (Underwood and Sealy 2006). Cowbird para- sitism often reduces the reproductive success of hosts (Lorenzana and Sealy 1999); there- fore, removal of cowbird eggs should benefit species that are parasitized frequently (Roth- stein 1990, Underwood and Sealy 2002). Hosts of the Brown-headed Cowbird (here- after “cowbird") that inhabit the grasslands of 667 668 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 the Great Plains have been presumed to pos- sess egg discrimination and ejection behavior because they have nested in sympatry with cowbirds for thousands of years (Friedmann 1929, Mayfield 1965). Moreover, cowbird par- asitism significantly reduces breeding success in parasitized grassland hosts (Hill 1976, El- liott 1978, Davis and Sealy 2000, Davis 2003). However, of the species regularly listed as grassland passerines in North America (Mengel 1970, Knopf 1994), only Eastern {Sturnella magna) and Western {S. neglecta) meadowlarks are known to regularly eject cowbird eggs (Peer et al. 2000). Instead, ac- ceptance of experimentally added cowbird eggs by several grassland passerines (Hill and Sealy 1994, Sealy 1999, Peer et al. 2000, Da- vis et al. 2002) and observations of cowbirds fledging from naturally parasitized nests (El- liott 1978, Davis and Sealy 2000, Davis 2003) suggest that most grassland species tolerate parasitism. Grassland passerines, according to the evo- lutionary lag hypothesis, may accept parasitic eggs because of insufficient selective pressure to evolve egg discrimination and ejection (Rothstein 1975a, 1990; Davies and Brooke 1989). Several studies have demonstrated that grassland passerines are infrequently parasit- ized compared with hosts that occupy adjacent habitats, although there is variability (Smith et al. 2000). Moreover, the selective pressure provided by parasitism is presumably higher today than in the past (Davis and Sealy 2000). This reflects recent increases in cowbirds and greater accessibility of grassland hosts that re- sult from habitat fragmentation and creation of suitable perches (Davis and Sealy 2000). Evolutionary lag may also arise if the muta- tions governing egg discrimination and ejec- tion have not developed or spread throughout the host’s population (Rothstein 1990). The evolutionary equilibrium hypothesis posits that even if egg ejection behavior has evolved, grassland passerines may accept cowbird eggs because ejection is too costly (Rohwer and Spaw 1988, Lotem and Naka- mura 1998). Ejection may be costly when par- asites have counter-evolved adaptations that cause hosts to make recognition errors, i.e., eject the wrong egg, or ejection errors, i.e., damage their own eggs during the ejection process (Davies et al. 1996, Lotem and Nak- amura 1998). Several authors have suggested that a generalized form of “mimicry” exists between color and maculation of cowbird eggs and those of grassland passerines, which may hinder egg discrimination and lead to rec- ognition errors (Peer et al. 2000, Davis et al. 2002, Klippenstine 2005). Cowbird eggs are also larger, rounder, and have thicker shells, which may impede egg ejection by intolerant hosts and result in ejection errors (Rohwer and Spaw 1988, Pieman 1989). Evolutionary equi- librium is supported by experimental evidence showing that several grassland passerines are more likely to eject foreign eggs if they are non-mimetic and undersized (Peer et al. 2000, Davis et al. 2002). Our goal was to examine which of these hypotheses better explains the apparent accep- tance of cowbird parasitism by grassland hosts. We experimentally parasitized six grassland species (Sprague’s Pipits [Anthus spragueii]. Vesper Sparrows [Pooecetes gra- mineus]. Savannah Sparrows [Passerculus sandwichensis], Baird’s Sparrows [Ammodra- mus bairdii]. Chestnut-collared Longspurs [Calcarius ornatus], and Western Meadow- larks) with real and model cowbird eggs and non-mimetic eggs to assess their (1) tolerance of parasitism, (2) ability to discriminate for- eign eggs, and (3) ability to remove foreign eggs from their nests. If these passerines ac- cept cowbird eggs because of evolutionary lag, we hypothesized that hosts would accept both cowbird and non-mimetic eggs (Table 1). If at equilibrium with cowbird parasitism, we hypothesized that differences would be re- corded in frequency and success of ejection TABLE 1. Possible outcomes of experimental parasitism on grassland passerines. Problems with . . . Egg type Discrimination only Removal only Discrimination and removal Neither Cowbird Non-mimetic Accepted Ejected Attempted ejection Attempted ejection Accepted Attempted ejection Ejected Ejected Klippenstine and Sealy • GRASSLAND PASSERINE ANTI-PARASITIC BEHAVIOR 669 with different egg treatments. If mimicry has evolved, non-mimetic eggs should be ejected more often than cowbird eggs, whereas diffi- culties in ejecting “cowbird-sized” eggs should result in more failed ejection attempts than successful ejections (Table 1). METHODS Fieldwork was conducted in south-central Saskatchewan, Canada, from 7 May to 1 July 2001 and 9 May to 15 July 2002. Sites were 60 km south of Regina and widely dispersed within a lOO-km^ area centered about the abandoned town of Dummer (49° 50' N, 104° 49' W). The 18 sites consisted of 7 privately owned pastures of mixed-grass prairie and 6 alfalfa (Medicago sativa), 1 fallow, 1 stubble, and 3 crested wheatgrass {Agropyron crista- tum) fields that ranged from 0.5 to over 500 ha. Most nests were found with a “dragging rope” pulled over the ground to flush incu- bating adults (Davis et al. 2002), although some nests were found while inspecting other nests. Each nest was experimentally parasitized with one of four egg treatments; real cowbird eggs, model cowbird eggs, real immaculate blue eggs, or model immaculate blue eggs (Fig. 1). Cowbird and blue eggs were used to assess host discrimination of foreign eggs be- cause cowbird egg treatments were considered mimetic in appearance to the eggs of all six grassland passerine species studied (Baicich and Harrison 1997); the immaculate blue egg treatments were considered non-mimetic (Fig. 1, Frontispiece). We note that despite recent evidence for birds ‘seeing’ objects in the UV spectrum, no evidence exists to suggest that acceptance or rejection is related to UV re- flectance (Underwood and Sealy 2008). We used model eggs as well as real eggs because no significant bias has been attributed to use of model eggs (Rothstein 1975a, Hill and Sea- ly 1994, Peer et al. 2000), and because they provided a means to assess the method of ejection used by grassland hosts. Preferential ejection of real eggs would signify puncture- ejection, as models cannot be pierced (Davis et al. 2002, Underwood and Sealy 2006). Conversely, the ejection of both model and real eggs would imply grasp ejection (Davis et al. 2002, Underwood and Sealy 2006). Real cowbird eggs were collected from nat- Cowbird egg treatments Blue egg treatments PIG. 1. Pour egg treatments used to parasitize nests. urally parasitized nests of Red-winged Black- birds (Agelaius phoeniceus) and Brewer’s Blackbirds {Euphagus cyanocephalus) and stored at —5° C for up to 5 days until they were used to test nests. Model eggs were made of plaster-of-Paris following Rothstein’s (1970) protocol and were (x ± SE) 20.84 ± 0.04 mm X 16.97 ± 0.03 mm (n = 101), which approximated the dimensions of real cowbird eggs (21.45 X 16.42 mm, n — 127; Fowther 1993); however, they averaged slightly heavier (3.40 ± 0.03 g, n = 101 vs. 3.12 g, /2 = 181; Fowther 1993). Folkart® and Fiquitex® acrylic paints were used to create the colors of the eggs. One part wicker white (#901 FA), one-quarter part medium gray (#425 FA), and one-quarter part burnt umber (#462 FA) were used in combination to sim- ulate the off-white to slightly brown ground color of cowbird eggs. The reddish-brown maculation consisted of one part burnt umber and one part wicker white, whereas the gray maculation was composed of one part medium gray and one part wicker white; both were then spattered on the surface of the egg with a toothbrush to resemble a typical cowbird egg. The solid blue ground color was one part cerulean blue (#470 L) and one part light blue violet (#680 L). Painted eggs were water- proofed with a coat of lacquer, which also gave them a realistic shine. Clutches during laying or in early incuba- tion (days 0 through 7-8) received one of the four egg treatments. It should be noted that, because real cowbird eggs were not available throughout the entire experiment, more model 670 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 egg treatments were used than real egg treat- ments; but as much as possible, treatments were assigned randomly. Each nesting pair was tested only once and we assumed that un- marked females at nests tested in 2001 were not tested again in 2002. Experimental eggs were left in nests for 5 days and inspected on days 1, 2, 4, and 5. Eggs were considered ac- cepted if they remained undamaged and tend- ed in nests for 5 days, and rejected if they disappeared from an otherwise active nest (successful ejection) or were damaged, i.e., chipped or scratched, but remained in the ac- tive nest (attempted ejection). A 5-day test pe- riod was considered sufficient, as most rejec- tions occur within 2 days and rarely after 5 days (Rothstein 1975b, Underwood and Sealy 2006). We removed accepted eggs following the 5-day test period and all nests, regardless of experimental outcome, were inspected until the host’s eggs hatched or the nest failed. Ejected eggs found within 5 m of the nest rim were recorded, as were host eggs damaged or missing following ejection. Models were cleaned after each experiment with a dilute alcohol solution and re-used; real cowbird eggs were used only once. All natu- rally parasitized nests also received experi- mental eggs, as ignoring nests already para- sitized may exclude individuals prone to ac- ceptance and bias the sample toward rejection. No host eggs were removed from nests tested. Statistical Analyses. — The Log-Likelihood Ratio was used to compare the frequency of ejection between egg types within a species. This test was selected over the Eisher Exact Test because it performs well with expected values less than five (Zar 1996). All tests used alpha = 0.05. Sprague’s Pipits received only model cowbird eggs and model blue eggs be- cause the expected number of nests was small. The experimental egg at 16 nests was re- moved after 4 days due to errors in estimating the hatching date; however, these did not af- fect the overall results and were included in the final analyses. Nests that were deserted or depredated during the 5 -day test period were excluded from analysis. RESULTS Natural cowbird parasitism was recorded for all six species and ranged from 5 to 28% with a pooled frequency of 15% for the entire community for both years {n = 306 nests): Sprague’s Pipit (16%, n = 19 nests). Vesper Sparrow (20%, n = 40), Savannah Sparrow (28%, n = 47), Baird’s Sparrow (16%, n = 70), Chestnut-collared Longspur (5%, n = 96), and Western Meadowlark (15%, n = 34). Of the 266 nests tested experimentally, 20% {n = 54) failed before a response could be measured, 4% (/z = 10) were excluded be- cause their fates could not be ascertained, and of the 76% (zz = 202) of nests that were suc- cessful, 16% {n = 33) rejected and 84% (zz = 169) accepted the experimental egg. Real versus Model Eggs. — No statistical tests were performed to investigate differences between model and real versions of either cowbird or blue eggs because so few rejec- tions were recorded in most treatments (Table 2). More real blue eggs than model blue eggs appeared to be rejected (attempted and suc- cessful combined) by most species (Table 2). We presume this difference resulted from the inability of hosts to make discernible marks TABLE 2. Frequency of rejection^ (%) of experimentally added model (^significant difference at alpha = 0.05). and real cowbird and blue eggs % Frequency of rejection (n) P value (cowbird vs. blue) Cowbird egg treatments Blue egg treatments Species Model Real Combined Model Real Combined Sprague’s Pipit 0.0 (8) 0.0 (8) 16.7 (6) 16.7 (6) 0.180 Vesper Sparrow 0.0 (11) 0.0 (4) 0.0 (15) 10.0 (1) 40.0 (5) 20.0 (15) 0.034* Savannah Sparrow 11.1 (9)*’ 0.0 (5) 7.1 (14) 0.0 (8) 28.6 (7) 13.3 (15) 0.581 Baird’s Sparrow 0.0 (20) 0.0 (6) 0.0 (26) 14.3 (14) 28.6 (7) 19.0 (21) 0.009* Chestnut-collared Longspur 0.0 (17) 0.0 (9) 0.0 (26) 0.0 (21) 27.3 (11) 9.4 (32) 0.054 Western Meadowlark 70.0 (10) 50.0 (2) 66.7 (12) 90.9 (11) 100.0 (1) 91.7 (12) 0.121 “ Includes both successful and failed ejection attempts. ^ Ejection was suspected given the timing of the event. Klippenstine and Sealy • GRASSLAND PASSERINE ANTI-PARASITIC BEHAVIOR 671 on the models during failed ejection attempts rather than due to the method of ejection. The actual frequency of rejection for each species, therefore, may have been higher than reported (Table 2). Cowbird versus Blue Eggs. — Sprague’s Pip- its, Vesper and Baird’s sparrows, and Chest- nut-collared Longspurs accepted all cowbird eggs (Table 2). Savannah Sparrows accepted all but one cowbird egg (Table 2), which dis- appeared between 4 and 5 days after “para- sitism.” This nest was later depredated and, given the timing of the event, we ascertained that a predator most likely removed the egg. Western Meadowlarks accepted 33% (n = 12) of the cowbird eggs and rejected the other eight within 48 hrs (Table 2). Sprague’s Pipits, Vesper, Savannah, and Baird’s sparrows, and Chestnut-collared Longspurs rejected a small proportion of blue eggs (Table 2). Sprague’s Pipits, Savannah Sparrows, and Chestnut-collared Longspurs rejected all blue eggs within 48 hrs, whereas Vesper Sparrows rejected one of three and Baird’s Sparrows rejected one of four blue eggs between 2 and 5 days, and 2 and 3 days, respectively. Western Meadowlarks rejected 92% {n = 12) of the blue eggs within 48 hrs (Table 2). Successful versus Failed Ejection At- tempts.— Seven (54%) of the 13 rejected blue eggs reported for Sprague’s Pipits, Vesper, Sa- vannah, and Baird’s sparrows, and Chestnut- collared Longspurs were classified as failed ejection attempts (Table 3). Of the six remain- ing ejections (46%) in which the blue egg was completely removed from the nest, two eggs were recovered within 5 m of the nest rims. Western Meadowlarks successfully ejected all egg treatments and none was recovered within 5 m of the nest rims. Egg Loss. — One Savannah Sparrow egg disappeared after a blue egg had been ejected (0.50 eggs lost/blue egg ejected) and three Western Meadowlark eggs disappeared during two separate ejections of blue eggs (0.27 eggs lost/blue egg ejected); no other egg loss was noted among the four remaining species. DISCUSSION Our findings suggest the acceptance or par- tial acceptance of cowbird eggs by Sprague’s Pipits, Vesper, Savannah, and Baird’s spar- rows, Chestnut-collared Longspurs, and West- ern Meadowlarks is best explained by the evo- lutionary equilibrium hypothesis. The ejection of foreign eggs reveals that all six species have evolved egg recognition and egg ejection in response to cowbird parasitism. Egg rec- ognition is implied because birds do not sim- ply remove discordant eggs (Rothstein 1975b, 1978) and the capacity to discriminate foreign eggs has little or no use outside the context of interspecific brood parasitism (Underwood and Sealy 2002, Kilner 2006). The preferen- tial ejection of real blue eggs over model blue eggs by five of the six species suggests punc- ture ejection; however, only one recovered real blue egg showed signs of puncture, and this occurred after attempts over several days by the Savannah Sparrow pair to remove the egg. In contrast, ejection of solid plaster eggs by Western Meadowlarks, Sprague’s Pipits, and Vesper Sparrows, and the recovery of an intact real blue egg ejected from a Chestnut- collared Longspur nest suggest grasp ejection. These findings do not support the evolution- ary lag hypothesis, which states that hosts ac- TABLE 3. Successful and failed ejections for both model and real blue eggs. Species in) Number successful ejections Number failed ejections Model eggs Real eggs Model eggs Real eggs Sprague’s Pipit ( 1 ) L 0 Vesper Sparrow (3) 1 0 0 2 Savannah Sparrow (2) 0 1 0 1 Baird’s Sparrow (4) 0 1 2 1 Chestnut-collared Longspur (3) 0 2*’ 0 1 Western Meadowlark (12) 10 1 0 0 “ Egg was recovered 15 cm from the nest rim. ^ One intact egg was found 20 cm from the nest rim; the other was nt>t recovered. 672 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 cept parasitic eggs because they have not yet evolved egg discrimination and ejection (Rothstein 1975a. 1990; Davies and Brooke 1989). More importantly we observed that the ability of all six grassland passerines to rec- ognize and eject a foreign egg depended on the egg's parameters. Such plasticity is the hallmark of the evolutionary equilibrium hy- pothesis. which states that hosts accept or re- ject parasitic eggs based on the probability of making recognition and ejection errors (Roh- wer and Spaw 1988. Lotem and Nakamura 1998). The differential ejection of blue eggs over cowbird eggs by all six species, for ex- ample. implies that hosts were hesitant to eject the latter, presumably because they are mi- metic and are more likely to induce a recog- nition error. The high number of failed ejec- tion attempts by five of the six species and the egg loss observed in Western Meadowlark and Savannah Sparrow nests following ejection of blue eggs implies that size and shape of cow- bird eggs has evolved to impede grasp ejec- tion by these hosts; although the influence of the thicker cowbird egg shell in preventing puncture ejection should be investigated fur- ther. The evolutionary lag h\pothesis only par- tially explains the pattern of ejection we ob- served. In particular, a mixture of acceptance and rejection could be observed in a popula- tion if egg discrimination and rejection have only recently evolved, and have begun to spread throughout the host's population (Peer et al. 2000). Western Meadowlarks examined in this study, for example, may consist of two- thirds rejecters and one-third accepters. The failed ejection attempts may reflect the recent development of a new skill. These ideas seem unlikely because we observed acceptance of naturally parasitized cowbird eggs by hosts that ejected an experimentally added cowbird egg or blue egg. which demonstrates that one host can be both tolerant and intolerant of for- eign eggs. The spread of rejection behavior does not explain the differential ejection of non-mimetic eggs over cowbird eggs ob- served in this study or the preferential ejection of undersized eggs reported by Peer et al. (2000); presumably a truly mixed population of rejecters and accepters would reject all for- eign eggs in roughly the same proportion. Our results suggest that grassland passerines tol- erate parasitism because cowbirds have coun- ter-evolved specific adaptations to the appear- ance and morphology of their eggs in response to the evolution of egg discrimination and ejection by host species. ACKNOWLEDGMENTS We are indebted to the landow ners who permitted us to use their land. We thank M. .A.. Talbot and N. L. Haalboom for their hard work and Cornelius Klippen- stine for use of his farmhouse. L. C. Graham. N. C. Kenkel. and M. D. Kujath commented on early drafts of the manuscript and offered their support throughout this project. D. E. Burhans and two anonymous re- viewers provided constructive comments that im- proved the manuscript. This study was funded by a Discovery Grant from the Natural Sciences and En- gineering Research Council of Canada to S. G. Sealy and through in-kind services provided by S. K. Davis and the Saskatchewan Watershed Authority (formerly Saskatchewan Wetlands Conservation Corporation). LITERATURE CITED Baicich. P. J. .vnd C. J. O. Rvrrison. 1997. Nests, eggs, and nestlings of North American birds. Sec- ond Edition. Academic Press. New York. USA. Dantes. N. B. .and M. de L. Brooke. 1989. An ex- perimental study of co-evolution between the Cuckoo. Cuciiliis canonis. and its hosts. I. Host egg discrimination. Journal of Animal Ecology 58^207-224. Davies. N. B.. M. de L. Brooke, .and A. Kvcelntk. 1996. Recognition errors and probability of para- sitism determine whether Reed Warblers should accept or reject mimetic cuckoo eggs. Proceedings of the Royal Society of London. Series B 263; 925-931. ' D.avis. S. K. 2003. Nesting ecology of mixed-grass prairie songbirds in southern Saskatchewan. Wil- son Bulletin 115:119-130. D.ants. S. K. .AST) S. G. SE.ALY. 2CKX). Cowbird para- sitism and nest predation in fragmented grasslands of southwestern Manitoba. Pages 220-228 in Ecology and management of cowbirds and their hosts: studies in the conservation of North Amer- ican passerine birds (J. N. M. Smith. T. L. Cook. S. I. Rothstein. S. K. Robinson, and S. G. Sealy. Editors). University of Texas Press. Austin. LSA. D.wts. S. K.. D. R. Klippenstine. ant) R. M. Brigram. 2(X)2. Does egg rejection account for the low in- cidence of cowbird parasitism in Chestnut-col- lared Longspurs (Calcarius omanisYl Auk 119: 556-560. Elliott. P. F. 1978. Cowbird parasitism in the Kansas tallgrass prairie. .Auk 95:161-167. Friedm.\nn. H. 1929. The cowbirds; a study in the bi- ology of social parasitism. C. C. Thomas. Spring- field! Illinois. US.A. Klippenstine and Sealy • GRASSLAND PASSERINE ANTI-PARASITIC BEHAVIOR 673 Hill, D. P. and S. G. Sealy. 1994. Desertion of nests parasitized by cowbirds; have Clay-colored Spar- rows evolved an anti-parasite defence? Animal Behaviour 48:1063-1070. Hill, R. A. 1976. Host-parasite relationships of the Brown-headed Cowbird in a prairie habitat of west-central Kansas. Wilson Bulletin 88:555-565. Kilner, R. M. 2006. The evolution of egg colour and patterning in birds. Biological Reviews 81:383- 406. Klippenstine, D. R. 2005. Can egg mimicry by Brown-headed Cowbirds explain the acceptance of brood parasitism by grassland passerines? The- sis. University of Manitoba, Winnipeg, Canada. Knopf, E L. 1994. Avian assemblages on altered grass- lands. Studies in Avian Biology 15:247-257. Lorenzana, J. C. and S. G. Sealy. 1999. A meta- analysis of the impact of parasitism by the Brown- headed Cowbird on its hosts. Studies in Avian Bi- ology 18:241-253. Lotem, a. and H. Nakamura. 1998. Evolutionary equilibria in avian brood parasitism: an alternative to the “arms race-evolutionary lag” concept. Pag- es 223-235 in Parasitic birds and their hosts: stud- ies in coevolution (S. I. Rothstein and S. K. Rob- inson, Editors). Oxford University Press, Oxford, United Kingdom. Lowther, P. E. 1993. Brown-headed Cowbird {Mol- othriis ater). The birds of North America. Number 47. Mayfield, H. 1965. The Brown-headed Cowbird, with old and new hosts. Living Bird 4:13-28. Mengel, R. M. 1970. The North American Central Plains as an isolating agent in bird speciation. Pages 280-340 in Pleistocene and recent environ- ments of the Central Great Plains (W. Dort and J. K. Jones, Editors). University of Kansas Press, Lawrence, USA. Peer, B. D. and S. G. Sealy. 2004. Correlates of egg rejection in hosts of the Brown-headed Cowbird. Condor 106:580-599. Peer, B. D., S. K. Robinson, and J. R. Herkert. 2000. Egg rejection by cowbird hosts in grasslands. Auk 117:892-901. PiCMAN, J. 1989. Mechanism of increased puncture re- sistance of eggs of Brown-headed Cowbirds. Auk 106:577-583. Rohwer, S. and C. D. Spaw. 1988. Evolutionary lag versus bill-size constraints: a comparative study of the acceptance of cowbird egg by old hosts. Evo- lutionary Ecology 2:27-36. Rothstein, S. I. 1970. An experimental investigation of the defenses of the hosts of the parasitic Brown-headed Cowbird {Molothms ater). Disser- tation. Yale University, New Haven, Connecticut, USA. Rothstein, S. I. 1975a. An experimental and teleo- nomic investigation of avian brood parasitism. Condor 77:250-271. Rothstein, S. I. 1975b. Mechanisms of avian egg-rec- ognition: do birds know their own eggs? Animal Behaviour 23:268-278. Rothstein, S. I. 1978. Mechanisms of avian egg rec- ognition: additional evidence for learned compo- nents. Animal Behaviour 26:671-677. Rothstein, S. I. 1990. A model system for coevolu- tion: avian brood parasitism. Annual Review of Ecology and Systematics 21:481-508. Sealy, S. G. 1999. Cowbird parasitism on Lark Bun- tings: frequency, acceptance, and fledging. Journal of Eield Ornithology 70:182-186. Smith, J. N. M., T. L. Cook, S. I. Rothstein, S. K. Robinson, and S. G. Sealy. (Editors). 2000. Ecol- ogy and management of cowbirds and their hosts: studies in the conservation of North American passerine birds. University of Texas Press, Austin, USA. Underwood, T. J. and S. G. Sealy. 2002. Adaptive significance of egg coloration. Pages 280-298 in Avian incubation: behaviour, environment, and evolution (D. C. Deeming, Editor). Oxford Uni- versity Press, Oxford, United Kingdom. Underwood, T. J. and S. G. Sealy. 2006. Grasp-ejec- tion in two small ejecters of cowbird eggs: a test of bill-size constraints and the evolutionary equi- librium hypothesis. Animal Behaviour 71:409- 416. Underwood, T. J. and S. G. Sealy. 2008. UV reflec- tance of eggs of Brown-headed Cowbirds (Mol- othnis ater) and accepter and rejecter hosts. Jour- nal of Ornithology 149:313-321. Zar, j. H. 1996. Biostatistical analysis. Third Edition. Prentice Hall, Upper Saddle River, New Jersey, USA. The Wilson Journal of Ornithology 120(4):674-682, 2008 BETWEEN AND WITHIN CLUTCH VARIATION OF EGG SIZE IN GREATER RHEAS GUSTAVO J. FERNANDEZ' - AND JUAN C. REBOREDA' ABSTRACT. — We describe egg characteristics, and analyze between and within clutch variation in egg size and mass in a natural population of Greater Rheas (Rhea americana). We assess the effect of this variation on nesting success and egg success. Yolk represented 29.5% of egg mass whereas albumen was 63.9%. Yolk mass increased with egg width but not with egg length, while mass of albumen increased mainly with egg length. The largest and smallest eggs were 10.3% larger and 25.3% smaller, respectively than mean intra-clutch values. The widest egg was 11.9% wider while the narrowest egg was 20.5% narrower than mean intra-clutch values. There was a significant decrease in egg size between clutches during the breeding season as a result of a decrease in egg length. There was no effect of laying order on intra-clutch variation in egg size, but we detected an increase in the variation of egg length within clutches with clutch size. We did not detect a relationship between egg size and nesting success, and between egg size and egg success. The relatively low intra-clutch variation in egg size and lack of effect of egg size on hatching success do not support the hypothesis that females invest in eggs according to expected chick fitness. Received 23 November 2007. Accepted 15 April 2008. Intraspecific egg size variation is relatively common in birds (Slagsvold et al. 1984, Wil- liams et al. 1993, Christians 2002). Most var- iation is among rather than within clutches and it is generally accepted this variation is the result of a strong genetic component of egg size (Christians 2002, Stryrsky et al. 2002, Valkama et al. 2002). However, varia- tion among or within clutches laid by the same female should be attributed to phenotyp- ic or environmental-mediated variation (Nager et al. 2006). Nutrient or energetic constraints could be responsible for egg size variation in some bird species (O’Connor 1979, Pierotti and Bellrose 1986). Egg size could also vary with parental age or experience with eggs laid by older and experienced females larger than those laid by younger females (Davis 1975, Blomqvist et al. 1997, Hipfner et al. 1997). Alternatively, egg size variation could respond to an adaptive investment of females, thereby influencing offspring quality. Larger eggs may contain more water and nutrients, which may benefit hatchlings by increasing their hatching success, growth rate, and survival (Parsons 1970, Reid and Boersma 1990, Magrath 1992, Williams 1994, Smith and Bruun 1998, Pelayo and Clark 2003). Egg size in some species ' Departamento de Ecologfa, Genetica y Evolucion, Eacultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, Pabellon 2, Ciudad Universitaria, Buenos Aires. Argentina. 2 Corresponding author; e-mail: gjf@ege.fcen.uba.ar decreases with laying order (e.g., Nisbet 1978, Shaw 1985, Custer and Erederick 1990). This decrease could be interpreted as an adaptive strategy because the reduction in size of last laid eggs could be a mechanism for shortening incubation period and favoring synchronous hatching (e.g., Birkhead and Nettleship 1982, Perrins 1996, Kennamer et al. 1997). Maternal investment in eggs also could be related to mate quality or prospects of success. Some studies have found that females mated with non-preferred males produced fewer and smaller eggs than females mated with the males they chose (Yamamoto et al. 1989, Cunningham and Russell 2000). Most studies of egg size variation were con- ducted on species where the clutch is laid by a unique female. In species with communal egg laying (where more than one female laid eggs into a single nest), differences in egg in- vestment by females could have more impor- tant consequences for offspring fitness. Eor example, larger eggs in Guira Cuckoo (Guira guira) produce heavier and larger chicks that may out compete smaller and lighter nest companions (Macedo et al. 2004). As a con- sequence, larger communal clutches in this species had eggs with more nutrients than in smaller clutches. Females of species with communal laying could manipulate the nutri- ent investment in their eggs depending upon social condition to enhance individual fitness (Macedo et al. 2004). We analyzed inter- and intra-clutch varia- 674 Fernandez and Reboreda • EGG SIZE VARIATION IN GREATER RHEAS 675 tion of egg size in a natural population of Greater Rheas {Rhea americana). Greater Rheas are communal nesters where two to 10 or more females lay eggs into a single nest that a male incubates. We analyzed the effect of season and clutch sizes (total number of eggs in the nest) on both inter- and intra- clutch variation of egg size to assess if egg size was associated with hatching success. The relationship between egg size and nesting success was analyzed because it is possible females reduce their investment in eggs laid for low quality males (i.e., young inexperi- enced or in poor physical condition) that are prone to desert the nest. We also analyzed the relationship between egg components (yolk and albumen) and egg size. We expected a positive relationship between egg size and yolk content, and between egg size and nest success, and hatching success. METHODS Study Site. — Data were collected during the 1992 and 1993 breeding seasons of Greater Rheas (Fernandez and Reboreda 1998, 2003). We searched for Greater Rhea nests in three contiguous cattle ranches of —3,500, 3,000, and 800 ha, near the town of General Lavalle in Buenos Aires Province, Argentina (36° 25' S, 56°56'W). Numbers of rheas on these ranches were —250, 100, and 150, respective- ly. The area is flat, low, and marshy with little of the land rising more than 10 m above sea level. Native vegetation is short grass with scattered patches of woodland in the higher areas (Fernandez and Reboreda 1998). Data Collection. — Greater Rhea nesting oc- curred from September to December, but most attempts were in November (Fernandez and Reboreda 1998). We found 99 nests, 41 in 1992, and 58 in 1993. We used data from a sample of 53 active nests where we measured all eggs (30 nests in 1992, 23 in 1993). We individually numbered the eggs in each nest with water proof ink and recorded clutch size (number of eggs in the nest). Nests were vis- ited (<2() min/visit) between ()9()0-l 700 hrs every 2-3 days until the eggs hatched or the nest failed. We estimated the date at which laying started either directly (we knew the date of laying of the first egg) or indirectly by the color of the eggs (light yellow when laid but white in —5 days) or by backdating (start of laying was estimated as date of hatching minus 40 days). We measured length and width of the eggs with calipers (±0.1 mm) and mass with a 1-kg Pesola spring scale (±5 g). Egg mass varies during incubation (Grant et al. 1982) and we corrected it using the relationship be- tween fresh egg weight and egg volume (length X width^; Hoyt 1979) estimated from a subsample of 24 eggs found before onset of incubation. This relationship was highly sig- nificant (Fi22 = 955.2, P < 0.0001) and the model equation was: fresh mass (g) = 11.8 + 0.55 X (length X width^) {R^ = 0.98). Egg volume was estimated using the water displacement for 39 fresh eggs collected from early deserted nests. We measured the volume displaced when the egg was immersed in wa- ter after its air cell was filled with water. The relationship between the product of F X W^ (length X width^) and the measured volume for these eggs was highly significant (simple regression analysis, Fj 37 = 101.6, P < 0.001). The equation was: volume (cm^) = -47.18 + 0.56 X (F X W2) (P2 = 0.73). We collected fresh eggs from nests deserted before the onset of incubation to study egg composition. We took one egg from four nests and two eggs from another four nests (total = 12 eggs) to minimize pseudoreplication. Eggs were measured, weighed, and boiled for —30 min to solidify the content (Carey 1996). We manually separated shell plus membranes, al- bumen, and yolk and recorded their wet mass with a Pesola spring scale (±5 g). Statistical Analyses. — We estimated the var- iance component of differences in egg size (width and length) within and between clutch- es. We performed a one-way ANOVA. with egg measurements as dependent variables and nests as a random factor. We assessed the relationship between lay- ing sequence and egg size using simple re- gression analyses with transformed data. We standardized egg length and width dividing 676 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 each by the mean clutch value to correct for differences among clutches (Kennamer et al. 1997). We only included eggs laid up to 12 days after the start of laying. We estimated intra-clutch variation in egg size by calculating the variation coefficients (CV) for each nest. We assessed variation of CVs with clutch size and season using general linear models (GLMs) with egg length and width as response variables, and season, clutch, and its interaction, as explanatory var- iables. We assigned nests to the date at which laying started and divided the breeding season into periods: 1—15 October; 16—31 October; 1-15 November; 16-30 November; 1-15 De- cember; and 1 6-3 1 December. We did not ob- serve laying after 31 December. We assumed a normal distribution of response variables, and checked it using residual and normal plots. We selected the final model after drop- ping all non-significant terms and checking for changes in deviance (Crawley 1993). We used linear mixed models to test sea- sonal variation of egg size with egg length, width, and mass as response variables; year, clutch size, time of the season, and their sec- ond term interactions were explanatory vari- ables. Nest identity was incorporated as a ran- dom factor to solve the problem of measuring eggs produced by the same female. We used residual and normal probability plots to check model assumptions. We selected the final model by sequentially dropping non-signifi- cant interactions and the non-significant main effects until only significant terms remained, and any additional factor deletion generated a significant change in the model. We assumed a normal distribution of residuals and used an identity link function for the analysis. The wet mass of egg components (albumen, yolk, shell, and membranes) was related to egg total mass, volume, length, and width using sim- ple regression analyses. Shell-free egg mass was estimated as the weight of the fresh egg minus the weight of the shell plus membranes. Yolk and albumen mass were expressed as the per- centage of shell-free egg mass, whereas mass of shell and membranes was expressed as percent- age of total weight of the fresh egg (including shell and membranes). We evaluated differences in size and weight among eggs in successful and deserted nests using a generalized linear model (GLM) in- cluding year, season, clutch size, and mean egg volume as explanatory variables, and nest fate (successful or deserted) as the response variable. We assumed a binomial distribution of residuals for the response variable and a logit link function for this analysis. We used residual and normal probability plots to check model assumptions. We selected a final model by sequentially dropping non-significant inter- actions and then non-significant main effects until only significant terms remained (Crawley 1993). The effect of egg characteristics on probability of hatching was tested using a log- it regression with a binary response (hatched- non-hatched) and egg length, egg width, and season as independent variables. Egg length and width were standardized to the mean val- ue of each nest. Hatching success can vary with egg size in a non-linear form (Deeming 1995), and we included the quadratic terms for egg width and length as predictor variables in an alternative model. We used Genstat DE2 (Release 4.2, VSN International Ltd., UK) to fit the models and perform all statistical anal- yses. All statistics presented are two-tailed and measures are mean ± SD. RESULTS Egg Size Variation. — We measured 1,226 Greater Rhea eggs from 53 nests. Egg volume varied from 255.4 to 788.0 cm^ and egg mass from 275 to 825 g (Table 1). Egg length was more variable than egg width (Table 1). The largest egg was 10.3% larger while the small- est egg was 25.3% smaller than mean values. The widest egg was 1 1 .9% wider while the narrowest egg was 20.5% narrower than mean values. Within clutch variation of egg width was higher than between clutch variation, while variation of egg length was similar within and between clutches. The variance in egg width explained by within clutch variation was 80.8%, whereas variance explained by between-clutches variation was 19.2%. Simi- larly, variation in egg length explained by within clutch variation was 88.7%, while be- tween clutches variation only explained 11.3% of the variance. Variation in egg length within clutches was related to clutch size (variance ratio V7?n,48 ~ 2.18, P = 0.04) but was independent from time of the season (AZ) = 5.43, df = 4, P == 0.02 for clutch size; and P > 0.05 for time of Fernandez and Reboreda • EGG SIZE VARIATION IN GREATER RHEAS 677 TABLE 1. Morphological characteristics of Greater Rhea eggs from 53 nests (mean clutch size SD = 10, range = 7-56 eggs), Buenos Aires Province, Argentina. = 22.6 eggs. Mean SD Range n Length (cm) 13.3 0.55 9.9-14.6 1,226 Width (cm) 9.3 0.38 7.39-10.4 1,225 Mass (g) 618.7 65.1 275-825 1,113 Volume (cm^) 595.4 66.5 255.4-788.0 1,125 Density (g/cm^)"* 1.0 0.06 0.75-1.3 1,107 Area (cm^)*’ 340.3 24.0 199.2-412.2 1,113 Shell density (g/cmO'" 2.1 0.003 2.1-2. 1 1,113 ^Density = 1.038 X weight*’*’®^ (Paganelli et al. 1974). Area = 4.835 X weight°-^^2 (Paganelli et al. 1974). Shell density = 1.945 X weight^ o^"^ (Paganelli et al. 1974). the season and the interaction term). The co- efficient of variation in egg length increased with clutch size. In contrast, within-clutch variation in egg width did not vary with clutch size or season (variance ratio V7?u48 = 0.67, P = 0.75). E^g Composition. — Yolk mass represented 29.5 ± 3.5% (n = 11, range = 24.1-35.6%) of egg mass and 34.7 ± 3.4% (range = 29.5- 42.5%) of shell-free egg mass. Mean yolk mass was 170 ± 18.1 g and varied linearly with egg mass (Fj lo =10.9, P = 0.01, = 0.58), and egg volume (F^ jq = 13.2, P = 0.04, = 0.34; Fig. 1). Yolk mass increased with egg width (regression analysis, Fj ,o =13.2, P = 0.005, R^ = 0.57), but was independent of egg length (regression analysis, Fj lo = 0.135, P = 0.72). Albumen mass represented 63.9 ± 3.5% (n = 10, range = 51.7-70.5 %) of egg mass and 66 ± 2.4% (range = 61.3-70.5%) of shell- free egg mass. Mean albumen mass was 330.5 ± 28.3 g and increased with egg length, width, mass, and volume (regression analyses; F,,8 = 7.23, P = 0.03, F2 = 0.47; F, « -5.31, P = 0.05, F2 = 0.32; F,,8 =18.7, P = 0.002, R^ = 0.64; and F, ^ - 37.9, P < 0.01, R^ = 0.83, respectively; Fig. 1). We did not detect a relationship between albumen and yolk mass (Pearson product-moment correlation, r = 0.42, P > 0.05). Mean mass of shell and membranes was 84.0 ± 15.1 g and represented 14.5 ± 2.4% of egg mass (range = 11.8-19.6%). We did not find any relationship among mass of shell plus membranes and egg mass, volume, length or width (regression analyses, P > 0.05). Effect of Laying Order and Time of Breed- ing on Egg Characteristics. — We found no ef- fect of laying order on egg length (Fj 135 = 1.54, P = 0.22), egg width (F^ 13^ = 0.13, P = 0.72), and egg mass (Fj ,35 = 0.81, P = 0.37). Similarly, egg characteristics were not affected by clutch size (Fj 35 = 1.45, P = 0.24 for length; Fj 35 = 0.74, P = 0.39 for breadth; Fj 35 = 0.05, P = 0.82 for mass). The minimal model for explaining between- clutch variation of egg length included season and the interaction between year and season (Wald/df = 3.74, df = 5, F = 0.002, and Wald/df = 2.67, df = 4, F = 0.03, respec- tively). Egg length decreased towards the end of the breeding season (Dec), but this decrease was most important during 1992 (Fig. 2 A). We did not detect any effect of season, clutch size, and year on egg width (Fig. 2B). Egg mass had a significant decrease with time of breeding, and the interaction between year and time of breeding (Wald/df = 2.80, df = 5, F = 0.02, and Wald/df = 3.31, df = 3, F = 0.02, respectively). Egg mass also decreased to- wards the end of the breeding season and the decrease was most prominent during 1992. Effect of Egg Size on Nest Late and Hatch- ing success. — Mean egg size, clutch size, and season did not affect nest fate (Deviance ratio = 0.84, df,,g = 8, dt;,, = 29, F = 0.57). Suc- cessful and deserted nests had similar mean egg sizes (r-test, a posteriori, t^j = -0.05, F = 0.96). Hatching success was not affected by egg size (A/) = 0.77, df = 1, F = 0.39) but was affected by season (AD = 5.43, df = 4, F = 0.01) although it does not vary in a pre- dictable manner (simple regression, F, -,0 = 0.015, F = 0.90). DISCUSSION Egg Size Variation. — Eggs of Greater Rhe- as varied considerably in size, differing in Yolk / Albumen mass (g) Yolk / Albumen mass(g) Yolk / Albumen mass (g) 678 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 Egg breadth (cm) LIG. 1. Variation of yolk and albumen mass with egg mass, egg length, and egg width. Lull lines rep- resent adjusted functions for albumen mass variation with these independent variables, whereas broken lines represent the adjusted functions for variation in yolk mass. mass by up to 550 g. Egg length and width also varied, as the largest rhea egg was 148% larger than the smallest, while the widest egg was 140% wider than the most narrow one. We detected differences in egg size between clutches associated with date of egg laying. There was a significant reduction in length but not in egg width during the breed- ing season. Other authors also found a nega- tive association between egg size and laying date (i.e.. Coulson and White 1958, Furness 1983, but see Perrins 1996). The reduction in egg size could affect embryo development and hatching success (Dzialowski and Soth- erland 2004) or post-fledgling survival (Dow and Fredga 1984, Newton and Marquiss 1984. Dzus and Clark 1998. but see Blomqvist et al. 1997 and Massaro et al. 2002). We did not detect differences in a previous study in short- term survival between chicks hatched early (Nov-Dec) and late in the breeding seasons (Jan-Feb) (Fernandez and Reboreda 2003). The lack of differences in egg size between successful and failed nests, and the absence of effects of egg size on the probability of hatch- ing appear to indicate that seasonal reduction in egg size in Greater Rheas does not affect female fitness. The decline in egg size could be the con- sequence of female differences in age and/or experience. Some authors have reported the ability of females to produce eggs improves with age, resulting in older females laying ear- lier and larger eggs (Hipfner et al. 1997, Mas- saro et al. 2002). Similarly, eggs laid by young Greater Rhea females in captivity are smaller than those laid by older, experienced females (Flieg 1973. Guittin 1985, Gunski 1992). Besides, nesting success increases dur- ing the breeding season (Fernandez and Re- boreda 1998). Therefore, young females would be expected to have higher fitness than older hens if the differences in egg size we found were age related. Another hypothesis for explaining seasonal variation in egg size postulates that females are constrained by laying late in the season (Lack 1968). Food supply, nutrient availabil- ity, or body reserves late in the season could affect the capability of females for laying (Williams 1994, 2005). In a related species. Emu {Dromaius novaehollandiae), clutch size appears to vary with amount of food available before laying (Davies 2002). Polyandry in rheas could similarly increase female demand for nutrients as the breeding season advances, limiting female capability for egg synthesis. Constraints in food quality or supply could af- Fernandez and Reboreda • EGG SIZE VARIATION IN GREATER RHEAS 679 O) c 0) D) O) LU 14.0 13.8 13.6 13.4 13.2 13.0 12.8 12.6 (A) o • o 9 0 1 o o t o o 9.8 9.7 9.6 9.5 ■g 9.2 m 9.1 8 o o o o o o 8 (B) o • o 0 s 1 o 8 • o o o > > o o O O o z O z CD Q CD Q If) T— CO 1 lO o lO 1 CO T— CO T — CD T — 1 T — 1 CD T — 1 1 CD T— FIG. 2. Seasonal variation in egg length (A), and width (B) in Greater Rheas. Open and black dots corre- .spond to egg mean values for clutches measured during 1992 and 1993, respectively. feet egg size without affecting hatching suc- cess and chick survival. We found that intra-clutch variation of egg size increased with clutch size. This variation could reflect the mating system of rheas. In this species, males defend a group of females that lay their eggs communally in a single nest. Thus, the variation detected may be the consequence of a larger number of different females laying their eggs in the same nest, as larger clutches are the result of larger harems laying in the same nest (Fernandez and Re- boreda 1998). Content and Size Variation. — Yolk mass represented —35% of egg content. This value was slightly smaller than predicted (39. 1 ) for precocial eggs using the equation of Sotherland and Rahn ( 1987). It was similar to 680 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 that measured in Common Ostrich {Struthio ccimelus, 37.8%; Sotherland and Rahn 1987), but smaller than that of Southern Cassowary (Casuarius casuarius, 42%; Sotherland and Rahn 1987), Emu (47%; Dzialowski and Soth- erland 2004), and Southern Brown Kiwi (Ap- teryx australis, 61%; Calder et al. 1978). The smaller content of yolk mass in rhea eggs is associated with a shorter incubation period (40 days) compared to other ratites (Emu; 56 days; Cassowary: 50-52 days; Kiwi: 70-90 days; Davis 2002) but similar to that of os- triches (40-45 days). The percentage of yolk mass of rheas that we measured was similar to that obtained in a previous study with eggs collected in captive and wild populations (Na- varro et al. 2001). That study found that wet yolk represented —34-36% of shell-free egg mass. Seasonal reduction in egg length and the increase in egg size differences within clutch- es as clutch size increases were not associated with changes in yolk content, but likely with albumen mass, as the latter was associated with egg length. Females under nutritional constraints may reduce investment by reduc- ing albumen content, while keeping yolk con- tent constant. If natural selection favors re- duction in variation of egg characteristics that affect chick fitness (Jover et al. 1993), females with nutritional constraints could vary invest- ment in egg components other than yolk con- tent (Carey et al. 1980). Thus, it would be expected that lower seasonal variation in egg width (which is associated with yolk content) than in egg length (which is associated with albumin content) would produce no effect on egg size on hatching success. Additional sup- port is provided by the absence of an associ- ation between variation in egg width and clutch size, which could reflect the high con- stancy of yolk contained within eggs. Variation of egg size in Greater Rheas was not associated with a decrease in hatching suc- cess or nesting success. Thus, there are no ap- parent benefits from laying larger eggs, as re- ported for other precocial species (i.e., Hepp et al. 1989). Seasonal variation of egg size in Greater Rheas would be the result of environ- mental constraints (nutrients or food quality or availability) and/or variation in quality, age or experience of laying females, rather than an adaptive strategy of females to maximize hatching success and chick survival. Further data about nutritional constraints and female investment in eggs are necessary to confirm this hypothesis. ACKNOWLEDGMENTS We thank John Boote and Horacio Martinez Guer- rero for allowing us to conduct this study at Estancias Los Yngleses and La Clementina, respectively, and Raul Paso, Jose Flores, and Angel Guzman for their field collaboration. Myriam Mermoz, Fernando Lor- enzini, and Silvia Rossi helped at different stages of field work. Mario Beade from Fundacion Vida Silves- tre Argentina provided logistical support during the field work. GJF and JCR are research fellows of Con- sejo Nacional de Investigaciones Cientificas y Tecni- cas de Argentina (CONICET). LITERATURE CITED Birkhead, T. R. and D. N. Nettleship. 1982. The adaptive significance of egg size and laying date in Thick-billed Murres Uria lomvia. Ecology 63: 300-306. Blomqvist, D., O. C. Johansson, and F Gotmark. 1997. Parental quality and egg size affect chick survival in a precocial bird, the Lapwing Vanellus vanellus. Oecologia 110:18-24. Calder, W. A., C. R. Parr, and D. P. Karl. 1978. Energy content of eggs of the Brown Kiwi Apter- yx australis; an extreme in avian evolution. Com- parative Biochemistry and Physiology 60:177- 179. Carey, C. 1996. Female reproductive energetics. Pag- es 324-375 in Avian energetics and nutritional ecology (C. Carey, Editor). Chapman and Hall, New York, USA. Carey, C., H. Rahn, and P. Parise 1980. Calories, water, lipid and yolk in avian eggs. Condor 82: 335-343. Christians, J. K. 2002. Avian egg size: variation with- in species and inflexibility within individuals. Bi- ological Reviews 77:1-26. COULSON, J. C. AND E. WHITE. 1958. The effect of age on the breeding biology of the Kittiwake Rissa tridactyla. Ibis 100:40—51. Crawley, M. 1993. Glim for ecologists. Blackwell Scientific Publishing, Oxford, United Kingdom. Cunningham, E. J. A. and A. F. Russell. 2000. Egg investment is influenced by male attractiveness in the Mallard. Nature 404:74-76. Custer, T. W. and P. C. Frederick. 1990. Egg size and laying order of Snowy Egrets, Great Egrets, and Black-crowned Night-herons. Condor 92: 772-775. Davis, J. W. F. 1975. Age, egg size and breeding suc- cess in the Herring Gull Lams argentatus. Ibis 117:460-473. Davis, S. J. J. E 2002. Ratites and tinamous. Oxford Unversity Press, New York, USA. Fernandez and Rehoreda • EGG SIZE VARIATION IN GREATER RHEAS 681 Deeming, D. C. 1995. Eactors affecting hatchability during commercial incubation of Ostrich (Struthio camelus) eggs. British Poultry Science 36:51-56. Dow, H. AND S. Fredga. 1984. Factors affecting re- productive output of the golden-eye duck Buce- phala clangula. Journal of Animal Ecology 53: 679-692. Dzialowski, E. M. and P. R. Sotherland. 2004. Ma- ternal effects of egg size on Emu Dromaiiis no- vaehollandiae egg composition and hatchling phe- notype. Journal of Experimental Biology 207: 597-606. Dzus, E. H. AND R. G. Clark. 1998. Brood survival and recruitment of Mallard in relation to wetland density and hatching date. Auk 115:311-318. Fernandez, G. J. and J. C. Reboreda. 1998. Effects of clutch size and timing of breeding on repro- ductive success of Greater Rheas. Auk 115:340- 348. Fernandez, G. J. and J. C. Reboreda. 2003. Male parental care in Greater Rheas {Rhea americana) in Argentina. Auk 120:418-428. Flieg, G. M. 1973. Nutritional problems in young rat- ites. International Zoo Yearbook 13:158-163. Furness, R. W. 1983. Variations in size and growth of Great Skua Catharacta skua chicks in relation to adult age, hatching date, egg volume, brood size and hatching sequence. Journal of Zoology 199: 101-116. Grant, G. S., C. V. Paganelli, T. N. Petit, G. C. Whittow, and H. Rahn. 1982. Determination of fresh egg mass during natural incubation. Condor 84:121-122. Guittin, P. 1985. Les strutioniformes en pare zoolo- gique, reproduction, croissance, elevage. These de Doctorat. Universite Paris, France. Gunski, R. j. 1992. Analise citogenetica e algunas consideragoes biologicas da especie Rhea ameri- cana - ema (aves:Rheidae). Thesis. Facultad de Ciencias Agrarias y Veterinaria. Universidad Es- tadual Paulista, Sao Paulo, Brasil. Hepp, G. R., R. a. Kennamer, and W. E Harvey. 1989. Recruitment and natal philopatry of Wood Ducks. Ecology 70:897-903. Hipfner, j. M., a. j. Ga.ston, and L. N. D. Fore.st. 1997. The role of female age in determining egg size and laying date of Thick-billed Murres. Jour- nal of Avian Biology 28:271-278. Hoyt, D. F. 1979. Practical methods of estimating vol- ume and fresh weight of bird eggs. Auk 96:73- 77. JovER, L., X. Ruiz, and M. Gonzalez-Martin. 1993. Significance of intraclutch egg size variation in the Purple Heron. Ornis Scandinavica 24:127- 134. Kennamer, R., S. K. Alsum, and S. V. Colwell. 1997. Composition of Wood Duck eggs in relation to egg size, laying sequence, and skipped days of laying. Auk 1 14:479-487. Lack, D. 1968. Ecological adaptations for breeding in birds. Methuen, London, United Kingdom. Macedo, R. H. E, M. O. Cariello, A. M. Pacheco, AND H. G. ScHWABL. 2004. Significance of social parameters on differential nutrient investment in Guira Cuckoo {Guira guira) eggs. Animal Behav- iour 68:685-694. Magrath, R. D. 1992. The effect of egg mass on the growth and survival of Blackbirds: a field exper- iment. Journal of Zoology 227:639-653. Massaro, M., j. T. Darby, L. S. Davis, K-A. Edge, AND M. J. Hazel. 2002. Investigation of interact- ing effects of female age, laying dates, and egg size in Yellow-eyed Penguins (Megadyptes antip- odes). Auk U9:U37-\\4\. Nager, R. G., P. Monaghan, D. C. Houston, K. E. Arnold, J. D. Blount, and N. Verboven. 2006. Maternal effects through the avian egg. Acta Zoologica Sinica 52(Supplement):658-661 . Navarro, J. L., M. L. Lopez, D. M. Maestri, and D. O. Labuckas. 2001. Physical characteristics and chemical composition of Greater Rhea {Rhea americana) eggs from wild and captive popula- tions. British Poultry Science 42:658-662. Newton, I. and M. Marquiss. 1984. Seasonal trend in the breeding performance of Sparrowhawks. Jour- nal of Animal Ecology 53:809-829. Nisbet, I. C. T. 1978. Dependence of fledging success on egg size, parental performance and egg com- position among Common and Roseate terns. Ster- na hirundo and S. dougallii. Ibis 120:207-215. O’Connor, R. J. 1979. Egg weight and brood reduc- tion in the European Swift {Apus apus). Condor 81:133-145. Paganelli, C., A. Oliszowka, and A. Ar. 1974. The avian egg: surface area, volume, and density. Con- dor 76:319-325. Parsons, J. 1970. Relationship between egg size and posthatching chick mortality in the Herring Gull {Larus argentatus). Nature 228:1221-1222. Pelayo, j. T. and R. G. Clark. 2003. Consequences of egg size for offspring survival: a cross-fostering experiment in Ruddy Ducks {Oxyura jamaicen- sis). Auk 102:384-393. Perrins, C. M. 1996. Eggs, egg formation and the tim- ing of breeding. Ibis 138:2-15. PiEROTTi, R. AND C. A. Bellrose. 1986. Proximate and ultimate causation of egg size and the “third-chick disadvantage” in the Western Gull. Auk 103:401- 407. Reid, W. V. and P. D. Boersma. 1990. Parental quality and selection on egg size in the Magellanic Pen- guin. Evolution 44:1780-1786. Shaw, P. 1985. Brood reduction in the Blue-eyed Shag Rhalacrocorax atriceps. Ibis 127:476-494. Slagsvold, T, j. Sandvick, G. Roe.stad. O. Loren- TSEN, AND M. Hlisby. 1984. On the adaptive value of intraclutch egg-size variation in birds. Auk 101 : 685-697. Smith, H. G. and M. Bruun. 1998. The effect of egg size and habitat on Starling nestling growth and survival. Oecologia 1 15:59-63. 682 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 SOTHERLAND, P. R. AND H. Rahn. 1987. On the com- position of bird eggs. Condor 89:48-65. Styrsky, J. D., R. C. Dobbs, and C. F. Thompson. 2002. Sources of egg-size variation in House Wrens {Troglodytes aedon): ontogenetic and en- vironmental components. Auk 119:800-807. Valkama, J., E. KorpimAcki, J. Wiehn, and T. Pak- KANEN. 2002. Inter-clutch egg size variation in Kestrels Falco tinnuncidus: seasonal decline un- der fluctuating food conditions. Journal of Avian Biology 33:426-432. Williams, T. D. 1994. Intraspecific variation in egg size and egg composition in birds: effects on off- spring fitness. Biological Review 68:35-59. Williams, T. D. 2005. Mechanisms underlying the costs of egg production. Bioscience 55:39-48. Williams, T. D., D. B. Lank, and F. Cooke. 1993. Is intraclutch egg-size variation adaptive in the Less- er Snow Goose? Oikos 67:250—256. Yamamoto, J. T, K. M. Shields, J. R. Milliam, T. E. Roudybush, and C. R. Grau. 1989. Reproductive activity of forced paired Cockatiels (Nymphicus hollandiciis). Auk 106:86-93. The Wilson Journal of Ornithology 1 20(4):683— 69 1 , 2008 DNA SEQUENCE ASSESSMENT OE PHYLOGENETIC RELATIONSHIPS AMONG NEW WORLD MARTINS (HIRUNDINIDAE: PROGNE) ROBERT G. MOYLE,' BETH SLIKAS,^ LINDA A. WHITTINGHAM,^ DAVID W. WINKLER,^ AND EREDERICK H. SHELDON^ ^ ABSTRACT. — The classification of New World martins (Progne) has a convoluted history because taxono- mists have relied on plumage traits that vary continuously across populations. We estimated the phylogeny of Progne by analyzing mitochondrial cytochrome h DNA sequences of 27 individuals of eight of the nine species (10 subspecies) and nuclear (B-fibrinogen intron 7 sequences of 20 individuals of six species (8 subspecies). The Brown-chested Martin (P. tapera) is sister to other Progne species. The Middle American taxa — Sinaloa Martin {P. sinaloae), Cuban Martin {P. cryptoleiica), Caribbean Martin (P. dominicensis), and Central American pop- ulations of Gray-breasted Martin (P. chalybea) — form a well supported clade. This group is distinct from Purple Martin (P. siihis), which has no particularly close relatives. All four Middle American taxa appear to be good species, although Cuban and Caribbean martins could be merged in view of their similar plumage and low genetic divergence (1.2%). Two of the South American taxa, the Peruvian Martin (P. murphyi) and Southern Martin (P. elegans), are also distinct species. We did not examine the Galapagos Martin (P. modesta) for lack of DNA, but it is likely to be a good species as well. An unexpected result of the study was that Gray-breasted Martin appears polyphyletic; its South American populations are closer to the Southern Martin than to its Central American populations. Received 4 September 2007. Accepted 1 February 2008. The New World martins are large, generally bluish or brownish swallows that range from Canada to southern Argentina. All taxa are morphologically similar and their plumages often grade into one another. Thus, the group’s classification is unsettled with names of spe- cies and subspecies having been shuffled in numerous ways to convey presumed relation- ships. Current classifications (AOU 1998, Dickinson 2003) generally group the martins in a single genus {Progne) consisting of nine species; P. tapera (Brown-chested Martin), P. siihis (Purple Martin), P. cryptoleuca (Cuban Martin), P. dominicensis (Caribbean Martin), P. sinaloae (Sinaloa Martin), P. chalybea (Gray-breasted Martin), P. modesta (Galapa- ' Natural Hi.story Museum and Biodiversity Re- search Center, and Department of Ecology and Evo- lutionary Biology, University of Kansas, Dyche Hall, Lawrence, KS 66045, USA. 2 Blood Systems Research Institute, 270 Masonic Avenue, San Francisco, CA 941 18, USA. ^ Department of Biological Sciences, University of Wisconsin-Milwaukee, Milwaukee, WI 53201. USA. Cornell University Museum of Vertebrates, De- partment of Ecology and Evolutionary Biology, Cor- nell University, Ithaca, NY 14853, USA. Museum of Natural Science and Department of Bi- ological Sciences, I 19 F4)ster Hall, Louisiana Stale University, Baton Rouge. LA 70803, USA. ^Corresponding author; e-mail: fsheld(«4su.edu gos Martin), P. elegans (Southern Martin), and P. murphyi (Peruvian Martin). This clas- sification facilitates discussion of the martins, but it conveys no phylogenetic structure and does not necessarily indicate appropriate spe- cies boundaries. At the generic level, the martins have been grouped in either Progne (Peters 1960, AOU 1998, Dickinson 2003, Turner 2004) or divid- ed between Progne and Phaeoprogne (Sibley and Monroe 1990). When recognized, Phaeo- progne consists of one species. Brown-chest- ed Martin, which is distinguished from all oth- er New World martins by its lack of irides- cence and sexual dimorphism, and its shorter and less forked tail (Turner 2004). Brown- chested Martin is also genetically distinct from other martin species, although within range of a typical passerine congener (Shel- don and Winkler 1993, Whittingham et al. 2002). Classilication of the eight remaining species of marlins {Progne sensu stricto) is unsettled. Their largely allopatric distributions and sim- ilar plumages have caused several authors to suggest these taxa might constitute a single polytypic species (Hellmayr 1935, Short 1975), but they have not been so classified. Some taxonomists have recognized four spe- cies (Peters I960, Turner 2004): P. suhis 683 684 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 (North America). P. dominicensis (Mexico and Caribbean, including dominicensis. ctyp- toleiica. and sinaloae). P. chcdybea (Central and South America), and P. tnodesta (South America and Galapagos). This arrangement emphasizes male plumage in defining species. P. dominicensis males are dark blue with white bellies. P. chalybea males have blue up- perparts and light underparts with a gray or bluish-gray throat. P. subis and P. modesta males are both entirely blue, but female P. subis are lighter than the chocolate-colored fe- male of P. modesta. Particular confusion has surrounded the re- lationships among the North and Middle American species. Hellmayr (1935) treated Caribbean cryptoleiica and dominicensis as subspecies of North American P. subis. He argued that plumages of these taxa grade into one another, Cuban cryptoleiica being inter- mediate between dominicensis and P. subis in the amount of white in its belly. He recog- nized P. sinaloae as a full species. Zimmer (1955) and Eisenmann (1959) believed that sinaloae, cryptoleiica. and dominicensis should be united in a single species based on the possession of a white belly and also be- cause they form a link between the North American P. subis and Central and South American P. chalybea. Zimmer (1955) noted that although male cryptoleiica resemble male P. subis, female cryptoleiica more closely re- semble female P. chalybea. Phillips (1986) formalized these arguments by grouping cryp- toleiica. dominicensis, and sinaloae in a single species. P. dominicensis (Snowy-bellied Mar- tin). This arrangement divided the North and Central American taxa into three distinct units. P. subis, P. dominicensis. and P. chal- ybea. The cohesiveness of this group was em- phasized by Sibley and Monroe (1990) in des- ignating the superspecies P. subis. Uncertainty also surrounds the “southern martins'’: P. modesta (Galapagos), P. rnurphyi (north coastal Peru), and P. elegans (east of the Andes). Observations on their relation- ships are relatively few in comparison to northern taxa because less is known about them. Hellmayr (1935). Zimmer (1955). Sib- ley and Monroe (1990), and Turner (2004) merged the three into P. modesta. but the AOU (1998) kept them as distinct species (also see Dickinson 2003). The taxa resemble one another with the centrally located P. rnur- phyi somewhat intermediate in plumage (Turner 2004). P. elegans is larger than P. rno- desta and P. rnurphyi and has more white in the female plumage, especially on the fore- head (Short 1975). The disjunct distribution of the southern martins begs the question of how they evolved. They might be isolated and di- verging populations of a single, once wide- spread species, or they might be independent- ly founded populations of northern taxa. An interesting clue, or complication, is that breed- ing ranges of P. elegans and P. chalybea over- lap in the southern Chaco, and the two hy- bridize there (Eisenmann and Haverschmidt 1970. Short 1975). We reconstructed the intrageneric phyloge- ny of New World martins using DNA se- quence data. The purposes of the study were to examine (1) whether the martins are divid- ed into clearly defined clades that may be con- sidered species, and (2) how these clades are related to one another. Knowledge of Progne phylogeny. in combination with geographic and plumage data, would provide the foun- dation for a stable classification useful to the many ecologists and behaviorists who study these birds. It would also provide the frame- work for future studies of population genetics of the genus. METHODS Martin phylogeny was estimated by com- paring mitochondrial cytochrome b (cytb) and nuclear p-fibrinogen intron 7 (pfib7) DNA se- quences (Table 1). DNA of most species was extracted from preserved tissues, but museum skins served as the source of DNA for P. sin- aloae. P. rnurphyi, and P. modesta. Swallows closely related to martins served as outgroups (Sheldon et al. 2005). These included South- ern Rough-winged Swallow {Stelgidopteryx nificollis). White-banded Swallow {Atticora fasciata), Blue-and-white Swallow (Pygoch- elidon cyanoleiica), and Golden Swallow {Tachycineta eiichrysea). Genomic DNA was extracted from pre- served tissue using the phenol/chloroform ex- traction method (Hillis et al. 1990, Slikas et al. 2000). DNA from museum skins was ex- tracted in a designated ancient DNA labora- tory and stringent precautions were followed to avoid contamination. All surfaces were Moyle et al. • NEW WORLD MARTIN PHYTOGENY 685 TABLE 1. List of taxa, specimens, and localities compared. Species^ Sample^ # Cytb length Cytb GenBank # pFibV length pFib? GenBank # Locality Prague tapera (Brown-chested Martin) B7297 955 AE074588 856 AY827412 Loreto Department, Peru P. siibis subis (Purple Martin) B25082 913 AY825996 855 AY827413 East Baton Rouge Parish, LA, USA B25083 863 EU427735 435 Y827413 East Baton Rouge Parish B5866 996 EU427734 435 EU427759 Cameron Parish B 30441 996 EU427733 435 EU427758 Cameron Parish B45004 863 EU427736 855 EU427755 Timbalier Islands, LA, USA B45475 996 EU427742 856 EU427754 Amazonas, Brazil P. subis arboricola (Purple Martin) B41548 996 EU42773 1 435 EU427757 Grant County, NM, USA B41549 AB B38690 996 996 996 EU427732 EU427730 EU427745 435 EU427756 Grant County Maplewood, BC, Canada Riverside County, CA, USA P. cryptoleuca (Cuban Martin) B52873 973 AY825945 856 EU427760 Cuba P. domiuicensis (Caribbean Martin) P. sinaloae (Sinaloa Martin) B22019 KMNH 40044* KMNH 40045* 951 537 537 AY825946 AY825947 EU427740 856 AY827415 Dominican Repub- lic Sinaloa, Mexico Sinaloa, Mexico P. chalybea chalybea (Gray-breasted Martin) B28808 995 AY825948 856 EU427750 Panama Province, Panama B28806 B27272 996 995 EU427738 EU427737 856 EU427749 Panama Province Alajuela Province, Costa Rica B9486 939 AE074583 856 AY827416 Pondo Department, Bolivia B9562 995 EU427739 856 EU427753 Pondo Department B7366 995 EU427740 856 EU427752 Loreto Department, Peru P. chalybea macrorhamplius (Gray-breasted Martin) B15315 996 EU427741 435 EU427760 Santa Cruz Depart- ment, Bolivia P. elegaus (Southern Martin) B25506 996 AY 8 25 949 856 AY827417 Amazonas, Brazil (austral migrant) PRS1690 996 EU427743 435 EU427761 Anelo Department, Argentina P. murphyi (Peruvian Martin) PRS1758 LSUMZ 1 14184* LSUMZ 114186* 996 891 891 EU427744 EU427746 AY825950 856 EU427748 Anelo Department Arequipa Depart- ment, Peru Arequipa Depart- ment, Peru Tachycineta eucluysea (Golden Swallow) B22018 969 AY05245 1 856 AY827409 Dominican Repub- lic Atticora fasciata (White-banded Swallow) B 12682 900 AE074584 856 AY827427 Santa Cruz Depart- ment, Bolivia Pygocheliilon cyanoleuca ( Blue-and-white Swallow) B12191 906 AE074586 857 AY826()03 Pichincha Province, Ecuador Stelgidopteryx ruficollis (Southern Rough-winged B 1 1 30 901 AE074589 856 AY82741 1 La Paz Department, Boli\ ia Swallow) “ Classiticalion follows AOU (1998) and Dickinson (2(K).t). AB, Allan Baker. Royal Ontario Museum; B. l.ouisiana .Stale Museum of Natural .Science Collection of (ienetic Resources; KMNM. Kansas linisersity Museum of Natural History; LSUMZ. LSU Museum of Natural Science Bird Collection; I’RS. Paul Sweet. American Museum of Natural History. *DN.\ extracted from museum skin. 686 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 TABLE 2. Primers used for DNA amplification and sequencing. Primer Sequence (5'-3') Source Cyt h HI 6065 H15710 CBINT LI 5656 LI 4993 LI 4999 L15191 L1531 1 cytbC LL5719 HI 5331 cytbE GGAGTCTTCAGTCTCTGGTTTACAAGAC GTAGGCGAATAGGAAGTATC GGTTGTTTGAGCCGGATTC AACCTACTAGGAGACCCAGA CCATCC7G\CATCTCAGCCTGATGAAACTT CTCATCTTGATGAAACTTCGGATC ATCTGCATCTACCTACACATCGG GCAAGCTTCTACCATGAGGACAAATATC CAAGCTTCTACCATGAGGAC AAAC C C GAATGAT AC TTC C T AAACTGCAGCCCCTCAGAATGATATTT ATAGGAAGTATCATTCGGGT Helm-Bychowski and Cracraft (1993) Helm-Bychowski and Cracraft (1993) Avise et al. (1994) Helm-Bychowski and Cracraft (1993) Modified from Kocher et al. (1989) Modified from Kocher et al. (1989) Modified from Kocher et al. (1989) Modified from Kocher et al. (1989) Modified from Kocher et al. (1989) Modified from Kocher et al. (1989) Modified from Kocher et al. (1989) Modified from Kocher et al. (1989) ppib7 FIB-BI7L FIB-BI7U FIB-BI7Lb FIB-BI7Ub TCCCCAGTAGTATCTGCCATTAGGGTT GGAGAAAACAGGACAATGACAATTCAC CAGTGCTCTATTATGTACTTTAC TGGGTCCTGAAGAAAGAGGC Prychitko and Moore (1997) Prychitko and Moore (1997) Sheldon et al. (2005) Sheldon et al. (2005) cleaned with 10% bleach solution prior to ex- tractions. Pipettors were cleaned with bleach solution following each set of extractions and filter-tips were used. All reagents were ex- posed to UV light (254 nm) for at least 20 min prior to use. Gloves, lab coat, and face mask were worn at all times during extraction and PCR preparation. DNA was extracted in sets of six, including hve samples and a blank extraction control. Samples were hrst chopped hnely with a sterile scalpel blade, then added to 750 fxl of digestion buffer (10 mg/ml DTT, 1 mg/ml Proteinase K, 1% SDS, 10 mM Tris, 2 mM EDTA, 10 mM NaCl2) and incubated overnight in an oven at 55° C in a rotating stand. Each sample was purified with two phe- nol and a single chloroform extraction follow- ing digestion. Each supernatant was washed with sterile, UV-treated H2O and concentrated using Centricon 30 columns (Amicon Inc., Beverly, MA, USA). Each sample was brought to a final volume of 120-160 p.1 with sterile, UV-treated H2O, aliquoted into four tubes of equal volume and heated at 65° C in a heat block for 10 min (to remove DNAse). Extracted DNA was stored in a —20° C freez- er in the ancient DNA laboratory. PCR reac- tions were prepared in an ancient DNA labo- ratory, but PCR amplification and post-PCR procedures were performed in the primary ge- netics laboratory. Amplification and sequenc- ing from skin specimens was done piecewise in fragments ranging from 150 to 400 base pairs. All DNA was sequenced in both directions using the primers (listed in Table 2) and either a Product Sequencing Kit (Version 2.0, U. S. Biochemical, Cleveland, OH, USA), or stan- dard cycle sequencing. DNA was sequenced either by hand on a 6% polyacrylamide gel or with an ABI 373 or ABI 3100 automated se- quencer. A maximum of 996 nucleotides of cytb, spanning sites 15046 to 16041 in the chicken mitochondrial genome (Desjardins and Morais 1990), was sequenced (Table 1). Shorter sequences were obtained for taxa whose DNA was extracted from museum skins. pfib7 lengths were either —435 nucle- otides (half sequences) or —856 nucleotides (whole sequences; Table 1). DNA sequence data were deposited at GenBank (Table 1). Base frequencies and sequence variation and divergence values were calculated using MEGA 3.1 (Kumar et al. 2004). Trees were visualized with TreeView (Page 1996). Max- imum likelihood (ML) analyses were per- formed using PAUP* 4.0b 10 (Swofford 2002) with models meeting the Akaike information eriterion in Modeltest 3.06 (Posada and Cran- dall 1998). Bootstrap analyses consisted of 10 heuristic searches with random addition of taxa and TBR branch swapping and reanalysis of the data 100 times. MrBayes 3.0 (Huelsen- beck and Ronquist 2001) was used to estimate Moyle et al. • NEW WORLD MARTIN PHYLOGENY 687 model parameters from the data and estimate the phylogeny with branch support values. Mixed models were used to build trees when cytb and pfib7 data were combined. Markov chains ran for 5 million generations. A shorter run of one million generations was used to assess stationarity (burn-in). Markov chains were sampled every 1,000 generations, yield- ing 5,000 point estimates of parameters. These subsamples, minus the burn-in generations, were used to create Bayesian consensus trees and to identify posterior probabilities for tree nodes. RESULTS Cytb of 27 individuals of eight Progne spe- cies and pfib7 of 20 individuals of six species were sequenced (Table 1). We failed to obtain cytb from museum specimens of P. modesta and pfib7 from museum specimens of P. mo- desta, P. murphyi, and P. sinaloae. Multiple individuals of most species were sequenced to examine intraspecific variation (Table 1). Cytb sequences of P. sinaloae and P. murphyi in- dividuals were invariable; thus, we used only one sequence of each for phylogenetic recon- struction. Martin cytb varied at 160 nucleotide sites and was potentially informative at 114 sites. pfib7 varied at 26 sites and was poten- tially informative at 1 1 sites. Most cytb nu- cleotide substitutions occurred at third codon positions (81%), followed by first positions (17%) and second positions (3%). Amino ac- ids varied at 17 sites and were potentially in- formative at nine sites. Cytb sequences ap- peared to be mitochondrial, not nuclear, be- cause amino acid substitutions were similar in chemical properties, no stop codons occurred, and third position changes predominated. The mean (range) of nucleotide composition was: thymine 24.8 (23.8-26.6), cytosine 34.4 (33.1-35.2), adenine 27.0 (25.9-27.4), and guanine 13.8 (13.3-14.3). A relatively even distribution of bases occurred at first codon positions (T 23.8, C 29.1, A 23.3, G 23.7); less at second positions (T 41.0, C 26.7, A 19.6, G 12.7), and even less at third positions (T 9.5, C 47.6, A 38.1 G 4.8). Cytb distances between Progne tapera and other Progne species averaged 7.6% (range: 6.7— 8.3%). Distances among the seven species of Progne (sensu stricto) averaged 4.9% (0.1 — 6.9%). An unexpectedly large cytb distance occurred between Central American and Am- azonian populations of P. chalybea (average 5.7%), and a remarkably short distance oc- curred between South American P. chalybea and P. elegans (0.8%). |3fib7 distances be- tween P. tapera and other Progne species av- eraged 1.8% (1. 3-2.0%). Distances among five species of Progne (sensu stricto) averaged O. 4% (0-1.0%). The average distance between Central American and Amazonian populations of P. chalybea was 0.4%, between Amazonian P. chalybea and P. elegans 0.3%, and between P. elegans and Central American P. chalybea 0.2%. Only nine sites were variable and only four were potentially parsimony informative in comparisons of P. chalybea and P. elegans. None of the informative sites was synapo- morphic for Central American and Amazonian populations of P. chalybea or for South Amer- ican P. chalybea and P. elegans. ML bootstrap and Bayesian analyses of cytb and combined cytb and [3fib7 data yield- ed consistent trees. However, the combined tree provided less resolution than the cytb tree among closely related taxa (e.g., among P. subis subspecies), presumably because of the influence of slower evolving (3fib7 sequences. We present the cytb tree as the optimal tree (Lig. 1), i.e., the tree that depicts basal rela- tionships found in the combined tree and dis- tal relationships provided by the fast-evolving cytb gene. All branches with less than 95% Bayesian or 87% bootstrap support in this tree can be collapsed. When this is done, unam- biguous relationships are derived. P. tapera is sister to the rest of Progne. Three clades emerge among the other taxa. P. subis is monophyletic and divided into western P. s. aboricola and eastern P. s. subis. Included in the eastern clade is a wintering specimen col- lected in Brazil. The Middle American taxa form a clade. This group includes P. crypto- leuca and its sister taxon P. doniinicensis, as well as P. sinaloae and P. chalybea from Cos- ta Rica and Panama. P. murphyi is distinct from P. elegans. The latter forms a clade with South American populations of P. chalybea, making that taxon polyphyletic. DISCUSSION We assumed at the outset that species of Progne would be substantially diverged from one another and, thus, appropriate for phylo- 688 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 1.00/93 1.00/95 P. tapera B7297 Peru — P. murphyi LSU1 1 41 86 Peru P. subis arboricola B41 549 New Mexico 1.00/94 1.00/97 P. subis arboricola B38690 California 0.44/57 P. subis aboricoia AB British Columbia 0.28/<50 P. subis arboricola B41 548 New Mexico 0.93/72 0.43/69 P. subis subis B25082 Louisiana P. subis subis B30441 Louisiana 1.00/92 P. subis subis B5866 Louisiana O. 34/<50 P. subis subis B45475 Brazil P. subis subis B25083 Louisiana 0.29/<50l p B45004 Gulf of Mexico 0.37/<50 I- P. chalybea chalybea B9486 Bolivia 0.98/82 1.00/100 P. elegans PRS1690 Argentina O. 98/88 P. elegans PRS1758 Argentina 0.53/<50 1.00/98 1.00/87 C 0.53/51 P. chalybea macrorhamphus B15315 Bolivia O. 64/53 P. elegans B25506 Brazil 0.55/63 P. chalybea chalybea B9562 Bolivia O. 32/<50 P. chalybea chalybea B7366 Peru P. cryptoleuca L52873 Cuba P. dominicensis B22019 Dominican Republic - P. sinaloae K40044 Mexico P. chalybea chalybea B28806 Panama 0.99/89 P. chalybea chalybea B28808 Panama O. 64/63 P. chalybea chalybea B27272 Costa Rica 0.1 FIG. 1. Estimate of Prague phylogeny from cytb data. Numbers on branches are Bayesian/ML bootstrap support. The ML model is GTR + I + G; -InL = 4020.2771; base frequencies: A = 0.30, C = 0.37, G = 0.11, T = 0.22; rate matrix: A-C = 16.9964, A-G = 248.2475, A-T = 23.0512, C-G = 6.0181, C-T = 101.1734, G-T = 1.0000; proportion of invariable sites = 0.52; and gamma distribution shape parameter = 0.64. Moyle et al. • NEW WORLD MARTIN PHYLOGENY 689 genetic study. Our comparisons, however, in- dicate remarkably close relationships between some species (e.g., P. dominicensis and P. cryptoleuca) and polyphyly and possibly hy- brid introgression in others {P. chalybea and P. elegans). These results suggest that ulti- mate understanding of relationships among some Progne taxa will require a population genetic rather than a phylogenetic perspective. Unfortunately, Progne tissues are poorly rep- resented in museum collections and we have not been able to sample the group adequately for a thorough analysis. However, our study provides insight into Progne phytogeny on several levels. It illuminates relationships among P. subis populations in the USA and Canada, largely supports the current nine spe- cies classification, and provides evidence of higher level relationships among some species in the genus, particularly those in Middle America. The phylogenetic analysis strongly supports the traditional division of New World martins into two groups (Fig. 1): P. tapera versus all other species {Progne sensu stricto). We sug- gest that P. tapera remain in Progne as cur- rently classified rather than in the monotypic genus Phaeoprogne. The current arrangement emphasizes its close relationship to the rest of the New World martin species as opposed to other groups of swallows (Sheldon and Wink- ler 1993, Sheldon et al. 2005). The taxonomically difficult part of the ge- nus, Progne sensu stricto, remains partially unresolved, but three well supported clades are evident (Fig. 1). (1) Progne subis of North America forms a distinct group. It is diverged by an average cytb distance of 5.6% from all other Progne species (sensu stricto) and is not particularly close to Middle American species. There is a clear division within P. subis between eastern and western subspecies, subis and arboricola. This division coincides with a boundary at the Rocky Mountains previously recognized based on morphological and life history dif- ferences. Birds from northern Arizona, Utah, northern Rockies, and the Pacific Northwest are generally larger than eastern birds, and fe- males of western birds have whiter underparts and foreheads than their eastern counterparts (Brown 1997). wSubspccific division of' western populations, particularly those in the northern Rockies and northern Pacific coast, has been uncertain (Brown 1997). Our data indicate a close genetic relationship between populations in British Columbia and those in southwestern New Mexico and southern California. Thus, these populations are appropriately classified as a single taxonomic group. Both subis and arboricola in northern Mexico are thought to occur in highland areas (Phillips 1986, Brown 1997), but we had no material from that re- gion for comparison. We also lacked samples of the third subspecies, Hesperia. This lowland taxon occurs in deserts of southern Baja Cal- ifornia, Sonora, and along the coast of Mexico to northern Sinaloa. It is well known for nest- ing in cacti. (2) Martins in Middle America form a clade consisting of Progne dominicensis, P. cryp- toleuca, P. sinaloae, and Central American P. chalybea. A close relationship among these taxa has long been suspected (Eisenmann 1959, Phillips 1986, Sibley and Monroe 1990). The cytb of P. dominicensis and P. cryptoleuca is diverged by only 1.2%, and these two taxa form a monophyletic group. They may be considered different species be- cause of their allopatry and reduced white on the belly of male P. cryptoleuca (McKitrick and Zink 1988), or they may be considered conspecific given their monophyly and genetic and plumage similarities (Johnson et al. 1999). P. sinaloae cytb is 3.5% diverged from that of Caribbean taxa and 3.4% diverged from that of P. chalybea. P. sinaloae appears to be a distinct species despite similarities in plumage to P. dominicensis, P. cry'ptoleuca, and P. chalybea (Zimmer 1955, Eisenmann 1959). (3) The southern martins appear not to be conspecific. P. murphyi of northwestern Peru is distinct from P. elegans, which occurs east of the Andes. P. murphyi is smaller, has less white in the female plumage, and differs from P. elegans in cytb sequence by 5.6%. We found that P. elegans is unexpectedly close to southern populations of P. chalybea (0.8% cytb divergence). P. chalybea, in turn, appears to be polyphyletic. Its Amazonian populations are close to P. elegans, whereas its Central American populations form a clade with Mid- dle American taxa (average cytb divergence 3.5%). Not only is cytb of Amazonian l\ chal- ybea substantially different from that of Cen- 690 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 tral American populations (5.7% divergence), but individuals from central Amazonia are darker-throated and males have a larger amount of blue on the sides of their breasts than individuals from trans-Andean and Cen- tral American populations (D. F. Lane, pers. comm.). Both cis- and trans-Andean P. chal- ybea are classified as the same subspecies (chalybea) despite these differences. The other subspecies, macrorhamphus from southeast- ern Amazonia, is strikingly pale-throated compared to chalybea of central Amazonia, but the genetic difference between these groups is negligible (0.1% cytb divergence). It is conceivable that Amazonian popula- tions of P. chalybea represent an unrecog- nized taxon that is closely related to P. ele- gans. This seems unlikely, however, because of consistent plumage differences between P. chalybea, which is steel-blue above and ex- tensively pale below, and P. elegans, whose males are all blue. Another explanation is in- terspecific hybridization between the two spe- cies. Hybridization is known to occur (Eisen- mann and Haverschmidt 1970, Short 1975) and may have resulted in introgression of P. elegans mtDNA into Amazonian P. chalybea. Introgression would explain why the mtDNA of Amazonian and Central American P. chal- ybea is so different. Unfortunately, nuclear pfib7 data do not support or refute this idea. P. elegans and the Central and South Ameri- can populations of P. chalybea have similar pfib7 sequences (0. 2-0.4% divergence), and no nucleotide substitutions are synapomorphic for any pair of populations. Future work on the evolution and bioge- ography of New World martins will require improved sampling. The position of the Ga- lapagos Martin (P. modesta) remains in doubt. It is usually considered to be closely related to P. murphyi and P. elegans based on plum- age, but this arrangement seems unlikely giv- en the genetic distinctness of P. murphyi and P. elegans from one another. A more likely explanation is that Galapagos Martin derives from errant migrant populations. The problem of polyphyly in P. chalybea needs to be ap- proached with greater sampling from northern South America to help gauge genetic variation across the entire range of the species. Several other relationships within Prague also merit examination, for example, the connection of the three subspecies of P. subis in Mexico and the genetic distinctiveness of P. chalybea war- neri in western Mexico. ACKNOWLEDGMENTS We thank A. J. Baker, E. A. Cardiff, S. W. Cardiff, Mario Cohn-Haft, D. L. Dittmann, N. K. Klein, M. B. Robbins, R. S. Ridgely, F. Sornoza Molina, Paul Sweet, Katherine Wallace, Bret Whitney, Academy of Natural Sciences of Philadelphia, American Museum of Natural History, Cornell University, LSU Museum of Natural Science, Museum of Vertebrate Zoology, University of Kansas Natural History Museum, Royal Ontario Museum, and Smithsonian Institution for help obtaining tissues. Daniel E Lane provided advice on plumage variation in Gray-breasted Martin. We also thank the Museo Ecuatoriano de Ciencias Naturales and the Ministerio de Agricultura, Quito; and institu- tions in Bolivia, Dominican Republic, Ecuador, Mex- ico, Panama, and Peru for their roles in making spec- imens and tissues available for study. This research was supported financially by the American Museum of Natural History, LSU Museum of Natural Science, Museum of Vertebrate Zoology, National Zoo, and NSF/LaSER 1 993-96- ADP-02. LITERATURE CITED American Ornithologists’ Union (AOU). 1998. Check-list of North American birds. Seventh Edi- tion. American Ornithologists’ Union, Washing- ton, D.C., USA. Avise, J. C., W. S. Nelson, and C. G. Sibley. 1994. DNA sequence support for a close phylogenetic relationship between some storks and New World vultures. Proceedings of the National Academy of Science of the USA 91:5173-5177. Brown, C. R. 1997. Purple Martin (Progne subis). The birds of North America. Number 287. Desjardins, P. and R. Morais. 1990. Sequence and gene organization of the chicken mitochondrial genome, a novel gene order in higher vertebrates. Journal of Molecular Biology 212:599-634. Dickinson, E. C. 2003. The Howard & Moore com- plete checklist of the birds of the world. Third Edition. Christopher Helm, London, United King- dom. Eisenmann, E. 1959. South American migrant swal- lows of the genus Progne in Panama and northern South America; with comments on their identifi- cation and molt. Auk 76:528-532. Eisenmann, E. and F. Haverschmidt. 1970. North- ward migration to Surinam of South American martins {Progne). Condor 72:368-369. Hellmayr, C. E. 1935. Catalogue of the birds of the Americas. Part 8. Field Museum of Natural His- tory Zoological Series 13:1—541. Helm-Bychowski, K. and J. Cracraft. 1993. Recov- ering phylogenetic signal from DNA sequences: relationships within the corvine assemblage (class Aves) as inferred from complete sequences of the Moyle et al. • NEW WORLD MARTIN PHYLOGENY 691 mitochondrial DNA cytochrome-/? gene. Molecu- lar Biology and Evolution 10:1 196-1214. Hillis, D. M., a. Larson, S. K. Davis, and E. A. Zimmer. 1990. Nucleic acids. III. Sequencing. Pages 318-370 in Molecular systematics (D. M. Hillis and C. Moritz, Editors). Sinauer Associates, Sunderland, Massachusetts, USA. Huelsenbeck, J. P. and E Ronquist. 2001. MRB AY- ES: Bayesian inference of phylogenetic trees. Bioinformatics 17:754-755. Johnson, N. K., J. V. Remsen, and C. Cicero. 1999. Resolution of the debate over species concepts in ornithology: a new comprehensive biologic spe- cies concept. Proceedings of the International Or- nithological Congress 22:1470-1482. Kocher, T. D., W. K. Thomas, A. Meyer, S. V. Ed- wards, S. PAABO, E X. ViLLABLANCA, AND A. C. Wilson. 1989. Dynamics of mitochondrial DNA evolution in animals: amplification and sequenc- ing with conserved primers. Proceedings of the National Academy of Sciences of the USA 86: 6196-200. Kumar, S., K. Tamura, and M. Nei. 2004. MEGA3: integrated software for molecular evolutionary ge- netics analysis and sequence alignment. Briefings in Bioinformatics 5:150-163. McKitrick, M. C. and R. M. Zink. 1988. Species con- cepts in ornithology. Condor 90:1-14. Page, R. D. M. 1996. TREEVIEW: an application to display phylogenetic trees on personal computers. Computer Applications in the Biosciences 12: 357-358. Peters, J. L. 1960. Family Hirundinidae. Pages 80- 129 in Check-list of birds of the world. Volume 9 (J. L. Peters, Editor). Museum of Comparative Zo- ology, Cambridge, Massachusetts, USA. Phillips, A. R. 1986. The known birds of North and Middle America. Part I. Allan R. Phillips, Denver, Colorado, USA. Posada, D. and K. A. Crandall. 1998. ModelTe.st: testing the model of DNA substitution. Bioinfor- matics 14:817-818. Prychitko, T. M. and W. S. Moore. 1997. The utility of DNA sequences of an intron from the beta- fibrinogen gene in phylogenetic analysis of wood- peckers (Aves: Picidae). Molecular Phylogenetics and Evolution 8:193-204. Sheldon, E H. and D. W. Winkler. 1993. Intergeneric phylogenetic relationships of swallows estimated by DNA-DNA hybridization. Auk 1 10:798-824. Sheldon, E H., L. A. Whittingham, R. G. Moyle, B. Slikas, and D. W. Winkler. 2005. Phylogeny of swallows (Aves: Hirundinidae) estimated from nuclear and mitochondrial DNA sequences. Mo- lecular Phylogenetics and Evolution 35:254-270. Short, L. L. 1975. A zoogeographic analysis of the South American Chaco avifauna. Bulletin of the American Museum of Natural History 154:165- 352. Sibley, C. G. and B. L. Monroe. 1990. Distribution and taxonomy of birds of the world. Yale Univer- sity Press, New Haven, Connecticut, USA. Slikas, B., I. Jones, S. Derrickson, and R. Fleischer. 2000. Phylogenetic relationships of Micronesian white-eyes based on mitochondrial sequence data. Auk 117:355-365. SwoFFORD, D. L. 2002. Phylogenetic analysis using parsimony (*and other methods). Version 4. Si- nauer Associates, Sunderland, Massachusetts, USA. Turner, A. K. 2004. Family Hirundinidae (swallows and martins). Pages 602-685 in Handbook of the birds of the world. Volume 9 (J. del Hoyo, A. Elliott, and D. Christie, Editors). Lynx Edicions, Barcelona, Spain. Whittingham, L. A., B. Slikas, D. W. Winkler, and E H. Sheldon. 2002. Phylogeny of the Tree Swal- low genus (Aves: Tachycineta) by Bayesian anal- ysis of mitochondrial DNA sequences. Molecular Phylogenetics and Evolution 22:430-441. Zimmer, J. T. 1955. Studies of Peruvian birds. Number 66. The swallows (Hirundinidae). American Mu- seum Novitates 1723:1-35. The Wilson Journal of Ornithology 1 20(4):692— 699, 2008 PRELIMINARY ANALYSIS OL THE ECOLOGY AND GEOGRAPHY OP THE ASIAN NUTHATCHES (AVES: SITTIDAE) SHAILY MENON,‘-5 ZAFAR-UL ISLAM,^'' JORGE SOBERON,^ AND A. TOWNSEND PETERSON’ ABSTRACT. — We explored distributions of Asian nuthatch species in ecological and geographic space using ecological niche modeling based on occurrence data associated with specimens and observations. Nuthatches represent a well-defined clade occurring throughout the Northern Hemisphere, but are most diverse in southern Asia where 15 of the 24 species occur and where the lineage is believed to have evolved. Species richness was focused in a narrow east-west band corresponding to the forested parts of the Himalayas with a maximum number of nine species predicted present in these foci. The distributional predictions have a mid-elevation focus with highest species diversity between 1,000 and 2,000 m. Niche breadth and volume were positively related, but accumulation of distributional area (niche volume) decreased with additional environmental combinations (niche breadth). The extent of potential range filling, a measure of distributional disequilibrium, was connected with montane habit (R^ = 0.422) indicating that montane situations limit the distributional potential of species. Received 13 September 2007. Accepted 1 February 2008. The Sittidae consists of 25 species in two genera; Sitta with 24 species (nuthatches) and Tichodroma with a single species (Wallcreep- er [T. muraria]). The family was conceived much more broadly (Mayr and Amadon 1951) to contain other, superficially similar groups, such as Daphoenositta (the sittellas of New Guinea and Australia), now known to have converged on a similar feeding niche and mor- phology from different ancestry (Sibley and Ahlquist 1990). True nuthatches occur throughout the Northern Hemisphere, but are most diverse in southern Asia where 1 5 of the 24 species occur (Harrap and Quinn 1995), and where the lineage probably evolved (Mat- thysen 1998). The co-occurrence of many closely-related species was noted by Ripley (1959) and Lack (1971). Matthysen (1998) observed that many nuthatch species have small ranges, although 5-7 nuthatch species occur at sites across southeast Asia, generally segregated by elevation, habitat, or both (Lack 1971, Matthyssen 1998). Nuthatch distributions range from narrow endemism to broad distributions crossing con- ‘ Biology Department, Grand Valley State Univer- sity, Allendale, MI 49401, USA. 2 Bombay Natural History Society, Mumbai, India. Natural History Museum and Biodiversity Re- search Center, University of Kansas, Lawrence, KS 66045, USA. Current address: National Wildlife Research Cen- ter, Taif, Saudi Arabia. -^Corresponding author; e-mail: menons@gvsu.edu tinents. For example, the White-browed Nut- hatch {Sitta victoriae) is microendemic, re- stricted to Mount Victoria in western Burma; in contrast, Eurasian Nuthatch (S. europaea) ranges across much of Eurasia. Nuthatches are typical in temperate and subtropical areas of the Northern Hemisphere with two species {S. europaea and S. canadensis) occurring well into the subarctic, almost to the northern edge of the boreal forest (Matthysen 1998). Eewer nuthatch species occur in tropical regions of southern Asia and only two {S. frontalis and S. azurea) reach the Equator. New informatic approaches offer novel in- sights into the interaction between ecology and geography in evolving lineages (Soberon and Peterson 2004, 2005). In particular, di- verse hypotheses relevant to distributional ecology and biogeography can be tested: con- servatism of ecological niche characteristics (Peterson et al. 1999), ecological innovation (Peterson and Holt 2003, Graham et al. 2004), distributional equilibrium with climate fea- tures (Svenning and Skov 2004), identification of barriers to dispersal (Peterson 2003), the role of interspecific competition in shaping species’ distributions (Anderson et al. 2002), and others. The objective of this paper is to analyze and explore the ecology and distri- butions of Asian nuthatch species as a first step toward a more integrative view of nut- hatch evolution and biogeography. METHODS Data. — We focused on 14 species in the ge- nus Sitta and one species in the genus Ti- 692 Menon et al. • ASIAN NUTHATCHES 693 TABLE 1. Species analyzed and occurrence data available for each. Crude estimates of actual distribution area, potential distribution area, and proportional range filling are included. Species Common name Number of occurrence points Actual distribution area (km^) Potential distribution area (km^) Proportional range filling Sitta cashmirensis Kashmir Nuthatch 17 5,387 21,739 0.25 S. castanea Chestnut-bellied Nuthatch 69 44,362 98,433 0.45 S. europaea Eurasian Nuthatch 40 240,678 275,212 0.87 S. formosa Beautiful Nuthatch 63 7,800 24,618 0.32 S. frontalis Velvet-fronted Nuthatch 60 41,620 83,444 0.5 S. himalayensis White-tailed Nuthatch 27 5,829 93,258 0.06 S. leucopsis White-cheeked Nuthatch 23 10,212 38,181 0.27 S. magna Giant Nuthatch 45 7,448 18,345 0.41 S. nagaensis Chestnut-vented Nuthatch 27 9,512 41,953 0.23 S. solangiae Yellow-billed Nuthatch 3 77 2,171 0.04 S. tephronota Eastern Rock Nuthatch 34 29,776 71,930 0.41 S. victoriae White-browed Nuthatch 2 48 48 1.00 S. villosa Chinese Nuthatch 10 11,779 46,344 0.25 S. yimnanensis Yunnan Nuthatch 7 3,005 11,830 0.25 Tichodroma muraria Wallcreeper 56 126,802 142,392 0.89 chodroma occurring in Eurasia. Occurrence information was accumulated from natural history museums across North America, in- cluding the Museum of Comparative Zoology, Field Museum of Natural History, University of Kansas Natural History Museum, and the U.S. National Museum of Natural History; data were also drawn from data bases devel- oped by BirdLife International (Collar et al. 2001). Textual descriptions of occurrence lo- calities were translated into geographic coor- dinates in decimal degrees using the GeoNet Names Server (National Geospatial Intelli- gence Agency 2007) and BioGeomancer (Chapman and Wieczorek 2006). The final data set consisted of 483 occurrence points with samples for individual species ranging from 2 to 69 (Table 1). The amount of occur- rence data available for Asian nuthatch spe- cies was variable. For most species, we had more than sufficient information to character- ize ecology and distribution. However, in two or three cases, sample sizes available were marginal (Sitta victoriae, 2 points; S. solan- gicie, 3 points; S. yunnanensis, 1 points). Some analyses suggest these sample sizes are prob- ably too low (Stockwell and Peterson 2()02b), but others indicate that such models may be viable (Peterson et al. 2006). These three spe- cies are genuinely microendemics and a few points are likely to be most of what is avail- able, particularly in the single mountain range endemic S. victoriae. Our analyses do not in- volve any projections to other time periods or other regions; we believe the effects of the potentially poor fit of these models are slight. Climate data (1960-1990) were drawn from the WorldClim climate data archive (Hijmans et al. 2005). We used a subset of the ‘biocli- matic’ coverages: annual mean temperature, mean diurnal temperature range, maximum temperature of warmest month, minimum temperature of coldest month, annual total precipitation, and precipitation of wettest and driest months. We supplemented these data sets with information from the U.S. Depart- ment of Interior, Geological Survey’s Hydro- IK data set (USDI 2001) for topography and landform (slope, aspect, compound topo- graphic index). We resampled all data sets to 0.17° resolution to avoid over interpretation of the precision of the point-occurrence data. Ecological Niche Modeling. — We used eco- logical niche modeling to provide a picture of likely distributional patterns for each species. This general class of procedures is based on known occurrences of species, as they relate to digital raster data coverages that summarize potentially relevant ecological parameters. The goal is to identify a suite of ecological conditions within which the species in ques- tion can likely maintain populations without immigration subsidy (Grinnell 1917). The re- sult is a picture of the species' potential geo- graphic distribution, defined as the area meet- ing the species’ ecological niche requirements 694 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 200H that characterize known distributional areas (Soberon and Peterson 2005). We used the Genetic Algorithm for Rule-set Prediction, or GARP (Stockwell and Peters 1999), which has seen extensive testing and application to such questions (Peterson and Cohoon 1999; Stockwell and Peterson 2003; Stockwell and Peterson 2002a, 2002b). GARP is an evolu- tionary-computing approach that relates known occurrences of species to raster data layers summarizing relevant environmental parameters to create a model of the ecological niche of the species, which can be used to identify a potential geographic distribution (Soberon and Peterson 2005). We used 19 bio- climatic variables from the WorldClim global 0.17° data set (Hijmans et al. 2005), plus data on topography and landform including eleva- tion, slope, aspect, and compound topographic index (USDI 2001) to characterize ecological landscapes. Our use of GARP is for visualization and interpolation purposes only, and the technique has been documented in detail elsewhere (An- derson et al. 2002, 2003; Illoldi et al. 2004; Martmez-Meyer et al. 2004; Ortega-Huerta and Peterson 2004; Soberon and Peterson 2004; Peterson 2005). Thus, we do not pro- vide a full, detailed description. We used half of the available occurrence data for training models, and half to provide test data sets that characterized model success in predicting in- dependent occurrence points. We developed 100 replicate models for each species using GARP and followed recent recommended pro- tocols (Anderson et al. 2003) in using inde- pendent measures of omission and commis- sion (Type I and Type II) prediction error to identify an optimal 10% of models from the 100 replicate models originally produced. The sum of these 10 models was taken as the best hypothesis of the species’ distribution. We inspected the distributional hypotheses for each species to establish a threshold for decisions of presence versus absence (lowest training presence threshold, Pearson et al. 2007). Initial model predictions in two cases {S. yunnanensis and S. nagaensis) were too broad and general and we used a tighter con- vergence criterion (convergence 0.001; maxi- mum iterations 10,000), which resulted in closer correspondence between model predic- tions and known distributional limits. No tests of model quality were developed given un- even sample sizes across species and known ability of Ecological Niche Models (ENMs) to reconstruct the generalities of species’ distri- butions at continental scales (Elith et al. 2006). We related model predictions for each spe- cies to the original ecological variables on which the models were based to reconstruct models in ecological space. We used the ‘Grid Combine’ option in ArcGIS 9.0 to create a grid with a distinct value for all unique com- binations of the environmental coverages across Eurasia. The attributes table associated with this grid yielded a matrix showing all unique environmental combinations and pre- dictions associated with each species, which permitted several visualization exercises. Niche breadth was reconstructed as the vari- ance of distributions in a standardized prin- cipal components analysis (Rotenberry and Wiens 1980, Carnes and Slade 1982, Litvak and Hansell 1990). Niche volume was calcu- lated as the spatial translation of the niche in terms of areal coverage of the distribution across the landscape of interest without atten- tion to which part is actually occupied by the species (Soberon 2007). We established which portions of the poten- tial distributional area are likely not to be in- habited to explore differences between actual and potential distributional areas (Svenning and Skov 2004). We were conservative and only eliminated as uninhabited those areas that were disjunct from areas of known oc- currence, and from which no occurrences were known. These steps resulted in maps of the likely actual area of occurrence (Soberon and Peterson 2005). Proportional range occu- pancy was calculated as the ratio of the areas covered by the actual and potential distribu- tional areas. We classed each species as oc- curring primarily in lowlands, foothills, or at high elevations based on published descrip- tions of nuthatch natural history (Matthysen 1998). RESULTS Niche Models. — These models produced re- alistic predicted distributions. Eor example, we used 17 available unique occurrence points for S. cashmirensis to characterize the species’ distribution (Fig. 1). Initial results from the Menon et al. • ASIAN NUTHATCHES 695 FIG. 1. Kashmir Nuthatch {Sitta cashmirensis) illustrating known occurrence points (dotted circles), crude potential geographic distribution (gray shading), and estimated actual distribution (black shading). ENM algorithm identified mainly highland ar- eas as potential distributional areas including the Himalayas of India, Nepal, and Pakistan (that constitute the known range of the spe- cies), as well as the highlands of Ethiopia and lower-elevation regions of Iran, Turkey, and Greece. A major disjunction in this distribu- tional prediction corresponds to the dispersal barrier likely constraining this species to its present distribution: the Central Persian Des- ert Basin in central and eastern Iran, and west- ern Afghanistan. Geographic patterns of species richness based on the modeled distributions of nut- hatch species were striking (Fig. 2). Spread generally across southern Asia, nuthatches show a dramatic richness focus in a narrow east- west band corresponding to the forested parts of the Himalayas, extending eastward into southwestern China (especially !S/.echuan and Yunnan provinces) and northern Burma. These foci reach predicted species richness up to nine species present (Fig. 2). However, the coarse resolution of our predictions (0.17°, or ~18.9 km) precludes separation of local spe- cies richness from high local-scale species turnover corresponding to local habitat diver- sity. These distributional predictions have a middle elevation focus (Fig. 3) with lowland areas and elevations above —3,000 m having relatively few species; intervening elevations (1,000-2,000 m), have the highest species di- versity. Niche Dimensions and Disequilibrium. — The relationship between niche breadth and extent was positive (Fig. 4) because wide niche breadths tend to map onto large geo- graphic areas, but accumulation of distribu- tional area (niche volume) decreases with ad- ditional environmental combinations (niche breadth). Proportional range-filling indices showed a negative relationship with montane habit (Fig. 5). An outlier in this relationship was the White-browed Nuthatch; regressions including this species explained 10% of the overall variation {R^ = 0.102; P < 0.05), whereas excluding it explained fourfold more variation {R^ = 0.422; P < 0.05). The prob- able relictual nature of this species needs fur- ther investigation, but montane characteristics clearly limit the distributional potential of nut- hatch species. 696 THE W ILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4. December 2008 FIG. 2. Overall panems of predicted species richness among Asian nuthatches (ramp from white = 0 species predicted present, darkest gray = 9 species predicted present). Sampling points (i.e.. occurrence records for the 15 species in the study) are shown as doned circles. DISCUSSION Niche brea(dth arud volume represent two measures of the ecological niche amplitude of a species, the former in terms of ecological combinations and the latter reflecting the spa- tial manifestation (Soberdn 2007). These quantities are only beginning to be character- ized and few studies have explored their re- lationships. The positive niche breadth and volume relationship observed in this study is of interest and should be explored in addition- al groups. Recent studies have explored the extent to which species fill their potential distributions, a measure of distributional disequilibrium (Svenning and Skov 2004). We generally found low proportional range filling among nuthatch species, particularly those with mon- tane distributions. This suggests that nuthatch distributions will frequently be constrained by limited dispersal abilities. Montane Ecology and Geography. — Nut- hatch diversity patterns are related to the ma- jor mountain systems and associated cool tem- :Le.3lar m FIG. 3. Relationship between number of nuthatch species predicted present and elevation across Asia. Menon et al. • ASIAN NUTHATCHES 697 FIG. 4. Relationship between niche volume and niche breadth. Niche breadth is measured as the average variance across the uncorrelated principal components, whereas niche volume is the geographic projection (in km-) of the ecological niche model. perate climates. Most nuthatch species have narrow geographic distributions in subtropical regions; those few with broader distributions range considerably farther and more broadly. The nuthatch diversity focus coincides closely with the Qinghai-Tibet Plateau region known to contain the largest concentration of Endem- ic Bird Areas in Asia (Long et al. 1996), as well as a focus of bird species richness (Ding et al. 2006). This association illustrates the seeming contradiction that most nuthatch spe- cies have small ranges, but the temperate-zone conditions they prefer are represented much more broadly farther to the north. These results raise issues of biogeography and the role of geography in subdividing geo- graphic distributions of evolving lineages. If species do not fill their ranges as completely in montane environments as in lowland envi- ronments, and if ecological niches are rela- tively conservative (Peterson et al. 1999), in- teractions may exist between geographic po- tential and ecological habit. Lowland species may have broader geographic distributions, but montane species may experience greater subdivision or have greater potential for iso- lation of populations that manage to colonize across dispersal barriers. Tests of these and FIG. 5. Relationship between proportion of potential distribution actually occupied and montane habit (0 = lowlands, 1 = foothills, 2 = high elevations). Shown are two simple linear regressions, one including (R- = 0.102) and the other excluding {R- = 0.422) Sitta victoriae (open box). 698 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 other hypotheses will be feasible once a robust phylogenetic hypothesis is available for the group. CONSERVATION IMPLICATIONS One fifth of all Asian nuthatch species are considered threatened (Collar et al. 1994, lUCN 2006). Among threatened Asian spe- cies, S. victoricie has an extremely narrow range, is listed as Endangered (lUCN 2006), and would clearly be threatened by any sig- nificant habitat destruction in the Mount Vic- toria region. Eour other species {S. solangiae, S. yunnanensis, S. formosa, S. magna) have wider ranges, but are threatened by habitat loss and degradation (Matthysen 1998, BirdLife International 2004, lUCN 2006). Sit- ta formosa and S. magna have small popula- tions, which are thought to be declining and severely fragmented, and are listed as Vulner- able. Sitta solangiae and S. yunnanensis are affected by ongoing habitat loss and degra- dation, but have larger populations and are listed as near Threatened (lUCN 2006). The area of highest nuthatch species diversity, pre- dicted by our analysis, closely matches the Sino-Himalayan Mountain Eorest region iden- tified by BirdLife International (2003) as one of nine key forest regions for threatened birds in Asia. The Chinese and Himalayan region consists of middle- and high-elevation forests, scrub, and grasslands on the southern slopes of the Himalayas and in the mountains of south- western China and northern Indochina. The natural habitat in this region is relatively se- cure at higher altitudes compared to middle elevations, which is under greater pressure from deforestation and fragmentation. Middle- elevation forests are also the areas where our analysis predicts highest nuthatch species di- versity. Habitat fragmentation further com- pounds the conservation implications for montane species, which have a greater poten- tial for isolation. Our results emphasize the importance of middle-elevation habitats in this region for biodiversity. ACKNOWLEDGMENTS This research was supported by an NSF-Research Opportunity Award. We gratefully acknowledge Asad Rahmani (Director, Bombay Natural History Society) for support, Monica Pape§ for technical assistance, and Jill Witt for introducing the Beautiful Nuthatch {Sitta formosa) to the Menon Laboratory. LITERATURE CITED Anderson, R. R, D. Lew, and A. T. Peterson. 2003. Evaluating predictive models of species’ distri- butions: criteria for selecting optimal models. Ecological Modelling 162:211-232. Anderson, R. R, A. T. Peterson, and M. G6mez-La- VERDE. 2002. Using niche-based GIS modeling to test geographic predictions of competitive exclu- sion and competitive release in South American pocket mice. Oikos 93:3-16. Birdlife International. 2003. Saving Asia’s threat- ened birds: a guide for government and civil so- ciety. Birdlife International, Cambridge, United Kingdom. Birdlife International. 2004. Threatened birds of the world 2004. CD-ROM. BirdLife International, Cambridge, United Kingdom. Carnes, B. and N. Slade. 1982. Some comments on niche in canonical space. Ecology 63:888-893. Chapman, A. D. and J. Wieczorek (Editors). 2006. Guide to best practices for georeferencing. Global Biodiversity Information Facility, Copenhagen, Denmark, www.gbif.org/prog/digit/data_quality/ BioGeomancerGuide.pdf (accessed 10 August 2007). Collar, N. J., M. J. Crosby, and A. J. Stattersfield. 1994. Birds to watch, 2. The world list of threat- ened birds. BirdLife Conservation Series Number 4. BirdLife International, Cambridge, United Kingdom. Collar, N. J., A. V. Andreev, S. Chan, M. J. Crosby, 5. SuBRAMANYA, AND J. A. ToBiAS. 2001. Threat- ened birds of Asia: the BirdLife International Red Data Book. BirdLife International, Cambridge, United Kingdom. Ding, T, H. Yuan, S. Geng, C. Koh, and P. Lee. 2006. Macro-scale bird species richness patterns of the East Asian mainland and islands: energy, area and isolation. Journal of Biogeography 33:683-693. Elith, j., C. H. Graham, R. P. Anderson, M. Dudik, S. Perrier, A. Guisan, R. J. Humans, E Huett- MAN, J. R. Leathwick, A. Lehmann, J. Li, L. G. Lohmann, B. a. Loiselle, G. Manion, C. Mo- ritz, M. Nakamura, Y. Nakazawa, J. M. Over- ton, A. T. Peterson, S. J. Phillips, K. Richard- son, R. Scachetti-Pereira, R. E. Schapire, J. SoBERON, S. E. Williams, M. S. Wisz, and N. E. ZiMMERMANN. 2006. Novel methods improve pre- diction of species’ distributions from occurrence data. Ecography 29:129-151. Graham, C. H., S. Ferrier, E Huettman, C. Moritz, AND A. T. Peterson. 2004. New developments in museum-based informatics and applications in biodiversity analysis. Trends in Ecology and Evo- lution 19:497-503. Grinnell, j. 1917. Field tests of theories concerning distributional control. American Naturalist 51: 1 15-128. Menon et al. • ASIAN NUTHATCHES 699 Harrap, S. and D. Quinn. 1995. Chickadees, tits, nut- hatches and treecreepers. Princeton University Press, Princeton, New Jersey, USA. Humans, R. J., S. Cameron, and J. Parra. 2005. WorldClim. Version 1.3. University of Califor- nia, Berkeley, USA. http://biogeo.berkeley.edu/ worldclim/worldclim.htm (accessed 10 August 2007). Illoldi, P, V. Sanchez-Cordero, and A. T. Peterson. 2004. Predicting distributions of Mexican mam- mals using ecological niche modeling. Journal of Mammalogy 85:658-662. lUCN 2006. 2006 lUCN Red List of threatened spe- cies. http://www.iucnredlist.org (accessed 10 Au- gust 2007). Lack, D. 1971. Ecological isolation in birds. Harvard University Press, Cambridge, Massachusetts, USA. Litvak, M. and R. I. Hansell. 1990. A community perspective on the multidimensional niche. Jour- nal of Animal Ecology 59:931-940. Long, A. J., M. J. Crosby, and A. J. Satterseield. 1996. Towards a global map of biodiversity: pat- terns in the distribution of restricted-range birds. Global Ecology and Biogeography Letters 5:281- 304. Marti'nez-Meyer, E., A. T. Peterson, and W. W. Har- grove. 2004. Ecological niches as stable distri- butional constraints on mammal species, with im- plications for Pleistocene extinctions and climate change projections for biodiversity. Global Ecol- ogy and Biogeography 13:305-314. Matthysen, E. 1998. The nuthatches. T. and A. D. Poyser, London, United Kingdom. Mayr, E. and D. Amadon. 1951. A classification of recent birds. American Museum Novitates 1946: 453-473. National Geospatial Intelligence Agency. 2007. GEOnet Names Server (GNS). http://earth- info.nga.mil/gns/html/index.html (accessed 10 August 2007). Ortega-Huerta, M. a. and A. T. Peterson. 2004. Modeling spatial patterns of biodiversity in north- eastern Mexico. Diversity and Distributions 10: 39-54. Pearson, R. G., C. Raxworthy, M. Nakamura, and A. T. Peter.son. 2007. Predicting species’ distri- butions from small numbers of occurrence rec- ords: a test ca.se using cryptic geckos in Mada- gascar. Journal of Biogeography 34:102-1 17. Peterson, A. T. 2003. Predicting the geography of spe- cies’ invasions via ecological niche modeling. Quarterly Review of Biology 78:419-433. Peter.son, A. T. 2005. Kansas Gap Analysis: the im- portance of validating distributional models before using them. Southwestern Naturalist 50:230-236. PiUERSON, A. T. AND K. C. CoHOON. 1999. Sensitivity of distributional prediction algorithms to geo- graphic data completeness. Ecological Modelling 117:159-164. Peterson, A. T. and R. D. Holt. 2003. Niche differ- entiation in Mexican birds: using point occurrenc- es to detect ecological innovation. Ecology Letters 6:774-782. Peterson, A. T, J. Soberon, and V. Sanchez-Cor- dero. 1999. Conservatism of ecological niches in evolutionary time. Science 285:1265-1267. Peterson, A. T, R. R. Lash, D. S. Carroll, and K. M. Johnson. 2006. Geographic potential for out- breaks of Marburg hemorrhagic fever. American Journal of Tropical Medicine and Hygiene 75:9- 15. Ripley, S. D. 1959. Character displacement in Indian nuthatches (Sitta). Postilla 42:1-1 1. Rotenberry, j. T. and J. A. Wiens. 1980. Habitat structure, patchiness, and avian communities in North American steppe vegetation: a multivariate analysis. Ecology 61:1228-1250. Sibley, C. G. and J. E. Ahlquist. 1990. Phylogeny and classification of birds: a study in molecular evolution. Yale University Press, New Haven, Connecticut, USA. Soberon, J. 2007. Grinnellian and Eltonian niches and geographic distributions of species. Ecology Let- ters 10:1115-1123. Soberon, J. and A. T. Peterson. 2004. Biodiversity informatics: managing and applying primary bio- diversity data. Philosophical Transactions of the Royal Society of London, Series B 359:689-698. Soberon, J. and A. T. Peterson. 2005. Interpretation of models of fundamental ecological niches and species’ distributional areas. Biodiversity Infor- matics 2: 1-10. Stockwell, D. R. B. and D. P. Peters. 1999. The GARP modeling system: problems and solutions to automated spatial prediction. International Jour- nal of Geographical Information Science 13:143- 158. Stockwell, D. R. B. and A. T. Peterson. 2()02a. Con- trolling bias in biodiversity data. Pages 537-546 in Predicting species occurrences: issues of scale and accuracy (J. M. Scott, P. J. Heglund. and M. L. Morrison, Editors). Island Press, Washington, D.C., USA. Stockwell, D. R. B. and A. T. Peter.son. 2()()2b. P2f- fects of sample size on accuracy of species distri- bution models. Ecological Modelling 148:1-13. Stockwi:ll, D. R. B. and A. T. Peti;rson. 2003. Com- parison of resolution of methods used in mapping biodiversity patterns from point occurrence data. Ecological Indicators 3:213-221. SviiNNiNG, J.-C. AND E Skov. 2004. Limited filling of the potential range in European tree species. Ideol- ogy Letters 7:565-573. U.S. Di;rar imi;n I oi lNri;RioR (USDl). 2001. Hydro- IK elevation derivative data base. USDl, Geolog- ical Survey. Washington. D.C'.. USA. http:// edcdaac.usgs.gov/gtopo30/hydro/ (accessed 1 0 August 2007). The Wilson Journal of Ornithology 1 20(4):700-707, 2008 DENSITY AND ABUNDANCE OE MOUNTAIN PLOVERS IN NORTHEASTERN MONTANA THERESA M. CHILDERS' ^ AND STEPHEN J. DINSMORE' ^ ABSTRACT. — Estimates of local abundance for declining species provide important information necessary for conservation measures. We estimated the density and abundance of Mountain Plover {Charadriiis montamis) in Phillips and Valley counties in north-central Montana in 2004 using distance sampling methodology. Sampling efforts were stratified to include active prairie dog (Cynomys sp.) colonies, an Area of Critical Environmental Concern (ACEC) specifically established for Mountain Plover, and all other habitats. The density of plovers was greatest on prairie dog colonies (7.20 ± 0.42 [SE] plovers/km-) and much lower on both the ACEC (1.60 ± 0.31 plovers/km-), and all other habitats (0.07 ± 0.01 plovers/km-). An estimated 1,028 (95% Cl = 903-1,153) plovers inhabited this region in 2004, most (74%) on prairie dog colonies. Our results highlight the importance of prairie dog colonies to plovers in this region and suggest that as much as 10% of their continental population may breed in north-central Montana. Received 10 September 2007. Accepted 14 February 2008. The Mountain Plover {Charadriiis montan- iis) is a declining species endemic to the west- ern Great Plains and Colorado Plateau (Knopf 1994, Knopf and Wunder 2006). Historically, its breeding range extended from Canada south along the eastern edge of the Rocky Mountains to New Mexico and east to eastern North Dakota and south to western Texas, in- cluding a vast region of short grass prairie, denuded plains, and semi-desert areas (Bent 1929, Knopf and Wunder 2006). Breeding Bird Surveys (BBS) indicate Mountain Plover declined in the 1966-1993 period (Knopf 1996, Knopf and Wunder 2006), although there are no other continent-wide monitoring data for comparison. The Mountain Plover was proposed in 1999 for federal listing as threatened due to concerns over population decline as a result of continued critical habitat loss; listing was denied in 2003 (USDI 2003). However, Mountain Plovers are still consid- ered a species of special concern throughout much of their breeding range (USDA 1994, USDI 2000, Brown et al. 2001) and are one of a suite of the Great Plains ecosystem in- dicator species that include black-footed ferret (Mustela nigripes) and Burrowing Owl {Athe- ne ciiniciilaria). Concerns with continental declines of Mountain Plovers have focused monitoring ef- ‘ Department of Natural Resource Ecology and Management, 339 Science II, Iowa State University, Ames, lA 50011, USA. -Current address: 24830 West Highway 50, #EC1- E, Gunnison, CO 81230, USA. Corresponding author; e-mail: cootjr@iastate.edu forts on estimating their abundance at breed- ing (Dinsmore et al. 2003, Wunder et al. 2003, Dreitz et al. 2006) and wintering (Knopf and Wunder 2006) sites. The current continental Mountain Plover population estimate is 11,000-14,000 birds (Plumb et al. 2005). Es- timates of local abundance are available for key breeding sites and are usually extrapolat- ed from density estimates. Adult density was 2.0 ± 0.46 (SE) plovers/km^ at the Pawnee National Grasslands, Colorado from 1990 to 1994 (Knopf and Wunder 2006) and an esti- mated 4,850 adult Mountain Plovers occur in eastern Colorado east of the Front Range (USDI 2003). Plover density on select prairie dog colonies in Phillips County, Montana ranged from 6.80 ± 1.61 (SE) plovers/km^ in 1991 to 1.28 ± 0.06 (SE) plovers/km^ in 1995 with an estimated 175 breeding adult Moun- tain Plovers (Dinsmore 2001, Dinsmore et al. 2003). The density of adult plovers in South Park, Colorado was 7.90 ± 0.90 (SE) plovers/ km^ and an estimated 2,310 breeding adult plovers (Wunder et al. 2003). Plumb et al. (2005) estimated there were 4.47 ± 0.55 (SE) plovers/km“ in Wyoming and a statewide adult population of 3,393 plovers. Populations in Colorado, Wyoming, and Montana combined comprise the majority of all known breeding Mountain Plovers (Knopf and Miller 1994, USDI 2003). Estimating and monitoring local abundance of Mountain Plovers throughout their range is important because it identifies concentrations of plovers, helps focus conservation efforts, and aids land management planning by natural 700 Childers and Dinsmore • MOUNTAIN PLOVER ABUNDANCE IN MONTANA 701 resource agencies. Our objectives were to: (1) estimate Mountain Plover density and abun- dance in 2004 in three habitat strata in south- ern Phillips and Valley counties, Montana, and (2) suggest how this information can aid man- agement and conservation planning activities to benefit Mountain Plovers. METHODS General Study Area. — We studied Mountain Plovers during the 2003 and 2004 breeding seasons in a 7,162-km2 area in Phillips and Valley counties, Montana (Fig. 1). We used the 2003 breeding season to calculate sam- pling effort from a pilot study and identify the sampling frame, and then conducted surveys during the 2004 breeding season. Mountain Plovers in Montana primarily select active black-tailed prairie dog (Cynomys ludovici- anus) colonies for nesting (Knowles et al. 1982, Olson and Edge 1985, Dinsmore 2001). Prairie dog colonies are one of the few re- maining suitable habitat types for Mountain Plovers in Montana, and have experienced de- clines throughout the last century due to large- scale poisoning and sylvatic plague (Olson and Edge 1985, Knowles 1999). Other habi- tats inhabited by breeding plovers in Montana include areas heavily grazed by domestic sheep in central Montana, and hardpan drain- ages and a former bentonite mining area in Valley County (Prellwitz 1993, Knowles and Knowles 1998). Plovers are either absent or occur in low densities in Montana outside of these habitats. Our study area consisted pri- marily of federal lands managed by the Bu- reau of Land Management (BLM, Glasgow Field Station and Malta Field Office) and the U.S. Fish and Wildlife Service (USFWS, Charles M. Russell National Wildlife Refuge). We divided the study area into three strata to facilitate surveying plovers. The strata were: (1) a Mountain Plover Area of Critical Environmental Concern (ACEC) in Valley County, (2) all active black-tailed prairie dog colonies in Phillips County (there was none in Valley County), and (3) all other habitats sur- rounding these strata south of U.S. Highway 2, east of Montana Highway 191, north of the Missouri River, and west of Montana High- way 24 (Fig. 1). Prior research suggested these areas were occupied by plovers at dif- fering densities and stratification was neces- sary to generate valid estimates of plover abundance. Mountain Plover ACEC. — The BLM (Glas- gow Field Station) established an Area of Critical Environmental Concern in the Little Beaver Creek drainage of Valley County to protect plover breeding habitat and considers the plover a species of special concern (USDI 2000). The ACEC stratum consisted of 10,007 ha delineated by existing roads and property lines, and managed by the BLM to protect Mountain Plover breeding habitat (USDI 2000). The ACEC was comprised of two pri- mary habitats: sparsely vegetated hardpan clay and bentonite soils in drainage bottoms, and densely vegetated gentle rises on either side of the drainages. Mountain Plovers in- habit hardpan soil valley bottoms within the ACEC, where dominant vegetation includes Nuttall’s saltbush {Atriplex nuttallii), Sand- berg bluegrass {Poa secunda), western wheat- grass {Pascopyrum smithii), plains prickly pear {Opuntia polycantha), wild onion {Allium spp.), and wild parsley (Lomatium foenicula- ceum) (USDI 2000). Plovers also use bentonic soils dominated by knotweed {Polygonum spp.), Sandberg bluegrass, blue grama {Bou- teloua gracilis), and western wheatgrass. Gen- tle rises on either side of valley bottoms are dominated by wild buckwheat {Polygonum convolvulus), horizontal juniper {Juniperus communis), basin big sagebrush {Artemisia tridentata tridentata), and western wheatgrass and are generally not used by Mountain Plo- vers. Prairie Dog Colonies. — Breeding Mountain Plovers in north-central Montana are known to selectively inhabit prairie dog colonies and are thought to be restricted to these sites in Phillips County (Knowles et al. 1982, Olson 1984, Dinsmore 2()()0). Our prairie dog colony stratum consisted of 334 active black-tailed prairie dog colonies ranging from < 1 to 308 ha in size comprising a total of 10,515 ha in 2002 (the most recent year colonies were sur- veyed; J. J. Grensten, pers. comm.). Dominant vegetation in this stratum included fringed sagewort {Artemisia frigida), biiffalograss {Buchloe dactyloides). club moss {Selaginella densa), plains prickly pear, blue grama, nee- dle-and-thread grass {Stipa comata), and Sandberg bluegrass. Other Ihdyitats. — All habitat types sur- 702 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 Other Habitats Stratum Phillips County Prairie Dog Colonies Valley County Glasgow Mountain Plover ACEC LIG. 1. Phillips and Valley counties. Montana showing strata where breeding Mountain Plovers were sur- veyed during the 2004 nesting season. The three strata consisted of the Mountain Plover Area of Critical Environmental Concern (ACEC) in Valley County (dark gray stippling), active black-tailed prairie dog colonies in Phillips County (black areas), and all other habitats surrounding the prairie dog colony and ACEC strata (light gray stippling). rounding the prairie dog colony and ACEC strata in southern Phillips and Valley counties were included in the ‘other habitats stratum’ (716.241 ha). We sampled all habitats in this stratum and made no attempt to partition the sample based on suitable habitat because: (1) we did not have access to CIS information with sufficient detail to identify suitable plo- ver habitat, and (2) logistical constraints pre- vented us from mapping suitable habitat in the Childers and Dinsmore • MOUNTAIN PLOVER ABUNDANCE IN MONTANA 703 field during this study. Our purpose for sam- pling this region was to provide baseline in- formation about Mountain Plover density out- side the other two strata. The other habitats strata varied in topography from the rolling Larb Hills to relatively flat open rangeland. Vegetation was generally taller and denser compared to that of the other two strata, but comprised of the same species. Small areas of habitat similar to those in the ACEC were scattered throughout both counties. Statistical Analyses. — We used radial dis- tance point count surveys (Buckland et al. 2001) to sample Mountain Plovers in each of the three strata. We used data from pilot stud- ies to calculate a suitable sample of points given a 15% desired coefficient of variation (CV). Distance sampling encounter rates from the 1991-2000 breeding seasons (SJD, un- publ. data) for the prairie dog colony stratum were used to calculate a suitable survey sam- ple {n = 105 points). We conducted randomly located point count surveys in early May 2004 in the ACEC stratum and used the resulting encounter rates to delimit the survey sample for the ACEC (/? = 110 points). The number of sample points in the other habitats strata was chosen to be logistically feasible and to provide baseline evidence re- garding plover density. We knew this stratum supported a low density of Mountain Plovers, based on personal observation and consulta- tion with regional biologists, and concluded that an excessive number of sample points was needed to generate a precise abundance estimate. Generating a plover abundance es- timate with reliable precision was not possible in this stratum given time, personnel, and lo- gistical constraints. However, we believed it was necessary to provide some evidence re- garding Mountain Plover density in the other habitats stratum. A generally recommended sample size for areas with relatively good probability of detection of animals is 40 (Buckland et al. 2001). We believed probabil- ity of detecting plovers was low to moderate in this stratum due to shrub cover and rolling topography. We chose to survey slightly more (/? = 50 points) than the recommended num- ber of sample points to provide evidence, rath- er than a precise estimate, of Mountain Plover density for this stratum. We overlaid a 500 X 500 m grid with a random starting point on the ACEC and other habitats strata, and randomly selected grid center point coordinates for survey points. We chose survey points differently for the prairie dog stratum because of how colonies were distributed on the landscape. The borders of all colonies were mapped in 2002 with GPS units by the BLM (Malta Eield Office). We selected colonies at random with replacement, assigned survey points proportional to colony size, and spaced points equidistant within a colony if it included >1 point. One observer (TMC) visited all points in the ACEC (mid-May through mid-Jun) and other habitats strata (mid-Jun through early Jul) while a second observer (SJD) visited all points on prairie dog colonies (mid-May through early Jun) to minimize observer bias. Timing of all surveys coincided with the nest- ing season in Montana (Dinsmore et al. 2002) but we visited the ACEC and other habitats strata at slightly different seasons for logisti- cal reasons. Individual points were located on the ground using a Garmin Explorer V GPS unit and were approached by vehicle. Plovers actively avoid a person on foot but appear to ignore vehicles (pers. obs.). We minimized vi- olation of the assumption that there was no avoidance behavior by approaching survey points in a vehicle. All surveys were con- ducted during daylight hours (0500 to 1200 hrs MDT) during standardized weather con- ditions (0-24 km/hr wind speed, no precipi- tation. and temperature <27° C). Surveys were conducted in the early portion of the breeding season to ensure that only breeding adults were included, and there were no prob- lems with post-breeding dispersal. We visited each point for 5 min. surveyed for plovers us- ing 10 X 40 binoculars, and measured the dis- tance (m) to each Mountain Plover using a NewCon Optic LRM 1500 laser rangefinder. All measurements were exact and we used ac- tual distances rather than placing observations into distance groups. Observations were treat- ed as statistically independent events, which is reasonable given that 43% of sightings were of a single bird. We used Program DISTANCE (Version 3.5: Thomas et al. 1998) to model detection rates of plovers and calculate stratum-specific den- sity estimates. We considered the four robust models best suited for detection functions sug- 704 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 TABLE 1. Model selection results from Program DISTANCE, and density (D; birds/km^) and abundance (yV) estimates for Mountain Plovers on the Area of Critical Environmental Concern (ACEC) and active black- tailed prairie dog colonies in north-central Montana, 2004. Model, expansion AIC AAIC D N cv ACEC Uniform, cosine® 658.76 0.00 1.60 160 24% Half-normal, hermite 666.74 7.98 Uniform, simple polynomial Hazard rate, cosine 751.16 777.27 92.40 118.51 Prairie dog colonies Hazard rate, cosine’’ 3792.75 0.00 7.20 758 6% Half-normal, hermite 3796.22 3.47 Uniform, simple polynomial Uniform, cosine 3870.75 17,453.01 78.00 13,660.26 ^ Four cosine expansion terms. Zero cosine expansion terms. gested by Buckland et al. (2001:155): (1) uni- form key function with a cosine expansion, (2) uniform key function with a simple poly- nomial expansion, (3) half-normal key func- tion with a hermite expansion, and (4) hazard rate key function with a cosine expansion. All models exhibit properties that meet the dis- tance sampling assumption that probability of detection of an animal declines with increas- ing distance from the point. Model fit was as- sessed using the Chi-square goodness-of-fit test in Program DISTANCE. We used AlC model selection (Burnham and Anderson 2002) to select the best approximating model for each stratum. There were insufficient plover detections ( 1 detection in 50 point surveys) in the other habitats stratum, and density and abundance estimates were calculated using the ACEC de- tection function. We chose this approach for three reasons: (1) we wanted to generate an estimate of abundance for this stratum for planning purposes and were unable to entirely ignore the small sample size; (2) both the ACEC and other habitats strata surveys were conducted with the same observer, and pool- ing these data into a single detection function seemed reasonable; and (3) portions of these two regions had similar habitat types, making them more similar than the other habitats stra- tum was to the prairie dog colony stratum. Abundance. — Plover density estimates in the prairie dog colony and ACEC strata were multiplied by their known areas to calculate stratum-specific estimates of Mountain Plover abundance. We estimated plover abundance differently for the other habitats stratum. Rather than estimate abundance for all habi- tats in this stratum, we chose to focus only on suitable habitat. We used previously classified (shrub, non-shrub, bare ground, and water) Landsat 7 (14 May 2003, path 37, rows 26 and 27) satellite imagery to identify potential Mountain Plover habitat. We considered po- tential Mountain Plover habitat at this scale to consist of either non-shrub or bare ground patches. Olson-Edge and Edge (1987) ob- served that Mountain Plovers in Montana se- lect prairie dog colonies >6 ha in size. Thus, we restricted potential habitat by selecting patches >6 ha (area = 157,573 ha). The veg- etation structure of non-shrub patches could not be calculated from satellite images, and not all of the area identified actually contained suitable plover nesting habitat. We estimated that 22% (157,573 ha) of this stratum con- tained suitable Mountain Plover habitat based on these criteria. RESULTS Model Selection and Detection Function. — The best model selected for the ACEC was a uniform key function with a cosine expansion and for the prairie dog colony stratum it was the hazard rate key function with a cosine ex- pansion (Table 1); no other models were con- sidered competing (AAIC < 2). Both of these models fit the data well {P > 0.15). We were unable to develop a detection function for the Childers and Dinsmore • MOUNTAIN PLOVER ABUNDANCE IN MONTANA 705 Other habitats stratum due to the low detection rate (1 detection in 50 points). Density and Abundance Estimates. — The density of Mountain Plovers was greater in the prairie dog colony stratum (D = 7.20 plovers/ km2, 95% Cl = 0.064-0.082, 6.41% CV) than in the ACEC stratum 0 == 1.60 plovers/km^ 95% Cl = 0.010-0.025, 24.03% CV). The majority (74%) of the Mountain Plovers with- in the study area occurred in the prairie dog colony stratum {N — 758, 95% Cl = 668-860) with smaller numbers in the ACEC (TV = 160, 95% Cl = 100-285). The density of Mountain Plovers in the other habitats stratum was two orders of magnitude less than the density in the other two strata 0 = 0.07 plovers/km^, 95% Cl = 0.0005-0.0010, 16.54% CV). An optimistic estimate of Mountain Plover abun- dance in this stratum was relatively low for such a large area {N = 110, 95% Cl = 78- 154). DISCUSSION We estimated Mountain Plover density in three habitats in north-central Montana. The density estimate for the prairie dog colony stratum had good precision (6.41% CV) due to robustness of the data used to generate the detection function. The precision of the ACEC density estimate was less than expected (24% CV vs. a desired 15% CV) and may have re- sulted because suitable habitat occurred in small and unevenly distributed patches. The density of Mountain Plovers in the other hab- itats stratum was thought to be low prior to this survey, and an infeasible number of point counts would have been necessary to produce an estimate with good precision. Our result for this stratum should be interpreted with cau- tion. Sampling issues caused us to make key as- sumptions about plover detection rates and suitable plover habitat in the other habitats strata. The ACEC and other habitats strata shared some habitat characteristics, e.g., densely vegetated low ridges interspersed with irregular patches of suitable Mountain Plover habitat. Thus, the calculated density for the other habitats stratum appeared to realistically represent Mountain Plover density in regions similar to the ACEC. The scale and frequency of suitable habitat patches varied between strata, and we corrected for an inflated abun- dance estimate in the other habitats stratum by constricting the inferential space to only the area of potential plover habitat. It is unlikely that all identified potential plover habitat (22%, or 157,573 ha) is actually suitable for nesting. There are few rigorous estimates of Moun- tain Plover abundance in Montana. A specific sampling protocol has been used on a portion of prairie dog colonies in southern Phillips County (Dinsmore 2001), but our study pre- sents the first formal sample of Mountain Plo- ver density and abundance throughout a sig- nificant proportion of its northeastern Mon- tana breeding range. Mountain Plover density in the prairie dog stratum was much greater than densities reported for grasslands in Col- orado (2.0 ± 0.46 birds/km^ to 4.7 ± 1.20 birds/km^; Knopf and Wunder 2006), and for grasslands (5.17 ± 1.06 birds/km^) and shrub- steppe habitats (4.23 ± 0.67 birds/km^) in Wyoming (Plumb et al. 2005). Our density es- timates were similar to those reported from Phillips County in the early 1990s (SJD, pers. obs.), despite an outbreak of sylvatic plague in the mid-1990s that decimated prairie dogs (and affected plover habitat) throughout this region (Dinsmore et al. 2005). The majority of Mountain Plovers currently breeding in southern Phillips and Valley counties (74%) inhabit active black-tailed prairie dog colo- nies. The strong link between plovers and prairie dogs in Montana (Dinsmore et al. 2005), which is not as prevalent elsewhere in the plover’s range suggests that size and mo- saic of suitable plover habitat patches may be more dynamic in short-grass prairie habitats at other sites. Other forms of disturbance such as fire and grazing by ungulates creates suit- able plover nesting habitat in areas where prairie dogs do not constantly maintain low vegetation. Our investigation of Mountain Plover abun- dance outside prairie dog colonies and ACEC strata partially confirms biologists’ contention that the other habitats stratum is mostly un- inhabited by plovers. An optimistic estimate of plover abundance in all potential nesting habitat within the other habitats stratum is 110 plovers, which may comprise up to 10% of the breeding population in southern Phillips and Valley counties. Our study revealed that plovers breeding 706 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 near the northern limit of their range in Mon- tana constitute a significant proportion of this species’ overall population and should remain a conservation priority. The total population estimate of Mountain Plovers in southern Phillips and Valley counties was 1,028 adults (95% Cl = 903-1,153, 6.18% CV), or less fhan 10% of the estimated 11-14,000 Moun- tain Plovers in North America (USDI 2003, Plumb et al. 2005). Most inhabit active prairie dog colonies in Phillips County. With removal of large-scale grazers and increased fire con- trol, sites maintained by prairie dog activities and unproductive soils have become the only suitable areas available to breeding Mountain Plovers in much of Montana. CONSERVATION IMPLICATIONS Our study provides critical baseline infor- mation about plover density and occurrence in Phillips and Valley counties in north-central Montana, but more investigation of plovers in this region is needed. Patches of breeding hab- itat outside of the prairie dog colony and ACEC strata should be identified and sampled more intensively to gain a more precise esti- mate of plover abundance. Our investigation of the other habitats stratum suggests it may support up to 10% of the Mountain Plover population in Phillips and Valley counties. Much of the suitable plover habitat in this stratum appears to be near the ACEC, sug- gesting that Mountain Plovers may benefit from an expansion of the ACEC. Mountain Plover abundance within the prairie dog col- ony and ACEC strata should be monitored at regular intervals. Plovers in both of these re- gions represent important contributions to the continental population, are considered indi- cators of the health of their respective habitats (Dinsmore 2000), and should continue to be the focus of local monitoring efforts and man- agement. Regional population estimates of breeding Mountain Plovers contribute useful informa- tion for conservation planning at both the re- gional and continental levels. Recently, re- searchers have estimated abundance of Moun- tain Plovers in South Park, Colorado (Wunder et al. 2003), eastern Colorado (Tipton 2007), and in Wyoming (Plumb et al. 2005). These estimates, combined with information from other breeding populations, suggest the con- tinental population of Mountain Plovers is greater than previously believed (Plumb et al. 2005, Knopf and Wunder 2006). The previous continental population estimate based on sam- ples of wintering plovers was 5,000-10,000 individuals; Plumb et al. (2005) suggests a re- vised population estimate of 11,000-14,000 individuals. The ACEC and active prairie dog colonies provide nesting habitat for the majority of Mountain Plovers in north-central Montana. Local managers should continue to ( 1 ) protect the ACEC and active prairie dog colonies on public lands from major human disturbances, such as mining, during the plover breeding season, and (2) monitor the size and health of suitable plover habitat in each region (hardpan flats in Valley County and active prairie dog colonies in Phillips County). ACKNOWLEDGMENTS We thank the U.S Bureau of Land Management (Glasgow Lield Office) for funding this study. David Waller, John Carlson, John Lahlgren, and many other staff from the Glasgow Lield Office provided field as- sistance. J. J. Grensten and C. T. Wilcox helped with fieldwork in Phillips County and M. D. Smith provided assistance with GIS work. LITERATURE CITED Bent, A. C. 1929. Life histories of North American shorebirds. U.S. National Museum Bulletin 146: 263-269. Brown, S., C. Hickey, B. Harrington, and R. Gill (Editors). 2001. The U.S. Shorebird Conservation Plan. Second Edition. Manomet Center for Con- servation Sciences, Manomet, Massachusetts, USA. Buckland, S. T, D. R. Anderson, K. P. Burnham, J. L. Laake, D. L. Borchers, and L. Thomas. 2001. Introduction to distance sampling: estimating abundance of biological populations. Oxford Uni- versity Press, Oxford, United Kingdom. Burnham, K. P. and D. R. Anderson. 2002. Model selection and multimodel inference: a practical in- formation-theoretic approach. Second Edition. Springer- Verlag, New York, USA. Dinsmore, S. J. 2000. Mountain Plover. Pages 213- 218 in Endangered animals: a reference guide to conflicting issues (R. P. Reading and B. Miller, Editors). Greenwood Press, Westport, Connecti- cut, USA. Dinsmore, S. J. 2001. Population biology of Mountain Plovers in southern Phillips County, Montana. Dissertation. Colorado State University, Fort Col- lins, USA. Dinsmore, S. J., G. C. White, and E L. Knopf. 2002. Childers and Dinsmore • MOUNTAIN PLOVER ABUNDANCE IN MONTANA 707 Advanced techniques for modeling avian nest sur- vival. Ecology 83:3476-3488. Dinsmore, S. J., G. C. White, and E L. Knopf. 2003. Annual survival and population estimates of Mountain Plovers in southern Phillips County, Montana. Ecological Applications 13:1013-1026. Dinsmore, S. J., G. C. White, and E E. Knopf. 2005. Mountain Plover population responses to black- tailed prairie dogs in Montana. Journal of Wildlife Management 69:1546-1553. Dreitz, V. J., P. M. Lukacs, and E L. Knopf. 2006. Monitoring low density avian populations: an ex- ample using Mountain Plovers. Condor 108:700- 706. Knopf, E L. 1994. Avian assemblages on altered grass- lands. Studies in Avian Biology 15:247-257. Knopf, E L. 1996. Prairie legacies — birds. Pages 135- 148 in Prairie conservation: preserving North America’s most endangered ecosystem (E B. Sam- son and E L. Knopf, Editors). Island Press, Co- velo, California, USA. Knopf, E L. and B. J. Miller. 1994. Charadrius mon- tanus — montane, grassland, or bare-ground plo- ver? Auk 1 1 1 :504-506. Knopf, E L. and M. B. Wunder. 2006. Mountain Plo- ver {Charadrius montaniis). The birds of North America. Number 211. Knowles, C. J. 1999. Selective use of black-tailed prairie dog colonies by Mountain Plovers — a sec- ond look. USDI, Bureau of Land Management, Malta Field Office, Malta, Montana, USA. Knowles, C. J. and P. R. Knowles. 1998. The historic and current status of the Mountain Plover in Mon- tana. USDI, Bureau of Land Management, Malta Field Office, Malta, Montana, USA. Knowles, C. J., S. P. Stoner, and S. P. Gieb. 1982. Selective use of black-tailed prairie dog towns by Mountain Plovers. Condor 84:71-74. Olson, S. L. 1984. Density and distribution, nest site selection, and activity of the Mountain Plover on the Charles M. Russell National Wildlife Refuge. Thesis. Montana State University, Bozeman, USA. Olson, S. L. and D. Edge. 1985. Nest site selection by Mountain Plovers in north central Montana. Journal of Range Management 38:280-282. Olson-Edge, S. L. and W. D. Edge. 1987. Density and distribution of the Mountain Plover on the Charles M. Russell National Wildlife Refuge. Prairie Nat- uralist 19:233-238. Plumb, R. E., F. L. Knopf, and S. H. Anderson. 2005. Minimum population size of Mountain Plovers breeding in Wyoming. Wilson Bulletin 117:15- 22. Prellwitz, D. M. 1993. Additional Mountain Plover sightings in Montana. Prairie Naturalist 25:23-26. Thomas, L., J. L. Laake, J. F. Derry, S. T. Buckland, D. L. Borchers, D. R. Anderson, K. P. Burnham, S. Strindberg, S. L. Hedley, M. L. Burt, F. F. C. Marques, J. H. Pollard, and R. M. Fewster. 1998. Distance 3.5. Release 6. Research Unit for Wildlife Population Assessment, University of St. Andrews, St. Andrews, United Kingdom. Tipton, H. C. 2007. Occupancy, abundance, and den- sity of Colorado breeding grassland birds: esti- mation and habitat correlations. Thesis. Colorado State University, Fort Collins, USA. U.S. Department of Agriculture (USDA). 1994. Fi- nal environmental impact statement for manage- ment strategy for Mountain Plover. USDA, Forest Service, Pawnee National Grasslands, Greeley, Colorado, USA. U.S. Department of Interior (USDI). 2000. Bitter Creek and Mountain Plover areas of critical en- vironmental concern plan amendment and envi- ronmental assessment. MT-096-99-04. USDI, Bu- reau of Land Management, Glasgow Field Station, Glasgow, Montana, USA. U.S. Department of the Interior (USDI). 2003. En- dangered and threatened wildlife and plants: with- drawal of the proposed rule to list the Mountain Plover as threatened. Federal Register 68:53083- 53101. Wunder, M. B., F. L. Knopf, and C. A. Pague. 2003. The high-elevation population of Mountain Plo- vers in Colorado. Condor 105:654-662. The Wilson Journal of Ornithology 1 20(4):708-7 16, 2008 LAND COVER ASSOCIATIONS OF NESTING TERRITORIES OF THREE SYMPATRIC BUTEOS IN SHORTGRASS PRAIRIE SCOTT McConnell, '•■*•5 timothy j. o’connell,^ and DAVID M. LESLIE JR.’ ABSTRACT. — Three species of Buteo hawks nest sympatrically in the southern Great Plains of the United States. Dietary overlap among them is broad and we tested the hypothesis these species partition their breeding habitat spatially. We compared land cover and topography around 224 nests of the three species breeding in shortgrass prairie in 2004 and 2005. Red-tailed Hawks {Buteo jamaicensis) nested almost exclusively in riparian timber surrounded by prairie (95% prairie land cover around nests) and disproportionately used areas with greater topographic relief within prairie landscapes. Swainson’s Hawks {B. swainsoni) commonly nested in low-relief areas dominated by small-grain production agriculture but generally used habitats in proportion to availability. Most nest sites of Ferruginous Hawks {B. regalis) were in prairie (78% prairie land cover around nests), but some were in areas that were at least partially agricultural. Ferruginous Hawks had at least two times more sand sagebrush {Artemisia filifolia) around their nests than their two congeners. We conclude that sympatric breeding Buteos on the southern Great Plains spatially partitioned nest sites according to subtle differences in land cover and topography. Received 12 March 2007. Accepted 17 February 2008. Species with similar ecological niches may compete for vital resources. Pianka (1974) warned against the assumption that niche overlap indicates competition because abun- dant resources and/or niche modification could make competition irrelevant. Jaksic and Braker (1983) concluded that diurnal raptors do not adjust their diets because of competi- tion, but rather partition their habitat use via agonistic encounters. This was supported by observations of sympatric Swainson’s {Buteo swainsoni) and Harris’s {Parabuteo unicinc- tus) hawks in New Mexico where overlap of prey species remained high even as prey bio- mass declined, and Swainson’s Hawks aggres- sively excluded Harris’s Hawks from their ter- ritories (Gerstell and Bednarz 1999). In western North America, Red-tailed {Bu- teo jamaicensis), Swainson’s, and Ferruginous {B. regalis) hawks breed sympatrically in ’ Department of Zoology, 430 Life Sciences West, Oklahoma State University, Stillwater, OK 74078, USA. 2 Department of Natural Resource Ecology and Management, 240 Agriculture Hall, Oklahoma State University, Stillwater, OK 74078, USA. ^ Oklahoma Cooperative Fish and Wildlife Research Unit, U.S. Geological Survey, 404 Life Sciences West, Oklahoma State University, Stillwater, OK 74078, USA. ^ Current address: 330A Pahlow Lane, Laramie, WY 82070, USA. Corresponding author; e-mail: scott.mcconnell 1 @yahoo.com shortgrass prairies in the southern Great Plains. These species are diurnal predators that seek elevated nest sites (Preston and Be- ane 1993, Bechard and Schmutz 1995, Eng- land et al. 1997); thus, they may compete for breeding territories. From studies in Oregon (Cottrell 1981), Washington (Bechard et al. 1990), and Alberta (Schmutz et al. 1980), au- thors concluded that sympatrically breeding Red-tailed, Ferruginous, and Swainson’s hawks coexisted by partitioning space, not prey. Habitat relationships in the southern Great Plains have received comparatively less attention, but Giovanni et al. (2007) found high dietary overlap between Swainson’s and Ferruginous hawks in southern Cimarron County, Oklahoma, and two adjacent counties in Texas (Dallam) and New Mexico (Union). The Red-tailed Hawk is a widely distrib- uted breeder in Oklahoma (Reinking 2004). Ferruginous and Swainson’s hawks are more restricted in their breeding range in Oklahoma with Swainson’s Hawks largely confined to the western half of the state and Ferruginous Hawks mostly in Texas and Cimarron counties in the Panhandle (Reinking 2004). Under- standing distributions and dynamics of these species is a goal of the Oklahoma Compre- hensive Wildlife Conservation Strategy (ODWC 2005) in which Swainson’s and Fer- ruginous hawks are listed as Tier I and II con- servation priorities, respectively. Swainson’s Hawk has been listed as a species of conser- 708 McConnell et al. • LAND COVER AND BUTEO IN SHORTGRASS PRAIRIE 709 vation concern at the continental scale (Rich et al. 2004). We examined the spatial partitioning of breeding territories among all three Buteos in shortgrass prairie due to the potential niche overlap and the propensity for Swainson’s Hawks to aggressively exclude other species (Schmutz et al. 1980, Janes 1994). Our objec- tives were to quantify and compare land cover, topography, and other features around nests to test the hypothesis the three sympatric Buteo species partition space in the southern Great Plains. METHODS Study Area. — We conducted this study in Cimarron County, Oklahoma within the Short- grass Prairie Bird Conservation Region (Rich et al. 2004) and the Western Short Grasslands Ecoregion (Ricketts et al. 1999). Cimarron County occupies 4,769 km^ at 1,090-1,518 m elevation, increasing from east to west. Most of the area is flat, but the northwestern part of the county is dominated by mesas, including Black Mesa, the highest point in Oklahoma. Average annual precipitation in the Boise City area was 47 cm from 1971 to 2000 (NOAA 2002). Ninety-six percent of Cimarron County was farmland in 2002 of which cropland occupied 40% and rangeland 56% of the land area. Crops and area harvested in 2002 were: sor- ghum, 236 km^; winter wheat, 228 km^; corn, 88 km^; and forage (hay, etc.), 65 km^ (USDA 2004). Many of the agricultural areas had cen- ter-pivot irrigation systems drawing on the Ogalalla Aquifer. Cimarron County had 641 km^ enrolled in the Conservation Reserve Program (CRP) in 2004 (P. G. Toon, USDA, pers. comm.). Historically, the Western Short Grasslands were characterized by buffalograss {Buchloe dactyloides) and blue grama (Bouteloua gra- cilis) prairie; pastures dominated by these grasses persist in Cimarron County (Ricketts et al. 1999). Sand sagebrush (hereafter sand- sage) {Artemisia filifolia) and plains yucca {Yucca glauca) are common on well-drained soils (Tyrl et al. 2002). Trees are limited to riparian zones and around human settlements. The Cimarron Riv- er flows through the northern half of the coun- ty and the Beaver River bisects the southern half (Fig. 1). Plains cottonwoods {Populus deltoides) and dense stands of salt cedar {Ta- marix gallica) occur in riparian areas; Siberian elms {Ulmus pumilla) have been planted as windbreaks around homesteads. Nest Searches. — We conducted opportunis- tic nest searches from early May to early July in 2004 and 2005, and located nests while driving county roads and walking ranches and river courses. We found some nests in the course of doing unrelated field work. Our total search effort included coverage of >50% of the sections (259 ha/section) in Cimarron County. We considered only active nests in our anal- ysis, defining “active” as the presence of ju- veniles or >1 incubating adult. A nest site used by the same species 2 years in succession was considered independent between years. Over 80% of the locations were plotted using a GPS unit (Garmin Geko 201, Garmin Inter- national, Olathe, KS, USA); we plotted the re- mainder in a GIS by finding the location of the nest tree on an aerial photograph after re- cording the legal location (township, range, 1/ 4-section) in the field. We recorded nest tree species at most locations. GIS Analysis. — We imported all Cimarron County nest locations into ArcMap 9 (ESRI, Redlands, CA, USA), and created buffers of 500-, 1,000-, and 2,000-m radii (79-, 31 4-, and 1,256 ha) around nest locations. We placed the same buffers around 54 random points generated in the GIS with a Hawth’s Tools function (Beyer 2004). We used these three spatial extents to permit comparisons ranging from the immediate vicinity of the nest (79 ha) to an extent (314-1,256 ha) roughly the size of the home ranges reported for the three hawk species (Bechard 1982, An- dersen and Rongstad 1 989, Woodbridge 1991, Bechard and Schmutz 1995, Leary et al. 1998). We imported the 1992 National Land Cover Dataset (NLCD) for Cimarron County from the USGS Seamless Data Distribution System (SDDS) web site (USGS 2004) into ArcMap and merged it with a layer delineating 2005 CRP land that was provided by the Cimarron County Farm Service Agency (NRCS) (Fig. 1 ). We combined potentially redundant fea- tures from the NLCD into a smaller number of more general categories. We renamed the 710 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 LIG. I. Land cover map of Cimarron County, Oklahoma, compiled from National Land Cover Dataset and Conservation Reserve Program data layers illustrating nest locations in 2004-2005 for Swainson’s (stars), Ler- ruginous (triangles), and Red-tailed hawks (hexagons). White areas represent prairie; gray indicates agricultural areas, with CRP areas darker; lightly stippled areas (e.g., northwestern corner) represent wooded vegetation. grasslands/herbaceous designation “prairie” and combined pasture/hay, row crops, and small grains into “agriculture,” and all for- ested and shrubland categories into “wood- ed.” The latter grouping was necessary be- cause many riparian areas in Cimarron County were heavily lined with large cottonwoods, some of which were classified as “shrubland” by the NLCD. We used Oklahoma Gap Analysis Project (GAP; Fisher and Gregory 2001) data in a separate analysis to analyze sandsage land cover around nests. There are large areas of sandsage in Cimarron County which, if rec- ognized by the NLCD, are generally classified as “shrubland;” there is no distinct category for sandsage prairie in the NLCD. We performed land cover analyses by clip- ping the nest-area and random point buffers to the NLCD/CRP and GAP layers and using a raster summary function tool in the GIS to derive the total number of the different land cover type pixels in each buffer. Raster sum- mary results for each buffer were summed to derive the total land cover proportion for each species/random point at each extent. We also compared the NLCD/CRP classifications of 100 GIS-generated random points with a Na- tional Agriculture Imagery Program (NAIP) 2003 photograph (USD A 2003) to test the ac- curacy of the NLCD/CRP land cover catego- ries. We used the same buffers (79-, 314-, and 1,256 ha) created for land cover analysis to investigate topography around each nest site and random point. We imported a Digital El- evation Model (DEM) from the USGS SDDS web site (USGS 2004) and analyzed the “mean maximum slope” for all buffers. In this operation, using Hawth’s Tools (Beyer 2004) and ArcMap’s Spatial Analyst, the GIS calculated a slope percentage for each 30 X McConnell et al. • LAND COVER AND BUTEO IN SHORTGRASS PRAIRIE 711 30 m pixel in each nest site (or random point) buffer. The largest value in the buffer was des- ignated the maximum slope percentage for that buffer. We calculated the mean of the maximum slope across all pixels in each buff- er for all nests of each species (and for all random points). The resulting mean of the maximum slope provided an indication of fea- tures such as cliffs and rock outcrops, and we compared the maximum slope percentage means for all species and random points. We also used the DEM to compare elevation at each nest site and random point. Statistical Analyses. — We analyzed mean percent cover for prairie, CRP, wooded, and, using GAP data, sandsage around Buteo nests and 54 random points. We did not include ag- ricultural land cover in our analysis because prairie and agricultural layers were strongly negatively correlated (r = —0.83) in a prelim- inary analysis of total Cimarron County land- cover. We tested for land cover and topogra- phy differences with one-way analyses of var- iance (ANOVA). We compared median values using Mann-Whitney or Kruskal-Wallis tests when Levene’s tests for equal variances indi- cated unequal variances that could not be sta- bilized with data transformations. All tests were performed at a = 0.05 in Minitab 13.20 (Minitab Inc. 2000). RESULTS We found 168 Swainson’s, 38 Red-tailed, and 18 Eerruginous hawk nests during 2004 and 2005 (Eig. 1). Fourteen of 18 Ferruginous (78%), 33 of 38 Red-tailed (87%), and 148 of 168 (88%) Swainson’s hawk nest locations an- alyzed represented locations recorded in only one of the two field seasons, because we made efforts to find new nest locations during the second field season. Land cover percentages surrounding nests and random points did not differ at the 79-, 31 4-, and 1,256-ha scales, except the wooded component around Red-tailed Hawk nests (^2.95 - P — 0.035). Percent cover of wooded habitat around Red-tailed Hawk nests, however, only ranged from 2 to 4% across the three scales. We selected the 314- ha scale for comparisons among nests of the different species and random points due to the overall similarities in landcovcr composition at the three scales. The land cover layer was 98% accurate in the test of 100 random points; there were two misclassifications and 12 in- stances where accuracy could not be measured by comparison with the NAIP photograph. Variances for Red-tailed Hawk nests were smaller than for the other two Buteos and ran- dom points because of similar land cover composition around all Red-tailed Hawk nests. Levene’s tests on transformed means in- dicated unequal variances when mean cover of CRP, prairie, and wooded vegetation for Red-tailed Hawk nests were included in the analysis; variances were equal when means for these three land cover types were exclud- ed. Thus, we tested for differences in mean CRP, prairie, and wooded vegetation between Swainson’s and Ferruginous hawk nests and random points with ANOVA and used non- parametric tests to examine differences with Red-tailed Hawk nests. Untransformed means for sandsage cover had equal variances with all three species and the random points in- cluded. We found no differences in percent CRP among random points and nest locations for Swainson’s and Ferruginous hawks (F2240 = 1.0, P = 0.37). A Kruskal-Wallis test includ- ing Red-tailed Hawk nests indicated a differ- ence (7/3 = 41.0; P < 0.001; df = 3) with less CRP surrounding Red-tailed Hawk nests (Fig. 2). Prairie cover differed among random points and locations for Swainson’s and Ferruginous hawk nests (P2.240 “ 5.9, P = 0.003); a Tu- key’s confidence interval comparison indicat- ed that, while nest buffers for neither species differed significantly from random sites, prai- rie cover around nests of the two species dif- fered from each other. A Mann-Whitney test for the difference in prairie cover between Red-tailed and Ferruginous hawks was not significant {V = 445, P = 0.24), but the dif- ference was significant comparing Red-tailed and Swainson’s hawk nests {U = 151 16, P < 0.001), and to random points {U = 2045, P < 0.001) (Fig. 2). The proportion of wooded cover was sim- ilar among random points and nest locations for Swainson's and Ferruginous hawks (7 2.240 = 0.6, P = 0.57). A Kruskal-Wallis test in- cluding Red-tailed Hawk nests indicated a dif- ference (/A = 49.8, P < 0.001 ). Wooded cov- er was highest around nests for Red-tailed 712 THE WILSON JOURNAL OL ORNITHOLOGY • VoL 120, No. 4, December 2008 100 90 80 70 60 50 40 30 20 10 0 CRP Prairie Wooded Sandsage PIG. 2. Mean r SE land cover percentages within 1 km (314 ha) of Swainson’s (n = 168), Red-tailed (n = 38), and Pemiginous (n = 18) hawk nests, and 54 random points in Cimarron County, Oklahoma. Sandsage percentages were calculated from a different GIS layer and totals can exceed 100%. Hawks because this species nested in relative- ly continuous riparian forest rather than in iso- lated trees as did the other two Biiteos (Fig. 2). Sandsage cover was higher around Ferru- ginous Hawk nests compared with the other Buteo species (F3278 = 3.6; P = 0.013; Fig. 2), but did not differ from sandsage cover around random points (Tukey’s confidence in- terval included zero). We found no differences between any of the other species or random points. Mean maximum slope differed within spe- cies at the different buffer scales (Table 1; F > 3.4; df > 53; P < 0.035; all tests summa- rized) and we compared the log-transformed means among species across all three scales independently. The mean maximum slope around Red-tailed Hawk nests was higher than for its congeners and random points at all scales; Swainson’s and Ferruginous hawks did not differ from random points (Table 2). Mean nest elevation ranged from 1,239 to 1,256 m and did not differ among species (P3.279 ~ 0-4; P = 0.76). We examined topographic relief around nests of Red-tailed Hawks to learn if the high- er maximum slope was an artifact of this spe- cies nesting disproportionately in prairie (higher relief) than in cultivated areas (lower relief), or if Red-tailed Hawks were associated with areas of especially high relief within prairie. We compared mean maximum slope around Red-tailed Hawk nests at the 314-ha scale with a “prairie” -classified subset {n = 57) of the 100 random points used to verify NLCD/CRP accuracy. Means for Red-tailed Hawk nests were higher (Pi,94 = 13.2; P < 0.001). Eight Swainson’s Hawk nests were <2 km TABLE I. Mean = {n = 168), Red-tailed {n County, Oklahoma. SE maximum slope percentages in 79-, 31 4-, and 1.256-ha buffers around Swainson s = 38), and Lerruginous {n = 18) hawk nests and random points {n = 54) in Cimarron Sites 79 ha 314 ha 1.256 ha Swainson’s Hawk 6.21 ± 0.58 8.96 ± 0.84 12.32 ± 1.09 Red-tailed Hawk 27.52 ± 3.11 35.80 ± 3.31 45.83 ± 3.85 Lerruginous Hawk 9.84 ± 1.49 11.94 = 1.51 16.76 ± 2.14 Random 9.92 ± 1.78 13.39 ± 2.10 17.90 ± 2.54 McConnell et al. • LAND COVER AND BUTEO IN SHORTGRASS PRAIRIE 713 TABLE 2. Between species comparisons for mean maximum slope in 79-, 31 4-, and 1,256-ha buffers around Swainson’s (n = 168), Red-tailed (n = 38), and Eerruginous (n = 18) hawk nests and random points (n = 54) in Cimarron County, Oklahoma. ANOVA comparisons between species and random points, and Tukey intervals, are given for each scale; * indicates P < 0.05. Comparison 79 ha 314 ha 1,256 ha ^3,278 = 35.9 ^3,278 ~ 37.7 ■^3,278 ~ 38.4 P < 0.001 P < 0.001 P < 0.001 Swainson’s vs. Red-tailed -0.8904 -0.8652 -0.8134 -0.5350* -0.5275* -0.4977* Swainson’s vs. Ferruginous -0.49667 -0.4472 -0.4465 -0.0059* 0.0191 -0.0104* Red-tailed vs. Ferruginous 0.1782 0.2133 0.1754 0.7446* 0.7514* 0.6787* Swainson’s vs. Random -0.2739 -0.2744 -0.2701 0.0356 0.0196 0.0048 Red-tailed vs. Random 0.3840 0.3699 0.3367 0.8032* 0.7681* 0.7091* Ferruginous vs. Random -0.1372 -0.1693 -0.1435 0.4016 0.3426 0.3352 from Boise City or Keyes, the two largest towns in the county. All Ferruginous Hawk nests were >8 km from those towns, and all Red-tailed nests were >12 km distant. Four Swainson’s Hawk nests were in “yards” <170 m from a house occupied occasionally (n — 1) or continually (n = 3). One Red-tailed Hawk nest was 40 m from an isolated, occu- pied house. Ten Swainson’s Hawk nests were along major highways (<20 m from the road) in Cimarron County; no Ferruginous or Red- tailed hawk nests were similarly located. We recorded tree species at 134 Swainson’s, 31 Red-tailed, and 14 Ferruginous hawk nest sites. Most Swainson’s Hawk nests were in Siberian elms, which were planted around homesteads in agricultural areas. Most Red- tailed Hawk nests occurred in riparian cotton- woods (Table 3). DISCUSSION Swainson’s, Red-tailed, and Ferruginous hawks selected nest sites with different pro- portions of adjacent land cover types and to- pographic relief. Red-tailed and Ferruginous hawks nested almost exclusively in prairie. Red-tailed Hawks disproportionately used ri- parian timber surrounded by prairie for nest- ing. Ferruginous Hawks nested disproportion- ately in prairie and were more likely to use TABLE 3. Number (%) of tree species used by Buteos in Cimarron County, Oklahoma, 2004-2005 (n — 179). Tree species/nest substrate Swainson’s (n = 134) Red-tailed (/i = 31) Ferruginous (n = 14) Siberian elm {Ulmiis pumilla) 95 (71.0) 1 (3.2) 5 (36.0) Plains Cottonwood (Populiis deltoides) 28 (21.0) 26 (84.0) 4 (29.0) Red mulberry {Morns rnhra) 4 (3.0) Hackberry {Celtis occidentalis) 3 (2.0) Osage orange (Maclnra pom if era) 1 (0.7) 1 (7.0) White mulberry {Moms alha) 1 (0.7) Oneseed Juniper {Jnniperns monosperma) 1 (0.7) 2 (14.0) Black walnut {Jnf>lcms nigra) 1 (3.2) Honey locust {Gleditsia triacanthos) 1 (3.2) 1 (7.0) Black locust {Rohinia pseudoacacia) 1 (0.7) Artihcial nest platform 1 (7.0) Cliff 2 (6.5) 714 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 areas with a high percentage of sandsage cov- er. Swainson’s Hawks were less selective in nesting habitat use than the other Biiteos. They tended to use habitats in proportion to their availability and were the only Buteo nesting in agricultural areas. This is similar to the report by Bechard et al. (1990) in Wash- ington, who also found Swainson’s Hawk nests more likely to be associated with agri- culture than either Ferruginous or Red-tailed hawks. Rothfels and Lein (1983) found Swainson’s Hawks in agricultural areas and Red-tailed Hawks in the foothill parklands of the Rocky Mountains in Alberta. Swainson’s Hawks were more likely to nest near towns, occupied homes, and along high- ways in CimaiTon County, and have been re- ported nesting in residential neighborhoods in California (England et al. 1995). FeiTuginous Hawks are known to avoid human habitations (Olendorff 1993). There was a potential sampling bias for finding higher numbers of Red-tailed Hawk nests than Swainson’s Hawk nests along riv- ers, which could have reduced percent of prai- rie cover around Swainson’s Hawk nests. Wooded floodplains of the Cimanon and Bea- ver rivers were frequently 0.50-0.75 km wide and examination of each tree was prohibitive. Red-tailed Hawks usually announced the pres- ence of a nest with loud protests directed at the nest searcher, but Swainson’s Hawks at times remained inconspicuous by sitting tight on a nest, even when the base of the nest tree was approached. We are confident that we lo- cated most, if not all. Red-tailed Hawk nests along the rivers investigated, but it is likely that at least some Swainson’s Hawk nests were missed. Cultivated areas in Cimarron County are flat, but prairie areas have more hills and cliffs. This may be a contributing factor to spatial partitioning of prairie and agricultural areas between Swainson’s and Red-tailed hawks. Janes (1985) reported that Red-tailed Hawk nest sites had higher topographic relief and higher perch (telephone poles or trees >2 m tall) density, on average, than Swainson’s Hawk nest sites. He attributed this to morpho- logical differences which made Red-tailed Hawks more dependent on slope soaring and perch-hunting. Significant differences in to- pographic relief among Buteo nest sites in Cimarron County indicate these species may select habitats similar to those reported by Janes (1985). The higher mean maximum slope in buffers around Red-tailed Hawk nests supports the hypothesis that this species se- lected greater topographic relief relative to random sites (even random sites within prai- rie) or to habitats of the two congeners. We found two Red-tailed Hawk nests on cliffs in areas of extreme topographic relief, but no cliff-nesting Ferruginous or Swainson’s hawks. We did not quantify perch density around each nest, but significantly higher mean cover of wooded habitat around Red-tailed Hawk nests indicated a higher density of trees in their nesting areas. Most (83%) of the Red- tailed Hawk nests were along the Cimarron and Beaver rivers and associated drainages. Vast stretches of both rivers were heavily lined with cottonwoods and suiTounded by prairie, enabling a hawk to access large tracts of prairie foraging areas from a constant sup- ply of perch trees. Most Swainson’s Hawk nests were in agricultural areas, where trees were more sparsely distributed and generally occurred as windbreaks around widely dis- persed homesteads. The low number of natu- ral perches in agricultural areas was aug- mented by utility poles, but they did not ap- proach the density of perch trees along rivers. We found species-specific differences in land cover around nest sites of Swainson’s, Red-tailed, and Fenuginous hawks. These dif- ferences contributed to a pattern of spatial segregation of nest locations that may reduce interspecific competition in relatively homo- geneous landscapes. We recognize that other landscape features around nests (e.g., fractal dimension, landscape diversity) may also be coiTelated with nest site choices with the spe- cies studied, but we focused on land cover and topography, which are the most readily avail- able information for land managers to assess. Nest selection by Red-tailed and Ferruginous hawks in Cimarron County apparently occurs in advance of spring arrival of Swainson’s Hawks. The potential interspecific nest site habitat partitioning among these species could be addressed in studies of diet overlap and nest productivity (e.g., Giovanni et al. 2007), and analyses of richness and habitat associa- McConnell et cil. • LAND COVER AND BUTEO IN SHORTGRASS PRAIRIE 715 tions for Buteo prey species. Behavioral stud- ies could examine the time allocation in each habitat type within a species’ home range, which may be as important in identifying spa- tial habitat use as the specific nest location. ACKNOWLEDGMENTS Ranchers in Cimarron County graciously allowed access to their properties to conduct nest searches. John Shackford conducted field work with SM for both seasons of the project and located some of the raptor nests. Bill Voelker freely shared knowledge of Cimar- ron County raptors gained during almost 40 years of study in the area. Martin Piorkowski provided software expertise, Mark Payton provided statistical advice, and David Walter and Mahesh Rao provided GIS expertise. Einancial support for this project was provided from State Wildlife Grants under Project T-4-P of the Oklahoma Department of Wildlife Conservation and Oklahoma State University administered through the Oklahoma Cooperative Pish and Wildlife Research Unit (Oklahoma Department of Wildlife Conservation, Oklahoma State University, U.S. Geological Survey, U.S. Fish and Wildlife Service, and Wildlife Manage- ment Institute cooperating). LITERATURE CITED Andersen, D. E. and O. J. Rongstad. 1989. Home- range estimates of Red-tailed Hawks based on random and systematic relocations. Journal of Wildlife Management 53:802-807. Bechard, M. J. 1982. Effect of vegetative cover on foraging site selection by Swainson's Hawk. Con- dor 84:153-159. Bechard, M. J. and J. K. Schmutz. 1995. Ferruginous Hawk {Buteo regal is). The birds of North Amer- ica. Number 172. Bechard, M. J., R. L. Knight, D. G. Smith, and R. E. Fitzner. 1990. Nest sites and habitats of sym- patric hawks (Buteo spp.) in Washington. Journal of Field Ornithology 61:159-170. Beyer, H. L. 2004. Hawth’s analysis tools for ArcGIS. Available at http://www.spatialecology.com/ h too Is CoTTREEL, M. J. 1981. Resource partitioning and re- productive success of three species of hawks (Bu- teo spp.) in an Oregon prairie. Thesis. Oregon State University, Corvallis, USA. Engi.and, a. S., j. a. E.step, and W. R. Hoei. 1995. Nest-site selection and reproductive performance of urban-nesting Swainson's Hawks in the Central Valley of California. Journal of Raptor Research 29:179-186. England, A. S., M. J. Bix hard, and C'. S. lloti.sTON. 1997. Swainson's Hawk (Buteo swainsoni). fhe birds of North America. Number 265. Fisher, W. L. and M. S. Gregory. 2001. Oklahoma Gap Analysis Project: a geographic apiiroach to planning for biological diversity. Oklahoma C'o- operative Fish and Wildlife Research Unit, Oklahoma State University, Stillwater, USA. Gerstell, a. T. and j. C. Bednarz. 1999. Competition and patterns of resource use by two sympatric rap- tors. Condor 101:557-565. Giovanni, M. D., C. W. Boal, and H. A. Whitlaw. 2007. Prey composition and provisioning by breeding Ferruginous Hawks and Swainson's Hawks on the southern Great Plains, USA. Wilson Journal of Ornithology 1 19:558-569. Jaksic, F. M. and H. E. Braker. 1983. Food-niche relationships and guild structure in diurnal birds of prey: competition versus opportunism. Cana- dian Journal of Zoology 61:2230-2241. Janes, S. W. 1985. Habitat selection in raptorial birds. Pages 159-188 in Habitat selection in birds (M. L. Cody, Editor). Academic Press, Orlando, Flor- ida, USA. Janes, S. W. 1994. Partial loss of Red-tailed Hawk territories to Swainson’s Hawks: relations to hab- itat. Condor 96:52-57. Leary, A. W., R. Mazaika, and M. J. Bechard. 1998. Factors affecting the size of Ferruginous Hawk home ranges. Wilson Bulletin 110:198-205. Minitab Inc. 2000. Minitab statistical software. Re- lease 13.20. Minitab Inc., State College, Pennsyl- vania, USA. National Oceanic and Atmospheric Administration (NOAA). 2002. Climatography of the United States Number 81. Monthly station normals of temperature, precipitation, and heating and cool- ing degree days 1971-2000. www5.ncdc.noaa. gov/climatenormals/clim81/OKnorm.pdf (accessed 24 February 2006). Oklahoma Department of Wildlife Conservation (ODWC). 2005. Oklahoma comprehensive wild- life conservation strategy. Oklahoma Department of Wildlife Conservation and Dynamic Solutions Inc. Report to the U.S. Fish and Wildlife Service. www.wildlifedepartment.com/CWCS.htm (ac- cessed 23 February 2006). Olendorff, R. R. 1993. Status, biology, and manage- ment of Ferruginous Hawks: a review. Special Re- port. Raptor Research and Technical Assistance Center, USDI, Bureau of Land Management, Boi- se, Idaho, USA. PiANKA, E. R. 1974. Niche overlap and diffuse com- petition. Proceedings of the National Academy of Science 71:2141-2145. Prp:,st()n, C. R. and R. D. Beani;. 1993. Red-tailed Hawk (Buteo jcunaicensis). The birds of North America. Number 52. Ri:inking, D. L. (Editor). 2004. Oklahoma Breeding Bird Atlas. University of Oklahoma Press, Nor- man, USA. Rich, T. D., C. .1. Bi:ardmori;, H. Bi ri anga, P. J. Bi anchi;r, M. S. W. Bradsiri:i:i, G. S. Butchi.r, D. W. Di.mari si, F. M. Dt nn, W. C. Hi nti r, E. F. 1nig()-Fj IAS, J. A. Ki:NNi:in , A. M. Martei i,, A. O. I’an.iabi, D. N. Pashi ey, K. V. Rose:nherg, ('. M. Rh.stay, j. S. Wend I . and T. C. Wii i . 716 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 2004. Partners in Flight North American landbird conservation plan. Cornell Laboratory of Orni- thology, Ithaca. New York. USA. Ricketts. T. H.. E. Dinerstein. D. M. Olson. C. L. Loucks, W. Eichb.aum. D. DellaSala, K. Ka- v.ANAGH. P. Hed.ao. P. T. Hurley, K. M. Carney, R. Abell, and S. Walters. 1999. Terrestrial ecoregions of North America: a conservation as- sessment. Island Press, Washington, D.C.. USA. Rothfels, M. .and M. R. Lein. 1983. Territoriality in sympatric populations of Red-tailed and Swain- son's hawks. Canadian Journal of Zoology 61:60- 64. SCHMLTZ. J. K., S. M. SCHMUTZ, AND D. A. BO.AG. 1980. Coexistence of three species of hawks {Bii- teo spp.) in the prairie-parkland ecotone. Canadian Journal of Zoology 58:1075-1089. Tyrl. R. J.. T. G. Bidwell, and R. E. Masters. 2002. Field guide to Oklahoma plants. Oklahoma State University, Stillwater, USA. U.S. Department of Agricultlre (USDA). 2003. Na- tional Agriculture Imagery Program. www2.ocgi. okstate.edu/2003imgl (accessed 22 November 2004). U.S. Department of Agriculture (USDA). 2004. 2002 Census of agriculture, Oklahoma state and county data, www.nass.usda.gov/census/census02/ volume 1/ok/OKVolume 104.pdf (accessed 24 Feb- ruary 2006). U.S. Geological Survey (USGS). 2004. The National Map seamless server, seamless.usgs.gov/website/ seamless/viewer.php (accessed 22 August 2005). WoODBRiDGE. B. 1991. Habitat selection by nesting Swainson's Hawks: a hierarchical approach. The- sis. Oregon State University, Corvallis. USA. The Wilson Journal of Ornithology 120(4):717— 724. 2008 SPECIES RECOGNITION IN A VOCAL MIMIC: REPETITION PATTERN NOT THE ONLY CUE USED BY NORTHERN MOCKINGBIRDS IN DISCRIMINATING SONGS OE CONSPECIEICS AND BROWN THRASHERS DUSTIN G. REICHARDi AND J. JORDAN PRICED ABSTRACT. — Vocal mimics that produce large repertoires of song types, such as in the Mimidae. have unique challenges discriminating songs of conspecifics from those of other mimids in areas where these species co- occur. We investigated cues used by Northern Mockingbirds {Minnis polyglottos) in discriminating their songs from songs of a sympatric mimid. the Brown Thrasher (Toxostoma rufiini). We presented territorial mockingbirds with four playback treatments in which either mockingbird song types or thrasher song types had either a standardized mockingbird repetition pattern (5 repetitions) or a standardized thrasher pattern (2 repetitions). Four measures (time within 2 m of speaker, latency to approach, closest approach, and number of flights) were used to estimate a subject's response to each playback. Subjects responded significantly more strongly to mockingbird song types in a mockingbird repetition pattern than to thrasher song types in either repetition pattern. Responses to mockingbird song types in a thrasher repetition pattern elicited intermediate responses. Thus, mockingbirds can distinguish conspecific songs from Brown Thrasher songs based on song types alone regardless of their repetition pattern, although repetition pattern still appears to have a role in conspecific recognition. Brown Thrasher song includes a significantly broader frequency range than mockingbird song, which may allow direct discrimination. Our results suggest cues used by mimids in species discrimination are not necessarily the same as those used by human observers. Received 29 October 2007. Accepted 23 April 2008. Discriminating songs of conspecifics from vocalizations of other species is fundamental- ly important for territorial songbirds. This dis- crimination in many species is facilitated by vocal features shared by individuals within a species but distinctive from sympatric species (reviewed in Becker 1982, Catchpole and Slater 1995, Marler 2004). The structure of many species’ songs, despite variation across broad geographic ranges and even between neighboring conspecifics, is easily recognized by certain species-distinctive characteristics (e.g., Walton et al. 2002, Kroodsma 2005). Discrimination in songbirds that regularly mimic the sounds of other species is not as easy (Baylis 1982). Species discrimination is likely to be especially challenging between vocal mimics that produce large repertoires of song types, particularly when multiple mim- icking species include imitations of the same sounds in their repertoires. Northern Mockingbirds (Mitnu.s polyglot- tos) and Brown Thrashers {To.xostotmi ritfum). ' Department of Biology. St. Mary'.s College of Maryland. St. Mary's City. MD 20686. USA. - Current address; Department of Biology. Indiana University, Bloomington. IN 47405. USA. Corresponding author; e-mail; dgreicha@indiana.edu Family Mimidae, provide an interesting illus- tration of the potential difficulties that differ- ent vocal mimics have in discriminating be- tween each other’s songs. Both species are highly versatile singers with large song rep- ertoires (—100-400 song types/individual in mockingbirds: Wildenthal 1965: Merrit 1985: Derrickson 1987, 1988: >1,000 song types/ individual in Brown Thrashers; Kroodsma and Parker 1977, Boughey and Thompson 1981). and repertoires of both species can include im- itations of the same sounds (Boughey and Thompson 1976). Northern Mockingbirds and Brown Thrashers have widely overlapping geographic ranges in eastern North America, occupy many of the same habitats, and often sing during the same seasons and same times of day (Derrickson and Breitwisch 1992. Cav- itt and Haas 2()()0). Both species defend ter- ritories. primarily against conspecifics. and there is little evidence that either species is interspecihcally territorial towards the other (Howard 1974, Boughey and Thompson 1976. Cavitt and Haas 2()(K)). Ornithologists have long recognized that songs of Northern Mockingbirds and Brown Thrashers can be reliably distinguished in the field by their repetition patterns (e.g.. Bent 1948, Walton et al. 2002). Mockingbirds typ- 717 718 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120. \o. 4, December 2008 ically repeat each song type 4-5 times se- quentially before producing the next song type in a sequence (range = 1-36 repetitions: Wil- denthal 1965). whereas Brown Thrashers typ- ically repeat each song type only twice in suc- cession (range = 1-4 repetitions; Boughey and Thompson 1976). Whether these different repetition patterns are used by mockingbirds and thrashers in species discrimination is not clear. Boughey and Thompson (1976) showed that Brown Thrashers respond more strongly to normal thrasher song than to thrasher song altered to have a longer mockingbird-like rep- etition pattern, as measured by singing and ap- proaches to a speaker. However, the same sub- jects did not respond differently to unaltered Brown Thrasher and mockingbird songs, which makes these results difficult to inter- pret. Responses of Northern Mockingbirds to similar manipulations of repetition pattern have not been previously reported. We investigated cues used by Northern Mockingbirds, and specifically the importance of repetition pattern as a cue. in discriminating mockingbird songs from those of Brown Thrashers. A variety of vocal features other than number of song repetitions might differ consistently between these species and could be used by mockingbirds in conspecific rec- ognition. including frequency and temporal characteristics of song types, intervals be- tween sounds, song amplitude, or the presence of song types that are species-distinctive (Fletcher and Smith 1978. Baylis 1982. Beck- er 1982). We presented subjects with songs of both species in which song repetition patterns were manipulated but other aspects of songs, including sound amplitudes and a variety of temporal characteristics, were standardized. Our study tested whether mockingbirds pref- erentially use either song repetition patterns or characteristics of the song types in discrimi- nating between species. METHODS Playback experiments were conducted from 25 September to 17 November 2(X)6 on the campus of St. Mary's College of Maryland and the neighboring grounds of historic St. Mary's Cit>. Maryland. USA (38“ 11' N. 76“ 25' W). The area includes abundant popula- tions of both Northern Mockingbirds and Brown Thrashers, and members of both spe- cies defend territories in spring that can over- lap interspecifically. Mockingbirds on our study site defend territories year-round and sing from mid-September through November as well as during the spring and summer, as reported for other mockingbird populations (Derrickson and Breitwisch 1992). Thrashers sing only during spring and most individuals leave the area during late fall (Hitchner 1996). We performed our study during fall when only mockingbirds were vocalizing and actively defending territories. This may have caused some of our playback treatments (e.g.. those with Brown Thrasher songs) to seem unnatu- ral. but it was unlikely to have influenced our subjects' abilities to discriminate songs of conspecifics from those of other species. Per- forming our experiments in fall also allowed us to minimize interference by Brown Thrash- ers during playback. Songs of Northern Mockingbirds and Brown Thrashers were recorded in the field by the authors using a Marantz PMD 670 dig- ital recorder and Telinga parabolic micro- phone or were obtained from the Macaulay Library of Natural Sounds (Cornell Labora- tory of Ornithology. Ithaca. NY. USA) or oth- er commercially available sources (Peterson 1990. Elliot et al. 1997). Song recordings used in the study (10 of mockingbirds and 7 of thrashers) were made in a variety of geograph- ic locations and all were recorded in spring between March and July (Table 1). Mocking- birds produce repertoires of different song tvpes in spring and in fall, and both repertoires have been shown to elicit agonistic responses in either season (Burnett 1978. Logan and Lulk 1984). We used spring songs because mockingbirds respond significantly more strongly to these vocalizations regardless of season (Logan and Lulk 1984). We generated onscreen spectrograms for each recording (digitized at either -14.1 kHz or 48 kHz) using Audacity 1.2.4 software (Maz- zoni et al. 2000) and selected 15 unique -30- sec song sequences from each species based on recording quality. No more than three song sequences were taken from any one recorded individual. Each sequence in all cases includ- ed unique song types, and did not contain chats or begging calls. We altered song type repetition patterns to produce two versions of each song sequence (Lig. 1); one with a stan- Reichard and Price • SONG RECOGNITION IN NORTHERN MOCKINGBIRDS 719 TABLE 1. Song recordings used in the study. (A) 15 10 . Species Recording source Mimas polyglottos Mimas polyglottos Mimas polyglottos Peter P. Kellogg, Richmond Air Force Base, EL, May 1950 Robert C. Stein, Rock Springs, TX, April 1961 William W. Gunn, Homestead, EL, March 1968 0 . (B) • 10 . 5 . Mimas polyglottos Mimas polyglottos Mimas polyglottos Mimas polyglottos Mimas polyglottos Mimas polyglottos Mimas polyglottos Toxostoma rafam Toxostoma rafam Toxostoma rafam Toxostoma rafam Toxostoma rafam Toxostoma rafam Toxostoma rafam Wilbur L. Hershberger, Ereder- ick, MD, July 1997 Peterson (1990) Elliot et al. (1997) Authors, St. Mary’s City, MD, April 2006 Authors, St. Mary's City, MD, April 2006 Authors, St. Mary’s City, MD, April 2006 Authors, St. Mary’s City, MD, April 2006 Geoffrey A. Keller, Ocala Na- tional Forest, EL, May 1994 Wilbur L. Hershberger, Freder- ick, MD, March 2000 Peterson (1990) Elliot et al. (1997) Authors, St. Mary’s City, MD, April 2006 Authors, St. Inigoes, MD, May 2006 Authors, St. Inigoes, MD, May 2006 1 I' I * r » - i » - T f f t 4 'f 4 -I 'I V W’ W‘ V''' J d J V.'’ M M ^ KA " ] I ’ t l! >Lt ^ 1' « i’’ (C) (D) FIG. 1. Spectrograms of the four types of song sequences presented to Northern Mockingbird sub- jects: (A) mockingbird song types in a standardized mockingbird repetition pattern (5 repetitions), (B) mockingbird song types in a standardized Brown Thrasher pattern (2 repetitions), (C) thrasher song types in a mockingbird pattern, and (D) thrasher song types in a thrasher pattern. dardized mockingbird pattern (5 repetitions/ bout) and one with a standardized thrasher pattern (2 repetitions/bout). We follow Der- rickson and Breitwisch (1992) in defining a “song type” as an acoustically distinct sound pattern, usually repeated more than once se- quentially, and in defining a “bout” as a group of repeated song types. Intervals be- tween song types and between bouts were not altered in these manipulations, and sounds that were not repeated in these recordings (short sounds that occurred in <25% of recordings) were deleted. All recordings were passed through a 10-band equalizer to remove low frequency background noise below 320 Hz and normalized to the same peak amplitude using wSound Studio 2.2.4 (Freeverse Inc., New York, NY, USA). Each song sequence used in playback experiments was repealed several times sequentially for a total playback duration of 3 min (mean ± SE number of song types/playback; NM = 14.2 ± 0.9. BT = IS. 6 ± 0.9; mean ± SE number of bouts/3 min playback: NM/NM = 73.5 ± 5.6, NM/BT = 136.7 ± 9.4, BT/BT = 128.9 ± 5.6, BT/NM = 66.9 ± 3.5). Each initial recording was used to create two treatment stimuli, one for each repetition pattern type, but both variations from the same original recording were not played to the same subject. Precautions were also taken to ensure that subjects would not hear their own songs or the songs of a nearby neighbor. Thus, each subject heard four song sequences with different song types. Fifteen territorial mockingbird subjects were each presented with four different treat- ments: (1) Northern Mockingbird song types in a standardized mockingbird repetition pat- tern (NM/NM), (2) mockingbird song types in a standardized Brown Thrasher repetition pat- tern (NM/BT), (3) thrasher song types in a standardized mockingbird repetition pattern (BT/NM), and (4) thrasher song types in a standardized thrasher repetition pattern (BT/ BT). Treatments for each subject were pre- sented in random order on the same day at 720 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 different locations in the territory, separated by a minimum of 15 min of silence to mini- mize habituation and fatigue. The speaker was relocated during these intervals within the subject’s territory for the next treatment. Small shrubs and trees with similar branch densities were chosen for each speaker loca- tion, and speaker locations were approximate- ly equidistant from one another as well as from estimated territorial boundaries to con- trol for speaker movement. Territorial bound- aries were estimated by a non-invasive obser- vation period of 15—30 min the day before playback was performed. Researchers mea- suring the responses of subjects did not know the order in which treatments were presented at the time of playback to minimize observer bias, although the repetition pattern was likely discernable after the first few song types were heard. Each treatment began only when the bird was visible, and a trial (playback of all 4 treat- ments) was aborted if the subject was lost from view for >15 min. Birds were not color banded for individual identification in our study; however, subjects were followed visu- ally throughout each trial to ensure the same individual was observed in all four treatments. Males and females are visually indistinguish- able in mockingbirds, and females are known to sing occasionally during fall (Derrickson and Breitwisch 1992). We attempted to focus on males by identifying them by their more frequent singing behaviors before each exper- iment. Songs were broadcast using an SME- AFS amplified field speaker (Saul Mineroff Electronics Inc., Elmont, NY, USA) connected to a 6 GB iPod Mini (Apple Inc., Cupertino, CA, USA). We standardized the volume of playback using a Realistic digital sound pres- sure level meter (fast response, C weighting) to approximate that of naturally singing birds. All experiments were conducted during the period of highest bird activity between 0700 and 1030 hrs EST and only under favorable weather conditions. We measured responses of subjects during each 3-min playback period by recording four measures: (1) the amount of time, in seconds, each bird spent within 2 m of the speaker, (2) the latency, in seconds, to approach towards the speaker >1 m after playback began, (3) the closest approach to the speaker, and (4) TABLE 2. Loadings of individual response mea- sures on the first principal component for playback treatments. Response measure Loading score Time within 2 m of speaker -0.769 Latency to approach 0.827 Closest approach to speaker 0.894 Number of flights -0.71 1 the number of flights >1 m. We conducted a principal components analysis using these four behavioral response measures as vari- ables to assess the overall strength of response to each treatment. The response measures loaded about equally on the first principal component (Table 2), which explained 64.5% of the total variation. We analyzed the result- ing response scores (PC I) with a one-way ANOVA and a Tukey’s HSD post hoc test to compare responses between the four treat- ments. We also compared each response mea- sure individually between treatments using a one-way repeated measures ANOVA with a Bonferroni post hoc test. We analyzed the song sequences used in playback following these experiments to in- vestigate the possibility that our treatments differed in aspects other than song type rep- etition pattern. We measured mean song type durations, bout durations, intervals between bouts, and song type repetition rates for each song sequence used in playback using Raven, Version 1.2.1 (Cornell Laboratory of Orni- thology, Ithaca, NY, USA). Five bouts were chosen randomly from each recording for these measurements, which were used to cal- culate a mean measurement for each song se- quence. We also measured the highest and lowest peak frequencies (“maximum frequen- cy” in spectrogram slices in Raven) that oc- curred in each song sequence to calculate the range of frequencies used (following Price et al. 2006). These data were analyzed using a two sample F-test for variances and an inde- pendent measures r-test. RESULTS A principal components analysis including time spent within 2 m of the speaker, latency to approach, closest approach, and number of flights (Fig. 2A) revealed significantly differ- Reichard and Price • SONG RECOGNITION IN NORTHERN MOCKINGBIRDS 721 (A) 0.8 0.6 0.4 £ 0.2 o 0 1 ° 1^.2 -0.4 -0.6 -0.8 -1 Song types/repetitions □ NM/NM ■ NM/BT ^ BT/NM □ BT/BT Time (s/10), Distance (m) or Number EIG. 2. (A) Mean (±SE) scores from a principal components analysis of mockingbird responses to playback treatments in which either Northern Mock- ingbird (NM) or Brown Thrasher (BT) song types were presented in either a NM or BT repetition pattern. (B) Mean (±SE) response measures included in the prin- cipal components analysis: time spent within 2 m of the speaker (Within 2 m), latency to approach >1 m towards the speaker (Latency), closest approach to the speaker (Closest), and number of flights (Flights) (n = 15 subjects). ent responses to mockingbircd song types in a mockingbird repetition pattern (NM/NM) than to Brown Thrasher song types with either rep- etition pattern (one-way ANOVA with Tu- key’s HSD post hoc, n = 15, ^3^^ = 5.177; BT/BT, P = 0.006; BT/NM, P = 0.01). Birds responded similarly to Brown Thrasher song types regardless of repetition pattern {P > 0.99). Mockingbird song types in a thrasher repetition pattern (NM/BT) elicited an inter- mediate response that was not signihcantly different from responses to any of the other three treatments (BT/BT, P = 0.26; BT/NM, P = 0.33; NM/NM, P = 0.41). Mockingbirds generally responded more strongly to playback of conspecihc song types than to Brown Thrasher song types, compar- ing individual response measures among treat- ments (Fig. 2B). For example, subjects spent significantly more time within 2 m of the speaker during both treatments with mocking- bird song types than during treatments with Brown Thrasher song types (one-way repeated measures ANOVA with Bonferroni post hoc, F342 = 13.328, NM/NM vs. BT/BT, P = 0.001; NM/NM vs. BT/NM, P = 0.002; NM/ BT vs. BT/BT, P = 0.011; NM/BT vs. BT/ NM, P = 0.048). Responses to the same song types with different repetition patterns did not differ in time spent near the speaker (NM/NM vs. NM/BT, P > 0.99; BT/BT vs. BT/NM, P > 0.99). Latency to approach did not differ significantly among treatments (F342 = 2.106, P = 0.13); however, 13 of 15 subjects re- sponded in 10 sec or less to NM/NM songs, whereas only 7 of 15 responded as quickly during NM/BT songs and 6 of 15 during both BT/NM and BT/BT treatments. Subjects ap- proached the speaker somewhat more closely during NM/NM playback than during treat- ments with BT/BT songs (F342 = 2.997, P = 0.03) or BT/NM songs {P = 0.098). Subjects also performed significantly more flights in re- sponse to NM/NM songs than to Brown Thrasher song types with either repetition pat- tern (F342 = 5.773; BT/BT, P = 0.031; BT/ NM, P = 0.029). No behavioral measures dif- fered significantly in response to the two treat- ments with Brown Thrasher song types (BT/ NM and BT/BT). Measurements of the song sequences used in our playback experiments revealed no sig- nificant differences in mean song type dura- tions between Northern Mockingbirds and Brown Thrashers (r-test; two sample assuming unequal variances, n — 15, P = 0.28). There was no significant difference in bout duration between treatments with the same repetition pattern {P = 0.88), but bout duration was sig- nificantly different between treatments with two repetitions/song type and five repetitions/ song type {P = 0.003). We found no differ- ences in the mean intervals between bouts with different repetition patterns {P = 0.61). Songs with different song types but the same repetition patterns did not differ in their mean song repetition rates (NM/NM vs. BT/NM, P = 0.48; NM/BT vs. BT/BT, P = 0.32) or in the total number of bouts included in each playback (NM/NM vs. BT/NM, P = 0.48; NM/BT vs. BT/BT, P = 0.32). Songs with mockingbird song types, however, exhibited a 722 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 significantly narrower range of mean peak fre- quencies (800-6,700 Hz) than songs with thrasher song types (600-9,250 Hz; P < 0.001). Brown Thrasher songs included both significantly lower frequencies (P = 0.008) and significantly higher frequencies (P < 0.001) than mockingbird songs, on average, which may be attributed to more energy in the upper harmonics of thrasher song (Fig. 1). DISCUSSION Our results indicate that Northern Mocking- birds can distinguish between conspecifics and Brown Thrashers based on their song types alone, regardless of the repetition pattern in which these songs are presented. Subjects re- sponded strongly to playback of mockingbird song types in a standardized mockingbird rep- etition pattern, but responded relatively little to Brown Thrasher song types even when these sounds were played in exactly the same pattern of five repetitions/bout. Mockingbird song types with a Brown Thrasher repetition pattern elicited intermediate responses, gen- erally stronger than responses to Brown Thrasher song types but not as strong as the responses to normal mockingbird song. Thus, although repetition pattern was not the prin- cipal cue used by our subjects to discriminate their own species’ songs from thrasher songs, it appeared to have a role in conspecific rec- ognition. Most North American field guides indicate that songs of species in Family Mimidae (e.g.. Northern Mockingbirds, Brown Thrashers, Gray Catbirds \Dumetella carolinensis]) can be most easily distinguished by their distinc- tive repetition patterns (e.g.. Bent 1948, Cim- prich and Moore 1995, Cavitt and Haas 2000, Walton et al. 2002); this cue is widely used by human observers in recognizing songs of these three species. Our results are surprising in demonstrating that mockingbirds preferen- tially use different vocal cues than what we typically use in discriminating species. None of our playback songs differed consistently in song type durations, amplitudes, or in inter- vals between song types or bouts. Our sub- jects appeared to recognize conspecific vocal- izations based on acoustic features of the song types themselves. Mockingbirds apparently have the ability to distinguish between two and five items, based on a study by Farnsworth and Smolinski (2006) which focused on visual discrimina- tion. In mockingbird song, however, the num- ber of times in which song types are repeated can potentially vary over a relatively wide range (1-36 times according to Wildenthal 1965), which could explain in part why rep- etition pattern is not used as a primary cue in conspecific recognition. Northern Mockingbirds increase the num- ber of repetitions/bout during countersinging between territorial males (Derrickson 1988). Experiments with a closely related species. Tropical Mockingbird {Mimus gilvus), show that territorial males respond more strongly to a higher number of repetitions/bout during playback (Botero and Vehrencamp 2007). Playing mockingbird song types with a higher number of repetitions in our study elicited a stronger response, but increasing the repeti- tion number of thrasher song types had no measurable effect on mockingbird responses. Repetition pattern in mockingbirds might have an important role in communication be- tween territorial males rather than simply ad- vertising species or individual identity. Mock- ingbirds are also known to decrease repeti- tions/bout during courtship and intersexual singing (Derrickson 1988). It is possible that our treatment of only two repetitions/bout was recognized as a mockingbird repetition pat- tern, although one atypical for the season. This may explain the intermediate response of our subjects to this treatment. Our findings agree somewhat with results of previous playback studies in which other mimid species were tested using songs with artificially altered patterns (Boughey and Thompson 1976, Fletcher and Smith 1978). For example. Gray Catbirds apparently do not distinguish their songs from those of other mimids based on repetition pattern alone. Al- tering repetition patterns of catbird, thrasher, or mockingbird songs has no apparent effect on a catbird’s ability to recognize species (Boughey and Thompson 1976); changing the order of catbird song types or even playing their songs backwards also has no effect (Fletcher and Smith 1978). Boughey and Thompson (1976) demonstrated that Brown Thrashers respond similarly to mockingbird song types with different repetition rates, in- cluding a thrasher-like repetition pattern. Reichard and Price • SONG RECOGNITION IN NORTHERN MOCKINGBIRDS 723 Brown Thrashers respond more strongly to Brown Thrasher songs in the normal repeti- tion pattern than to thrasher songs in a mock- ingbird pattern, suggesting that repetition pat- tern is important for species recognition in thrashers. However, these experiments also showed that Brown Thrashers do not discrim- inate between normal conspecific song and normal mockingbird song, which makes these results difficult to interpret. How our mockingbird subjects were able to distinguish mockingbird song types from Brown Thrasher song types with the same rep- etition pattern is not clear. One possibility is that our subjects were familiar with song types we included in the study or that mockingbird songs generally include song types that are species-specific and indicate species identity. This seems unlikely, however, as most of our subjects appeared to make this discrimination after hearing only a few different song types (i.e., within the first 10 sec of playback). Fur- thermore, mockingbirds include an extensive variety of mimicked sounds in their song rep- ertoires (Derrickson and Breitwisch 1992) and composition of these repertoires can vary geo- graphically (Thompson et al. 2000), with age and social context (Derrickson 1987, 1988), and even seasonally between spring and fall (Burnett 1978). Our study was conducted in fall and included mockingbird songs recorded in spring from widely different locations and different years (Table 1). Some songs used in our study had been recorded from birds in our study population (4 Northern Mockingbird songs and 1 Brown Thrasher song), but we made sure that subjects did not hear their own song types or song types recorded from a nearby territory. A more likely explanation for our results is that mockingbird and Brown Thrasher song types differ consistently in certain acoustic characteristics. We found that Brown Thrash- ers use a significantly wider range of sound frequencies than mockingbirds, as has been noted in previous studies (Wildenthal 1965, Boughey and Thompson 1976; Fig. 1). It is possible that mockingbirds are unable to pro- duce the extremely high and low frequency whistles of which Brown I’hrashers are ca- pable, and studies investigating the limitations on song performance in mockingbirds are consistent with this possibility (ca. 750 11/ — ca. 7,000 Hz in Zollinger and Suthers 2004). Other attributes that might differ between these species, not investigated in our study, include rate and pattern of frequency changes within song types, production of two sounds simultaneously, relative amplitudes of har- monics, and overall song amplitude. Further studies will be needed to identify the principal cues used by these birds and by other vocal mimics in recognizing conspecifics by song. ACKNOWLEDGMENTS We thank Megan Stallman, Mary Clapp, Cathy Brandt, Emily Myron, and Katie Zdilla for assistance in the field, and Kevin Boyle for help in developing the project and field assistance. We also thank Dawn O’Neal, Sara Schrock, Amy Skypala, and four anon- ymous reviewers for comments on the manuscript. Funding for this study was provided by a grant from the Maryland Ornithological Society and a Myron Marlay Award from St. Mary’s College of Maryland to D. G. Reichard. LITERATURE CITED Baylis, J. R. 1982. Avian vocal mimicry; it’s function and evolution. Pages 51-83 in Acoustic commu- nication in birds. Volume 2 (D. E. Kroodsma and E. H. Miller, Editors). Academic Press, New York, USA. Becker, P. H. 1982. The coding of species-specific characteristics in bird sounds evolution. Pages 213-252 in Acoustic communication in birds. Volume 1 (D. E. Kroodsma and E. H. Miller, Ed- itors). Academic Press, New York, USA. Bent, A. C. 1948. Life histories of North American nuthatches, wrens, thrashers, and their allies. U.S. National Museum Bulletin. Number 195. Botero, C. a. and S. L. Vehrencamp. 2007. Respons- es of male Tropical Mockingbirds {Minins gilvns) to variation in within-song and between-song ver- satility. Auk 124:185-196. Boughey, M. J. and N. S. Thomf\son. 1976. Species specificity and individual variation in the songs of the Brown Thrasher (Toxostonui rnfnni) and Cat- bird {Diiinetella carolinensis). Behaviour 57:64— 90. BouciHEY, M. j. and N. S. Thomp.son. 1981. Song va- riety in the Brown Thrasher {Toxostonui rnfnni). Zeitschrit't fiir Tierpsychologie 56:47-58. Bt'RNiUT, L. J. 1978. Mockingbird song {Minins jioly- glottos): an investigation within and across sea- sons. Thesis. Lhii\ersity ol North Carolina, Greensboro, USA. Caichpoi i;, C'. K. and P. .1. B. Si An:R. 1995. Bird song: biological themes and \ariations. Cam- bridge University Press. New York. USA. (’aviei, j. E and ('. A. Haas. 2000. Brown Thrasher {'Toxostonui rnfnni). The birtls of North America. Nutnber 557. 724 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 CiMPRiCH, D. A. AND F. R. Moore. 1995. Gray Catbird (Dumetella carolinensis). The birds ot North America. Number 167. Derrickson. K. C. 1987. Yearly and situational chang- es in the estimate of repertoire size in Northern Mockingbirds (Mimiis polyglottos). Auk 104:198— 207. Derrickson. K. C. 1988. Variation in repertoire pre- sentation in Northern Mockingbirds. Condor 90: 592-606. Derrickson. K. C. .\nd R. Breitwisch. 1992. Northern Mockingbird (Mimiis polyglottos). The birds of North America. Number 7. Elliot. L., D. Stokes. a.nd L. Stokes. 1997. Stokes field guide to bird songs, eastern region. Time Warner Audiobooks. New York. USA. Farnsworth. G. L. and J. L. Smolinski. 2006. Nu- merical discrimination by wild Northern Mock- ingbirds. Condor 108:953-957. Fletcher. L. E. and D. G. Smith. 1978. Some param- eters of song important in conspecific recognition by Gray Catbirds. Auk 95:338-347. Hitchner. S. 1996. Brown Thrasher. Pages 298-299 in Atlas of the breeding birds of Maryland and the District of Columbia (C. S. Robbins, Editor). Uni- versity of Pittsburgh Press. Pittsburgh. Pennsyl- vania. USA. Howard, R. D. 1974. The influence of sexual selection and interspecific competition on Mockingbird song {Mimiis polyglottos). Evolution 28:428—438. Kroodsma, D. E. 2005. The singing life of birds: the art and science of listening to birdsong. Houghton Mifflin Company. Boston, Massachusetts, USA. Kroodsma, D. E. and L. D. Parker. 1977. Vocal vir- tuosity in the Brown Thrasher. Auk 94:783—785. Logan. C. A. and K. R. Fulk. 1984. Differential re- sponding to spring and fall song in Mockingbirds {Mimiis polyglottos). Journal of Comparative Psy- chology 98:3-9. Marler, P. 2004. Science and birdsong: the good old days. Pages 11-18 in Nature's music (P. Marler and H. Slabbekoom. Editors). Elsevier Academic Press. New York. USA. M.xzzonl D.. M. Brubeck. J. Crook. V. Johnson. L. Lucius, and M. Meyer. 2000. Audacity, Version 1.2.4. SourceEorge.net. Fremont. California. USA. Merrit, P. G. 1985. Song function and the evolution of song repertoires in the Northern Mockingbird {Mimiis polyglottos). Dissertation. University of Miami, Coral Gables. USA. Peterson. R. T. 1990. A field guide to bird songs: eastern and central North America. Third Edition. Houghton Mifflin Company, Boston. Massachu- setts. USA. Price. J. J.. S. M. Earnshaw, and M. S. WTbster. 2006. Montezuma Oropendolas modify a compo- nent of song constrained by body size during vo- cal contests. Animal Behaviour 71:799—807. Thompson, N. S., E. Abby, J. Wapner. C. Logan, P. G. Merritt, and A. Pooth. 2000. Variation in the bout structure of Northern Mockingbird {Mimiis polyglottos) singing. Bird Behavior 13:93-98. Walton. R. K., R. W’. Lawson, and R. T. Peterson. 2002. Birding by ear: eastern and central North America. Houghton Mifflin Company, Boston, Massachusetts, USA. W'lLDENTHAL, J. L. 1965. Structure in primary song of the Mockingbird {Mimiis polyglottos). Auk 82: 161-189. Zollinger, S. A. and R. A. Suthers. 2004. Motor mechanisms of a vocal mimic: implications for birdsong production. Proceedings of the Royal So- ciety of London Series B Biological Sciences 271: 483-491. The Wilson Journal of Ornithology 120(4);725-73 1 , 2008 LESSER SNOW GEESE AND ROSS’S GEESE FORM MIXED FLOCKS DURING WINTER BUT DIFFER IN FAMILY MAINTENANCE AND SOCIAL STATUS JON EINAR JONSSON' ^-* AND ALAN D. AFTON^ ABSTRACT. — Smaller species are less likely to maintain families (or other forms of social groups) than larger species and are more likely to be displaced in competition with larger species. We observed mixed-species flocks of geese in southwest Louisiana and compared frequencies of social groups and success in social encounters of Lesser Snow Geese {Chen caerulescens caerulescens; hereafter Snow Geese) with that of the smaller, closely- related Ross’s Geese (C. rossii). Less than 7% of adult and <4% of juvenile Ross’s Geese were in families, whereas 10-22% of adult and 12-15% of juvenile Snow Geese were in families. Snow Geese won 70% of interspecific social encounters and had higher odds of success against Ross’s Geese than against individuals of their own species. The larger Snow Geese maintain families longer than Ross’s Geese, which probably contrib- utes to their dominance over Ross’s Geese during winter. Predator vigilance probably is an important benefit of mixed flocking for both species. We suggest the long-standing association with Snow Geese (along with asso- ciated subordinate social status) has selected against family maintenance in Ross’s Geese. Received 23 August 2007. Accepted 27 February 2008. Body size seemingly influences social be- havior and foraging behavior in many animals (Calder 1996). Body size has important phys- iological implications for birds: (1) rate of heat loss increases with decreasing body size because of increasing surface to volume ratio (Goudie and Ankney 1986, Calder 1996); (2) mass-specific metabolic rate is inversely re- lated to body mass (Kendeigh 1970, Calder 1996); (3) gut size scales linearly with body size and partly affects the rate of energy ex- traction from food (Demment and Van Soest 1985); and (4) larger species generally have greater fasting endurances than smaller spe- cies (Goudie and Ankney 1986, Calder 1996, Jonsson et al. 2007). Smaller species are rel- atively less likely to maintain social groups, generally select more sheltered habitats, and consume more specialized diets (Jarman 1974, Shelley et al. 2004). Smaller species also are more likely to be displaced in competition with larger species, regardless of numbers present (Shelley et al. 2004) and often use scramble tactics in competition for food ' School of Renewable Natural Resources, Louisiana State University, Baton Rouge, LA 70803, USA. - Current address; University of Iceland, Snafiellsnes Research Centre, Hafnargata 3. 340 Stykkisholmur, Iceland. ^ USGS, Louisiana Cooperative Fish and Wildlife Research Unit, Louisiana State University. Baton Rouge, LA 70803. USA. ••Corresponding author; e-mail; joneinar^^hi.is (Krause and Ruxton 2002). Smaller species generally are more vulnerable to predator at- tacks than larger species, but can benefit by forming mixed flocks with larger species, which at times have better predator detection capabilities (McWilliams et al. 1994, Kristian- sen et al. 2000, Randier 2004). Most geese, including Lesser Snow Geese {Chen caerulescens caerulescens', hereafter Snow Geese) maintain families from one breeding season to the beginning of the next (family social system) (Boyd 1953; Raveling 1970; Prevett and Macinnes 1980; Black and Owen 1989a, b; Gregoire and Ankney 1990; Kalmbach 2006). Larger goose families gen- erally are dominant over smaller families, pairs, and lone geese (Loonen et al. 1999, Stahl et al. 2001, Kalmbach 2006). Parents ap- parently profit from juvenile assistance when defending patches of food from other flock members (contributor effect hypothesis. Black and Owen 1989b). Conversely, smaller goose species may not maintain families in winter (McWilliams and Raveling 1998). In California, Ross's Geese (C. rossii) form denser flocks than larger goose species when foraging on grasslands where they often associate with Cackling Geese {Branta hutchinsii) (Johnson and Rav- eling 1988, McWilliams and Raveling 1998). Only a small percentage of Ross's Geese in California are paired or in families (dense- flock social system) (Johnson and Raveling 725 726 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. J20, No. 4, December 2008 1988, McWilliams and Raveling 1998). Fam- ily maintenance varies in relation to food choices (seeds vs. leafs and tubers) and fine- scale spatial distribution of selected food plants among habitats (McWilliams and Rav- eling 1998). Ross’s Geese probably are more at risk from avian predators than are larger goose species and predation pressure may have been an important evolutionary factor (albeit not the only factor) influencing their social system during winter (McWilliams et al. 1994). In ad- dition, denser flocks may convey benefits of decreased nearest neighbor distance, i.e., birds in denser flocks are able to spend more time feeding and less time vigilant (Fernandez-Jur- icic et al. 2004, 2007). Southwest Louisiana is a historical winter- ing area for Snow Geese (Jonsson and Afton 2006 and citations therein), but Ross’s Geese only began wintering in Louisiana during the last decade (Jonsson 2005). Ross’s Geese in Louisiana occur only in rice-prairies (cultivat- ed former tail-grass prairie) where they pri- marily forage in rice fields in mixed flocks with Snow Geese. Ross’s Geese have small bills that are adapted for grazing on grass (Ry- der and Alisauskas 1995). Thus, they are rare- ly found in marshes along the Gulf Coast, which are historical habitats of Snow Geese (which have larger bills adapted for excava- tion of marsh plants) (Alisauskas 1998). Dominant goose species can affect feeding behavior, distribution, and food selection of less aggressive species when feeding in mixed flocks (Kristiansen and Jarrett 2002). Interspe- cific dominance relationships often are affect- ed by the number of individuals present from each species (i.e., the more numerous species is dominant within mixed flocks; Fox and Madsen 1981, Madsen 1985, Gawlik 1994), although exceptions are known (Kristiansen and Jarrett 2002). We present the first quantitative comparison of (1) frequencies of pairs and families, and (2) frequencies and outcomes of intra- and in- terspecific social encounters of Ross’s Geese and Snow Geese in mixed wintering flocks. Our study provides a comparison of two closely-related species of varying size forag- ing together on the same plant species and controls for effects of macro habitat, geo- graphical location, season, and climate (Jons- son et al. 2007). METHODS Study Area. — We observed Snow Geese and Ross’s Geese in the rice prairie region of southwest Louisiana in winters 2002-2003 and 2003-2004 (Jonsson 2005; Jonsson and Afton 2006, 2008). Rice prairies are former tail-grass prairies which are extensively cul- tivated and managed, mostly for rice, but also as pasture for cattle (Alisauskas et al. 1988, Bateman et al. 1988). This area has been de- scribed by Alisauskas et al. (1988) and Bate- man et al. (1988). We exclusively observed mixed flocks comprised of Snow Geese and Ross’s Geese using foraging habitats, i.e., non-flooded rice- fields, which were uncut, stubble, tilled, or fal- low (Alisauskas et al. 1988, Jonsson 2005). Ross’s Geese comprised, on average, 7% of observed mixed white goose flocks during our study. Estimated combined Snow Goose and Ross’s Goose numbers on our study area were 257,119 in 2002-2003 and 360,487 in 2003- 2004 (Eronczak 2004). Sampling of Focal Geese. — Three trained observers and JEJ collected behavioral data in winters 2002-2003 and 2003-2004; JEJ was the only observer present in both winters and trained other observers, until results of obser- vation of the same focal birds were nearly identical among observers (Jonsson and Afton 2006, 2008). We are confident that inter-ob- server variation between or within years did not bias our results. We sampled goose behavior 3-4 days/week from 10 November until 10 Eebruary each winter. Observations were made during day- light between 0800 and 1700 hrs CST Mixed flocks were large (a few hundred to a few thousand), mobile, and flushed often; thus, risk of repeated sampling of individuals was minimal. Observers used 20X spotting scopes and collected 5 to 10-min focal sampling obser- vations of randomly selected individuals (Alt- mann 1974). We used sequences of 20 random numbers to select focal geese within a field of vision, counting from left or right until a goose was located that corresponded to each random number. Time of day was not a variable of biological Jonsson and Afton • MIXED FLOCKS AND FAMILY MAINTENANCE IN WINTERING GEESE 727 interest in this study. We attempted a priori to control for time of day variation in behavior by alternating between species during field observations to ensure that comparisons be- tween species were unbiased (Jonsson and Af- ton 2006, 2008). Both species were sampled equally during mid-day (1100-1300 hrs) when geese were relatively prone to cease ac- tivities and rest. We assigned age classes to Snow Geese and Ross’s Geese based on plumage color and pat- terns (Ryder and Alisauskas 1995, Mowbray et al. 2000). We assigned pair and family sta- tus to individuals under observation based on mutual participation in social encounters, mu- tual chasing or avoiding other geese, and co- ordinated directions of locomotion (Raveling 1970, Black and Owen 1989a, Gregoire and Ankney 1990). We categorized focal individ- uals into five social groups (after Boyd 1953, Raveling 1970, Gregoire and Ankney 1990): (1) lone adult, a lone after-hatch-year goose; (2) parent, adult goose bonded (i.e., paired) with another adult goose, accompanied by at least one hatch-year bird; (3) paired non-par- ent, adult goose bonded with another adult goose without hatch-year birds; (4) juvenile in family, hatch-year goose accompanied by adult parents; and (5) lone juvenile, a lone hatch-year goose. We recorded frequencies of social encoun- ters between focal geese and other geese, scor- ing wins if opponents responded to interac- tions by avoiding or fleeing focal geese; focal birds were assigned a loss if an opponent chased them (Raveling 1970, Gregoire and Ankney 1990). We only recorded social en- counters directly involving focal geese, their mates, parents or offspring. Statistical Analyses. — We used a general- ized linear model in PROC GENMOD (Agresti 1996, SAS Institute 1999) to estimate whether frequencies of social groups (parents, non-parental pairs, and lone geese) differed between species, age groups, and winters, which were categorical explanatory variables. Final models were selected using backwards stepwise model selection (Agresti 1996), ex- cept the age X social group interaction was fixed (regardless of significance) in this model because pairs without juveniles were not ob- served in the Juvenile category. We constructed generalized linear models based on normal and Poisson distributions; the Poisson log-linear model is equivalent to a lo- gistic regression based on the multinomial dis- tribution (Agresti 1996). We evaluated good- ness of fit for these models by comparing ra- tios between degrees of freedom (df) and de- viance of the models; a ratio of deviance/df close to 1 .0 indicates a good model ht (Agresti 1996). A linear model based on the normal distribution fit the data reasonably well (de- viance = 24.0, df = 15), whereas the Poisson model gave a poorer fit (deviance = 149.6, df = 15). We calculated probabilities of winning en- counters against the other species (Pother) compared odds of success (Osuccess) inter- specific social encounters versus intraspecific social encounters for both species. Our goal was to quantify success in interspecific en- counters, using success in intraspecific en- counters as a baseline value. We calculated odds ratios of winning against the other spe- cies over the odds of winning against a con- specific (Oown) as: {Osuccess against other species = Probability of winning (Po,he.V(l “ Pother)} { O^uccess against own species = Probability of winning (P^^J/(1 - Po^n)}- We assumed that differing odds of success (unequal odds ratio) between Snow Geese and Ross’s Geese indicated that one species was dominant over the other species, whereas odds ratios of ~1 indicated equal success in inter- specific social encounters and equal social sta- tus for the species. RESULTS Frequencies of social groups differed be- tween species (y^ — 6.12, P = 0.013) and age groups (x“ — 35.55, P < O.OOl), but not be- tween winters (x“ = 0.53, P — 0.466). The ratio of juveniles to adults was relatively high- er for Snow Geese in both winters (Table 1 ). Less than 7% of adult and <4% of juvenile Ross’s Geese were in families, whereas 10- 22% of adult and 12-15% of juvenile Snow Geese were in I'amilies ( fable 1). On average, focal Snow Geese were three to 10 limes more likely to have intraspecilic encounters than interspecific encounters (Ta- 728 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 TABLE 1. rice prairies < Age and social groups (% of observations) of focal Lesser Snow Geese and Ross’s Geese in the of southwest Louisiana during winters, 2002-2003 and 2003-2004. Age Social group Lesser Snow Geese Ross’s Geese 20()2-20()3 2003-2004 2002-2003 2003-2004 Adults Lone 40.5 31.1 58.9 40.7 Paired parents 9.8 22.1 0.0 6.9 Paired non-parents 26.0 25.4 29.5 41.6 Juveniles Lone 11.7 6.7 1 1.3 7.2 In a family 12.1 14.7 0.3 3.6 Age ratio"* 23.8 21.4 11.6 10.8 A^b 405 302 319 305 ^ Percent juveniles within N. Number of focal individuals. ble 2). In contrast, focal Ross’s Geese engaged in intra- and interspecific social encounters with equal frequency in 2003-2004, but had 3 times more interspecific social encounters than intraspecific social encounters in 2002- 2003 (Table 2). Focal Snow Geese were more likely to win social encounters with Ross’s Geese than with other Snow Geese (Table 2). Focal Snow Geese were relatively more suc- cessful in interspecific social encounters; Snow Geese won 30 of 52 social encounters in 2002-2003, and 32 of 33 social encounters in 2003-2004 (Table 2). Snow Geese won 63 of 87 (72.4%) interspecific social encounters observed when all focal observations of both species were combined. Focal birds of both species were more successful in intraspecific social encounters in 2003-2004 than in 2002- 2003 (Table 2). Overall, focal Snow Geese lost only 10 so- cial encounters against Ross’s Geese; all Ross’s Goose wins were against lower ranked Snow Geese (i.e., non-parental pairs and lone birds); six were against lone juvenile Snow Geese, three were against lone adult Snow Geese, and one win was against an adult pair. Focal Ross’s Geese did not win social en- counters against Snow Geese in families. DISCUSSION Family Maintenance. — Our data from Lou- isiana, combined with that from other loca- tions, indicate Ross’s Geese maintain families for shorter periods than Snow Geese through- out their current wintering ranges. Timing of family break-up is known to vary among spe- cies, individuals, and years (Prevett and Mac- Innes 1980, Black et al. 2007). Eighty percent of all juvenile Snow Geese wintering in the Mississippi Flyway are in families from 20 December to 15 March, whereas <50% of all goslings are in families on staging areas in late March and early April (Prevett and Mac- Innes 1980). Generally, over 50% of all gos- ling Barnacle Geese {Branta leucopsis) leave their parents by December and <20% remain with their parents until April; however, this distribution varies among years (Black et al. 2007). We observed more families in winter 2003-2004; family breakup on our study area TABLE 2. Frequencies of social encounters (n/hr) of focal Lesser Snow Geese and Ross’s Geese, and odds of their success in social encounters in the rice prairies of southwest Louisiana during winters, 2002—2003 and 2003-2004. Lesser Snow Geese Ross’s Geese Types and success of social encounters 2002-2003 2003-2004 2002-2003 2003-2004 Intraspecific social encounters/hr Percentage of intraspecific social encounters won Interspecific social encounters/hr Percentage of interspecific social encounters won Odds of interspecific success^ 0.9 1.0 0.2 0.5 27.8 45.1 50.0 72.0 0.3 0.1 0.6 0.5 57.6 97.0 42.4 3.0 3.53 39.36 0.74 0.01 “ Odds of interspecific success = odds of winning against other species/odds of winning against own species. Jonsson and Afton • MIXED FLOCKS AND FAMILY MAINTENANCE IN WINTERING GEESE 729 may have occurred earlier for both species in winter 2002-2003 than in winter 2003-2004. Interspecific Dominance and Mixed Flocks. — Snow Geese were dominant over Ross’s Geese, as indicated by their relatively higher odds of winning against Ross’s Geese. We did not observe a Ross’s Goose win a so- cial encounter against a Snow Goose in a fam- ily group. The relatively higher success of Snow Geese in interspecific encounters in 2003-2004 corresponded to a higher frequen- cy of families in that year. Effects of species and family maintenance on outcomes of social encounters probably are confounded; Snow Geese may be more successful in interspecific social encounters because they maintain fam- ilies. Similarly, family maintenance, rather than species or body size may explain differ- ent time-budgets of the two species (Jonsson and Afton 2008). Single-species flocks are known for both species in other locations (Johnson and Raveling 1988, Ryder and Ali- sauskas 1995, Mowbray et al. 2000), but we only observed one single-species flock of Ross’s Geese, in winter 2003-2004 (Jonsson 2005). Predator vigilance (via “many eyes” and dilution effect) probably is an important ben- efit of mixed flocking in both species (Kris- tiansen et al. 2000, Krause and Ruxton 2002, Beauchamp 2003, Randier 2004). Predators may find it increasingly difficult to select prey, when prey choice requires choice of differing types and each prey type has differing cost- benefit relationships for the predator (confu- sion effect; Sinclair 1985, FitzGibbon 1990, Krause and Ruxton 2002). Snow Geese may have relatively better predator detection ca- pacities because they are taller (cf. Randier 2004) and may have greater visual acuity, giv- en acuity is positively correlated with eye size, which scales positively with body size (Fernandez-Juricic et al. 2004). We often ob- served Red-tailed Hawks [Buteo jamaicensis) fly by the geese, causing them to respond by becoming alert. Bald Eagles {Haliaeetiis leii- cocephalns) also attacked goose Hocks (Jons- son 2005). Ross’s Geese within mixed flocks evade predators by (1) remaining close to Snow Geese, thus, exposing Snow Geese lo avian predators polenlially chasing Ross’s Geese (cf. Sinclair 1985, FitzGibbon 1990), and (2) re- maining well within flock boundaries (JEJ, pers. obs.; R. C. Drewien, pers. comm.). Thus, dominance of Snow Geese seemingly does not drive Ross’s Geese towards flock edges; in- dividuals on flock edges often are subordi- nates (Black et al. 1992). Mixed flocks probably have been common throughout evolutionary history of these goose species; recent genetic studies show that gene flow is frequent between Snow Geese and Ross’s Geese over historical time (Weekstein et al. 2002). New pairs are formed on wintering grounds or during spring migra- tion (Ganter et al. 2005); thus, the two species exchange genetic material via mutual winter- ing areas (Mowbray et al. 2000). We suggest that along with predation pressure (cf. Mc- Williams et al. 1994), the historical associa- tion of Ross’s Geese with Snow Geese, along with the former’s associated subordinate so- cial status, has selected against family main- tenance in Ross’s Geese. However, we ob- served a small proportion of Ross’s Geese in families each winter. Thus, family mainte- nance probably represents individual choice within a species, rather than a species-fixed evolutionary constraint (Black et al. 2007). ACKNOWLEDGMENTS Our study was funded by the Canadian Wildlife Ser- vice, Louisiana Department of Wildlife and Fisheries (LDWF), Delta Waterfowl Foundation, Rockefeller Scholarship program, a Research Partnership Proposal (RPP) Grant from Cameron Prairie National Wildlife Refuge (NWR) and the U.S. Fish and Wildlife Service, and by the U.S. Geological Survey-Louisiana Coop- erative Fish and Wildlife Research Unit, Graduate School, School of Renewable Natural Resources at Louisiana State University (LSU), and LSU AgCenter. Rockefeller State Wildlife Refuge, Cameron Prairie NWR, and LDWF provided housing and valuable lo- gistical support. We thank C. J. Michie, Brandt Meix- ell, M. G. Pollock, T. W. Blair, and J. M. Yurek for help with data collection, and D. C. Blouin and M. D. Kaller for advice on statistical analyses. We thank R. N. Helm, Guthrie Perrie, T. J. Hess, J. T. Linscombe, and Daniel Gary for valuable assistance with our proj- ect. We acknowledge D. G. Homberger, W. G. Henk, M. J. Chamberlain, S. R. McWilliams, and an anony- mous referee for suggestions that improved this paper. LITERATURE CITED A(iKi:sri. A. 1996. An introduction lo categorical data analysis, .lohn Wiley and Sons, New Yt)rk, USA. Ausauskas, R. T. 1998. Winter range expansion and relationships between landscape and morphomet- 730 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 rics of midcontinent Lesser Snow Geese. Auk 1 15:851-862. Alisauskas, R. T, C. D. Ankney, and E. E. Klaas. 1988. Winter diets and nutrition of midcontinental Lesser Snow Geese. Journal of Wildlife Manage- ment 52:403-414. Altmann, J. 1974. Observational study of behaviour: sampling methods. Behaviour 49:227-267. Bateman, H. A., T. Joanen, and C. D. Stutzenbaker. 1988. History and status of midcontinent Snow Geese on their Gulf Coast winter range. Pages 495-515 in Waterfowl in winter (M. W. Weller, Editor). University of Minnesota Press, Minne- apolis, USA. Beauchamp, G. 2003. Group-size effects on vigilance: a search for mechanisms. Behavioural Processes 63:1 1 1-121. Black, J. M. and M. Owen. 1989a. Agonistic behav- iour in Barnacle Goose flocks: assessment, in- vestment and reproductive success. Animal Be- haviour 37:199-209. Black, J. M. and M. Owen. 1989b. Parent-offspring relationships in wintering Barnacle Geese. Animal Behaviour 37:187-198. Black, J. M., C. Carbone, and M. Owen. 1992. Lor- aging dynamics in goose flocks: the cost of living on the edge. Animal Behaviour 44:41-50. Black, J. M., J. Prop, and K. Larsson. 2007. Wild goose dilemmas. Branta Press, Groningen, The Netherlands. Boyd, H. 1953. On encounters between wild White- fronted Geese in winter flocks. Behaviour 5:85- 129. Calder III, W. A. 1996. Size, function and life history. Second Edition. Dover Publications, Mineola, New York, USA. Demment, M. W. and P. j. Van Soest. 1985. A nutri- tional explanation for body-size patterns of rumi- nant and nonruminant herbivores. American Nat- uralist 125:641-672. Lernandez-Juricic, E., J. T. Erichsen, and A. Kacel- NiK. 2004. Visual perception and social foraging in birds. Trends in Ecology and Evolution 19:25- 31. Lernandez-Juricic, E., G. Beauchamp, and B. Bas- TAIN. 2007. Group-size and distance-to-neighbor effects on feeding and vigilance in Brown-headed Cowbirds. Animal Behaviour 73:771-778. FitzGibbon, C. D. 1990. Mixed species grouping in Thompson and Grant gazelles: the anti-predator benefits. Animal Behaviour 39:1 116-1 126. Fox, A. D. AND J. Madsen. 1981. The pre-nesting be- haviour of the Greenland White-fronted Goose. Wildfowl 32:48-54. Fronczak, D. 2004. Waterfowl harvest and population survey data. USDI, Fish and Wildlife Service, Di- vision of Migratory Bird Management, Columbia, Missouri, USA. Ganter, B., W. S. Boyd, V. V. Baranyuk, and F. Cooke. 2005. First pairing in Snow Geese Anser caerulescens: at what age and at what time of year does it occur? Ibis 147:57-66. Gawlik, D. E. 1994. Competition and predation as processes affecting community patterns of geese. Dissertation. Texas A&M University, College Sta- tion, USA. Goudie, R. I. AND C. D. Ankney. 1986. Body size, activity budgets, and diets of sea ducks wintering in Newfoundland. Ecology 67:1475-1482. Gregoire, P. E. and C. D. Ankney. 1990. Agonistic behavior and dominance relationships among Lesser Snow Geese during winter and spring mi- gration. Auk 107:550-560. Jarman, P. J. 1974. The social organization of antelope in relation to their ecology. Behaviour 48:215- 267. Johnson, J. C. and D. G. Raveling. 1988. Weak fam- ily associations in Cackling Geese during winter: effects of body size and food resources on goose social organization. Pages 71-89 in Waterfowl in winter (M. W. Weller, Editor). University of Min- nesota Press, Minneapolis, USA. JONSSON, J. E. 2005. Effects of body size on goose behavior: Lesser Snow Geese and Ross’s Geese. Dissertation. Louisiana State University, Baton Rouge, USA. JoNSSON, J. E. AND A. D. Afton. 2006. Different time and energy budgets of Lesser Snow Geese in rice prairies and coastal marshes in southwest Louisi- ana. Waterbirds 29:451-458. JONSSON, J. E. AND A. D. Afton. 2008. Time budgets of Snow Geese and Ross’s Geese in mixed flocks: implications of body size, ambient temperature and family associations. Ibis 150:In press. JoNssoN, J. E., A. D. Afton, and R. T. Alisauskas. 2007. Does body size influence nest attendance? A comparison of Ross’s Geese {Chen rossii) and the larger, sympatric Lesser Snow Geese (C. ca- erulescens caerulescens). Journal of Ornithology 148:549-555. Kalmbach, E. 2006. Why do goose parents adopt un- related goslings? A review of hypotheses and em- pirical evidence, and new research questions. Ibis 148:66-78. Kendeigh, S. C. 1970. Energy requirements for exis- tence in relation to size of a bird. Condor 72:60- 65. Krause, J. and G. D. Ruxton. 2002. Living in groups. Oxford series in ecology and evolution. Oxford University Press, Oxford, United Kingdom. Kristiansen, J. N. and N. S. Jarrett. 2002. Inter-spe- cific competition between Greenland White-front- ed Geese Anser albifrons flavirostris and Canada Geese Branta canadensis interior moulting in West Greenland: mechanisms and consequences. Ardea 90:1-13. Kristiansen, J. N., A. D. Fox, H. Boyd, and D. A. Stroud. 2000. Greenland White-fronted Geese Anser albifrons flavirostris benefit from feeding in mixed-species flocks. Ibis 142:139-158. Loonen, M. j. j. E., L. W. Bruinzeel, J. M. Black, Jonsson ami Aft on • MIXED FLOCKS AND FAMILY MAINTENANCE IN WINTERING GEESE 731 AND R. Drent. 1999. The benefit of large broods in Barnacle Geese: a study using natural and ex- perimental manipulations. Journal of Animal Ecology 68:753-768. Madsen, J. 1985. Habitat selection of farmland geese in West Jutland, Denmark: an example of a niche shift. Ornis Scandinavica 16:140-144. McWilliams, S. R. and D. G. Raveling. 1998. Hab- itat use and foraging behavior of Cackling Canada and Ross’s geese during spring: implications for the analysis of ecological determinants of social behavior. Pages 167-178 in Biology and manage- ment of Canada Geese (M. D. Samuel, Editor) Proceedings of the International Canada Goose symposium, Milwaukee, Wisconsin, USA. McWilliams, S. R., J. P. Dunn, and D. G. Raveling. 1994. Predator-prey interactions between eagles and Cackling Canada and Ross’s geese during winter in California. Wilson Bulletin 106:272— 288. Mowbray, T. B., F. Cooke, and B. Ganter. 2000. Snow Goose {Chen caenilescens). The birds of North America. Number 514. Prevett, j. P. and C. D. MacInnes. 1980. Family and other social groups in Snow Geese. Wildlife Monographs 71. Randler, C. 2004. Coot benefit from feeding in close proximity to geese. Waterbirds 27:240-244. Raveling, D. G. 1970. Dominance relationships of ag- onistic Canada Geese in winter. Behaviour 37: 291-319. Ryder, J. P and R. T. Alisauskas. 1995. Ross’s Goose (Chen rossii). The birds of North America. Num- ber 162. SAS Institute. 1999. SAS/SYSTAT users guide. Ver- sion 8. SAS Institute, Cary, North Carolina, USA. Shelley, E. L., M. Y. U. Tanaka, A. R. Ratnathicam, AND D. T. Blumstein. 2004. Can Lanchester’s laws help explain interspecific dominance in birds? Condor 106:395-400. Sinclair, A. R. E. 1985. Does interspecific competi- tion or predation shape the African ungulate com- munity? Journal of Animal Ecology 54:899-918. Stahl, J., P. H. Tolsma, M. J. J. E. Loonen, and R. H. Drent. 2001. Subordinates explore but domi- nants profit: resource competition in high Arctic Barnacle Goose flocks. Animal Behaviour 61: 257-264. Weckstein, j. D., a. D. Alton, R. M. Zink, and R. T. Alisauskas. 2002. Hybridization and popula- tion subdivision within and between Ross’s Geese and Lesser Snow Geese: a molecular perspective. Condor 104:432-436. The Wilson Journal of Ornithology 1 20(4);732-742, 2008 NESTING ECOLOGY OF COMMON GOLDENEYES AND HOODED MERGANSERS IN A BOREAL RIVER SYSTEM HELENE SENECHAL,' GILLES GAUTHIER,' ^ AND JEAN-PIERRE L. SAVARD^ ABSTRACT. — Common Goldeneyes (Bucephala clangula) and Hooded Mergansers (Lophodytes cucullatus) are common cavity-nesting ducks but the importance of fast-flowing rivers as suitable nesting habitat may have been overlooked. We monitored the use of >90 nest boxes installed along a boreal forest river over a 5-year period. A high nest box occupancy rate was reached in the second year (40%) and was maintained thereafter (48 to 55%). On average, 35 nest boxes were occupied by goldeneyes and 1 1 by mergansers each year. Laying date was similar between the two species but merganser nests hatched slightly later. Both species had similar clutch sizes but merganser nests contained more eggs than goldeneye nests when heterospecific parasitic eggs were included. On average, 16% of goldeneye nests were parasitized by mergansers, and 49% of merganser nests were parasitized by goldeneyes. Density of suitable natural cavities in the area was relatively low suggesting the high occupancy rate of nest boxes may be a response to lack of suitable cavities. Nest box use was positively related to the total surface area of ponds in the vicinity and negatively to distance to the river. Use of nest sites along fast-flowing rivers appears to be an opportunistic strategy and may be dependant on the presence of nearby ponds and lakes. Received 1 October 2007. Accepted 23 February 2008. Selection of a suitable nest site is of critical importance for breeding birds because it can influence their reproductive success. A good nest site must be safe from predators, and pro- vide shelter from bad weather and extreme en- vironmental conditions. Ideally, it should also be in an area with abundant food resources. A strategy used by many bird species is to nest in tree cavities. This strategy benefits these species by providing a favorable microclimate for eggs and young, and decreases nest pre- dation risks (Nilsson 1986). Common Goldeneye {Bucephala clangula) and Hooded Merganser {Lophodytes cuculla- tus) are secondary cavity-nesting ducks that breed in the boreal forest (Dugger et al. 1994, Eadie et al. 1995). These species are relatively large and require large cavities, which repre- sent only a small proportion of all cavities available in the boreal forest (Gauthier and Smith 1987). Intra- and interspecific compe- tition for these large cavities is therefore se- vere (Erskine 1990, Semel and Sherman 2001). Selection of a suitable nest site may be especially critical in regions of intensive for- ‘ Departement de Biologic and Centre d’etudes Nor- diques. Pavilion Vachon, 1045 Avenue de la Medeci- ne, Universite Laval, Quebec, PQ GIV 0A6, Canada. - Science and Technology, Environment Canada, 1141, Route de I’Eglise, C.R 10100, 9"' etage, Quebec, PQ, GIV 4H5, Canada. Corresponding author; e-mail; gilles.gauthier@bio.ulaval.ca est management because forestry practices can reduce the availability of suitable cavities. Many studies have shown that addition of nest boxes results in increases in breeding popu- lations of secondary cavity-nesters (Savard 1988, Newton 1994, Poysa and Poysa 2002) suggesting that nest site availability was lim- iting these populations at least locally. Common Goldeneyes and Hooded Mergan- sers commonly breed on ponds and lakes of the boreal forest. Fast-flowing rivers, although abundant throughout the boreal forest, are generally not considered an important breed- ing habitat for these species (Eadie et al. 1995). Nest site selection of Common Gold- eneyes has been previously studied in lake systems (Lumsden et al. 1980, 1986), but the nesting biology of these species in river sys- tems is little known. The overall objective of this study was to evaluate the potential of riv- ers as nesting habitat for Common Golden- eyes and Hooded Mergansers in the boreal forest. We studied the nesting ecology of both species using nest boxes in a fast-flowing river system in the Quebec boreal forest. We also assessed the potential for nest site competition between both species and investigated wheth- er occupancy of nest boxes along the river could be associated with a scarcity of natural cavities in the area. Finally, we identified river characteristics and components of the riparian habitat influencing nest site selection and nest- ing success of these species in this habitat. 732 Senechal et al. • NESTING ECOLOGY OF GOLDENEYES AND MERGANSERS 733 METHODS Study Area. — This study was conducted in the Ste. Marguerite River Valley (48°23'N, 70° 12' W), 200 km northeast of Quebec City, Canada. This is a meandering but fast flowing river typical of the boreal forest in Quebec (mean annual discharge = 58 mVsec). The river flows for —100 km in a 0.5- 1.0 km wide valley, often bordered by steep mountain slopes rising to 350 m above the valley bot- tom. About 53 ponds (including oxbows) oc- cur in the valley of the studied section and can be used by breeding adults or broods after hatch. The Ste. Marguerite River is a prime Atlantic salmon {Salmo salar) and brook trout {Salvelinus fontinalis) fishery managed by a private fishing association. Mature stands of balsam fir {Abies balsa- mea) and yellow birch {Betula alleghaniensis) dominate the forest of the valley. The forest in the upper half of the study area has re- mained almost untouched for at least 100 years and large yellow birch trees are abun- dant. However, clear-cutting occurred in the early 1980’s on about 6% of the catchment area, mainly in the lower half of the valley; this area still has some 80-year-old trees. Experimental Design. — Ninety-two nest boxes were erected in fall 1997 along a 50- km stretch of the river, encompassing a diver- sity of river habitats (rapids, fast current, slow current, meanders, etc.); 10 additional boxes were installed in fall 2000. The number of boxes varied annually from 82 to 93 because some nest boxes were damaged by falling trees or disappeared. Their distribution along the river followed fishing access paths in most cases but boxes were isolated from fishing pools to minimize human disturbance. Boxes were separated by at least 30 m and nailed to trees close to the water edge (range = 0 to 29 m) with the entrance hole facing the river, be- tween 2 and 5 m above the ground. Boxes were made of plywood (inside dimensions: 23 X 26 X 58 cm; entrance: 10 X 12 cm; depth below the entrance: 45 cm) and a 10-cm layer of wood shavings was added as nest material. We removed branches in front of the boxes to ensure visibility from the river and ease of entry for waterfowl. Boxes were cleaned at the end of each breeding season (abandoned eggs removed) and wood shavings were changed. Nest Box Monitoring and Measurement of Nesting Parameters. — Nest boxes were mon- itored from early May to early July weekly in 2001, three times in 2000 and 2002, and twice in 1998 and 1999. We recorded the number of eggs of each species and boxes containing eggs of more than one species were identified as interspecific nest parasitism. The species that first started the nest or, if unknown, the incubating species was considered the host. Eggs were numbered with a permanent mark- er, and measured (length and width, ±0.1 mm) with a caliper and weighed (±1 g) with a spring scale. These measurements were used to calculate the Maximum Euclidian Distance (MED), a criterion used to identify intraspe- cific nest parasitism (Eadie 1989, Poysa et al. 2001). Nests were classified as parasitized by conspecifics when at least one of the follow- ing criteria was met: (1) the MED among the eggs was >2.5, (2) clutch size was >12 eggs, (3) more than one egg was laid per day, and (4) eggs were laid two or more days after the start of incubation. More than one of these criteria was met in most nests where parasit- ism was detected. These criteria should yield a minimum estimate of nest parasitism (An- dersson and Ahlund 2001). We calculated the date on which the first egg of the clutch was laid (i.e., laying date) for nests discovered during laying by back- dating using the number of eggs of the host species at the time of discovery (maximum of 12 eggs; additional eggs were considered par- asitic) and assuming a mean laying rate of one egg every 1 .4 days. The laying date for nests discovered during incubation was backdated from the hatching date assuming an incuba- tion period of 30 days. We calculated hatching date when it was unknown (for some nests in 2000 and 2002), using the relationship be- tween incubation stage and egg density (Se- nechal 2003). We estimated the age of embry- os (Caldwell and Snart 1974) for deserted nests with unknown laying dates and back- dated the laying date. Clutch size was defined as the highest number of eggs recorded in an incubated nest. Nesting success was the proportion of suc- cessful nests (i.e., those in which at least one egg hatched). Deserted nests were those aban- 734 THE WILSON JOURNAL. OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 denned during laying or incubation with no signs of predation and were included in the calculation of nesting success. Hatching suc- cess, defined as the proportion of eggs that hatched within a clutch, was calculated for successful nests. Availahility of Nat u ml Cavities. — We esti- mated the availability of living trees and snags sufficiently large to support cavities and of cavities along the river using line transect sampling (Buckland et al. 1993). We posi- tioned 17, 300-m transects every 3 km at right angle to the river axis. We recorded the po- sition and perpendicular distance from the transect of all trees and snags >30 cm in di- ameter at breast height (DBH) encountered. We ascertained species, height, DBH, and presence of cavities with entrance diameter >8 cm (visual estimate from the ground) for these trees. We recorded the cavity type (hol- low top of standing trees, woodpecker hole, broken tree limbs, or crack on the side of a tree), its height above ground, an index of its accessibility from the air (under the canopy, in the canopy, over the canopy, or open for- est), and the approximate size of the entrance hole. Any signs of waterfowl use (e.g., pres- ence of down at the entrance) were noted. Cavities with entrances >8 cm, >1.8 m above ground, and easily accessible (i.e., above the forest canopy or in an open stand) were con- sidered suitable (Bergeron et al. 1997). Sam- pling was conducted in late May-early June and at the end of July 2001. Nest Site Characterization. — We sampled the habitat surrounding each nest box in late July-early August 2001. We measured river depth at the deepest point with a graduated rod, river width with a measuring tape (at times estimated from the shore), and surface water velocity at the deepest point using a portable current meter (Marsh McBirney, model 20 IM; Frederick, MD, USA). We com- bined velocity measurements and the general aspect of the river to create a new variable that categorized river sections as rapids (>60 cm/ sec with riffles at the surface), fast-flowing (>60 cm/sec, water surface still) or slow- flowing (<60 cm/sec). We sampled 12 terrestrial habitat variables at each nest box. These variables included: distance to water edge, to nearest lake or pond, to nearest used nest box, and to nearest tree >10 cm in diameter; entrance height above the river and above the ground; com- pass direction faced by the entrance (catego- rized as North, East, South, or West); and can- opy height. Distances to nearest pond or nest box were obtained from a digital map (scale: 1:20,000) whereas other distance variables were measured in the field. We also sampled density of trees with a DBH of >20 cm in a semi-circle centered in front of the nest box (11.3-m radius) and evaluated shrub height. Stems >4 m in height were defined as trees and those <4 m as shrubs. We also counted the number of lakes and ponds in the vicinity on the digital map, i.e., those included in a 2 X 6 km ellipse (long axis parallel to the river) centered at the nest box, and calculated the total surface area of these water bodies. Data Analysis. — Statistical analyses were completed with software SAS 8.0 (SAS Insti- tute 1999). Interactions in analyses with more than one factor were examined and removed from the model when non significant {P > 0.05). We examined the effects of species and year on laying date, hatching date, and clutch size with two-way ANOVAs. Dates were rank-transformed and we used the ART pro- cedure to test the interaction as recommended by Salter and Fawcett (1993). We examined the effects of year, species, and parasitic egg laying on nesting success using a log-linear analysis. We analyzed the effect of egg status (host species in a parasitized nest, heterospe- cific parasite in a parasitized nest, or in a nest non-parasitized by a heterospecific parasite), year, and species on hatching success using a three-way ANOVA on rank-transformed data. We examined the effects of species and year on the rate of parasitism using logistic regres- sions, and on the number of heterospecific parasitic eggs using two-way ANOVAs. We compared the laying date of parasitized and non-parasitized nests of each species using ANOVA. We estimated densities of living trees (>30 cm DBH), snags, and cavities with DIS- TANCE 3.5 (Buckland et al. 1993, Research Unit for Wildlife Assessment 1999). We com- pared DBH and height between snags and liv- ing trees using Mann-Whitney tests, and the proportion of trees with cavities using x“- examined if snag and living tree densities. Senechal et al. • NESTING ECOLOGY OE GOLDENEYES AND MERGANSERS 735 1998 1999 2000 2001 2002 EIG. 1. Nest boxes used by waterfowl in the Ste. Marguerite River Valley, Quebec, Canada, 1998-2002. Percentages are based on the total number of boxes used by waterfowl and numbers in parentheses are the annual number of boxes available. Data for 1998 and 1999 are from Nadia Aubin-Horth (pers. comm.). cavity abundance (because of the small num- ber of suitable cavities found), and tree char- acteristics (DBH, height) were related to dis- tance to the river (in 50-m classes) with linear regression. We generated three independent data sets to analyze characteristics of nest sites used by each species from 1999 to 2002. Boxes were classified whether they were used: ( 1 ) at least once versus not used, (2) at least twice versus not used, and (3) at least three times versus not used. Separate analyses with each data set generally yielded similar results (Senechal 2003), and we present only the results based on boxes used at least twice because this avoided using data from boxes rarely used (i.e., only once) while retaining sufficient sample size. The effect of the 13 quantitative habitat variables on nest site use was first an- alyzed using a stepwise discriminant analysis, followed by a logistic regression including variables retained by the first analysis {P < 0.15) and qualitative variables. We used the same procedure to test the relationship be- tween nesting success and habitat character- istics for goldeneyes (sample sizes were too small for mergansers). We combined data from 2000 to 2002 and, for nests used more than once, we assigned success or failure based on the occurrence that was most fre- quent, excluding those with the same number of successes and failures. RESULTS Nest Box Use and Nesting Parameters. — The number of nest boxes used by waterfowl was low the first year but increased in the sec- ond year (Fig. 1). Occupancy rate stabilized at 51-55% by the fourth year of the program. Goldeneye use of nest boxes outnumbered mergansers three to four times in all years but the first. Both species started their clutches at about the same time from 2000 to 2002 (F, 109 = 2.0, P = 0.17) with no differences between years (^2,109 1.8, F = 0.18). Overall, mean (±SE) laying date was 4 May ± 0.9 days (range = 14 Apr-24 May, n = 81) for goldeneyes and 7 May ± 1.9 days (21 Apr-5 Jun, n = 34) for mergansers. Goldeneye nests hatched slightly earlier than mergansers (F, qq = 5.8, P = 0.02). Overall, the mean hatching date was 15 Jun ± 0.9 days (31 May-2 Jul, n = 68) for goldeneyes and 20 Jun ± 1.9 (31 May- 13 Jul, n = 28) for mergansers. Mean (±SE) clutch size in incubated gold- eneye nests (9.8 ± 0.4 eggs, n — 78) was sim- ilar to that of mergansers (9.2 ± 0.5, n = 33; ^1.107 “ 0.8, P = 0.36) when eggs resulting from interspecific parasitism were excluded and did not vary among years (F2 107 = 0.2, P = 0.83; Table 1). The total number of eggs (i.e., including parasitic eggs) was higher for TABLE 1. Mean |± SE| clutch size (sample size) of the host species only and total clutch size (including interspecific parasitic eggs) in incubated nests of Common G(7ldeneyes and Hooded Mergansers, Ste. Marguerite River, Quebec, Canada, 2()()()-20()2. Host species clutch (inly Total clutch Species 2(K)0 2(K)I 2tK)2 2(KM) 2(K)I 2002 Common Goldeneye Hooded Merganser 9.2 ± ().."i (23) 10.0 ± 1.2 (10) 9.7 ± 0..S (31) 8.7 ± 0..3 (13) 10.3 ± 0.8 (24) 8.9 ± 0.7 (10) 9.8 ± O.b (23) 12.7 ± 0.9 (10) 10.0 ± 0.6 (31) 1 1 ..‘i ± 1 .0 (13) 1 1.0 ± 0.9 (24) 10.4 ± 1.2 (10) 736 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 TABLE 2. Percent nesting success (sample size) of Common Goldeneyes and Hooded Mergansers in nest boxes, Ste. Marguerite River, Quebec, Canada, 2000-2002. Parasitized nests® Non-parasitized nests All nests Host species 2000 2001 2002 2000 2001 2002 2000 2001 2002 Common Goldeneye Hooded Merganser 86 (7) 67 (15) 85 (13) 100 (8) 88 (8) 75 (4) 79 (24) 64 (22) 100 (2) 67 (6) 50 (24) 67 (6) 81 (31) 100 (10) 65 (37) 79 (14) 62 (37) 70 (10) Intraspecific and interspecific nest parasitism. merganser nests (11.6 ± 0.6) than for gold- eneye nests (10.2 ± 0.4; = 3.7, P = 0.06). Nesting success of both species decreased between 2000 and 2002 (x^2 ~ 16.1, P < 0.01), and was lower for goldeneyes than for mergansers (x^i — 3.3, P = 0.07; Table 2). Non-parasitized nests were less successful than those parasitized (64 vs. 82%, x^i = 8.8, P < 0.01). Desertion was the main cause of nesting failures of goldeneyes (54%, n = 28). Other causes included desertion due to mark- ing (18%), death of incubating females (7%), clutch predation (7%), other human distur- bance (7%), and unknown (4%). The main cause of nest failure for mergansers was de- sertion due to human disturbance other than marking (33%, n = 6). Other causes included clutch predation (17%) and unknown (50%). Overall, heterospecific parasitic eggs did not fail more often than eggs from host spe- cies or eggs in non-parasitized nests (F2 107 = 0.2, P = 0.80; Table 3). Goldeneyes and mer- gansers had similar hatching success (90 vs. 88% overall, Fj ,07 = 0.4, P — 0.56) and there were no differences between years (F2 107 = 2.3, P- 0.11). Pattern of Nest Parasitism. — Interspecific nest parasitism started in the second year (1999) of the nest box program and appeared to increase with the proportion of nest boxes used (Fig. 2; year effect: x^3 ^ 5.6, P = 0.13). A much greater proportion of merganser nests were parasitized by goldeneyes than the re- verse (x^i = 8.6, P < 0.01). Parasitized mer- ganser nests also contained on average more goldeneye eggs (4.7 ± 0.6 eggs, n - \1) than the reverse (2.8 ± 0.3, n = \6\ Pj 29 = 7.2, P = 0.01). Parasitism by conspecifics was also com- mon in both species (species effect: x^i = 0-5, P = 0.49), ranging from 9 to 32% of gold- eneye and 7 to 30% of merganser nests during 2000-2002 (year effect: x"2 = 3.2, P = 0.20). Among the total number of nests initiated by each species (i.e., nests where each species laid at least 1 egg), mergansers tended to lay more often as a parasite (intra- and interspe- cific egg laying combined) than goldeneyes (X^ = 3.2, P = 0.07; Fig. 3). On average, 25 ± 5% (range = 18-35%) of all nests contain- ing goldeneye eggs were associated with par- asitic events by this species compared to 39 ± 8% (22-50%) for mergansers. Nests para- sitized by goldeneyes (intra- and interspecifi- cally) were initiated earlier (30 Apr ± 1.3 days, n — 31) than non-parasitized nests (6 May ± 1.1, n - 50; = 9.1, P < 0.01); this was not the case for mergansers (P129 = 2.2, P = 0.15). TABLE 3. Mean [± SE] percent hatching success (sample size) of Common Goldeneyes and Hooded Mer- gansers in successful nest boxes, Ste. Marguerite River, Quebec, Canada, 2000-2002. Parasitized nests® Host species eggs Heterospecific parasitic eggs Non-parasitized nests Species 2000 2001 2002 2000 2001 2002 2000 2001 2002 Common 84 ± 16 83 ± 4 79 ± 8 100 76 ± 10 67 ± 17 97 ± 2 91 ± 2 96 ± 3 Goldeneye (5) (2) (7) (6) (7) (2) (17) (22) (15) Hooded 85 ± 9 85 ± 9 100 93 ± 7 83 ± 17 87 ± 10 90 ± 5 89 ± 8 90 ± 5 Merganser (6) (7) (2) (5) (2) (7) (4) (4) (5) ® Interspecific nest parasitism. Senechal et al. • NESTING ECOLOGY OF GOLDENEYES AND MERGANSERS 737 S 70 1 60 2 50 CO Q. O 40 o o 30 CO B 20 c o 10 0) ■co PI QC 0 FIG. 2. Common Goldeneye and Hooded Mergan- ser nests parasitized by a heterospecific, Ste. Margue- rite River, Quebec, Canada, 1998-2002. Numbers in parentheses are number of nests. Common Goldeneye (10) Hooded Merganser (7) (30) (2) (2) (27) 1(37) Am (14) (10) 1998 1999 2000 2001 2002 Availability and Suitability of Natural Cav- ities.— Of 688 trees (> 30 cm DBH) recorded along transects, 17% were snags and 83% were alive. Large snag density was estimated at 6.5 ± 0.8 snags/ha, almost six times lower than large living trees (36.6 ± 2.0 trees/ha). Snags were slightly larger (mean DBH: 42.4 ± 0.8 cm, n = 114) than living trees (40.4 ± 0.4 cm, n = 574; U = 5.9, P = 0.02), but shorter (7.8 ± 0.7, ^ = 45 vs. 18.1 ± 0.3 m, n = 204, respectively; U = 82.6, P < 0.01). Suitable cavities were present in 6% of the snags (n = 114; 1 with 2 cavities) but were rarely found in living trees (\%, n = 574; x^i = 32.3, P < 0.01). We found 16 suitable cav- ities during sampling for an estimated density of 0.92 ± 0.31 cavity/ha, but none had signs of waterfowl use. Thirty-seven cavities suffi- ciently large for ducks were found (1.71 ± 0.37 cavities/ha) when the accessibility crite- rion was not considered. Pileated Woodpecker (Dryocopus pileatus) holes were the most abundant cavity type (50%), followed by chimney cavities (38%) and natural cracks (12%). Most suitable cavities were in snags (69%) at a mean height of 6.2 ± 1.0 m {n = ID- The dominant species of large trees were yellow birch (57%), balsam fir (19%), eastern white cedar {Thuja occidental is\ 7%), and white spruce (Picea glauca\ 6%). Black ash (Fraxinus nigra), eastern white pine (Pinus strobus), paper birch {Petula papyrifera), trembling aspen (Populus treniuloides), bal- sam poplar {P. balsamifera), and white elm 80 ^ mtm Common Goldeneye 'T I. .1 Hooded Merganser 2000 2001 2002 FIG. 3. Nests that contained eggs that were laid parasitically (either intra- or interspecifically) by Com- mon Goldeneyes among all nests initiated by this spe- cies (i.e., nests where goldeneyes laid at least 1 egg), and vice-versa for Hooded Mergansers, Ste. Margue- rite River, Quebec, Canada, 2000-2002. Numbers in parentheses are number of nests. {Ulmus americana L.) each accounted for <3% of the large trees. The 16 suitable cav- ities were mainly in yellow birch (81%); the rest were equally distributed among elm, ash, and fir (6% each). Yellow birch was the only tree species for which density increased with distance to the river up to 300 m ((3 = 3.73 ± 0.47 trees/ha/ 50 m, = 0.94, P < 0.01, n = 6). Density of large snags decreased away from the river ((3 = —0.87 ± 0.37 snags/ha/50 m, r- - 0.58, P = 0.01). There was no relationship between the abundance of suitable cavities and dis- tance to the river {r- = 0.17, P = 0.41), nor between tree height or DBH and distance to the river {P > 0.12). Nest Site Characteristics. — Only 19 of all nest boxes available were not used. Boxes in the upper half of the river were generally used more often than those in the lower half. Few habitat variables explained nest box use by ducks. Boxes used by goldeneyes had a great- er total surface area of ponds in the vicinity (40.7 ± 6.1 ha) than those not used (8.3 ± 4.0 ha) and goldeneyes used boxes predomi- nantly oriented toward the south (Table 4). Boxes not used by this species were usually farther from water than those used (9.4 ± 1 .2 vs. 6.2 ± 0.7 m, respectively). Used boxes were farther from trees >10 cm (2.4 ± 0.3 m) than those not used (1.5 ± 0.2 m). Goldeneye nesting success was affected mainly by the to- tal surface area of ponds in the vicinity (x^i = 738 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 TABLE 4. Characteristics explaining use by Com- mon Goldeneyes and Hooded Mergansers for nest box- es used at least 2 years, Ste. Marguerite River, Quebec, Canada, 1998-2002. ‘ + ’ indicates a positive effect and ‘ — ’ indicates a negative effect. Numbers in parenthe- ses are the number of used vs. unused boxes. Species Variables^ Effect p Common SURFPOND + 0.003 Goldeneye ORIEN s+ 0.047 DISWAT 0.004 DISTREE • (37 vs. 27) 0.032 Hooded CANOP -b <0.001 Merganser DISTREE + (8 vs. 55) 0.006 ^ SURFPOND: total surface area of ponds. ORIEN: compass direction faced by the nest box entrance (S = South). DISWAT: distance to the water edge. DISTREE: distance to the nearest tree >10-cm diameter, and CAN- OP: canopy height. 13.2, P < 0.01); failed nests were surrounded by a larger surface of ponds than those that were successful (61.6 ± 15.0 ha, n = 14 vs. 35.5 ± 7.9, n = 36, respectively). Boxes used by mergansers were also farther from the nearest tree >10 cm in diameter than those not used (3.3 ± 0.7 vs. 2.0 ± 0.2 m), and tree canopy around used boxes (21.8 ± 2.4) was higher than around boxes not used (14.5 ± 0.7, Table 4). DISCUSSION Nesting Biology. — Common Goldeneye and Hooded Merganser nesting biology in river habitat appears similar to that in the more usu- al lacustrine habitats of northeastern North America. Nest initiation dates (16 Apr-27 May) reported for goldeneyes in Ontario (Mallory et al. 1994, Eadie et al. 1995) are similar to those in our study. Mean Clutch siz- es reported in Ontario (7.4-9.4 eggs). New Brunswick (9.0 eggs), and Minnesota (10.2 eggs) (Eadie et al. 1995) are also similar. Nesting success in our study was higher than reported by Bouvier (1974) in southern Que- bec (57%) but lower than observed in Maine (87%) (Allen et al. 1990). Clutch size of Hooded Mergansers usually ranges from nine to 13 eggs and nesting success ranges from 66 to 82% (Dugger et al. 1994, Mallory et al. 2002). This suggests that nesting in habitats adjacent to fast-flowing rivers does not im- pose severe constraints on nesting phenology or success for these species. The similarity in the observed breeding phenology of Common Goldeneyes and Hooded Mergansers is not surprising given the similarities in life history traits of the two species (Dugger et al. 1994, Eadie et al. 1995). The few merganser nests initiated after 25 May were possibly renesting attempts or nests of hrst-time breeders that nest later than older, experienced females (Dow and Fredga 1984). Desertion was a major cause of nesting fail- ure for goldeneyes. This may be due to a high proportion of hrst-time breeders in the area, as inexperienced birds are more prone to de- sertion (Eriksson and Andersson 1982, Gau- thier 1989). A high proportion of hrst-time breeders may result from a recent increase in the breeding population possibly due to in- stallation of nest boxes. Human disturbance may also be a cause of nest desertion of gold- eneyes. This species shows relatively weak nest attentiveness and defense relative to Hooded Mergansers, hushing more readily when disturbed (Mallory et al. 1993a, 1998). Greater sensitivity to disturbance may contrib- ute to the high rate of nest abandonment of goldeneyes. It is also possible that more fre- quent visits to nest boxes in 2001, and more capture and handling of females, contributed to high nest desertion that year. Nest parasitism often has a negative impact on nesting success of waterfowl due to dis- turbance by parasitic females (Semel et al. 1988, 1990). This was not the case in our study, as parasitized nests generally had great- er nesting success than non-parasitized nests (also reported by Eadie 1989). Large clutches resulting from nest parasitism at times also have reduced hatching success (Mallory et al. 2002), but normal hatching success has also been reported (Eriksson 1979, Eriksson and Andersson 1982, McNicol et al. 1997). Large clutches (>20 eggs) were rare in our study, which may explain why parasitism had little effect on hatching success. Nest Parasitism and Nest Site Competi- tion.— Intra- and interspecihc nest parasitism were both common in our study. The frequen- cy of conspecihc egg laying is likely a mini- mum value because Andersson and Ahlund (2001) showed with biochemical techniques that it is underestimated with traditional meth- ods. Our results support the suggestions of Ea- die (1989), Semel et al. (1988), and Haramis Senechal et al. • NESTING ECOLOGY OE GOLDENEYES AND MERGANSERS 739 and Thompson (1985) that high breeding den- sities of cavity-nesting waterfowl and relative scarcity of suitable nest sites may contribute to nest parasitism. Almost half of our nest boxes were unused, apparently suggesting that nest sites were not limiting. However, this as- sumes that all nest boxes and sites selected for their placement were of equal quality, which is unlikely to be true. We found evidence that goldeneyes preferred some boxes over others. Moreover, boxes were distributed over 50 km of river that may vary greatly in quality for ducks. For example, boxes in the lower half of the river received relatively little use. Thus, even though only 55% of the boxes were used, there could be competition for high quality nest sites. Merganser nests were more likely to be par- asitized by goldeneyes and received a larger num.ber of parasitic eggs than goldeneye nests, as reported in other studies (Bouvier 1974, Mallory et al. 1993a). The greater proportion of parasitized merganser nests may be related to the smaller size of the merganser popula- tion rather than a lower propensity for para- sitic egg-laying. Mergansers tended to lay more often as a parasite than goldeneyes (also reported by Mallory et al. 1993a) when we compared the number of nests where one spe- cies laid parasitically in relation to the number of nests initiated by that species. Intraspecihc nest parasitism in cavity-nesting ducks may be an adaptive strategy (Semel et al. 1988, Poysa 1999, Andersson 2001, Ahlund 2005), but it can also be the outcome of two individuals inadvertently choosing the same nest site to lay their eggs (Erskine 1990). Semel and Sherman (2001) supported this hypothesis for Wood Ducks {Aix sponsci) and interspecihc nest parasitism is perhaps more likely to result from such nest site competition. Hooded Mer- gansers are probably dominated in competi- tive interactions by the larger, more aggressive goldeneye. Thus, they may be more easily evicted from a nest box, giving the impression that mergansers “parasitize” more often than goldeneyes. Availability and Suitability of Natural Cav- ities.— No breeding pair surveys were con- ducted in the area before nest boxes were in- stalled and we do not know with certainty if high use of boxes by goldeneyes and mergan- sers reflects a real increase of their local pop- ulations, as reported for other populations (Fredga and Dow 1984, Savard 1988, Poysa and Poysa 2002). One explanation for the in- crease in nest box use may be that local fe- males moved from nearby natural cavities to the more conspicuous nest boxes near the riv- er (Gauthier and Smith 1987). Our sampling showed that density of suitable cavities was low in our study area considering the high density of large snags and living trees present. Our density estimate for natural cavities may also be a maximum value as we were not able to inspect the interior of cavities. Similar cav- ity densities have been reported elsewhere (Senechal 2003), but the density was lower than expected for a mature forest (Prince 1968, Gilmer et al. 1978). None of the cavities found appeared to be used by waterfowl. Suit- able cavities may have been beyond our 300- m transects, but steep slopes or cliffs were of- ten encountered beyond this distance, which would reduce their attractiveness because of predation risks associated with long and dif- ficult overland travel by broods. Common Goldeneyes appear flexible in the type of cavity used. Carter (1958) mentioned they use mainly lateral openings in New Brunswick, whereas Prince (1968) identified chimney cavities as the main type used. Mai- sonneuve et al. (2002) reported that half of Hooded Merganser and Common Goldeneye nests found in Quebec boreal forest were in old Pileated Woodpecker cavities. Abandoned Pileated Woodpecker holes in our study were the main source of suitable cavities and this woodpecker may be a key species for cavity- nesting ducks (Bonar 2000). Nest Box Use ami Habitat Characteris- tics.— The preference of goldeneyes for nest boxes in areas with a high surface area of ponds suggests that many pair territories were on ponds. Mallory et al. (1993b) reported goldeneyes preferred nest boxes in wetlands without fish isolated from other water bodies in acid-stressed areas of Ontario. Many lakes and ponds in our study area supported fish and were little acid-stressed, but may pro\ ide more invertebrate food than the river where Atlantic salmon were present. An abundance of ponds in the vicinity may also provide many potential brood-rearing sites, as ponds are important brood habitats tor females nest- ing along the ri\er (Senechal 2003). G(4den- 740 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 eyes preferred nest boxes closest to water, pre- sumably because proximity to water minimiz- es overland travel for broods after hatch (Morse et al. 1969, Poysa et al. 1999) and decreases predation risks. Nest boxes near wa- ter are also more conspicuous than those fur- ther in the forest and are more easily found by ducks. The preference for nest boxes far from large trees is another indication that goldeneyes prefer nesting in open habitats. Mergansers also selected nest sites in open habitats with a high canopy. The preference of goldeneyes for boxes with a southward ori- entation may be related to a more favorable micro-climate (i.e., away from prevailing westerly wind and toward the sun). No char- acteristics of the river influenced nest box use by either species, but this may not be surpris- ing if pair territories were mostly away from the river. Presence of a suitable cavity (nest box) along the river may be the main stimulus for nesting in this habitat. Few characteristics of the habitat surround- ing nest sites affected goldeneye nesting suc- cess. Surprisingly, goldeneye nests with a high surface area of ponds in their vicinity were less successful. Pond area positively influ- enced nest site use, and the attractiveness of nest sites in these areas may have increased disturbance from prospecting females (Eadie and Gauthier 1985) and may have decreased nesting success. Mallory et al. (1993b) re- ported that lake isolation was an important factor in nest site selection by goldeneyes in Ontario and suggested that high bird density may lead to interference in this highly terri- torial species (Savard 1984). Eadie (1989) found no differences in habitat characteristics between successful and failed Barrow’s {Bu- cephala islandica) and Common goldeneye nests in British Columbia. ACKNOWLEDGMENTS We acknowledge the Fondation de la faune du Que- bec, Fonds Quebecois de Recherche sur la Nature et les Technologies, Natural Sciences and Engineering Research Council of Canada, Canadian Wildlife Ser- vice, Ducks Unlimited, and the Centre Interuniversi- taire de Recherche sur le Saumon Atlantique for finan- cial support. We also thank Nadia Aubin-Horth, Dom- inic Savard, Gerald Picard, Gregory Bourguelat, Je- rome Leger, M.-C. Cadieux, Antoine Morrissette, J.-F Savard, Karine Plante, Catherine Bonenfant, and Felix Ledoux for help in the field, and Jean Huot and Cyrille Barrette for comments on the manuscript. LITERATURE CITED Ahlund, M. 2005. Behavioural tactics at nest visits differ between parasites and hosts in a brood-par- asitic duck. Animal Behaviour 70:433-440. Allen, R. B., P. O. Corr, and J. A. Dorso. 1990. Nesting success and efficiency of waterfowl using nest boxes in central Maine: a management per- spective. Pages 291-296 m Proceedings of the 1988 North American Wood Duck Symposium (L. H. Fredrickson, G. V. Burger, S. P. Havera, D. A. Graber, R. E. Kirby, and T. S. Taylor, Editors). St. Louis, Missouri, USA. Andersson, M. 2001. Relatedness and the evolution of conspecific brood parasitism. American Natu- ralist 158:599-614. Andersson, M. and M. Ahlund. 2001. Protein fin- gerprinting: a new technique reveals extensive conspecific brood parasitism. Ecology 82:1433- 1442. Bergeron, D., M. Darveau, A. Desrochers, and J.- P. L. Savard. 1997. Impact de I’abondance des chicots sur les communautes aviaires et la sauva- gine des forets coniferiennes et feuillues du Que- bec meridional. Technical Report Series Number 27 IF. Canadian Wildlife Service, Quebec City, Quebec, Canada. Bonar, R. L. 2000. Availability of Pileated Wood- pecker cavities and use by other species. Journal of Wildlife Management 64:52-59. Bouvier, J. M. 1974. Breeding biology of the Hooded Merganser in southwestern Quebec, including in- teractions with Common Goldeneyes and Wood Ducks. Canadian Field-Naturalist 88:323-330. Buckland, S. T, D. R. Anderson, K. P. Burnham, AND J. L. Laake. 1993. Distance sampling: esti- mating abundance of biological populations. Chapman and Hall, London, United Kingdom. Caldwell, P. J. and A. E. Snart. 1974. A photograph- ic index for aging Mallard embryos. Journal of Wildlife Management 38:298-301. Carter, B. C. 1958. The American Goldeneye in cen- tral New Brunswick. Wildlife Management Bul- letin 2. Number 9. Canadian Wildlife Service, Ot- tawa, Ontario, Canada. Dow, H. AND S. Fredga. 1984. Factors affecting re- productive output of the goldeneye duck Bu- cephala clangiila. Journal of Animal Ecology 53: 679-692. Dugger, B. D., K. M. Dugger, and L. H. Fredrick- son. 1994. Hooded Merganser Lophodytes cucul- latus. The birds of North America. Number 98. Eadie, J. M. 1989. Alternative reproductive tactics in a precocial bird: the ecology and evolution of brood parasitism in goldeneyes. Thesis. University of British Columbia, Vancouver, Canada. Eadie, J. M. and G. Gauthier. 1985. Prospecting for nest sites by cavity-nesting ducks of the genus Bu- cephala. Condor 87:528-534. Senechal et al. • NESTING ECOLOGY OE GOLDENEYES AND MERGANSERS 741 Eadie, J. M., M. L. Mallory, and H. G. Lumsden. 1995. Common Goldeneye Biicephala clangiila. The birds of North America. Number 170. Eriksson, M. O. G. 1979. Clutch size and incubation efficiency in relation to nest-box size among gold- eneyes Bucephala clangula. Ibis 121:107—109. Eriksson, M. O. G. and M. Andersson. 1982. Nest parasitism and hatching success in a population of goldeneyes Bucephala clangula. Bird Study 29: 49-54. Erskine, a. J. 1990. Joint laying in Bucephala ducks — “parasitism” or nest-site competition? Omis Scandinavica 21:52-56. Fredga, S. and H. Dow. 1984. Factors affecting the size of a local population of goldeneye Bucephala clangula (L.) breeding in Sweden. Viltrevy 13: 225-255. Gauthier, G. 1989. The effect of experience and tim- ing on reproductive performance in Buffleheads. Auk 106:568-576. Gauthier, G. and J. N. M. Smith. 1987. Territorial behaviour, nest-site availability, and breeding den- sity in Buffleheads. Journal of Animal Ecology 56:171-184. Gilmer, D. S., I. J. Ball, L. M. Cowardin, J. E. Mathisen, and j. H. Riechmann. 1978. Natural cavities used by Wood Ducks in north-central Minnesota. Journal of Wildlife Management 42: 288-298. Haramis, G. M. and D. Q. Thompson. 1985. Density- production characteristics of box-nesting Wood Ducks in a northern greentree impoundment. Jour- nal of Wildlife Management 49:429-436. Lumsden, H. G., R. E. Page, and M. Gauthier. 1980. Choice of nest boxes by Common Goldeneyes in Ontario. Wilson Bulletin 92:497-505. Lumsden, H. G., J. Robinson, and R. Hartford. 1986. Choice of nest boxes by cavity-nesting ducks. Wilson Bulletin 98:167-168. Maisonneuve, C., R. McNicoll, A. Desrosiers, and G. Lupien. 2002. Caracterisation de 1’ habitat de reproduction des canards arboricoles. Societe de la faune et des pares du Quebec, Direction de la recherche sur la faune et Direction de I’amenagement de la faune du Saguenay-Lac Saint-Jean, Quebec City, Quebec, Canada. Mallory, M. L., H. G. Lumsden, and R. A. Walton. 1993a. Nesting habits of Hooded Mergansers Mergu.s cucullatus in northeastern Ontario. Wild- fowl 44:101-107. Mallory, M. L., D. K. McNicol, and P. J. Weath- ERHEAD. 1994. Habitat quality and reproductive ef- fort of Common Goldeneyes nesting near Sud- bury, Canada. Journal Wildlife Management 58: 552-560. Mallory, M. L., D. K. McNicol, R. A. Walton, and M. Wayland. 1998. Risk-taking by incubating Common Goldeneyes and Hooded Mergansers. Condor 100:694-701. Mallory, M. L., A. Taverner. B. Bower, and D. Crook. 2002. Wood Duck and Hooded Merganser breeding success in nest boxes in Ontario. Wild- life Society Bulletin 30:301-316. Mallory, M. L., P. J. Weatherhead, D. K. McNicol, AND M. E. Wayland. 1993b. Nest site selection by Common Goldeneyes in response to habitat features influenced by acid precipitation. Ornis Scandinavica 24:59-64. McNicol, D. K., R. A. Walton, and M. L. Mallory. 1997. Monitoring nest box use by cavity-nesting ducks on acid-stressed lakes in Ontario, Canada. Wildlife Biology 3:1-12. Morse, T. E., J. L. Jakabosky, and V. P. McCrow. 1969. Some aspects of the breeding biology of the Hooded Merganser. Journal of Wildlife Manage- ment 33:596-604. Newton, I. 1994. The role of nest sites in limiting the numbers of hole-nesting birds: a review. Biologi- cal Conservation 70:265-276. Nilsson, S. G. 1986. Evolution of hole-nesting in birds: on balancing selection pressures. Auk 103: 432-435. POysa, H. 1999. Conspecific nest parasitism is asso- ciated with inequality in nest predation risk in the Common Goldeneye {Bucephala clangula). Be- havioural Ecology 10:533-540. POysa, H. and S. Poysa. 2002. Nest-site limitation and density dependence of reproductive output in the Common Goldeneye Bucephala clangula: impli- cations for the management of cavity-nesting birds. Journal of Applied Ecology 39:502-510. Poysa, H., M. Milonoff, V. Ruusila, and J. Virta- NEN. 1999. Nest-site selection in relation to habitat edge: experiments in the Common Goldeneye. Journal of Avian Biology 30:79-84. Poysa, H., P. Runko, V. Ruusila, and M. Milonoff. 2001. Identification of parasitized nests by using egg morphology in the Common Goldeneye: an alternative to blood sampling. Journal of Avian Biology 32:79-82. Prince, H. H. 1968. Nest sites used by Wood Ducks and Common Goldeneyes in New Brunswick. Journal of Wildlife Management 32:489-500. Research Unit for Wildlife Assessment. 1999. DIS- TANCE. Version 3.5. University of St. Andrews, Fife, Scotland. Salter, K. C. and R. F. Fawcett. 1993. The ART test of interaction: a robu.st and powerful rank test of interaction in factorial models. Communications in Statistics: Simulation and Computation 22:137- 153. SAS In.stitlite. 1999. SAS/STAT user's guide. Version 8.0. SAS Institute Inc., Cary, North Carolina, USA. Savard, J.-P. 1.. 1984. Territorial behaviour of Com- mon Goldeneye. Barrow's Goldeneye and BufUe- head in areas of sympatry. Ornis Scandinavica 15: 211-216. Savard, J.-P. L. 1988. Use of nest boxes by Barrou 's Goldeneyes: nesting success and effect im the breeding population. Wildlife Society Bulletin 16: 125-132. 742 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 Shmel, B. and R W. Sherman. 2001. Intraspecific par- asitism and nest-site competition in Wood Ducks. Animal Behaviour 61:787-803. Semel, B., P. W. Sherman, and S. M. Byers. 1988. Effects of brood parasitism and nest-box place- ment on Wood Duck breeding ecology. Condor 90:920-930. Semel, B., P. W. Sherman, and S. M. Byers. 1990. Nest boxes and brood parasitism in Wood Ducks: a management dilemma. Pages 163-170 in Pro- ceedings of the 1988 North American Wood Duck Symposium (L. H. Lredrickson, G. V. Burger, S. P. Havera, D. A. Graber, R. E. Kirby, and T. S. Taylor, Editors). St. Louis, Missouri, USA. Senechal, H. 2003. Etude comparative de I’ecologie de nidification et d’elevage des couvees de Gar- rot a oeil d’or (Bucephala clangula) et de Harle couronne (Lophodytes cucullatus) dans un habi- tat de riviere. Thesis. Universite Laval, Quebec, Canada. The Wilson Journal of Ornithology 1 20(4):743-754, 2008 MOVEMENTS AND HABITAT USE BY RED-BREASTED MERGANSER BROODS IN EASTERN NEW BRUNSWICK SHAWN R. CRAIK'-2 AND RODGER D. TITMAN' ABSTRACT. — Red-breasted Mergansers (Mergus serrator) commonly breed in estuaries, but little is known about their brood-rearing in coastal environments. We measured daily movements and habitat use of radio- marked {n = 17) female Red-breasted Mergansers with broods originating from coastal barrier islands at Ko- uchibouguac National Park, New Brunswick, Canada from 2002 to 2004. Primary brood movements from nest sites to initial rearing areas were often extensive, averaging 3.5 km {n — 15), since many broods crossed Saint- Louis Lagoon to continental rearing sites. Broods remained mobile throughout the rearing period and there was little difference in daily movements between age class I (days 1-10 post nest exodus), class II (days 11-20), and class III (>20 days) broods. Broods frequented shallow (x = 51 cm, 95% Cl: 44-58 cm, n = 191 locations), nearshore (x = 47 m, 95% Cl: 33-60 m, n = 157 locations) waters that often supported submergent eel grass (Zostera marina). Broods selected estuarine intertidal regions in Saint-Louis and Kouchibouguac lagoons, as well as wetlands at the mouths of tidal streams. Few broods were found in tidal river and marine habitats. Continental estuarine intertidal, tidal stream, and saltmarsh habitats were preferred by age class I broods whereas estuarine intertidal and subtidal habitats were preferred by age classes II and III. This study highlights the importance of estuarine habitats in lagoons and tidal streams for brood-rearing Red-breasted Mergansers in eastern New Brunswick. Received 22 June 2007. Accepted 31 January 2008. Adult waterfowl may enhance juvenile re- cruitment by assisting young in locating suit- able habitats throughout the brood-rearing pe- riod (Afton and Paulus 1992). Parents select wetlands where food quality and abundance are greatest to meet the nutritional require- ments of developing young (Minot 1980, Tal- ent et al. 1982), and use areas that provide loafing sites for resting broods (Bengtson 1971b). Movements and habitat selection by brood-rearing females have been documented for a variety of dabbling (Anatini) and diving (Aythyini) duck species, including Mallard (Anas platyrhynchos) (Krapu et al. 2006), Wood Duck (Aix sponsa) (Granfors and Flake 1999), Redhead (Aythya americana) (Yerkes 2000a, b), and Canvasback (Aythya valisiner- ia) (Austin and Serie 1991). Few studies have investigated daily movements and patterns of brood habitat selection by sea ducks (Mergini) breeding at inland and, in particular, at coastal sites, despite this tribe representing over 40% of duck species breeding in North America (Palmer 1976, Pdysa and Virtanen 1994). The brood-rearing activities of many sea duck populations are poorly known because sea ' Department of Natural Resource Sciences, McGill University, Ste-Anne-de-Bellevue, QC M9X 3V9, Can- ada. “ Corresponding author; e-mail: shawn.craikC7%), but widgeon grass {Ruppia maritima) dominated brackish waters (Beach 1988). Average water depths in the es- tuary were shallow (<1 m) except for several channels (2-6 m) and over unvegetated inter- tidal flats (0.3 m) (Beach 1988). Saltmarsh, consisting of two zones, fringed much of the continental side of the lagoons. The low Craik and Titman • RED-BREASTED MERGANSER BROOD ECOLOGY 745 marsh was flooded daily by tides (<0.7 m tid- al amplitude) and was dominated by saltwater cordgrass {Spartina alterniflora), whereas the upper marsh was only infrequently flooded by the highest tides and saltmeadow cordgrass {Spartina patens), saltmarsh sedge (Carex pa- leacea), and prairie cordgrass {Spartina pec- tinata) were most common (Beach 1988). Red-breasted Mergansers nested colonially on Tern Islands, a 2.8-ha island complex con- sisting of three barrier islands in Saint-Louis Lagoon (Fig. 1). The islands are composed of sand and primarily stabilized by marram grass {Ammophila breviligulata). Potential duckling predators included Great Black-backed {Larus marinus) and Herring (L. argentatus) gulls, American Crows {Corvus brachyrhynchos). Common Ravens (C. corax). Northern Harri- ers {Circus cyaneus), and red fox {Vulpes vul- pes). Data Collection. — Merganser nests were lo- cated each year by systematically searching Tern Islands on foot between 0700 and 1100 hrs AST once a week from late May to early July. Nest coordinates were obtained using a global positioning system (GPS) with 4 m ac- curacy (model eTrex, Garmin Ltd., Olathe, KS, USA). Hatching date was estimated for each nest by floating 2-3 incubated eggs in water and assuming an incubation period of 30 days (Westerskov 1950). Females were captured on the nest during the last week of incubation using a dip net or automatic nest trap (Weller 1957), and were equipped with radio transmitters subcutaneously implanted immediately posterior to the nape (model PD- 2, Holohil Systems Ltd., Carp, ON, Canada) (Korschgen et al. 1996). Transmitters weighed between 3.1 and 3.8 g, which represented <0.6% of the total body mass of the lightest marked female (680 g). Females were cap- tured and radiomarked throughout the hatch- ing season (early Jul-early Aug). We attempted to locate marked broods (hen and >1 duckling) once daily post-nest exodus (White and Garrott 1990). Locations were ob- tained throughout the diurnal period (0600- 2100 hrs) for each brood from a sea kayak; however, a small motorboat (3 m) was occa- sionally used. Efforts were made to monitor broods to 45 days post-nest exodus, after which brood-hen bonds generally deteriorated. Total brood mortality or abandonment was as- sumed when a marked hen was observed post- nest exodus either alone or with other female mergansers. We excluded travel locations be- tween rearing sites {n = 2) and locations that may have been influenced by our activity {n = 8). We assumed investigator disturbance occurred when broods were initially spotted swimming directly away from the boat. Ef- forts were made upon obtaining a telemetry signal from a brood to observe the brood with binoculars from as great a distance as possible (>500 m). This was feasible given the open environment and often linear shorelines of the estuarine and marine habitats at Kouchiboug- uac. We approached brood locations only when broods had moved >200 m to further reduce the possibility of investigator influence on brood movements. We recorded coordinates (location of brood where initially observed) at each location with a GPS and broods were classified to age class as: (I) days 1-10 post nest exodus, (II) days 11-20, and (III) >20 days (Austin and Serie 1991). We measured water depth (cm) and distance to shore (m) at each location, and es- timated proportion of cover of three submer- gent vegetation types (eelgrass, widgeon grass, macro algae) within a 5-m radius cen- tered on the location. Habitat Map. — Coastal habitat types within the study area were delineated with use of the Wetlands of the Maritime Provinces classifi- cation system (Hanson and Calkins 1996). Habitats were classified as: (1) tidal river, (2) tidal stream (1-3 order), (3) saltmarsh, (4) es- tuarine intertidal flat in the lagoons, (5) estu- arine subtidal flat in the lagoons, (6) marine intertidal flat, and (7) marine subtidal flat. We divided estuarine intertidal flat into continen- tal estuarine intertidal and barrier island es- tuarine intertidal. Estuarine intertidal habitats were distinguished from estuarine subtidal by the absence of extensive eelgrass meadows at intertidal areas (Hanson and Calkins 1996). Marine intertidal habitat included nearshore, foreshore, and backshore regions (Godfrey 1976). We excluded barrier island and upland habitats (forests, freshwater wetlands) from analyses given complete avoidance of these habitats by merganser broods. Candidate hab- itats (n = 8) were identified on a mosaic of digital geo-referenced orthophotographs pro- vided by Parks Canada and habitats were dig- 746 THE WILSON JOURNAL OF ORNITHOLOGY • VoL 120, No. 4, December 2008 itized from the mosaic using Cartalinx 1.04 (Clark Laboratories, Worcester, MA, USA). Data Analyses. — Primary and secondary movements for each brood were measured as straight-line distances (km) using ArcView GIS 3.1 (Environmental Systems Research In- stitute Inc., Redland, CA, USA). Secondary movements were calculated only when a brood was located on consecutive days. Movement data were summarized by pooling locations among broods so that daily move- ments were compared between each of the three brood age classes. Data were square-root transformed to improve normality (Shapiro- Wilk Test = 0.95) prior to a one-way analysis of variance (ANOVA). We defined a home range as the area where broods restricted their activities during observation (Mauser et al. 1994). Home range size (ha) was calculated for broods that were assumed to have fledged at least one duckling (>13 locations) using the minimum convex polygon (MCP) method with the Animal Movement Program 2.0 (Hooge and Eichenlaub 2000) in ArcView. Before estimating home ranges, 5% of outly- ing locations were removed from each brood using the harmonic mean method (Dixon and Chapman 1980). We measured the total dis- tance moved (km) throughout the observation period for these broods. Habitat use by broods was compared to availability at two spatial scales (Johnson 1980). Eirst, proportions of each habitat with- in brood home ranges were compared to those available (second-order selection). Second, we compared proportions of telemetry locations that occurred in each habitat type to those within home ranges (third-order selection). Habitat availability was calculated for each brood by creating a thematic map in ArcView containing used habitats surrounding the nest within a radius equivalent to the distance be- tween a brood’s nest and its farthest location from the nest (Granfors and Elake 1999, Chouinard and Arnold 2007). The thematic map was clipped in the habitat map to esti- mate the proportion of each habitat type avail- able to individual broods. Thematic maps con- taining brood home ranges were also clipped in the habitat map to estimate the proportion of each habitat type that occurred within in- dividual home ranges. Telemetry locations were overlaid onto the habitat map to estimate the proportion of locations that occurred with- in each habitat for each brood. The Friedman (1937) ranking procedure was used for each brood to examine trends in differences in availability and use of each hab- itat type across broods, which resulted in hab- itats being ranked from one to eight at the second and third orders of selection. Compo- sitional analysis (CA) was used to examine habitat selection at the two spatial scales (Ae- bischer et al. 1993, Smith 2003). Habitat pro- portions with zero values were replaced with 0.00001, a value an order of magnitude small- er than the smallest non-zero value used or available, to calculate log-ratios (Aebischer et al. 1993). Randomization procedures were used to avoid inflated Type I error rates pro- duced by CA (Bingham and Brennan 2004, Thomas and Taylor 2006). We generated 2,500 data sets of the same size (number of broods) and number of locations as the orig- inal data but under the hypothesis of random habitat use (Pendleton et al. 1998, Granfors and Flake 1999). A m.atrix of r-tests was con- structed if nonrandom selection of habitats oc- cun*ed (P < 0.05) by using differences of log- ratios between habitats of interest to rank hab- itat preferences and examine where ranks dif- fered (Aebischer et al. 1993). Broods tracked for fewer days provided less information, thus, observations were weighted by the square-root of the number of days observed (Granfors 1996). We considered broods that were assumed to have fledged young {n = 6) or were observed on at least eight occasions {n - 2) prior to total brood mortality or brood abandonment for analyses of habitat selection. Effects of brood age on habitat use were evaluated using a Chi-square goodness of fit test for each of the three brood-age classes (White and Garrott 1990). Few broods {n = 6) provided information about each age class and many broods had few locations (<6) within a given age class, thus, locations were pooled among broods and assigned to one age class (Alldredge and Ratti 1986, Gammonley 1990). Available habitat was estimated by cre- ating a thematic map in ArcView containing used habitats surrounding the nest of the brood with the largest radius between the brood’s nest and its farthest location from the nest. The thematic map was clipped in the habitat map to calculate the proportion of each Craik and Titman • RED-BREASTED MERGANSER BROOD ECOLOGY 747 habitat type available. Brood locations were overlaid onto the habitat map to calculate the proportion of locations that occurred within each habitat for each age class. We used an exact test to compare use of habitats within each age class because of many small expect- ed frequencies (<5) in the Chi-square table (Sokal and Rohlf 1981). Bonferroni confi- dence intervals were calculated if nonrandom selection occurred within an age class (P < 0.05) to identify which habitat types were pre- ferred or avoided (Byers and Steinhorst 1984). Data analyses were conducted using SPSS 1 1.5 (SPSS Inc., Chicago, IL, USA) and SAS 9.1 (SAS Institute Inc., Cary, NC, USA). RESULTS Marking and Tracking. — Twenty-seven fe- male mergansers were fitted with transmitters. Nests of three marked females were unsuc- cessful, and location data were unavailable for three females without broods post-nest exodus and four females that were not found follow- ing nest exodus. We obtained 233 locations {x = 13.7 locations/brood, range 1 to 45) from 17 broods. The mean brood age at last obser- vation was 16.2 days post-nest exodus and ranged from 0 to 5 1 days (day 0 = day of nest exodus). Ninety (38%) locations were record- ed for age class I broods, 49 (21%) for class II broods, and 95 (41%) for class III broods. Movements. — Primary movements averaged 3.5 km (95% Cl: 2.8-4.2 km, n = \5) and ranged from 0.6 to 5.9 km. Most broods crossed Saint-Louis Lagoon from Tern Islands to the continent and subsequently traveled north or south along shorelines to estuarine bays or mouths of tidal streams. One hen in 2003 led her brood <1 km from her nest to initial rearing habitat along the shorelines of Tern Islands. One female, marked in both 2003 and 2004, led her broods to Anse a Si- mon Michel, a shallow bay in Saint-Louis La- goon, in both years (Pig. 1 ). Secondary movements averaged 1 .0 km/ day (95% Cl: 0.8-1. 2 km, n = 157 locations) and ranged from localized movements (<(). 1 km) to 7.9 km. Secondary movements were typically not directional (consistently moving away from Tern Islands) as broods often re- turned to previous rearing sites. Overall, brood movements were greatest immediately post-nest exodus (Table 1 ), but there were no TABLE 1 . Distances moved by Red-breasted Mer- ganser broods for three age classes at Kouchibouguac National Park, New Brunswick, 2002-2004. Age class^ Mean distance moved/ day (km) 95% Cl (km) Range (km) Broods^’ Observations I 1.3 1.0-1. 7 0.0-5.9 15 76 II 1.1 0.7-1. 4 0. 1-7.9 5 52 III 1.1 0.9-1. 3 0.0-3.5 6 77 ^ I = days 1-10 post-nest exodus, II = days 1 1-20, III = >20 days. Number of broods that provided movement data within each brood age class. differences in average daily movements be- tween age class I, class II, and class III broods (^2.202 = 0.48, P = 0.62). Mean home range was 806 ha (95% Cl: 40-1,572 ha) and ranged from 196 to 2,199 ha for those broods that were assumed to have fledged young (n = 6). Total distance traveled by these broods aver- aged 44.0 km (95% Cl: 20.2-67.7 km) and ranged from 17.4 to 81.2 km. Microhabitat. — Microhabitat at brood lo- cations varied by habitat type (Table 2). Over- all, broods frequented shallow waters (T = 51 cm, 95% Cl: 44-58 cm, n = 191 locations) near shorelines (x = 47 m, 95% Cl: 33-60 m, n = 157 locations). The proportion of sub- mergent vegetation at locations averaged 38.3% (95% Cl: 33.8-42.9%, n = 212 loca- tions). Eelgrass occurred at 84.7% of 216 brood locations whereas macro algae and wid- geon grass each occurred at <20% of loca- tions. Eelgrass cover at brood locations was considerably higher {x = 35.2%, 95% Cl: 30.8-39.4%, n = 216 locations) than that of macro algae (T = 3.4%, 95% Cl: 1.9— 4.9%, n = 216 locations) or widgeon grass (.v = 0.6%, 95% Cl: 0-1.3%, n = 216 locations) (X^ = 299.49, df = 2, P < O.OOl). Broods loafed on shore typically <1 m from water. Broods on the continental side of lagoons rested on sandy intertidal areas sur- rounded by saltwater cordgrass and on peat deposits consisting of scattered cordgrass. Broods on the barrier island side of lagoons loafed on intertidal beaches that lacked veg- etative concealment. Habitat Use and Selection. — We observed little variability across broods when differenc- es of availability and use lor each habitat were examined (Table 3). Estuarine subtidal was 748 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 TABLE 2. Habitat characteristics at Red-breasted Merganser brood locations within four habitat types at Kouchibouguac National Park, New Brunswick. 2002-2004. Data were obtained from 17 broods. Habitat type“ STR ES IFB IFC Mean water depth, cm 30.3 65.3 38.2 33.3 95% Cl 20.4-40.2 52.7-77.8 24.9-51.3 27.2-39.5 Range 4.9-64.1 2.7-475.0 2.2-122.6 2.0-91.8 n 20 86 20 46 Mean distance to shore, m 22.0 77.5 13.3 16.0 95% Cl 6.3-37.7 53.5-101.3 7.6-18.9 12.4-19.7 Range 2.1-45.3 0.9-771.0 0.3-39.2 0.1-45.1 n 7 80 22 41 Mean vegetative cover, %'" 33.8 44.4 20.2 36.7 95% Cl 20.5-47.0 37.5-51.3 7.4-33.0 28.0-45.4 Range 5.0-95.0 0.0-100 0.0-100 0.0-100 n 20 94 24 55 Occurrence, %*’ Eelgrass 100 (20/20) 89.4 (84/94) 65.2 (15/23) 84.9 (45/53) Macro algae‘s 30.0 (6/20) 7.4 (7/94) 13.0 (3/23) 9.4 (5/53) Widgeon grass 15.0 (3/20) 2.1 (2/94) 0.0 (0/23) 3.8 (2/53) Cover, %'° Eelgrass Mean 26.3 42.8 18.1 33.4 95% Cl 17.0-35.5 35.9-49.8 6.4-29.8 25.5-41.3 Range 5.0-70.0 0.0-100 0.0-100 0.0-100 n 20 94 24 55 Macro algae‘s Mean 5.8 0.9 2.1 3-4 95% Cl 0.0-12.6 0.0-1. 8 0.0-5.5 0.7-6. 1 Range 0.0-50.0 0.0-40.0 0.0-40.0 0.0-50.0 n 20 94 24 55 Widgeon grass Mean 1.8 0.3 0.0 0.0 95% Cl Range n 0.0-3.7 0.0-15.0 20 0.0-0.7 0.0-20.0 94 24 55 3 STR = tidal stream. ES = estuarine subtidal. IFB = barrier island estuarine intertidal, and IFC = continental estuarine intertidal. Other habitats excluded since few locations (<6) were available. Measured within a 5-m radius centered on brood location. Included live and dead vegetation. Families Phaeophyta. Rhodophyta. Cyanophyta. and Chlorophyta. ranked first (most used) for all eight broods at the second order of selection and continental estuarine intertidal was ranked either first or third for all broods at the third order of selec- tion (Table 3). Use of habitats was propor- tional to availability at the second order of selection (Wilk’s X = 0.002, = 50.79, df = 7, randomized P = 0.26). However, marine (n = 5 locations) and river (n = 4 locations) hab- itats were generally not important to brood home ranges (Table 4), and all river locations were within 2.8 km of a lagoon. Marine and tidal river habitats were exclud- ed from analyses of third-order selection as at least six non-missing log-ratio differences, necessary for randomization to show a differ- ence from zero at P < 0.05 (df ^ 5 for t- tests), were unavailable. Removal of habitats did not affect ranking among remaining hab- itats. Habitat use was nonrandom (weighted mean Wilk’s X = 0.13, X" = 16.51, df = 4, randomized P = 0.03) at the third order of selection (Table 4). Continental estuarine in- tertidal was ranked first but was not used more (P = 0.52) than tidal stream. Broods used shallow intertidal areas in several continental bays, including Anse a Simon Michel and Grande Anse, and at the mouths of tidal Craik and Titrnan • RED-BREASTED MERGANSER BROOD ECOLOGY 749 TABLE 3. Habitat ranks for eight Red-breasted Merganser broods at Kouchibouguac National Park, New Brunswick, 2002-2004. Ranks calculated at the second and third orders of habitat selection. Numbers in paren- theses represent number of locations. Habitat type“ Brood ID STR ES IFB IFC MI MS RIV SM 150 in = 39) 5/2^ 1/6 H5 4/1 6/n 8/n 7/3 3/4 131 in = 13) 6/2 1/7 3/1 5/3 HA 8/6 7/n 4/5 711 in = 8) All 1/5 HA 2/1 5/n 8/n 6/n 3/3 638 in = 29) 4/n 1/6 2/1 5/3 HI 8/4 6/n 7/5 500 in = 38) 5/3 1/6 HA 4/1 7/n 8/n 6/2 3/5 209 in = 22) 5/3 Ml 2/6 3/1 6/5 8/n 7/4 4/2 529 in = 31) 5/2 1/8 2/6 4/1 HA 8/3 6/5 3/7 120 in = 8) 4/2 1/3 7/n 3/1 5/n 8/n 6/n 2/4 ^ STR = tidal stream, ES = estuarine subtidal, IFB = barrier island estuarine intertidal, IFC = continental estuarine intertidal, MI = marine intertidal, MS = marine subtidal, RIV = tidal river, and SM = saltmarsh. ^>5/2, where 5 is habitat rank (observed - expected frequencies) at second order of selection and 2 is rank at third order of selection. Rank of 1 = most used, (n) denotes no rank since habitat was not available. streams, such as Major Kollock Creek and Duck Brook (Fig. 1). Brood locations that oc- curred in streams were near stream mouths (<675 m) that emptied into the lagoons. Es- tuarine subtidal was ranked last and was used less {P = 0.008) than continental estuarine in- tertidal and barrier island estuarine intertidal {P = 0.02). The relatively little use of estua- rine subtidal reflected large areas of deep {>15 cm), offshore (>100 m) subtidal habitat that was not used in brood home ranges. Most broods regularly used shallow, nearshore es- tuarine subtidal habitats (Table 4). Use of habitats was nonrandom for brood age class I (x" - 1,318.03, df = 7, P < 0.001). Continental estuarine intertidal, tidal stream, and saltmarsh habitats were preferred (Table 5). Habitat use was also nonrandom for brood age classes II (x^ = 452.21, df = 7, P < 0.001) and III (x" = 1,723.03, df = 7, P < 0.001) as continental estuarine intertidal, bar- rier island estuarine intertidal, and estuarine subtidal habitats were preferred for these age classes (Table 5). DISCUSSION Movements. — Distances traveled by Red- breasted Merganser broods from nests on Tern Islands to rearing areas were variable but of- ten exceeded 3 km. Extensive travels are com- mon for waterfowl broods moving from nest- ing islands to continental wetlands (Munro and Bedard 1977, Sayler and Willms 1997). In Lake Michigan, Braun et al. (1980) report- TABLE 4. Mean (± 95% Cl) percentages of available habitats, habitats within home ranges, and habitats at locations for eight Red-breasted Merganser broods at Kouchibouguac National Park, New Brunswick, 2002- 2004 in = 188 locations). Habitat type*' .STR ES IFB IFC MI MS RIV SM Available 0.4 ± 0. 1 24.0 ± 6.2 4.2 ± 2.1 1.0 ± 0.1 1 .4 ± 0.3 62.8 ± 8.1 2.5 ± 0.6 'jj be 1 + be Home range 0.6 ± 0.5 65.9 ± 13.7 10.7 ± 6.2 6.0 ± 5.9 3.0 ± 2.5 4.5 ± 10.7 0.9 ± 1.1 9.9 ± 6.5 Locations 1 1 .0 ± 11 .0 26.1 ± 17.8 17.7 ± 15.9 29.4 ± 1 1.8 0.9 ± 2.0 2.9 ± 6.8 2.1 ± 2.9 9.9 ± 8.1 Rank: 2nd order^ N/A Rank: 3rd order IP'C > .STR > ILB > SM > E S'> “.STR = tidal stream. FS = estuarine subtidal. II•M = barrier island estuarine intertidal. IIC eontmental estuarine intertidal. MI marine intertidal. MS = marine subtidal, RIV = tidal river, and SM saltmarsh. Home ranges vs. available habitat. Randomized /’ 0.26 and habitats were not ranked. '•'Telemetry locations vs. home ranges. Habitats are presented in rank order where they are separateil with > symbols, those to the lelt being of higher rank than those to the right. Habitats that share an underline were used equally (/’ O.O.'S). ^ Randomization indicated greater (P - 0.02) use of II M in comparison to liS. 750 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 c3 3 DX) 3 O XI ^ 2: d t: I 1 km/day. This mobility may be linked to foraging since mergansers are very mobile while searching for small fish (Wood 1985), often using extensive paths along shorelines in search of prey (Wood and Hand 1985). Merganser broods may move large distances in search of prey, particularly when previous sites are unprofitable (Wood 1985). Multiple merganser broods often occupied nearby loafing and foraging sites, which sug- gested that encounters between broods were likely common. It was generally unknown whether brood movements were triggered by agonistic behavior between broods, but on 22 August 2004, an amalgamated brood of 23 young was aggressively displaced from the mouth of Major Kollock Creek by the hen of an older brood of 7 young. Aggression be- tween brood-rearing female Red-breasted Mergansers was reported in Denmark but there was no indication these behaviors re- sulted in brood movements (Kahlert 1993). We eliminated brood locations that were ap- parently influenced by investigator activity (e.g., boat tracking). However, our activities may have resulted in undetected brood dis- placement on a small number of occasions, such as during motorboat departures from brood locations following habitat measure- ments. Fast-moving motorboats in Denmark were twice as likely to disturb merganser broods than smaller boats, but these broods quickly resumed their activities once the dis- turbance had passed (Kahlert 1993). It is un- likely that our tracking efforts had a major role in affecting brood movements, given that disturbed broods did not move large distances (>200 m) and that some broods made exten- sive secondary movements (>2 km) regard- less of whether they were followed by kayak or motorboat. Habitat Use and Selection. — Merganser broods predominantly used shallow, nearshore intertidal and subtidal estuarine habitats throughout Saint-Louis and Kouchibouguac lagoons. Several broods also used the mouths of tidal streams that emptied into the lagoons. These habitats provided an abundance of small fish throughout the brood-rearing peri- od. Threespine {Gasterosteus aculeatns)., blackspotted (G. wheatlandi), fourspine (Apel- tes c/iiadracus), and ninespine (Pungitius pun- gitiiis) sticklebacks, mummichog {Fundnlns fieteroclitus). Atlantic silverside {Menidia menidia), and cunner {Tautogolahrns adsper- sns) were common in the lagoons throughout summer and represented much of the total fish biomass present within the park's estuarine system (Klassen 2001, Joseph et al. 2006). Broods were observed foraging (/; = 36 lo- cations) over stands of submergent eelgrass in shallow (<1 111) subtidal waters near intertidal areas. Lelgrass meadows provide sticklebacks and other small fish with nursery habitat that 752 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 generally produces more adult recruits than other juvenile fish habitats (Beck et al. 2001, Joseph et al. 2006). Habitat selection by mergansers is affected by availability and abundance of fish (Kamin- ski and Weller 1992). Abundance of Common Merganser {Mergus merganser) broods in British Columbia was linked to abundance of Juvenile Pacific salmon {Oncorhyncus spp.) (Wood 1986) and distribution of Red-breasted Merganser broods is closely related to the dis- tribution and abundance of small fish (<15 cm) (White 1957, Bengtson 1971a, Rad 1980). Sjoberg (1989) suggested the breeding period for Red-breasted Mergansers is syn- chronized with movements of threespine sticklebacks where broods forage on juvenile fish that move downstream to coastal regions. Sticklebacks were likely the most abundant fish in the lagoons at Kouchibouguac through- out the merganser brood-rearing period (Klas- sen 2001). Continental estuarine intertidal, tidal stream, and saltmarsh habitats were preferred by age class I broods. These continental hab- itats provided newly-hatched broods with shallow waters (<50 cm) that offered shelter from wind and larger waves, which likely im- proved foraging given the limited diving abil- ities of young mergansers (Beard 1964). These areas also provided young broods with loafing sites in saltwater cordgrass. Vegetative concealment may have reduced predator de- tection of age class I broods during a period when duckling mortality is often highest (Flint and Grand 1997, Guyn and Clark 1999). We observed little use of stream and saltmarsh habitats for brood age classes II and III as more open estuarine intertidal and subtidal habitats were used by each of these age clas- ses. Bengtson (1971b) observed young mer- ganser broods in restricted areas near conti- nental and island shorelines whereas older broods occurred in more open areas of Lake M’yvatn. Older merganser broods (>2 weeks) may use deeper, more open habitats since shallow waters and concealment are likely of less importance than for younger broods. Age-related shifts in habitat use may be in- fluenced by changes in duckling food require- ments (Beard 1964, Gammonley 1990, Afton and Paulus 1992). The diet of merganser young at Kouchibouguac has not been docu- mented, so it is not known whether temporal variation in habitat use is influenced by chang- es in duckling food requirements (Bengtson 1971a) or by variability in distribution and abundance of food resources at brood-rearing sites. ACKNOWLEDGMENTS We are grateful to Kouchibouguac National Park, particularly Eric Tremblay and Audrey Beaudet, for providing logistical support. M. J. Bouchard, G. O. Fraser, Mireille Gravel, P.-E. Hebert, L. N. Leger, R. M. Leighton, B. J. Martin, E. J. Reese, E. A. Titman, and especially Amelie Rousseau assisted with data col- lection. We thank K. P Kenow for teaching S. R. Craik the transmitter implantation technique and G. O. Fraser for providing sea kayaks. L. A. Egan-Mitton assisted with surgeries. Guillaume Larocque and Bernard Pel- letier provided GIS and statistical assistance, respec- tively. Female capture and transmitter attachment pro- cedures were in compliance with the Canadian Council on Animal Care (CCAC). Funding for this project was provided by Bird Protection Quebec, Bishop’s Univer- sity, Canadian Wildlife Federation, and New Bruns- wick Wildlife Trust Fund. R. G. Clark, C. E. Done- hower, M. A. Gahbauer, M.-A. R. Hudson, M. A. J. O’Connor, M. J. Richardson, J.-P. L. Savard, J. T. Ratti, and an anonymous reviewer provided helpful com- ments on earlier versions of the manuscript. LITERATURE CITED Aebischer, N. j., P. a. Robertson, and R. E. Ken- ward. 1993. Compositional analysis of habitat use from radio-tracking data. Ecology 74:1313-1325. Afton, A. D. and S. L. Paulus. 1992. Incubation and brood care. Pages 62-108 in Ecology and man- agement of breeding waterfowl (B. D. J. Batt, A. D. Afton, M. G. Anderson, C. D. Ankney, D. H. Johnson, J. A. Kadlec, and G. L. Krapu, Editors). University of Minnesota Press, Minneapolis, USA. Alldredge, J. R. and J. T. Ratti. 1986. Comparison of some statistical techniques for analysis of re- source selection. Journal of Wildlife Management 50:157-165. Austin, J. E. and J. R. Serie. 1991. Habitat use and movements of Canvasback broods in southwestern Manitoba. Prairie Naturalist 23:223—228. Beach, H. 1988. The resources of Kouchibouguac Na- tional Park: resource description and analysis. Kouchibouguac National Park, Environment Can- ada, New Brunswick, Canada. Beard, E. B. 1964. Duck brood behavior at the Seney National Wildlife Refuge. Journal of Wildlife Management 28:492-521. Beck, M. W., K. L. Heck, K. W. Able, D. L. Childers, D. B. Eggleston, B. M. Gillanders, B. Halpern, C. G. Hays, K. Hoshino, T. J. Minello, R. J. Orth, P. F. Sheridan, and M. P. Weinstein. 2001. Craik and Titman • RED-BREASTED MERGANSER BROOD ECOLOGY 753 The identification, conservation, and management of estuarine and marine nurseries for fish and in- vertebrates. Bioscience 51:633—641. Bengtson, S.-A. 1971a. Eood and feeding of diving ducks breeding at Lake M’yvatn, Iceland. Ornis Eennica 48:77-92. Bengtson, S.-A. 1971b. Habitat selection of duck broods in Lake M’yvatn area, north-east Iceland. Ornis Scandinavica 2:17-26. Berg, P E 1956. A study of waterfowl broods in east- ern Montana with special reference to movements and the relationship of reservoir fencing to pro- duction. Journal of Wildlife Management 20:253- 262. Bingham, R. L. and L. A. Brennan. 2004. Compari- son of Type I error rates for statistical analyses of resource selection. Journal of Wildlife Manage- ment 68:206-212. Braun, B. M., P. A. Heinz, and G. H. Heinz. 1980. Herring Gull predation on Red-breasted Mergan- ser ducklings. Wilson Bulletin 92:403. Byers, C. R. and R. K. Steinhorst. 1984. Clarifica- tion of a technique for analysis of utilization- availability data. Journal of Wildlife Management 48:1050-1053. Chouinard, M. P. and T. W. Arnold. 2007. Survival and habitat use of Mallard {Anas platyrhynchos) broods in the San Joaquin Valley, California. Auk 124:1305-1316. Decarie, R., E Morneau, D. Lambert, S. Carriere, AND J.-P. L. Savard. 1995. Habitat use by brood- rearing waterfowl in subarctic Quebec. Arctic 48: 383-390. Dixon, K. R. and J. A. Chapman. 1980. Harmonic mean measure of animal activity areas. Ecology 61:1040-1044. Erskine, a. J. 1992. Atlas of breeding birds of the Maritime Provinces. Nimbus Publishing and Nova Scotia Museum, Halifax, Nova Scotia, Canada. Flint, P. L. and J. B. Grand. 1997. Survival of Spec- tacled Eider adult females and ducklings during brood rearing. Journal of Wildlife Management 61:217-221. Friedman, M. 1937. The use of ranks to avoid the assumption of normality implicit in the analysis of variance. Journal of the American Statistical Association 32:675-701. Gammonley, j. H. 1990. Habitat use and survival of Wood Duck broods in Mingo Swamp. Thesis. University of Missouri-Columbia, Columbia, USA. Godfrey, P. J. 1976. Barrier beaches of the East Coast. Oceanus 19:27-40. Granfor.s, D. a. 1996. Wood Duck brood movements, habitat use, and survival rales on two prairie rivers in South Dakota. Dissertation. South Dakota Stale University, Brookings, USA. Granfors, D. a. and L. D. Flakf;. 1999. Wood Duck brood movements and habitat use on prairie rivers in South Dakota. Journal of Wildlife Management 63:639-649. Guyn, K. L. and R. G. Clark. 1999. Factors affecting survival of Northern Pintail ducklings in Alberta. Condor 101:369-377. Hanson, A. R. and L. Calkins. 1996. Wetlands of the Maritime Provinces: revised documentation for the Wetlands Inventory. Technical Report Number 267. Canadian Wildlife Service, Atlantic Region, Sackville, New Brunswick, Canada. Hooge, P. N. and B. Eichenlaub. 2000. Animal move- ment extension for AreView. Version 2.0. USDI, Geological Survey, Alaska Science Center, Bio- logical Science Office, Anchorage, USA. Johnson, D. H. 1980. The comparison of usage and availability measurements for evaluating resource preference. Ecology 61:65-71. Joseph, V, A. Locke, and J.-G. J. Godin. 2006. Spatial distribution of fishes and decapods in eelgrass {Zoster marina L.) and sandy habitats of a New Brunswick estuary, eastern Canada. Aquatic Ecol- ogy 40:111-123. Kahlert, j. 1993. Breeding biology of the Red-breast- ed Merganser {Mergus serrator). Thesis. Univer- sity of Arhus, Arhus, Denmark. Kaminski, R. M. and M. W. Weller. 1992. Breeding habitats of Nearctic waterfowl. Pages 568-589 in Ecology and management of breeding waterfowl (B. D. J. Batt, A. D. Afton, M. G. Anderson, C. D. Ankney, D. H. Johnson, J. A. Kadlec, and G. L. Krapu, Editors). University of Minnesota Press, Minneapolis, USA. Klassen, G. j. 2001. Fish and fisheries of Kouchi- bouguac National Park. Unpublished Report. Memramcook, New Brunswick, Canada. Korschgen, C. E., K. P. Kenow, W. L. Green, M. D. Samuel, and L. Sileo. 1996. Technique for im- planting radio transmitters in day-old ducklings. Journal of Field Ornithology 67:392-397. Krapu, G. L., P. J. Pietz, D. A. Brandt, and R. R. Cox. 2006. Mallard brood movements, wetland use, and duckling survival during and following a prairie drought. Journal of Wildlife Management 70:1436-1444. Mauser, D. M., R. L. Jarvis, and D. S. Gilmer. 1994. Movements and habitat use of Mallard broods in northeastern California. Journal of Wildlite Man- agement 58:88-94. Minot, E. O. 1980. Tidal, diurnal and habitat infiu- ences on Common Eider rearing activities. Ornis Scandinavica 11:165-172. Munro, j. and j. Bedard. 1977. Creche formation in the Common Eider. Auk 94:759-771. Palmer, R. S. 1976. Handbook of North American birds. Volume 3. Yale University Press, New Ha- ven, Connecticut, USA. PiiNDLinoN, G. W., K. Titus, E. DeGayner, C. J. Plat- tf;n. and R. E. Lowell. 1998. Compositional analysis and GIS for study of habitat selection by goshawks in southeast Alaska. Journal of Agri- cultural. Biological, and Environmental Statistics 3:280-295. I’oysA, H. and j. Viktanhn. 1994. Habitat selection 754 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 and survival of Common Goldeneye (Bucephala clangLila) broods - preliminary results. Hydro- biologia 279/280:289-296. Rad, O. 1980. Breeding distribution and habitat selec- tion of Red-breasted Mergansers in freshwater in western Norway. Wildfowl 31:53-56. Rotella, J. J. and J. T. Ratti. 1992. Mallard brood movements and wetland selection in southwestern Manitoba. Journal of Wildlife Management 56 508-515. Sayler, R. D. and M. a. Willms. 1997. Brood ecol- ogy of Mallards and Gadwalls nesting on islands in large reservoirs. Journal of Wildlife Manage- ment 61:808-815. Sjoberg, K. 1989. Time-related predator/prey interac- tions between birds and fish in a northern Swedish river. Oecologia 80:1-10. Smith, R G. 2003. Compos analysis. Version 5.1 stan- dard. Smith Ecology Ltd., Abergavenny, United Kingdom. SoKAL, R. R. AND L. J. Rohlf. 1981. Biometry. Second Edition. W.H. Lreeman and Company, San Lran- cisco, California, USA. Talent, L. G., G. L. Krapu, and R. L. Jarvis. 1982. Habitat use by Mallard broods in south central North Dakota. Journal of Wildlife Management 46:629-635. Thomas, D. L. and E. J. Taylor. 2006. Study designs and tests of comparing resource use and avail- ability. II. Journal of Wildlife Management 70: 324-336. Titman, R. D. 1999. Red-breasted Merganser {Mergus serrator). The birds of North America. Number 443. Weller, M. W. 1957. An automatic nest-trap for wa- terfowl. Journal of Wildlife Management 21:456- 458. Westerskov, K. 1950. Methods for determining the age of game bird eggs. Journal of Wildlife Man- agement 14:56-67. White, G. C. and R. A. Garrott. 1990. Analysis of wildlife radio-tracking data. Academic Press, San Diego, California, USA. White, H. C. 1957. Pood and natural history of mer- gansers on salmon waters in the Maritime Prov- inces of Canada. Pisheries Research Board of Canada Bulletin. Number 116. Wood, C. C. 1985. Pood-searching behaviour of the Common Merganser {Mergus merganser). II. Choice of foraging location. Canadian Journal of Zoology 63:1271-1279 Wood, C. C. 1986. Dispersion of Common Merganser {Mergus merganser) breeding pairs in relation to the availability of juvenile Pacific salmon in Van- couver Island streams. Canadian Journal of Zo- ology 64:756-765. Wood, C. C. and C. M. Hand. 1985. Pood-searching behaviour of the Common Merganser {Mergus merganser) I. Punctional responses to prey and predator density. Canadian Journal of Zoology 63: 1260-1270. Yerkes, r. 2000a. Influence of female age and body mass on brood and duckling survival, number of surviving ducklings, and brood movements in Redheads. Condor 102:926-929. Yerkes, T. 2000b. Nest-site characteristics and brood- habitat selection of Redheads: an association be- tween wetland characteristics and success. Wet- lands 20:575-580. Yolng, a. D. and R. D. Titman. 1986. Costs and benefits to Red-breasted Mergansers nesting in tern and gull colonies. Canadian Journal of Zo- ology 64:2339-2343. The Wilson Journal of Ornithology 1 20(4):755-766, 2008 DIFFERENCES IN GROWTH OE BLACK BRANT GOSLINGS BETWEEN A MAJOR BREEDING COLONY AND OUTLYING BREEDING AGGREGATIONS CHRISTOPHER A. NICOLAI,‘-'^ JAMES S. SEDINGER,i-^ AND MICHAEL L. WEGE2 ABSTRACT. — We compared growth rates of Black Brant {Branta bernicla nigricans) goslings from two dispersed nesting aggregations to those from the large Tutakoke River Colony on the Yukon-Kuskokwim Delta, Alaska during summers, 1999 and 2000. Approximately 20% of the Black Brant population on the Yukon- Kuskokwim Delta, Alaska nests outside of the four major colonies. Dispersal to these outlying breeding locations is hypothesized as a mechanism by which individuals reduce the negative effects of density dependence asso- ciated with major colonies. Growth rates of goslings varied among brood rearing areas both associated with dispersed nesting aggregations {n = 4) and associated with the Tutakoke River Colony (n = 7). Mean mass of goslings, adjusted for age, from brood rearing areas associated with dispersed nesting aggregations ranked sixth, eighth, and ninth of nine brood rearing areas sampled in 1999, and sixth and ninth of nine brood rearing areas sampled in 2000. Mean age-adjusted mass of goslings with the largest mass were 198 and 139 g lighter, respectively, from brood rearing areas associated with a dispersed nesting aggregation than those from brood rearing areas associated with the Tutakoke River colony in 1999 and 2000. Our findings suggest that goslings from dispersed nesting aggregations we sampled are unlikely to have an advantage over goslings from a major colony with respect to survival, adult body size, and recruitment. Received 7 June 2007. Accepted 9 February' 2008. Gosling growth rate is highly variable (Cooch et al. 1991, Sedinger and Flint 1991, Aubin et al. 1993) and is governed by both quality and quantity of forage plants (Larsson and Forslund 1991, Cooch et al. 1993, Loonen et al. 1997, Person et al. 2003, Sedinger et al. 2001). Growth rates of goslings have been re- lated to availability of high quality foods at different temporal and spatial scales: (1) among brood-rearing areas associated with a colony (Cooch et al. 1993, Herzog 2002), (2) among colonies (Aubin et al. 1993, Sedinger at al. 2001), (3) among years (Cooch et al. 1991, Sedinger et al. 1998, Person et al. 2003), and (4) among hatch dates (Cooch et al. 1991, Sedinger and Flint 1991, Lindholm et al. 1994). Food quality (measured as % ni- trogen) generally is inversely correlated with plant biomass; when biomass is greater, per- ' Department of Biology and Wildlife and Institute of Arctic Biology, University of Alaska Fairbanks, 21 1 Irving I, Fairbanks, AK 99775, USA. 2U.S. Fish and Wildlife Service. Yukon Delta Na- tional Wildlife Refuge. P. O. Box 346, Bethel, AK 99559, USA. ^ Current address: Department of Natural Resi)urces and Environmental Science. University of Nevada Reno. 1 ()()() Valley Road. Reno. NV 89512. USA. ^Corresponding author; e-mail: nicolai^^unr.nevada. edu cent nitrogen content is lower (Sedinger and Raveling 1986, Person et al. 1998, Sedinger et al. 2001). Growth rates of goslings in one study where both quantity and quality were measured (Sedinger et al. 2001) were posi- tively associated with quantity but not quality of grazed vegetation. Strong preference for plant foods with the highest nitrogen content within brood rearing areas, despite substantial biomass of alternative plant foods of lower ni- trogen content (Harwood 1977, Sedinger and Raveling 1984), indicates that plant foods be- low some minimum level of nitrogen concen- tration will not sustain high rates of growth in goslings. Density of broods is often negatively cor- related with availability of high quality foods (Cooch et al. 1993; Person et al. 1998; Sedin- ger et al. 1998, 2001 ) and growth rates of gos- lings are negatively correlated with local pop- ulation density (Cooch et al. 1991. Person et al. 2003). Reduced gosling growth at higher population densities is a principal mechanism regulating local breeding populations of geese because hrst-year survival is highly correlated with gosling body size late in their hist sum- mer (Francis and Cooke 1992. Schmutz 1993. Sedinger and Chelgren 2007). Growth rate may also affect adult body size (Cooch et al. 155 756 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 1991, Larson and Forslund 1991, Sedinger et al. 1995) and future fecundity (Sedinger et al. 1995). Density-dependent regulation of pop- ulations of arctic nesting geese is largely me- diated through increased competition for suf- ficiently high quality food for growing gos- lings (Cooch et al. 1993, Black et al. 1998, Sedinger et al. 2001). Thus, understanding how individuals respond to local density or population size is important. Density-dependent effects could be reduced by (1) parents from a particular nesting area dispersing with broods to brood-rearing areas of lower density (e.g., Cooch et al. 1993, Slat- tery 2000, Mainguy et al. 2006), or (2) natal or breeding dispersal to new nesting areas away from major colonies (e.g.. Black et al. 1998) and using brood-rearing areas associ- ated with these dispersed breeding locations. Either mechanism could expose goslings to improved foraging conditions and potentially result in increased recruitment and future fe- cundity (Black et al. 2007). Approximately 75% of the breeding Black Brant {Branta hernicla nigricans; hereafter Brant) population nests on the Yukon-Kus- kokwim (Y-K) Delta (Sedinger et al. 1993). Most (80%) Brant on the Y-K Delta nest in four major colonies and the remainder nest in small aggregations (Sedinger et al. 1993), which we define as dispersed nesting aggre- gations. Numbers of nests in the four major colonies increased from 12,000 to >25,000 between 1982 and 2000 (Sedinger et al. 1993; Anthony et al. 1995; R. M. Anthony, USGS, unpubl. report Anchorage, AK), although numbers in major colonies are generally be- low historic levels. Numbers of nests in dis- persed aggregations range from <10 nests to several hundred nests and Brant may also nest solitarily. Local density, through its impact on food abundance (Person et al. 1998), has re- duced gosling growth rates at the two largest colonies on the Y-K Delta (Sedinger et al. 1998, 2001). Concurrently, numbers of Brant nesting in dispersed nesting aggregations have increased (Robert Stehn, U.S. Fish and Wild- life Service, unpubl. report). Natal philopatry by females to main colonies is high (0.83) but not absolute for female Brant (Lindberg et al. 1998, Sedinger et al. 2008), suggesting that dispersal from major colonies could contribute to growth of dispersed nesting aggregations. We did encounter individual female Brant (n < 100) originally marked in major colonies nesting in dispersed nesting aggregations (pers. obs.). Yet, if goslings produced in dis- persed nesting aggregations grew sufficiently rapidly, recruitment by these goslings would also provide a mechanism explaining growth of satellite breeding aggregations. Our objective was to compare growth rates of Brant goslings from a major colony (Tu- takoke River) with those from two dispersed nesting aggregations to test the hypothesis that individuals hatched outside the largest colo- nies experienced improved foraging condi- tions relative to those from the main colonies. We include measures of culmen and tarsus growth as these skeletal measures tend to be more conserved when nutrients are limiting (Sedinger et al. 2001). Responses of these two measures provide a stronger indication of growth responses to variation in nutrient availability among brood rearing areas than does mass (Cooch et al. 1991). We address the potential benefits to gosling growth that would result from immigrating to dispersed nesting aggregations. Our study differs from that of Cooch et al. (1993) who studied benefits of moving to new brood-rearing areas by Lesser Snow Geese {Chen caerulescens caerules- cens) that had not changed nesting location. METHODS Field Methods. — We monitored goslings originating from one large colony (Tutakoke River Colony; hereafter TRC) and two dis- persed nesting aggregations (Big Slough and Aknerkochik River) on the outer coastal fringe in the central part of the Y-K Delta in western Alaska (Fig. 1). The Big Slough dis- persed nesting aggregation had 347 nests in 1992 (M. S. Lindberg, pers. comm.). This number increased to 749 nests in 1999 (Ni- colai 2003). Color banding has been conducted at TRC since 1984. All nests in which at least one parent was marked were visited at least every other day after hatch began at TRC. All nests at Big Slough and Aknerkochik River were checked daily for hatching goslings. We ap- plied individually coded, fish-fingerling tags to the foot webbing of hatched goslings and goslings in pipping eggs (Alliston 1975). We were able to assign gosling ages to within ± 1 Nicolai et al • GROWTH RATES OF DISPERSED NESTING BRANT 757 FIG. 1. Study and brood rearing areas on the Yukon-Kuskokwim Delta, Alaska. Shaded areas reler to brood rearing areas. Hatched areas and bold type refer to nesting areas where web-tags were attached. 758 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 day because hatching of individual clutches requires —24 hrs. Following hatch, broods move to rearing areas up to 40 km from nest sites (Lindberg et al. 1997) where they have relatively w^ell defined home ranges (Flint and Sedinger 1992). Tracking radio-tagged adult females at hatch in 1999 indicated that most broods at Big Slough moved to one general brood rearing area (Horseshoe Lake), whereas broods from Aknerkochik River displayed a more widespread movement of broods (un- publ. data). We recaptured goslings in corral traps during adult remigial molt (Sedinger et al. 1997: Fig. 1). We captured goslings over a range of ages (16-36 days of age) when growth w'as approximately linear. We weighed goslings using spring scales (±5 g) and mea- sured diagonal tarsus, and total culmen using dial calipers (±0.1 mm; Dzubin and Cooch 1992). Statistical Analysis. — Actual density of broods on brood-rearing areas is rarely mea- sured because of logistical difficulties associ- ated with estimates for mobile species (but see Herzog 2002). Rather, size of goslings near fledging, which is affected by their rate of growth, is used as a surrogate for density of broods relative to availability of sufficiently high quality food (e.g., Cooch et al. 1993). We used the same approach because the key ques- tion was, can individuals improve foraging conditions and fitness for their goslings by moving to dispersed nesting aggregations, where competition for food might be reduced? We used the mixed model procedure in SAS (SAS Institute 1989) to assess variation in gosling measurements among brood rearing areas. We included year, gender, and brood rearing area as fixed effects and age as a co- variate. We included brood identity as a ran- dom variable in the analysis to control for within-brood effects. This model structure treated individual broods (not individual gos- lings) as the independent sampling unit for as- sessment of the fixed effects of year, gender, and brood rearing area. Brood mates were not considered independent when evaluating var- iation associated with fixed effects. We did not include hatch dates in the anal- yses for three reasons. First, hatch date rela- tive to other goslings using the same brood- rearing area (rather than absolute hatch date) was likely the component of hatch date most relevant to within brood rearing area variation in growth (Cooch et al. 1991. Sedinger and Flint 1991, Sedinger et al. 1997). Failure to account for relative hatch dates increased the within-brood rearing area residual variance, rendering our analysis more conservative, un- less there was substantial intermixing of gos- lings from two colonies with substantially dif- ferent mean hatch dates. Second, if differences in hatch date between major colonies and dis- persed nesting aggregations created an advan- tage for goslings from major colonies, this ad- vantage would be one mechanism nullifying the benefit of dispersal. Third, and most im- portant. both application of web-tags and re- capture of web-tagged gosling occur over rel- atively short periods within years and colo- nies: — 10 days for application of web-tags and 4 days for recapture. Models containing a brood-rearing area*year interaction created nearly complete confounding between gosling age and hatch date because each brood-rearing area was sampled on only 1 day each year (Herzog 2002). The single annual sample from each brood-rearing area, combined with the negative covariance between gosling age and hatch date and the narrow range of gos- ling ages and hatch dates, made it virtually impossible to simultaneously estimate the ef- fects of both gosling age and hatch date in models. Candidate models included all possible combinations of year, gender, brood-rearing area and age, including all tw'o-way interac- tions. We used a maximum likelihood ap- proach to generate parameter estimates and the Satterthwaite method (Kuehl 2000) to cal- culate denominator degrees of freedom. We used an information theoretic approach based on Aikaike's Information Criterion adjusted for small sample size (AICc) (Burnham and Anderson 1998. Anderson et al. 2000) to as- sess hypotheses represented by candidate models. We examined sources of variation in se- lected models by using estimates of covari- ance from results of the mixed model to es- timate variation wdthin random variables. Co- variance parameters estimated the variance as- sociated with random effects within residual error. We analyzed the same data set using a general linear model (SAS Institute 1989) and estimated variance for fixed effects and error. Nicolai et al. • GROWTH RATES OF DISPERSED NESTING BRANT 759 TABLE 1. Number of web-tagged goslings and broods captured in banding drives. Two areas were sampled in each of the two dispersed nesting aggre- gations and seven areas were sampled associated with the Tutakoke River Colony. Brood rearing areas not sampled are designated N/S. 1999 2000 Nesting area/ Brood area Goslings Broods Goslings Broods Aknerkochik River Camp 8 5 N/S N/S Peninsula 3 2 N/S N/S Big Slough Horseshoe Lake 53 33 14 12 Opagaryak River N/S N/S 13 11 Tutakoke River Camp 38 24 13 9 Bend Colony 10 5 3 3 Kash-Tut 124 72 40 27 Bend Slough 138 91 42 30 Hock Slough 107 64 80 60 Onumtuk 6 5 2 2 Emperor Bend 0 0 5 4 We allowed sums of the covariance parameter estimates from the mixed model to equal error sums of squares from the linear model while maintaining the proportions of the covariance parameter estimates. We present variance components from mixed miodel analysis of variance to allow an assessment of the contri- butions of variables to the overall variance in the dependant variable. Hypothesis tests were not consistent with the information theoretic approaches we used to assess models and we did not use variance components to test hy- potheses in the traditional sense. We present least square means ± SE for mass, culmen, and tarsus. We used model-averaged parame- ter estimates to depict variation in mass, cul- men, and tarsus with respect to age, year, and brood-rearing area. RESULTS We recaptured 487 and 212 web-tagged goslings from 301 and 158 broods in 1999 and 2000, respectively (Table 1 .). Tracking radio- tagged adult females at hatch in 1999 indicat- ed that most broods at Big Slough moved to one general brood rearing area (Horseshoe Lake), whereas broods from Aknerkochik River displayed a more widespread move- ment. We captured goslings on two brood rearing areas associated with the Aknerkochik River dispersed nesting aggregation, two areas associated with the Big Slough dispersed nest- ing aggregation, and seven areas associated with TRC. Twelve goslings from TRC were recaptured at brood rearing areas principally associated with Big Slough and six goslings from Big Slough were recaptured at brood rearing areas associated with TRC, which rep- resented 2 and 8%, respectively, of goslings captured from the two breeding locations. Broods associated with specific brood rearing areas were primarily associated with the breeding locations we assumed in our analy- ses. Mean hatch dates for captured web- tagged goslings among breeding colonies (in- cluding TRC and both dispersed nesting ag- gregations) and years varied by less than 0.5 days. Modal hatch dates were similar in 2000 among breeding colonies; however modal hatch dates varied slightly in 1999: Julian date 176 for Aknerkochik River, Julian date 177 for Big Slough, and Julian date 173 for Tu- takoke River. Mean hatch dates of recaptured goslings followed a similar pattern to modal hatch dates among breeding colonies (mean difference 0 days), indicating that hatch dates of our samples relative to their colony means were similar among our samples from differ- ent breeding locations. Eight models for gosling mass were within four AAICc units of each other (Table 2). These eight models contained effects of brood rearing area, year, gender, age, and brood, but had different combinations of interactions. All competing models of mass contained the ef- fects of brood-rearing area, gender, and age, and a year*brood-rearing area interaction (Ta- ble 2). Mass of goslings, after adjusting for age, from brood-rearing areas associated with dispersed breeding aggregations, ranked sixth, eighth, and ninth of nine areas in 1999, and sixth and ninth of nine areas sampled in 2()()() (Table 3, Fig. 2). Age-adjusted masses for goslings from the dispersed breeding aggre- gation with the largest size were 198 and 139 g smaller than those from TRC brood rearing areas which produced the largest goslings in 1999 and 2()()(), respectively. Goslings from brood-rearing areas associated with dispersed breeding aggregations were 28 and 44 g smaller than goslings from TRC brood rearing areas producing the smallest goslings in 1999 and 2()()0, respectively (Fig. 2). 760 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 TABLE 2. AICc model selection for gosling mass. All models with AAICc <4.0 and the general model are presented. ModeF np*’ AlCf AAICc Model weight Y S BRA A Y*S Y*BRA S*A B 24 791 1.5 0.0 0.2047 Y S BRA A Y*BRA S*A B 23 791 1.6 0.1 0.1947 Y S BRA A Y*BRA B 22 7912.0 0.5 0.1594 Y S BRA A Y*S Y*A Y*BRA S*A B 25 7912.9 1.4 0.1017 Y S BRA A Y*A Y*BRA S*A B 24 7913.0 1.5 0.0967 Y S BRA A Y*A Y*BRA B 23 7913.4 1.9 0.0792 Y S BRA A Y*S Y*BRA B 23 7914.0 2.5 0.0587 Y S BRA A Y*S Y*A Y*BRA B 24 7915.5 4.0 0.0277 Y S BRA A Y*S Y*BRA Y*A S*BRA S*A BRA*AGE B 43 7934.9 23.4 0.0000 Y = Year. S = Gender, BRA = Brood Rearing Area. A = Age. and B = Brood. ^ np = Number of parameters in model. Model selection for tarsus length resulted in four models with AAICc ^ 4.0, all contained variation among brood-rearing area, years, genders, and ages with different combinations of two-way interactions (Table 4). Tarsus length of goslings from dispersed breeding ag- gregations, after adjusting for age, ranked sixth, eighth, and ninth in 1999, and eight and ninth in 2000 (Table 3, Fig. 3). Seven models for gosling culmen length had AAICc < 4.0 and all models contained brood-rearing area, years, gender, and age except for the seventh ranked model (AAICc = 2.6) which did not contain a gender effect (Table 5). Age-adjust- ed estimates of culmen length for goslings from dispersed breeding aggregations ranked fourth, sixth, and ninth in 1999 and seventh and ninth in 2000 (Table 3, Fig. 4). Males weighed 31.6 ± 5.6 g more than fe- males after adjusting for age and had tarsi and culmens that were 2.3 ± 0.3 mm and 0.21 ± 0.12 mm longer than those of females, re- spectively (Figs. 2-4). Goslings of the same age weighed 298.9 ± 64.8 g more and had tarsi and culmens that were 6.2 ±3.8 mm and 2.3 ± 1.6 mm longer in 1999 than in 2000, respectively (Figs. 2—4). We estimated that gosling age accounted for 36% of the variance in gosling mass, followed by the random ef- fect brood (16% of variance), year (13.5%) and brood-rearing area (9%) (Table 6). Vari- ables we considered explained 77% of the var- iance in gosling mass. DISCUSSION Gosling growth varied among brood rearing areas both within and among areas primarily associated with TRC and two dispersed nest- ing aggregations. Variation in gosling growth among brood rearing areas also occurs in oth- er goose populations (Larsson and Forslund 1991, Aubin et al. 1993, Sedinger et al. 2001) and has been related to availability and quality of forage (Cooch et al. 1991, Sedinger and Flint 1991, Manseau and Gauthier 1993). Per- son et al. (1998) found forage quantity varied substantially among brood rearing areas as- sociated with the TRC. Areas with the highest forage quantity (Person et al. 1998) supported the largest goslings in the current study. Gosling size on brood rearing areas asso- ciated with dispersed nesting aggregations was smaller than or similar to that of goslings from TRC. These hndings indicate that gos- lings produced in dispersed nesting aggrega- tions at the time of this study did not experi- ence reduced competition for food, as indi- cated by growth rates. Sedinger et al. (2001) showed that Brant goslings from the Colville River Delta (70° N 149° W) on Alaska’s arctic coast were —300 g larger (adjusted for age) than goslings from TRC and the Kochechik Bay colony (62° N 166° W), another major colony on the Y-K Delta. These results indi- cate that growth rate of goslings at dispersed nesting aggregations on the Y-K Delta is be- low the maximum possible, which is also the case at the major colonies on the Y-K Delta. Gosling growth directly affects first-year survival and consequently recruitment into the breeding population (Owen and Black 1989, Cooch et al. 1991, Sedinger et al. 1995). Our findings suggest that recruitment rates are un- likely to be greater for goslings from dis- Nicolai et al. • GROWTH RATES OF DISPERSED NESTING BRANT 761 TABLE 3. Rankings of brood rearing areas by measurement and year based on Least Squares Means. Lowest values indicate largest size among brood rearing areas within measurement and year. Brood rearing areas not sampled or for which no web-tagged goslings were captured are denoted N/S. Mass Tarsus Culmen Colony/ Brood rearing area 1999 2000 1999 2000 1999 2000 Aknerkochik River Camp 8 N/S 8 N/S 6 N/S Peninsula 6 N/S 6 N/S 4 N/S Big Slough Horseshoe Lake 9 9 9 9 9 / Opagaryak River N/S 6 N/S 8 N/S 9 Tutakoke River Camp 4 8 4 6 7 5 Bend Colony 5 3 5 7 1 6 Kash-Tut 7 7 7 4 8 8 Bend Slough 3 5 3 5 5 3 Hock Slough 2 4 2 3 3 1 Onumtuk 1 2 1 2 2 2 Emperor Bend N/S 1 N/S 1 N/S 4 800 600 400 200 Age (days) FIG. 2. Model averaged Least Squares Means for gosling mass (g) vs. age (days). Filled symbols and heavy lines represent brood rearing areas for dispersed nesting aggregations. Open symbols and thin lines represent brood rearing areas from the Tutakoke River Colony. Only individual broods are presented as symbols. Gosling mass was averaged among brood mates within the same year and gender. 762 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 TABLE 4. AICc model selection for gosling tarsus length. All models with AAICc <4.0 and the general model are presented. ModeP np^ AICc AAICc Model weight Y S BRA A Y*BRA A*BRA B 30 3810.1 0.0 0.3844 Y S BRA A Y*BRA S*A A*BRA B 31 3812.2 2.1 0.1345 Y S BRA A Y*A Y*BRA A*BRA B 31 3812.2 2.1 0.1345 Y S BRA A Y*S Y*BRA A*BRA B 31 3812.2 2.1 0.1345 Y S BRA A Y*S Y*BRA Y*A S*BRA S*A BRA*AGE B 43 3832.1 22.0 0.0000 ^ Y = Year, S = Gender, BRA = Brood Rearing Area, A = Age, and B = Brood, np = Number of parameters in model. persed nesting aggregations compared to those from major colonies. Additionally, nest suc- cess is typically lower in dispersed breeding aggregations than major colonies (0.2— 0.5 vs. 0.7-0. 8) (Raveling 1989; J. S. Sedinger, un- publ. data), which was also true during our study. Nest success in the dispersed nesting aggregations averaged 65% between 1998 and 2000 compared to 80% in the TRC (unpubl. data). We suggest that natal or breeding dis- persal from major colonies on the Y-K Delta to dispersed nesting areas would not improve foraging conditions experienced by goslings during the growth period. The hypothesis that dispersers reduce the influence of local den- sity dependence during brood rearing depends 80 70 60 50 40 I 30 « 80 H CO ^ 70 60 50 40 30 • □ 0 0 g I Wi * □ □ 1999 Females Age (days) LIG. 3. Model averaged Least Squares Means for gosling tarsal length (mm) vs. age (days). Lilled symbols and heavy lines represent brood rearing areas for dispersed nesting aggregations. Open symbols and thin lines represent brood rearing areas from the Tutakoke River Colony. Only individual broods are presented as symbols. Gosling tarsal length was averaged among brood mates within the same year and gender. Nicolai et al. • GROWTH RATES OF DISPERSED NESTING BRANT 763 TABLE 5. AICc model selection for gosling culmen length, model are presented. All models with AAICc <4.0 and the general ModeF np*^ AIQ- AAICr Model weight Y S BRA A Y*BRA A*BRA B 30 2644. 1 0.0 0.2710 Y BRA A Y*BRA S*A A*BRA B 31 2644.6 0.5 0.21 10 Y S BRA A Y*A Y*BRA A*BRA B 31 2646.0 1.9 0.1048 Y S BRA A Y*S Y*BRA S*A A*BRA B 32 2646.3 2.2 0.0902 Y S BRA A Y*S Y*BRA A*BRA B 31 2646.3 2.2 0.0902 Y S BRA A Y*A Y*BRA S*A A*BRA B 32 2646.6 2.5 0.0776 Y BRA A Y*A Y*BRA A*BRA B 30 2646.7 2.6 0.0738 Y S BRA A Y*S Y*BRA Y*A S*BRA S*A BRA*AGE B 43 2664.0 19.9 0.0000 a Y = Year, S = Gender, BRA = Brood Rearing Area, A = Age, and B - Brood, np = Number of parameters in model. on their use of brood rearing areas distinct from those used by broods originating from major colonies. We observed a relatively small amount of mixing on brood rearing ar- eas; <8% of goslings from each colony were captured on brood-rearing areas used primar- ily by broods from the other colony. Even if mixing was substantial, it would be another mechanism reducing the benefits of breeding dispersal. We cannot assess the most interesting hy- pothesis: what would have been the fates of the young of dispersing individuals had they not dispersed? It is possible dispersing indi- FIG. 4. Model averaged Least Squares Means for gosling culnien length (inni) vs. age (days), l illed symbols and heavy lines represent brood rearing areas for dispersed nesting aggregations. Open symbols aiul thin lines represent brood rearing areas from the Tutakoke River Colony. Only individual broods are presented as symbols. Gosling culmen length was averaged among brood mates within year aiul gentler. 764 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 TABLE 6. Mixed model variance partitioning for three measurements of growth. Values are percentages of total variance within each measurement. Measurement Effect Mass Culmen Tarsus Year 13.5 1 1.2 I 1.8 Gender 1.3 0.5 1 1.0 Age 36.1 1 1.7 7.0 Brood Rearing Area 9.1 7.2 4.6 Year X Brood Rearing Area 1.4 5.5 6.4 Age X Brood Rearing Area 7.4 5.6 Brood 16.0 9.4 20.8 Residual 22.7 47.1 32.8 viduals achieved some benefit by dispersing, relative to not doing so. It is also possible that, overall, the population recruited more young than would have occurred without dispersal, because individuals within the population showed a combination of fidelity to major col- onies and dispersal to dispersed breeding ag- gregations. For Lesser Snow Geese, however, dispersal provides an absolute benefit: growth rates are higher for individuals that dispersed than for those that did not (Cooch et al. 1993). A reasonable question, therefore, is whether such a pattern exists for Brant? We could not assess parental quality for goslings originating from dispersed nesting aggregations because most adults from these aggregations were unmarked. Thus, we could not evaluate the potential that lower quality parents nested in dispersed nesting aggrega- tions, resulting in slower growth of their gos- lings relative to those from major colonies. Lindberg et al. (1998) showed natal and adult dispersal from TRC, and we encountered marked adults originally marked in areas not associated with these dispersed nesting aggre- gations. Effects of parental quality on gosling growth are typically believed to involve social interactions between broods and variation in dominance (Williams et al. 1994, Loonen et al. 1997, Black et al. 2007). Thus, quality of parents relative to each other, not absolute quality, influences growth rate of goslings. Even if parents from dispersed nesting aggre- gations were of lower quality than those from the major colony, we would not expect re- duced growth rates of goslings from dispersed nesting aggregations as long as they were not competing primarily with broods from the ma- jor colony, for which we found no evidence. Dispersal from major colonies to dispersed nesting aggregations occurs, but goslings of emigrants apparently do not benefit from re- duced density dependence mediated through reduced competition for food. Masses of gos- lings from dispersed nesting aggregations would have reduced first-year survival by about 20%, relative to those from TRC (Sed- inger and Chelgren 2007). Modeling of the Y- K Delta Brant population (J. S. Sedinger and N. D. Chelgren, unpubl. report) indicates that low nest success of Brant in dispersed nesting aggregations is sufficient to reduce X (per ca- pita rate of increase) in these areas to <1. Goslings from dispersed nesting aggregations were relatively small with low first-year sur- vival; our data suggest that recent increases in numbers of Brant nesting in dispersed nesting aggregations must result at least partially from immigration from major colonies. We suggest the hypothesis that dispersed nesting aggre- gations may not be sustained by in situ re- cruitment; rather, they are likely augmented by immigration from major colonies. Dis- persed nesting aggregations may provide some release for major colonies when nesting densities become high, producing higher fit- ness for individual breeders than they would have experienced had they not dispersed, and a greater overall rate of population increase than would have occurred in the absence of dispersal. ACKNOWLEDGMENTS The U.S. Fish and Wildlife Service, Yukon Delta National Wildlife Refuge, U.S. Park Service, National Science Foundation (OPP-9985931), Ducks Unlimited Inc., and the University of Alaska Fairbanks provided funding for this project. J. P. Arturo, M. R. Axelson, T. W. Bentsen, L. M Bernstein, C. J. Brown, B. C. Comstock, T. R. Dixon, T. R. Jones, B. A. Lorenz, R. L. McQuire, William Naneng, K. L. Nicholson, A. J. Nicolai, T. J. Olson, B. T. Person, D. R. Ruthrauff, M. B. Sato, J. L. Schamber, and C. R. Villa provided field assistance. Logistical support was provided by Yukon Delta National Wildlife Refuge. M. P. Herzog provided unpublished data and invaluable suggestions for statis- tical analysis. Graduate committee members, R. W. Ruess and P. S. Barboza provided valuable insight throughout this project. P. L. Flint, R. F. Rockwell, and J. A. Schmutz commented on an earlier draft of the manuscript. Nicolai et al. • GROWTH RATES OF DISPERSED NESTING BRANT 765 LITERATURE CITED Alliston, W. G. 1975. Web tagging ducklings in pipped eggs. Journal of Wildlife Management 39: 625-628. Anderson, D. R., K. P. Burnham, and W. L. Thomp- son. 2000. Null hypothesis testing: problems, prevalence, and an alternative. Journal of Wildlife Management 64:912-923. Anthony, R. M., W. H. Anderson, J. S. Sedinger, and L. L. McDonald. 1995. Estimating populations of nesting Brant using aerial videography. Wildlife Society Bulletin 23:80-87. Aubin, a. E., a. Dzubin, E. H. Dunn, and C. D. Mac- Innes. 1993. Effects of summer feeding area on gosling growth in Snow Geese. Ornis Scandinav- ica 24:255-260. Black, J. M., J. Prop, and K. Larsson. 2007. Wild goose dilemmas. Branta Press, Groningen, The Netherlands. Black, J. M., E. G. Gooch, M. J. J. E. Loonen, R. H. Drent, and M. Owen. 1998. Body size variation in Barnacle Goose colonies: evidence for local saturation of habitats. Norsk Polarinstitutt Skrifter 200:129-140. Burnham, K. P. and D. R. Anderson. 1998. Model selection and inference: a practical information theoretic approach. Springer-Verlag, New York, USA. Gooch, E. G., R. L. Jeefries, R. F. Rockwell, and F. Gooke. 1993. Environmental change and the cost of philopatry: an example in the Lesser Snow Goose. Oecologia 93:128-138. Gooch, E. G., D. B. Lank, A. Dzubin, R. E Rockwell, AND E Gooke. 1991. Body size variation in Lesser Snow Geese: environmental plasticity in gosling growth rates. Ecology 72:503-512. Dzubin, A. and E. G. Gooch. 1992. Measurements of geese: general field methods. Galifornia Water- fowl Association, Sacramento, Galifornia, USA. Flint, P. L. and J. S. Sedinger. 1992. Reproductive implications of egg size in Black Brant. Auk 109: 896-903. Francis, G. M. and E Gooke. 1992. Sexual differences in survival and recovery rates of Lesser Snow Geese. Journal of Wildlife Management 56:287- 296. Harwood, J. 1977. Summer feeding ecology of Lesser Snow Gee.se. Journal of Wildlife Management 41: 48-55. Herzog, M. R 2002. Environmental regulation of growth in Black Brant. Dissertation. University of Alaska, Fairbanks, USA. Kuehl, R. O. 2000. Design of experiments: statistical principles of research design and analysis. Second Edition. Duxbury Press, Pacific Grove. Galifornia. USA. Lar.s.son, K. and P Forslund. 1991. Environmentally induced morphological variation in the FEirnacle Goose, Branta lencopsis. Journal of Evolutionary Biology 4:619-636. Lindberg, M. S., j. S. Sedinger, and P. L. Flint. 1997. Effect of spring environment and nesting phenol- ogy on clutch size in Black Brant. Gondor 99: 381-388. Lindberg, M. S., J. S. Sedinger, D. V. Derksen, and R. F. Rockwell. 1998. Natal and breeding philo- patry in a Black Brant, Branta bernicla nigricans, metapopulation. Ecology 79:1893-1904. Lindholm, a., G. Gauthier, and A. Desrochers. 1994. Effects of hatch date and food supply on gosling growth in arctic-nesting Greater Snow Geese. Gondor 94:898-908. Loonen, M. J. J. E., K. Oosterbeek, and R. H. Drent. 1997. Variation in growth of young and adult size in Barnacle Geese Branta lencopsis: evidence for density dependence. Ardea 85:177-192. Mainguy, j., G. Gauthier, J. F. Giroux, and J. Bety. 2006. Gosling growth and survival in relation to brood movements in Greater Snow Geese {Chen caeriilescens atlantica). Auk 123:1077-1089. Manseau, M. and G. Gauthier. 1993. Interactions be- tween Greater Snow Geese and their rearing hab- itat. Ecology 74:2045-2055. Nicolai, G. A. 2003. Reproductive decisions by Black Brant: mechanisms to synchronize hatch and spa- tial variation in growth rates of goslings. Thesis. University of Alaska, Fairbanks, USA. Owen, M. and J. M. Black. 1989. Factors affecting the survival of Barnacle Geese on migration from the breeding grounds. Journal of Animal Ecology 58:603-617. Person, B. T, G. A. Babcock, and R. W. Ruess. 1998. Forage variation in brood-rearing areas used by Pacific Black Brant Geese on the Yukon-Kusko- kwim Delta, Alaska. Journal of Ecology 86:243- 259. Person, B. T, M. P. Herzog, R. W. Ruess, J. S. Sed- inger, R. M. Anthony, and G. A. Babcock. 2003. Feedback dynamics of grazing lawns: cou- pling vegetation change with animal growth. Oec- ologia 135:583-592. Raveling, D. G. 1989. Nest-predation rates in relation to colony size of Black Brant. Journal of Wildlife Management 53:87-90. SAS Institute. 1989. SAS U.ser’s Guide. Version 6. Fourth Edition. SAS Institute Inc., Gary, North Garolina, USA. ScHMUTZ, J. A. 1993. Survival and pre-fledging body mass in juvenile Emperor Geese. Gondor 95:222- 225. Sedinger, J. S. and N. D. Ghelgren. 2007. Survival and breeding advantages of larger Black Brant goslings: within and among cohort variation. Auk 124:1281-1293. Sedingi-:r. j. S. and P. L. Flint. 1991. Growth rate is negatively correlated with hatch date in Black Brant. Ecology 72:496-502. Si;dingi;r, J. S. and D. G. Ravi;ling. 1984. Dietary selectivity in relation to availability and quality of footl for goslings of Gackling Geese. Auk 101: 295-306. 766 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 Sedinger, J. S. and D. G. Raveling. 1986. Timing of nesting by Canada Geese in relation to the phe- nology and availability of their food plants. Jour- nal of Animal Eeology 55:1083-1 102. Sedinger, J. S., R L. Llint, and M. S. Lindberg. 1995. Environmental influence on life-history traits: growth, survival, and fecundity in Black Brant (Branta heniicla). Ecology 76:2404-2414. Sedinger, J. S., N. D. Chelgren, D. H. Ward, and M. S. Lindberg. 2008. Lidelity and breeding prob- ability related to population density and individual quality on Black Brant Geese {Branta hernicla ni- gricans). Journal of Animal Ecology 77: In press. Sedinger, J. S., M. S. Lindberg, M. W. Eichholz, and N. Chelgren. 1997. Influence of hatch date versus maternal and genetic effects on growth of Black Brant goslings. Auk 114:129-132. Sedinger, J. S., C. J. Lensink, D. H. Ward, R. M. Anthony, M. L. Wege, and G. V. Byrd. 1993. Status and recent dynamics of the Black Brant breeding population. Wildfowl 44:49—59. Sedinger, J. S., M. S. Lindberg, B. T. Person, M. W. Eichholz, M. P. Herzog, and P. L. Flint. 1998. Density-dependent effects on growth, body size, and clutch size in Brant. Auk 115:613-620. Sedinger, J. S., M. P. Herzog, B. T. Person, M. T. Kirk, T. Obritchkewitch, P. P. Martin, and A. A. Stickney. 2001. Large-scale variation in growth of Black Brant goslings related to food availability. Auk 118:1088-1095. Slattery, S. M. 2000. Factors affecting first-year sur- vival in Ross’ Geese. Dissertation. University of Saskatchewan, Saskatoon, Canada. Williams, T. D., M. J. J. E. Loonen, and E Cooke. 1994. Fitness consequences of parental behavior in relation to offspring number in a precocial spe- cies - the Lesser Snow Goose. Auk 1 1 1:563-572. The Wilson Journal of Ornithology 120(4):767— 777, 2008 DIET OF THE YELLOW-KNOBBED CURASSOW IN THE CENTRAL VENEZUELAN LLANOS CAROLINA BERTSCH' 2 AND GUILLERMO R. BARRETO' ABSTRACT. — Curassows (Cracidae) are important components of the avian biomass in neotropical frugiv- orous bird communities. However, their feeding habits and ecological role remain unclear. We identified the diet of wild Yellow-knobbed Curassow {Crax daubentom) based on analyses of feces and direct observations from November 2001 to July 2002 in a tropical dry forest in central Venezuela. We also analyzed stomach contents from specimens collected in different localities throughout the Llanos region. The diet of curassows included fruits (41 and 49% of dry weight in feces and stomach contents, respectively), seeds (15 and 48%), leaves (39 and 0.7%), minerals (stones, earth; 4.3 and 1.1%), and small proportions of flowers, roots, fungus, seedlings, and invertebrates (insects. Order Coleoptera), each <1% of total dry weight. Curassows fed on 26 plant species from 21 families. When food resources for frugivores are scarce during the dry season (Nov-Apr), 47-50% of the diet was a single species {Guazuma iilmifolia, Sterculiaceae) indicating this species can be critical for curassow survival. An increase in consumption of leaves and invertebrates was observed in the rainy season (May-Jul). Most seeds observed in feces (93%; n = 5,408; range = 1-10 mm) were intact suggesting that curassows could have an important role as seed dispersers in this tropical ecosystem. Received 20 November 2007. Accepted 10 April 2008. Food is important in the life history of most organisms. Thus, identification of key food re- sources of a particular species can assist in revealing its habits and behavioral patterns, and its ecological role in the community. This knowledge is also important when selecting conservation areas for endangered species (Jimenez et al. 2001). Cracids are the most threatened family of birds in the Americas (Brooks and Strahl 2000, Brooks 2006); data on diet and feeding habits of this group have been reported for only a few species. Cracids are considered mainly herbivorous, consuming fruits, seeds, and leaves (Delacour and Amadon 2004). Cu- rassows, in particular, appear to consume mostly fruits and seeds but occasionally feed on flowers, leaves, invertebrates, and soil {Mini .salvini, Santamaria and Franco 1994, 2000; Jimenez et al. 1998; Yumoto 1999; M. mini, Torres 1989; M. tuherosa, Gutierrez 1997; Crax alector, Erard and Sabatier 1989, Erard et al. 1991, Erard and Thery 1994, Thery et al. 1994; C. rubra, Sermeno 1997). The diet of curassows may vary throughout the year but fruit consumption is typically ' Laboralorio de Con.servacion y Manejo de Fauna, Departamento de Biologi'a de Organismo.s, Universi- dad Simon Bolfvar, Apartado 89()0(), Caraca.s 1 080- A Venezuela. “Corresponding author; e-mail: daubentoni@gmail. com high (Erard et al. 1991, Santamaria and Fran- co 2000). Variation in consumption of fruits, seeds, and invertebrates has been detected on a daily basis for Mitu salvini and Crax alector (Jimenez et al. 1998, 2001). Unfortunately, few data are available to describe general pat- terns. Identifying curassow food resources is rel- evant to understanding the role these species have in the dynamics of forest ecosystems. Cracids comprise an important component of the biomass of avian communities in neotrop- ical ecosystems (Terborgh 1986a, Strahl and Grajal 1991). They also may constitute key elements in maintenance of plant communities (Begazo and Bodmer 1998), as many curas- sow species move considerably while feeding, and defecate or regurgitate intact seeds (Wen- ny 1993, Caziani and Protomastro 1994, Yu- moto 1999, Mamani-F 2001). The decline of many curassow populations has caused con- cern for possible effects of their extinction on structure and dynamics of neotropical forests (Levey 1994). The Yellow-knobbed Curassow {Crax daii- hentoni) occurs locally in the Llanos, dry woodlands, and gallery and deciduous forests in northern Venezuela and adjacent Colombia, being one of larger frugivores within its dis- tribution (Schaefer 1953, Strahl and Silva 1997, Hilty 2003, Delacour and Amadon 2004). Buchholz and Bertsch (2006) desig- 767 768 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 nated the status of C. daubentoni as Globally Vulnerable (VU A3a, c, d) due to accelerated urban and agrarian growth in northern Vene- zuela, and increasing hunting pressure. How- ever, little is known about the status, biology, habitat requirements, and feeding habits of this species (Delacour and Amadon 2004). Our objectives were to: (1) identify the composition and diversity of the diet of Yel- low-knobbed Curassows in central Venezuela, and (2) examine its ecological role as possible dispersers of the plants they consume. METHODS Study Area. — The study was conducted on a 75,000-ha private cattle ranch (Hato Pinero, 08° 56' N, 68° 05' W) in the central Llanos of Venezuela (State of Cojedes) from November 2001 to July 2002. The study area included savannas, pastures, and deciduous dry and gallery forests. The specific study area was a 30,000-ha dry forest at the center of the ranch. The average annual temperature and precipi- tation are 27.5° C and 1,469.6 mm, respec- tively. There is strong seasonality with a rainy season (Apr-Nov) and a dry season (Dec- Mar) (Barreto and Hernandez 1988, Hato Pi- nero 2005). Different conservation practices (closure to hunting, fire control, protection of forested areas, and maintenance of corridors between pastures) have been established at this ranch since 1953. Consequently, healthy populations of many species occur, which has permitted development of an ecotourism pro- gram. The only other cracid species in this area is the Rufus-vented Chachalaca {Ortalis ruficauda). Diet Composition and Diversity. — We esti- mated composition and diversity of curassow diets based on: (1) fecal analyses, and (2) di- rect observations in the field. Feces were col- lected and observations were made during eight field trips (15-20 days each) during No- vember 2001 to July 2002 (except Jan). Di- versity was estimated by identifying the dif- ferent species comprising the diet of curas- sows. We also analyzed contents of 15 stom- achs from specimens collected between 1961 and 1980 by personnel of the Estacion Bio- logica de Rancho Grande Museum (Ministry of Environment) in different localities of the central Llanos (Portuguesa, Cojedes, and Gua- rico states) and northern Venezuela (Falcon State). Fecal Analyses. — Feces were collected by searching the area around water sources used by curassows, and at feeding and resting sites. Fifteen samples were collected on a monthly basis (90 samples total; 45/season). No sam- ples were collected in July due to extensive flooding which prevented location of intact fe- ces. Fecal samples >2 m apart were selected whenever possible to increase the probability of originating from different individuals. Fe- ces within a 2-m circle were considered one sample as they likely belonged to the same individual. Samples were oven dried at 70° C for 72 hrs and dry mass was measured to the nearest 0.01 g. We separated and weighed the differ- ent items for further identification. Seeds were counted and their status (intact or damaged) recorded. We considered intact seeds as those not showing evident damage (e.g., broken seeds). We also measured the seeds and relat- ed their size to their status through regression analyses to test the suggestion by Santamaria and Franco (2000) that fate of consumed seeds (predated or dispersed) depends on their size. Direct Obserx'ations in the Field. — Curas- sows are elusive and shy; it was rarely pos- sible to track individuals for periods longer than 2 min. Occasionally we could follow in- dividuals while foraging for 10-20 min. We recorded every food the birds consumed whenever possible during these events. We catalogued items as: seeds, leaves, fruits, flowers, and took samples for further identi- fication. Occasionally, we observed curassows eating soil. This was recorded as soil. Monthly composition of the diet was expressed as per- cent of observations in which every item was recorded in relation to the total number of ob- servations. Stomach Contents Analyses. — Stomach contents were washed with water and sieved (1 mm). Items in each sample were separated under a stereoscopic microscope and cata- logued as seeds, leaves, fruits (only in the case of Guazuma sp., the only fruit for which we could separate the seed from the pulp), flow- ers, animal material, minerals (pebbles of dif- ferent sizes differing from soil which could not be identified in feces or stomach contents), etc. Both wet and dry mass of every compo- Bertsch and Barreto • YELLOW-KNOBBED CURASSOW DIET 769 TABLE 1. Proportion and frequency of foods in Yellow-knobbed Curassow feces (n (n = 15), and proportion of observations in which each food category was recorded (n = = 90) and stomachs 192 feeding bouts). Food categories % Total dry weighU Frequency'’ ” Field observations Feces Stomach contents Feces Stomach contents % (# of feeding bouts) Fruits'" 41.39 49.40 0.57 0.53 57.8 (111) Seeds 15.35 48.20 0.92 1.00 16.7 (32) Leaves 38.62 0.65 0.83 0.33 3.1 (6) Minerals 4.30 1.07 0.59 0.40 8.9 (17) Flowers 0.00 0.36 0.00 0.27 12.5 (24) Animals 0.34 0.008 0.11 0.13 0.5 (1) Seedlings & cotyledons 0.00 0.03 0.00 0.07 0.5 (1) Roots 0.00 0.07 0.00 0.13 0.0 (0) Fungi 0.00 0.004 0.00 0.07 0.0 (0) Othei^ 0.00 0.21 0.00 0.47 0.0 (0) ^ (Dry weight of each category/total dry weight) X 100. ^ Proportion of feces or stomachs in which the food occurred (range = 0 to 1 ). Including fruits consumed with or without seeds. Unidentified material. nent were measured to the nearest 0.01 g; items were identified to the species level whenever possible. The data are expressed as percent dry weight (mass) of every item (leaves, fruits, seeds, flowers, animal material) in relation to total mass and frequency (proportion of feces or stomachs in which the item was present). We note the amount of fruit may be under- estimated due to the difficulty of identifying pulp as a separate component. We correlated these variables (Pearson coefficient with arc- sine transformed data) to investigate whether the heaviest components were also the more frequent. We compared proportions in feces with stomachs using a Spearman test. A sig- nificant correlation indicates items with higher proportions in feces will also have higher pro- portions in stomachs (Barreto et al. 1997). Statistical analyses were performed using STATISTICA (V. 5.0) software (StatSoft 1984-2008). RESULTS Fecal Analyses. — The diet of Yellow- knobbed Curassows was mostly fruits (41% dry weight), leaves (39%), seeds (15%), min- erals (4%), and small proportions of flowers, roots, fungus, seedlings, and invertebrates (in- sects, Order Coleoptera), each < 1 % of total dry weight (Table 1). Fruits (mainly Guaznma ulmifolia) and leaves were abundant in the diet. G. ulmifolia seeds and minerals, although not important in terms of weight, were fre- quent in the diet. Guazuma ulmifolia fruits were the most im- portant item from December to April (dry sea- son) and accounted for 45-82% of the total dry weight. Leaves became the main part of the diet (48-89% dry weight) with an increase in consumption of minerals in June (rainy sea- son) (Fig. 1). We found 12 different seed mor- photypes in feces of which four could be iden- tified to species and one to Family. We found 5,815 seeds in the 90 feces analyzed with size ranging from 1 to 10 mm. Ninety-three per- cent {n = 5,408) were intact and the remain- der were partially damaged. We did not find a relationship between damaged seeds and their size {R^ = 0.2476; P = 0.15; n = 12); however, small seeds (1-5 mm) had a low per- cent of damage or no damage at all. Direct Observations in the Field. — One hundred and ninety-two observations of for- aging curassows were made from November 2001 to July 2002. Fruits (dry and fleshy) and seeds were consumed most frequently (Table 1 ). Dry fruits were mainly G. ulmifolia (26.6% of observations) and Samanea saman (7.8%). Fleshy fruits most frequently con- sumed were Mangifera indica (9.9%) and Thalia geniculata (9.4%) (Table 2). Most fleshy fruits were indehiscent (the seeds re- main in the fruit after it has been shed from the parent plant). Curassows consumed a wide array of fruit shapes (round, ellipsoidal, cylin- 770 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 LIG. 1. Monthly variation in diet composition of Yellow-knobbed Curassows based on fecal analyses {n - 90). The number of fecal samples was 15/month. drical, reniform, straight, curved), types (drupes and legumes), and colors (green, brown, yellow, red, white, and orange) mainly from trees and shrubs and, less frequently, from herbs. Curassows also ingested soil and pebbles (8.8%). All other components ap- peared in <5% of observations. Curassows were observed consuming seeds and seedlings embedded in livestock feces, and a curassow was observed consuming the tail of an un- identified lizard. Curassows were observed throughout the study (Nov-Jul) eating mainly fruits (Fig. 2). Seeds were observed being eat- en from February to April. Both seedlings and animal material were consumed during the rainy season. Stomach Content Analyses. — Stomachs contained mainly fruits and seeds with G. ul- mifolia representing nearly 50% dry weight of the diet. Other components represented <1% dry weight (Table 1) including leaves from Hecastotemon completus and Capparis odor- atissima, flowers of a Caesalpinaceae, and root fragments of an unidentified herb. A piece of a Dyscomicetes fungus was detected in one stomach. We found 13 morphotypes of seeds, seven identified to species and three to Family. Seed size ranged from 1 to 20 mm. G. idmifolia and an unidentified Euphorbi- aceae were the most frequent and dominant dry weight components (Pearson; r- = 0.22; P = 0.008; n = 31). Stomach contents dif- fered from fecal contents (Spearman; R = 0.70; gl = 8; P = 0.19) although both stom- achs and fecal contents had G. idmifolia as the most important item. Stomachs contained more seeds, but fewer leaves, compared to fe- ces. Diet Diversity. — The diet of curassows in- cluded 26 plant species in 21 families in the Hato Pinero. Overall, curassows consumed 48 plant species in 29 families (Table 2). Thirty- six types of plant parts could be identified (25 to species, 6 to genus, and 5 to Family). Of these, 16 species provided fruits, 17 seeds, 10 flowers, 6 leaves, and 1 seedling. Legumes (8 species) were most frequently represented in the diet followed by Rubiaceae, Verbenaceae, and Boraginaceae (2 species each). Species consumed varied throughout the study period (Fig. 3). DISCUSSION Diet Composition. — Yellow-knobbed Cu- rassows are frugivorous based on fecal and Bertsch and Barreto • YELLOW-KNOBBED CURASSOW DIET 771 TABLE 2. Diet of Yellow-knobbed Curassows in the Llanos of Venezuela based on fecal and stomach analyses, and field observations. Family and species Food category^ *-' Found in^-^* Family Mimosaceae Acacia glomerosa Se*" Ob, Fe cv. Mimosa pellita Se^^ Fe Enterolobium cyclocarpwn Se, S&C Ob, Sc^i Samanea saman Fr, Se, FI Ob, Sc Family Bignoniaceae Arrabidaea moUisima Le Ob Sp 1 Not identified Se Sc Sp 2 Not identified Se Sc Family Verbenaceae Lantana camara Fr, FI Ob L. trifolia Fr, FI Ob Family Rubiaceae cv. Guettarda sp. Se^^ Sc Genipa americana Fr, Le, Se Ob, Fe*! Family Boraginaceae Cordia sp. FI Ob Cordia tetranda Fr Ob Family Caesalpiniaceae Caesalpinia coriaria Fr Ob Sp 1 Not identified FI Sc Family Sterculiaceae Guaziana idmifolia Fr Fe, Sc, Ob Family Marantaceae Thalia genicidata FI, Fr Ob Family Nyctacinaceae Giiapira olfersiana Le Ob Family Sapindaceae Allophyhis racemosiis Fr, Le Ob Family Apocynaceae Tabernaemontana cymosa Fr, Le Ob Family Combretaceae Combretum alternifolium Fr, FI Ob Family Flacourtiaceae Hecatostemon completus Fr. Le Ob, Sc Family Rhamnaceae Zyzyphus cyclocordia Fr Ob Family Anacardiaceae Mangifera indica Frp Ob Family Chrysobalanaceae Licania pyrifolia Frp Ob Family Capparidaceae Capparis odoratissima Fr, FI, Le Ob. Sc Family Arecaceae (Palmae) Copernicia tectorum Fr Ob Family Alismotaceae Echinodorus panicnlatus Fr Sc Family Polygonaceae Coccoloba ca racasa na Se Ob, Sc Family Cecropiaceae Cecropia sp. FI Ob TABLE 2. Continued. Family and species Food category** Found in^ ** Family Fabaceae cv. Sesbania sp. Se*^ Sc Family Convolvulaceae cv. Convolvus sp. Se^ Sc Family Passifloraceae cv. Passiflora sp. Se*-' Sc Family Cyperaceae cv. Scleria sp. Infrut Sc Family Euphorbiaceae Sp 1 not identified Se*^ Sc Family Poaceae Sp 1 not identified Fr Fe ^ Se = Seed; Fr = Fruits (including seeds); Frp = Fruit pulp (seeds not consumed); Infrut = Infrutescence; FI = Flowers; L = Leaves; S&C = Seedlings and Cotyledons. •’ Ob = field observation; Fe = Feces; Sc = Stomach content. ^ Not certain if the seed was consumed alone or with the fruit. Not certain about the identification of the item. Stomach content analyses, and (direct obser- vations in the fiel(d. They also consumed seeds, leaves, flowers, roots, fungi, animal matter, and soil as reported for other curas- sows (Torres 1989, Erard et al. 1991, Calchi and Perez 1997, Gutierrez 1997, Renjifo and Renjifo 1997, Santamarfa and Franco 2000. Jimenez et al. 2001). Fruits, the most frequently ingested food item, are nutritious and high energy foods. They represent a rich source of carbohydrates while seeds (most consumed with the fruit) contain protein and lipids (Howe and Westley 1988). Despite these features, it is not clear which traits present in fruits affect their choice by birds (reviewed by Levey and Martinez del Rio 2001). Leaves are low calorie foods that are difficult to digest due to high content of structurally complex carbohydrates, including cellulose, hemicellulose and pectins, phenolic polymers such as lignin and tannins, and a variety of toxic secondary compounds includ- ing alkaloids, terpenoids, and cyanogenic compounds (Howe and Westley 1988). These compounds make leaves a low-intake item tor birds. Morton (1978) and Jimenez et al. (2001) observed that consumption of leaves for nutrients (e.g.. nitrogen) can be expected in frugivorous birds. Galliformes (Cracidae. Phasianidae) are among birds which feed upon leaves most frequently with some spe- cies entirely foliovorous (I,(ii»()/)ns. Canachi- 772 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 0 Fruits (including seeds) □ Seeds m Flowers H Fruits (only pulp) □ Leaves ^ Cotyledons, seedlings ■ Animals b Minerals FIG. 2. Monthly variation in diet composition of Yellow-knobbed Curassows based on field observations {n = 192 feeding bouts). Number of observations differed between months and is represented by the monthly number of feeding bouts. Nov Per Thalia geniculata Guazuma ulmifolia — Acacia glomerosa Guapira olfersiana Enterolobium cyclocarpum cv. Mimosa pellita Coccoloba caracasana Hecatostemon completus Caesaipinia coriaria Combretum alternifolium Samanea saman Capparis odoratissima Poaceae sp. 1 Genipa americana Lantana camara Licania pyrifolia Arrabidaea mollisima Allophylus racemosus Lantana trifolia Cordia sp. Zyzyphus cyclocordia Mangifera indica Copernicia tectorum Cordia tetranda Cecropia sp. Tabernaemontana cymosa Months Jan Feb Mar Apr May Jun Jul ? = even when it was not observed for that period, it is probable the item was consumed, because we observed curassows eating it during the month before and after. FIG. 3. Monthly variation in diet diversity (species or Family level) of Yellow-knobbed Curassows based on field observations and fecal analysis. Bertsch and Barreto • YELLOW-KNOBBED CURASSOW DIET 773 tes, and Tetrao\ Morton 1978). Some species of cracids are reported to feed on leaves: the diet of Mitu mitu is 68% leaves in Peru (Torres 1989), and leaves comprised 39 and 27% of the diet of Ortalis canicolis and Penelope per- spicax in Argentina and Colombia, respective- ly (Caziani and Protomastro 1994, Munoz 2004). Consumption of soil (geophagy) by animals has been attributed to requirements for spe- cific nutrients, anti-acids, or detoxifying sub- stances, or to enhance grinding of food in the gizzard (Diamond et al. 1999). Soil comprised 2-10% of the diet of cracids. Nine percent of the birds we observed ingested soil and it comprised 1 and 4% dry weight in stomachs and feces, respectively. Curassows have been observed eating small pebbles which may be related to the need to grind material inside their strong gizzard (Delacour and Amadon 2004). Several species of cracids have been re- ported to consume animal material including insects (ants, butterflies, caterpillars; Caziani and Protomastro 1994), earthworms (Erard et al. 1991, Erard and Thery 1994, Arriaga and Bermudez 1997, Gutierrez 1997, Renjifo and Renjifo 1997), spiders, terrestrial crabs, and snails (Santamaria and Franco 2000), live adult frogs, lizards, and snakes (Torres 1989), bird eggs, chicks, and carcasses of armadillos (Dasypus sp.), pacas (Cnniculus paca), rats (a number of Muridae species), and bats (Chi- roptera) (Renjifo and Renjifo 1997, Santa- maria and Franco 2000). This consumption, in all reported cases, coincided with the repro- ductive period when adults feed chicks a high protein-based diet (Morton 1973). We found small quantities of animal material (0.52% of observations and < 1 % dry weight in both fe- ces and stomachs). This material may have been consumed incidentally while eating fruits, although it coincided with hatching of eggs in April (Kvarnbiick et al. 2008). Seeds as a separate item from fruit may be overestimated as it is difficult to recognize fruit pulp in stomachs or feces. Our observa- tions indicate curassows ingest mainly fruits, but were observed eating seeds on a few oc- casions. Thus, the importance of seeds in fe- ces or stomachs (Table 1) may be an artifact, underestimating the real importance of fruits in the diet. Different methods (fecal and stomach anal- yses, and direct observations) produced dif- ferent results concerning importance of com- ponents other than fruits (Fig. 4). Fecal anal- yses produced proportions of seeds and leaves inverse from those resulting from stomach analyses. A richer diet was revealed by stom- ach content analysis compared to any other method, partially because the food is less pro- cessed making it possible to distinguish fea- tures that are not distinguishable in feces. However, stomach analysis requires collecting the animal. Diet Seasonality. — The diet of Yellow- knobbed Curassows varied throughout the year with a predominance of legumes and dry fruits during the dry season, and a shift to drupes, flowers, leaves, and small proportions of animal material during the rainy season. Food resources and water are scarce during the dry season, which is a particularly difficult time for frugivorous, folivorous, and insectiv- orous animals (Robinson 1986, Nino P 1994). Curassows consumed large quantities of G. ul- mifolia during the dry season, one of the few species with fruits at this time of the year, as most species produce fruits by the onset of the rainy season (Robinson 1986; pers. obs.). G. ulmifolia, a common species in the Llanos, is consumed by a number of frugivores and has been reported as the most important item in the diet of peccaries (Tayassu spp.) in the study area (Barreto et al. 1997); G. ulmifolia represents a keystone species in this environ- ment {sensu Terborgh 1986b). Other cracids have also been reported to depend on key plant species: Mitu mitu in Peru (4 species of Moraceae: Brosinum sp., Clarisia racemosa. Ficus sp., and Pseudolmenia sp.) (Torres 1989), M. salvini in Colombia (Giiarea gui- donia) (Santamaria and Franco 2()()0), and Or- talis canicollis in Argentina {Schinus poyga- nnis and Rivina humilis) (Caziani and Proto- mastro 1994). Availability of resources is higher during the rainy season and curassows had a diverse diet during that period. It is possible that cu- rassows behave as opportunistic foragers, con- suming a wide array of resources as they be- come available. Curassows were observed shifting foraging areas if they contained more seeds, leaves, llowers, and fruits (Bertsch and Barreto 2008). 774 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 Fruits Leaves Seeds Minerals Others FIG. 4. Food categories in the diet of Yellow-knobbed Curassows: feces (A), stomach analysis (B), and field observations (C). Seed Dispersers or Predators? — Are curas- sows dispersers or predators of the seeds they consume? Too few data are available to prop- erly answer this question. Diets vary among species (Levey 1994, Thery et al. 1994) with guans and chachalacas serving as efficient seed dispersers (Terborgh 1986a, Strahl and Grajal 1991, Wenny 1993, Caziani and Pro- tomastro 1994, Mamani-F 2001, Brooks and Fuller 2006), while curassows serve as dis- persers and predators (Erard and Sabatier 1989, Erard and Thery 1994, Yumoto 1999). The difference may be attributed to their di- gestive physiology, as guans and chachalacas can regurgitate seeds they consume and have relatively weaker gizzards than those of cu- rassows, leaving seeds to pass intact through their digestive tract (Delacour and Amadon Bertsch and Barreto • YELLOW-KNOBBED CURASSOW DIET 775 2004). Some studies suggest that seed size can affect seed fate. Larger seeds (6-30 mm) were completely destroyed in the gastrointestinal tracts of Mitu salvini and Crax alector com- pared to smaller (2-5 mm) seeds (Yumoto 1999, Santamaria and Franco 2000). Larger seeds comprised most of the diet of these spe- cies, making them seed predators, but they oc- casionally ingested smaller seeds which were dispersed. Seeds consumed by C. daubentoni in our study were mostly intact in feces, es- pecially those <1-5 mm, in agreement with Santamaria and Franco (2000). Preliminary experiments showed that seeds of G. ulmifolia and S. saman germinated after passing through the digestive tract of birds, suggesting the Yellow-knobbed Curassow may be an im- portant disperser of some of the plants they consume. This study demonstrated that, although C. daubentoni has been considered a forest dwelling species (Schaefer 1953); it also uses resources at the forest border (e.g., G. ulmi- folia) and open areas (e.g., Mangifera indica, Licania pyrifolia). The current habitat of the Yellow-knobbed Curassow is highly frag- mented, and this behavior may be advanta- geous if this species is tolerant of fragmenta- tion (Strahl and Grajal 1991, Borges 1999). Illegal hunting may be the most serious factor threatening Yellow-knobbed Curassow popu- lations throughout their range. ACKNOWLEDGMENTS We thank the owners and personnel of Hato Pinero for facilities and allowing us to conduct this study on their property. We also thank Gertrudis Gamarra for assistance in the field. John Kvarnback, Gabriela Ech- evarria, Loto Kanta Chandra, Oskar Nilsson, Alejandro Nagy, and Francia Medina helped in the field during different times. Francisco Delascio and Elizabeth Perez helped identify botanical samples. Francisco Bisbal from Estacion Biologica de Rancho Grande kindly lent the stomach samples. Comments by Carlos Bosque, Dan Brooks, Pablo Lau, Doug Levey, and Virginia Sanz were useful in clarifying our ideas. This Project was possible thanks to the financial support of Fun- dacion Branger and Organizacion para Estudios Trop- icales (OET) in Costa Rica. LITERATURE CITED Arriaga, I. and S. Bermudkz. 1997. Dicta de la Gua- characa {Orta! is rnficanda) en el Estado Cojedes. Pages 34-47 in The Cracidae: their biology and conservation (S. D. Strahl, S. Beaujon, D. M. Brooks, A. J. Begazo, G. Sedaghatkish, and F. Ol- mos. Editors). Hancock House Publishers, Blaine, Washington, USA. Barreto, G. R. and O. E. Hernandez. 1988. Aspectos bioecologicos de los Baquiros (Tayassu tajacu y T. pecari) en el Estado Cojedes: estudio compar- ativo. Thesis. Universidad Central de Venezuela, Caracas. Barreto, G. R., O. E. Hernandez, and J. Ojasti. 1997. Diet of peccaries {Tayassu tajacu and T. pecari) in a dry forest of Venezuela. Journal of Zoology (London) 241:279-284. Begazo, A. and R. Bodmer. 1998. Use and conser- vation of Cracidae (Aves: Galliformes) in the Pe- ruvian Amazon. Oryx 32:301-309. Bertsch, C. and G. R. Barreto. 2008. Abundancia y area de accion del pauji de copete {Crax dauben- toni) en los llanos centrales de Venezuela. Orni- tologia Neotropical 19:287-293. Borges, S. H. 1999. Relative use of secondary forests by cracids in central Amazonia. Ornitologia Neo- tropical 10:77-80. Brooks, D. M. (Editor). 2006. Conserving cracids: the most threatened family of birds in the Amer- icas. Miscellaneous Publication Number 6. Hous- ton Museum of Natural Science, Houston, Texas, USA. Brooks, D. M. and R. A. Fuller. 2006. Biology and conservation of cracids. Pages 9-21 in Conserv- ing cracids: the most threatened family of birds in the Americas (D. M. Brooks, Editor). Miscella- neous Publication Number 6. Houston Museum of Natural Science, Houston, Texas, USA. Brooks, D. M. and S. D. Strahl. 2000. Curassows, guans and chachalacas. Status survey and conser- vation action plan for cracids 2000-2004. lUCN, Gland, Switzerland. Buchholz, R. and C. Bertsch. 2006. Crax dauben- toni. Pages 91-94 in Conserving cracids: the most threatened family of birds in the Americas (D. M. Brooks, Editor). Miscellaneous Publication Num- ber 6. Houston Museum of Natural Science, Hous- ton, Texas, USA. Calchi, R. and N. Perez. 1997. Notas sobre dos es- pecies pertenecientes a la Farnilia Cracidae. Page 12 in The Cracidae: their biology and conserva- tion (S. D. Strahl, S. Beaujon, D. M. Brooks, A. J. Begazo, G. Sedaghatkish, and E Olmos, Edi- tors). Hancock House Publishers, Blaine, Wash- ington, USA. Caziani, S. and j. Protomastro. 1994. Diet of the Chaco Chachalaca. Wilson Bulletin 106:640-648. Delacour, j. and D. Amadon. 2004. Curassows and related birds. Second Edition. Lynx Edicions and American Museum of Natural History, Barcelona, Spain and New York, USA. Diamond, J., K. D. Bishop, and J. D. Gilardi. 1999. Gcophagy in New Guinea birds. Ibis 141:181- 193. Erard, C. and D. Sabatier. 1989. Role des oiseaux frugivores terrestres dans la dynamique forestiere 776 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 en Guyane frangaise. Proceedings of the Interna- tional Ornithological Congress 19:803-815. Erard, C. and M. Thery. 1994. Frugivorie et ornith- ochorie en foret guyanaise: I’exemple des grands oiseaux terrestres et de la Penelope marail. Alau- da 62:27-31. Erard, C., M. Thery, and D. Sabatier. 1991. Regime alimentaire de Tinamus major (Tinamidae), Crax alector (Cracidae) et Psophia crepitans (Psophi- idae), en foret guyanaise. Gibier Faune Sauvage 8:183-210. Gutierrez, R. 1997. Observaciones de la poblacion de Mitu tuberosa (Linnaeus, 1766) en el Parque Na- cional El Manu, Peru. Pages 80-88 in The Cra- cidae: their biology and conservation (S. D. Strahl, S. Beaujon, D. M. Brooks, A. J. Begazo, G. Sedaghatkish, and F. Olmos, Editors). Hancock House Publishers, Blaine, Washington, USA. Hato Pinero. 2005. Wildlife paradise- Venezuela, http: // www.hatopinero.com (accessed 2 December 2005). Hilty, S. 2003. Birds of Venezuela. Christopher Helm Publishers, London, United Kingdom. Howe, H. and L. Westley. 1988. Ecological relation- ships of plants and animals. Oxford University Press, New York, USA. Jimenez, L, J. Aldana, C. D. Cadena, and J. Forero. 1998. How does the diet of a curassow vary with- in a week? Field Studies of New World Monkeys, La Macarena, Colombia 12:33-40. Jimenez, I., J. L. Parra, M. Agudelo, G. Londono, AND Y. Molina. 2001. Temporal variation in the diet of Black Curassows {Crax alector, Cracidae). Pages 195-211 in Cracid ecology and conserva- tion in the new millenium (D. M. Brooks and F. Gonzalez-Garcia, Editors). Miscellaneous Publi- cation Number 2. Houston Museum of Natural Sciences, Houston, Texas, USA. Kvarnback, j., C. Bertsch, and G. R. Barreto. 2008. Nest-site selection and nesting success of the Yellow-knobbed Curassow {Crax daubentoni) in a fragmented landscape in the Venezuelan lla- nos. Ornitologia Neotropical 19:347-352. Levey, D. 1994. Cracids as seed dispersers: what we need to know and why it is important. Cracid Newsletters 3. Levey, D. and C. Martinez del Rio. 2001. It takes guts (and more) to eat fruit: lessons from avian nutritional ecology. Auk 118:819-831. Mamani-F, a. M. 2001. Historia natural y evaluacion de la densidad poblacional de Ortalis canicollis (Charata) en Izozog, Provincia Cordillera del Dpto. Santa Cruz, Bolivia. Pages 87-100 in Cra- cid ecology and conservation in the new millen- nium (D. M. Brooks and F. Gonzalez-Garcia, Ed- itors). Miscellaneous Publication Number 2. Houston Museum of Natural Sciences, Houston, Texas, USA. Morton, E. S. 1973. On the evolutionary advantages and disadvantages of fruit eating in tropical birds. American Naturalist 107:8-22. Morton, E. S. 1978. Avian arboreal folivores: why not? Pages 123-130 in The ecology of arboreal folivores (G. G. Montgomery, Editor). Smithson- ian Institution Press, Washington, D.C., USA. Munoz, M. 2004. Caracteristicas morfologicas, nutri- cionales y de disponibilidad de los frutos en la dieta de Penelope perspicax (Aves, Cracidae), en el Santuario de Flora y Fauna Otiin, Colombia. Boletm del Grupo de Especialistas en Cracidos. Number 19. Nino P, I. O. 1994. Correlaciones de estructura y pre- ferencia de habitat del Pauji Copete de Plumas {Crax daubentoni) en los llanos del Estado Gua- rico. Thesis. Universidad Central de Venezuela, Caracas. Renjifo, j. and j. T. Renjifo. 1997. Pauxi unicornis: biologia y ecologfa. Pages 89—92 in The Cracidae: their biology and conservation (S. D. Strahl, S. Beaujon, D. M. Brooks, A. J. Begazo, G. Sedagh- atkish, and F. Olmos, Editors). Hancock House Publishers, Blaine, Washington, USA. Robinson, J. G. 1986. Seasonal variation in use of time and space by wedge-capped capuchin monkey Ce- bus olivaceus: implications for foraging theory. Smithsonian Contributions to Zoology Number 431. Santamaria, M. and a. M. Franco. 1994. Historia natural del Paujil Mitu salvini y densidades pob- lacionales de Cracidos en el Parque Nacional Nat- ural Tinigua-Amazonia Colombiana. Informe Fi- nal. Wildlife Conservation Society, Bogata, Co- lombia. Santamaria, M. and A. M. Franco. 2000. Frugivory of Salvin’s Curassow in a rainforest of the Col- ombian Amazon. Wilson Bulletin 112:473-481. Schaefer, E. 1953. Estudio bio-ecologico comparativo sobre algunos Cracidae del norte y centro de Ve- nezuela. Boletm de la Sociedad Venezolana de Ciencias Naturales 15:30-63. Sermeno, a. 1997. Alimentacion y reproduccion del Pau- jil {Crax rubra) en El Salvador. Pages 71-79 in The Cracidae: their biology and conservation (S. D. Strahl, S. Beaujon, D. M. Brooks, A. J. Begazo, G. Sedaghatkish, and E Olmos, Editors). Hancock House Publishers, Blaine, Washington, USA. STATSOFT Inc. 1984-2008. Statistica v 5.0. StatSoft Inc., Tulsa, Oklahoma, USA. Strahl, S. D. and A. Grajal. 1991. Conservation of large avian frugivores and the management of neotropical protected areas. Oryx 25:50-55. Strahl, S. D. and J. L. Silva. 1997. The status of the Family Cracidae in Venezuela. Pages 383-393 in The Cracidae: their biology and conservation (S. D. Strahl, S. Beaujon, D. M. Brooks, A. J. Be- gazo, G. Sedaghatkish, and F. Olmos, Editors). Hancock House Publishers, Blaine, Washington, USA. Terborgh, j. 1986a. Community aspects of frugivory in tropical forests. Pages 371-383 in Frugivores and seed dispersal (A. Estrada and T. H. Fleming, Editors). Dr. W. Junk Publishers, The Hague, Netherlands. Terborgh, J. 1986b. Keystone plant resources in the Bertsch and Barreto • YELLOW-KNOBBED CURASSOW DIET 111 tropical forest. Pages 330-344 in Conservation bi- ology: the science of scarcity and diversity (M. Soule, Editor). Sinauer Associates, Sunderland, Massachusetts, USA. Thery, M., C. Emid, and D. Bate 1994. Diets of the Marail {Penelope marail) and the Black Curassow (Crax alector). Cracid Newsletters 3. Torres, B. 1989. La dieta del paujil {Mini mitu): o las vicisitudes de ser frugivoro. Boletm de Lima 66: 87-90. Wenny, D. 1993. Black Guans and seed dispersal in Costa Rica. Cracid Newsletters 2. Yumoto, T. 1999. Seed dispersal by Salvin’s Curas- sow, Mitu salvini (Cracidae), in a tropical forest of Colombia: direct measurements of dispersal distance. Biotropica 31:654-660. The Wilson Journal of Ornithology 1 20(4):778-783, 2008 LANDSCAPE CONFIGURATION EFFECTS ON DISTRIBUTION AND ABUNDANCE OF WHIP-POOR-WILLS MICHAEL D. WILSON* 2 AND BRYAN D. WATTS* ABSTRACT. — We examined the influence of landscape configuration created by forest regeneration practices on distribution of Whip-poor-wills {Caprimiilgus vociferous) during the breeding season by comparing relative abundance and space use between forest areas (stands > 17 years of age) and regenerating forest edges (regen- eration stand < 6 years of age adjacent to forest area). Regenerating forest edges contained greater (P < 0.001) abundance of Whip-poor-wills (x ± SE = 2.4 ± 0.30 birds/10 ha) than forest areas (0.8 ± 0.1 1 birds/10 ha). Eighty-four percent of detections at regenerating forest edges were from within the regenerating stand. However, Whip-poor-wills within regenerating stands were detected within 100 m of the forested edge with a greater probability (P < 0.001) than expected by chance. The positive response of Whip-poor-wills to forest edges is likely due to proximity and use of foraging habitats. The relatively high number of habitat openings created by some forest regeneration practices provide Whip-poor-wills with foraging opportunities not present in less in- tensively managed forest systems. Eorest management for Whip-poor-wills should consider harvest strategies that maintain the availability of regenerating patches in close proximity to mature forests. Received 18 August 2006. Accepted 30 January 2008. The distribution and abundance of species that depend on resources that are not con- tained within a single patch type are likely influenced by the spatial organization of all required habitat types (Szaro and Jackie 1985, Dunning et al. 1992, Pearson 1993, Watts 1996, Sisk et al. 1997, Ries and Sisk 2004). This is because species that depend on more than one habitat type must broaden their use of a landscape to include all patch types need- ed to meet their requirements. The use of a habitat patch within a heterogeneous land- scape for these species is influenced both by the characteristics of the patch (e.g., food sup- ply, predation risk, competitive pressure, be- havioral constraints) and the characteristics of surrounding patches (Hansson 1977, Poster and Gaines 1991, Johnson et al. 1992, Rode- wald and Yahner 2001). Ultimately it is likely that spatial association of required habitat patches serves as the most important charac- teristic in affecting distribution and abundance of a species in a landscape (Szaro and Jackie 1985, Dunning et al. 1992, Pearson 1993, Watts 1996, Sisk et al. 1997, Ries and Sisk 2004). Whip-poor-wills (Caprirnulgus vociferous) are nocturnal insectivorous birds that typically nest in forested habitat but frequently use open habitats including fallow fields, crop- ' Center for Conservation Biology, College of Wil- liam and Mary, Williamsburg, VA 23187, USA. 2 Corresponding author; e-mail: mdwils@wm.edu land, shrubland, and regenerating pine (Pinus spp.) stands for foraging (Bent 1940, Cooper 1981, Eastman 1991, Peterjohn and Rice 1991). That Whip-poor-wills use resources that can occur in distinctly different habitat types implies their distribution and abundance may be influenced by the spatial configuration of patches within a broader landscape. Thus, Whip-poor-will abundance would be predicted to respond positively to landscapes where the required patches are in close proximity (Ries and Sisk 2004). Lorested landscapes under intensive man- agement often contain a spatial mosaic of dif- ferent age stands where adjacent stands (i.e., patches) are separated by abrupt edges (Thompson et al. 1995). Landscapes that are frequently disturbed by forest regeneration practices may provide Whip-poor-wills with habitat opportunities not available in less-in- tensively managed forest systems. The objec- tive of our study was to investigate the effect of the spatial arrangement of habitat patches created by regeneration forest management on the distribution and abundance of Whip-poor- wills. METHODS Study Area. — Our study was conducted on a 30,000 ha forested tract in eastern North Carolina (~35° 30' N, 76° 60' W) that is man- aged primarily as a loblolly pine {Pinus taeda) plantation. The tract is divided among 1,010 778 Wilson and Watts • WHIP-POOR-WILL DISTRIBUTION 779 forest stands that are individually managed on a 30-35 year rotation schedule. Pine stands were initially planted as seedlings in parallel rows with a stocking level of 800 to 1,200 pine stems/ha. Young regenerating stands (1- 5 years after planting) are characterized by a dense cover of shrubby plants, a high per- centage of ground cover of grasses and forbs, and no overstory (i.e., open). Dominant plants include switch cane {Arundinaria gigantea), sweet pepperbush (Clethra anifolia), highbush blueberry {Vaccinium corymhosum), and blackberry (Rubus sp.)- Pines reach a height of —2.5 m after 6 years. Stands are commer- cially thinned two times (at —12-15 years and 19-21 years after planting) before final har- vest. Mature stands are characterized by dense understory vegetation dominated by switch cane, sweet pepperbush, highbush blueberry, fetterbush {Lyonia lucida), and gallberry {Ilex glabra). Dominant hardwood trees include red maple {Acer rubrum), sweetgum {Liquidam- bar styracifiua), red bay {Persea borbonia), sweet bay {Magnolia virginiana), and tulip poplar {Lirodendron tulipifera). Mature stands are harvested by clear-cutting. A network of logging roads cross the landscape. The stag- gered regime of harvesting and thinning cre- ates a spatial mosaic of different age stands separated by distinct boundaries (Wilson and Watts 1999). Study Design.— -T\\q influence of the config- uration on stands of different age was exam- ined by comparing the number of Whip-poor- wills at survey points along the boundary be- tween two mature forest stands (forested area) to survey points along the boundary between one mature forest stand and one open, regen- erating stand (regenerating edge). All mature forests had been planted 17-28 years prior and had been commercially thinned at least once. Regenerating stands were planted with pines within the last 1-5 years. Twenty-nine forest- ed areas and 28 regenerating edges were se- lected as spatial replicates for study. Criteria for selection included stand size and shape. Small, narrow stands were avoided to reduce contagious effects from edges of other stands not selected for study. We selected a replicate for survey only if adjacent stands shared at least 350 m of edge. Each spatial replicate contained one point count station positioned on a logging road and along the edge between the two focal stands. The location of survey points was chosen to maximize the distance between the boundaries of other stands not se- lected for study. Bird Surveys. — The 57 survey points were divided into two groups (31 and 26 points, respectively) that were sampled on different nights between 13 May and 17 July 1999. All points were visited seven times during this pe- riod. Whip-poor-will activity is known to vary with lunar light intensity (Mills 1986, Wilson and Watts 2006); thus the two groups of points were surveyed on sequential nights over dif- ferent periods of the lunar phase (new moon, first quarter, full moon, last quarter). Surveys began at 0.5 hrs after dusk and ended at least 1 hr before sunrise, and were conducted by driving between points on a predetermined route. The order in which points were visited was reversed between successive visits to re- duce bias due to time of night. Surveys were not conducted during rain or winds >15 km/ hr. One observer stood at the survey point dur- ing each point visit and recorded all birds that performed at least one syllable of the ono- matopoeic “whip-poor-will” vocalization for a period of 5 min. Surveys may be biased to detect calling males because females are not known to routinely use this specific vocali- zation (Cink 2002). Visual detections and oth- er vocalizations, such as grunt calls (Cink 2002), were not recorded to avoid biasing counts near the road edge or by stand type. The positions of all birds were mapped so de- tections could be summarized by the distance from the observer, distance from the edge of the two focal stands, or distance from the edge of any other stand within the vicinity. We attempted to verify the observer’s ac- curacy of estimating the location of Whip- poor-wills from call-count surveys by com- paring these data to locations estimated from radio-marked birds at this site the following year (Wilson 2003). This is not a direct mea- sure an observer’s ability to estimate location because it compares data collected indepen- dently, but we believe il offers the only op- portunity to validate an observer's ability to identify the location of Whip-poor-wills at night. We summarized the estimated distance of Whip-poor-wills to the nearest forest edge for all observations in each data set into 100- m intervals and compared them for indepen- 780 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 dence using analysis. There was no signif- icant difference (x^3 = 3.27, P > 0.10) be- tween locations estimated by using call counts and locations estimated by using radio telem- etry. We assume that Whip-poor-wills used space in a consistent manner between years and this space use can be measured effectively using each technique. Statistical Analyses. — Density values {x ± SE) used to compare abundance between for- est areas and regenerating edges were calcu- lated using only the detections within a spec- ified transect. The width of each transect was standardized among all survey points to limit detections used for this anaylsis to those ob- served within a 100 m perpendicular distance measured from the boundary between stands selected for study. The length of each transect was allowed to vary between individual sur- vey points to concentrate the investigation within a zone between stands selected for study, and to eliminate observations that might have been influenced by other stand edges not selected for study. This was accom- plished by changing the lengths of each tran- sect so it terminated at a distance of 150 m from the edges between other stands not se- lected for study. The survey visit with the highest recorded abundance within each tran- sect was used for the density estimate. Density values were standardized to number of birds/ 10 ha (i.e., 500 m length X 200 m width rect- angular transect) before analysis to accom- modate variation in transect length. Density was compared between edge types using a one-way ANOVA (Statistica 6.1, StatSoft Inc. 1984-2005). Whip-poor-will distribution was also com- pared between two stands that shared a com- mon edge by summing all detections collected during the entire study into 100-m distance classes. This overall sum represents an accu- mulated pattern of space use through time rather than a population index since the sam- ples are not independent. The total number of detections varied among survey nights due to changes in lunar light conditions (Wilson and Watts 2006). However, detection frequencies between distance classes were statistically in- distinguishable across lunar conditions (i.e., <25%, 26-50%, and >50% of moon face il- luminated) (3X2 contingency tables for all X^ values <4.5, all P > 0.20) and the data were pooled. We also compared the influence of stand age on edge use (Whip-poor-will detections) between forested stands and regenerating stands using x^ contingency tables. We used detections for this comparison only within 100 m of the boundary between two adjacent focal stands. Stands of each forested area were ran- domly assigned to one of two groups for this comparison. We examined the influence of stand type on relative edge use by testing for independence between detections 0-100 m and 101-200 m from the edge of adjacent for- ested stands, and adjacent forest and regen- erating stands using x^ contingency tables. The distribution of Whip-poor-wills within regenerating stands was further examined by comparing the distance of all observations from the nearest forest edge to an expected random distance using x^ analysis. This com- parison included the entire perimeter of the regenerating stand and not only the boundary that was selected for density comparisons. Distances were expected to vary geometrical- ly with stand area and shape; thus, separate statistical comparisons were made from five stand size classes; (1) <20 ha {n = 2), (2) 21- 30 ha {n = 4), (3) 31-40 ha {n = 8), (4) 41- 50 ha {n = 4), and (5) stands >51 ha {n = 10). Expected values were generated by cre- ating 30 random locations within each open stand from digitized stand maps using Arc- View 3.2 GIS software (ESRI 1992) and the animal movement extension of Hooge et al. (1999). The distance of random points to the nearest stand edge was measured from GIS coverage. The distances of both observed and random locations from the nearest forested stand edge were classified by 100-m intervals before comparison. RESULTS Regenerating edges supported a significant- ly greater (F, 55 = 27.6, P < 0.001) average density of Whip-poor-wills compared to for- ested areas (T ± SE, 2.4 ± 0.30 birds/10 ha and 0.8 ± 0.1 1 birds/ 10 ha, respectively). The distribution of Whip-poor-wills between ad- jacent stands was influenced by edge type (x^i = 46.4, P < 0.001). The frequency of all Whip-poor-wills detected in forest areas was evenly distributed between adjacent mature Wilson and Watts • WHIP-POOR-WILL DISTRIBUTION 781 50 -40 t ^ ^ ^ ^ ^ 0-100 101-200 202-300 301-400 401-500 >500 Distance class (m) FIG. 1. Deviation (%) between observed locations of Whip-poor-wills within regenerating forest stands and an expected random distribution after classification into distance categories from the nearest forested edge. Positive deviations indicate distance categories where Whip-poor-wills were detected with greater frequency than expected by chance and negative deviations indicate the distance categories where Whip-poor-wills were detected less frequently than by chance. forest stands (x^i = 0.74, P = 0.61) with 45% and 55% of detections occurring in one of two randomly assigned groups, respectively. Whip-poor-wills in regenerating edges were over five times more likely (x^i = 100.4, P < 0.001) to be detected within the regenerating stand (183 of 218 observations, 84%) com- pared to the adjacent forest stand. Stand age had a significant influence on space use between distance observations of 0- 100 m and 101-200 m (x^ = 3.8, P = 0.050) from the stand edge. Whip-poor-wills at re- generating edges had a greater probability (xu = 26.6, P < 0.001) of being delected within a distance of 0-100 m from the forest edge than a distance of 101-200 m from the nearest forest edge. Sixty-four percent (217 of 339) of observations occurred within 100 m of the forest edge. The distribution of Whip-poor- wills did not differ (x^ = 14, P = 0.76) be- tween 0-100 m and 10 1-200 m in forest ar- eas. Fifty-four percent (86 of 157) of obser- vations were recorded within a distance of 100 m from the stand edge. The distribution of Whip-poor-wills within regenerating stands was highly skewed to for- est edges. Whip-poor-wills within regenerat- ing stands were more likely to be detected within 200 m of a forest edge than further away. This result was most pronounced in patches >20 ha where birds were more likely to be detected within 200 m of the forest edge compared to an expected random distribution among all available elasses (all x~\ ~ 91.6, < 0.001: XU = 76.2, P < 0.001; xu = 79.9. P <0.001: and X“s = <55.8, P < ().()() 1 for stands that were 21-30 ha, 31-40 ha, 41-50 ha. and > 50 ha, respectively) (big. 1). De- tections within regenerating stands < 20 ha did not differ from random when compared 782 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 between 0—100 m and 101—200 m distance classes (x^ = 0.30, P = 0.58). The shortest distance between opposing stand edges in patches < 20 ha was generally not > 400 m. DISCUSSION The presence of open regenerating stands had a positive effect on the density of Whip- poor-wills using this forested landscape. This suggests that regenerating stands provide some resource that Whip-poor-wills prefer for foraging or breeding. One possible explana- tion for high use of open habitats is that it provides greater access to prey. Both abun- dance and richness of adult lepidopteran in- sects, the primary food item of Whip-poor- wills (Cink 2002), can be influenced by timber management and have been shown to be greater in clear-cuts than forest interiors (Joki- maeki et al. 1998, Summerville and Crist 2002). Whip-poor-wills forage on aerial prey almost exclusively using a visual field with short, upward-directed flights initiated from on or near the ground. Foraging activity in- creases with lunar light intensity for this spe- cies (Mills 1986), possibly an adaptation for finding food by exploiting back-lit insects. Open habitats receive more lunar illumination and probably provide better opportunities for visual detection of prey. Regenerating stands may therefore provide both a richer prey base and better foraging conditions compared to mature forest stands with well-developed can- opies. Whip-poor-wills are ground nesters that re- quire forested habitats (Bent 1940, Peck and James 1983, Cink 2002) with ground vegeta- tion for nesting. The density of ground vege- tation in our study area varied markedly be- tween forest and regenerating patches. Forest patches contained dense, continuous thickets of ground-level vegetation and regenerating stands had sparse patchy ground vegetation. Regenerating pine stands had a positive ef- fect on use of adjacent mature stands by Whip-poor-wills. Responses to habitat adja- cency often result from resource complemen- tation when two adjacent patches contain dis- tinctly different resources (Forman and God- ron 1986, Dunning et al. 1992, Ries and Sisk 2004). This is consistent with the general be- lief that Whip-poor-wills require forested patches for nesting but use open areas includ- ing agricultural fields, scrub, and marshes for foraging (Cooper 1981, 1982; Alexander and Creswell 1990; Eastman 1991; Peterjohn and Rice 1991; Wang and Brigham 1997). Regen- erating forests that are directly adjacent to ma- ture stands may provide Whip-poor-wills the opportunity to exploit foraging and nesting habitats in close proximity. The distribution of Whip-poor-wills near re- generating stands is consistent with a resource complementation argument. Whip-poor-will distribution within large regenerating stands was skewed to the edges that were adjacent to mature forest areas suggesting that territory placement was influenced by openings created by forest harvesting. This pattern would allow birds to effectively exploit resources from both habitat types. Observations of radio- tracked Whip-poor-wills from a separate study (Wilson 2003) support this suggestion. Males with both forested and regenerating habitats in close proximity typically roosted in forest habitat during the day, although they spent a disproportionately high amount of time for- aging in regenerating stands at night. The response of Whip-poor-wills to forest edges has direct implications for management of Whip-poor-will breeding populations in forested landscapes. Forest harvesting strate- gies that provide sustained yield of edge hab- itat by interspersing harvested patches with mid-rotation patches will favor Whip-poor- will populations. The amount of edge habitat available on a landscape scale is sensitive to both patch size and the level of interspersion between patch types (Franklin and Forman 1987). Management scenarios that use small to moderate patch sizes and spatially orches- trate harvests to maximize interspersion of patch types should be most useful. ACKNOWLEDGMENTS This study was funded through a cooperative agree- ment between Weyerhaeuser Company and the College of William and Mary. We thank M. A. Melchiors and D. A. Miller for the opportunities and administrative support to conduct the study, and Wiliam Barber and Kevin O’ Kane for assistance in selecting study sites. We thank J. B. Cameron for field assistance. This man- uscript was greatly improved after reviews by R. M. Brigham, C. R Woods, and C. E. Braun. Finally, we are grateful to L. C. Whitaker, C. A. Adams, R. L. Peace, Anne Womack, G. D. Sciole, M. E. Roberts, Wilson and Watts • WHIP-POOR-WILL DISTRIBUTION 783 and C. A. Pope for administrative support from the College of William and Mary. LITERATURE CITED Alexander, I. and B. Cresswell. 1990. Foraging by Nightjars Caprimulgus europaeus away from their nesting areas. Ibis 4:568-574. Bent, A. C. 1940. Life histories of North American cuckoos, goatsuckers, hummingbirds and their al- lies. U.S. National Museum Bulletin 176. CiNK, C. L. 2002. Whip-poor-will {Caprimulgus vocif- erous). The birds of North America. Number 620. Cooper, R. J. 1981. Relative abundance of Georgia caprimulgids based on call-counts. Wilson Bulle- tin 93:363-371. Cooper, R. J. 1982. Range expansion of the Whip- poor-will in Georgia. The Oriole 47:1-9. Dunning, J. B., B. J. Danielson, and H. R. Pulliam. 1992. Ecological processes that affect populations in boundary landscapes. Oikos 65:169-175. Eastman, J. 1991. Whip-poor-will. Pages 252-253 in The atlas of breeding birds of Michigan (R. Brew- er, G. A. McPeek, and R. J. Adams Jr., Editors). Michigan State University Press, East Lansing, USA. Environmental Systems Research Institute (ESRI). 1992. ArcView 3.2. Environmental Systems Re- search Institute, Redlands, California, USA. Eorman, R. T. and M. Gordon. 1986. Landscape ecol- ogy. John Wiley and Sons, New York, USA. Eoster, j. and M. S. Gaines. 1991. The effects of a successional habitat mosaic on a small mammal community. Ecology 72:1358-1373. Eranklin, j. E and R. T. Eorman. 1987. Creating land- scape patterns by forest cutting: ecological con- sequences and principles. Landscape Ecology 1: 5-18. Hansson, L. 1977. Spatial dynamics of field voles, Mi- crotus agrestis, in heterogeneous landscapes. Oi- kos 29:539-544. Hooge, P. N., W. Eichenlaub, and E. Solomon. 1999. The animal movement extension to Arcview. Ver- sion 2.0. USDI, Geological Survey, Alaska Sci- ence Center, Biological Science Office, Anchor- age, USA. Johnson, A. R., J. A. Wiens, B. T. Milne, and T. O. Crist. 1992. Animal movements and population dynamics in heterogeneous landscapes. Landscape Ecology 7:63-75. JOKIMAEKI, J., E. HUHTA, J. ItAEMIES, AND R. RaHOKO. 1998. Distribution of arthropods in relation to for- est patch size, edge, and stand characteristics. Ca- nadian Journal of Forest Research 28: 1068— 1072. Mills, A. M. 1986. The inlluence of moonlight on the behavior of goatsuckers. (Caprimulgidae). Auk 103:370-378. Pearson, S. M. 1993. The spatial extent and relative influence of landscape-level factors on wintering bird populations. Landscape Ecology 3:13-18. Peck, G. K. and R. D. James. 1983. Breeding birds of Ontario: nidiology and distribution. Volume 1. Non-passerines. Royal Ontario Museum of Life Sciences Miscellaneous Publication. Toronto, On- tario, Canada. Peterjohn, B. G. and D. L. Rice. 1991. The Ohio breeding bird atlas. Ohio Department of Natural Resources, Columbus, USA. Ries, L. and T. D. Sisk. 2004. A predictive model of edge effects. Ecology 85:2917-2926. Rodewald, a. D. and R. H. Yahner. 2001. Influence of landscape composition on avian community structure and associated mechanisms. Ecology 82: 3493-3504. Sisk, T. D., N. M. Haddad, and P. R. Ehrlich. 1997. Bird assemblages in patchy woodlands: modeling the effects of edge and matrix habitats. Ecological Applications 7:1 170-1 180. STATSOFT Inc. 1984-2005. Statistica (Data Analysis Software System). Version 6.1. StatSoft Inc., Tul- sa, Oklahoma, USA. Summerville, K. S. and T. O. Crist. 2002. Effects of timber harvest on forest lepidoptera: community, guild, and species responses. Ecological Applica- tions 12:820-835. SzARO, R. C. AND M. D. Jackle. 1985. Avian use of a desert riparian island and its adjacent scrub hab- itat. Condor 87:51 1-519. Thompson III, E R., J. R. Probst, and M. G. Raphael. 1995. Impacts of silviculture: overview and man- agement recommendations. Pages 201-219 in Ecology and management of neotropical migra- tory birds: a synthesis and review of critical issues (T. E. Martin and D. M. Einch, Editors). Oxford University Press, New York, USA. Wang, K. and R. M. Brigham. 1997. Roost-site char- acteristics of Common Poorwills, Phalaenoptilus nuttallii, in Saskatchewan. Canadian Field-Natu- ralist 1 1 1 :543-547. Watts, B. D. 1996. Landscape configuration and di- versity hotspots in wintering sparrows. Oecologia 108:512-517. Wilson, M. D. 2003. Distribution, abundance, and home range of the Whip-poor-will {Caprimulgus vociferous) in a managed forest landscape. Thesis. College of William and Mary. Williamsburg. Vir- ginia. USA. Wilson, M. D. and B. D. Waits. 1999. Response of Brown-headed Nuthatches to thinning of pine plantations. Wilson ILilletin 1 1 1:56-60. Wilson, M. D. and B. D. Waits. 2006. Effect of moonlight on detection of Whip-poor-wills: im- plications for long-term monitoring strategies, .lournal of Piekl Ornithology 77:207 21 1. The Wilson Journal of Ornithology 120(4):784— 792, 2008 IMPACT OF HURRICANE WILMA ON MIGRATING BIRDS: THE CASE OF THE CHIMNEY SWIFT MARK DIONNE,’ 2 ' CELINE MAURICE,' JEAN GAUTHIER,' AND FRANCOIS SHAFFER' ABSTRACT. — We documented the effects of hurricane Wilma (2005) on Chimney Swift (Chaetura pelagica) using data from the Quebec Chimney Swift Survey Program and observations of swift mortality during migra- tion. Hurricane Wilma developed in the Caribbean and followed the eastern coast of North America, moving over areas used extensively by migrating birds. Thousands of birds and, among them. Chimney Swifts, were caught and carried by the storm as far as Atlantic Canada and western Europe. At least 727 swifts were reported dead. Chimney Swift numbers in the province of Quebec, Canada, declined significantly the following year, suggesting adverse consequences of the hurricane on this population over a large area. Roost counts declined by an average of 62%; the total Chimney Swift population decreased by approximately 50%. These results suggest that hurricanes can reduce the breeding population size of some migratory bird species. Received 20 August 2007. Accepted 30 January 2008. There is increasing interest in the effect that extreme weather events can have on ecologi- cal processes and wildlife populations (Par- mesan et al. 2000, Jentsch et al. 2007). Ex- treme weather conditions, including those caused by hurricanes, affect not only humans and their structures, but also wildlife by either altering or destroying important habitats (Wauer and Wunderle 1992) or by causing di- rect mortality (Dinsmore and Farnsworth 2006, Newton 2007). Migrating birds origi- nating from eastern North America cross ac- tive hurricane paths and are at risk of encoun- tering violent storms. Existing information on interactions between birds and hurricanes is rare and mostly anecdotal (Newton 2007). Typically, hurricanes, along the Atlantic coast of North America displace birds northward along the Atlantic coast or inland and are be- lieved to kill some birds (Robertson and Mull- er 1961, Tuck 1968, Mills 1969, Anonymous 1998, Brinkley et al. 1997, Brinkley 1999, D’Anna 2004, Pulcinella and Lockyer 2004, NOAA 2005, Dinsmore and Farnsworth 2006). Inclement weather conditions are known to cause significant mortality in birds. (Gessa- man and Worthen 1982, Newton 2007). These ' Environment Canada, Canadian Wildlife Service, Quebec Region, 1141 Route de I’Eglise, P. O. Box 10100, Quebec, QC, GIV 4H5, Canada. -Current address; 541, rue Jerome Apartment 3, Quebec, QC, GIK 5R5, Canada. -"'Corresponding author; e-mail: calumix@hotmail. com events, when they occur during migration, can cause daily mortality rates much greater than the mean daily rate during nesting and win- tering periods. The relative proportion of mor- talities attributable to severe weather during migration is unknown and difficult to evaluate (Newton 2007). The effects of storms during migration are thought to be among possible factors affecting population levels of eastern North American songbirds (Butler 2000). Se- vere weather can also reduce numbers of some local breeding bird populations (Newton 2007). Understanding the adverse conse- quences of weather on birds is critical for de- clining species, species at risk, or species hav- ing low numbers and small, local distribu- tions. A record number of hurricanes occurred in the Atlantic Basin in 2005, causing consider- able damage in many parts of eastern North America (NCDC 2006). We pursued two lines of evidence to examine the effect of hurricane activity on Chimney Swift {Chaetura pelagi- ca) populations. First we summarized existing information on hurricane Wilma and Chimney Swift fallouts following this event to establish direct mortality associated with the hurricane. Second, we used data from a survey program established in 1998 by the Canadian Wildlife Service (CWS) to quantify the impacts of Wil- ma on Chimney Swift breeding population in Quebec. METHODS Hurricane Wilma. — This hurricane coincid- ed with migration of Chimney Swifts across 784 Dionne et al • IMPACTS OF HURRICANES ON CHIMNEY SWIETS 785 EIG.l. Trajectory and wind swaths of hurricane Wilma along the eastern Atlantic coast during October 2005 based on advisories 1 through 43 of the National Weather Service (National Hurricane Center) from the National Oceanographic and Atmospheric Administration. Black dots represent the position of the storm’s eye. The figure was modified from: http://www.nhc.noaa.gov/archive/2005AVILMA_graphics.shtml. Position true at 30.00° N. its path. Wilma began as a tropical depression on 15 October 2005 near Jamaica and moved towards the southwest, becoming a tropical storm on 17 October (Fig. 1; Pasch et al. 2006). The tropical storm proceeded in a west-north-west direction on 1 8 October, gain- ing strength and developing into a hurricane. Wilma became a category 5 hurricane (Saffir- Simpson Hurricane Scale) on 19 October with wind speeds increasing to 278 km/hr. Wind speeds peaked at 296 km/hr, while the esti- mated minimum central pressure intensity reached —882 mb, a new record low for hur- ricanes in the Atlantic Basin. Hurricane Wil- ma lost strength and became a category 4 hur- ricane on 20 October with winds speeds reaching 241 km/hr. The eye’s radius in- creased to 74 km and ranged between 74 and 1 1 1 km for most of the storm’s lifetime. Wil- ma crossed and caused considerable damage to the northeastern Yucatan Peninsula on 22 October, emerging into the Gulf of Mexico on 23 October. After losing some vigor over land, the hurricane increased again in strength until it reached southern Florida, with winds reach- ing 194 km/hr (category 3 hurricane) on 24 October. Maximum winds decreased to 176 km/hr (category 2 hurricane) over Florida, causing extensive damage to the area. Wind speeds increased to 204 km/hr on 25 October. The hurricane lost strength on 26 October and became an extratropical cyclone with its eye —370 km southeast of Halifax, Nova Scotia, Canada. The cyclone was then absorbed by another cyclone over eastern Nova Scotia on 27 October (Fig. 1; Pasch et al. 2006). Hur- ricane-force winds reached —125 km in radius all through the storm, while tropical-force winds varied between 500 and 1,000 km in radius (Pasch et al. 2006). Data Collection. — Hurricane data were ob- tained from the National Hurricane Center (NHC) of the United States' National Weather Service's website; the NHC is part of the Na- tional Oceanographic and Atmospheric Ad- ministration (NOAA). We collected observa- tions of Chimney Swifts after hurricane Wil- ma from a variety of sources to document di- 786 THb W 1 1 SON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 rect hurricane-associated mortalities and sightings. All known birding and “news- group” websites were browsed for potential information on bird fallouts. Some sightings were validated by communicating directly with the originating persons. Concerned or- ganizations in northeastern North America (U.S. Fish and Wildlife Service, university or- nithology departments, birding associations and non-profit organizations providing care or rehabilitation services to injured birds) were contacted for information about unusual bird observations or mortalities after the passage of hurricane Wilma. Some birdwatchers re- ported unusual bird sightings directly to CWS. We used data from the Chimney Swift Sur- vey Program to quantify effects of the hurri- cane on Chimney Swift breeding populations. The Chimney Swift Survey Program was ini- tiated in 1998 by CWS (Quebec Region) to identify and characterize Chimney Swift roosting and nesting sites, and to monitor pop- ulation trends. The study area covers the southern portion of the province; below the species’ northern breeding range limit (Fig. 1). Over 331 different parishes were visited between 1998 and 2006 by volunteers (—100 people every year) and CWS employees. A parish is a religious territory, varying in both size and number of people. A small village usually has one parish while a town or a city may have two or more parishes. Every parish has a church, rectory, and school with old large masonry chimneys often used by Chim- ney Swifts. All potential Chimney Swift roost- ing and nesting sites in every parish were vis- ited systematically starting in late May when swifts begin arriving from wintering areas. Roost surveys were conducted once or twice per week, throughout the breeding season and swifts were counted as they entered the roost. Monitoring started 30 min before sunset and lasted until birds were no longer visible. Sur- veys were conducted during clement weather (no rain and temperatures >15° C). The an- nual maximum number of birds seen at a giv- en roost was compiled for two different time periods; at the beginning and the end of the breeding season. Spring counts conducted too close to the migration period were excluded because daily numbers can vary significantly and do not represent local swifts. Data from 1998 were not used because surveys effort was too low. Survey effort represents the number of roosts surveyed annually. Statistical Analyses. — We calculated a “be- tween-year change index” (BYCI): the max- imum number of swifts recorded in year t + 1 divided by the number recorded at year t for each roost (Peach et al. 2004). Indices greater than one indicated an increase in swift num- bers at a roost; those smaller than one, a de- crease. Counts from a roost at the beginning and the end of the same breeding season were considered as distinct in the analysis, because birds adopt different flocking behavior after the breeding season and hatch-year birds may be present. Due to missing data, normality, and variance assumption violations, annual changes in Chimney Swift numbers at roosts were analyzed using a Friedman’s test (non- parametric two-way ANOVA on ranked val- ues). The ‘between-year change index’ (BYCI) was the response variable while ‘roost’ and ‘period’ were used as the explan- atory categorical variables. Pairwise differ- ences between 2005-2006 and other periods (1999-2000, 2000-2001, 2001-2002, 2002- 2003, 2003-2004, 2004-2005) were tested us- ing the least square means of ranked values. P-values for significant effects were adjusted using a layered Bonferroni correction to ac- count for non-orthogonal multiple compari- sons (Darlington 1990). RESULTS Bird Fallouts. — Unusual Chimney Swift sightings along the northeastern coast of North America began on 26 October, imme- diately after hurricane Wilma lost strength and became an extra-tropical cyclone. Birds were observed in many areas of northeastern North America at times when they are not usually present (Table I). Thousands of Chimney Swifts were observed in Atlantic Canada (Nova Scotia, New Brunswick, Newfound- land), in Saint-Pierre and Miquelon (France), and in Maine (USA). Some notable bird counts per day (sum of the maximum count from different areas) were: 672 (Nova Scotia) and 325 (Saint-Pierre and Miquelon) on 27 October, 600 (New Brunswick) and 577 (Nova Scotia) on 28 October, thousands (Nova Scotia) on 30 October, more than 400 (Nova Scotia) on 31 October, 350 (Nova Sco- tia) on 1 November and —2,000 (Maine) in TABI.K I. Unusual Chimney Swifl sightings in October-November 2005 following passage of hurricane Wilma based on web sites (a/.ores. seawatching. net, birdingt)nthe.net, birdpac.org, birdsireland.com, surfbirds.com, ornithomedia.com, perso.orange.fr,), newsgroups (nf.birds, NatureNB, Nova Scotia Rare Birds Alerts), Alfrey 2005, Htcheberry 2005, Jiguet and Zucca 2005, SIGHS 2005, Dinsmore and Farnsworth 2006, and personal communications. Dionne et al. • IMPACTS OF HURRICANES ON CHIMNEY SWIFTS 787 TABLE 2. Annual results of the Quebec Chimney Swift Survey Program showing maximum bird counts and number of surveyed sites (effort), including nest- ing sites and roosts. Year # Birds Surveyed sites 1999 3,568 58 2000 3,775 49 2001 2,197 36 2002 3,679 79 2003 4,044 52 2004 3,221 107 2005 4,847 183 2006 2,480 177 r-- r-. r- oc 'O, r', ir. ri n r- IT', ri 'I : > :? = _»Oo- I Z Z Z ^ < early November (Table 1). Chimney Swifts were also observed in areas of western Europe including the Azore Islands (Portugal), Ca- nary Islands (Spain), and the Atlantic coasts of France, Ireland, and England (Table 1 ). Many observers also reported Chimney Swift mortalities: at least 727 were reported in northeastern North America (Table 1). Chimney Swift Survey Program. — The total maximal annual count of Chimney Swifts in the province of Quebec, Canada generally in- creased as new roosts were being discovered and monitoring efforts increased (Table 2). Survey efforts in 2005 were at their highest level and counts reached 4,847 birds. Total numbers of Chimney Swifts in 2006 declined by —50% compared to 2005; survey efforts were similar and at their highest level during these 2 years. The Between-Year Change In- dex (BYCI) in Chimney Swift counts among roosts varied with time {F = 4.52; df — 6, 82; P = 0.0005). Chimney Swift numbers at roosts mostly increased prior to hurricane Wil- ma, but showed no consistent trend (Fig. 2). Unadjusted pairwise comparisons showed no signihcant difference in the periods prior to the hurricane, suggesting the mean number of Chimney Swifts/roost remained similar be- tween 1999 and 2005. The BYCI subsequent to hurricane Wilma was signilicantly lower compared to previous periods, except 2000- 2001 (Table 3). These results suggest a de- crease in the mean number of Chimney Swifts/roost after hurricane Wilma. Roost numbers declined on average by 62% (Fig. 2) between 2005 and 2006. 788 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 X 0> ■D c 0> U) c n SI 0 n o> c 1 CD 3 1 1999 - 2000 - 2001 - 2002 - 2003 - 2004 - 2005 2000 2001 2002 2003 2001 2005 2006 Period LIG. 2. Mean between-year index change in Chimney Swift numbers at known roosts in the province of Quebec, Canada. The number of surveyed roosts for each period is indicated between parentheses under the plots. Dotted line represents the mean. The median corresponds to full line within the box’s boundaries which represent the 25th and 75th percentiles. Whiskers represent the 10th and 90th percentiles, and outliers are indicated by dots. DISCUSSION Thousands of Chimney Swifts were ob- served in Atlantic Canada after passage of hurricane Wilma. Swifts are usually well on their way to wintering areas in South America at the time hurricane Wilma reached the Mar- itime Provinces. They normally leave this re- TABLE 3. Between-year change index in roosts (least square means pairwise differences on ranked in- dices) between 2005-2006 and other periods in Que- bec, Canada. Period Adjusted P value* 1999-2000 0.005 2000-2001 0.192 2001-2002 0.007 2002-2003 0.005 2003-2004 0.006 2004-2005 0.002 * P values were adjusted using the layered Bonferroni method. gion to migrate at the end of August and, by September, most have left (Tuft 1986). Chim- ney Swifts depart northernmost range by Sep- tember and southernmost range by mid- to late October (Kyle and Kyle 2005). This species can be observed during migration in the West Indies from August to October (Raffaele et al. 1998). Chimney Swifts are considered to be uncommon to fairly common autumn tran- sients in the northeastern Yucatan Peninsula from late August to early November (Howell and Webb 1995); the species is common on the Caribbean coasts of Costa Rica and Pan- ama in October and November (Ridgely and Gwynne 1989, Stiles and Skutch 1989) and some have been reported as far as Peru in ear- ly November (Plenge et al. 1989). Late re- cords of Chimney Swift sightings in north- eastern North America have been scarce (Er- skine 1992, David 1996). Reports from the Quebec data base Etude des populations Dionne et al • IMPACTS OF HURRICANES ON CHIMNEY SWIETS 789 d’oiseaux du Quebec (EPOQ) show 1,649 Chimney Swift sightings in August, 239 in September, 14 in October, and none in No- vember (Larivee 2007). EPOQ is a data set, which contains over 6 million bird reports made by birdwatchers year round in Quebec. These include 11,968 sightings referring to Chimney Swifts between the end of April to 18 October from 1901 to 2006. No quantita- tive information is available for the Maritime Provinces, but Tuft (1986) mentions that swift sightings later than September are exception- al. Reverse migrations in the Maritime Prov- inces have, on occasion, been attributed to late summer dispersal and dawn reorientation to- wards the coast (Richardson 1982), but the fall event of 2005 strongly suggests that birds were carried off course by hurricane Wilma. This hypothesis is supported by the number of birds observed, timing of their arrival, and the condition in which many of them were found. A similar hurricane (Gladys) in 1968, in terms of intensity, timing, and route, carried Chim- ney Swifts as far as Nova Scotia (Mills 1969). Birds displaced by hurricanes are probably either carried in the eye of the hurricane or blown by strong winds in the outer bands of the storm (Kaufman 1977). Many birds ap- parently become trapped in the eye of the hur- ricane, where the weather is calm, forcing them along the hurricane’s path (Mills 1969, Jones 1999). There are several reports origi- nating from hurricane hunter pilots (Philipps- born 1999, Halverson 2004), cargo ship pas- sengers (Mayhew 1949, NOAA 2005) and in- land hurricane witnesses (Theiss 2005, Dins- more and Earnsworth 2006) of several birds being trapped inside the eye of a hurricane. The mechanisms by which birds become trapped inside a hurricane are not well under- stood. Most birds usually avoid storms and remain grounded until they pass (Sutton 1945, Richardson 1978, Butler 2000). These obser- vations suggest that landbirds displaced by hurricanes are caught when flying over water where they cannot land nor hnd shelter (New- ton 2007). Chimney Swifts from northern and eastern areas of their breeding range converge during fall migration in the Mississippi Valley and, although some fly over Florida, most migrate across the Gulf of Mexico towards the Yuca- tan Peninsula (Lowery 1943). We hypothesize that swifts displaced by hurricane Wilma were caught while crossing the Gulf of Mexico. This explanation is not entirely satisfactory, however, because birds could have landed and sought shelter when the eye of the hurricane passed over southern Florida. Swifts could also have been caught over land either blown by the powerful winds from rain bands around the storm’s center or driven towards the center by the air flowing near the surface and con- verging in the storm’s eye. We suspect that most displaced birds were caught on the Yu- catan Peninsula and over the Gulf of Mexico. Reports of thousands of Chimney Swifts in the storm’s eye as it passed over Martin Coun- ty, Florida on 24 October (Dinsmore and Farnsworth 2006) lend further weight to this hypothesis. We believe that hurricane Wilma caused massive mortality of Chimney Swifts. When land birds are caught in hurricanes, some probably remain in constant flight during the storm’s course and may be unable to feed dur- ing this time. When birds escaped the storm in Atlantic Canada, many were probably in poor physical condition. This suggests that many birds were unable to restore their body condition to resume migration and eventually died of starvation. Average temperature in Nova Scotia (Halifax) from 26 October to 1 1 November 2005 was 6.7 ± 3.1° C (Environ- ment Canada 2007), making flying insects scarce and thermoregulation cost high. Two Chimney Swifts found dead on Grand Manan (New Brunswick) had lost between 30 and 35% of their average normal body weight (Brian Dalzell, Fundy Bird Observatory, pers. comm.). Starvation experiments revealed that adult Common Swifts {Apiis cipiis) usually die when their weight decreases by 30% or more (Koskomies 1950). Many Chimney Swifts were reported to be in an advanced state ol exhaustion (Kyle and Kyle 2006). The total number of deaths is probably largely under- estimated, as many birds may have perished at sea. Numerous Chimney Swifts also died indirectly from the hurricane as a result of seeking shelter in active chimneys. Three hun- dred Chimney Swifts were found dead inside a chimney in use in Charlotte, New Bruns- wick on 28 October 2005 (Daniel Busby, CWS, pers. comm.). riic results IVoin the Quebec Chimney 790 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 200H Swift Survey Program suggest that hurricane Wilma reduced breeding populations of Chim- ney Swifts by as much as 50% over a large area. The Breeding Bird Survey (BBS) results also suggest a strong population decline for Canada between 2005 and 2006 with an an- nual index 50% lower in 2006 than in 2005 (Downes and Collins 2007). Other BBS re- sults suggest that declines may have extended beyond Canada, into northeastern United States (Sauer et al. 2007). Data from the Que- bec Chimney Swift Program agree with exist- ing information that suggests adverse weather during migration can cause decreases between 25 and 90% in breeding bird population num- bers the following year (Newton 2007). Most studies, however, relate to smaller, local bird populations, which experience only temporary declines (Newton 2007). The survey area is large for the Quebec Chimney Swift breeding population, covering —200,000 km^ (Fig. 1). Hurricane Wilma may possibly exacerbate the decline of Chimney Swifts. The BBS results indicate that Chimney Swifts have been de- clining in North America since 1966 (Sauer et al. 2007) and in Quebec since 1968 (Downes and Collins 2007). The Canadian breeding Chimney Swift population is estimated at 10,000 breeding birds with 2,500 in Quebec (Gauthier et al. 2007). The Committee on the Status of Endangered Wildlife in Canada (CO- SEWIC) has recommended granting the status of ‘threatened’ to the species (COSEWIC 2007). Hurricanes could potentially accelerate declines of species such as Chimney Swift. Long term monitoring is necessary to learn if breeding numbers of Chimney Swift will re- cover from the effects of hurricane Wilma. Our results demonstrate that hurricanes dur- ing migration can significantly impact breed- ing populations of some landbirds over wide areas. However, more typical hurricanes and severe storms may also affect population lev- els of some eastern songbirds (Butler 2000). Eactors such as intensity, trajectory, frequen- cy, and timing of storms and hurricanes are likely important in affecting birds. Extreme weather conditions including those caused by hurricanes could increase in frequency with climate change (Parry et al. 2007), potentially affecting more birds during migration. Further research is needed to identify species at most risk of suffering from severe weather condi- tions during migration, and to understand why and how birds get trapped by hurricanes. A continental network of observers is also nec- essary to standardize information on bird fall- outs to increase the analytical power of ac- quired data and detecting actual impacts of hurricanes on different bird species. ACKNOWLEDGMENTS The authors thank all volunteer observers from the Quebec Chimney Swift Survey Program, including Regroupement QuebecOiseaux, and Quebec’s minis- tere des Ressources naturelles et de la Faune, which provided reliable data on the species. We also thank J.-P. L. Savard, Pierre Laporte, and Melanie Cousineau for constructive comments on this manuscript, as well as Michel Melan^on for help on GIS mapping. We are grateful to all the birdwatchers who reported their valuable observations. Funding for this project was provided by the Canadian Wildlife Service, Quebec Region (Environment Canada). LITERATURE CITED Alfrey, P. 2005. American vagrants on the Island of Corvo, Azores, in October 2005. Birding World 18:465-474. Anonymous. 1998. Hurricane Bonnie on Sable Island, Nova Scotia. Birders Journal 7:254. Brinkley, E. S. 1999. The changing seasons: low pres- sure. North American Birds 53:12-19. Brinkley, E. S„ T Hass, and J. Lockyer. 1997. The storms of 96', Part 1: the storms and their asso- ciated birds. Audubon Society Field Notes 5 1 : 819-829. Butler, R. W. 2000. Stormy seas for some North American songbirds: are declines related to severe storms during migration? Auk 1 17:518-522. COSEWIC. 2007. COSEWIC report on the status of the Chimney Swift {Chaetura pelagica) in Cana- da. Committee on the Status of Endangered Wild- life in Canada, Ottawa, Ontario, Canada. D’Anna, W. 2004. Hurricane Isabel (September 2003) in New York. Kingbird 54:14-19. Darlington, R. B. 1990. Regression and linear mod- els. McGraw-Hill, New York, USA. David, N. 1996. Liste commentee des oiseaux du Que- bec. Association quebecoise des groupes d’ornithologues, Montreal, Quebec, Canada. Dinsmore, S. J. and a. Farnsworth. 2006. The changing season: weatherbirds. North American Birds 60:14-26. Downes, C. M. and B. T Collins. 2007. Canadian bird trends web site. Version 2.2. Canadian Wild- life Service, Environment Canada, Gatineau, Que- bec, Canada. Environment Canada. 2007. Online historical daily temperatures from 26 October to 1 1 November 2005 for Halifax International Airport (Nova Scotia). Dionne et al. • IMPACTS OF HURRICANES ON CHIMNEY SWIETS 791 http://climate.weatheroffice.ec.gc.ca/climateData/ almanac_e.html (accessed 13 December 2007). Erskine, a. J. 1992. Atlas of breeding birds of the Maritime Provinces. Nimbus Publication Ltd. and Nova Scotia Museum, Halifax, Nova Scotia, Can- ada. Etcheberry, R. 2005. The Wilma bird fallout, late Oc- tober 2005 in Saint-Pierre et Miquelon, http:// grandcolombier.com/geographie/nature/pdf-nature/ falLmigration_2005_Wilmas_bird.pdf (accessed 13 December 2007). Gauthier, J., M. Dionne, C. Maurice, J. Potvin, M. Cadman, and D. Busby. 2007. Status of the Chimney Swift (Chaetura pelagica) in Canada. Technical Report Series Number 477. Canadian Wildlife Service, Environment Canada, Quebec City, Quebec, Canada. Gessaman, j. a. and G. L. Worthen. 1982. The effect of weather on avian mortality. Utah State Printing Services, Logan, Utah, USA. Halverson, J. 2004. The eyes of hurricane Isabel. Weatherwise, January/Eebruary 2004, Washing- ton, D.C., USA. Howell, S. N. G. and S. Webb. 1995. A guide to the birds of Mexico and northern Central America. Oxford University Press, New York, USA. Jentsch, a., j. Kreyling, and C. Beierkuhnlein. 2007. A new generation of climate-change exper- iments: events, not trends. Erontiers in Ecology and the Environment 5:365-374. JiGUET, E AND M. ZuccA. 2005. Report on a birdwatching trip to the Azores, 25 Oct-4 Nov 2005. http:// azores.seawatching.net/reports/azores_2005_fjmz. pdf (accessed 13 December 2007). Jones, R. M. 1999. Seabirds carried inland by tropical storm Nora. Western Birds 30:185-192. Kaufman, K. 1977. The changing seasons: an intimate look at Kathleex and Oilier avian phenomena of autumn, 1976. American Birds 31:142-152. Koskomies, j. 1950. The life of the swift. Micropus apus (L.), in relation to weather. Annales Acade- miae Scientiarum Eennicae, Series A 4:1-151. Kyle, P. D. and G. Z. Kyle. 2005. America’s myste- rious birds above the fireplace. Texas A&M Uni- versity Press, College Station, USA. Kyle, P. D. and G. Z. Kyle. 2006. Avian aftermath of the 2005 hurricanes. Chaetura 1 1:12-13. Larivee, j. 2007. Etude des populations d’oiseaux du Quebec (EPOQ). Ornithological data base, Re- groupement QuebecOiseaux, Rimouski, Quebec, Canada. Lowery Jr., G. H. 1943. The dispersal of 21,414 Chimney Swifts banded at Baton Rouge, Louisi- ana, with notes on probable migration routes. Pro- ceedings of the Louisiana Academy of Science 7: 56-74. Mayhfiw, j. R. 1949. Atmospheric pressure and bird night. Science 109:403. Mills, E. L. 1969. Hurricane “Gladys” and its orni- thological effect on the Maritimes provinces. Nova Scotia Bird Society Newsletter I 1:6-16. NCDC. 2006. Climate of 2005, Atlantic hurricane sea- son. National Climatic Data Center, http:// www.ncdc.noaa.gOv/oa/climate/research/2005/ hurricanes05.html (accessed 21 August 2006). Newton, I. 2007. Weather-related mass-mortality events in migrants. Ibis 149:453-467. NOAA. 2005. National Oceanographic and Atmo- spheric Administration, http://www.srh.noaa.gov/ jetstream/tropics/tc_structure.htm (accessed 19 October 2005). Parmesan, C., T. L. Root, and M. R. Willig. 2000. Impacts of extreme weather and climate on ter- restrial biota. Bulletin of the American Meteoro- logical Society 81:443-450. Parry, M. L., O. E Canziani, J. P. Palutikof, and co- authors. 2007. Technical Summary. Pages 23-78 in Climate change 2007: impacts, adaptation and vulnerability. Contribution of Working Group II to the Eourth Assessment Report of the Intergov- ernmental Panel on Climate Change (M. L. Parry, O. P. Canziani, J. P. Palutikof, P. J. van der Linden, and C. E. Hanson, Editors). Cambridge University Press, Cambridge, United Kingdom. Pasch, R. j., E. S. Blake, H. D. Cobb III, and D. P. Roberts. 2006. Tropical cyclone report, hurri- cane Wilma 15-25 October 2005. National Hur- ricane Center, http://www.nhc.noaa.gov/pdf/ TCR-AL252005_Wilma.pdf (accessed January 28 2008). Peach, W. J., S. R. Baillie, and S. T. Buckland. 2004. Current practices in the British Trust for Orni- thology constant effort site scheme and compari- sons of temporal changes in mist-net captures with changes in spot-mapping counts at the extensive scale. Studies in Avian Biology 29:46-56. Philippsborn, R. 1999. Perspective from a NOAA hur- ricane hunter. NOAA News Online Story Number 272. National Oceanographic and Atmospheric- Administration. http://www.noaanews.noaa.gov/ stories/s272.htm (accessed 13 December 2007). Plenge, M. a., T. a. Parker III, R. A. Hlighes, and J. P. O’Neill. 1989. Additional notes on the dis- tribution of birds in west-central Peru. Le Gerfaut 79:55-68. PuLCiNELLA, N. AND J. LocKYER. 2004. Hunicane Is- abel’s birds with notes on previous tropical storms and storm related birds. Pennsylvania Birds 17:1 — 33. Raffaele, H., j. Wiley, O. Garrido, A. Kf;u h, and J. Raffaele. 1998. A guide to the birds of the West Indies. Princeton University Press. Princeton. New Jersey. USA. Richardson. J. W. 1978. Timing and amount of birtl migration in relation to weather: a review. Oikos 30:224-272. Richard.son. j. W. 1982. Northeastward reverse mi- gration of birds in Nova Scotia. C'anada. in au- tumn - a radar stmly. Behavioral Ecology and .So- ciobiology 10:193-206. Rii)Gf;i y. R. S. and J. A. Gwynnf; Jr. 1989. A guide to the birds of Panama with ('osta Rica. Nicaragua 792 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 and Honduras. Second Edition. Princeton Univer- sity Press, Princeton, New Jersey, USA. Robertson Jr., W. B., and H. B. Muller. 1961. Wild winds and wildlife. Audubon Magazine 63:308- 311. Sauer, J. R., J. E. Hines, I. Thomas, and J. Fallon. 2007. The North American Breeding Bird Survey, results and analysis 1966-2006. Version 10.13.2007, uses, Patuxtent Wildlife Research Center, Laurel, Maryland, USA. SIGHS. 2005. Wilma birds on Sable Island, October 2005. Sable Island Green Horse Society, Nova Scotia, Canada. http://www.greenhorsesociety. com/Birds %20-%20Wilma/Birds% 20from%20Hurricane%20Wilma.htm (accessed 13 December 2007). Stiles, F. G. and A. F. Skutch. 1989. A guide to the birds of Costa Rica. Cornell University Press, Ith- aca, New York, USA. Sutton, G. M. 1945. Behavior of birds during a Flor- ida hurricane. Auk 62:603-606. Theiss, M. 2005. Hurricane Rita chase account, http:// www.mthurricane.com/HurricaneJRita.htm (accessed 13 December 2007). Tuck, L. M. 1968. Recent Newfoundland bird records. Auk 85:304-31 1. Tuft, R. W. 1986. Birds of Nova Scotia. Third Edition. Nimbus Publishing and the Nova Scotia Museum, Halifax, Canada. Wauer, R. H. and j. M. Wunderle Jr. 1992. The effect of hurricane Hugo on bird populations on St. Croix, U.S. Virgin Islands. Wilson Bulletin 104:656-673. The Wilson Journal of Ornithology l20(4):793-800, 2008 RED-COCKADED WOODPECKER HOME RANGE USE AND MACROHABITAT SELECTION IN A LOBLOLLY-SHORTLEAF PINE FOREST DOUGLAS R. WOOD,' -'* FRANCISCO J. VILELLA,^ AND L. WESLEY BURGER JR.> ABSTRACT. — We examined annual and seasonal home ranges of 41 Red-cockaded Woodpecker {Picoides borealis) groups from 1997 to 1999 in a Mississippi loblolly (Finns taeda) and shortleaf (P. echinata) pine forest. Adaptive kernel annual home-range estimates (x = 43.1 ± 6.3 ha) were more conservative than maximum convex polygon estimates (x = 58.4 ± 4.5 ha). Mean non-nesting season home ranges were 15-20 ha greater than nesting season home ranges. Home ranges were smaller during nesting and increased during the post- fledging period. Compositional analysis revealed that Red-cockaded Woodpeckers selected habitats dispropor- tionate to their availability annually and seasonally. Red-cockaded Woodpeckers selected pine sawtimber, pine poletimber, pine regeneration, and hardwood sawtimber habitats in that order. Home range appears to be a factor of landscape composition and inversely related to habitat quality. Received 28 June 2004. Accepted 30 January 2008. The Red-cockaded Woodpecker (Picoides borealis) is an endangered species endemic to pine ecosystems in the southeastern United States (Jackson 1994). Red-cockaded Wood- peckers excavate cavities in old, live pines {Piniis spp.) and home ranges generally in- clude a component of older pine stands (Jack- son et al. 1979, Conner and O’Halloran 1987, DeLotelle and Epting 1988, Conner et al. 1994). They prefer foraging stands with large, old (>60 yrs) pines and a limited hardwood component (Hooper and Harlow 1986, Jones and Hunt 1996, Zwicker and Walters 1999). Davenport et al. (2000) demonstrated a rela- tionship in North Carolina between reproduc- tive success and habitat quality. Group fitness was related to the number of cavities, limited under-story vegetation, and large diameter pine trees in the canopy. Similarly, Walters et al. (2002) reported that Red-cockaded Wood- pecker fitness was related to moderate densi- ties of large diameter pines, herbaceous ground cover, and removal of mid-story. This species serves as an indicator species tor ma- ture pine ecosystems of the southeastern Unit- ed States (Jackson 1971, Escano 1995). ' Box 9690, Department of Wildlife and lusherie.s, Mississippi State, MS 39762, USA. 2 Current address: PMB 4068, Southeastern Oklahoma State University, 1405 North 4"' Avenue, Durant, OK 74701, USA. ^ Box 9691, Mississippi Cooperative I'ish and Wild- life Research Unit, Mississippi .State, M.S 39762, U.SA. ■•Corresponding author; e-mail: dwood(?Ae.edu Studies throughout the species geographic range demonstrate extensive variation in mean home range size (14-225 ha; Table 1). Doster and James (1998) calculated a range-wide mean of 76.1 ha from published studies. Red- cockaded Woodpeckers also exhibit seasonal shifts in home range. Jerauld et al. (1983) re- ported that woodpeckers constricted their home ranges by 40-60% during nesting. Con- versely, home ranges may be substantially ex- panded during the non-nesting season; Sko- rupa and McFarlane (1976) reported they dou- bled their home range size during the non- nesting season. Habitat loss or habitat quality may influence home range size (Nesbitt et al. 1983, Jackson and Parris 1995, Doster and James 1998). Previous studies of Red-cockaded Wood- pecker home ranges used different field meth- ods, which makes comparisons among .studies difficult. Timing of observations (seasonal or annual), small numbers of groups ob.served/ year, duration of observations, and time be- tween locations all varied among home range studies. For example, several studies used a diel (dawn-dusk) approach (Sherrill and Case 1980, DeLotelle et al. 1995), whereas others used observation periods of 2-14 hrs (Hooper et al. 1982, Hardesty et al. 1997, Doster and James 1998). Similarly, several studies used hourly locations (Hooper et al. 1982), whereas others used locations recorded in 5-min inter- vals (DeLotelle et al. 1987). Different methods have also been used to 793 794 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 TABLE 1. Home range studies throughout the geographic range of the Red-cockaded Woodpecker. Study State Season Forest type # groups X (ha) Bradshaw (1995) VA Annual Loblolly 6 91.3 Crosby (1971) EL Nesting Longleaf 2 17.2 DeLotelle et al. (1995) EL Annual Longleaf 13 129 Doster and James (1998) AR Annual Shortleaf 5 24.8 Epting et al. (1995) FL/GA Annual Mixed 18 79.8 Hardesty et al. (1997) EL Annual Longleaf 20 108.8 Hooper et al. (1982) SC Annual Mixed 24 86.9 Jackson and Parris (1995) LA Annual Mixed 8 135 Nesbitt et al. (1978) EL Non-nesting Mixed 3 69.8 Sherrill and Case (1980) SC Non-nesting Longleaf 4 32.1 Skorupa and McFarlane (1976) SC Annual Mixed 2 41.9 Wood (1983) OK Nesting Slash 1 44.1 Wood et al. (this study) MS Annual Loblolly 41 58.4 calculate home ranges including minimum convex polygon (Sherrill and Case 1980, DeLotelle et al. 1995), maximum convex polygon (Hooper et al. 1982, Jackson and Par- ris 1995), and adaptive kernel (Hardesty et al. 1997). Each estimation method has advantag- es and disadvantages. Minimum or maximum convex polygon methods often inflate home range estimates by including areas not actu- ally used by woodpeckers. Conversely, adap- tive kernel methods incorporate repeated lo- cations and actual use of inclusive habitat types, which may provide more conservative estimates of home range. Animals exhibit habitat selection at multi- ple spatial scales and may use different prox- imate cues at different spatial scales (Johnson 1980, Orians and Whittenberger 1991). Most woodpecker habitat selection studies focus on microhabitat selection because woodpeckers typically occur in relatively homogenous ma- ture pine stands (Jones and Hunt 1996, Zwick- er and Walters 1999). However, some studies have examined within-stand type and age- class selection by Red-cockaded Woodpeck- ers. Hooper and Harlow (1986) in South Car- olina and Jones and Hunt (1996) in Louisiana found that woodpeckers selected pine stands >30 years of age, but avoided pine stands <30 years of age and mature hardwood stands. Many populations of woodpeckers oc- cur in relatively homogenous landscapes, but others occur in highly fragmented landscapes which provide an opportunity to examine ma- crohabitat selection in heterogeneous loblolly {Pinus taeda) and shortleaf {P. echinata) pine forests (USDI 2003). Our objectives were to: (1) estimate Red- cockaded Woodpecker home ranges at three temporal scales (annual, nesting, and non- nesting seasons) in a loblolly-shortleaf pine forest, (2) estimate home range size for a large number (^12) of groups annually, (3) calcu- late home ranges using adaptive kernel and maximum convex polygon methods for com- parison with other studies, and (4) examine macrohabitat selection within home ranges us- ing compositional analysis. METHODS Study Area. — The Bienville National Forest in central Mississippi is 72,216 ha of pine, pine-hardwood, and hardwood stands of dif- fering age classes in a fragmented landscape. Dominant tree species include loblolly and shortleaf pine intermixed with low densities of longleaf pine {Pinus palustris), slash pine {P. ellioti), oak {Quercus spp.), hickory (Carya spp.), and sweetgum {Liquidambar styraci- flua) (Wood 2001). Field Procedures. — We randomly selected, without replacement, a subsample of 15 Red- cockaded Woodpecker groups/year for visual observation from the 104-106 groups at the study site. However, due to the dynamic na- ture of Red-cockaded Woodpecker groups (e.g., groups disband due to predation or dis- persal), some groups were omitted during the course of the study. We observed 41 different groups (1997, n — 15; 1998, n — 14; 1999, n Wood et al. • RED-COCK ADED WOODPECKER HOME RANGE 795 = 12) at Bienville National Forest. We banded 1 15 adult and 54 nestlings to identify individ- uals in each group (Wood et al. 2001); this represents 70% of the individuals in our study groups. We conducted —1,925 hrs of observations from 1997 to 1999 on 41 groups {x = 47 ob- servation hrs/group). Five-hour visual obser- vation periods were performed on each group for 1 year beginning in January and conclud- ing in December. Observation periods were performed sequentially by group throughout the year to approximate equal effort, and each group was followed at least once per month. Each observation period began at first light and contact was maintained for 5 hours (Hooper and Harlow 1986, Engstrom and Sanders 1997). Each observation period was divided into 6-min periods consisting of a 1- min visual observation period followed by a 5-min waiting period when no data were col- lected (Sherrill and Case 1980, Hooper et al. 1982, DeLotelle et al. 1983). Erequent obser- vation periods enabled us to detect rapid changes in habitat use in the fragmented land- scape of Bienville National Forest. One wood- pecker was selected at the beginning of the 1- min observation period; all locations of that bird were marked with flagging and subse- quently georeferenced with a differentially- corrected Global Positioning System unit. Red-cockaded Woodpeckers typically forage in proximity and the whole group can not be observed simultaneously; we believe that se- lecting an individual from the group was rep- resentative of the groups’ location at any giv- en time period. We defined the group as the sampling unit for home-range analysis and pooled all locations from all individuals in each group observed during the visual obser- vation periods. Statistical Analyses. — We used the animal movement extension in ArcView version 3.2 to estimate annual (1 Jan-1 Dec), non-nesting (1 Jan-3 1 Mar and 1 Aug-3 1 Dec), and nesting ( 1 Apr-31 Jul) season home ranges. We calculated 100% maximum convex polygon (MCP) home ranges and 95% adaptive kernel home ranges. Kernel methods account for repeated locations whereas MCP methods do not (Seaman and Powell 1996, Seaman et al. 1999). Adaptive kernel methods are better suited for analysis ol location data collected in a truncated manner (e.g., 5-hr visual observation periods; Swihart and Slade 1985). However, adaptive kernel methods may underestimate home ranges of species that have a defended resource centrally located in their territory (Worton 1989). A bal- anced ANOVA model with group as a random effect (SAS 1985) was used to test for differ- ences in annual, non-nesting, and nesting season home ranges. We used compositional analysis to test for differences between habitat use and availabil- ity within annual, non-nesting, and nesting season home ranges (Aebischer et al. 1993). Habitat types within home ranges were de- fined as those habitats and age classes avail- able at Bienville National Forest. Habitat types included pine regeneration (0-20 yrs), pine poletimber (21-50 yrs), pine sawtimber (51-110 yrs), and hardwood sawtimber (51- 100 yrs). Woodpecker locations were assigned to habitat type using points derived by GPS locations collected in the field. Locations were imported into ArcView and combined with a map layer of habitat types to identify which locations occurred in different habitat types. We summed the proportion of locations for each habitat to 1 .0 for compositional analysis. Spatial Analyst in ArcView was used to cal- culate the proportion of locations available for each habitat type. Compositional analysis was performed using the program RESOURCE SE- LECTION for Windows 1.00b8.4© (Leban 1999). Compositional analysis uses a M ANO- VA on differences in the log ratios ot use and availability proportions to test for a main ellect (Aitchison 1986). Habitat use and availability proportions are log-ratio transformed which makes the observations linearly independent (Aitchison 1986). Compositional analysis uses a generalized likelihood ratio statistic to calculate an overall Chi-square test statistic. Pair-wise comparisons were used, il a main ellect was detected, to test and rank habitat preferences. All statistical tests were evaluated for significance at a = 0.05; home range means ± ,SE are pre- sented. Ri:SUITS Home Rani>e. — MC’P home-range estimates were generally greater than adaptive kernel estimates, although not for the nesting season (Table 2). Both methods detected similar dil- ferences between annual anti seasonal home- 796 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, Ao. 4, December 2008 TABLE 2. Maximum convex polygon (MCP) and adaptive kernel (AK) methods for calculating Red- cockaded Woodpecker annual, non-nesting, and nest- ing season home ranges (ha) at Bienville National For- est. Mississippi. 1997-1999. Method Season SE Range MCP Annual 584 4.5 14.4-201.5 Non-nesting 43.7 3.5 9.8-122.0 Nesting 24.1 5.6-66.0 AK .Annual 43.1 6.3 3.6-252.8 Non-nesting 39.1 2.7 10.0-79.4 Nesting 24.4 5.2 1.0-214.9 range sizes. Annual home ranges (.v = 58.4 r 4.5 ha) using 100% MCP home ranges were larger than nesting season home ranges (_v = 24.^1 = 2.2 ha: Fj J9 = 51.5. P = 0.047). How- ever. annual home ranges were not larger than non-nesting season home ranges i.x = 43.7 = 3.5 ha. Fi 39 = 29.2. P = 0.15). Red-cocka(ie(d \Voo(ipeckers increased their non-nesting sea- son home-range size by —20 ha compared to average nesting season home ranges (F139 = 2147.3. P = 0?02). Adaptive kernel estimates of annual home ranges (.v = 43.1 ± 6.3 ha) were larger than nesting season (.v = 24.4 =: 5.2 ha) home ranges (F139 = 28.47. P = 0.04). but were not larger than non-nesting season home ranges (.v = 39.1 = 2.7 ha: ^1.39 = 2.40. P = 0.34). Red- cockaded Woodpeckers increased their non- nesting season home ranges by —15 ha com- pared with the average nesting season home range (F1.39 = 599.7. P = 0.03). The sum of the nesting and non-nesting sea- son home ranges with both methods often ex- ceeded annual home-range size. Examination of ArcView products of individual home rang- es showed that woodpeckers used core areas near the cavity trees in all seasons, but fre- quently incorporated habitat patches outside the nesting season home range during the non- nesting season. This produced annual home range estimates larger than either seasonal home range estimate for some woodpecker groups. Macrohabitat Selection. — Red-cockaded Woodpeckers selected habitats disproportion- ate to their availability annually (xa ^ 27.0. P < 0.001 ) and during the nesting (xa = 13.4. P = 0.001) and non-nesting (xa = 21.0. P < 0.001) seasons (Table 3). Woodpeckers se- lected (1) pine sawtimber. (2) pine poletimber. (3) pine regeneration, and (4) hardwood saw- timber in that order. Regardless of season, pair-wise comparisons indicated that wood- peckers disproportionately selected pine saw- timber over the other habitat types. Xo differ- ences were detected for habitat selection among pine poletimber. pine regeneration, and hardwood sawtimber. DISCUSSION Home Range. — Both adaptive kernel (.x = 43.1 ha) and MCP (.v = 58.4 ha) estimates of Red-cockaded Woodpecker home ranges at Bienville National Forest were smaller than the 76.1 ha range-wide estimate calculated by Doster and James ( 1998). MCP methods pro- vided a realistic estimate of the furthest extent of Red-cockaded Woodpecker activity. How- TABLE 3. Red-cockaded Woodpecker annual and seasonal habitat use and availability at Bienville National Forest. Mississippi. 1997-1999. Season Habitat t>pe 9t Use (range 1 Availabilits ( range t .Annual Pine sawtimber 87 (18-100) 74 (28-100) Pine poletimber 6 (0-77) 7 (0-31) Pine regeneration 5 (0-24) 14 (0-35) Hardwood sawtimber 2 (0-21) 5 (0-67) Nesting Pine sawtimber 88 (6-100) 82 (28-100) Pine poletimber 7 (0-90) 6 (0-63) Pine regeneration 4 (0-25) 11 (0-53) Hardwood sawtimber 1 (0-24) 1 (0-12) Non-nesting Pine sawtimber 83 (29-100) 77 (37-100) Pine f)oletimber 7 (0-61) 6 (0-38) Pine regeneration 6 (0-26) 14 (0-50) Hardwood sawtimber 4 (0-43) 3 (0-30) Wood et al. • RED-COCKADED WOODPECKER HOME RANGE 797 ever, stands or areas are included in the home- range estimate that may not have been used by Red-cockaded Woodpeckers. Conversely, maximum convex polygon methods did not account for repeated uses of locations. Adap- tive kernel methods more accurately reflected actual habitat use by Red-cockaded Wood- peckers and accounted for repeated locations. Adaptive kernel methods may be more accu- rate in estimating Red-cockaded Woodpecker home range and reflect a more conservative estimate of home range. j Mean Red-cockaded Woodpecker home- range size in the Bienville National Forest was similar to the mean breeding season home range (46.5 ha) in a Florida longleaf pine forest (DeLotelle and Epting 1992) and in southwest j Georgia (47.1 ha) (Engstrom and Sanders 1997). However, mean home range at Bienville National Forest was greater than (24.8 ha) in an Arkansas shortleaf pine forest (Doster and James 1998), where habitat availability may have restricted Red-cockaded Woodpecker home-range size. Similarly, (Bradshaw 1995) at- tributed Red-cockaded Woodpecker home-range size in Virginia to the extent of isolation of ma- ture loblolly pine stands compared to the sur- rounding landscape matrix of pine plantations and agricultural areas. Hooper et al. (1982) reported a mean Red- cockaded Woodpecker home range of 87 ha in a South Carolina longleaf pine forest using the maximum convex polygon method, ap- proximately twice the mean home range at Bienville National Forest, but with a similar range (34-225 ha) to those at Bienxille Na- tional Forest. Hardesty et al. (1997) conducted a similar study in longleaf pine forests at Eglin Air Force Base in Florida using adaptive kernel methods and documented a mean home range of 108.8 ha (range = 64-387 ha). Seasonal Home Range Shifts. — Red-cock- aded Woodpeckers on average increased their non-nesting season home ranges by 15-20 ha; 85% (35/41) of Red-cockaded Woodpecker groups had larger non-nesting season than nest- ing season home ranges. Three of six groups that had larger nesting season than non-nesting season home ranges were in seed tree cuts (i.c., the cavity trees functioned as the seed trees in a pine regeneration stand) or were adjacent to young pine plantations. This frequently required woodpeckers to traverse longer distances to ob- tain prey, especially while feeding nestlings, than groups in mature pine stands (Wood 2001). One group shifted from their active cluster ai'ea to a recently abandoned cluster area to nest, yet returned to their active cavities for roosting dur- ing the nesting season inflating their apparent home-range size. Two other groups exhibited larger nesting season than non-nesting season home ranges. Both groups appeared to occupy high-quality habitat and there is no obvious ex- planation for the larger nesting season home range. Red-cockaded Woodpeckers in a Virginia loblolly pine forest expanded their home rang- es from 46 to 75 ha during the nesting season and from 84 to 167 ha during the non-nesting season (Bradshaw 1995). Similarly, in a South Carolina longleaf pine forest, Skorupa and McFarlane (1976) reported an increase of 72- 113% in Red-cockaded Woodpecker home- range size during the non-nesting season. Conversely, Jerauld et al. (1983) reported that Red-cockaded Woodpeckers used only 36- 47% of their annual home range during the nesting season. However, Hooper et al. (1982) documented larger home ranges during nest- ing than during the non-nesting season. Non- nesting home ranges were 21% smaller than nesting season home ranges. Home Range and Habitat Quality'. — An in- verse relationship between habitat quality and home-range size has been documented for species including Belted Kingfisher (Mega- cerxle alcyon) (Davis 1982), Lapland Long- spurs (Calcarius lapponicus) (Seastedt and MacLean 1979). and warblers (Parulinae) (Morse 1976). Does an inverse relationship between foraging quality and home-range size exist for Red-cockaded Woodpeckers? Theo- retically, Red-cockaded Woodpecker groups in higher quality foraging habitat would have smaller home ranges than woodpecker groups in lower quality habitats (DeLotelle et al. 1987). Red-cockaded Woodpeckers with prox- imate access to suitable foraging habitat would expend less energy to obtain resources than those that must traverse or circumvent unsuitable habitats to meet the same energetic requirements (Doster and James 1998). DeLotelle et al. (1987) reported greater Red-cockaded Woodpecker home-range size in a longleaf pine forest in Florida than home ranges in a longleaf pine forest in South Car- 798 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 olina which had better microhabitat quality. Epting et al. (1995) reported a mean home range of 67 ha in a mixed longleaf-loblolly pine forest at Fort Stewart in Georgia. Home ranges at Fort Stewart were correlated with habitat quality including increased pine basal area and decreased area of hardwood stands. Fpting et al. (1995) compared the Fort Stewart population with one in longleaf pine in Florida that had a mean home range of 1 1 6 ha. They suggested that Red-cockaded Woodpeckers in Georgia had higher quality habitats than those in Florida. Nesbitt et al. (1983) reported large nesting season home ranges (—144 ha) in Florida and speculated that large Red-cockaded Wood- pecker home ranges were due to low habitat quality. However, Nesbitt et al. (1983) could not establish a significant relationship between habitat quality and home range. Hooper et al. (1982) reported a strong relationship between presence of suitable foraging habitat near the nest cluster and Red-cockaded Woodpecker home-range size. Generally, presence of hard- wood stands or pine plantations may increase home-range size without contributing signifi- cantly to prey availability. Macrohabitat Selection. — Compositional analysis at Bienville National Forest revealed that pine sawtimber stands were strongly pre- ferred, even though pine sawtimber stands had high availability. This relationship may par- tially be a function of Red- cockaded Wood- pecker groups within pine sawtimber stands and proximity to pine sawtimber stands man- aged intensively for Red-cockaded Wood- pecker foraging habitat (USD A 1995). Most Red-cockaded Woodpecker nest clusters were surrounded by suitable patches of pine saw- timber, but several were adjacent to pine re- generation, pine poletimber, hardwood saw- timber stands, and agricultural areas. We fre- quently observed Red-cockaded Woodpeckers bypassing these less suitable stands to forage in 90-1 10 year old pine stands. Pine poletimber was the second most fre- quently selected habitat annually and season- ally. However, pine poletimber stands were not used disproportionately to availability and were not considered suitable for cavity place- ment due to their young age (25-50 yrs of age) and high stocking densities (USDI 1985). However, pine poletimber can serve as suit- able foraging habitat (Hooper and Harlow 1986, Wigley et al. 1999, Wood 2001). Pine regeneration stands ranked third in or- der of habitat selection. Pine regeneration stands had a relatively high availability (11- 14%, range = 0-53%), due primarily to in- tensive timber harvest and even-aged silvicul- tural methods (USD A 1995). However, Red- cockaded Woodpecker use (4-6%, range = 0- 26%) of pine regeneration stands was regularly less than its availability. Red-cock- aded Woodpeckers generally avoided pine re- generation stands, even when a group’s cavi- ties were within a pine regeneration stand. Specifically, in the mid-1980s, —25 Red-cock- aded Woodpecker cluster areas were regener- ated as seed tree cuts with active Red-cock- aded Woodpecker cavity trees left as the seed trees. These woodpeckers had to traverse open areas to access foraging sites. Red-cockaded Woodpeckers used pine regeneration stands at low rates, but many of the stems selected for foraging were large dbh seed trees within the regeneration stand. The mean dbh of stems selected for foraging in pine regeneration stands was 47 ± 15 cm (range = 7.5-79 cm; Wood 2001). Thus, many of the stems were large dbh pines; small dbh pines were rarely selected for foraging by Red-cockaded Wood- peckers (Zwicker and Walters 1999). Hardwood sawtimber was the least pre- fened habitat type in the Bienville National Forest. Several hardwood sawtimber stands were proximally located to woodpecker groups and received high seasonal proportion of foraging by Red-cockaded Woodpeckers; however, large pine stems were overwhelm- ingly selected by Red-cockaded Woodpeckers (Wood et al. 2005). Due to overall low avail- ability of hardwood sawtimber, pine poletim- ber, and pine regeneration, use and availability proportions may have exerted leverage on habitat selection models. Thus, the order of selection for habitat types with low availabil- ity and use may be a function of the sensitivity of the statistical models rather than biological selection by Red-cockaded Woodpeckers. ACKNOWLEDGMENTS We thank our field technicians for their dedicated work under difficult conditions. We thank M. C. Brit- tingham, R. N. Conner, K. G. Smith, H. D. Wilkins, and 2 anonymous reviewers for providing comments Wood et al. • RED-COCKADED WOODPECKER HOME RANGE 799 on the manuscript. We are grateful to the U. S. Eorest Service, National Council of the Paper Industry for Air and Stream Improvement, Forest and Wildlife Re- search Center at Mississippi State University, and the Mississippi Cooperative Fish and Wildlife Research Unit for financial support for this research. We thank the Mississippi Department of Wildlife, Fisheries, and Parks for logistical support. LITERATURE CITED Aebischer, N. J., P. a. Robertson, and R. E. Ken- ward. 1993. Compositional analysis of habitat use from animal radio-tracking data. Ecology 74: 1313-1325. Aitchison, j. 1986. The statistical analysis of compo- sitional data. Chapman and Hall, London, United Kingdom. Bradshaw, D. S. 1995. Habitat use by a relict popu- lation of Red-cockaded Woodpeckers in south- eastern Virginia. Pages 482-488 in Red-cockaded Woodpecker: recovery, ecology, and management (D. L. Kulhavy, R. G. Hooper, and R. Costa, Ed- itors). Center for Applied Studies in Forestry, Col- lege of Forestry, Stephen E Austin University, Na- codoches, Texas, USA. Conner, R. N. and K. A. O’Halloran. 1987. Cavity- tree selection by Red-cockaded Woodpeckers as related to growth dynamics of southern pines. Wilson Bulletin 99:398-412. Conner, R. N., D. C. Rudolph, D. Saenz, and R. R. Schaefer. 1994. Heartwood, sapwood, and fungal decay associated with Red-cockaded Woodpecker cavity trees. Journal of Wildlife Management 58: 728-734. Crosby, G. T. 1971. Home range characteristics of the Red-cockaded Woodpecker in north-central Flor- ida. Pages 60-73 in Ecology and management of the Red-cockaded Woodpecker (R. L. Thompson, Editor). USDI, Bureau of Sport Fisheries and Wildlife and Tall Timbers Research Station, Tal- lahassee, Florida, USA. Davenport, D. E., R. A. Lancia, J. R. Walters, and P. D. Doerr. 2000. Red-cockaded woodpeckers: a relationship between reproductive fitness and hab- itat in the North Carolina sandhills. Wildlife So- ciety Bulletin 28:426-434. Davis, J. D. 1982. Territory size in Megaceryle alcyon along a stream habitat. Auk 99:353-362. DeLotelle, R. S. and R. J. Epting. 1988. Selection of old trees for cavity excavation by Red-cockaded Woodpeckers. Wildlife Society Bulletin 16:48-52. DeLotelle, R. S. and R. J. Epting. 1992. Reproduc- tion of the Red-cockaded Woodpecker in central Florida. Wilson Bulletin 104:285-294. DeLotelle, R. S., J. R. Newman, and A. E. Jerauld. 1983. Habitat use by Red-cockaded Woodpeckers in central Florida. Pages 59-67 in Red-cockaded Woodpecker Symposium II (D. A. Wood. Editor). Florida Game and Freshwater Fish Commission. Tallahassee. USA. DeLotei.i.e, R. S.. R. j. Epting. and J. R. Niavman. 1987. Habitat use and territory characteristics of Red-cockaded Woodpeckers in central Florida. Wilson Bulletin 99:202-217. DeLotelle, R. S., R. J. Epting, and G. Demuth. 1995. A 1 2-year study of Red-cockaded Woodpeckers in central Florida. Pages 259-269 in Red-cockaded Woodpecker: recovery, ecology, and management (D. L. Kulhavy, R. G. Hooper, and R. Costa, Ed- itors). Center for Applied Studies in Forestry, Col- lege of Forestry, Stephen F. Austin University, Na- codoches, Texas, USA. Doster, R. H. and D. a. James. 1998. Home range size and foraging habitat of Red-cockaded Wood- peckers in the Ouachita Mountains of Arkansas. Wilson Bulletin 110:110-117. Engstrom, R. T. and E J. Sanders. 1997. Red-cock- aded Woodpecker foraging ecology in an old- growth longleaf pine forest. Wilson Bulletin 109: 203-217. Epting, R. J., R. S. DeLotelle, and T. Beaty. 1995. Red-cockaded Woodpecker territory and habitat use in Georgia and Florida. Pages 270-276 in Red-cockaded Woodpecker: recovery, ecology, and management (D. L. Kulhavy, R. G. Hooper, and R. Costa, Editors). Center for Applied Studies in Forestry, College of Forestry, Stephen F. Austin University, Nacodoches, Texas, USA. Escano, R. E. F. 1995. Red-cockaded Woodpecker ex- tinction or recovery: summary of status and man- agement of our national forests. Pages 28-35 in Red-cockaded Woodpecker: recovery, ecology, and management (D. L. Kulhavy, R. G. Hooper, and R. Costa, Editors). Center for Applied Studies in Forestry, College of Forestry, Stephen E Austin University, Nacodoches, Texas, USA. Hardesty, J. L., K. E. Lucas, and H. F. Percival. 1997. Ecological correlates of Red-cockaded Woodpecker {Picoides borealis) foraging prefer- ence, habitat use, and home range size in north- west Florida (Eglin Air Force Base). Final Report. Florida Cooperative Fish and Wildlife Research Unit, University of Florida, Gainesville, USA. Hooper, R. G. and R. F. Harlow. 1986. Forest stands selected by foraging Red-cockaded Woodpeckers. USDA, Forest Service, Research Paper SE-259. Asheville, North Carolina, USA. Hooper, R. G., L. J. Niles, R. F. Harlow, and G. W. Wood. 1982. Home ranges of Red-cockaded Woodpeckers in coastal South Carolina. Auk 99: 675-682. Jackson. J. A. 1971. The evolution, taxonomy, distri- bution. past populations, and current status of the Red-cockaded Woodpecker. Pages 4-29 in Ecol- ogy and management t>f the Red-cockaded Wood- pecker (R. L. Thompstin, Editor). USDI, Bureau of Sport Fisheries and Wildlife and Tall Timbers Research Station. Tallahassee. Florida, USA. Jackson, J. A. 1994. Red-cockaded Woodpecker (Pi- coidcs borealis). The birds of North America. Number 85. Jackson, .1. A. and S. D. Parris. 1995. The ecology 800 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 of Red-cockaded Woodpeckers associated with construction and use of a multi-purpose range complex at Lort Polk, Louisiana. Pages 277-282 in Red-cockaded Woodpecker: recovery, ecology, and management (D. L. Kulhavy, R. G. Hooper, and R. Costa, Editors). Center for Applied Studies in Lorestry, College of Lorestry, Stephen L. Austin University, Nacodoches, Texas, USA. Jackson, J. A., M. R. Lennartz. and R. G. Hooper. 1979. Tree age and cavity initiation by Red-cock- aded Woodpeckers. Journal of Lorestry 77:102-103. Jerauld, A. E., R. S. DeLotelle, and J. R. Newman. 1983. Restricted Red-cockaded Woodpecker clan movement during reproduction. Pages 97-98 in Red-cockaded Woodpecker Symposium II (D. A. Wood, Editor). Llorida Game and Lreshwater Lish Commission, Tallahassee, USA. Johnson, D. H. 1980. The comparison of usage and availability measurements for evaluating resource preference. Ecology 61:65-71. Jones, C. M. and H. E. Hunt. 1996. Eoraging habitat of the Red-cockaded Woodpecker on the D’Arbonne National Wildlife Refuge, Louisiana. Journal of Eield Ornithology 67:51 1-518. Leban, E a. 1999. Resource selection for Windows: user’s guide. E A. Leban, University of Idaho, Moscow, USA. Morse, D. H. 1976. Variables affecting the density and territory size of breeding spruce-woods warblers. Ecology 57:290-301. Nesbitt, S. A., D. T. Gilbert, antd D. B. Barbolin. 1978. Red-cockaded Woodpecker fall movements in a Rorida flatwoods community. Auk 95:145-151. Nesbitt, S. A., B. A. Harris, A. E. Jerauld, P. J. Skoog, C. B. Brownsmith, and G. L. Evink. 1983. Response of a male Red-cockaded Wood- pecker to drastic habitat alteration. Pages 101-104 in Red-cockaded Woodpecker Symposium II (D. A. Wood, Editor). Elorida Game and Freshwater Fish Commission, Tallahassee. USA. Orians, G. H. and j. F. Whittenberger. 1991. Spatial and temporal scales in habitat selection. American Naturalist 137:29-49. SAS Institute. 1985. SAS user’s guide: statistics. Ver- sion 6 Edition. SAS Institute Inc., Cary, North Carolina. USA. Seaman, D. E. and R. A. Powell. 1996. An evaluation of the accuracy of kernel density estimators for home range analysis. Ecology 77:2075-2085. Seaman, D. E., J. J. Millspaugh, B. J. Kernohan, G. C. Brundige, K. j. Raedeke, and R. A. Gitzen. 1999. Effects of sample size on kernel home range estimates. Journal of Wildlife Management 63: 739-747. Seastedt, T R. and S. F. Maclean. 1979. Territory size and composition in relation to resource abun- dance in Lapland Longspurs breeding in arctic Alaska. Auk 96:131-142. Sherrill, D. M. and V. M. Case. 1980. Winter home range of 4 clans of Red-cockaded Woodpeckers in the Carolina Sandhills. Wilson Bulletin 92:369- 375. Skorupa, j. P. and R. W. McFarlane. 1976. Seasonal variation in foraging territory of Red-cockaded Woodpeckers. Wilson Bulletin 88:662—665. SwiHART. R. K. AND N. A. Slade. 1985. Influence of sampling interval on estimates of home range size. Journal of Wildlife Management 49:1019-1025. U.S. Department of Agriculture (USDA). 1995. Fi- nal environmental impact statement for the man- agement of the Red-cockaded Woodpecker and its habitat on national forests in the southern region. USDA, Forest Service, Management Bulletin R9- MB 73. Asheville, North Carolina, USA. U.S. Department of Interior (USDI). 1985. Red- cockaded Woodpecker recovery plan. USDI, Fish and Wildlife Service, Southeast Region, Atlanta, Georgia, USA. U.S. Department of Interior (USDI). 2003. Red- cockaded Woodpecker recovery plan. USDI, Fish and Wildlife Service, Southeast Region, Atlanta, Georgia, USA. W.xLTERS, J. R., S. J. Daniels, J. H. Carter III. and P. D. Doerr. 2002. Defining quality of Red-cock- aded Woodpecker foraging habitat based on hab- itat use and fitness. Journal of Wildlife Manage- ment 66:1064-1082. WiGLEY. T. B., S. W. Sweeney, and J. R. Sweeney. 1999. Habitat attributes and reproduction of Red- cockaded Woodpeckers in intensively managed forests. Wildlife Society Bulletin 27:801-809. Wood, D. A. 1983. Foraging colony habitat character- istics of the Red-cockaded Woodpecker in Oklahoma. Pages 51-58 in Red-cockaded Wood- pecker Symposium II (D. A. Wood, Editor). Flor- ida Game and Freshwater Fish Commission. Tal- lahassee. USA. Wood, D. R. 2001. Multi-resolution assessment of the foraging and reproductive ecology of Red-cock- aded Woodpeckers in a Mississippi loblolly-shor- tleaf pine forest. Dissertation. Mississippi State University, Starkville, USA. Wood, D. R., F. J. Vilella. and L. W. Burger Jr. 2001. Red-cockaded Woodpecker banding at Bienville National Forest, Mississippi. North American Bird Bander 26:16-19. Wood, D. R., L. W. Burger Jr., and F. J. Vilella. 2005. Red-cockaded Woodpecker {Picoides bo- realis) behavior in a Mississippi loblolly-shortleaf pine forest. Journal of the Mississippi Academy of Sciences 50:168-176. WORTON, B. J. 1989. Kernel methods for estimating the utilization distribution in home-range studies. Ecology 70:164—168. ZwiCKER, S. M. AND J. R. WALTERS. 1999. Selection of pines for foraging by Red-cockaded Wood- peckers. Journal of Wildlife Management 63:843- 852. The Wilson Journal of Ornithology 120(4):80 1—812, 2008 QUALITY OF ANTHROPOGENIC HABITATS FOR GOLDEN- WINGED WARBLERS IN CENTRAL PENNSYLVANIA JACOB E. KUBELi 2^ AND RICHARD H. YAHNER' ABSTRACT. — Populations of Golden-winged Warblers (Vermivoro chrysoptera) are declining dramatically in eastern North America. Success of conservation efforts will rely largely upon identification and management of suitable anthropogenic habitats (e.g., harvested forest and utility rights-of-way [ROWs]). We assessed habitat quality among three habitat types in central Pennsylvania by comparing population density, nesting success, and productivity among 1-ha patch clearcuts (clearcut area), a 60-m-wide utility ROW (wide ROW), and a 20-m- wide utility ROW (narrow ROW) in 2002 and 2003. Golden-winged Warblers did not use any portion of the narrow ROW that occurred outside the clearcut area. Density (territories/ha) did not differ either year (all P > 0.20) between used sectors of the clearcut area (0.47, 0.50) and the wide ROW (0.71, 0.79). Overall nesting success (successful nests/total nests) was not independent {P = 0.012) of habitat type (58% in the clearcut area, 15% in the wide ROW). Overall productivity (young fledged/nesting territory) was greater (P = 0.026) in the clearcut area (2.38) than in the wide ROW (0.57). Our study raises important questions about the suitability of utility ROWS for Golden-winged Warblers. Received 25 August 2006. Accepted 22 March 2008. Long-term data from the North American Breeding Bird Survey (BBS) indicate the Golden-winged Warbler {Vermivora chrysop- tera) has been one of the most rapidly declin- ing neotropical migrants in eastern North America over the past 40 years (Sauer et al. 2007). The species is listed internationally as “Near Threatened” (BirdLife International 2004) and nationally as a “Bird of Conser- vation Concern” (USDI 2002). Regionally, the BBS reports that Golden- winged Warblers have declined in the northeastern United States (USFWS Region 5) at a rate higher than any other bird since 1966 (Sauer et al. 2007). Declines have been especially alarming during the past 15 years (Buehler et al. 2007). A major cause of decline is decreased avail- ability of early successional forest (Confer 1992, Buehler et al. 2007). The amount of this habitat in eastern North America has de- creased during the past 40-60 years through forest maturation, changes in forest-manage- ment practices (i.e., shifts from even-aged to uneven-aged management), decreased rates of farm abandonment (reducing the abundance of old-held habitats), and suppression of natural disturbances (e.g., wildhre, tree-felling and ' School of Fore.sl Resources, Pennsylvania State University, University Park, PA 16802, USA. ^ Current address: Natural tteritage and Paidangered Species Program, Division of Fisheries and Wildlife. I Rabbit Hill Road, Westborough, MA 01581, USA. ^Corresponding author; e-mail; Jacob. kubelC?^state. ma.us flooding by American beaver [Castor cana- densis]) (Askins 2001, Lorimer 2001, Trani et al. 2001). Wildfire policy is unlikely to change in the foreseeable future, beaver are removed from areas near human habitation, and major windstorms (e.g., hurricanes and tornadoes) have long recurrence intervals in the north- eastern United States (Lorimer and White 2003). Thus, extensive increases in early suc- cessional forest are unlikely to occur by “nat- ural” means. Successful conservation of Golden-winged Warblers and other early suc- cessional wildlife will rely, in part, on im- proved quality of existing habitats or in- creased availability of anthropogenic habitats, such as utility rights-of-way (ROWs) and re- cently harvested forest stands (Klaus and Buehler 2001, DeGraaf and Yamasaki 2003, Buehler et al. 2007). No published studies have attempted to as- sess relative qualities of anthropogenic habi- tats for breeding populations of Golden- winged Warblers. Therefore, we examined habitat use and reproductive success at a site in central Pennsylvania during 2002 and 2003 to provide preliminary information about the subject. The objective of our study was to evaluate habitat quality among three anthro- pogenic types by measuring population den- sity, nesting success, and productivity. Our a priori assumplion was that these measures were indicative of greater habitat quality with the latter two probably of greater importance (Van Horne 1983, Hobbs and Hanley 1990, Brawn and Robinson 1996, Jones et al. 2005). 801 802 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 State Game Lands 176 boundary Scotia Range Road (unimproved/dirt) Barrens Grouse Habitat Management Area Reference sector of Barrens GHMA Treated sector of Barrens GHMA (clearcut area) — . . — Electric transmission line (wide ROW) Electric distribution line (narrow ROW) '//////////////////////^^^////^^//^^^^^ •yyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyy^ ‘10-Acre Pond’ •^j^^>^/yyyyyyyyyyyyyyyyyyyyy. , 'y/^^yyyyyyyyyyyyyy///y///^////^^/* \ 'ffyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyy \ ■yyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyy. \ ■/yyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyy* \ v/yyyyyyyyyyyyyyyyyyyyyyyyyyyyy//^///// \ 'ryyyyyyyyyyyyyyyyyyyyyyyyyyyyyy//////^- \ }fyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyy* \ yyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyy/yy \ ryyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyy. j 'yyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyy> \ •yyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyy v yyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyy.s ^/yyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyy* \ vyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyy \ yyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyy/y. \ ^/yyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyy* ryyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyr 'yyyyyyyyyyyyyyyyyyyyyyyyyyyy////^///'J'^ ^yyyyyyyyyyyyyyyyyy^y-r^. ~ ' ^yyyy>u^^’A - - " “ 2000 FIG. 1. Habitat types (clearcut area, wide ROW, and narrow ROW) and other features at State Game Lands 176, Centre County, Pennsylvania during 2002 and 2003. METHODS Study Area. — Our study was conducted at State Game Lands (SGL) 176 (40° 47' N, 77° 57' W) in Centre County, Pennsylvania. This 2,060-ha site has sandy clay soils (Conklin 1943) and consists mainly of aspen {Populus tremuloides and P. grandidentata) and oak (Quercus spp.) cover types (Yahner 1986). Low ridges and a valley characterize the to- pography, and elevation ranges from a low of —350 m in the valley to a high of 470 m on the northwest ridge. Numerous forest distur- bances have occurred at the site since the 1800s (Conklin 1943) and Golden-winged Warblers have been present since at least 1974 (Yahner 2000, 2003b). The study area is sur- rounded by agriculture and suburban devel- opment. Habitat Types. — The Pennsylvania Game Commission (PGC) established the Barrens Grouse Habitat Management Area (GHMA) in a 1,126-ha portion of SGL 176 to study the response of Ruffed Grouse (Bonasa umbellus) to manipulation of habitat via small-scale, ro- tation forest-harvesting (Palmer 2003, Yahner 2003b). The Barrens GHMA consists of a 544-ha treated (systematic, 1-ha patch clear- cutting) sector and a 582-ha reference (uncut) sector (Palmer 2003; Fig. 1); early succes- sional habitat in 136 recent clearcuts of the treated sector (Fig. 2) represents the hrst hab- itat type studied. Each of these 1-ha plots was cut during winters 1999-2001 and retained —5-25 overstory trees, representing a two- aged regeneration method known as “clear- cutting with reserves” (Helms 1998). We refer to the treated sector of the Barrens GHMA as “the clearcut area” to reflect the type of early successional habitat present (i.e., clearcuts and associated access roads and log landings). We defined early successional habitat as having an open canopy; this stage of succession is con- sistent with that described for habitat require- ments of Golden-winged Warblers (i.e., patch- es of herbs, shrubs, and scattered trees adja- cent to a forest [closed-canopy] edge [Confer Kiihel ami Yahner • HABITAT QUALITY FOR GOLDEN-WINGED WARBLERS 803 wwwww w WwwWW WWWWW rrrrr w n; Err-- •' '■) 1- — t7~a 'IT? jj y^yyA -----WW- n n P- P . . . WWWWW i i wwwww i m. WWWWW f 1 t wwwww 1 1 200 m y 1976-1977 1985-1987 0 m □ ^ Uncut ^y 1999-2001 1976-1977 1999-2001 D 1980-1981 B 1985-1987 Scotia Range Road (unimproved/dirt) Recent (1-3 years) clearcut Closed-canopy (15-26 yr-old) aspen forest Closed-canopy (15-90 yr-old) oak forest FIG. 2. Schematic and cutting cycles of the clearcut area at the Barrens Grouse Habitat Management Area. Centre County, Pennsylvania. 2()()2-2()()3. Management units consisted of 1.36 blocks subdivided into 1-ha plots (A-D); dates indicate the winter .season(s) during which plots were harvested. I992|). Major vegetation in the regenerating elearciits included aspen, scrub oak {Q. ilici- foiui), dwarf chinkapin oak (Q. prinoidcs), ta- tarian honeysuckle (Loniccra tatarica), grass- es, and forbs (especially goldenrod ISolidai^o spp.l) in aspen forest, and oak. red maple {Acer ruhrum), grasses, and lerns (especially hayscented fern [Dcnnslacdlia pimcfilohnla]) in oak forest (Kubel 2005). The second habitat type was a 3.7-km por- tion of a 60-m-wide electric transmission line ROW, 2 km west of the clearcut area and termed “the wide ROW” (Fig. I). Penelec OlKMating C'ompany of F'irstldiergy C\trpora- lion iiKiinlains the witle ROW by selective herbicide-aiiidictuion at 5-yetir inteiwals. For- est types bordering the wide ROW' are aspen (sotithern of the ROW) and oak (northern 804 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 40%). Vegetation within the ROW consists primarily of scrub oak, American hazelnut (Corylus americana), tatarian honeysuckle, hawthorns {Craetagus spp.), grasses, and forbs in aspen forest, and brambles (Rubus spp.), blueberry {Vciccinium sp.), sweetfern {Comptonia pere grind), grasses, and ferns in oak forest (Kubel 2005). A 20-m-wide electric distribution line ROW bisects the clearcut area and continues in a northerly direction through the northeast corner of SGL 176 (Fig. 1). We considered the portion of the ROW in the clearcut area as part of the clearcut habitat type because the ROW occurred within and immediately adja- cent to harvested stands. The portion outside the clearcut area (beginning —250 m east of the eastern boundary of the clearcut area and continuing 2.4 km northeastward past 10- Acre Pond; Fig. 1) served as the third habitat type and is “the narrow ROW.” The narrow ROW has been maintained by selective herbicide- application at 8-year intervals. Vegetation in the narrow ROW is similar to that of aspen forest in the wide ROW and the clearcut area. Warbler Density. — We used three methods to sample male Golden-winged Warblers dur- ing peak territorial and nesting behavior (mid- May through mid-Jun in Pennsylvania; McWilliams and Brauning 2000) in 2002 and 2003. We used each method throughout the three habitat types studied and throughout the sampling season. Particular surveys for each method were conducted independently. The first sampling method combined point- count surveys (variable circular plot; Reyn- olds et al. 1980) with playback recordings us- ing a technique modified from Yahner and Ross (1995). A point count comprised (1) a 3.0-min pre-playback period, (2) a 1.3-min playback stimulus period, and (3) a 3.0-min post-playback period (total = 7.3 min), whereupon all male warblers observed by sight and sound during the count were noted. The playback period consisted of seven Type I (Gill and Lanyon 1964, Ficken and Ficken 1967) Golden-winged Warbler songs broad- cast at 10-sec intervals with the use of a cus- tom-made compact disc provided by the Ma- caulay Library of Natural Sounds (Ithaca, NY, USA; MLNS Job #2002073) and a small com- pact disc player with portable speakers. Each song consisted of one zee note and three bee notes because this song structure produces a relatively strong response from Golden- winged Warblers (Ficken and Ficken 1973). Songs were broadcast at a volume comparable to that of male warblers on the study area. We conducted point counts in the clearcut area at stations {n ~ 64 in 2002 and 44 in 2003) distributed among four transects with each transect spaced 400 m apart. Stations in the wide ROW {n ~ 15 in 2002 and 18 in 2003) and the narrow ROW (n = 12 in 2002 and 0 in 2003) were distributed along a single transect positioned along the central axis of each ROW. Point-count stations were spaced 200 m apart along a given transect. Lengths of transects in 2002 and 2003, respectively, were 3.2 and 2.0 km in the clearcut area, 2.9 and 3.5 km in the wide ROW, and 2.2 and 0.0 km in the narrow ROW. We defined a point- count survey as the sampling of all stations at SGL 176 once within 7 days; JEK conducted four and three surveys in 2002 and 2003, re- spectively. Surveys were conducted during 20 May- 17 June between dawn and 1000 hrs EST when wind velocity was <16 km/hr and precipitation was absent or limited to light, intermittent rain. The second and third sampling methods were spot-mapping methods and supplement- ed the point-count surveys. The first spot- mapping method, termed the “opportunistic” method, was similar to the “Combined Ver- sion” mapping method (Paul and Roth 1983) and was used daily from 4 May to 26 June in 2002 and 2003. This method consisted of five or more visits (0. 2-3.0 hrs each) over the course of the breeding season to specific areas where Golden-winged Warblers were known to occur. We recorded movements and behav- iors of individual warblers, giving particular attention to interactions among males that might help identify boundaries of territories (Robbins 1970). The second spot-mapping method, termed the “census” method, was modified from that recommended by the International Bird Cen- sus Committee (Robbins 1970) and was used eight and seven times (~twice/week, 22 May- 18 Jun) in the clearcut area and the wide ROW, respectively, in 2003. Each census en- tailed walking slowly (<3.2 km/hr) along pre- determined routes and recording location, time, gender, and behavior of all Golden- Kuhel and Yahner • HABITAT QUALITY FOR GOLDEN-WINGED WARBLERS 805 winged Warblers seen or heard in an attempt to obtain a complete count of males within the point-count survey area. Censuses occurred between dawn and 1000 hrs EST during days with little or no precipitation and winds <16 km/hr. Observer, starting point, and direction of travel for each route alternated throughout the season, and we took special precautions to avoid double-counting individual birds locat- ed between routes (Kubel 2005). We were usually able to differentiate indi- viduals of neighboring territories during a giv- en sampling method by observing singing males simultaneously. We also gave particular attention to plumage (e.g., hybrid vs. “pure”) and song (e.g., pitch, length, and number of notes) of each individual observed. For ex- ample, several individuals having distinctly “raspy” songs relative to most warblers on the study area were encountered consistently in particular clearcuts and portions of the wide ROW. Several other individuals having unusu- ally “fast” songs (i.e., especially short notes without compensatory increases in length of interval between notes) were encountered consistently in particular areas. One territory consisted of a male that consistently sang an unusually high number (6-7) of bee notes, and the bee notes of males at several other terri- tories seemed to be higher-pitched than the in- troductory zee note {bee notes were lower- pitched than zee notes in all other males). We combined results from point-count and spot-mapping surveys to estimate total number of territories in each habitat type. In designating a territory, the International Bird Census Com- mittee (Robbins 1970) recommended that of 8- 10 visits to a survey area, at least three obser- vations of an individual in a particular region of the area should be registered to count the region as a territory. We made more than 10 visits (combining point counts and spot-mapping) to survey areas of each habitat type and arbitrarily required at least five observations of a male Golden-winged Warbler in a pailicular region for it to be designated a territory. We also re- quired the live observations to occur during a period of ^2 weeks (to account for possible mi- grants) because visits often occurred at short in- tervals ( 1-5 days). We did not detect Golden-winged Warblers throughout all sectors of each habitat type. We were unable to ascertain whether apparent ab- sence from certain sectors was due to those sec- tors being unsuitable (not available) versus not preferred (available but not selected; Johnson 1980). We were unable to identify what consti- tuted available habitat and based estimates of density on the used sector of each habitat type. We defined “used sector” as the area inclusive of all known warbler territories and exclusive of areas where warblers were not observed. We calculated density for each habitat type as num- ber of territories/ha of early successional habitat present in the used sector. Nesting Success and Productivity. — We conducted nest searching in the clearcut area and the wide ROW during 2002 and 2003 by observing warbler behaviors indicative of mated status or locations of nests (Baird 1967; Ficken and Ficken 1967, 1968a, 1968b; Con- fer 1992) and by physically searching poten- tial nest substrates (e.g., patches of herbaceous vegetation) with a three-pronged stick (J. F. Confer, pers. comm.). We coordinated nest searching with opportunistic spot-mapping and usually spent 0. 5-2.0 hrs/visit searching a territory for nests. Locations of nests discovered by observers other than JFK were marked temporarily (<24 hrs) by attaching plastic flagging to a shrub >10 m from the nest; no other nests were marked. JFK monitored each nest through completion (i.e., fledging or failure) by visit- ing the nest at 3-day intervals (Martin and Geupel 1993) and observing its contents; spe- cial care was taken to avoid leaving visual and olfactory clues to nest location. Nests were checked daily when nestlings were near fledg- ing (within 1-3 days, based on expected Hedg- ing date of 8—1 1 days post-hatch [Will 1986. Confer et al. 2003]) to maximize chances of confirming successful fledging by observation of fledged individuals. We estimated nesting success for each hab- itat type using apparent success (Shaffer 2004) and nest survival rate (Mayfield 1961. 1975). We calculated apparent success by di- viding number of successful nests (nests that fledged >1 young) by total number of nests for which nest fate was known. We included nests where we failed to observe eggs or nest- lings in unsucccssrul nests (those not Hedging young) in the total if we had obseiwed them during the building stage aiul nest construc- tion had been completed (Klaus and Buehler 806 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 TABLE 1 . Amount (ha) of early successional habitat, number of observed territories, and density of Golden- winged Warblers in used sectors of the clearcut area and the wide ROW at State Game Lands 176, Centre County, Pennsylvania, 2002 and 2003. Goodness-of-fit tests were used to compare observed vs. expected number of territories between used sectors of the clearcut area and the wide ROW each year. Amount of habitat Observed territories (n) Density^* Year Clearcut ROW Clearcut ROW Clearcut ROW P 2002 6L9 1T6 29 9 OT7 OJl L25 0.264 2003 31 10 0.50 0.79 1.64 0.200 ^ Density = territories/ha of early successional habitat. 2001) or there was physical evidence the nest had been depredated (e.g., nest was over- turned or destroyed). We calculated nest sur- vival rate by multiplying survival rates of egg- laying, incubation, and nestling stages of the nesting cycle (Mayfield 1961, 1975). We as- signed 3, 1 1, and 8 exposure days to the egg- laying, incubation, and nestling stages, re- spectively (Will 1986, Confer 1992, Confer et al. 2003, Kubel 2005), based on a half-day for the first and last days of each stage and a full day for each day in between. We used data only from nests in which eggs and/or nestlings were observed in calculating survival rates. We estimated productivity for each habitat type by calculating mean number of fledglings produced per nesting territory, which we de- fined as a territory within which at least one nesting attempt occurred. We assumed when a nest was successful that number of young fledged was equal to the number of nestlings last observed in the nest (Confer et al. 2003). Statistical Analyses. — Statistical compari- sons of density, nesting success, and produc- tivity among habitat types were restricted to the clearcut area and the wide ROW because we did not detect Golden-winged Warblers in the narrow ROW. Statistical tests were con- sidered significant at P ^ ().()5. We compared density between used sectors of the clearcut area and the wide ROW by com- paring observed versus expected numbers of ter- ritories with Chi-square goodness-of-fit tests (Sokal and Rohlf 1995). We calculated 95% confidence intervals (Cl) for apparent success, nest survival rate, and productivity in the clear- cut area and the wide ROW. We used Pearson Chi-square tests-of-independence (Minitab Inc. 1998), program CONTRAST (Hines and Sauer 1989), and Mann-Whitney f/-tests (Sokal and Rohlf 1995, Systat Software Inc. 2002) to com- pare apparent success, nest survival rate, and productivity, respectively, between the clearcut area and the wide ROW. RESULTS Density. — We did not detect Golden-winged Warblers in the narrow ROW and we assumed density in the habitat type was 0.00 territories/ ha. We estimated that 29 and 3 1 territories oc- curred in the clearcut area, and 9 and 10 ter- ritories occuned in the wide ROW in 2002 and 2003, respectively; almost all territories occurred in aspen forest. Density was 0.47 and 0.50 territories/ha in the clearcut area and 0.7 1 and 0.79 in the wide ROW in 2002 and 2003, respectively, based on early successional cov- er in used sectors of each habitat type (Table 1). Observed versus expected number of ter- ritories did not differ between used sectors of the clearcut area and the wide ROW in either year (x~ — 1.64, df = 1, F ^ 0.20; Table 1). Nesting Success. — We analyzed data for 9 and 15 nests monitored in the clearcut area and 6 and 7 nests in the wide ROW in 2002 and 2003, respectively. Apparent success was 33% in 2002 and 73% in 2003 in the clearcut area, and 17% in 2002 and 14% in 2003 in the wide ROW (Fig. 3). Success during both years combined was not independent (y^ = 6.34, df = 1, F = 0.012) of habitat type (58% [95% Cl = 37-78%] in the clearcut area and 15% [95% Cl = 2-45%] in the wide ROW). We excluded five nests from calculations of nest survival rate because no exposure days were observed (4 nests [wide ROW, 2002] ob- served during the building stage appeared to have been depredated before eggs were ob- served, and the other [clearcut area, 2003] ap- peared to have been abandoned [with eggs] prior to discovery). Survival rates in 2002 and 2003, respectively, were 4% {n = 9 nests) and Kubel and Yahner • HABITAT QUALITY FOR GOLDEN-WINGED WARBLERS 807 0.8 0.7 0.6 0.5 (/) 0) o 9. 0.4 CO 0.3 0.2 0.1 1 1 2002 ■ Clearcut - apparent success HClearcut - survival rate □ ROW - apparent success S ROW - survival rate 2003 Year Overall FIG. 3. Estimates of nesting success in the clearcut area and the wide ROW at State Game Lands 176, Centre County, Pennsylvania during 2002 and 2003. Estimates included apparent success (proportion of nests fledging young) and nest survival rate (Mayfield 1961, 1975). Apparent success was a more reliable estimate in 2002 because nest survival rate was overestimated in the wide ROW (4 depredated nests excluded from analysis) and underestimated in the clearcut area (3 nests survived to fledging, and 4% survival would imply 75 nesting attempts among only 8 breeding pairs). Apparent success overall differed {P = 0.012) between habitat types. 73% {n = 14) in the clearcut area, and 48% {n = 2) and 15% (n = 1) in the wide ROW (Table 2; Fig. 3). Survival rates differed (x^ = 8.75, df - 1, P = 0.003) between habitat types in 2003 but did not differ (x^ = 2.08, df == 1, P = 0.15) over both years combined (46% [95% Cl == 28-76%] in the clearcut area and 20% [95% Cl = 6-67%] in the wide ROW) (Table 2). Survival rate did not differ between years in the wide ROW (x^ = 0.74, df = \, P = 0.39) but was greater in 2003 than in 2002 in the clearcut area (x^ = 18.5, df = 1, P < 0.001). Nest predation was the primary cause of nest failure in both habitat types in 2002, ac- counting for hve of six (83%) failures in the clearcut area and all five (100%) failures in the wide ROW. Nest predation in 2003 caused only one of four (25%) failures in the clearcut area compared to six of six ( 100%) failures in the wide ROW. Productivity. — We estimated there were 28 nesting territories in the clearcut area and the wide ROW combined over both years of the study. Productivity in 2002 and 2003, respec- tively, was 1.13 (/7 = 8 nesting territories) and 3.15 {n = 13) fledglings/nesting territory in the clearcut area and 0.67 (/? = 3) and 0.50 (/? = 4) in the wide ROW. Productivity during both years combined was different {V = 114, P - 0.026) between habitat types (2.38 [95% Cl = 1.53-3.24] in the clearcut area and 0.57 1 95% Cl = 0.00-1.47] in the wide ROW). Productivity was similar between years in the wide ROW {U = 6.5, P = 0.83) but greater in 2003 than in 2002 in the clearcut area {U = 82.5, P = 0.018). DISCUSSION l.ack of replication and short duration of the study limit the inteipretation and application of our results. However, our intensive survey ef- 808 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 forts likely produced accurate estimates of hab- itat use and reproductive success at SGL 176 during the study, as we thoroughly monitored the entire breeding population within each hab- itat type and found a high percentage of nests (likely >85% during 2003, based on ability to identify mated status of males by singing be- havior; Ficken and Ficken 1967). Close inspec- tion of our results yields interesting hypotheses that have range-wide implications and warrant further investigation, especially considering the importance of appropriate habitat management for Golden-winged Warblers and other early- successional, ground-nesting birds in eastern North America. Clearciit Area and Wide Densities of singing males observed within the used sectors of the clearcut area and the wide ROW in 2002 and 2003 suggest the two habitat types are suitable when appropriate nesting substrates and other resources are present (e.g.. in aspen forest; Kubel 2005). However, reproductive success was vastly different be- tween types during our study because rates of nest predation were much lower in the clear- cut area than in the wide ROW. This result is surprising because the tw’o habitat types were <2 km apart and in the same patch of forest (SGL 176 is surrounded by agriculture and de- velopment); we would not expect predator communities to differ between those sectors of the study area. We can only speculate about the identity of nest predators and why nest predation differed between the clearcut area and the wide ROW, but there is fairly strong evidence to suggest that nest predators differed between habitat types. Almost all depredated nests in the clearcut area were undamaged by the predator. Observations of snakes and of avian predators at ground level were rare, but eastern chip- munks (Tamias striatiis) were observed fre- quently near (1-10 m) nests. Chipmunks were unusually abundant in 2002 both at SGL 176 (Yahner 2003a; JEK, pers. obs.) and region- ally (J. L. Confer, pers. comm.), but appeared to be much less abundant in 2003. Major in- ter-annual population fluctuations are common in small mammals, which might explain why nest predation in the clearcut area was consid- erably lower in 2003 (7%) relative to 2002 (56%). Nest predation in the wide ROW was high Kiibel and Yahner • HABITAT QUALITY FOR GOLDEN-WINGED WARBLERS 809 in both 2002 (83%) and 2003 (86%) and did not follow the pattern observed in the clearcut area. Physical evidence at almost half of the depredated nests in the wide ROW suggested predators other than small mammals and snakes, as five nests were severely damaged (e.g., pulled apart, flipped over, and/or pressed into the ground) and vegetation surrounding an additional nest was trampled. Larger avian predators (e.g., American Crow [Cor\'iis bra- chyrhynchos^ and Blue Jay [Cyanocitta cris- tata]) were rarely seen at the wide ROW. Therefore, mid- to large-sized mammals likely had an important role in nest predation at this habitat type. Nest predation may have been high in the wide ROW because long-term management of this habitat by selective herbicide treatment has resulted in a static vegetation structure with abrupt edges between patches of shrubs and herbs. This simple structure is easier to traverse than one with gradual edges where plant suc- cession occurs at high levels of spatial hetero- geneity and poses a physical barrier to move- ment (Bowman and Harris 1980, Yahner and Wright 1985, Suarez et al. 1997). We observed animal trails in the wide ROW, and the trails tended to avoid patches of shrubs and travel through open patches of herbs. Hence, mid- to large-sized mammals at the wide ROW were often within close proximity of nests of Golden- winged Warblers, as nests were usually in patch- es of herbaceous vegetation (Kubel 2005). An- imal trails were not encountered in harvested stands of the clearcut area where these relatively new, ephemeral shrublands tended to have greater interspersion of shrubs and herbs. This complex vegetation structure is unlikely to im- pede movement of small mammals, which we suspect were the primary nest predators in the clearcut area. Proximity to agriculture may be an addi- tional explanation for differences in nest pre- dation between the clearcut area and the wide ROW during our study. Predation pressure may be greater in agricultural landscapes than in forested landscapes (Andren and Angelstam 1988, Andren 1992, Rudnicky and Hunter 1993), or greater at agricultural edges than at silvicultural edges (Bayne and Hobson 1997. Rodewald and Yahner 2001). Abundance of meso-carnivores in particular can be relatively high in agricultural areas (Pedlar et al. 1997, Dijak and Thompson 2000). Stands used by Golden-winged Warblers in the clearcut area were 1.0-2. 6 km from agriculture, whereas used portions of the wide ROW were only 0.0-0. 6 km distant. The high rate of nest success in the clearcut area in 2003 is similar to the >83% success observed for ground nests (<0.5 m) of mul- tiple avian species in young (<10 yrs), even- aged aspen stands at the same site during 1985-1987 (Yahner 1991). King et al. (2001) observed a daily survival rate of 0.990 for 290 nests of multiple species in clearcuts in New Hampshire, which is comparable to the daily survival rate of nests of Golden-winged War- blers in the clearcut area in 2003. Thus, clear- cutting at SGL 176 and elsewhere in the northeastern United States may create high- quality habitat for Golden-winged Warblers and other ground-nesting birds. Some studies of avian ecology in utility ROWs have reported higher nest success for shrubland birds than we observed for Golden- winged Warblers in the wide ROW at SGL 176. Reported estimates of nest success in ROWs throughout the northeastern United States include 83% for Chestnut-sided War- blers (Dendroica pensylvanica) in Massachu- setts (King and Byers 2002), 56 and 61% for 20 species in Maryland (Chasko and Gates 1982), 55% for multiple species in Maine, Massachusetts, and New York (Confer and Pascoe 2003), and 39 and 65% for 10 species in Pennsylvania (Yahner et al. 2004). These estimates of nest success include data from numerous shrub-nesting species and do not necessarily reflect success of ground-nesting species like the Golden-winged Warbler. We are not aware of any published studies of nest success specific to Golden-winged Warblers in ROW habitat, but estimates from research through 2005 included 36% in West Virginia (R. A. Canterbury, unpubl. data), 41% in New York (J. L. Confer, unpubl. data), and 44% in New Jersey (Sharon DeFalco, un- publ. data) with almost all failures the result of nest predation. As in our study, these esti- mates are lower than those reported for Gold- en-winged Warblers in other habitats, which include 55% in mowed lield-hnest edges and rocky outcrops of the Canadian Shield (Ra- chel Vallender and R. J. Robert.son, unpubl. data), 71% in upland mine fields and edges 810 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 (R. A. Canterbury, unpubl. data), 73% along access roads and logging decks in young (<20 yrs) forest stands (Klaus and Buehler 2001), and 85% in brushy fields adjacent to forest (Will 1986). An additional estimate from up- land mine fields (Bulluck and Buehler 2008) ranges from 50% (overall nest survival based on daily survival rate of 0.973) to 59% (ap- parent nest success). Confer et al. (2003) re- ported only 38% nest success in shrubby fields of abandoned croplands but, perhaps success there was influenced by the agricultural his- tory and setting of the study sites. Narrow ROW. — Structure and species com- position of vegetation in the narrow ROW were similar to that observed in aspen sectors of the wide ROW and the clearcut area (Kubel 2005). However, our surveys in 2002 failed to detect Golden-winged Warblers in the portion of the narrow ROW that served as our third habitat type. Preliminary surveys in 1998 and 1999 (RHY, unpubl. data) failed to detect Golden-winged Warblers in the narrow ROW, and we concluded prior to 2003 the species does not typically establish territories there. Warblers did establish territories and nests in portions of the same ROW where it extended into the clearcut area (surveyed incidentally as part of the clearcut area in 2002 and 2003), suggesting that additional openings contigu- ous with the ROW were required for breeding. We are not aware of any published studies that have documented territories of Golden- winged Warblers in narrow (<20-m) ROWs bordered on both sides by relatively mature (>40 yrs) forest. However, this warbler does occur along a portion of a 20-m-wide ROW in southern New York (J. L. Confer, pers. comm.) where there are additional openings in the forest canopy adjacent to both sides of the ROW (Kubel 2005). Rossell (2001) suggested Golden-winged Warblers choose song perches at positions in the forest canopy that enhance their ability to display vocally and visually to attract mates. Most territories in the wide ROW at SGL 176 were 100-175 m long; thus, a territory in the narrow ROW would have to be 300-525 m long to contain as much early successional cover. Assuming there is a minimum threshold for amount of early successional habitat in a territory, a non-linear habitat would optimize the ability of a male to be conspicuous in all parts of the territory by visual and vocal dis- plays (e.g., for mate attraction or territory de- fense). This may explain why Golden-winged Warblers at SGL 176 and elsewhere (e.g., southern New York) use portions of 20-m- wide ROWs only where they are contiguous with or adjacent to other forest openings. CONSERVATION IMPLICATIONS Small-scale, rotation clearcutting may be a favorable management technique for Golden- winged Warblers when suitable regeneration (i.e., small, interspersed patches of herbs and multi-stemmed shrubs or root-suckering trees [Kubel 2005]) can be assured. However, cau- tion should be used when assessing the poten- tial for ROWs to provide suitable habitat. Nar- row ROWs (<20-m wide) in the setting of mature (80-90 yrs) forest and in the absence of additional open areas should not be regard- ed as potential sources of habitat during con- servation planning. Wider (e.g., 60-m wide) ROWs appear to provide habitat, but our study raises questions about whether nests of Golden- winged Warblers (and perhaps other ground-nesting birds) may be especially vul- nerable to nest predation by mid- to large- sized mammals in certain types of ROWs. Further, ROWs do not appear to be preferred habitats in some areas (Bulluck and Buehler 2006). Our results are not necessarily representa- tive of long-term conditions at SGL 176, but the low reproductive success we observed in the wide ROW appears to be consistent with relatively low success observed for Golden- winged Warblers in ROWs elsewhere in the eastern United States. Further research is needed to learn if ROWs may act as habitat sinks for Golden-winged Warblers, especially where ROWs are maintained in a manner that creates abrupt edges between patches of shrubs and herbs. We recommend long-term monitoring of reproductive success coupled with experimental manipulations of vegetation structure. ACKNOWLEDGMENTS bunding was provided by the Teresa Heinz Scholars for Environmental Research and the Pennsylvania Ag- ricultural Experiment Station. We thank J. A. De- Cecco, M. E. McDermott, D. C. Rabbers, K. E. Vil- berg, and P. L. Howell for field assistance, M. R. Mar- Kuhel and Yahner • HABITAT QUALITY FOR GOLDEN-WINGED WARBLERS 811 shall and C. B. Goguen for help in study design and data analysis, J. L. Confer for advice on nest-searching techniques, and M. C. Brittingham, K. C. Steiner, and two anonymous reviewers for helpful comments on earlier versions of this manuscript. We appreciate as- sistance from Loren Zirkle of Allegheny Power, and Matt Belinda and Charles Olenik of FirstEnergy. LITERATURE CITED Andren, H. 1992. Corvid density and nest predation in relation to forest fragmentation: a landscape perspective. Ecology 73:794-804. Andren, H. and P. Angelstam. 1988. Elevated pre- dation rates as an edge effect in habitat islands: experimental evidence. Ecology 69:544-547. Askins, R. a. 2001. Sustaining biological diversity in early successional communities: the challenge of managing unpopular habitats. Wildlife Society Bulletin 29:407-412. Baird, J. 1967. Some courtship displays of the Gold- en-winged Warbler. Wilson Bulletin 79:301-306. Bayne, E. M. and K. A. Hobson. 1997. Comparing the effects of landscape fragmentation by forestry and agriculture on predation of artificial nests. Conservation Biology 11:1418-1429. BirdLife International. 2004. Vennivora chrysoptera. In 2007 lUCN Red List of Threatened Species. www.iucnredlist.org (accessed 24 February 2008). Bowman, G. B. and L. D. Harris. 1980. Effect of spatial heterogeneity on ground-nest depredation. Journal of Wildlife Management 44:806-813. Brawn, J. D. and S. K. Robinson. 1996. Source-sink population dynamics may complicate the interpre- tation of long-term census data. Ecology 77:3-12. Buehler, D. a., a. M. Roth, R. Vallender, T. C. Will, J. L. Confer, R. A. Canterbury, S. Bark- er SwARTHOUT, K. V. Rosenberg, and L. P. Bul- LUCK. 2007. Status and conservation priorities of Golden-winged Warbler (Vermivora chrysoptera) in North America. Auk 124:1439-1445. Bulluck, L. P. and D. a. Buehler. 2006. Avian use of early successional habitats: are regenerating forests, utility right-of-ways and reclaimed surface mines the same? Forest Ecology and Management 236:76-84. Bulluck, L. P. and D. A. Buehler. 2008. Factors in- fluencing Golden-winged Warbler [Vermivora chrysoptera) nest-site selection and nest survival in the Cumberland Mountains of Tennessee. Auk 125: In Press. Chasko, G. G. and j. E. Gates. 1982. Avian habitat suitability along a transmission-line corridor in an oak-hickory forest region. Wildlife Monographs 82. Confer, .1. L. 1992. Golden-winged Warbler [Venni- vora chrysoptera). The birds of North America. Number 20. Confer, .1. L. and S. M. Pascof;. 2003. Avian com- munities on utility rights-of-ways and other man- aged shrublands in the northeastern United States. Forest Ecology and Management 185:193-205. Conif;r, j. L., j. L. Larkin, and P. E. Ai.lf;n. 2003. Effects of vegetation, interspecific competition, and brood parasitism on Golden-winged Warbler {Vennivora chrysoptera) nesting success. Auk 120:138-144. Conklin, W. G. 1943. Portion of famous “Barrens” in Centre County acquired as state game lands. Pennsylvania Game News 14:6-7, 24-25. DeGraaf, R. M. and M. Yamasaki. 2003. Options for managing early-successional forest and shrubland bird habitats in the northeastern United States. Eorest Ecology and Management 185:179-191. Dijak, W. D. and F. R. Thompson III. 2000. Landscape and edge effects on the distribution of mammalian predators in Missouri. Journal of Wildlife Man- agement 64:209-216. Ficken, M. S. and R. W. Ficken. 1967. Singing be- haviour of Blue-winged and Golden-winged war- blers and their hybrids. Behaviour 28:149-181. Ficken, M. S. and R. W. Ficken. 1968a. Courtship of Blue-winged Warblers, Golden-winged Warblers, and their hybrids. Wilson Bulletin 80:161-172. Ficken, M. S. and R. W. Ficken. 1968b. Territorial relationships of Blue-winged Warblers, Golden- winged Warblers, and their hybrids. Wilson Bul- letin 80:442-451. Ficken, M. S. and R. W. Ficken. 1973. Effect of num- ber, kind and order of song elements on playback responses of the Golden-winged Warbler. Behav- iour 46:114-128. Gill, F. B. and W. E. Lanyon. 1964. Experiments on species discrimination in Blue-winged Warblers. Auk 81:53-64. Helms, J. A. 1998. The dictionary of forestry. Society of American Foresters, Washington, D.C., USA. Hines, J. E. and J. R. Sauer. 1989. Program CON- TRAST: a general program for the analysis of sev- eral survival or recovery rate estimates. Fish and Wildlife Technical Report 24. USDI, Fish and Wildlife Service, Washington, D.C., USA. Hobbs, N. T. and T. A. Hanley. 1990. Habitat evaluation: do use/availability data reflect cairying capacity? Journal of Wildlife Management 54:515-522. Johnson, D. H. 1980. The comparison of usage and availability measurements for evaluating resource preference. Ecology 6 1 :65-7 1 . Jones, J., P. J. Doran, L. R. Nagy, and R. T. Holmes. 2005. Relationship between Mayfield nest-surviv- al estimates and .seasonal fecundity: a cautionary note. Auk 122:306-312. King, D. I. and B. E. Byers. 2002. An evaluation of powerline rights-of-way as habitat for early-suc- cessional shrubland birds. Wildlife Society Bul- letin 30:868-874. King, D. I., R. M. Di:Graaf, and C. R. Griffin. 2(X)1. Productivity of early successional shmbland birds in clearcuts anti groupcuts in an eastern tleciduous for- est. Journal t)f WikIlife Management 65:345-350. Klaus, N. A. and D. A. Bi f:hi i r. 2001. Golden- wingetl Warbler breeding habitat characteristics and nest success in clearcuts in the southern Ap- 812 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 palachian Mountains. Wilson Bulletin 113:297- 301. Kubel, J. E. 2005. Breeding ecology of Golden- winged Warblers in managed habitats ot central Pennsylvania. Thesis. Pennsylvania State Univer- sity, University Park, USA. Lorimer, C. G. 2001. Historical and ecological roles of disturbance in eastern North American forests: 9,000 years of change. Wildlife Society Bulletin 29:425-439. Lorimer, C. G. and A. S. White. 2003. Scale and frequency of natural disturbances in the north- eastern US: implications for early successional forest habitats and regional age distributions. For- est Ecology and Management 185:41-64. Martin, T. E. and G. R. Geupel. 1993. Nest-monitor- ing plots: methods for locating nests and moni- toring success. Journal of Field Ornithology 64: 507-519. Mayfield, H. F. 1961. Nesting success calculated from exposure. Wilson Bulletin 73:255-261. Mayfield, H. F. 1975. Suggestions for calculating nest success. Wilson Bulletin 87:456-466. McWilliams, G. M. and D. W. Brauning. 2000. The birds of Pennsylvania. Cornell University Press, Ithaca, NY, USA. Minitab Inc. 1998. MINITAB for Windows. Version 12.2. Minitab Inc., State College, Pennsylvania. USA. Palmer, W. L. 2003. The Barrens Grouse Habitat Man- agement Study, State Game Lands Number 176. Pennsylvania Game Commission Project Annual Job Report. Job Code 29001. www.pgc.state.pa.us/pgc/ lib/pgc/reports/2003_wildlife/2900 l-02z.pdf (ac- cessed 6 August 2006). Paul, J. T. and R. R. Roth. 1983. Accuracy of a ver- sion of the spot-mapping census method. Journal of Field Ornithology 54:42-49. Pedl.ar, j. H., L. Fahrig, and H. G. Merriam. 1997. Raccoon habitat use at 2 spatial scales. Journal of Wildlife Management 61:102-112. Reynolds, R. T, J. M. Scott, and R. A. Nussbaum. 1980. A variable circular-plot method for estimat- ing bird numbers. Condor 82:309-313. Robbins, C. S. 1970. An international standard for a mapping method in bird census work. Internation- al Bird Census Committee. Audubon Field Notes 24:722-726. Rodewald, a. D. and R. H. Yahner. 2001. Influence of landscape composition on avian community structure and associated mechanisms. Ecology 82: 3493-3504. Rossell Jr., C. R. 2001. Song perch characteristics of Golden-winged Warblers in a mountain wetland. Wilson Bulletin 113:246-248. Rudnicky, T. C. and M. L. Hunter. 1993. Avian nest predation in clearcuts, forests, and edges in a for- est-dominated landscape. Journal of Wildlife Management 57:358-364. Sauer, J. R., J. F. Hines, and J. Fallon. 2007. The North American Breeding Bird Survey, results and analysis 1966-2(X)6. Version 10.13.2(X)7. USDI, Geological Survey, Patuxent Wildlife Research Center. Laurel, Maryland. USA. www.mbr-pwrc.usgs.gov/bbs/bbs. html (accessed 24 February 2008). Shaffer, T. L. 2004. A unified approach to analyzing nest success. Auk 121:526-540. SOKAL, R. R. AND F. J. Rohlf. 1995. Biometry. Third Edition. W. H. Freeman, New York, USA. Suarez, A. V., K. S. Pfennig, and S. K. Robinson. 1997. Nesting success of a disturbance-dependent songbird on different kinds of edges. Conserva- tion Biology 11:928-935. Systat Software Inc. 2002. SYSTAT grad pack for Windows. Version 10.0. Systat Software Inc., Richmond, California, USA. Trani, M. K., R. T. Brooks, T. L. Schmidt, V. A. Rudis, ANT) C. M. Gabbard. 2001. Patterns and trends of early successional forests in the eastern United States. Wildlife Society Bulletin 29:413-424. U. S. Department of Interior (USDI). 2002. Birds of conservation concern 2002. USDI, Fish and Wild- life Service, Division of Migratory Bird Manage- ment. Arlington, Virginia, USA. Van Horne, B. 1983. Density as a misleading indi- cator of habitat quality. Journal of Wildlife Man- agement 47:893-901. Will, T. C. 1986. The behavioral ecology of species replacement: Blue-winged and Golden-winged warblers in Michigan. Dissertation. University of Michigan, Ann Arbor, USA. Yahner, R. H. 1986. Structure, seasonal dynamics, and habitat relationships of avian communities in small even-aged forest stands. Wilson Bulletin 98: 61-82. Yahner, R. H. 1991. Avian nesting ecology in small even-aged aspen stands. Journal of Wildlife Man- agement 55:155-159. Yahner, R. H. 2000. Long-term effects of even-aged management on bird communities in central Penn- sylvania. Wildlife Society Bulletin 28:1 102-1 1 10. Yahner, R. H. 2003a. Population trends in eastern chipmunks: implications to avian nesting success. Journal of the Pennsylvania Academy of Science 76:62-65. Yahner, R. H. 2003b. Responses of bird communities to early successional habitat in a managed land- scape. Wilson Bulletin 115:292—298. Yahner, R. H. and B. D. Ross. 1995. Seasonal re- sponse of Wood Thrushes to taped-playback songs. Wilson Bulletin 107:738—741. Yahner, R. H. and A. L. Wright. 1985. Depredation on artificial ground nests: effects of edge and plot age. Journal of Wildlife Management 49:508-513. Yahner, R. H., B. D. Ross, R. T. Yahner, R. J. Hut- NIK, AND S. A. Liscinsky. 2004. Long-term effects of rights-of-way maintenance via the wire-border zone method on bird nesting ecology. Journal of Arboriculture 30:288—294. The Wilson Journal of Ornithology 120(4):813— 819, 2008 AVIAN CELL-MEDIATED IMMUNE RESPONSE TO DROUGHT J. M. FAIR' 2 AND S. J. WHITAKER' ABSTRACT. — Regional droughts have far-reaching, substantial, and easily recognizable impacts on popula- tions and the environment. One component of these impacts that is not widely recognized is impairment of immune function by drought-related physiological stress. We studied cell-mediated immune function of cavity- nestling Western Bluebirds {Sialia mexicana), Ash-throated Flycatchers [Myiarchus cinerascens), and Violet- green Swallows {Tachycineta thalassino) at Los Alamos, New Mexico. There was a dramatic decrease in the cell-mediated immune responsiveness of developing nestlings associated with unusually dry conditions. Adult Western Bluebirds captured in 2002 weighed 1% less than in all previous years and average clutch size for all three species was reduced by 21% in 2002. Nestling body mass was also reduced for flycatcher and bluebird nestlings in 2002 compared to all other years. Survival to fledging age was lower overall during the drought years of 2000-2002 compared to the first 3 years of the study. Received 17 April 2006. Accepted 14 Fehruaiy 2008. Regional droughts have far-reaching, sub- stantial, and easily recognizable impacts on populations (George et al. 1992) and the en- vironment (Allen and Breshears 1998). One component of these impacts that is not widely recognized is impairment of immune function by drought-related physiological stress. There is ample evidence that stress affects immune system function (Apanius 1998), but a reduc- tion in immunity has not been documented di- rectly with drought conditions. One conse- quence, global warming, as frequency and se- verity of drought increases, may be increased vulnerability of populations to disease accom- panied by more frequent disease epidemics (Epstein 2001; Harvell et al. 2002; Mpller and Erritz0e 2003; Shaman et al. 2003, 2005). Ris- ing temperatures could expand the distribution of vector-bome pathogens transmitted by ar- thropods, including mosquitoes, sand flies, midges, and ticks (Shope 1992, Epstein and Defilippo 2001, Shaman et al. 2005); many arthropod-borne disease epidemics are asso- ciated with droughts (Epstein and Defilippo 2001). The drought of 20()0-2()()2 in the south- western United States, although not unprece- dented (Allen and Breshears 1998), was one of the most severe in 50 years. We assessed cell-mediated immune response from 1997 to 2002 using a phytohaemagglutinin (PH A) in- jection in nestlings of three species of cavity- ' Los Alamos National Laboratory, Atmospheric, Climate, and Environmental Dynamics, M.S .1495, Los Alamos, NM 87545. USA. -Corresponding author; e-mail; jmfair^^lanl.gov nesting birds: Ash-throated Flycatcher {Myiarchus ciuerascens). Western Bluebird {Sialia mexicana), and Violet-green Swallow {Tachycineta thalassina). PH A response has been associated with body condition and sur- vival (Horak et al. 1999, Alonso- Alvarez and Telia 2001). PHA injected for localized in vivo inflammatory response in birds has long been used to measure cell-mediated immunity (Sta- decker et al. 1977, Lamont and Smyth 1984) and has been found to not impose additional stress or survival costs (Merino et al. 1999, Smits and Williams 1999). PHA has been shown to increase energy expenditure and food intake (Martin et al. 2003, Barbosa and Moreno 2004). Martin et al. (2006) found that PHA-induced swelling is related to heightened immune cell activity in House Sparrows {Passer cloniesticiis) and that PHA-induced swelling is a tradeoff with other physiological functions. Cell-mediated immunity is particularly im- portant during the breeding season, because the likelihood of injury during sexual com- petition is high and cell-mediated immunity is essential for healing wounds and resisting in- fection (Zuk and Johnsen 1998). Our objec- tives were to: (1) compare three similar spe- cies in relation to nestling cell-mediated im- mune response, condition, and survival; and (2) examine the effects of severe drought on both nestlings and adults during the breeding season. METHOD.S Study Area. — 1'his study was cmiducted at I.os Alamos National Laboratory (l.ANL) in 813 814 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 north-central New Mexico. The 111-km^ fa- cility is on the Pajarito Plateau and consists of a series of relatively narrow mesas sepa- rated by deep, steep-sided canyons that extend west to east-southeast along an elevation gra- dient from the Jemez Mountains to the Rio Grande. Vegetation community types range from open ponderosa pine (Pinus ponderosa) forest to pihon-juniper {P. ediilis, Jiiniperus osteospenna) woodland. We placed 433 blue- bird boxes on LANL property during winter 1997. Nest boxes were placed ~2 m above ground level on trees and spaced —50-75 m apart. All nest boxes were visited every 2 weeks starting in May 1997; nests with eggs were considered active and visited every 2 days un- til the first eggs hatched (day = 0). Nests were visited again on days 3, 10, 15 and 16. Ten percent of the Western Bluebird pairs had sec- ond broods, which did not differ between years and were not included in the analysis. All nests included in the analysis were new pairs not previously in the study. Each nest- ling was handled for <8 min/visit in accor- dance with the Guidelines for the Use of Wild Birds in Research (Gaunt and Oring 1997). Measurements. — All nestlings in selected nests were injected intradermally on day 15 in the wing-web with either 0.05 ml (1.0 mg/ml) PHA in phosphate-buffered saline (PBS) (right side) or 0.05 ml PBS only (left side). This date was chosen, as it was latest date before fledging when this could be completed. The amount of swelling in the wing-web 24 hrs after inoculation was measured with a pocket dial gauge micrometer (L.S. Starrett® #1010, Athol, MA, USA) to the nearest 0.001 mm by the same person. The lower variance within individuals for PHA measurements us- ing the left wings of Western Bluebirds and Ash-throated Flycatchers in 2002 with wing webs not injected with PBS (post 24 hrs) was derived from a one-way ANOVA using the GEM (general linear model) in SAS (SAS In- stitute 1998). Our estimates of individual var- iance, consistent with Lessells and Boag (1987), used the correlation coefficient cal- culated using the variance component of the data set. The variability of measuring PHA with just one observer was estimated to be 0.94 and associated F, gj ratio of 1 127. A PHA index (Fair et al. 1999) to compare species was computed as the thickness of the PHA- inoculated wing-web minus the thickness of the opposite wing-web and standardized by the average wing thickness before inoculation. All PHA measurements were made by one person (JMF). Nestlings were weighed at 15 days and sur- vival of 864 nestlings was calculated up to the maximum of 14 days of age after hatching over 6 years. The length of the right ninth primary was measured with a ruler to the nearest mm using the flattened wing method (Svensson 1984). The left and right tarsi were measured with digital calipers to the nearest 0.1 mm. All birds were weighed using a dig- ital balance to the nearest 0.01 g. Predation was relatively low with most known dead birds being found in the nest (Fair et al. 2003). Probability of fledging for all three species after 14 days was 68% for all years combined. PHA-induced wing web swelling was mea- sured for 225 Western Bluebird (WEBF), 40 Ash-throated Flycatcher (ATFF), and 33 Vi- olet-green Swallow (VGSW) chicks. The ma- jority of the flycatchers and swallows were measured for PHA in 2001 and 2002. Three hundred flycatchers were additionally mea- sured for morphological parameters. Approx- imately 70-[jl1 of blood was collected on day 15 of age from the brachial vein of the wing in heparinized microcapillary tubes, which were kept in a cooler. Heparinized capillary tubes were spun for 10 min in a micro-capil- lary centrifuge within 1 hr of collection. He- matocrit estimates from the blood collected were obtained directly using a microhemato- crit reader. Data Analysis. — We used SAS statistical software (SAS Institute 1987) for all analyses. PHA measurements were heteroscedastic be- tween years {F = 5.50, P < 0.0003, Fevene’s Test), the data within nests were not indepen- dent, and we used the means of the PHA mea- surements for each nest in a nonparametric Kruskal-Wallis Test. Weight data were also not independent but had similar variances be- tween years and we used a mixed general lin- ear model (PROC MIXED). The nest box was included as a random factor to account for non-independence among young within a brood and fixed factors included nestling var- iables of body mass, species, and year. Com- parisons of bluebird and flycatcher nestling Fair and Whitaker • IMMUNE FUNCTION AND DROUGHT 815 survival rates were analyzed using the log- rank procedure LIFETEST (SAS Institute 1987), which allows for right censoring of data points where the failure times to the right are missing. The log-rank procedure LIFE- TEST computes nonparametric estimates of the survivor function and rank tests for asso- ciation of the response variable with other var- iables; it is more robust than the Wilcoxon statistic in detecting differences between groups. RESULTS Precipitation for this region was 25% below average during 2000 and 2001 and 65% below average through summer (Aug 2002). All three species had a similar reduction in the PHA-induced wing-web swelling response in 2001 and a more pronounced reduction in 2002 compared to the first 3 years of the study (WEBL, = 40.3 df = 4, P < 0.0001; ATEL, = 10.2, df = 3, P = 0.0002; VGSW, x" = 13.4, df = 1, P = 0.0003, Kruskal- Wallis) (Eig lA). Response to PHA increased with an- nual precipitation for all three species (WEBL [n = 225], u = 0.23, P = 0.003; ATEL [n = 40], u = 0.73, P <0.001; VGSW [n = 33, 2 yrs] u = 0.42, P = 0.01, Pearson Rank Cor- relation) (Pig. IB). Western Bluebird and Ash- throated Plycatcher did not differ in response to PHA for the first 3 years of the study. How- ever, Western Bluebird had a higher response in 2001 and 2002 to PHA than Ash-throated Plycatcher and Violet-green Swallow (x^ = 8.8, df = 2, P = 0.01; x" = 14.3, df = 2, P = 0.0008), respectively. PHA response was not related to clutch size in any of the three species for any year except for Western Blue- bird where birds with smaller clutch sizes (2 and 3) had lower responses to PHA than those with larger clutches (4 and 5) (x^ = 3.7, df = 1, P = 0.05). One hundred and thirty-four birds received an antigenic challenge when they were 5 days of age; we tested for the any correlations between PHA response and the three antigens and control, and found no re- lationship (P = 0.98, df = 4 and 133, P = 0.43). This was not a confounding factor in the analysis. Nestling body mass (g) was reduced for lly- catcher and bluebird nestlings which weighed less in 2002 compared to all other years (/' = 18.02, df = 3 and 36, P < 0.0001 P = 31.02, df = 4 and 220, P < 0.0001, respectively). The average body mass for Ash-throated Ply- catcher was significantly lower in 2002 than all other years (LSD test) and was signifi- cantly lower for Western Bluebird in both 2001 and 2002 than in other years. Violet- green Swallows did not have previous years for comparison of body mass. Similarly, the right wing length of the ninth primary was significantly shorter in 2001 and 2002 for Western Bluebird (P = 14.3, df = 4 and 220, P < 0.000 T), Ash-throated Plycatcher (P = 12.1, df = 4 and 87, P < 0.0001) and, in 2002, for Violet-green Swallow (P = 7.5, df = 1 and 31, P < 0.01). Length of the right tarsus did not differ between any years for Western Bluebird (P = 0.62, df = 4 and 220, P < 0.0001). Survival to fledging age was not lower in 2002 (x^ = 54.3, df = 5, P < 0.0001, log rank test) than in the previous 2 years, but was low- er overall during the drought years of 2000- 2002 compared to the first 3 years of the study (Pig. 2). Percent of eggs that hatched per clutch did vary between years for Western Bluebird with a reduced clutch size in 2002 from all other years (P = 12.86, df = 5 and 632, P < 0.0001). This was also similar for Ash-throated Plycatcher (P = 10.38, df = 5 and 223, P < 0.0001) and for Violet-green Swallow in 2002 (P = 5.47, df = 2 and 26, P = 0.01). Average clutch size for all three species was reduced by 21% in 2002. There was variation in hematocrit between years (P = 3.32, df = 5 and 128, P = 0.007) in adult Western Bluebirds that were captured, bled and banded (/? = 133); but there was no clear pattern for year differences. There was also variability in mass (g) for adult bluebirds be- tween years (2002, x ± SE - 23.9 ± 0.87 g; other years 25.8 ± 0.44 g) with adults cap- tured in 2002 weighing 7% less than in all previous years (P = 4.50, df = 5 and 112, P = 0.0009). DISCUSSION There was a dramatic decrease in the PHA- induced wing-web swelling in nestlings as- sociated with unusually dry conditions. Adult Western Bluebirds captured in 2002 weighed 7% less than in all previous years. Average body mass was also reduced for flycatcher and bluebird nestlings, which weighed less in 816 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 1997 1998 1999 2001 2002 Year 2.5 2 X ■g 1-5 C < I 1 Q. 0.5 0 0 10 20 30 40 50 60 70 Precipitation (cm) LIG. 1. (A). Cell-mediated response to PHA was significantly reduced in 2002 from all other years (PHA index [mean ± SE]). Morphological and immune data were not collected during 2000 due a large-scale fire adjacent to the study area in May 2000. (B). Cell-mediated response to PHA for annual average precipitation from 1997 to 2002. B ♦WEBL □ □ ATFL 1997 1999 1998 2002 2001^^ ^ ♦ □ ♦ □ □ 2002 compared to all other years, indicating nutritional stress was potentially related to drought conditions. Survival to fledging age in Western Bluebird nestlings was not lower in 2002 than in the previous 2 years, but was lower overall during the drought years of 2000-2002 compared to the first 3 years of the study (Table 1). Hematocrit did not pro- vide additional information on impacts of drought on condition due to variability be- tween individuals and years. There is no direct published evidence for effects of drought on cell-mediated response, but numerous studies document the cell-me- diated immune function is strongly correlated with nestling body mass (Saino et al. 1997, Christe et al. 1998, Telia et al. 2001, Westneat et al. 2004), availability and quality of food (Lochmiller et al. 1993, Birkhead et al. 1998, Gonzales et al. 1999, Hoi-Leitner et al. 2001, Lifjeld et al. 2002) and the overall environ- ment (Brinkhof et al. 1999, Telia et al. 2000). Lifjeld et al. (2002) found that cell-mediated immune response is affected by weather con- Fair and Whitaker • IMMUNE FUNCTION AND DROUGHT 817 0.8 1997 1998 1999 2000 2001 2002 Year FIG. 2. Probability of nestling survival to day 14, before fledging, for 1997-2002. ditions in particular, and that PHA response is influenced by short-term fluctuations in ener- gy balance. The PHA response in this study was measured over the entire summer and the measurements should include daily variation during the breeding season. We demonstrate wide variation in PHA between years that co- incides with a significant drought, but do not have food limitation data to identify the exact relationship. It is also not clearly understood how much variation in PHA response is her- itable, as there has been conflicting evidence (Brinkhof et al. 1999, Telia et al. 2000, Martin et al. 2006). Other environment conditions such as increased exposure to predators have also been shown to decrease the avian cell- mediated response (Navarro et al. 2004). The mostly likely mechanism for reduction in cell-mediated immune response, reduced weight, and nestling survival would be food availability and parental effort. There has been relatively little research documenting impacts of drought on arthropod communities, but one study experimentally showed that drought negatively impacted farmland arthropod com- munities (Frampton et al. 2000). Studies on the impacts of drought on arthropods or other prey in relation to parental effort and devel- opment would help in understanding the exact mechanism involved in impacts of severe drought on birds and immune function. One of our most significant findings was that clutch size for all three species was re- duced by 21% in 2002 — a result that follows TABLE 1. Average individual and nest characteristics for three cavity-nesting species for 1997-1999 (av- erage precipitation) and 2000-2002 (below average precipitation). Species and year Body mass, g Tarsus length, mm Survival to day 14, PHA Index (n) Clutch si/e Hatch date Percent hatch Number of individuals for PHA Number of nests Ash-throated Flycatcher 1997-1999 25.6 26.3 50.7 1.7 (10) 3.9 175.8 75.7 10 53 2()()()-20()2 25.2 25.5 23.0 0.51 (20) 3.4 170.8 67.1 30 50 Violet-green Swallow 1997-1999 2()()()-2()()2 17.5 13.6 44.8 0.51 (33) 3.3 183.3 70.2 33 30 Western Bluebird 1997-1999 26.5 23.0 74.5 1.2 (197) 4.5 165.4 70.6 197 170 2()()()-20()2 22.7 22.8 53.2 0.76 (28) 4.1 164.2 72.8 28 90 818 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 Telia et al. (2000) who found that cell-medi- ated immune response was higher in nestlings from larger clutches. However, this does not follow the studies of Horak et al. (1999) and Ilmonen et al. (2003) who found an increase in brood size reduced the cell-mediated im- mune response. We found no relationship be- tween timing of breeding during the season (hatch date) and response to PHA. Many of the tradeoffs between brood size, develop- ment, parental effort, and food availability re- main to be understood, but the impact of ex- treme environmental conditions can be mea- sured in immune function. Martin et al. (2006) noted that PHA-induced swelling does not ap- pear to be an unambiguous index of T-cell- mediated immunity, but rather a multifaceted index of cutaneous immune activity that may be affected by the physiological state of the individual as well as the life-history traits of the species. Our study documents a decrease in PHA-induced swelling over several years on the same day of nestling development within the same species that is associated with a reduction in precipitation and nestling mass and survival. ACKNOWLEDGMENTS We thank R C. Beeson, K. L Colestock, L. A. Haus- samen, Tina Sommer, and C. E. Talus for excellence in field assistance. We thank C. D. Allen, D. B. Bre- shears, M. H. Ebinger, H. A. Hinojosa, J. M. Heikoop, L. K. Marsh, P. M. Rich, and R. E. Ricklefs for com- ments on an earlier draft. We also thank B. K. San- dercock for helpful comments on the manuscript. We thank O. B. Myers for initiation of the nest box-mon- itoring network. The animal care and use committees of both LANL and the University of Missouri-St. Lou- is approved all protocols. This research was funded by the Environmental Restoration Program through Los Alamos National Security LLC, operator of the Los Alamos National Laboratory under Contract DE- AC52-06NA25396 with the U.S. Department of En- ergy. LITERATURE CITED Allen, C. and D. Breshears. 1998. Drought-induced shift of a forest- woodland ecotone: rapid land- scape response to climate variation. Proceedings of the National Academy of Sciences of the USA. 95:14839-14842. Alonso- Alvarez, C. and J. Tella. 2001. Effects of experimental food restriction and body mass changes on the avian T-cell-mediated immune re- sponse. Canadian Journal of Zoology 79:101 — 105. Apanius, V. 1998. Stress and immune response. Pages 133-153 in Advances in the study of behavior. Volume 27. Stress and behavior (A. P. Mpller, M. Milinski, and P. Slater, Editors). Academic Press, San Diego, California, USA. Barbosa, A. and E. Moreno. 2004. Cell-mediated im- mune response affects food intake but not body mass: an experiment with wintering Great Tits. Ecoscience 11:305—309. Birkhead, T, E Fletcher, and E. Pellat. 1998. Sex- ual selection in the Zebra Finch {Taeniopygia gut- tata): condition, sex traits and immune capacity. Behavioral Ecology and Sociobiology 44:179- 191. Brinkhof, M. W., P. Heeb, M. Kolliker, and H. Richner. 1999. Immunocompetence of nestling Great Tits in relation to rearing environment and parentage. Proceedings of the Royal Society of London, Series B 266:2315-2322. Christe, P, a. P. M0LLER, AND F. DE LoPE. 1998. Im- munocompetence and nestling survival in the House Martin: the tasty chick hypothesis. Oikos 83:175-179. Epstein, P. 2001. Climate change and emerging infec- tious diseases. Microbes and Infection 3:747-754. Epstein, P. and J. Defilippo. 2001. West Nile virus and drought. Global Change and Human Health 2: 105-107. Fair, J. M., E. Hansen, and R. E. Ricklefs. 1999. Growth, developmental stability and immune re- sponse in juvenile Japanese Quail. Proceedings of the Royal Society of London, Series B 266:1735- 1742. Fair, J. M., O. B. Myers, and R. E. Ricklefs. 2003. Immune and growth response of Western Blue- birds and Ash-throated Flycatchers to soil contam- inants. Ecological Applications 13:1817—1829. Frampton, G. K., P. j. van den Brink, and P. J. L. Gould. 2000. Effects of spring drought and irri- gation on farmland arthropods in southern Britain. Journal of Applied Ecology 37:865-883. Gaunt, A. and L. Oring. 1997. Guidelines to the use of wild birds in research. The Ornithological Council, Washington, D.C., USA. George, T, A. Fowler, R. Knight, and L. McEwen. 1992. Impacts of a severe drought on grassland birds in western North Dakota. Ecological Appli- cations 2:275-284. Gonzales, G., G. Sorci, A. P. Mpller, P. Ninni, C. Haussy, and F. de Lope. 1999. Immunocompe- tence and condition-dependent sexual advertise- ment in male House Sparrows {Passer dornesti- cus). Journal of Animal Ecology 68:1225-1234. Harvell, C. D., C. E. Mitchell, J. R. Ward, S. Al- TizER, A. P. Dobson, R. S. Ostfeld, and M. D. Samuel. 2002. Climate warming and disease risks for terrestrial and marine biota. Science 296: 2158-2162. Hoi-Leitner, M., M. Romero-Pujante, H. Hoi, and A. Pavlova. 2001. Food availability and immune capacity in Serin {Seriniis seriniis) nestlings. Be- havioral Ecology and Sociobiology 49:333-339. Fair and Whitaker • IMMUNE FUNCTION AND DROUGHT 819 Horak, R, L. Tegelmann, I. Ots, and A. P. M0ller. 1999. Immune function and survival of Great Tit nestlings in relation to growth conditions. Oecol- ogia 121:316-322. Ilmonen, R, D. Hasselquist, A. Langefors, and J. WiEHN. 2003. Stress, immunocompetence and leu- kocyte profiles of Pied Flycatchers in relation to brood size manipulation. Oecologia 136:148-154. Lamont, S. and J. Smyth. 1984. Effect of selection for delayed amelanosis on immune response in chickens. 2. Cell-mediated immunity. Poultry Sci- ence 63:440-442. Lessells, C. M. and P. T. Boag. 1987. Unrepeatable repeatabilities: a common mistake. Auk 104:116- 121. Lifjeld, j., P. Dunn, and L. Whittingham. 2002. Short-term fluctuations in cellular immunity of Tree Swallows feeding nestlings. Oecologia 130: 185-190. Lochmiller, R. L., M. R. Vestey, and J. C. Boren. 1993. Relationship between protein nutritional status and immunocompetence in Northern Bob- white chicks. Auk 110:503-510. Martin, L. B., A. Scheuerlein, and M. Wikelski. 2003. Immune activity elevated energy expendi- ture of House Sparrows: a link between direct and indirect costs? Proceedings of the Royal Society of London, Series B 270:153-158. Martin II, L. B., P. Han, J. Lewittes, J. R. Kuhlman, K. C. Klasing, and M. Wikelski. 2006. Phyto- hemagglutinin-induced skin swelling in birds: his- tological support for a classic immunoecological technique. Functional Ecology 20:290-299. Merino, S., J. Martinez, A. P. Mpller, S. Sanabria, F. DE Lope, J. Perez, and F. Rodriguez-Caabeiro. 1999. Phytohaemagglutinin injection assay and physiological stress in nestling House Martins. Animal Behavior 58:219-222. M0LLER, A. P. AND J. Erritz0e. 2003. Climate, body condition, and spleen size in birds. Oecologia 137: 621-626. Navarro, C., F. de Lope, A. Marzal, and A. M0ller. 2004. Predation risk, host immune response, and parasitism. Behavioral Ecology 15:629-635. Saino, N., S. Calza, and A. M0ller. 1997. Immu- nocompetence of nestling Barn Swallows in rela- tion to brood size and parental effort. Journal of Animal Ecology 66:827—836. SAS IN.STITUTE Inc. 1987. SAS/STAT guide for per- sonal computers. Sixth Edition. SAS Institute Inc., Cary, North Carolina, USA. Shaman, J., J. F. Day, and M. Stieglitz. 2003. St. Louis encephalitis virus in wild birds during the 1990 south Florida epidemic: the importance of drought, wetting conditions, and the emergence of Culex nigripalpus (Diptera: Culicidae) to arbovi- ral application and transmission. Journal of Med- ical Entomology 40:547-554. Shaman, J., J. F. Day, and M. Stieglitz. 2005. Drought-induced amplification and epidemic transmission of West Nile virus in southern Flor- ida. Journal of Medical Entomology 42:134-141. Shope, R. E. 1992. Impacts of global climate change on human health: spread of infectious disease. Pages 363-70 in Global climate change: impli- cations, challenges and mitigation measures (S. K. Majumdar, L. S. Kalkstein, B. Yarnal, E. W. Mill- er, and L. M. Rosenfeld, Editors). The Pennsyl- vania Academy of Science, Easton, USA. Smits, j. and T. Williams. 1999. Validation of the im- munotoxicology techniques in passerine chicks exposed to oil sand tailings water. Ecotoxicology and Environmental Safety 44:105-112. Stadecker, M., M. Lukic, A. Dvorak, and S. Leas- KOWiTZ. 1977. The cutaneous basophil response to phytohemagglutinin in chickens. Journal of Im- munology 118:1564-1568. SvENSSON, L. 1984. Identification guide to European passerines. Second Edition. Stockholm and British Trust for Ornithology, Thetford, United Kingdom. Tella, j. L., G. Bortolotti, M. G. Forero, and R. Dawson. 2000. Environmental and genetic varia- tion in T-cell-mediated immune response of fledg- ling American Kestrels. Oecologia 123:453-459. Tella, J. L., M. G. Forero, G. Bortolotti, J. A. Don- AZAR, G. Blanco, and O. Cellallos. 2001. Off- spring body condition and immunocompetence are negatively affected by high breeding densities in colonial seabirds: a multiscale approach. Pro- ceedings of the Royal Society of London, Series B 268:1455-1461. Westneat, D., j. Weiskittle, R. Endenfield, T. Kin- NARD, AND J. PosTON. 2004. Correlates of cell-me- diated immunity in nestling House Sparrows. Oec- ologia 141:17-23. ZuK, M. AND T. JoHNSEN. 1998. Seasonal changes in the relationship between ornamentation and im- mune response in Red Jungle Fowl. Proceedings of the Royal Society of London, Series B 265: 1631-1635. The Wilson Journal of Ornithology 1 20(4 );820— 829, 2008 INFLUENCE OE GRAZING AND AVAILABLE MOISTURE ON BREEDING DENSITIES OE GRASSLAND BIRDS IN THE CENTRAL PLATTE RIVER VALLEY. NEBRASKA DANIEL H. KIM.'-* WESLEY E. NEWTON.- GARY R. LINGLE.' ’ AND FELIPE CHAVEZ-RAMIREZ" .ABSTR.ACT. — We investigated the relationship between grassland breeding bird densities and both grazing and available moisture in the central Platte River Valley. Nebraska between 1980 and 1996. We also compared species richness and community similarity of breeding birds in sedge (Carex spp.) meadows and mesic grass- lands. Densities of two species had a significant relationship with grazing and six of seven focal species had a significant relationship with available moisture. Bobolink (Dolichonyx oryzivorus) and Brown-headed Cowbird {Molothriis ater) densities were lower in grazed plots compared to ungrazed plots, whereas Red- winged Black- bird (Agelaius phoeniceus) densities were greater in sedge-meadow plots compared to mesic grassland plots. Bobolink. Dickcissel (Spiza arnericana). and Brown-headed Cowbird were negatively associated with available moisture with breeding densities peaking during the driest conditions. Our results suggest that wet conditions increase species richness for the community through addition of wetland-dependant and wetland-associated birds, but decrease densities of ground-nesting grassland birds in wet-meadow habitats, whereas dry conditions reduce species richness but increase the density of the avian assemblage. We propose that wet-meadow habitats serve as local refugia for grassland-nesting birds during local or regional droughts. Received 9 October 2007. Accepted 21 March 2008. Grasslanid birds have experienced greater population declines over the past 40 years than any other avian group in North America (Askins 1993. Sauer et al. 2005). These de- clines in the Midwest and Great Plains have been attributed to loss of grasslands in breed- ing areas (Knopf 1994. Herkert 1995). Many recent studies of grassland birds have exam- ined the effects of habitat loss and fragmen- tation on avian occurrence and abundance (e.g.. Herken et al. 2003). Several studies also have examined the impacts of natural ecolog- ical drivers, such as burning and grazing, but only a few studies have evaluated the dynamic wet-dry cycles associated with grasslands and their influence on avian densities or nest suc- cess (Cody 1985. George et al. 1992. Zim- merman 1992, Igl and Johnson 1999, Fuhl- endorf et al. 2006). The dominant historical ecological factors ' Platte River Whooping Crane Maintenance Trust Inc.. 6611 West Whooping Crane Drive. Wood River. NE 68883. US.\. - U.S. Geological Sur\ey. Northern Prairie Wildlife Research Center. 871 1 37th Street SE. Jamestown, ND 58401. US.\. ^Current .Address: .Aim Consulting. 1568 L Road. Minden. NE 68959. US.A. ' Corresponding author: e-mail: Dkim@whoopingcrane.org were fire and grazing in the eastern mesic tail- grass prairies (Steinauer and Collins 1996), while the historical ecological drivers in the shortgrass prairies in the western Great Plains were climate and grazing (Bragg and Steuter 1996). The mixed-grass prairie ecosystem rep- resents the integration of the characteristics of the tallgrass and shortgrass prairies as influ- enced by climate and soils (Bragg and Steuter 1996). As the climate becomes more arid from the eastern tallgrass prairies to the western shortgrass plains, drought supersedes fire as the primary ecological factor. Vegetation structure may be the most im- portant aspect of habitat selection and suit- ability for grassland birds (Wiens 1969. 1973; Cody* 1981; George et al. 1992; Winter et al. 2005). The forb component in mixed-grass systems, which is responsible for the majority of the structural \ ariation. is directly affected by precipitation (Bragg and Steuter 1996). Grassland bird populations respond quickly to and are relatively tolerant of changes in veg- etation structure resulting from natural (e.g.. drought, wildfire) and anthropogenic distur- bances (e.g.. mowing, prescribed fire) (De- chant et al. 2003a. b. c; Herkert 2003). In ad- dition. fluctuations in precipitation affect food resource availability through reductions in in- sect biomass during drought years (Witten- berger 1980. George et al. 1992). 820 Kim et al. • GRAZING, MOISTURE, AND GRASSLAND BIRDS 821 Vegetation structure within the Platte River Valley of south-central Nebraska is influenced by both vegetation association and manage- ment practices. Vegetation associations in ri- parian wet-meadow pastures in this region vary from wetland emergent vegetation to dry-ridge grasslands based on elevation above the water table (Henszey et al. 2004). Sedge {Carex spp.) meadows and mesic tallgrass prairies are the dominant plant communities in native and restored areas within the riparian corridor and provide patches of suitable hab- itat for grassland birds in a mosaic of crop fields with strips of gallery forests. We evaluated the influence of grazing and available moisture on breeding densities of grassland birds in mesic prairies and sedge meadows. We had two major objectives: (1) document differences in avian assemblages between sedge-meadow and mesic prairie plots using community metrics (richness and similarity), and (2) examine the impacts of grazing and fluctuating wet-dry cycles on the density of seven grassland bird species. METHODS Study Area and Management. — Research occurred on lands owned and managed by the Platte River Whooping Crane Maintenance Trust Inc. (hereafter “the Trust”). The Trust owns and manages >4,000 ha of cropland, wet-meadow grasslands, and gallery forests in Hall, Buffalo, and Phelps counties along the Platte River in central Nebraska. The study area was Mormon Island Crane Meadows in Hall County, Nebraska (40°48'N, 98° 26' W). Wet meadows were managed using a three-pasture grazing rotation. Each pasture was grazed either in the early season (May- late Jun), mid season (late Jun-mid Aug) or late season (mid Aug-mid Oct). Rotation or- der changed annually (Table 1), and manage- ment included periodic prescribed burns to discourage woody plant encroachment (Cur- rier et al. 1985). We categorized pastures as grazed only if cattle were present in the early grazing season (i.e., before or during the bird censuses). Four permanent 16-ha plots were estab- lished on Mormon Island Crane Meadows (MICM) in Hall County, Nebraska and clas- sified as either sedge meadow or mesic prairie following recommendations of Henszey et al. TABLE 1. Study design and classification of Palmer Drought Severity Index (PDSI) into five cate- gories. Grazing treatments included grazing ( 1 ) and no grazing (0). No data (ND) were collected in 1991 and 1992 or in sedge plot 4 in 1981. Year PDSI (Aug) PDSI category (midpoint) Mesic plots 1 2 Sedge plots 3 4 1980 -1.89 A (-0.91) 1 1 1 1 1981 1.42 B (1.06) u 0 1 ND 1982 2.54 C (3.03) 0 0 0 1 1983 3.74 C (3.03) 1 0^' u 1 1984 4.59 D (4.99) u u 0 1 1985 4.20 D (4.99) u 0“ 0 1 1986 3.24 C (3.03) 0 0^‘ 0 1 1987 4.10 D (4.99) 0“ 0 0^' 1 1988 1.64 B (1.06) 0 u 0^' u 1989 0.60 B (1.06) 1 0 0“ u 1990 0.54 B (1.06) 0 1 0 0 1991 ND ND ND ND ND ND 1992 ND ND ND ND ND ND 1993 7.94 E (6.96) 1 0 0 1 1994 4.23 D (4.99) 0^' 1 0 0^‘ 1995 -0.59 A (-0.91) 0 0^' 0 0 1996 2.21 C (3.03) 0 u 0 0 ^ Indicates that plot was burned during spring. Most prescribed fire re- sulted in patchy bums resulting in no detectable affects of fire to species density in this study. (2004). Plots 1 and 2 were classified as mesic prairie and had greater topographic relief, rais- ing them above the water table and resulting in greater densities of goldenrod {Solidago spp.), Maximilian sunflower {Helianthus niax- imiliani), and yellow sweetclover {Melilotus officinalis). Plots 3 and 4 were classified as sedge meadow, were 15-60 cm closer to the water table, and included expansive seasonal wetland habitat. Plots 3 and 4 were character- ized by intermittent relic channels dominated by aquatic sedge (Carex aqiiatilis). which formed hummocks as a result of grazing and the presence of indigo bush (Amorpha fruti- cosa) near sloughs and relic channels. All plots occurred within the largest contiguous tract (>1,000 ha) of wet-meadow habitat in the central Platte River Valley (Lingle 1981). Spatially, no plot was further than 2.0 km from any other plot and all plots occurred within the same landscape. The plots do not repre.sent true replicates, but they v\ere on dif- ferent grazing schedules. Palmer Drought Severity Index. — We downloaded Palmer Drought Severity Index (PD.SI) data for central Nebraska (State Cli- 822 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 TABLE 2. Densities (males/ 10 ha) of species detected during breeding bird censuses at Mormon Island Crane Meadows in Hall County, Nebraska, 1980-1996. Densities are by plot and averaged across years. Plots 1 and 2 represent mesic mixed-grass prairie, and plots 3 and 4 represent sedge-meadow grasslands. Soecies Plot 1 Plot 2 Plot 3 Plot 4 Least Bittern {Ixohrychus exilis) Wood Duck {Aix sponsa) Mallard {Anas platyrhynchos) Northern Pintail (A. acuta) Blue-winged Teal (A. discors) Virginia Rail (Rallus limicola) Ring-necked Pheasant {Phasiamis colchiciis) Northern Bobwhite {Col inns virginianus) Killdeer {Charadrius vociferous) Upland Sandpiper {Bartramia longicauda) Wilson’s Snipe {Gallinago delicata) Wilson’s Phalarope {Phalaropus tricolor) Mourning Dove {Zenaida macroura) Sedge Wren {Cistothorus platensis) Common Yellowthroat {Geothlypis trichas) Grasshopper Sparrow {Ammodramus savannarum) Swamp Sparrow {Melospiza georgiana) Dickcissel {Spiza americana) Bobolink {Dolichonyx oryzivoriis) Red-winged Blackbird {Agelaius phoeniceus) Eastern Meadowlark {Sturnella magna) Western Meadowlark {S. neglecta) Yellow-headed Blackbird {Xanthocephalus xanthocephalus) Brown-headed Cowbird {Molothrus ater) Focal species density Mean species richness Species richness 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.125 0.042 0.000 0.000 0.042 0.000 0.021 0.042 0.229 0.000 1.615 1.500 0.000 0.000 0.000 0.000 0.094 0.125 0.208 0.000 0.000 0.000 1.823 3.208 0.000 0.000 2.354 1.396 5.313 4.938 1.333 0.167 0.083 0.167 1.729 1.573 0.000 0.000 1.917 1.542 16.083 14.323 6.8 6.1 14 11 0.042 0.021 0.000 0.021 0.250 0.417 0.000 0.000 0.323 0.313 0.000 0.042 0.000 0.000 0.000 0.000 0.313 0.477 1.302 1.208 0.063 0.042 0.438 0.813 0.198 0.510 0.146 0.271 0.000 0.063 0.792 1.052 0.000 0.083 3.031 1.271 5.677 4.969 7.323 7.208 0.083 0.021 0.948 1.271 0.000 0.083 2.563 2.625 21.640 19.604 8.1 10.0 16 21 mate Division 5) from the National Climate Data Center web site (http://www.ncdc.noaa. gov/oa/climate/research/monitoring.html). The PDSI is a long-term drought index cal- culated monthly and incorporates both precip- itation and temperature data (Palmer 1965). The PDSI expresses the severity of a wet (pos- itive values) or dry (negative values) period by factoring in both past and present condi- tions. More specihcally, values of zero to —0.5 indicate normal moisture conditions, -0.5 to -1.0 indicate incipient drought, -1.0 to -2.0 indicate mild drought, -2.0 to -3.0 indicate moderate drought, -3.0 to -4.0 in- dicate severe drought, and less than —4.0 in- dicate extreme drought. Similar adjectives are associated with positive values and wet peri- ods. Each state is divided into climate regions and PDSI values are calculated independently for each region. We evaluated the relationship between PDSI values from each month in the breeding season (May— Aug) and annual breeding bird abundances using simple cor- relation coefficients for each individual spe- cies and all focal species combined. PDSI val- ues from August had the highest correlations with all species except Grasshopper Sparrow (scientific names of birds are in Table 2), which was more highly correlated with PDSI values from May. We used PDSI values from August for further evaluation of species re- sponse to climatic factors with August values representing cumulative conditions through the breeding season. Avian Siirx'eys. — We conducted spot map surveys on each of the four 16-ha wet-mead- ow plots following standard Breeding Bird Census techniques (Van Velzen 1972). We systematically walked each plot on a mini- mum of eight mornings during the breeding season. Each territorial male present, its be- havior, and the location of all nests found were mapped. Only female Brown-headed Cowbirds were counted, except in 1980, when Kim et al. • GRAZING, MOISTURE, AND GRASSLAND BIRDS 823 they were present but not counted. Eight sur- veys for each plot were conducted annually between 0505 and 1059 hrs CST from 23 May to 25 June from 1980 to 1990, and from 1993 to 1996 for a total of 15 years of data on plots 1-3. Plot 4 had 14 years of data, because no data were collected in that plot in 1981. An- nual reports for these study sites were pub- lished in American Birds and Journal of Field Ornithology (Hay and Lingle 1982; Lingle and Whitney 1983a, b; 1991a, b; Lingle and Haugh 1984a, b; Lingle and Bedell 1989a, b; 1990a, b; Lingle et al. 1994a, b; Lingle 1995a, b; 1996a, b). Statistical Analyses. — We report species richness for each plot for all 15 years and av- eraged across years. We compared avian com- munities using Morisita’s (1959) Similarity Index, which ranks community similarity from 0.0 (no species overlap) to 1.0 (complete overlap). We calculated densities for each spe- cies on each plot (males/ 10 ha). We selected seven focal species (Upland Sandpiper, Grass- hopper Sparrow, Dickcissel, Bobolink, Red- winged Blackbird, Western Meadowlark, and Brown-headed Cowbird) that were either ob- ligate grassland nesting birds or strongly as- sociated with grassland habitats and which oc- curred consistently across all four plots. We used analysis of variance (ANOVA) to assess the effects of PDSI and grazing on the density of each bird species and overall avian diversity. We considered the design structure to be in the form of a strip-plot (Milliken and Johnson 1984). Given that PDSI is completely aliased with year (i.e., regional PDSI values are unique to each year but the same for all plots within a year), we first placed PDSI val- ues into five categories (0-20, 21-40, 41-60, 61-80, and 81-100 percentiles). We consid- ered these five levels of PDSI to be applied across each of the four plots simultaneously as “strips” with grazing applied randomly within each plot by PDSI combination. The results of this categorization and randomiza- tion varied (Table 1) with midpoint values of each PDSI category indicating reasonable spread and randomness of each factor level across years. We considered each of the four plots as random blocks for each PDSI by graz- ing level combination. Multiple occurrences of each PDSI category by grazing level com- bination were considered as sub-samples in time (Steel and Torrie 1980). We used PROG MIXED (SAS Institute Inc. 2004) to conduct the ANOVAs by considering three error terms and by computing Type III sums-of-squares using the most appropriate F-tests based on expected mean squares (Littell et al. 2006). We used Fisher’s protected LSD test for pair- wise comparisons between means for signifi- cant main effects and interactions at F = 0.05 (Milliken and Johnson 1984). All means re- ported are least squares means (Ismeans) with standard errors. We performed a similar anal- ysis for prescribed burns and PDSI, but no significant associations were found. Thus, we do not report means or statistics for burn ef- fects. RESULTS Plot Similarity and Species Richness. — Spe- cies richness (5) was greater for sedge-mead- ow plots {s = 18.5) compared to mesic-grass- land plots {s = 12.5) (Table 2). Mesic-prairie plots were more similar to each other (C^^ = 0.83) than to sedge-meadow plots (0.72 > C;, > 0.51) using Morisita’s Similarity Index (C;,); however, sedge-meadow plots had the greatest similarity (C^, = 0.90). PDSI Values. — Moisture conditions varied over the 15 years of study from mild drought (PDSI = —1.89) to extremely wet (PDSI = 7.94, flooded fields with only patches of dry ground available for nesting). Many of the bird species commonly associated with wet meadows and wetlands were detected only during extremely wet or flood conditions. For example, Virginia Rail, Wilson’s Snipe, and most waterfowl species were detected only on wet-meadow plots during relatively wet sea- sons (PDSI > 3.0). Bird Densities. — ANOVA results indicated the main effects of grazing and available moisture (i.e., PDSI), and their interaction, varied for each focal species and for all focal species combined (Table 3, Fig. 1 ). Data for all focal species combined indicated a signif- icant relationship with PDSI main effect but with no grazing or interaction effects. The densities for all focal species combined, av- eraged across PDSI levels, were similar be- tween ungrazed plots (32.1 ± 4.00 males/ 10 ha) and grazed plots (28.5 ± 3.92 male.s/lO ha). The densities of all focal species com- bined, averaged across the two grazing re- 824 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 TABLE 3. Analysis of variance Index) on species densities. assessing effects of grazing and available moisture (Palmer Drought Severity Species Grazino“ F^y-tes?’ PDSI Interaction F^y-test^ Local species G.3 = 7.34 F4.8 = 30.52** Gs = 0.13 Upland Sandpiper F, 3 = 9.49 G, = 9.31** = 5.59* Grasshopper Sparrow Fi3 = 3.56 G.8 = 3.81 G,8 — 3.18 Dickcissel G'3 = 0.03 G, = 68.21** F4 8 = 2. 1 5 Bobolink G3 = 15.33* G.8 = 8.82** g.8 = 1.51 Red-winged Blackbird Fi3 = 0.73 G.8 = 8.88** G,8 = 0.96 Western Meadowlark = 5.77 G, = 32.15** g.8 — 3.82 Brown-headed Cowbird F,3 = 11.62* Gs = 13.31** G.5 = 0.63 a Grazed fields had cattle in the plots during spot mapping; fallow fields were not grazed during the census period, b Significance level: * P < 0.05; ** P < 0.01. gimes, declined with increasing PDSI from 41.9 ± 4.78 males/ 10 ha at the lowest level of PDSI (-0.91) to 14.0 ± 8.15 males/10 ha at the highest level of PDSI (6.96). Bobolink and Brown-headed Cowbird den- sities indicated signihcant main effects for grazing and PDSI but no interaction between the two effects. Bobolink densities, averaged across all PDSI levels, were higher on un- grazed plots (11.5 ± 1.06 males/ 10 ha) than on grazed plots (5.3 ± 1.08 males/10 ha). Bobolink densities, averaged across the two grazing regimes, declined from 11.8 ± 1.15 males/10 ha at the lowest level of PDSI (-0.91) to 6.4 ± 1.50 males/10 ha at the high- est level of PDSI (6.96). Brown-headed Cow- bird patterns paralleled those of Bobolink. Densities were higher on ungrazed plots (4.3 ± 0.61 females/ 10 ha) than on grazed plots (2.6 ± 0.59 females/ 10 ha) and tended to de- cline with increasing levels of PDSI (3.9 ± 0.59 females/10 ha at PDSI 1.51 to 3.0 ± 0.99 females/10 ha at PDSI 6.96). Upland Sandpiper had a significant grazing by PDSI interaction. Their densities on grazed plots were near constant with a mean of 2.9 ± 0.34 males/10 ha across the PDSI levels. Densities on ungrazed plots peaked at inter- mediate levels of PDSI (2.6 ± 0.99 males/ 10 ha) with low densities at both the lowest PDSI level (0.6 ± 0.58 males/10 ha) and at the high- est PDSI level (0.0 ± 0.76 males/10 ha). Dickcissel, Western Meadowlark, and Red- winged Blackbird densities varied signihcant- ly with PDSI with no indication of grazing main effects or grazing by PDSI interactions. Dickcissel densities declined with increasing PDSI levels when averaged across both graz- ing treatments from 5.9 ± 1.05 males/ 10 ha (PDSI -0.91) to 0.3 ± 1.60 males/10 ha (PDIS 6.96). Western Meadowlark densities were low (all <4.0 males/ 10 ha) with no significant interaction occurring between PDSI and graz- ing, but Western Meadowlark experienced the most erratic patterns in densities across PDSI levels. Western Meadowlark declined signifi- cantly across the PDSI levels from 2.8 ± 0.29 males/10 ha (PDSI —0.91) to 2.0 ± 0.53 males/ 10 ha (PDSI 6.96). Red-winged Black- bird densities varied significantly with PDSI but had no grazing effects; the species de- clined only slightly from 8.6 ± 3.30 males/10 ha to 5.2 ± 3.60 males/ 10 ha with increasing PDSI. Grasshopper Sparrow had non-significant increases in densities under grazed treatments for the lowest PDSI level (-0.91), but had no difference between grazing treatments or across PDSI levels. Grasshopper Sparrow mean density was 2.76 ± 0.37 males/10 ha across all levels of PDSI and the two grazing treatments. DISCUSSION All study plots were in close proximity, but small variations in elevation (at the decimeter scale) influenced the vegetative communities (Henszey et al. 2004) and, concomitantly, avi- an communities. Sedge-meadow pastures had a higher water table and were more suscepti- ble to flooding. Thus, sedge-meadow plots had greater species richness due to wetland- associated and -dependant bird species using inundated fields during wet periods (Table 1). The extent of inundation depended on both local precipitation and river flood stage; in Territorial males/10 ha Kirn et al. • GRAZING, MOISTURE, AND GRASSLAND BIRDS 825 PDSl FIG. 1. Effects of moisture availability (Palmer Drought Severity Index or PDSl) on density of grassland birds. Open circles represent grazed plots, and closed circles represent ungra/ed pk)ts. I be PDSl ranges from — 2 (drier than normal, minimal standing water) to 8 (very wet. plots flooded). 826 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 some years, inundation may result from up- stream releases from reservoirs on the Platte River rather than local precipitation patterns. The Bobolink’s predilection for ungrazed pastures is supported by hndings from other studies (Bollinger 1995, Fritcher et al. 2004, Winter et al. 2005), and likely reflects this species’ preference for moderate to tall, dense vegetation (Dechant et al. 2003b). The nega- tive relationship between Bobolink densities and PDSI suggests that nest sites may be lim- ited during wet years or there are moisture- related movements of Bobolink between ri- parian and upland grasslands. Some studies of banded Bobolink populations noted site fidel- ity increases with habitat quality in stable en- vironments (Bollinger and Gavin 1989, Fletcher et al. 2006). Grazing and climatic fluctuations might remove predictability from riparian grasslands at our study sites, poten- tially facilitating movements at the patch scale. We did not find a strong relationship be- tween grazing and densities of Grasshopper Sparrow, Western Meadowlark, and Upland Sandpiper. Previous studies have reported these three species breed in actively managed (burned, grazed or hayed) grasslands (Dechant et al. 1999, 2003a, d), but the low overall den- sities for all three species limited the statistical power to detect a strong grazing-treatment ef- fect. Upland Sandpipers typically are associ- ated with early serai stages in grasslands, and Western Meadowlarks display intermediate habitat preferences with regional variation in their responses to disturbances (e.g., associ- ated with recent disturbance in tallgrass prai- rie, associated with rested pastures in short- grass prairie; Renfrew and Ribic 2001, Fritch- er et al. 2004). The Upland Sandpiper was the only species that had an interaction between grazing and PDSI; however, the species had minimal variation in densities on grazed plots compared to ungrazed plots. Standing water on the ungrazed plots may have reduced suit- ability of ungrazed plots during extremely wet years for Upland Sandpipers or resulted in their movements to upland grasslands Dickcissel and Red-winged Blackbird also displayed negative relationships with PDSI. Vegetation associations for Red-winged Blackbird appear more important than envi- ronmental conditions during selection of breeding habitat (Fletcher and Koford 2004). The association with PDSI may result from potential fluctuations in habitat suitability or in nest-site or prey availability while Dickcis- sels may respond to environmental conditions at either patch or regional scales. Drought conditions have been implicated as a proxi- mate cause of Dickcissel irruptions in some areas (Taber 1947, Igl 1991). In addition, Dickcissel densities during dry or average PDSI years may be affected by interspecific interactions with Red-winged Blackbirds. To our knowledge, no studies have reported di- rect behavioral dominance of Red-winged Blackbirds over Dickcissels (Yasukawa and Searcy 1995, Temple 2002), although both species share similar nest-site selection (i.e., above ground nesters in grasslands). Dickcis- sels in the Platte River Valley initiate nesting 2-3 weeks after Red-winged Blackbirds, and the larger blackbird males often chase Dick- cissel males from their territories (D. H. Kim, unpubl. data). Cowbird densities in our study were higher in ungrazed plots compared to grazed plots, which is counter to information in the litera- ture (e.g., Kostecke et al. 2003). Possible ex- planations include ( 1 ) regional cowbird den- sities are near capacity (Jensen and Cully 2005), (2) the spatial proximity of grazed pas- tures to ungrazed plots (<2 km) exert no en- ergetic costs on female cowbirds moving among breeding and feeding areas (Goguen and Mathews 2001), or (3) host species’ den- sities declined with grazing, offsetting the in- creased foraging opportunities for cowbirds. In our system, cowbird densities tracked Dick- cissel and Red-winged Blackbird densities, but especially Bobolink densities, suggesting cowbirds may exploit the most common host species in the study area or that these four species share similar habitat preferences in this region Available moisture may have profound ef- fects on the avian community in sedge mead- ows and mesic grasslands. Declining species richness associated with drier conditions was consistent with other grassland bird studies (George et al. 1992, Zimmerman 1992), but declining densities with higher PDSI values suggests that sedge meadows and mesic grass- lands may be less suitable for grassland birds during wet conditions. The pattern of increas- Kim et al. • GRAZING, MOISTURE, AND GRASSLAND BIRDS 827 ing density associated with drought conditions in the Platte River Valley appears opposite of other reported studies (e.g., Igl and Johnson 1999). Individual species have different hab- itat requirements; therefore, changes in vege- tation structure and amount of standing water should affect species individually. For exam- ple, Igl and Johnson (1999) found abundance of Le Conte’s Sparrow (Ammodramus lecon- teii) increased with moisture availability in grasslands enrolled in the Conservation Re- serve Program in the northern Great Plains. This unexpected pattern from our study could result from riparian wet-meadow grasslands acting as a refugia for species escaping drier conditions in regional upland areas. Dry con- ditions may cause birds to move from upland sites as vegetation structure and habitat suit- ability deteriorates with drought conditions. During dry years, wet-meadow habitats should remain suitable longer during the breeding season given the higher water table and greater soil moisture compared to upland sites. Van Horne (1983) cautioned against using density as an indicator of habitat quality, es- pecially in disturbed habitats. Additional re- search focusing on reproductive success over several years will be required to measure the importance of both climate and grazing on re- productive success for all seven species. For example, in Oregon, drought lowered annual reproductive success for Bobolinks (Witten- berger 1982). Dry conditions also may limit food availability for adults and young. Al- though some insects, such as orthopterans, may increase during dry conditions, the avail- ability and biomass of lepidopteran larvae may decrease, restricting growth rates for young nestlings unable to efficiently process chitonous prey. In either case, birds may be able to assess the potential of food or nest-site availability based on the condition of grass- lands at the beginning of the breeding season (early May in Nebraska) and to adjust their breeding densities accordingly. ACKNOWLEDGMENTS We thank all of the people who helped collect the breeding bird data, especially T. E. Labed/. L. D. Igl. N. D. Niemuth, T. L. Shaffer, and 2 anonymous re- viewers provided insightful comments that improved earlier versions of this manuscript. Work could not have been completed without the support of the U.S. Geological Survey’s Northern Prairie Wildlife Re- search Center and Platte River Whooping Crane Trust. This is Platte River Whooping Crane Maintenance Trust manuscript # 0702. LITERATURE CITED Askins, R. a. 1993. Population trends in grassland, shrubland, and forest birds in eastern North Amer- ica. Current Ornithology 11:1-34. Bollinger, E. K. 1995. Successional changes and hab- itat selection in hayfield bird communities. Auk 112:720-730. Bollinger, E. K. and T A. Gavin. 1989. The effects of site quality on the breeding-site fidelity in Bob- olinks. Auk 106:584-594. Bragg, T. B. and A. A. Steuter. 1996. Prairie ecol- ogy— the mixed prairie. Pages 53-65 in Prairie conservation: preserving North America’s most endangered ecosystem (E B. Samson and E L. Knopf, Editors). Island Press, Washington, D.C., USA. Cody, M. L. 1981. Habitat selection in birds: the roles of vegetation structure, competitors, and produc- tivity. BioScience 31:107-113. Cody, M. L. 1985. Habitat selection in grassland and open-country birds. Pages 191-226 in Habitat se- lection in birds (M. L. Cody, Editor). Academic Press, Orlando, Elorida, USA. Currier, R, G. Lingle, and J. VanDerwalker. 1985. Migratory bird habitat on the Platte and North Platte rivers in Nebraska. Platte River Whooping Crane Trust, Grand Island, Nebraska, USA. Dechant, J. A., M. E Dinkins, D. H. Johnson, L. D. Igl, C. M. Goldade, B. D. Parkin, and B. R. Eltiss. 1999. Effects of management practices on grassland birds: Upland Sandpiper. Version 12 August 2004. USDI. Geological Survey, Northern Prairie Wildlife Research Center, Jamestown, North Dakota. USA. www.npwrc.usgs.gov/resource/literatr/grasbird/ upsa/upsa.htm (acces.sed 25 September 2(X)7). Dechant, J. A., M. L. Sondreal, D. H. Johnson. L. D. Igl, C. M. Goldade, M. P. Nenneman, and B. R. Euless. 2003a. Effects of management practic- es on grassland birds: Grasshopper Sparrow. Ver- sion 12 August 2004. USDI. Geological Survey. Northern Prairie Wildlife Research Center. James- town, North Dakota, USA. www.npwrc.usgs.gov/ resource/1 iteratr/ grasbird/grsp/grsp.htm (accessed 25 September 2007). Dechant. J. A., M. I.. Sondreal. D. H. John.son. L. D. Igi., C. M. Goldade. A. L. Zimmi;rman. and B. R. Eui.i.ss. 2(K)3b. Effects of management practices on grassland birds: Bobolink. Version 12 August 2(X)4. USDI. Geological Survey. Noilhern Prairie Wildlife Research C’enter. Jamestown. NoHli Dakota. USA. w w w’.npw rc.usgs.gov/resource/literatr/grasbi 1x1/ bobo/ bobo.htm (accessed 25 .Septemix'r 2(K)7). DtiCHANT. .1. A., M. L. .SONDRI AI . D. M. Johnson. L. D. kiL, C\ M. Got DADr. A. I., /immlrman, and B. R. Et I iss. 2(X)3c. Effects of management practices on 828 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 grassland birds; Dickcissel. Version 12 August 2004. USDL Geological Survey, Northern Prairie Wildlife Research Center. Jamestown, North Dakota. USA. ww’w.npwrc.usgs.gov/resource/literatr/grasbird/ dick/ dick.htm (accessed 25 September 2007). Dech.\nt. J. a.. M. L. SONDRE.VL, D. H. Johnson. L. D. IGL. C. M. GoLD.-\DE. a. L. ZiMMERMVN, ANT» B. R. Euliss. 2003d. Effects of management practices on grassland birds; Western Meadowlark. Version 12 August 2004. USDI, Geological Survey. Northern Prairie Wildlife Research Center. Jamestown. North Dakota. USA. www.npwrc.usgs.gov/resource/literatr/ grasbird/weme/weme.htm (accessed 27 September 2007). Lletcher. R. J. and R. R. Koford. 2004. Consequenc- es of rainfall variation for breeding wetland black- birds. Canadian Journal of Zoology 82; 1316- 1325. Lletcher. R. J.. R. R. Koford, and D. A. Seaman. 2006. Critical demographic parameters for declin- ing songbirds breeding in restored grasslands. Journal of Wildlife Management 70;145-157. pRiTCHER. S. C.. M. A. Rumble, .and L. D. Ll.vke. 2004. Grassland bird densities in serai stages of mixed-grass prairie. Journal of Range Manage- ment 57;35 1-357. Luhlendorf. S. D., W. C. H.arrell, D. M. Engle, R. G. H.xmilton. C. a. Davis, .and D. M. Leslie. 2006. Should heterogeneity be the basis for con- servation? Grassland bird response to fire and grazing. Ecological Applications 16; 1706-1716. George, T. L.. A. C. Powler, R. L. Knight, and L. C. McEwen. 1992. Impacts of a severe drought on grassland birds in western North Dakota. Eco- logical Applications 2;275-284. Goguen, C. B. and N. E. M.xthews. 2001. Brown- headed Cowbird behavior and movements in re- lation to livestock grazing. Ecological Applica- tions 1 1 ; 1533-1544. Hay. M. a. .\nd G. R. Lingle. 1982. 1981 Breeding bird censuses along the Platte River in south-cen- tral Nebraska. American Birds 36; 105— 106. Henszey. R. J., K. Pfeiffer, .xnd J. R. Keough. 2004. Linking surface- and ground-water levels to ripar- ian grassland species along the Platte River in central Nebraska. USA. Wetlands 24;665-687. Herkert, j. R. 1995. Analysis of Midwestern breeding bird population trends — 1966—1993. American Midland Naturalist 134;41-50. Herkert. J. R. 2003. Effects of management practices on grassland birds; Henslow’s Sparrow. Version 12 August 2004. USDI, Geological Survey, Northern Prairie Wildlife Research Center. James- town, North Dakota, USA. http;// www.npwrc.usgs.gov/resource/literatr/grasbird/ hesp/hesp.htm (accessed 24 September 2007). Herkert, J. R., D. L. Reinking. D. A. Wiedenfeld, M. Winter, J. L. Zimmerman, W. E. Jensen, E. J. Finck. R. R. Koford, D. H. Wolfe. S. K. Sher- rod. M. A. Jenkins, J. Faaborg, and S. K. Rob- inson. 2003. Effects of prairie fragmentation on the nest success of breeding birds in the midcon- tinental United States. Conservation Biology 17; 587-594. iGL, L. D. 1991. The role of climate and mowing on Dickcissel {Spiza americana) movements, distri- bution, and abundance. Thesis. Iowa State Uni- versity, Ames, USA. iGL, L. D. A.ND D. H. Johnson. 1999. Le Conte’s Spar- rows breeding in Conservation Reserve Program fields; precipitation and patterns of population change. Studies in Avian Biology 19; 178—196. Jensen, W. E. and J. F. Cully. 2005. Geographic var- iation in Brown-headed Cowbird {Molothnis ater) parasitism on Dickcissels {Spiza americana) in Great Plains tallgrass prairie. Auk 122;648-660. Knopf, F. 1994. Avian assemblages on altered grass- lands. A century of avifaunal change in western North American. Studies in Avian Biology 15; 232-246. Kostecke, R. M., j. a. Koloszar, and D. C. Dear- born. 2003. Effect of a reduction in cattle stock- ing rate on Brown-headed Cowbird activity. Wild- life Society Bulletin 31; 1083—1091. Lingle, G. 1981. Mormon Island Crane Meadows — Protecting habitat for cranes along the Platte Riv- er, Nebraska. Pages 17-21 in Proceedings 1981 Crane Workshop (J. C. Lewis, Editor). National Audubon Society, Tavernier, Florida, USA. Lingle, G. 1995a. 1994 Breeding bird census — Sub- irrigated grassland and native hay. Journal of Field Ornithology 66(supplement); 1 1 1-1 12. Lingle, G. 1995b. 1994 Breeding bird census — Wet- land sedge meadow' I and II. Journal of Field Or- nithology 66(supplement); 100— 101 . Lingle, G. 1996a. 1995 Breeding bird census — Sub- irrigated grassland and native hay. Journal of Field Ornithology 67(supplement);86-87. Lingle, G. 1996b. 1995 Breeding bird census — Wet- land sedge meadow I and II. Journal of Field Or- nithology 67(supplement);76-77. Lingle, G. and P. Bedell. 1989a. 1988 Breeding bird census — Subirrigated native hay and grassland. Journal of Field Ornithology 60(supplement);60- 61. Lingle, G. .and P. Bedell. 1989b. 1988 Breeding bird census — Wetland sedge meadow' I and II. Journal of Field Ornithology 60(supplement);65-66. Lingle, G. and P. Bedell. 1990a. 1989 Breeding bird census — Subirrigated grassland and native hay. Journal of Field Ornithology 61(supplement);72- 73. Lingle. G. and P. Bedell. 1990b. 1989 Breeding bird census — Wetland sedge meadow I and II. Journal of Field Ornithology 61(supplement);78-79. Lingle, G. R. and K. L. H.augh. 1984a. 1983 Breeding bird census — Subirrigated grassland and native hay. American Birds 38; 120. Lingle, G. R. and K. L. H.augh. 1984b. 1983 Breeding bird census — Wetland sedge meadow I and II. American Birds 38; 124-1 25. Lingle. G. and W. S. Whitney. 1983a. 1982 Breeding Kim et al. • GRAZING, MOISTURE, AND GRASSLAND BIRDS 829 bird census — Subirrigated grassland and native hay. American Birds 37:101. Lingle, G. and W. S. Whitney. 1983b. 1982 Breeding bird census — Wetland sedge meadow I and II. American Birds 37:104. Lingle, G. and W. S. Whitney. 1991a. 1990 Breeding bird census — Subirrigated grassland and native hay. Journal of Lield Ornithology 62(supplement): 84-85. Lingle, G. and W. S. Whitney. 1991b. 1990 Breeding bird census — Wetland sedge meadow I and II. Journal of Lield Ornithology 62(supplement):77- 78. Lingle, G., S. Bergman, and J. Liske. 1994a. 1993 Breeding bird census — Subirrigated grassland and native hay. Journal of Lield Ornithology 65(sup- plement): 122-123. Lingle, G., S. Bergman, and J. Liske. 1994b. 1993 Breeding bird census — Wetland sedge meadow I and II. Journal of Lield Ornithology 65(supple- ment): 107-108. Littell, R. C., G. a. Milliken, W. W. Stroup, R. D. Wolfinger, and O. Schabenberger. 2006. SAS for mixed models. Second Edition. SAS Institute Inc., Cary, North Carolina, USA. Milliken, G. A. and D. E. Johnson. 1984. Analysis of messy data. Volume I. Designed experiments. Van Nostrand Reinhold Co., New York, USA. Morisita, M. 1959. Measuring of interspecific asso- ciation and similarity between communities. Memoirs of the Laculty of Science, Kyushu Uni- versity Series E 7:65-80. Palmer, W. C. 1965. Meteorological drought. Re- search Paper Number 45. U.S. Department of Commerce, Weather Bureau, Washington, D.C., USA. Renfrew, R. B. and C. A. Ribic. 2001. Grassland birds associated with agricultural riparian practices in southwestern Wisconsin. Journal of Range Man- agement 54:546-552. SAS Institute Inc. 2004. SAS Online Doc® 9.1.2. SAS Institute Inc., Cary, North Carolina, USA. Sauer, J. R., J. E. Hines, and J. Ballon. 2005. The North American Breeding Bird Survey, results and analysis 1966-2004, Version 2004.1. USDl, Geological Survey, Patuxent Wildlife Research Center, Laurel, Maryland, USA. www.mbr- pwrc.usgs.gov/bbs/bbs.html. (accessed 25 Sep- tember 2007). Steel, R. G. D. and J. H. Torrie. 1980. Principles and procedures of statistics: a biometrical approach. Second Edition. McGraw-Hill Book Co., New York, USA. Steinauer, E. M. and S. C. Collins. 1996. Prairie ecology-the tallgrass prairie. Pages 39-52 in Prai- rie conservation: preserving North America’s most endangered ecosystem. (L B. Samson and L L. Knopf, Editors). Island Press, Washington, D.C., USA. Taber, R. D. 1947. The Dickcissel in Wisconsin. Pas- senger Pigeon 9:39-46. Temple, S. A. 2002. Dickcissel (Spiza americana). The birds of North America. Number 703. Van Horne, B. 1983. Density as a misleading indi- cator of habitat quality. Journal of Wildlife Man- agement 47:893-901. Van Velzen, W. 1972. Breeding-bird census instruc- tions. American Birds 26:1007-1010. Wiens, J. A. 1969. An approach to the study of eco- logical relationships among grassland birds. Or- nithological Monographs 8:1-93. Wiens, J. A. 1973. Pattern and process in grassland bird communities. Ecological Monographs 43: 237-270. Winter, M., D. H. Johnson, and J. A. Shaffer. 2005. Variability in vegetation effects on density and nesting success of grassland birds. Journal of Wildlife Management 69:185-197. WiTTENBERGER, J. L 1980. Vegetation structure, food- supply, and polygyny in Bobolinks {Dolichonyx oryzivonis). Ecology 61:140-150. WiTTENBERGER, J. L. 1982. Lactors affecting how male and female Bobolinks apportion parental invest- ments. Condor 84:22-39. Yasukawa, K. and W. a. Searcy. 1995. Red-winged Blackbird {Agelaius phoeniceus). The birds of North America. Number 184. Zimmerman, J. L. 1992. Density-independent factors affecting the avian diversity of the tallgrass prairie community. Wilson Bulletin 104:85-94. The Wilson Journal of Ornithology 1 20(4);830— 839, 2008 RESPONSE OF SONGBIRDS TO RIPARIAN WILLOW HABITAT STRUCTURE IN THE GREATER YELLOWSTONE ECOSYSTEM BRIAN F. M. OLECHNOWSKF'^ AND DIANE M. DEBINSKF ABSTRACT. — We compared the structure of riparian willow {Salix spp.) habitat and songbird diversity across two regions of the Greater Yellowstone Ecosystem: Gallatin National Forest and Yellowstone National Park in the north and Grand Teton National Park in the south. The average height of willows was greater (151 vs. 65.9 cm) in the Teton region, and the average density of willows was greater (45.2 vs. 26.1%) in the Gallatin. The average height of willows was the most important variable explaining songbird species richness and abundance across these two regions. Songbird richness and abundance was greater (6.1 vs. 3.1 species; 11.7 vs. 5.6 indi- viduals) in the Teton region. Larger patch size in the Tetons could be a factor contributing to the higher level of diversity but was not statistically significant. Individual species responses to habitat structure varied based on the nesting height preference of the species. Species that nest above the ground or in taller vegetation had abundance positively correlated with average willow height (Yellow Warbler [Dendroica petechia] P < 0.001; Fox Sparrow [Passerella iliaca] P = 0.001). Yellow Warbler, Willow Flycatcher (Empidonax trailii). Fox Spar- row, and Common Yellowthroat (Geothylpis trichas) all had higher abundances in the Teton sites. The difference in willow habitat structure across these regions is likely influenced by historic differences in elk (Cervus elaphus) browsing in the northern regions of the Greater Yellowstone Ecosystem. Received 4 September 2007. Accepted 26 February 2008. Many studies have examined how avian di- versity responds to structural characteristics of habitat (Mac Arthur and Mac Arthur 1961, Jackson 1992, Yamaura et al. 2006). Most of this work has focused on bird species inhab- iting prairie and forested ecosystems in North America (Cody 1974, Zimmerman 1992, Powell 2006, Sallabanks et al. 2006). These studies confirmed the importance of preserv- ing and restoring prairie and forested ecosys- tems, but few studies have considered how avian biological diversity may respond to hab- itat changes in other environments. Montane riparian willows (Salix spp.) support the great- est diversity of songbird species in the semi- arid portions of the western United States (Knopf et al. 1988, Dobkin et al. 1998). Thus, they warrant examination with respect to how their structure affects songbird diversity and abundance patterns. Riparian plant communities dominated by willows in the mountain regions of the west- ern United States provide a number of impor- tant ecological functions including stabilizing stream banks, moderating water temperatures and the surrounding microclimate, providing riparian corridors for dispersal of plants and ' Department of Ecology, Evolution, and Organis- mal Biology, 253 Bessey Hall, Iowa State University, Ames, lA 50014, USA. 2 Corresponding author; e-mail: bolech@gmail.com vertebrates, and supporting a rich array of spe- cies from insects to birds to large ungulates and predators (Wingington and Beschta 2000, NRC 2002, Ripple and Beschta 2004b). Some willow habitats have been in decline for a number of decades (Singer et al. 1998). Pei- netti et al. (2002) documented a 20% decline in riparian shrub cover and a 55% decline in tall willow cover in two valleys of Rocky Mountain National Park, Colorado. Ammon and Stacey (1997) and Berger et al. (2001) identified a number of songbird species in- cluding Yellow Warblers (Dendroica pete- chia), Willow Flycatchers (Empidonax trailii). Fox Sparrows (Passerella iliaca), and Calli- ope Hummingbirds (Stellula calliope) that were less abundant in areas where vegetative structure of willows had been altered by browsing. Elk (Cervus elaphus) are the major native browsers of willows in the mountain west in the United States, and a number of studies in the Greater Yellowstone Ecosystem (GYE) have demonstrated decreased willow growth, reproduction, and cover in areas where elk are abundant (Kay and Chadde 1992; Ripple and Beschta 2004a, b). Competition for food, es- pecially in the winter, has caused ungulate populations to browse on willows, aspen (Po- pulus tremuloides), and cottonwood (Populus spp.) stands in the ecosystem (Romme et al. 1995, Ripple et al. 2001, Smith 2001). Singer 830 Olechnowski and Debinski • SONGBIRD RESPONSE TO WILLOW HABITAT STRUCTURE 83 1 et al. (1994) found that 47% of willow com- munities in the winter range of elk in northern Yellowstone National Park had been height suppressed. Kay and Chadde (1992) studied long-term exclosures and found that elk re- duced potential willow seed production on the northern range of Yellowstone by 100%. They concluded that some willows have not been able to produce seeds for the last 50 years. The objectives of our study were to: (1) compare willow habitat structure in the north- ern and southern portions of the Greater Yel- lowstone Ecosystem, (2) examine how song- bird communities respond to structure of wil- low habitat in this montane riparian system, and (3) discuss conservation implications of the relationship between avian diversity and willow structure in the GYE. We expected willow habitat to be more structurally com- plex in the Teton Valley (southern GYE) com- pared to the Gallatin (northern GYE) because of high levels of elk herbivory in the Gallatin River Valley (Singer et al. 1998; Ripple and Beschta 2004a, b). We hypothesized that songbird diversity would be directly correlat- ed with the variables used to describe the veg- etative structure of the willow habitat (height and density) of the vegetation. METHODS Study Area. — Our research was conducted in two areas of the Greater Yellowstone Eco- system: Gallatin National Forest and north- west Yellowstone National Park (northern re- gion of the GYE, hereafter referred to as the “Gallatin region”), and in Grand Teton Na- tional Park (southern region of the GYE, here- after referred to as the “Teton region”). Non- forest cover types within the ecosystem range from hydric willow and sedge (Carex spp.) meadows to alpine rock meadows. The two regions have distinct landscapes and differ in average patch size, but support similar ripar- ian willow habitat, plant, and songbird species (Debinski et al. 2001). Hydric meadows in the Gallatin region occur near creeks and rivers flowing along narrow valleys. The largest val- leys are 1-3.3 km wide and up to 12.5 km long (Saveraid et al. 2001). The Teton region consists of large areas of sagebrush (Artemisia spp.) and willow flats with forested regions generally occurring on buttes, foothills, and mountains (Saveraid et al. 2001). 3'he feton Valley lies east of Jackson Lake and is 5-19 km wide and 76 km long (Saveraid et al. 2001). Sampling Sites. — Sampling sites were es- tablished using remotely-sensed classification of the montane meadow habitats in the Great- er Yellowstone Ecosystem to identify a mois- ture gradient in montane meadows. A map was produced that displayed this moisture gra- dient across meadows within the ecosystem. Hydric meadows, composed of riparian veg- etation including willows, were identified us- ing this map in 1997. Sites were considered suitable for sampling if they were at least 100 X 100 m in size, at least 500 m or farther from other sites, and within 8 km of a road or trail (Debinski et al. 2000). Willows had to be present at five of the 15 points (stops) at each site where we measured habitat structure to be included in our analysis. Eleven willow sites in the Gallatin and 14 sites in the Tetons met our criterion for analysis. Sites in the Gallatin were in the northwest region of Yellowstone National Park and within 10 km of the Gal- latin River in the southeast region of Gallatin National Forest. Sites in the Tetons were with- in 8 km (east) of Jackson Lake in Grand Teton National Park near Moran, Wyoming. A cen- ter point was established at each site and was marked with a 1.25-m fiberglass or wooden post driven into the ground. UTM coordinates were taken at each center-point using a Global Positioning System to allow relocation of the points. Willow' Habitat Structure. — We assessed the three-dimensional structure of the willow veg- etation at 25 sampling sites in July 2()()(). Veg- etation sampling was limited to 1 year because of the intense effort required to obtain these data (4 people for 4 hrs/site). We assumed there were no major changes in grazing or browsing pressure, or effects of fire across years. We sampled the avifauna at each site annually between 1997 and 2001, and did not observe any qualitative differences in willov\ structure across these years (DMD, pers. obs.). Three 100-m transects were established al each site labeled "west", “center”, and “east” and were aligned north-south al a dis- tance of 30 m from each other. J'hc center transect was placed through the pre-estab- lished site center. Five points. 20 m apart, were established along each transect, lor a to- 832 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 tal of 15 points at each site. Data collected at each point included: vegetation type (grass, forb, shrub), maximum height of each vege- tation type, and percent cover within each 10- cm vertical increment of the pole by that veg- etation (modified from Robel et al. 1970). The pole used for density measurements was ver- tically partitioned into 10-cm segments up to 125 cm. Each segment was considered a “zone” and the percent density was estimated as: 0 = 0; 1-20 = 1; 21-40 = 20; 41-60 = 40, etc. Measurements were taken 1 m from the pole at each of the stops at a site. Percent density for all the points at a site was obtained by averaging across zones and points to pro- vide an overall density value of the willows for that site. High density values were indic- ative of habitat that was consistently dense, while lower obscurity values were indicative of habitat that consists of “spaces” in the wil- low structure. The presence of willows could be assessed from 5 to all 15 of the possible points at each study site. We also recorded the specific willow species found at each site. The three major species in the Greater Yellowstone region are S. wolfii, S. boothii, and S. bebbi- ana. Salix boothii was the tallest of these three species, reaching 7 m at maturity, S. bebbiana reaches 3-4 m at maturity, and S. wolfii ranges from 0.5 to 1.5 m at maturity (USDA 2007). Relative abundance of these three species did not differ across the Gallatin and Teton re- gions (DMD and BFO, unpubl. data). Our analysis focuses on bird responses to vegeta- tion structure rather than willow species com- position because birds primarily choose where to nest or forage based on the physical struc- ture of the habitat (Yamaura et al. 2006). Bird Surveys. — Songbird density and abun- dance data were collected using 50-m radius point count surveys. Surveys were conducted at all 25 of our willow sites in 1999 (n = 1 1 in the Gallatin region, n = 14 in the Teton region), and at 12 of the 25 sites across the Gallatin region {n = 5) and Teton region {n = 7) in 1997-1998 and 2000-2001. These surveys were performed one to three times per year during 1997-2001 and from 0530-1030 hrs for each study site from late-May/early- June to mid-July. We only consider the first round of these surveys (during late May/early Jun) at each of our sampled sites due to un- even sampling effort (we were not able to re- peat surveys at all sites after mid-June; the highest sampling size for sites was attained using only the first round of surveys conduct- ed in late May/early Jun). Each survey in- volved two people recording and observing for 15 min. All researchers were trained in identifying birds by sight and sound. Birds were not surveyed if it was snowing or rain- ing, but surveys were conducted if there was light frost or snow on the ground because bird activity was not reduced under these condi- tions (DMD, pers. obs.). Each individual bird seen and/or heard within the 50-m radius dur- ing a survey was recorded, and its location and distance from the observer were mapped. Flyovers were not included in the analysis be- cause we were interested in those individuals that gave evidence of using the habitat for for- aging or reproduction. Behaviors including singing, chirping, carrying nesting material, and feeding fledglings were also recorded (Gill 1995). Statistical Analyses. — We used one-way analysis of variance (ANOVA) to assess dif- ferences in the three-dimensional structure of willow habitat (patch size, heights and densi- ties of willows) between sites in the Gallatin region versus those in the Teton region. Sim- ple and multiple linear regressions were used to examine the relationships of the vegetation variables and songbird species richness and abundance. Patch size was also considered in these analyses, and we compared patch sizes for our study sites using FRAGSTATS spatial analysis program (McGarigal and Marks 1994). We analyzed the songbird data for 1999 when all 25 sites were surveyed and for 1997-2001 when 12 sites were surveyed for multiple years. We averaged species richness and abundance at each site across years to ex- amine how overall levels of diversity corre- lated with habitat measurements. These aver- ages were also used to compare songbird spe- cies richness and abundance between the Gal- latin and Teton regions. The effects of site and year were considered as co-factors when com- paring abundances of nine individual species that used willow habitat across both regions of the GYE. We calculated the average annual abundance by region for each species, cor- recting for the number of sites surveyed. We considered year as a co-factor in this analysis because individual species did not show time Olechnowski and Debinski • SONGBIRD RESPONSE TO WILLOW HABITAT STRUCTURE 833 TABLE 1. Willow vegetation measurements in the Gallatin region (northern GYE) and the Teton region (southern GYE). Measurements were taken during summer 2000. Category Gallatin region (n = 11) X ± SE Teton region (n = 14) X ± SE F p Willow height, cm 65.9 ± 13.0 151.6 ± 11.6 24.2 <0.0001 Willow density, % 45.2 ± 4.1 26.1 ± 3.8 11.5 0.003 Patch size, ha 1.8 ± 80.5 252.8 ± 95.2 4.0 0.07 Extent of willows at site, % 70.8 ± 6.7 61.8 ± 6.0 1.0 0.33 trends in abundance during the years we re- corded our observations. Multiple compari- sons of species abundances across these two regions were corrected using the Bonferroni method. We calculated the average Shannon- Weaver Diversity Index (Shannon and Weaver 1962) across the 12 sites that were sampled between 1997 and 2001 to compare overall levels of songbird diversity between the two regions of the GYE. Analyses were performed using the statistical software package S-Plus Version 7.0 (Insightful Corp 2005). RESULTS Willow Structure in the Gallatin and the Te- tons. — The overall three-dimensional structure of willow habitat was strikingly different when comparing the northern and southern re- gions of the GYE (Table 1). Willows in the Gallatin region were shorter and more dense, whereas willows in the Teton region were tall- er and less dense. Average willow heights and TABLE 2. Multiple linear regression models showing (A) Response of songbird abundance and richness to willow structure in 1999 across all 25 wil- low sites in the Greater Yellowstone Ecosystem. (B) Response of average songbird abundance and average songbird richness across 12 sites in which birds were surveyed from 1997 to 2001. Songbird abundance Songbird richness Category F P F p (A) Willow height 6.2 0.022 12.4 0.002 Willow density 0.4 0.52 1.3 0.67 Height X density 0.5 0.48 0.05 0.82 Patch size 0.8 0.41 2.8 0. 1 3 (B) Willow height 4.7 0.06 6.1 0.040 Willow density 0.7 0.44 0.1 0.73 Height X density 1.7 0.22 2.3 0.17 Patch size 2.1 0. 1 9 3.5 0.10 densities were negatively correlated {F = 14.8, df = 23, = 0.40, P = 0.0009). Av- erage patch size of willow sites surveyed was much greater in the Teton region, but patch sizes of sites surveyed in this region varied considerably, and this difference was not sig- nihcant {F = 4.0, df = 1, P = 0.07). The extent of willows surveyed at each site, (mea- sured as the number of points where willows were present divided by 15) did not differ be- tween the Gallatin and the Tetons (n = 11 for the Gallatin and n = 14 for the Tetons). Willow Structure and Songbird Species Di- versity in Willow Habitat of the GYE. — Two separate analyses were conducted to examine songbird abundance and richness versus veg- etative structure. The hrst used all 25 sites where willows were measured and in which one round of bird surveys was completed in 1999. Average willow height was positively correlated with both songbird abundance (F = 6.2, df = 19, P = 0.022) and richness (P = 12.4, df = 19, P = 0.002) across sites in the multiple linear regression model (Table 2A). Patch size differences did not have a signih- cant influence on either abundance or richness at our sites (P = 0.8, df = 19, P = 0.41; P = 2.8, df = 19, P = 0.13, respectively). The second analysis used the 12 willow sites in the Gallatin and Tetons that were surveyed over multiple years (1997—2001), including the data collected in 1999. Species richness and abundance were averaged at each site across this time period. Average willow height was positively correlated with average species richness (P = 6.1, df = 6, P = 0.04). but not with average species abundance (P = 4.7, dl = 6, P = 0.06) (Table 2B). Differences in patch size at our sites were not significant in explaining abundance or richness of bird spe- cies across multiiTle years of surveys (/•' = 2.1, df = 6, P = 0.19; P = 3.5, df = 6. P = 0.10. 834 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 TABLE 3. (A) Average songbird species richness and abundance across 25 willow sites in the Greater Yellowstone Ecosystem in 1999. (B) Average songbird species richness, abundance, and the Shannon-Weaver Index across 12 sites in which birds were surveyed from 1997 to 2001. Gallatin region x ± SE Teton region x ± SE F p (A) (n = 11) (n = 14) Abundance 5.4 ± 1.1 8.8 ± 0.9 4.7 0.040 Richness 3.2 ± 0.8 5.8 ± 0.6 5.3 0.030 (B) (n = 5) (n = 7) Abundance 5.6 ± 2.0 11.7 ± 1.4 38.0 0.0001 Richness 3.1 ± 1.3 6.1 ± 1.1 19.3 0.001 Shannon-Weaver Index 0.9 ± 0.4 1.66 ± 0.2 15.7 0.003 respectively). None of the habitat variables (willow height, willow density, patch size, ex- tent of willows) was significantly related to average species abundance across multiple years, but average willow height was the most important variable in this model when ex- plaining both songbird species richness and abundance. Songbird Diversity in the Tetons and Gal- latin.— Two separate analyses were conducted to examine differences in overall songbird richness and abundance in the Teton and Gal- latin regions. The first used all 25 sites where willows were measured and for which one round of bird surveys was completed in 1999. The Tetons had higher levels of both songbird species richness {F = 5.3, df = 1, P = 0.030) and abundance {F = 4.7, df = 1, P = 0.040) (Table 3 A). The second analysis used the 12 willow sites in the Gallatin and Tetons that were surveyed across multiple years (1997- 2001). The Tetons had higher levels of both average songbird species richness (F = 19.3, df = 1, P = 0.001) and average abundance (P = 38.0, df = 1, P = 0.0001). The average Shannon-Weaver Index of Diversity was also greater in the Teton region (P = 15.7, df = 1, P = 0.003) (Table 3B). Individual Species Responses. — Four of nine individual species across the Gallatin and Teton regions (data pooled from the sites within each region and across years in which surveys were conducted) were more abundant in the Teton region (Table 4). These four spe- cies were Yellow Warbler (P = 46.2, df = 1, P < 0.0001), Fox Sparrow (P = 12.0, df = 1, P = 0.008), Common Yellowthroat (Geoth- lypis trichas) (P = 9.6, df = 1, P = 0.030), and Willow Flycatcher (P = 7.5, df = 1, P = 0.040). Abundance of three of these four spe- cies was significantly correlated with average willow height at 12 sites across both regions of the GYE: Yellow Warblers (P < 0.001, P = 0.26), Fox Sparrows (P - 0.001, P = 0.16), and Willow Flycatchers (P = 0.045, P = 0.07) (Fig. 1.). The average abundance of three other species did not differ across the two regions of the GYE, but still had a posi- TABLE 4. Average annual abundance by region for nine species of songbirds that used willow habitats in the Greater Yellowstone Ecosystem from 1997 to 2001, corrected for number of sites surveyed in each region. Gallatin region (n = 5) X ± SE Teton region (n = 7) X ± SE F p Yellow Warbler 0.2 ± 0.1 2.5 ± 0.3 46.2 <0.0001 Lincoln’s Sparrow 2.0 ± 0.4 1.7 ± 0.2 0.4 1.0 Willow Flycatcher 0.1 ±0.1 0.5 ± 0.1 7.5 0.040 Fox Sparrow 0 ± 0 0.5 ± 0.1 12.0 0.008 Common Yellowthroat 0.4 ± 0.1 1.5 ± 0.3 9.6 0.030 Dusky Flycatcher 0.1 ± 0.4 0.3 ± 0.6 1.0 1.0 American Robin 0.2 ± 0.1 0.3 ± 0.1 0.5 1.0 Song Sparrow 0.3 ± 0.2 0.5 ± 0.2 0.6 1.0 White-Crowned Sparrow 0.1 ± 0.1 0.5 ± 0.2 3.0 0.81 Olechnowski and Debinski • SONGBIRD RESPONSE TO WILLOW HABITAT STRUCTURE 835 FIG. 1. Relationship of individual species abundances witli average willow heights (cm) at 12 sites in the Greater Yellowstone Ecosystem. Songbird abundance data are pooletl across P)97 to 2001. 836 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 live correlation with willow height. These spe- cies include Dusky Flycatcher (Empidonax oherholseri) (P = 0.018, E = 0.09), American Robin {Turdus migratorius) {P — 0.013, E = 0.10), and White-crowned Sparrow (Zonotri- chia leucophrys) {P — 0.012, E = 0.10). DISCUSSION Dijferences in Willow Habitat Structure and Songbird Diversity. — Willow habitat was more structurally complex in the Teton than the Gallatin region and this habitat complexity was reflected in abundance and distribution of willow-dependent songbirds. Willows in the Teton region were taller and were less dense. Levels of overall songbird diversity and abun- dance were positively correlated with height of willows, and a greater number of individual songbirds and species were recorded in the Teton region. We acknowledge that larger patch size of willows in the Tetons could be a contributing variable for greater songbird species richness and abundance in the Teton region (despite the lack of statistical signifi- cance). However, other studies have also re- ported vegetation height as an important var- iable in explaining songbird species richness and abundance in forested (Barbaro et al. 2005) and prairie systems (Zalba and Cozzani 2004). The abundance of several individual species had a significant positive relationship with average willow height across the Gallatin and Teton regions. Yellow Warblers and Fox Sparrows had the strongest positive relation- ship with average willow height. Our results build on those of Berger et al. (2001) who conducted a similar study in the GYE and re- ported that structure of willow vegetation in- fluences songbird species richness and abun- dance. Our study strengthens the biological signal of these patterns because we conducted bird surveys over multiple years (vs. 1 yr, Ber- ger et al. 2001 ) and we surveyed a larger num- ber of sites than Berger et al. (2001 [n = 6]). High levels of herbivory can have cascad- ing effects on species diversity and abundance across a variety of taxa, including plants, birds, and herbivores (McLaren and Peterson 1994; Ripple et al. 2001; Ripple and Beschta 2004a, b; Beyer 2007). Elk numbers and her- bivory increased during the period of time gray wolves (Canis lupus) were absent and elk control efforts (shooting and trapping) were relatively low in the GYE (mid 1920s to 1995) and, as a result of overbrowsing, willow recruitment decreased (Kay 1990; Kay and Chadde 1992; Singer et al. 1994, 1998; Ripple and Larsen 2000; Ripple et al. 2001; Ripple and Beschta 2004a, b; Beyer 2007). Our re- search demonstrates that willow habitats char- acterized by shorter willow vegetation have lower levels of songbird species richness and abundance. Songbirds use these habitats for foraging and nesting, and these activities may be affected by herbivory. Historically, winter populations of elk have been lower in the Te- ton region because many elk in the southern region of the GYE spend their winters on the National Elk Refuge (NER) in Jackson Hole, where supplemental food sources are provided (Smith et al. 2004). The additional food pro- vided at the refuge has likely decreased browsing on willows within boundaries of Grand Teton National Park. Anderson (2007) reported that feed stations at the NER attract elk that browse on willows in the surrounding area, and effectively decrease willow height, songbird species richness, and songbird abun- dance. Our study sites were not near these feeding stations, but the Anderson (2007) study emphasizes that intense elk browsing negatively affects willow habitat and songbird communities. These historical patterns in elk abundance and grazing across the GYE are likely contributing factors in explaining the striking structural difference in willow habitat across our two regions of study. Ungulate browsing in the northern GYE has also affected American beaver {Castor cana- densis) populations, which may be affecting riparian willow habitats. No beaver were doc- umented on the northern range of the ecosys- tem during 1996, shortly after gray wolf re- introduction (Smith et al. 2003). Collins (1976) and Gribb (2004) documented active beaver dens in the Teton region during their studies. Ripple and Beschta (2004a) suggested that as elk increased in abundance in the northern range of the GYE through the mid to late 1900s, beaver populations declined and became practically extinct. Beaver occurred in the northern GYE before wolves were extir- pated in 1926 (Warren 1926, Jonas 1955). The overall vigor and success of willows is related to activity of beaver populations because bea- ver affect the level of surface water in the ri- Olechnowski and Debinski • SONGBIRD RESPONSE TO WILLOW HABITAT STRUCTURE 837 parian environment (Ripple and Beschta 2004b). Medin and Clary (1990) also reported that avian species richness in riparian habitat of the mountain west was greater in ponds with beaver activity. Naiman et al. (1988) and Pollock et al. (1995) indicated that beaver can increase plant, invertebrate, and vertebrate di- versity in the system. Individual Species Responses. — We found varying abundances of nine songbird species that use riparian willow habitats, but those species that nest higher in the willow vege- tation showed consistent trends. Yellow War- blers, Fox Sparrows, Willow Flycatchers, and Common Yellowthroats were significantly more abundant in the Tetons (where willow heights were greater), and other species (e.g., Lincoln’s Sparrow [Melospiza lincolnii], American Robin, White-crowned Sparrow) had no clear difference in abundance levels. Berger et al. (2001) found that Yellow War- blers, Fox Sparrows, and Willow Flycatchers declined in areas that were heavily grazed by moose {Alces alces). White-crowned Sparrow and Lincoln’s Sparrow were found in greater abundances in more heavily grazed areas in Berger’s (2001) study. Additionally, Anderson (2007) found that Fox Sparrows and Willow Flycatchers were especially sensitive to habi- tat changes due to elk browsing. Reasons for differential responses among songbirds may be explained by nest placement. Salt (1957) indicated Lincoln’s Sparrow and White- crowned Sparrow nest closer to the ground level in the GYE (and generally reside in low- er vegetation) where herbivory impacts would presumably not be as great. Data from Am- mon and Stacey (1997) and Berger et al. (2001) suggests that ground nesting birds are less affected by grazing than species that nest above ground. In contrast. Yellow Warblers, Willow Flycatchers, and Fox Sparrows prefer to reside and nest in taller vegetation. Yellow Warblers, in particular, nest much higher above ground level. Thus, it is not surprising that Yellow Warblers had the most pro- nounced difference in abundance across the Gallatin and Teton regions, and had the stron- gest positive relationship with willow height. If tall willows are not available, songbird spe- cies that nest high in willows may be absent from these ecological communities. Species that are more generalist in habitat selection and tend to nest lower in willows (e.g., Amer- ican Robin, White-crowned Sparrow, Song Sparrow [Melospiza melodia]) had fewer dif- ferences in abundance levels across the Gal- latin and Teton regions. These species choose between a number of different habitats to for- age and nest, and their patterns of presence in riparian willow habitat may be more haphaz- ard than a willow specialist, such as Willow Flycatcher. CONSERVATION IMPLICATIONS Willows in the Teton region were taller and less dense than those in the Gallatin region. This difference in willow height and density had strong correlations with differences in songbird abundance and diversity patterns. The most reasonable mechanism causing this difference has been high levels of herbivory by elk in the northern part of the system, es- pecially during winter when food is scarce. A number of other studies have examined how populations of ungulates can cause woody vegetation in the Rocky Mountain region of the United States to decline (White et al. 1998, Ripple and Larsen 2000, Beschta 2003). Riparian songbird communities are influenced by the structural characteristics of the habitat, and this structure could differentially affect avian species based on their specific life his- tory characteristics. Our findings suggest that songbirds that nest in the uppermost portions of tall willows may be particularly affected by elk herbivory. ACKNOWLEDGMENTS We thank the University of Wyoming, National Park Service Research Center (AMK Ranch; H. E Harlow, Director) for funding and accommodating our research team. Data collection was funded by a grant from the Environmental Protection Agency (EPA) through their Ecological Assessment and Restoration program. Prod- ucts of this grant, although funded by the EPA (through grant 96-NCERQA-lA Debinski et al.) have not been subjected to the Agency's peer re\ iew and dti not necessarily reHect the views of the Agency and no official endorsement should be inferred. .Statistical consulting was provided by D. H. Cook, (iiaphics as- sistance was provided by .1. W. Morphew. We thank the many field technicians who helped over the years. UTERATURF Cl H:D Ammon, A. M. and P. B. .Siacia. 1997. A\ian nest success in relation to past grazing regimes in a montane riparian system. Condor 99:7 13. 838 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 Anderson. E. M. 2007. Changes in bird communities and willow habitats associated with fed elk. Wil- son Journal of Ornithology 119:400 — 109. Barb.aro, L., L. Pontcharraud. F. Vetillard. D. Gu- ^ ON. AND H. J.A.CTEL. 2005. Comparative responses of bird, carabid. and spider assemblages to stand and landscape diversity in maritime pine planta- tion forests. Ecoscience 12:110-121. Berger. J.. P. B. Stacey. L. B. Bellis. and M. P. John- son. 2001. A mammalian predator-prey imbal- ance: grizzly bear and wolf extinction affect avian neotropical migrants. Ecological Applications 1 1 : 947-960. Beschta. R. L. 2003. Cottonwoods, elk, and wolves in the Lamar Valley of Yellowstone National Park. Ecological Applications 13:1295-1309. Be^'ER. H. L. 2007. W illow on Yellowstone's northern range: evidence for a trophic cascade? Ecological Applications 17:1563-1571. Cody. M. L. 1974. Competition and structure of bird communities. Princeton University Press. Prince- ton, New Jersey. USA. Collins. T. C. 1976. Population characteristics and habitat relationships of beavers. Castor canaden- sis. in northwest Wyoming. Dissertation. Univer- sity of Wyoming. Laramie. USA. Debinski. D. M., M. E. Jakubauskas. and K. Kindscher. 2000. Montane meadows as indicators of environmental change. Environmental Moni- toring and Assessment 64:213-225. Debinski. D. M.. C. Ray. and E. H. Saveraid. 2001. Species diversity and the scale of the landscape mosaic: do scales of movement and patch size af- fect diversity? Biological Conservation 98:179- 190. Dobkin. D. S., a. C. Rich, and W. H. Pyle. 1998. Habitat and avifaunal recovery from livestock grazing in a riparian meadow system of the north- western Great Basin. Conservation Biology 12: 209-221. Gill. F. B. 1995. Ornithology. Second Edition. W. H. Freeman and Company. New York. USA. Gribb. W. J. 2004. Development of a potential beaver habitat model for the Snake River Basin. Grand Teton National Park. Wyoming. Pages 31 — lO in 28th Annual Report, UW-NPS Research Station. Jackson Lake. Wyoming. USA. Insightfll Corp. 2005. S-Plus Version 7.0. Seattle. Washington. USA. Jackson. S. G. 1992. Relationships among birds, wil- lows. and native ungulates in and around northern Yellowstone National Park. Thesis. Utah State University. Logan. USA. JoN.AS. R. J. 1955. A population and ecological study of the beaver {Castor canadensis) in Yellowstone National Park. Thesis. Uni\ ersity of Idaho. Mos- cow. USA. Kay. C. E. 1990. Yellowstone's northern elk herd: a critical evaluation of the “natural regulation" par- adigm. Dissertation. Utah State University, Logan. USA. Kay. C. E. and S. Ch.xdde. 1992. Reduction of willow seed production by ungulate browsing in Yellow- stone National Park. Pages 92-99 in Proceedings: symposium on ecology and management of ripar- ian shrub communities. 29-31 May. 1991 (W. P. Clary, E. D. McArthur, D. Bedunah. and C. L. Wambolt. Editors). General Technical Report INT- 289. USDA, Forest Service. Intermountain Forest and Range Experiment Station. Ogden. Utah. USA. Knopf, F. L., J. A. Sedgwick, .and R. W. Cannon. 1988. Guild structure of riparian avifauna relative to seasonal cattle grazing. Journal of Wildlife Management 52:280-290. MacArthur, R. H. .and J. W. MacArthur. 1961. On bird species diversity. Ecology 42:594—598. McG.arig.al, K. .ANT) B. J. Marks. 1994. FRAG- STATS: spatial pattern analysis program for quan- tifying landscape structure. Version 2.0. Oregon State University. Corvallis. USA. McLaren. B. E. and R. O. Petersen. 1994. Wolves, moose and tree rings on Isle Royale. Science 266: 1555-1558. Medin. D. E. .and W. P. Clary. 1990. Bird populations in and adjacent to a beaver pond ecosystem in Idaho. Research Paper INT-432. USDA, Forest Service, Intermountain Forest and Range Experi- ment Station, Ogden, Utah, USA. N.ALM.AN, R. J., C. A. Johnston, .and J. C. Kelley. 1988. Alteration of North American streams by beaver {Castor canadensis). BioScience 38:753- 762. National Rese.arch ColtsCil (NRC). 2002. Ecologi- cal dynamics on Yellowstone's northern range. National Academy Press. Washington D.C., USA. Pelnetti, H. R., M. a. Kalkh.an. and M. B. Cough- ENOUR. 2002. Long-term changes in w illow spatial distribution on the elk range of Rocky Mountain National Park. USA. Landscape Ecology 17:341- 354. Pollock, M. M., R. J. N.aim.an, H. E. Erickson, C. A. Johnston, J. Paster, .and G. Pinay. 1995. Beaver as engineers: influences of biotic and abiotic char- acteristics of drainage basins. Pages 117-126 in Linking species and ecosystems (C. G. Jones and J. H. Lawton, Editors). Chapman and Hall. New York, USA. Powell. A. F. L. A. 2006. Effects of prescribed burns and bison {Bos bison) grazing on breeding bird abundances in tallgrass prairie. Auk 123:183-197. Ripple, W. J. .and R. L. Beschta. 2004a. Wolves and the ecology of fear: can predation risk structure ecosystems? BioScience 54:755-766. Ripple, W. J. .and R. L. Beschta. 2004b. Wolves, elk, willows, and trophic cascades in the upper Gal- latin Range of southwestern Montana. Forest Ecology and Management 200:161-181. Ripple. W. J. and E. J. L.arsen. 2000. Historic aspen recruitment, elk, and wolves in northern Yellow- stone National Park. Forest Ecology and Manage- ment 184:299-313. Olechnowski and Debinski • SONGBIRD RESPONSE TO WIEEOW HABITAT STRUCTURE 839 Ripple. W. J„ E. J. Larsen, R. A. Renkin, and D. W. Smith. 2001. Trophic cascades among wolves, elk, and aspen on Yellowstone National Park’s north- ern range. Biological Conservation 102:227-234. Robel, R. J., J. N. Briggs, A. D. Dayton, a.nd L. C. Hulbert. 1970. Relationships between visual ob- struction measurements and weight of grassland vegetation. Journal of Range Management 23: 295-297. Romme, W. H., M. G. Turner, L. L. Wallace, and J. S. W.ALKER. 1995. Aspen, elk, and fire in northern Yellowstone National Park. Ecology 76:2097- 2106. S.\LLAB.A.NKS, R., R. A. RiGGS, L. E. COBB, AND S. W. Dodson. 2006. Bird-habitat relationships in grand fir forests of the Blue Mountains, Oregon. Eorest Science 52:489-502. Salt, G. W. 1957. An analysis of avifaunas in the Te- ton Mountains and Jackson Hole, Wyoming. Con- dor 59:373-393. S.AVERAiD, E. H., D. M. Debinski, K. Kindscher, and M. E. JAKUBAUSK.AS. 2001. A comparison of sat- ellite data and landscape variables in predicting bird species occurrences in the Greater Yellow- stone Ecosystem, USA. Landscape Ecology 16: 71-83. Sh.annon, C. E. .\nd W. Weaver. 1962. The mathe- matical theory of communication. University of Illinois Press, Urbana. USA. Singer, E J., L. M.vck, and R. G. Cates. 1994. Un- gulate herbivory of willows on Yellowstone’s northern winter range. Journal of Range Manage- ment 47:435-443. Singer, E J., L. C. Zeigenfuss, R. G. Cates, and D. T. Barnett. 1998. Elk, multiple factors, and per- sistence of willows in national parks. Wildlife So- ciety Bulletin 26:419-428. S.MITH, B. 2001. Winter feeding of elk in western North America. Journal of Wildlife Management 65: 173-190. Smith, B., E. K. Cole, and D. S. Dobkin. 2004. Im- perfect pasture: a century of change at the Na- tional Elk Refuge in Jackson Hole Wyoming. Grand Teton National History Association, Moose, Wyoming, USA. Smith, D. W, R. O. Peterson, .and D. B. Houston. 2003. Yellowstone after wolves. BioScience 53: 330-340. U.S. Dep.artmen"t of Agricu'ltlre (US DA). 2007. L^SDA plants database, http://plants.usda.gov/ (ac- cessed 16 January 2007). W.ARREN, E. R. 1926. A study of beaver in the Yancey region of Yellowstone National Park. Roosevelt Wild Life Annals 1:1-191. White, C. A., C. E. Olmsted, and C. E. Kay. 1998. Aspen, elk. and fire in the Rocky Mountain na- tional parks of North America. Wildlife Society Bulletin 26:449-462. WINGINGTON Jr., P. j. and R. L. Beschta. 2000. Ri- parian ecology and management in multi-land use watersheds. American Water Resources Associa- tion, Middleburg, Virginia, USA. Yamaura. Y, K. K.ATOH, .AND T. Takahashi. 2006. Re- versing habitat loss: deciduous habitat fragmen- tation matters to birds in a larch plantation matrix. Ecography 29:827-834. Z.A.LBA, S. M. ANT) N. C. Cozz.ANi. 2004. The impact of feral horses on grassland bird communities in Argentina. Animal Conservation 7:35 — 14. Zimmerman, J. L. 1992. Density-independent factors affecting the avian diversity of the tallgrass prairie community. Wilson Bulletin 104:85-94. The Wilson Journal of Ornithology 120(4):840— 855, 2008 REGIONAL ANALYSIS OF RIPARIAN BIRD SPECIES RESPONSE TO VEGETATION AND LOCAL HABITAT FEATURES NADAV NUR,'^ GRANT BALLARD,' AND GEOFFREY R. GEUPEL' ABSTRACT. — We investigated relationships between riparian bird abundance and local vegetation character- istics and habitat features across the Sacramento/San Joaquin Valley, California. Number of detections was analyzed for each of 21 species from point count surveys over a 4-year period at 22 sites from three regions (Sacramento River, Cosumnes River, and San Joaquin River) in relation to 16 measures of habitat and vegetation composition within 50 m of 1 84 survey points. Tree variables, including tree height and trunk diameter, were often important, as was specific composition of tree species, especially Fremont cottonwood (Populus fremontii) and valley oak (Quercus lobata). Effects of mugwort {Artemisia douglasiana) and blackberry {Rubus spp.) were generally positive. The median partial due to vegetation/habitat characteristics was 16% after controlling for regional differences in abundance per species. Comparisons of model results at the local versus regional scale revealed spatial variation in bird abundance was independent of spatial variation in habitat variables. The effect of a habitat variable differed among the three regions for 1 1 of 16 variables. Models that used one or more of the first three principal components (extracted from the 16 vegetation and habitat variables) had substantially lower predictive ability than models built using individual variables. The results emphasize the importance of both understory vegetation and tree characteristics at different spatial scales. Local vegetation and habitat char- acteristics are important in explaining variation in local abundance, but there is a need to develop models specific to each subregion. Received 28 July 2006. Accepted 1 March 2008. Riparian habitat in the western United States is one of the most productive and valu- able habitats for all wildlife, especially in Cal- ifornia (Knopf 1985, Rich 2002, Faber 2003). It is also one of the most threatened habitats, with only 5% of California’s original riparian habitat remaining (Katibah 1984, Abell 1989). Riparian restoration and management of ri- parian habitat have become management pri- orities for agencies and non-governmental or- ganizations throughout California (RHJV 2004). Successful management and conservation of birds using riparian habitat requires infor- mation on how birds respond to habitat char- acteristics, including changes resulting from habitat degradation or restoration. Specific in- formation is needed regarding elements or at- tributes of habitats used by avian species. Ear- lier studies on bird-habitat associations em- phasized general structural characteristics of vegetation (Lack 1933, Hilden 1965, Wiens 1969, Willson 1974, Cody 1985), but more recent studies have identified the importance of specific tree species for riparian-dependent birds (Strong and Bock 1990, Saab 1999). This is consistent with findings of Wiens and ' PRBO Conservation Science, 3820 Cypress Drive, #11, Petaluma, CA 94954, USA. ^ Corresponding author; e-mail: nnur@prbo.org Rotenberry (1981) that shrub-steppe birds re- spond more strongly to specific shrub species than to general vegetation and habitat struc- ture. Riparian vegetation is composed of ground cover, herbs and forbs, shrubs, and trees; few studies have examined the impor- tance of all components (Heath and Ballard 2003). We also do not know the spatial scale at which birds respond. Previous riparian studies have either had broad spatial coverage (e.g., Tewksbury et al. 2002) or have examined fac- tors influencing abundance at the local, terri- tory-level scale (e.g.. Strong and Bock 1990), but have rarely combined both components. Saab (1999) concluded that regional-scale fac- tors were more important than local habitat factors in explaining variation in riparian bird abundance, but this conclusion has not been supported in other studies (Scott et al. 2003). Little information is available as to whether bird-habitat correlations at a site or cluster of sites apply across larger spatial scales (e.g., across different watersheds or regions). This limits our ability to successfully generalize management recommendations based on one local study to that of other populations. Riparian bird studies have often examined community-wide metrics (i.e., species diver- sity or richness), but relatively few have an- alyzed species-specific patterns of abundance 840 Nur et al. • SONGBIRD ABUNDANCE AND RIPARIAN VEGETATION 841 for a wide variety of species. Bird species di- versity is a useful metric (MacArthur and MacArthur 1961), but does not indicate how individual species respond to specific habitat features, which may vary by species (Howell et al. 2000). Studies focused on a few “focal” species can be informative, but raise the ques- tion of how representative findings or predic- tions are for the entire community (Chase et al. 2000). We investigated relationships between the abundance of 21 landbird species (20 passer- ines and 1 near-passerine) and local habitat or vegetation characteristics at 22 sites spanning 400 km of the Central Valley of California, including the Sacramento, San Joaquin, and Cosumnes rivers. All sites were classified as “valley-foothills riparian habitat” (Mayer and Laudenslayer 1988). Specific objectives were to develop and evaluate statistical predictive models for each of the 21 species to: (1) iden- tify which vegetation or habitat variables (hereafter habitat variables) explain variation in abundance in each species at the local scale, (2) characterize commonalities and differenc- es among bird species in the variables of in- fluence and direction of their influence, (3) ex- amine the consistency of effects of these var- iables for each species across three different regions of the Central Valley, (4) quantify the magnitude of variation in abundance of indi- vidual species explained by variation in hab- itat characteristics at the local scale, and (5) compare predictive abilities of models based on principal components with models based on the individual habitat variables. METHODS Study Sites. — There were three study areas: along the Middle Sacramento River, Cosum- nes River, and San Joaquin River, all within the Sacramento/San Joaquin River watershed (Fig. 1). Studies in the Sacramento River re- gion were conducted at 10 sites, spanning 160 km between Colusa and Red Bluff (n = 84 point count stations). Data were collected at six sites in the Cosumnes River region span- ning — 15 km (n = 44 point count stations). Data in the San Joaquin River region were collected at six sites within the wSan Luis Na- tional Wildlife Refuge spanning —20 km (// = 54 point count stations). Field Methods. — Point count surveys (Ralph et al. 1993, 1995) were conducted three times per breeding season during May and early June. Surveys were conducted in 4 years, 1995-1998 for Cosumnes and Sacra- mento river sites, and 1995-1997 for San Joa- quin sites. All species detected within 50 m of the observer were recorded over a 5 -min period at each point count station. Surveys were conducted beginning 15 min after sun- rise and concluding within 4 hrs of sunrise. Surveys were not conducted during rainy or excessively windy conditions. Stations were 200-250 m apart and in all habitats within a site; only those in riparian habitat (vs. those in crops, grasslands, orchards, etc.) were in- cluded in this study. Fine-scale vegetation and habitat assess- ment were conducted at all point count sta- tions at least once during 1995-1997. We used the most recent year of data if vegetation was surveyed in more than 1 year; there was no assessment of change in vegetation during the survey period. We used a modified version of the releve (Estimation of Stand Characteris- tics) method of vegetation assessment follow- ing Ralph et al. (1993). Vegetation at each point count station was assessed using a 50- m radius (0.785 ha) plot centered on the point count survey station. General characteristics of the plot were recorded (including maxi- mum tree dbh, tree height, presence of water, etc.), and cover, abundance, and height of each vegetation stratum (tree, shrub, herb) were estimated. These were defined following Ralph et al. (1993) with respect to height from ground, not botanically. The tree layer was >5 m, the shrub layer was 0.5 to 5 m, and the herb layer was <50 cm. Species composition and species richness of trees and shrubs were ascertained within each vegetation stratum. Percent absolute cover was recorded for each species. All willows {Scdix spp.) were pooled for analysis. A given species could be record- ed either as shrubs or trees depending on height. Presence of water was defined with re- spect to open water within the 50-m radius (1 if water present within 50 m, 0 otherwise). Selection of Habitat Variables. — We con- sidered 16 habitat variables, with region as an additional analytical variable (Table 1 ). We chose variables that we considered potentially biologically relevant and/or of management interest (Heath and Ballard 2003). A principal 842 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 "A /’■ -o o N 25 50 HA \ 100 I Kilometers LIG. 1. Study area, showing 3 regions (Sacramento, Cosumnes, and San Joaquin rivers). Sites within region with four letter abbreviations, from north to south: RYAN = Ryan; LABA = Labaranca; OHM = Ohm; HALE = Haleakala; LLYN = Llynn; KOSL = Kopta Slough; RIVI = River Vista; STCR = Stony Creek; SUNO = Sul Norte; CODO = Codora; VALE = Valensin; TAEO = Tall Lorest; WERO = Wendell’s Road; WELL = Wendell’s Levee; WISL = Wilson’s Slough; DWRE = DWR East; WBCR = West Bear Creek; NOSA = North San Joaquin; MISA = Middle San Joaquin; SASL = Salt Slough; ELKP = Elk Pasture; and SOSA = South San Joaquin. criterion for selection of vegetation variables was adequacy of representation in at least two of the three regions. Shrub, tree, or herb spe- cies prevalent in only one region were not in- cluded, nor were plant species that were un- common. Preliminary analyses indicated that transformations of maximum tree height and dbh provided better predictors of bird abun- dance than untransformed values for each var- iable. We used tree height index (Table 1) and the natural log of maximum tree dbh for anal- yses. The transformation allowed for non-lin- ear, monotonic effects of tree height or max- imum tree dbh on abundance. Nur et al. • SONGBIRD ABUNDANCE AND RIPARIAN VEGETATION 843 TABLE 1. Principal component analysis {n = 178) for 16 vegetation and habitat variables. Loadings for the first five principal components. Variables with loadings of absolute value >0.35 indicated by*. Eigenvectors Variable Mean I II III IV V Water (Yes = 1 ; No = 0) 0.55 0.046 0.285 0.445* -0.166 0.109 Tree cover (%) 35.1 0.318 -0.317 0.075 -0.217 -0.264 Shrub cover (%) 24.5 0.289 0.288 -0.127 0.395* -0.108 Herb cover (%) 34.1 -0.383* 0.116 0.029 -0.162 0.290 Tree species richness 3.42 0.357* 0.177 -0.109 0.097 0.079 Shrub species richness 4.31 0.305 0.266 -0.172 0.044 0.168 Tree dbh index"* 3.68 0.287 -0.124 0.353* -0.081 0.242 Tree height index** 3.20 0.340 -0.269 -0.056 -0.029 0.243 Willow shrub cover (%) 7.88 -0.012 0.165 0.362* 0.478* 0.007 Willow tree cover (%) 12.7 -0.047 0.014 0.537* -0.011 -0.354* Cottonwood tree cover (%) 4.23 0.116 -0.470* -0.193 -0.015 -0.178 Valley oak shrub cover (%) 0.84 -0.106 0.2556 -0.329 0.175 -0.381* Valley oak tree cover (%) 6.13 0.225 0.264 -0.073 -0.401* 0.226 Blackberry cover (%) 3.12 0.269 0.302 -0.053 -0.221 -0.089 Mugwort cover (%) 2.40 0.008 -0.228 0.003 0.471* 0.498* Grass cover (%) 15.0 -0.315 0.074 -0.192 -0.181 0.246 ^ Index = InCmaximum tree dbh + 1 ); maximum tree dbh measured in cm, 0 if no trees present. Scored 0 if no trees; otherwise 1, 2, 3, 4, or 5, classifying height into groups corresponding to: <9, 9 to <13, 13 to <18, 18 to <25, and >25 m. Distinguishing the potential effect of Cali- fornia blackberry (Rubiis ursinus) from Hi- malayan blackberry {R. discolor) on bird spe- cies was of management interest (native vs. non-native invasive), but California blackber- ry was either uncommon or absent from most sites. Thus, we analyzed only the summed blackberry cover. Selection of Avian Species for Analysis. — We focused on landbird species known to breed in the study region (presence throughout the breeding season, mist net captures, and nest-monitoring at nearby sites). We selected the 20 most abundant species based on detec- tions at point count stations in the riparian sites. Four of the 20 species were identified as focal species by the Riparian Habitat Joint Venture (RHJV 2004). We also included as many additional RHJV-designated riparian fo- cal species as possible with adequate sample size. Only Blue Grosbeaks (scientihc names for bird species analyzed are in the Appen- dix), among species not in the top 20, met the requirement of at least 100 detections. Bank Swallows (Riparia riparia) were detected in similar numbers to Blue Grosbeaks but the species is colonial and not reliably surveyed using point counts (Ralph ct al. 1993, 1995). Thus, this species was not included. Twenty- one species were analyzed: 20 passerines spe- cies and Nuttall’s Woodpecker (Appendix). Treatment of Dependent Variables and Principal Components. — We calculated an abundance index for species by summing the number of detections per species over the three surveys in a year. We then analyzed ln(mean detections + constant), where the mean was calculated over all years and the constant was the smallest non-zero abundance value (Nur et al. 1999). Log-transformation of mean total detections per species was used be- cause it yielded better predictive relationships with independent variables than use of un- transformed mean detections and served to better normalize residuals. Residuals of the re- sultant models in all cases but one were ap- proximately normally distributed {sktest pro- cedure and examination of quantile-quantile plots for a normal distribution) (Nur et al. 1999, StataCorp. 2003). The exception was Bullock’s Oriole, but log transformation was still preferred to no transformation. Analysis of log-transformed data is preferred if birds respond in a multiplicative manner to habitat factors, which is reasonable (i.e., with every unit increase in the independent variable, abundance increases by a fixed proportion or decreases by the same proportion) (Nur et al. 1999). 844 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 Mean values analyzed were based on a dif- ferent number of years per point count station. Analyses were weighted by the square-root of the number of years of data (mostly 4 yrs/ point, but in some cases 2 or 3 yrs) because the standard error of the mean detection per point is inversely proportional to the square- root of the number of years of data (Neter et al. 1990). Abundance values derived from fewer years of data contributed less to model results than those with more years of data. We conducted a principal component analysis (PCA) of the habitat variables to examine how well the 16 variables were represented by a smaller number of principal components (Le- gendre and Legendre 1998, Lichstein et al. 2002). Model Selectioji and Statistical Analysis. — We used backward variable selection, consid- ering each species one at a time, following Catchpole et al. (2004:15) for identifying the best statistical model. We started with 16 in- dependent, quantitative variables for each spe- cies, plus region, treated as a categorical var- iable in the model (Sacramento, Cosumnes, or San Joaquin rivers). We used a two-stage model selection process (Zuur et al. 2007); first using a backward stepwise-procedure to choose the model with lowest Akaike Infor- mation Criterion (AIC) value. We then used a stepwise procedure to drop, one at a time, any additional variables from the AlC-preferred model that were not significant at the P < 0.10 level, using an F-test. This second stage was necessary because AIC is often over inclusive, selecting models that include nonsignificant variables, even at the P < 0.1 level. Model selection using AIC, especially with large sample sizes, leads to inclusion of extraneous (ecologically unrelated) variables (Brooks et al. 2000, Hansen and Yu 2001, Davison 2003: 404). We considered an alternative approach, the Bayes Information Criterion (BIC; Schwarz 1978), which has been criticized for being too stringent (i.e., setting thresholds too high for including additional variables; Hansen and Yu 2001). The approach we used was intermedi- ate between AIC alone and BIC. We evaluated the consistency of the effect of each variable identified in the optimal sta- tistical model for each species by testing for an interaction of that variable with region. while including other identified variables in the statistical model. We did not evaluate in- teractions of variables not included in the final species-specific model. We evaluated the sig- nificance of the specified interaction while in- cluding the second (or third) interaction in the model where more than one interaction was identified for a species at P < 0.1. We com- pared total R- and partial R- (Neter et al. 1990) for models that explain variation in abundance in relation to: (1) habitat variables only, (2) regional differences only (region treated as a categorical variable), and (3) both sets of var- iables, considered simultaneously. The first three principal components from the habitat variable PCA were used to statis- tically model the response of individual spe- cies’ abundance. We compared models using the 16 habitat variables with models using the first three principal components with respect to P2 and partial P". We consider P > 0.1 to indicate lack of statistical significance and P < 0.05 to indicate statistical significance. We consider the results to be inconclusive where 0.05 < P < 0.10, and report the P value. RESULTS Habitat Variables.— The overall 16-dimen- sional variation among independent variables was not well captured by a small number of easily interpretable principal components (Ta- ble 1). Sixty-one percent of the variance of the independent variables was accounted for by the first three principal components (pro- portion of the variance explained by each of the first 5 components was 0.301, 0.158, 0.151, 0.081, and 0.064, respectively). The first component contrasted tree species richness with herb cover as those two vari- ables loaded most strongly but in opposite di- rections. Thus, a survey station scored high on the first principal component if tree species richness was high and herb cover was low. The second component reflected Fremont cot- tonwood {Populus fremontii) tree cover more than any other variable. The third component loaded most strongly on willow tree cover, willow shrub cover, presence of water, and tree dbh index. The fourth component con- trasted mugwort {Artemisia douglasiana) and willow shrub cover (both positive) with valley oak (Quercus lobata) tree cover (negative). The fifth component contrasted mugwort cov- Nur et al. • SONGBIRD ABUNDANCE AND RIPARIAN VEGETATION 845 er (positive) with valley oak shrub cover and willow tree cover. Tree height index, tree cov- er, shrub species richness, and blackberry cov- er did not load strongly on any of the five components. The first three principal compo- nents were used in statistical analyses of pre- dictors of species-specific abundance. Species-specific Abundance Models in Re- lation to Vegetation and Region. — Number of detections for each species varied more than 10-fold with Spotted Towhee having the most detections and Blue Grosbeak the least (Ap- pendix). Region was included in 17 of 21 spe- cies-specific preferred models, demonstrating variation in abundance due to region that could not be accounted for by differences due to habitat variables (Table 2). Regional differ- ences were not apparent for House Finch, Tree Swallow, Nuttall’s Woodpecker, or Bullock’s Oriole. At least one habitat variable was re- tained for all species. Seventeen of the 21 spe- cies’ models included at least three habitat variables and six species’ models included at least six habitat variables (Brown-headed Cowbird, Red-winged Blackbird, House Wren, American Robin, Common Yellow- throat, and Western Wood-Pewee). Overall R^ varied from 21 to 58% for most (17 of 21) species abundance models, but for two species overall R^ was less than 20% (Bullock’s Ori- ole and House Finch) and for two species overall R^ was greater than 70% (Song Spar- row, Western Wood-Pewee). Fifteen of the 16 habitat variables were in- cluded in at least one species model (Table 2). Grass cover was not included, either positive- ly or negatively, for any species. Tree species richness was only included in models for two species (Spotted Towhee, negative; Bushtit, positive). Willow tree cover, was included only for House Wren. In contrast, willow shrub cover (i.e., which includes trees <5 m in height) was included in six species’ models. Thirteen of the 16 variables were included in models for at least four species. The three variables included in the most species models were herb cover (1 1 species), presence of water (9 species), and tree dbh index (8 species). Herb cover was often in- cluded, yet grass cover was not included. These two variables were the only two vari- ables that tracked the herb stratum. Variables reflecting tree cover or tree size were often included in the models. These included spe- cies-specific tree variables as well as tree dbh index (8 species) and tree height index (7 spe- cies). Tree dbh index and tree height index were not included in the same model. The ef- fect of tree size could be captured by one of the two indices, although which index de- pended on the species. Most habitat variables had both positive and negative effects, depending on bird spe- cies. The two exceptions were blackberry and mugwort, for which only positive effects were demonstrated (5 and 4 species, respectively). Variability in the Effect of Habitat and Veg- etation Across Regions. — The majority of var- iables, when examined species by species, did not demonstrate heterogeneity of slopes among regions (Table 3). However, most var- iables (11 of 15 examined) demonstrated a significant interaction (P < 0.05) with region for at least one species. Four variables (tree cover, shrub cover, herb cover, and tree dbh index) had significant interactions for two bird species and none had significant interactions for more than two species. Slopes differed quantitatively, not qualita- tively, in eight cases (Table 3). In these in- stances, slopes for the variable in question were in the same direction (all positive or all negative) when examined region by region, but because the region-specific coefficients were sufficiently dissimilar, an interaction was detected (coded “1” in Table 3). The evidence in six cases indicated the effect of a variable was confined to a single region (coded “2” in Table 3) with no effect apparent in the other two regions. There were four cases where the estimated slopes differed quantitatively and qualitatively among regions (coded “3” in Ta- ble 3). Marked contrasts were observed for the effect of tree dbh index for Common Yellow- throats (significantly positive for Sacramento sites, but significantly negative slope for Co- SLimnes sites), and for cottonwood tree cover for Western Kingbirds (significantly positive for Sacramento sites, but significantly nega- tive slope for San Joaquin sites). House Wrens had a significant positive effect for mugwort cover at Sacramento sites but an apparently negative effect (P = 0.085) at San Joaquin sites; Blue Grosbeaks had a significant nega- tive effect for valley oak tree cover at Co- TAHI.B 2. Spccies-specilic s.aCislical models for abundance of 21 species in rclalion lo 16 hahilal/vcgelalion variables in Ihe Ceniral Valley of California, 846 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 2 I 0- Z Z > 4 OK >■ + * * o _ r', 2u z z P r- -1- * * * * * -t- * * * * 1 § IT-, z z 2- z z > (N + + + + + + + + * — + + 2k + 2- * z z ^ + * + * + + * * * + * -1- * ^ + v: * * * + * + + + O oc oc N N z Z 2- 2. A >- ^ ■S’ + * * + * * * + * r^, Z + z Z — + + * * * + -H * * * + + * * * + + y ^ sC z 2- z 2- z - >- + * + Z 2- Z 2- 2- * + + Z 2k 2- z z re 'i' 3 g- ' € a I 18^:=:- ^ P ^ ^ K. 3 is il ; I > o c > ■y y vC >■ (N > ■ ir-, II 5 5 = ^ i U 'j SXj :? ;> a I i 5 5 — ''i CXj — O -N 2i 2C TABLE 2. Continued. Nur et al. • SONGBIRD ABUNDANCE AND RIPARIAN VEGETATION 847 — (NO O — -^OOO— 'CN(^iOOO (NCNrJin-^ — (N0-^0(N000 ONor-- — cN'(j-oor'0 ^ in ^ o * * Z Z t + Z Cl >" r- + + + + + + O * * + + + + * * + c/5 CTN o u * + + + + * + A* (N * z CL CL Cl CL Z Cl o + + d OQ N + + + V < * + + + y O .J * + + + - in lO d + + y + + V tu + + O '■5 + + + + in Cl a. Cl Cl >“ (N + * * Z Z r- ■3 z ^ >- (N £ ^ O '' o c “ X ~ y. V C (u o •- III y. 1) m -D O % S £ L> >, u. \- OJ s: L» O LI y: 5IJ XH75f-h->:^0>>cc Clj — S O c^ ^ Oil E 1) > relers to /’ • O.Ol; Mticl (Moderate) refers to 0.01 < f* < 0.10. 848 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 * * * * * V ^ ^ i r- CJ V: ■- 'i ^ ^ ^ O > ^ I 9, ± O = -E U ^ p c/5 I H o > 4J O >> • — > r; C '■1 i; ^ = J I - P P u > > m 2 O V ^ ^ i Niir et al. • SONGBIRD ABUNDANCE AND RIPARIAN VEGETATION 849 TABLE 4. Coefficients of determination and partial R- due to habitat variables only and to regional main effect only. No significant region main effects indicated by*. Species /?2 Full model R- Model with habitat variables; no region R- Model with region only Partial R- due to habitat variable, for full model Black-headed Grosbeak 0.574 0.400 0.413 0.161 Blue Grosbeak 0.257 0.009 0.15 0.107 Spotted Towhee 0.448 0.354 0.214 0.234 Ash-throated Flycatcher 0.296 0.038 0.143 0.153 Brown-headed Cowbird 0.264 0.194 0.168 0.096 Red-winged Blackbird 0.576 0.415 0.259 0.317 House Finch* 0.184 0.184 0 0.184 House Wren 0.494 0.366 0.179 0.315 Tree Swallow* 0.291 0.291 0 0.291 American Goldfinch 0.209 0.157 0.130 0.079 Song Sparrow 0.712 0.217 0.687 0.025 Bushtit 0.377 0.252 0.301 0.076 Western Kingbird 0.457 0.29 0.348 0.109 Bewick’s Wren 0.44 0.423 0.036 0.404 American Robin 0.291 0.182 0.117 0.174 Nuttall’s Woodpecker* 0.432 0.432 0 0.432 Bullock’s Oriole* 0.015 0.015 0 0.015 Western Scrub-jay 0.252 0.205 0.128 0.124 Lazuli Bunting 0.552 0.354 0.454 0.098 Common Yellowthroat 0.329 0.086 0.108 0.221 Western Wood-Pewee 0.771 0.214 0.531 0.240 sumnes sites, but an apparently positive effect at San Joaquin sites {P = 0.060). Variance in Bird Abundance Explained by Habitat Variables, Compared to Region. — We compared the magnitude of variance in abun- dance explained by habitat variables with that explained by region main effects for each of the 21 species, (Table 4). The 7?^ in 13 cases was 20% or more and in 7 cases was 35% or more for habitat-only statistical models, ex- cluding region main effects, (Table 4). The amount of variance explained by region dif- fered even more among species than variance explained by habitat variables. There were four species for which region main effects, in the absence of habitat variables, accounted for 35% or more of the variance in species abun- dance: Song Sparrow, Western Wood-Pewee, Lazuli Bunting, and Black-headed Grosbeak. However, in two of the four cases (Lazuli Bunting and Black-headed Grosbeak), region- al differences were reduced once habitat dif- ferences were controlled. Regional differences were paramount for Western Wood-Pewee and Song Sparrow, and consideration of habitat variables did not alter this finding. The effects of region were negligible for four species. The most anomalous species was Bullock’s Oriole for which a habitat model could only explain 1.5% of the variance and region was not in- cluded. The median partial due to habitat vari- ables was 16%; for four species the partial R- due to habitat variables was over 30% (Red- winged Blackbird, Bewick’s Wren, House Wren, Nuttall’s Woodpecker). The median partial R^ due to region effects was 12.5%; the partial R^ due to region was over 30% for Song Sparrow and Western Wood-Pewee. Species-specific Models Using Principal Compojients of Vegetation and Habitat Fea- tures.— We compared the performance of models for the 21 species using the first three principal components (extracted from the 16 habitat variables; Table 1 ) with models using individual habitat variables (Table 2). At least one of the first three principal components was at least moderately associated {P < 0.10) with the abundance index in a model that in- cluded region main effects for each species with the exception of Brown-headed Cowbird. Song Sparrow, and Commo?i Yellowthroat (Table 5). The P-values for those three species as.sociated with the best-fitting principal com- 850 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 TABLE 5. Coefficients of determination for models with region and first three principal components of habitat variables. Results are shown for models with retained principal components^ Partial after controlling for region, is shown for principal component model and for individual habitat variables. Species R- for model w ith region and principal component(s)® Partial R- due to principal component(s)^ Partial R- due to individual habitat variables Principal components retained^' Black-headed Grosbeak 0.519 0.106 0.161 I Blue Grosbeak 0.186 0.036 0.107 I Spotted Towhee 0.382 0.168 0.234 I Ash-throated Flycatcher 0.237 0.094 0.153 I Brown-headed Cowbird 0.170 0.002 0.096 (II) Red-winged Blackbird 0.531 0.272 0.317 I, III House Finch* 0.056 0.056 0.184 I House Wren 0.343 0.164 0.315 L II Tree Swallow* 0.126 0.126 0.291 I, IL III American Goldfinch 0.145 0.015 0.079 II Song Sparrow 0.700 0.013 0.025 (II) Bushtit 0.359 0.058 0.076 I Western Kingbird 0.376 0.028 0.109 I Bewick’s Wren 0.348 0.312 0.404 I, III American Robin 0.171 0.054 0.174 I Nuttall’s Woodpecker* 0.378 0.378 0.432 I, II Bullock’s Oriole* 0.005 0.005 0.015 III Western Scrub-jay 0.157 0.029 0.124 I Lazuli Bunting 0.495 0.041 0.098 I Common Yellowthroat 0.116 0.008 0.221 (I) Western Wood-Pewee 0.752 0.221 0.240 I, II, III a Principal component retained if P < 0.1 in model with region; or if no component P < 0.1, then the one with lowest P value, b Model includes region main effects, except for asterisked species, c Principal component I. II, and/or III retained; shown in parentheses if P > O.I. ponent was P > 0.2 in each case. PC I (Table 1) was included in the hnal principal-compo- nent model for sixteen species, PC II was in- cluded for five species, and PC III was in- cluded for hve species. The partial R- due to individual habitat variables, after controlling for region main effects, was greater, and usu- ally much greater, compared to the partial R- due to the retained principal components (Ta- ble 5). Partial R^ due to habitat variables for eight species was at least three times that of partial R- due to the principal components; partial R^ due to habitat variables for seven species was between 50 and 150% greater than that due to principal components. DISCUSSION Importance of Habitat Differences at the Local Scale. — Variation in abundance of ri- parian-associated songbird species in the Cen- tral Valley of California was strongly associ- ated with local-scale vegetation characteristics for each species examined, except one. Only for Bullock’s Oriole was abundance not ex- plained well by any habitat variable. Variables in the statistical models included both species- specific variables (e.g., cottonwood cover or mugwort cover) as well as more general var- iables (shrub cover or tree height index). Wiens and Rotenberry (1981) reported that vegetation composition was more predictive of bird abundance in shrub-steppe habitat than were measures of habitat structure (physiog- nomy). Our findings emphasize the importance of including species-specific vegetation compo- sition in analyses of habitat features influenc- ing bird species presence or abundance. There is extensive literature reporting on habitat-as- sociations for terrestrial bird species (Vemer et al. 1986, McCullough and Barrett 1992, Scott et al. 2002), but most studies have an- alyzed general habitat characteristics and not species-specific vegetation composition, as exemplified by Block et al. ( 1986), Larson and Bock (1986), Sanders and Edge (1998), Sal- labanks et al. (2000), Vernier et al. (2002), and Miller et al. (2004). Nur et al. • SONGBIRD ABUNDANCE AND RIPARIAN VEGETATION 851 The omission of potentially important veg- etative characteristics has implications for studies that contrast the importance of local habitat features with that of landscape-scale features in predicting species-specific patterns of avian abundance or presence. Lichstein et al. (2002) and Miller et al. (2004) summarize recent studies, which have generally been mixed: some indicate that local scale is more important than landscape scale and others have come to the opposite conclusion. The problem of insufficient characterization of vegetation may explain the different conclu- sions reached by Saab (1999) and Scott et al. (2003), both of which focused on riparian songbirds in the Idaho/Montana region. The former concluded that landscape variables were more important than local variables, while the latter concluded that local variables were more important. However, only Scott et al. (2003) included understory vegetation in the analysis. Specific Vegetation Features of Importance to Birds. — The two variables that were most often included with a positive effect on spe- cies-specific abundance were both measures of tree size: tree height index and tree dbh index. We also demonstrated that shrubs and, more specifically, understory vegetation, were predictive of the abundance of the 21 study species. This result is of interest as many stud- ies of habitat associations for riparian song- birds have only examined plant species com- position with regard to trees (e.g.. Strong and Bock 1990, Saab 1999). Robertsen et al. (2002) also identified several songbird species in mixed forest habitat that responded to un- derstory vegetation. Krueper et al. (2003) found that removal of cattle benefited riparian understory, which in turn influenced bird abundance. Regional Variation and Spatial Scale. — We found marked variation in abundance among regions within a single, large watershed (Sac- ramento/San Joaquin Valley). All sites were within a single habitat classification (“valley- foothill riparian”), but a large amount of var- iation remained in bird species abundance at- tributable to region alone, after controlling for differences in vegetation composition. Important differences in the effect of habitat variables were revealed across regions, similar to the findings of Heath and Ballard (2003). In some cases, birds in one region do not ap- pear to respond to the same habitat variable as conspecifics in another region, but in a few cases responses appeared to differ qualitative- ly from one region to another. The causal ba- sis for this heterogeneity is not clear. Site to site variation may reflect the rarity of a plant species; for example, a bird species may be selecting habitat that contains that plant spe- cies if the species is rare, but avoiding the habitat if the plant species is common. Asso- ciations among plant species may also differ by region. For example, at the Cosumnes sites, presence of valley oak trees was asso- ciated with continuous valley oak forest cov- er; valley oak trees in other locations were often a component of a mixed forest type in- cluding cottonwoods, walnuts {Juglans spp.), or black willow {Salix nigra). Thus, the pres- ence of a species (such as valley oak) implied a different set of associated plants, depending on region. The influence of particular habitat variables also depends on spatial scale (Heath and Bal- lard 2003). This scale-dependence was well illustrated by studies in shrub-steppe birds; Rotenberry and Wiens (1980) demonstrated the importance of general habitat-structural variables at a large (continental) scale, but Wiens and Rotenberry (1981) demonstrated the greater importance of vegetational, spe- cies-specific composition at a more local (sin- gle-region) scale. Response of Birds to Multivariate Fac- tors.— Some studies have analyzed the effects of habitat variables using ordination to reduce a rich multivariate data set of independent var- iables to a few factors, for example, by using principal components or non-metric multidi- mensional scaling and analyzing bird response to multivariate factors (Lichstein et al. 2002). Our results raise concern about this practice. The factors or axes so obtained may concisely capture a large portion of variation with re- spect to vegetation and habitat among study sites (Legendre and Legendre 1998), but there is no assurance these axes reflect a coherent set of elements to which birds respond. An example of this problem was evident for Com- mon Yellow throat where the individual vari- able model performed well, but the principal component model did not (partial R- = 0.221 vs. ().()()8). Two of the variables from the in- 852 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 dividual variable model (tree cover and herb cover) for this species loaded moderately strongly on principal component I, but with opposite signs. However, in the individual var- iable model, the coefficients for both variables were positive. Thus, Common Yellow throats were more abundant where tree cover and herb cover were both high, but because the first principal component weighted these two variables with opposite signs, their effects tended to cancel out in the principal compo- nent statistical model. Other studies have also demonstrated stronger correlations between bird abundance and individual variables than with principal component or other multivari- ate factors (Wiens and Rotenberry 1981, Am- buell and Temple 1983). Limitations of the Study. — We did not ana- lyze differences in detection of birds due to habitat differences. We restricted detections to a distance of 50 m to reduce this problem (Ralph et al. 1993), but even within this dis- tance there could be salient differences in de- tection which would be confounded with dif- ferences in abundance. All points surveyed were within a single, broad habitat type and this might be expected to reduce differences in detection rate within our sample. One should be cautious in inferring causa- tion from patterns of association among bird and habitat variables unless one analyzes bird response to a change in habitat characteristics (which w^e have not done). Each variable iden- tified in a species-specific model (Table 2) has statistical support for its inclusion, but we rec- ognize the ubiquity of model uncertainty (Burnham and Anderson 2002). Recently de- veloped data-mining techniques, such as clas- sification and regression trees (Hastie et al. 2001) are likely to be of great heuristic value, especially when one is analyzing 21 different dependent variables. We restricted our analyses to species that were moderately abundant to improve statis- tical power. We may have failed as a result to capture characteristics of uncommon species. For example, some avian components of the ecosystem we investigated have been extir- pated or drastically reduced w ithin human his- tory (e.g.. Yellow-billed Cuckoo [Coccyzus americanus]. Yellow Warbler [Dendroica pe- techia], and Bell’s Vireo [Vireo bellii]), and we were unable to create habitat models for them. It is possible the habitat features they require are drastically reduced or even absent within the region. CONSERVATION IMPLICATIONS Nearly all species studied were shown to respond predictably to local variation in veg- etation and habitat, at the scale of a territory or home range. Obtaining information on veg- etation composition is labor-intensive and re- quires in-the-field data collection; one cannot rely on remotely-sensed data. The second im- plication is that one cannot rely on a few' focal species to capture the diverse response of the full set of species of interest (Chase et al. 2000, Chase and Geupel 2005). Conservation planners and managers must consider a large, diverse set of target species. The final impli- cation is that recommendations and models need to be developed on a region-by-region basis, especially if one is drawing inferences based on patterns of correlation alone. ACKNOWLEDGMENTS Funding was provided by the California Bay-Delta Authority (principally Contract ERP 02-P17), David and Lucile Packard Foundation, The Nature Conser- vancy, ESDI Fish and Wildlife Service, William and Flora Hewlett Foundation, Morgan Stanley Dean Whitter & Co., the National Fish and Wildlife Foun- dation, and the National Science Foundation (DBI- 0542868). We thank numerous PRBO project leaders, field biologists, and intern biologists, especially Anne King, Ryan Degaudio, Tonya Haff, Stacy Small, and Mike Lynes. We thank Mike Eaton, Ramona Swenson. Joe Silveira. Ramon Vega. Tara Zimmerman, Dawit Zeleke, Dennis Woolington, Valerie Calegari, and Becky Waegell, for support. We thank Tom Gardali, Viola Toniolo. Chrissy Howell, and Mike Green for helpful comments and other assistance and especially Greg Golet for support and insightful comments. This is PRBO Contribution Number 1618. LITERATURE CITED Abell, D. L. (Editor). 1989. Proceedings of the Cali- fornia Riparian Systems Conference: protection, management, and restoration for the 1990’s. USD A, Forest Service, General Technical Report PSW-110. Pacific Southwest Forest and Range Experiment Station, Berkeley, California. USA. Ambl-ell, B. .xnd S. a. Temple. 1983. Area-dependent changes in the bird communities and vegetation of southern Wisconsin forests. Ecology 64:1057- 1068. Blcxtk. W. M., L. a. Brennan, ant) R. J. Gltierrez. 1986. The use of guilds and guild-indicator spe- cies for assessing habitat suitability . Pages 109— Niir et al. • SONGBIRD ABUNDANCE AND RIPARIAN VEGETATION 853 113 in Wildlife 2000: modeling habitat relation- ships of terrestrial vertebrates (J. Vemer, M. L. Morrison, and C J. Ralph, Editors). University of Wisconsin Press, Madison, USA. Brooks, S. R, E. A. Catchpole, B. J. T. Morgan, and S. C. Barry. 2000. On the Bayesian analysis of ring-recovery data. Biometrics 56:951-956. Burnham, K. P. and D. R. Anderson. 2002. Model selection and inference: a practical information- theoretic approach. Second Edition. Springer Ver- lag. New York, USA. Catchpole, E. A., Y. Ean, B. J. T. Morgan, T. H. Clutton-Brock, and T. Coulson. 2004. Sexual dimorphism, survival and dispersal in red deer. Journal of Agricultural, Biological, and Environ- mental Statistics 9:1-26. Chase, M. and G. R. Geupel. 2005. The use of avian focal species for conservation planning in Cali- fornia. Pages 130-142 in Bird conservation im- plementation and integration in the Americas: Proceedings of the Third International Partners in Plight Conference (C. J. Ralph and T. D. Rich, Editors). USDA, Porest Service, General Techni- cal Report PSW-GTR-191. Pacific Southwest Por- est and Range Experiment Station, Albany, Cali- fornia, USA. Chase, M. K., W. B. Kristan III, A. J. Lynam, M. V. Price, and J. T. Rotenberry. 2000. Single species as indicators of species richness and composition in California coastal sage scrub birds and small mammals. Conservation Biology 14:474-487. Cody, M. L. 1985. Habitat selection in birds. Academ- ic Press, Orlando, Florida, USA. Davison, A. C. 2003. Statistical models. Cambridge University Press, Cambridge, United Kingdom. Faber, P. (Editor). 2003. Proceedings of the riparian habitat and floodplains conference. Riparian Hab- itat Joint Venture, Sacramento, California, USA. Hansen, M. H. and B. Yu. 2001. Model selection and the principle of minimum description length. Jour- nal of the American Statistical Association 96: 746-774. Hastie, T, R. Tibshirani, and J. Friedman. 2001. The elements of statistical learning: data mining, in- ference, and prediction. Springer Verlag, New York, USA. Heath, S. K. and G. Ballard. 2003. Patterns of breeding songbird diversity and occurrence in ri- parian habitats of the eastern Sierra Nevada. Pro- ceedings of the riparian habitat and floodplains conference (P. Faber, Editor). Riparian Habitat Joint Venture, Sacramento, California, USA. Hilden, O. 1965. Habitat selection in birds. Annales Zoologici Fennici 2:53-75. Howell, C. A., S. C. Latta, T. M. Donovan, P. A. PoRNELUZi, G. R. Parks, and J. Faaborg. 2000. Landscape effects mediate breeding bird abun- dance in midwestern forests. Landscape Ecology 15:547-562. Katibah, E. F. 1984. A brief history of riparian forests in the Central Valley of California. Pages 23-29 in California riparian systems: ecology, conser- vation, and productive management (R. E. Warner and K. M. Hendrix, Editors). University of Cali- fornia Press, Berkeley, USA. Knopf, F. L. 1985. Significance of riparian vegetation to breeding birds across an altitudinal dine. Pages 105-111 in Riparian ecosystems and their man- agement: reconciling conflicting uses (R. R. John- son, C. D. Ziebell, D. R. Patten, P. F. Folliot, and R. H. Hamre, Editors). USDA, Forest Service, General Technical Report RM-120. Rocky Moun- tain Station, Fort Collins, Colorado, USA. Krueper, D., j. Bart, and T. D. Rich. 2003. Response of vegetation and breeding birds to the removal of cattle on the San Pedro River, Arizona (U.S.A.). Conservation Biology 17:607-615. Lack, D. 1933. Habitat selection in birds, with special reference to the effects of afforestation on the Breckland avifauna. Journal of Animal Ecology 2: 239-262. Larson, D. L. and C. E. Bock. 1986. Determining avian habitat preferences by bird-centered vege- tation sampling. Pages 37-43 in Wildlife 2000: modeling habitat relationships of terrestrial ver- tebrates (J. Verner, M. L. Morrison, and C. J. Ralph, Editors). University of Wisconsin Press, Madison, USA. Legendre, P. and L. Legendre. 1998. Numerical ecol- ogy. Second English Edition. Elsevier, Amster- dam, The Netherlands. Lichstein, j. W., T. R. Simons, and K. E. Franzreb. 2002. Landscape effects on breeding songbird abundance in managed forests. Ecological Appli- cations 12:836-857. MacArthur, R. H. and J. W. MacArthur. 1961. On bird species diversity. Ecology 42:594-598. Mayer, K. E. and W. F. Laudenslayer (Editors). 1988. A guide to wildlife habitats of California. California Department of Fish and Game, Sacra- mento, USA. McCullough, D. R. and R. H. Barrett (Editors). 1992. Wildlife 2001: populations. Elsevier Ap- plied Science, New York, USA. Miller, J. R., M. D. Dixon, and M. G. Turner. 2004. Response of avian communities in large-river floodplains to environmental variation at multiple scales. Ecological Applications 14:1394-1410. Neter, j., W. Wasserman, and M. H. Kutner. 1990. Applied linear statistical models: regression, anal- ysis of variance, and experimental designs. Third Edition. Irwin Press, Homewood, Illinois, USA. NUR, N., S. L. Jones, and G. R. Geupel. 1999. Statis- tical guide to data analysis of avian monitoring programs. Biological Technical Publication BTP- R6()01-1999. USDI, Fish and Wildlife Service. Washington, D.C., USA. Ralph, C. J., S. Droege. and J. R. Sauer. 1995. Mon- itoring bird populations by point counts. USDA. Forest Service, General Technical Report PSW- GTR-149. Pacific Southwest Forest and Range Experiment Station, Albany, California, USA. 854 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 Ralph, C. J., G. R. Geupel, R Pyle, T E. Martin, AND D. E DeSante. 1993. Handbook of field methods for monitoring landbirds. USDA, Forest Service, General Technical Report PBSD-GTR- 144. Pacific Southwest Forest and Range Experi- ment Station, Albany, California, USA. Riparian Habitat Joint Venture (RHJV). 2004. Ver- sion 2.0. The riparian bird conservation plan: a strategy for reversing the decline of riparian as- sociated birds in California. Available from http:// www.prbo.org/calpif/plans.html. Rich, T. D. 2002. Using breeding land birds in the assessment of western riparian systems. Wildlife Society Bulletin 30:1128-1139. Robertsen, M. J., S. a. Temple, and J. Coleman. 2002. Predicting the distribution of songbirds in forests of central Wisconsin. Pages 399-410 in Predicting species occurrences (J. M. Scott, P. J. Heglund, M. L. Morrison, J. B. Hauffer, M. G. Raphael, W. A. Wall, and F. B. Sampson, Editors). Island Press, Washington, D.C., USA. Rotenberry, j. T. and J. A. Wiens. 1980. Habitat structure, patchiness, and avian communities in North American steppe vegetation: a multivariate analysis. Ecology 61:1228-1250. Saab, V. A. 1999. The importance of spatial scale to habitat use by breeding birds in cottonwood for- ests: a hierarchical analysis. Ecological Applica- tions 9:135-151. Sallabanks, R., j. R. Walters, and J. A. Collazo. 2000. Breeding bird abundance in bottomland hardwood forests: habitat, edge, and patch size ef- fects. Condor 102:748-758. Sanders, T. A. and W. D. Edge. 1998. Breeding bird community composition in relation to riparian vegetation structure in the western United States. Journal of Wildlife Management 62:461-473. Schwarz, G. 1978. Estimating the dimension of a model. Annals of Statistics 6:461-464. Scott, J. M., P. J. Heglund, M. L. Morrison, J. B. Hauffer, M. G. Raphael, W. A. Wall, and F. B. Sampson (Editors). 2002. Predicting species oc- currences. Island Press, Washington, D.C., USA. Scott, M. L., S. K. Skagen, and M. F. Merigliano. 2003. Relating geomorphic change and grazing to avian communities in riparian forest. Conserva- tion Biology 17:284-296. StataCorp. 2003. Stata statistical software. Release 8.0. Stata Corporation, College Station, Texas, USA. Strong, T. R. and C. E. Bock. 1990. Bird species distribution patterns in riparian habitats in south- eastern Arizona. Condor 92:866-885. Tewksbury, J. J., A. E. Black, N. Nur, V. A. Saab, B. D. Logan, and D. S. Dobkin. 2002. Effects of anthropogenic fragmentation and livestock graz- ing on Western riparian bird communities. Studies in Avian Biology 25:158-202. Verner, j., M. L. Morrison, and C. J. Ralph (Edi- tors). 1986. Wildlife 2000: modeling habitat rela- tionships of terrestrial vertebrates. University of Wisconsin Press, Madison, USA. Vernier, P. R., F. K. A. Schmiegelow, and S. G. Cum- mings. 2002. Modeling bird abundance from for- est inventory data in the boreal mixed-wood for- ests of Canada. Pages 559-571 in Predicting spe- cies occurrences (J. M. Scott, P. J. Heglund, M. L. Morrison, J. B. Hauffer, M. G. Raphael, W. A. Wall, and F. B. Sampson, Editors). Island Press, Washington, D.C., USA. Wiens, J. A. 1969. An approach to the study of eco- logical relationships among grassland birds. Or- nithological Monographs 8:1-93. Wiens, J. A. and J. T. Rotenberry. 1981. Habitat as- sociation and community structure of birds in shrubsteppe environments. Ecological Mono- graphs 51:21-41. Willson, M. 1974. Avian community organization and community structure. Ecology 66:1211-1214. ZuuR, A. E, E. N. Ieno, and G. M. Smith. 2007. An- alyzing ecological data. Springer Science, New York, USA. Nur et al. • SONGBIRD ABUNDANCE AND RIPARIAN VEGETATION 855 APPENDIX. Common and scientific names of bird species analyzed and species code. Species identified by RHJV (2004) as riparian focal species and total detections by species at all point count stations {n — 184) shown. Common name Scientific name Code Riparian focal species Total detections Nuttall’s Woodpecker Picoides nuttallii NUWO 475 Western Kingbird Tyrannus verticalis WEKI 581 Ash-throated Flycatcher Myiarchus cinerascens ATFL 630 Western Wood-Pewee Contopus sordidulus WEWP 335 Western Scrub-jay Aphelocoma californica WESJ 435 Brown-headed Cowbird Molothrus ater BHCO 1 106 Red- winged Blackbird Agelaius phoeniceus RWBL 726 Bullock’s Oriole Icterus bullockii BUOR 416 House Finch Carpodacus cassinii HOFI 572 American Goldfinch Car due Us tristis AMGO 415 Song Sparrow Melodia melospiza SOSP X 757 Spotted Towhee Pipilo maculatus SPTO 1425 Black-headed Grosbeak Pheucticus melanocephalus BHGR X 638 Blue Grosbeak Guiraca caerulea BLGR X 1 1 1 Lazuli Bunting Passerina arnoena LAZB 406 Tree Swallow Tachycineta bicolor TRES X 680 Common Yellowthroat Geothlypis trichas COYE X 403 Bewick’s Wren Thryomanes bewickii BEWR 767 House Wren Troglodytes aedon HOWR 848 Bushtit Psaltriparus minimus BUSH 605 American Robin Turdus migratorius AMRO 208 Short Communications The Wilson Journal of Ornithology 120(4):856-862, 2008 First Description of the Breeding Biology and Natural History of the Ochre-breasted Brush Finch (Atlapetes semirufus) in Venezuela Luis Biancucci’’^ and Thomas E. Martin* ABSTRACT. — We provide the first description of the eggs, breeding biology, and natural history of the Ochre-breasted Brush Finch {Atlapetes semirufus). We found 37 nests over four breeding seasons (2004- 2007) in Yacambu National Park, Venezuela. Nesting activity started in late April and continued until early June suggesting single-brooded behavior despite a typ- ical tropical clutch size of two eggs {x = 1.89) that were laid on consecutive days. Egg mass averaged 3.38 g and 1 1.6% of adult female mass. The incubation and nestling periods averaged 14.9 and 10.5 days, re- spectively. Only females incubated and the percent time they spent incubating did not change between ear- ly and late incubation. Females brooded 42.7% of the time when nestlings were 2 days of age and 20.5% when 9 days of age. Both parents provisioned young at a low rate (3.9 trips/hr) and nestling growth rate {k = 0.45) was also slow. Nest predation rates were rel- atively high with daily mortality rates of 0.058 and 0.067 during incubation and nestling stages, respec- tively. Received 31 January 2007. Accepted 18 Feb- ruary 2008. The Ochre-breasted Brush Finch {Atlapetes semirufus) is restricted to Colombia and Ve- nezuela (Hilty and Brown 1986, Hilty 2003) with an altitudinal range of 1,600-3,500 m in Colombia and 600-2,700 m in Venezuela (Hilty and Brown 1986, Remsen and Graves 1995, Stotz et al. 1996, Hilty 2003). This spe- cies is protected in several national parks in both countries and does not present immediate conservation challenges (lUCN 2007). De- spite being a common bird (Ridgely and Tu- dor 1994), its breeding biology and life his- tory remain largely unknown. The objectives of our paper are to: (1) pro- vide data on previously unknown life history traits, and (2) add to the knowledge of the natural history and breeding biology of the ' USGS, Montana Cooperative Wildlife Research Unit, University of Montana, Missoula, MT 59812, USA. 2 Corresponding author; e-mail: luis.biancucci @gmail.com Ochre-breasted Brush Finch. Data were col- lected by field and video observations at 37 nests in four consecutive breeding seasons from 2004 to 2007. The work was conducted in Yacambu National Park at the northern end of the Andes in Venezuela (09° 42' N, 69° 42' W) at 1,350-2,000 m elevation. The park is in a mountainous area and has primary cloud forest and a small area of secondary forest (Fierro-Calderon and Martin 2007). The latter is where Ochre-breasted Brush Finches occur most often. OBSERVATIONS Nests and Nest Placement. — We searched for nests from the beginning of March until early July each year using two methods: (1) behavioral observation of parents, and (2) sys- tematic search (Martin and Geupel 1993). Systematic search implies a thorough inspec- tion through all suitable nesting habitat. It was usually conducted after seeing a pair exhibit- ing nesting behavior. Ochre-breasted Brush Finches were fairly common in our study site and were easily seen and heard, and typically occurred as pairs. Our study plots ranged in altitude from 1,350 to 2,000 m elevation; however, we only detected this species below 1 ,600 m. Nests were in dis- turbed habitats: trail sides and secondary growth forests in well-lighted, grassy or vine- covered environments. Nests were usually concealed in grasses, vines or bushes, between 0.2 and 3 m above ground. Foraging activity occurred in the same habitat. Breeding Season. — The breeding season starts in March, when the dry season is end- ing, and lasts until July, in about the middle of the rainy season. The earliest nest was found on 21 March and contained two fresh eggs, based on candling. Breeding activity was highest in late April through late May (Fig. 1). Gender Difference and Roles. — Gender of 856 SHORT COMMUNICATIONS 857 “O 0 -I— » 05 FIG. 1. Timing of bi-weekly nest initiation {n = 19) from early March to late June, 2004-2007 for Ochre- breasted Brush Finch at Yacambii National Park, Lara, Venezuela. Ochre-breasted Brush Finches cannot be ac- eurately assigned based on plumage. Roles of males and females during different periods of nesting were identified using behavior, video recording, and data from mist-netting. Pairs were together during nest building, but only one adult actually collected material and eon- structed the nest while the other bird followed and apparently guarded its mate. Only one adult ineubated based on video reeordings, and only females had developed brood patch- es among mist-netted birds. Thus, we assume that only the female builds the nest and in- eubates. Nest Building and Laying. — The time the female used to build a nest and lay eggs varied throughout the breeding season. It usually took ~ 1 week to build a complete nest at the beginning of the season and another week to lay the first egg. A female took only 8 days to build a nest and lay a first egg towards the end of the season, after having lost her pre- vious nest to predation. The nest of Atlapetes semirufus is an open- cup usually built with thick grasses on the out- side, at times complemented with small sticks on the outer edge. The lining consisted of thin grasses and rootlets. The inside color varies between yellow and brown, depending on ma- terials used (Fig. 2). We measured outer diameter (from edge to edge), inner diameter (cup), outer height (ex- terior bottom-to-top) and inner height (bot- tom-to-top of eup) of 29 nests. Nests averaged [± SE] 6.79 ± 0.12 cm in inner diameter, 12.59 ± 0.32 em in outer diameter, 4.94 ± 0.13 cm in inner height, and 9.24 ± 0.32 cm in outer height. Clutch Size and Eggs. — Clutch size was measured only for nests found during nest building or laying stages and was the number that did not ehange on subsequent days. Cluteh size was two eggs in all but two nests and averaged [± SE] 1.89 ± 0.06 eggs (n = 28). Eggs were laid on eonsecutive days and were white with reddish-brown spots (Fig. 2). Spots were usually concentrated at the large end of the eggs, but may be more uniformly distributed; spots of a few eggs were faint and almost lacking. We weighed 24 eggs at 13 nests between day zero and day 3 of the in- cubation period using an ACCULAB portable electronic scale (precision 0.001 g). Fresh mass of eggs averaged 1± SEl 3.38 ± 0.04 g. Adult male and female mass averaged 29.25 ± 0.33 g (// = 32) based on birds captured in mist nets throughout the breeding season. Thus, fresh egg mass represented 1 1 .6% of the adult mass. Incubation. — The incubation period was 858 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 FIG. 2. Nest and eggs of Ochre-breasted Brush Finch in Yacambu National Park. Photograph by A. M. Niklison. measured as the number of days between the last egg laid and the last egg hatched (Briskie and Sealy 1990, Martin 2002). This period was measured only for nests found during the pre-laying or laying period and for which hatching day was observed. Nests were checked every other day, except at stage- changing events, such as start of incubation, hatching, and fledging, when nests were checked daily or twice daily. The mean [± SE] incubation period was 14.86 ± 0.26 days (n = 7). We assessed incubation behavior and nest attentiveness (percent of time adults spend on the nest incubating) by video filming nests for 6-8 hrs starting within 30 min of sunrise (Martin and Ghalambor 1999, Martin et al. 2007). Nests were filmed once during early and once during late incubation when possible. The mean duration of on- and off- bouts in early incubation was 34:17 and 15: 34 (min:sec), respectively, and during late in- cubation was 44:49 and 15:24 (min:sec). Nest attentiveness averaged 69.4% (n = 1) on the third day of incubation, 70.6% (n = 1) in mid- dle incubation (day 6), and 74.4% ± 0.35 (n — 3) during late incubation (last 4 days). Nestling Period. — Hatching was synchro- nous in all nests of the Ochre-breasted Brush Finch. The nestling period length was 10.5 days {n = 2). Nestlings are able to jump from the nest when they feel threatened, at times as early as day 7, when their primary pin feathers are just breaking their sheaths (i.e., the day that feathers of the 8* primary break their sheaths). Brooding behavior (percent of time females spent brooding nestlings) and parental provi- sioning rates were measured using video re- cordings for 6-8 hrs beginning within 30 min of sunrise. Females brooded 42.7% {n = 1) of the time on day 2 of the nestling period and 20.5% (n = 1) of the time on day 9. Both males and females provision the young based on vid- eos showing both adults at the nest simulta- neously. The provisioning rate on day 2 of the nestling stage was 3.66 trips/hr {n = 1) and was similar on day 9 at 4.20 trips/hr {n — 1). We measured tarsus length of nestlings us- ing Mitutoyo digital calipers and body mass SHORT COMMUNICATIONS 859 Nestling age (day) FIG. 3. Ochre-breasted Brush Finch nestling growth of mass and tarsus length with age. Growth rate constant (k) and asymptotic size (A) are indicated in each plot. using ACCULAB digital electronic scales. Nestlings were measured every other day. We estimated growth rate following Remes and Martin (2002) using the logistic growth func- tion to estimate the growth rate constant (k) and asymptotic size (A). The estimated growth rate based on body mass was faster than when based on tarsus length (Fig. 3). Nest Predation. — Predation accounted for all nesting mortality. We calculated daily nest predation rates using the May held method (Mayheld 1975, Hensler and Nichols 1981). Daily predation rates (x ± SE) were ().07() ± 0.016 for the incubation period, 0.068 ± 0.025 for the nestling period, and 0.069 ± 0.013 (// = 37 nests; 378.5 exposure days) for the total nesting period including egg laying. DISCUSSION Ochre-breasted Brush Finches have an ar- ray of breeding features similar to those of congeneric species. Nest placement and shape 860 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 were similar to that of the Pale-headed Brush Finch (A. pallidiceps) (Oppel et al. 2003, 2004) , White-naped Brush Finch (A. albinu- cha) (Cisneros Palacios 2005), and Yellow- striped Brush Finch (A. citrinellus) (Luis Biancucci, pers. obs.). The clutch size was within the range reported for the genus; clutch size for A. albinucha in Mexico was two eggs based on a single nest (Cisneros Palacios 2005) , one to three eggs for A. pallidiceps in Ecuador (Oppel et al. 2003), one egg for A. leucopis (n = 1; Salaman et al. 1998), and two eggs for A. citrinellus in Argentina {n = 3; T. E. Martin, unpubl. data). Only the female in A. pallidiceps appeared to build the nest and incubate the eggs (Oppel et al. 2003), similar to our observations for A. semirufus. Two closely related species. Chestnut-capped Brush Finch {Buarremon brunneinucha) and Stripe-headed Brush Finch {B. torquatus) also occur in our study site. The former usually pre- fers more forested habitat with good canopy cover. The latter nests in vines and bushes, but does not hide the nest between grass clumps as we observed for the Ochre-breasted Brush Finch. However, there is some nest site overlap between Buarremon torquatus and A. semirufus (Luis Biancucci, pers. obs.). Our data are sparse but suggestive of a sin- gle brood per season; we did not observe birds attempting a second brood after a successful previous nest, and the season is sufficiently short that single-brooded behavior seems most likely. Single-brooded behavior was also ob- served for the closely-related Pale-headed Brush Finch (Oppel et al. 2003). This con- trasts with most emberizines in North Amer- ica, which are more commonly double-brood- ed (Martin 1995). The related genera, Junco and Pipilo (Klicka et al. 2000, Yuri and Min- dell 2002, Carson and Spicer 2003) are both multi-brooded (Martin 1995). This pattern is opposite to the generally accepted view that tropical birds have more broods per season than north temperate relatives (Martin 1996). The incubation period was similar to related tropical and subtropical species. The incuba- tion period of the congener A. pallidiceps in Ecuador was 14 days (Oppel et al. 2003), sim- ilar to the 14.9 days that we found for A. se- mirufus. In Argentina, the Stripe-headed Brush Einch had an incubation period of 15.75 ±0.17 days (Auer et al. 2007), and the Saf- fron-billed Sparrow (Arremon flavirostris) av- eraged 15.31 ± 0.25 days (Martin 2002, Auer et al. 2007). Related emberizine species in high elevation Arizona have shorter incuba- tion periods: 12.6 days for Green-tailed To- whee {Pipilo chlorurus) and 12.4 days for Gray-headed Junco {Junco hyemalis caniceps) (Martin 2002). Nest attentiveness was higher than for related species in Argentina where the same sampling protocols were used. The Stripe-headed Brush Einch averaged 54.8% and Saffron-billed Spar- row averaged 62.5% (Auer et al. 2007), while attentiveness averaged 71.5% for A. semirufus. Nest attentiveness of tropical species is com- monly lower than north temperate relatives (Martin 2002, Chalfoun and Martin 2007), but the Ochre-breasted Brush Einch was similar to related species in Arizona; Green-tailed Towhee averaged 70.7% and Gray-headed Junco aver- aged 75.1% (Martin 2002). Nest attentiveness can affect length of the incubation period (Price 1998, Martin 2002, Martin et al. 2007), but can- not explain the difference in incubation period between A. semirufus and northern relatives. The nestling period (10.5 days) was shorter than for Stripe-headed Brush Finch (12.75 ± 0.21 days) and Saffron-billed Sparrow (12.16 ± 0.24 days) in Argentina (Auer et al. 2007), but similar to related species in Arizona where it was 11.1 days for Green-tailed Towhee and 11.0 days for Gray-headed Junco (T. E. Mar- tin, unpubl. data). Provisioning rates were about half the rate that parents fed nestlings in related species in Arizona based on the same sampling methods (Martin et al. 2000). Growth rates were slow compared with north temperate emberizines. For example, the growth rate constant based on mass was k = 0.45 ± 0.02 compared with 0.53 ± 0.01 for 3 1 north temperate emberizine species (Remes and Martin 2002). This slow growth matches general expectations for tropical birds (Rick- lefs 1968, 1976). The Ochre-breasted Brush Finch has a small- er clutch size, longer incubation period, and lower provisioning rates than related north tem- perate species, as is common for tropical birds (Martin et al. 2000, 2007). However, the Ochre- breasted Brush Finch also exhibits several traits that are not typical of tropical birds and are more similar to north temperate relatives such as similar nestling period and nest attentiveness. SHORT COMMUNICATIONS 861 In contrast to conventional views (Martin 1996), it also appeared to have fewer broods than re- lated north-temperate speeies despite a smaller cluteh size and similar breeding season length. These data suggest that more work is needed on breeding biology of tropieal birds to clarify per- ceived patterns of life history differences among latitudes. ACKNOWLEDGMENTS We thank A. M. Niklison and K. L. Decker for help- ful comments on the manuscript, and several people for help in finding nests. This study was made possible in part by support under NSF grants DEB-9981527 and DEB-0543178 to T E. Martin. Permit numbers are DM/0000237 from EONACIT, PA-INP-005-2004 from INPARQUES, and 01-03-03-1147 from Ministerio del Ambiente. LITERATURE CITED Auer, S. K., R. D. Bassar, J. J. Fontaine, and T. E. Martin. 2007. Breeding biology of songbirds in a subtropical montane forest in northwestern Ar- gentina. Condor 109:321-333. Briskie, j. and S. G. Sealy. 1990. Evolution of short incubation periods in the parasitic cowbirds, Mo- lothrus spp. Auk 107:789-793. Carson, R. J. and G. S. Spicer. 2003. A phylogenetic analysis of the emberizid sparrows based on three mitochondrial genes. Molecular Phylogenetics and Evolution 29:43-57. Chaleoun, a. and T. E. Martin. 2007. Latitudinal variation in avian incubation attentiveness and a test of the food limitation hypothesis. Animal Be- havior 73:579-585. Cisneros Palacios, E. 2005. First nesting record of White-naped Brush-finch {Atlapetes albimicha). Huitzil 6:6-7. Fierro-Calderon, K. and T. E. Martin. 2007. Repro- ductive biology of the Violet-chested Humming- bird in Venezuela and comparisons with other tropical and temperate hummingbirds. Condor 109:680-685. Hensler, G. L. and j. D. Nichols. 1981. The Mayfield method of estimating nesting success: a model, estimators and simulation results. Wifson Bulletin 93:42-53. Hilty, S. L. 2003. Birds of Venezuela. Second Edition. Princeton University Press. Princeton, New Jer- sey, USA. Hilty, S. L. and W. L. Brown. 1986. A guide to the birds of Colombia. Princeton University Press, Princeton, New Jersey, USA. International Union for the Conservation of Na- ture (lUCN). 2007. lUCN Red list of threatened species, www.iucnredlist.org (accessed 28 Octo- ber 2007). Klicka, j., K. P. John.son, and S. M. Lanyon. 2000. New World nine-primaried oscine relationships: constructing a mitochondrial DNA framework. Auk 117:321-336. Martin, T. E. 1995. Avian life history evolution in relation to nest sites, nest predation and food. Eco- logical Monographs 65:101-127. Martin, T. E. 1996. Life history evolution in tropical and south temperate birds: what do we really know? Journal of Avian Biology 27:263-272. Martin, T. E. 2002. A new view for avian life history evolution tested on an incubation paradox. Pro- ceedings of the Royal Society of London Series B 269:309-316. Martin, T. E. and G. R. Geupel. 1993. Nest-monitor- ing plots: methods for locating nests and moni- toring success. Journal of Eield Ornithology 64: 507-519. Martin, T. E. and C. K. Ghalambor. 1999. Males helping females during incubation. Required by microclimate or constrained by nest predation? American Naturalist 153:131-139. Martin, T. E., S. K. Auer, R. D. Bassar, A. M. Nik- lison, AND P. Lloyd. 2007. Geographic variation in avian incubation periods and parental influenc- es on embryonic temperature. Evolution 61:2558- 2569. Martin, T. E., P. R. Martin, C. R. Olson, B. J. Hei- DiNGER, AND J. J. PoNTAiNE. 2000. Parental care and clutch sizes in North and South American birds. Science 247:1482-1484. Mayfield, H. 1975. Suggestions for calculating nest success. Wilson Bulletin 87:456-466. Oppel, S., H. M. Schaefer, and V. Schmidt. 2003. Description of the nest, eggs, and breeding behav- ior of the endangered Pale-headed Brush-finch {Atlapetes pallidiceps) in Ecuador. Wilson Bulle- tin 115:360-366. Oppel, S., H. M. Schaefer, V. Schmidt, and B. Schroder. 2004. Habitat selection by the Pale- headed Brush-finch {Atlapetes pallidiceps) in southern Ecuador: implications for conservation. Biological Conservation 118:33-40. Price, T. 1998. Maternal and paternal effects in birds: effects on offspring fitness. Pages 202-226 in Ma- ternal effects as adaptations (T. A. Mousseau and C. W. Fox, Editors). Oxford University Press, New York, USA. Remes, V. AND T. E. Martin. 2002. Environmental in- fluences on the evolution of growth and devel- opmental rates in passerines. Evolution 56:2505- 2518. Remsen, j. V. AND W. S. Graves. 1995. Distribution patterns and zoogeography of Atlapetes Brush- finches (Emberizinae) of the Andes. Auk 112: 210-224. Rickleis, R. 1968. Patterns of growth in birds. Ibis 1 1 0:4 1 9-45 1 . Ricklei s, R. 1976. Growth rates of birds in the humid New World tropics. Ibis 1 18:179-207. RiDCiiii.Y, R. S. AND G. fuDOK. 1994. Birds of South America. Volume 2. University of Texas Press, Austin, USA. 862 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 Sal AM AN, R G. W., L. Davalos, and G. M. Kir wan. 1998. The first breeding records of White-rimmed Brush-finch Atlapetes leucopis with ecological notes. Cotinga 9:24-26. Stotz, D. L, J. W. Litzpatrick, T. A. Parker, and D. K. Moskovits. 1996. Neotropical birds: ecology and conservation. University of Chicago Press, Chicago, Illinois, USA. Yuri, T. and D. P. Mintdell. 2002. Molecular phyloge- netic analysis of Lringillidae, ‘’New World nine-pri- maried oscines” (Aves: Passeriformes). Molecular Phylogenetics and Evolution 23:229-243. The Wilson Journal of Ornithology 120(4):862— 867, 2008 Reproductive Biology of the Red-ruffed Fruitcrow {Pyroderus scutatus granadensis) James A. Muir,i — Diane Licata,i and Thomas E. Martin^ ABSTRACT. — We provide a detailed report on the reproductive biology of the Red-ruffed Lruitcrow (Pv- roderus scutatus granadensis). Eight nests were found between 2003 and 2007 in tropical montane cloud for- est in Yacambu National Park. Lara, Venezuela. All nests were near streams in steep drainages. Nests con- sisted of twigs arranged in a cupped platform. Clutch size was a single egg and the average incubation pe- riod {n = 3) was 22.3 days. Nest attentiveness during incubation averaged [± SE] 76.3 ± 1.86% and in- creased only slightly across stages (early, middle, late). On-bout and off-bout durations were relatively similar across incubation stages. A nestling period of 35 days was recorded for one nest and feather pin-break was estimated to occur at day 19. Brooding attentiveness during the early nestling period averaged 62.5 ± 6.41%, and the adult ceased brooding at about feather pin-break. Food delivery rates increased with nestling age. Food provisioning consisted mostly of insects (66.7%) and lizards (25%) with fruit comprising only 8.3% of the nestling diet at early stages. Provisioning changed to mostly fruit (82.4%) and some insects (17.6%) in late stages of the nestling period. Received 16 January 2007. Accepted 31 January 2008. The reproductive biology of endemic trop- ical birds is poorly known. Information on the life history of the Red-ruffed Fruitcrow (Py- roderus scutatus), despite having a broad neo- tropical distribution, is symptomatic of the poor knowledge of tropical endemics. P. scu- ’ uses, Montana Cooperative Wildlife Research Unit, University of Montana, Missoula, MT 59812, USA. 2 Current address: Ensis, Biodiversity Division, For- estry Road, University of Canterbury, Christchurch, New Zealand. ^ Corresponding author; e-mail: muir.j.a@gmail.com tatiis is a locally rare and uncommon bird of the Cotinga family, a diverse family of pas- serines restricted to the Neotropics. Red- ruffed Fruitcrows are predominately frugivo- rous and inhabit mountainous regions of Ve- nezuela, Colombia, Brazil, Paraguay, Peru, Ecuador, and Argentina. Most records and species accounts for P. scutatus are from Bra- zil and Colombia (Serrano 1994, Pizo et al. 2002). Pyroderus scutatus granadensis occurs in the Coastal Cordillera and lower Andean regions of Venezuela where it inhabits wet, humid, old growth forests (Hilty 2003). No comprehensive descriptions of the breeding biology of P. scutatus granadensis exist. The Red-ruffed Fruitcrow is among the larg- est (300-390 g) passerines in the world (Snow 1982, Hilty 2003) and is considered rare within its range (Stotz et al. 1996). Consequently, life history data are essential for development of conservation strategies that may protect this spe- cies. The objectives of this paper are to describe aspects of the breeding biology of P. scutatus for the first time, and to discuss food provisioning during the nestling period. All data are from montane, wet, primary forest at Yacambu Na- tional Park, Lara State, Venezuela (09° 24' N, 69° 30' W; 1,750-2,000 m elevation). Data were collected from March through June 2002 through 2006. OBSERVATIONS Nest Locations and Descriptions. — We found eight nests of P. scutatus granadensis in tall forest areas of steep-sided river valleys and adjacent ridgelines within primary forest SHORT COMMUNICATIONS 863 (canopy height = —35 m). Nests were found by thoroughly searching throughout the forest or by following adults to the nest site. The eight nests were all on the edge of streams. This is congruent with the description of three lek sites and six nest sites in a river drainage with steep-sided valleys at Ucumari National Park, Colombia (Serrano 1994). Our study sites ranged in elevation from 1,400 to 2,000 m. Red-ruffed Fruitcrows were observed and nests located in higher elevation plots (1,650- 1,850 m). Seven of the eight nests that we found occurred within an area of ~ 6 ha in a major river drainage; the same general area was used for nesting among years. The earli- est nest building behavior was observed on 27 March and the last was on 26 May. Most breeding activity appeared to occur during April, May, and June. Nest heights were visually estimated be- tween 3.5 and 11 m in mid-canopy trees (5- 14 m in height). Three nests were in trees (So- larium spp.), one was in a tree fern (Cyathea- ceae), and the remainder were in unknown tree species. Nests in Solarium spp. were placed in horizontal, forked, mossy branches in positions where leaves provided substantial cover for the nest. Nests were large, untidy structures consist- ing of leafless twigs forming a shallow cup. On one occasion an individual brought sticks to the nest while another adult observed the nest from nearby. All other observations of nest building efforts were of solitary individ- uals. We prepared a detailed description of one nest. The nest was circular to elliptical in shape with an external diameter of 21.5-22.5 cm and an inner diameter of 14.8-15.6 cm. It was close to the trunk of the tree and sup- ported by one main branch with a few large sticks spanning the gap to other branches. The main branch had a diameter of 3.0 cm. The nest was composed of two types of sticks. The outer part of the nest was composed of —45 large, dark brown, dead sticks arranged to form a platform; the sticks were 30-60 cm long and averaged 4-5 mm in diameter. Some sticks had lichen, moss, or black rootlets growing on them. The inner sticks were light reddish-brown and flexible. The longest piec- es were 25 cm in length, but most were —10 cm with a diameter of 1-1.5 mm. The flexible sticks were woven to form a distinct cup placed on and woven into the large stick plat- form. The cup was tightly woven and more dense on the side closest to the tree trunk; it was not possible to see through it. The other side was more loosely woven. Nests measured in Colombia had an external diameter of 37.0 ± 7.5 cm and were lined with leaves (Serrano 1994, Snow 2004). The nest that we measured differed in external diameter and had no inner lining of leaf litter. Clutch Size and Nest Success. — Each active nest had one egg. Six of eight nests were dep- redated; four during incubation and two dur- ing the nestling stage. One nestling fledged from one nest and the other nest was still ac- tive with a 20-day old nestling when we stopped monitoring at the end of June. Overall daily mortality (±SE) rate at these eight nests calculated using the Mayfield method was 0.032 ± 0.013 (exposure days = 190) (Hen- sler and Nichols 1981). Laying and Incubation Behavior. — Red- ruffed Fruitcrows delayed laying eggs follow- ing nest completion; the time between nest completion and laying was 8 days at one nest checked daily and 6-8 days at two other nests. We measured incubation periods as the num- ber of days between the last egg laid and last egg hatched (Briskie and Sealy 1990) based on visual inspection of nests. One of three nests where both laying and hatching were ob- served had an incubation period of 21 ± 1 day and two had an incubation period of 23 ± 1 day each. Nest attentiveness (percent time spent on the nest incubating eggs) during incubation was measured by video-taping nests at early (day 1 to 3 of the incubation period), middle (day 8 to 13 of the incubation period), and late (day 17-22 of the incubation period) stages. The camera ran for 6-8 hrs and started within 1 hr of sunrise (Ghalambor and Martin 2001, Martin 2002). Only the female incubated, as males were not recorded attending the nest. Nest attentiveness remained similar across stages with a slight increase through incuba- tion (Fig. lA). Nest attentiveness averaged 76.3 ± 1.86% (n = 13) across all nests and stages. Duration of incubation on-bouts was similar across stages of the incubation period and averaged 60.0 ± 5.51 min (/? = 13) across all nests and periods (Fig. IB). Off-bout du- ration also was similar among periods (Fig. 864 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 LIG. 1. Behaviors during early (day 1-3), middle (day 8-13), and late (day 17-22) stages of the incubation period for the Red-ruffed Lruitcrow; (A) nest attentiveness and (B) duration of on and off-bouts. Sample sizes ifi) represent numbers of nests sampled and error-bars represent ± SE. IB) and averaged 13.4 ± 0.59 min (n = 13) across all nests and periods. Nestling Development. — Nestling develop- ment was documented by video taping a sin- gle nest in 2004. The nestling remained cov- ered in thin feathery yellow-orange down from day 2 through day 15. Black feathers first developed on the wings, and pin feathers broke their sheaths about day 19 (±1 day) of the nestling period. The nestling had lost al- most all of the down by day 21 and was stand- ing and flapping in the nest. Only residual down remained on the eyebrow region by day 24 and the nestling had a well developed bill. The nestling at day 31 had adult-like black plumage and was about three quarters the size of the adult; the ‘Red Ruff’ throat plumage that is most distinctive in adults was extreme- ly faint and almost non existent in the nestling at this stage. Serrano (1994) noted the ‘Red Ruff’ does not develop until the fledgling is 2 months of age. Video analysis indicated the nestling did not appear to vocalize and beg- ging displays were infrequent and subdued. Serrano (1994) reported a soft ‘mooing’ vo- calization between adult and nestling. We fol- SHORT COMMUNICATIONS 865 FIG. 2. Brooding attentiveness of the Red-ruffed Fruitcrow as a function of age of the nestling based on repeated video observations at two nests. lowed one nest through to fledging with a nestling period of 35 days. Brooding attentiveness (percent time spent brooding the nestling) during the nestling pe- riod was recorded by video-taping two nests starting on day 0 (hatching) and repeated once a week for 6-8 hrs per recording event. Brooding attentiveness decreased with age of the nestling and appeared to stop about the time of feather pin-break (Fig. 2). Adults pro- visioned nestlings an average of 0.18 ± 0.42 times/hr with little variation by age, varying from 0.14 to 0.22 visits/hr across stages. Nestling Diet. — Food items brought by par- ents for nestlings were identified by viewing videos in two nests filmed at early (day 0- 16), middle (day 17-22), and late (day 23-35) nestling stages. The diets of both nestlings were combined to characterize food type (Fig. 3). Food type was recorded if it was positively identifiable and was used to estimate the rel- ative abundance of foods brought to nestlings at different ages. The incidence of fruit in the nestling’s diet increased strongly from early to middle to late nestling stages (Fig. 3). Corre- spondingly, the incidence of insects in the nestling’s diet decreased from early to late in the nestling stages (Fig. 3). DISCUSSION Male Red-ruffed Fruitcrows form commu- nal leks (Hilty 2003) and do not help with parental care at any stage, leaving females to raise the young alone. Many species with fe- male-only nest care are frugivorous or necti- vorous (Beehler 1986, Cockburn 2006). A co- evolutionary interaction between food plant and avian disperser has been proposed by Snow (1976) to be a causal factor influencing male emancipation from nesting duties (Beeh- ler 1986, Cockburn 2006). Snow (1976) uses the Cotingas as exemplary subjects because the female appears to rear the nestling on lipid rich lauraceous fruits unassisted by the male. However fruit may not provide sufficient pro- tein to allow rapid growth (Morton 1973). Data presented here and by Serrano (1994) in- dicate Fruitcrow nestlings do not have a sig- nificant fruit diet until late in the nestling pe- riod. Serrano (1994) reported that nestlings were fed only insects until day 10 of the nest- ling period and only fruit thereafter. In our study, insects were part of the nestling's diet throughout the nestling period in decreasing amounts with nestling age. The high protein diet of P. scutatiis nestlings does not yield rapid growth but may be essential for devel- opment. Brietwisch et al. (1984) reported that nestling frugivory only becomes profitable as nestlings develop cndothermy. P. scutatns adults may feed nestlings an optimal mixture of animal protein for growth, and fruit car- 866 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 16 Early Middle Late Nestling stage FIG. 3. Foods provided to Red-ruffed Fruitcrow nestlings during early (day 0-16), middle (day 17-22), and late (day 23-35) stages of the nestling period. Two nests were sampled in each period. bohydrates for other energetic needs, despite adults being almost solely frugivorous. Ex- tremely low provisioning rates may prolong the nestling period but may also reduce risk of nest predation (Skutch 1949). Development rate of nestlings may represent a compromise among energy delivery, physiological trade- offs, and predation risk. The Red-ruffed Fruitcrow is a rare neotrop- ical endemic and it is essential that its ecolog- ical requirements and reproductive biology be better understood. Further study into the life history and habitat requirements of the Red- ruffed Fruitcrow is clearly needed. ACKNOWLEDGMENTS We acknowledge the Venezuela field crews that aid- ed in data collection. Specifically we thank Brian Schwartz, Elena Arriero, Karie Decker, Bruce Robert- son, Sophie Walker, and the two anonymous reviewers for helping with manuscript review. This study was made possible by support under NSF grants DEB- 9981527 and DEB-0543178 to T. E. Martin. Permit numbers are DM/0000237 from FONACIT, PA-INP- 005-2004 from INPARQUES, and 01-03-03-1147 from Ministerio del Ambiente. LITERATURE CITED Beehler, B. M. 1986. Birds of Paradise and mating system theory, predictions and observations. Emu 87:78-89. Breitwisch, R., P. G. Merritt, and G. H. Whitesides. 1984. Why do Northern Mockingbirds feed fruit to their nestlings? Condor 86:281—287. Briskie, J. V. AND S. G. Sealy. 1990. Evolution of short incubation periods in the parasitic cowbirds, Molothrus spp. Auk 107:789-793. CocKBURN, A. 2006. Prevalence of different modes of parental care in birds. Proceedings of the Royal Society of London, Series B 273:1375-1383. Ghalambor, C. K. and T. E. Martin. 2001. Fecundity survival trade-offs and parental risk taking in birds. Science 292:494-497. Hensler, G. L. and J. D. Nichols. 1981. The Mayfield method of estimating nesting success: a model, estimators and simulation results. Wilson Bulletin 93:42-53. Hilty, S. L. 2003. Birds of Venezuela. Second Edition. Princeton University Press, Princeton, New Jer- sey, USA. Martin, T. E. 2002. A new view of avian life history evolution tested on an incubation paradox. Pro- ceedings of the Royal Society of London, Series B 269:309-316. Morton, E. S. 1973. On the evolutionary advantages and disadvantages of fruit eating in tropical birds. American Naturalist 107(953):8-22. Pizo, M. A., W. R. Silva, M. Galetti, and R. Laps. 2002. Frugivory in Cotingas of the Atlantic Forest of southeast Brazil. Ararajuba 10:177-185. Serrano, D. 1994. Seleccion de Habitat, Ciclo Re- productive y Sistema Lek de Apareamiento de Py- roderus scutatus. Thesis. Universidad de Valle, Cali, Colombia. SHORT COMMUNICATIONS 867 Skutch, a. F. 1949. Do tropical birds rear as many young as they can nourish? Ibis 91:430-455. Snow, D. W. 1976. The web of adaptation. Quadrangle Press, New York, USA. Snow, D. W. 1982. The cotingas: bellbirds, umbrella birds and other species. Comstock Press, Ithaca, New York, USA. Snow, D. W. 2004. Family Cotingidae (cotingas). Pag- es 578-579 in Handbook of the birds of the world. Volume 9. Cotingas to pipits and wagtails (J. del Hoyo, A. Elliot, and D. Christie, Editors). Lynx Editions, Barcelona, Spain. Stotz, D. E, J. W. Eitzpatrick, T. A. Parker III, and D. K. Moskovits. 1996. Neotropical birds: ecol- ogy and conservation. University of Chicago Press, Chicago, Illinois, USA. The Wilson Journal of Ornithology 120(4):867-871, 2008 Nests and Nesting Behavior of Golden Swallow (Tachycineta euchrysea) in Abandoned Bauxite Mines in the Dominican Republic Jason M. Townsend,*’^ Esteban Garrido,^ and Danilo A. Mejia^ ABSTRACT. — Little information is available on nesting of Golden Swallow {Tachycineta euchrysea), a threatened species that occurs only on the island of Hispaniola. We report on six nests discovered and monitored in abandoned bauxite mines in the Sierra de Bahoruco of the Dominican Republic. Nests were in cavities of the vertical walls of these pit mines. Clutch sizes consisted of 2-4 eggs and the nestling stage last- ed between 21 and 24 days with both parents provi- sioning the brood. Three of the six nests were depre- dated by introduced mammals. We compare our ob- servations of Golden Swallow nesting success to nest- ing studies of congeneric swallows and emphasize the potential conservation importance of a nest box place- ment and monitoring program on Hispaniola. Received 3 January 2008. Accepted 6 May 2008. The Golden Swallow {Tachycineta euchry- sea) is a rare and poorly described species that occurs only on Hispaniola, the Greater Antil- lean island politically divided between Do- minican Republic and Haiti. Historically, the range of Golden Swallow also included Ja- maica, but it has not been reported there since 1989 and causes of this local extirpation are unknown (Raffaelle et al. 1998). On Hispan- ' Department of Environmental and Forest Biology, State University of New York, College of Environ- mental Science and Forestry, Syracuse, NY 13210, USA. ^ Grupo Jaragua, Calle El Vergel 36, Ensanche El Vergel, Santo Domingo, Dominican Republic. Sociedad Ornitologica de la Hispaniola, Galen'as Comerciales Suite 202, Avenida 27 de Febrero 54, Santo Domingo, Dominican Republic. •’Corresponding author; e-mail: Jatownse@syr.edu iola, where populations have declined over the last 30 years, the species is increasingly re- stricted to isolated remnant patches of mon- tane forest from 750 m elevation to the high- est forested peaks (Birdlife International 2000, Keith et al. 2003, Latta et al. 2006). Golden Swallows appear to prefer Hispaniolan pine (Pinus occidentalis) and mixed pine-broadleaf forests, but occasionally forage over open ag- ricultural areas and natural savannahs (JMT, pers. obs.). The species occurs in the Massif de la Hotte of Haiti, Sierra de Bahoruco/Mas- sif de la Selle along the border between the two nations, and in Sierra de Neiba and Cor- dillera Central of the Dominican Republic (Keith et al. 2003). The Golden Swallow is a cavity-nesting species, reported to occasionally nest in col- onies (Dod 1992), but detailed descriptions of its nesting ecology are lacking. The species has been observed in the Cordillera Central nesting in cavities of Cecropia schreheriana with 12 nests reported in one stand of four trees during May 1975 (Dod 1992). Multiple adults were observed delivering insects to these nests but the author did not report cavity size, number of nestlings, or structure of the nest. Individuals in the Sierra de Bahoruco have been observed constructing nests in cav- ities of living and dead emergent snags of I)i- dymopana.x tremulus, at heights estimated to be 14-16 m (JMT, pers. obs.). Similar obser- vations of individuals investigating cavities in emergent hardwoods and hardwood snags 868 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 TABLE 1. Golden Swallow nest and nest-site characteristics in abandoned bauxite mines, Sierra de Baho- ruco, Dominican Republic, 2004. Characteristic Nest 1 Nest 2 Nest 3 Nest 4 Nest 5 Nest 6 Mean ± SD Cavity substrate soil/rock soil/rock Height above ground (m) 2 2.7 Depth of cavity (cm) 16 18 Width of cavity (cm) 9 9.8 Height of cavity (cm) 8 9.9 Width of entrance (cm) 4 5.7 Height of entrance (cm) 3 2.6 Elevation (m) 1,092 1,424 Direction of exposure (“) 340 80 soil rock soil/rock rock 0.3 3.5 5 5 3.1 -h 1.8 28.5 23.6 21.5 -h 5.7 11.9 9.2 10.0 -F 1.3 5.2 5.8 7.2 H- 2.2 7.7 9.2 6.7 -F 2.3 5.2 5.8 4.2 -F 1.6 ,306 1,306 1,424 1,395 1,325 -F 126 210 32 254 172 181 -F 113 have been made in the Sierra de Neiba and Massif de la Hotte (Rimmer et al. 2004, 2005). These cavities are often originally ex- cavated by the endemic Hispaniolan Wood- pecker (Melanerpes striatus) (Bond 1943). Additionally, there are reports of nesting in snags of Hispaniolan pine, especially in re- cently burned stands (E. M. Fernandez, pers. comm.); in caves (Fernandez and Keith 2003); and under the eaves of houses (Bond 1943). We report six nests of Golden Swallows found in 2004 while conducting field work on other resident species. These nests occurred in abandoned bauxite pit mines surrounded by pine forest between 1,000 and 1,425 m in the Sierra de Bahoruco (18°07'N, 71°33'W). Pits were generally rounded with a diameter of —200-500 m and vertical walls that ex- tended 5-35 m below the natural surface of the land. Some forest regeneration is occur- ring in and around these pits with scattered Hispaniolan pines ranging from 2 to 5 m height and up to 15 cm diameter at breast height. The mines were created with heavy machinery between 1952 and 1982, and were abandoned in 1985. There are multiple earthen depressions and rock-pile crevices along the walls of these pits that provide cavity spaces for nest construction. OBSERVATIONS Overall Phenology. — Observations of Golden Swallow nesting activity occurred be- tween early February and 25 July 2004. Pairs of swallows were observed investigating cav- ities in pit mines and in snags in early Feb- ruary. Nest construction was observed in pit mines on 5 May (entering with pine needles) and in cavities of emergent hardwood species on 10 May (entering with pine needles) and on 16 May (entering with moss). Between 8 May and 25 July we discovered and moni- tored six active Golden Swallow nests in pit mine habitat. One of the observed nests was depredated on 9 May and an embryo recov- ered from an intact egg was estimated to be 2-4 days old. Ranges of hatch dates calculat- ed for three of these nests were 25-27 May, 30-31 May and 3 June. Hatch dates for the remaining two nests were estimated to be within 15-30 June and 20-28 June. It is not known if these later nests were re-nests or if Golden Swallows nest asynchronously. Fledg- ing dates for the three successful nests were 17 June, 19 July, and 22 July. Physical Placement of Nests. — The ob- served nests were constructed deep within the mining pits and occurred within 5 m of the floor of the pits (Table 1). Nests were con- structed in cavities along the vertical walls where rocks had tumbled and formed layers of crevices and openings. The inner walls and ceilings of cavities consisted of limestone boulders while the floors of the cavities were platforms of soil and rock. The soil in this area is a red, bauxite-rich clay with a dry, fine layer at the surface, and all incubating indi- viduals had red-stained ventral body feathers from contact with this soil. Four of the six nest cavities were accessible for interior measurement and had mean height of 7.2 cm, mean width of 10.0 cm, and mean depth of 21.5 cm (Table 1). Two of the four nest cavities had openings narrower than the interior cavity height with small triangular en- trance holes measuring 4.0-5. 7 cm at the base and 1. 0-2.0 cm at the apex of the triangular openings. SHORT COMMUNICATIONS 869 TABLE 2. Incubation and provisioning observations at Golden Swallow nests in abandoned bauxite mines. Sierra de Bahoruco, Dominican Republic, 2004. Incubation Feeding Nest # # min incubating/ # min observed Mean ± SD duration of incubation period (min) # min observed Mean ± SD # provisionings/hr 1 82/136 11.3 ± 3.9 b 2 234/426 5.6 ± 3.3 c 3 167/397 7.6 ± 3.0 22F 11.4 ± 1.5 4 a 9L 11.9 5 a 390‘ 12.5 ± 4.2 6 74/120 8.2 ± 3.1 C ^ Discovered post-hatching. Depredated during incubation stage. No data collected during provisioning. Data collected over 4 days between nestling days 2 and 10. e Data collected on 1 day = nestling day 6. ^ Data collected over 5 days between nestling days 8 and 16. Nest Materials. — One of the six nests ob- served was removed and collected intact after a depredation event. This nest was ovular in shape to accommodate the shape of the cavity and measured 15 X 10 cm. The long end of the oval nest was from the front to the back of the nest cavity. Nest structure consisted pri- marily of dry pine needles from Hispaniolan pines. These needles were interlaid to form the structural body of the nest and also to level any low points in the earthen floor of the cav- ity. The inner lining of the nest cup was a tightly woven amalgam of dry hanging moss (40-45% of total material), cotton-like plant material (40-45%), body feathers ranging from 1 to 4 cm in length and oriented with quills buried in the plant material (—5%), wo- ven pine needles (—5%), and flakes of bark (<1%). Golden Swallows were frequently ob- served aerially collecting moss from hanging branches of Hispaniolan pines. One solitary individual was observed repeatedly feather- dropping outside of a nest cavity before en- tering the cavity and apparently depositing the feather. Eggs and Clutch Size. — Clutch size was two to four eggs (2 eggs: n = 2; 3 eggs: n = 2; 4 eggs: n = 1 ; at least 2 eggs (observed after depredation]: n = 1). The natural color ot eggs was difttcult to ascertain because all those observed appeared to have been stained with the same red soil that stained the body feathers of the incubating adult. Natural egg color appeared to be white to creamy-white with minimal spotting. One egg, recovered in- tact from a depredated nest, was 18.8 mm in length X 12.8 mm in width at its widest point, smaller than the mean egg size reported for a Caribbean congener, Bahama Swallow (Ta- chycineta cyaneoviridis) (19.4 mm length X 13.9 mm width) (Allen 1996). Parental Behavior. — Nest observations took place on clear, rain-free days between 0900 and 1500 hrs at a distance from the nest of at least 100 m. All observations were by JMT from a position considered to be hidden from view of the birds where he remained sta- tionary for the duration of the observation (mean duration of observation periods = 76 min, range = 32 - 136 min). Nesting pairs were dispersed throughout mine habitat and there was no evidence of overlapping territories, communal breeding or coloniality. Four of the six nests were ob- served during the incubation period (Table 2). Apparently, only one member of the pair in- cubates, based upon observations of the red, soil-stained ventral body feathers of the en- tering and exiting individual, and the clean, white body feathers of the other individual ob- served in the area, but not in the nest cavity. The individual with clean, white body leathers frequently circled within 20 m ol the nest cav- ity vocalizing loudly in the minutes before the red-stained individual exited the cavity, and then foraged with the red-stained individual. Incubating individuals spent 52% ot their ob- served activity on the nest. Provisioning of young was observed at three of the six nests (Table 2). Both members of the pair provisioned nestlings and were ob- served removing fecal sacks. We could not as- 870 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 certain if brooding behavior was shared be- tween the male and female because the feath- ers of both pair members became soil-stained during the nestling period. The mean provi- sioning rate at the three observed nests, each with two nestlings, was 11.9/hr (SD = 0.6). Time to Fledging and Nesting Success. — Three of the six monitored nests successfully fledged young and three were depredated. The three successful nests hatched and fledged dif- ferent numbers of young: one fledgling from four eggs, two fledglings from two eggs, and three fledglings from three eggs. Young fledged at 21-24 days of age at two of the three successful nests, based on known hatch dates, within a range of 1 to 3 days; at the third nest, hatch date was only known within a range of 15 days. The three unsuccessful nests were found with disturbed nest cups and the near-complete feather sets of an adult Golden Swallow strewn within and immedi- ately around the cavity area. We assume the incubating or brooding individual was killed during each depredation event. DISCUSSION The rate of nest depredation (50%) was higher than reported for other congeners. Dep- redation was reported for two of eight nests (25%) of Mangrove Swallows (Tachycineta albilinea) in Panama (Dyrcz 1984) but was not reported as a cause of failure in 10 nests of Bahama Swallows in Grand Bahama (Allen 1996) nor in 128 nests of White-rumped Swal- lows (T. leucorrhoa) in Argentina (Massoni et al. 2007). Nesting success for 3,458 nests of Tree Swallows {T. bicolor) in 10 eastern North American populations was 78.8% and nest predation was not reported as a significant cause of failure (Robertson et al. 1992). Both White-rumped and Tree swallows make exten- sive use of nest boxes which may be consid- ered less vulnerable to predators than natural cavities. Potential mammalian nest predators on His- paniola include Black rats (Rattiis rattiis), Norway rats (R. norvegicus), feral cats (Felis catus), and Indian mongoose (Herpestes ja- vanicus), all invasive species that have been introduced within the last 500 years. An In- dian mongoose was observed entering the nest cavity at one Golden Swallow nest within 12 hrs of nest predation, apparently in an attempt to extract the remaining egg in the far depths of the cavity (Townsend 2006). The Indian mongoose is known to be expanding in pop- ulation on Hispaniola (Horst et al. 2001), and is capable of easily accessing the narrow cav- ities chosen by Golden Swallows. Placement of nests on or near the ground in open pit mine areas may leave Golden Swallows es- pecially vulnerable to predation by mamma- lian predators. It is unknown if Golden Swal- lows nesting in cavities of snags are exposed to high rates of predation, but rats on Hispan- iola are known to be highly arboreal and have depredated nests of other resident cavity nest- ing species (JMT, pers. obs.). Future detailed and more extensive studies should focus on the breeding biology of Golden Swallows in all available habitats to examine nest preda- tion and to place this species’ breeding biol- ogy in the broader ecological context of a ge- nus that spans much of the Americas. That Golden Swallows are nesting in mines abandoned within the last 25 years attests to the species’ willingness to explore novel hab- itats. A network of well-placed nest boxes with proper predator protection could be a key component to this species’ continued exis- tence. Mangrove, Bahama, and White-rumped swallows have all recently been observed nesting in artificial nest boxes and gourds (Dyrcz 1984, Allen 1996, Massoni et al. 2007). However, nest boxes randomly placed in pine forest and mixed pine-broadleaf hab- itats of Hispaniola have not been occupied (JMT, pers. obs.). A more focused nest-box placement effort, possibly concentrating on the open mining areas, may result in greater success. A general program of nest box con- struction, focused nest box placement, inten- sive monitoring, and predator control could provide a wealth of data on this declining spe- cies. The proposed program could also be the basis for potential reintroduction of Golden Swallows into appropriate habitat in Jamaica. ACKNOWLEDGMENTS We are grateful for funding support from C. C. Rim- mer and the Vermont Center for Ecostudies. We thank V. M. Mejia, M. M. Paulino, and A. K. Townsend for assistance in the field and D. W. Winkler for helpful advice. The manuscript was greatly improved by the constructive comments of D. J. Cerasale and L. J. Woolaver. Logistical support and permission to con- duct research in the Dominican Republic was gener- SHORT COMMUNICATIONS 871 ously provided by the Subsecretaria de Areas Prote- gidas y Biodiversidad and Fundacion Moscoso Puello. LITERATURE CITED Allen, P E. 1996. Breeding biology and natural his- tory of the Bahama Swallow. Wilson Bulletin 108; 480-495. Birdlife International. 2000. Threatened birds of the world. Lynx Editions and BirdLife International, Barcelona, Spain and Cambridge, United King- dom. Bond, J. 1943. Nidification of the passerine birds of Hispaniola. Wilson Bulletin 55:115-125. Dod, a. S. 1992. Endangered and endemic birds of the Dominican Republic. Cypress House Press, Fort Bragg, California, USA. Dyrcz, a. 1984. Breeding biology of the Mangrove Swallow and the Gray-breasted Martin at Barro Colorado Island, Panama. Ibis 126:59-66. Fernandez, E. M. and A. R. Keith. 2003. Three un- usual bird nests from the Dominican Republic. Journal of Caribbean Ornithology 16:73-74. Horst, G. R., D. B. Hoagland, and C. W. Kilpatrick. 2001. The mongoose in the West Indies: the bio- geography and population biology of an intro- duced species. Pages 409-424 in Biogeography of the West Indies: patterns and perspectives (C. A. Woods and F. E. Sergile, Editors). CRC Press, Boca Raton, Florida, USA. Keith, A. R., J. W. Wiley, S. C. Latta, and J. A. Ottenwalder. 2003. The birds of Hispaniola: Haiti and the Dominican Republic. An annotated checklist. British Ornithologists’ Union and Brit- ish Ornithologists’ Club, Tring, Herts, United Kingdom. Latta, S., C. Rimmer, A. Keith, J. Wiley, H. Raf- FAELE, K. McFarland, and E. Fernandez. 2006. Birds of the Dominican Republic and Haiti. Princeton University Press, Princeton, New Jer- sey, USA. Massoni, V., F. Bulit, and J. C. Reboreda. 2007. Breeding biology of the White-rumped Swallow in Buenos Aires Province, Argentina. Ibis 148: 10-17. Raffaele, H. a., j. W. Wiley, O. H. Garrido, A. R. Keith, and J. Raffaele. 1998. A guide to the birds of the West Indies. Princeton University Press, Princeton, New Jersey, USA. Rimmer, C. C., J. Almonthe, E. Garrido, D. A. Mejia, M. Milagros, and P. R. Wieczoreck. 2004. Bird records in a montane forest fragment of western Sierra de Neiba, Dominican Republic. Journal of Caribbean Ornithology 16:55-60. Rimmer, C. C., J. M. Townsend, A. K. Townsend, E. M. Fernandez, and J. Almonte. 2005. Avian di- versity, abundance, and conservation status in the Macaya Biosphere Reserve of Haiti. Ornitologia Neotropical 16:219-230. Robertson, R. J., B. J. Stutchbury, and R. R. Cohen. 1992. Tree Swallow (Tachycineta bicolor). The birds of North America. Number 1 1. Townsend, J. M. 2006. Predation of a Golden Swallow nest by the Indian mongoose in the Sierra de Ba- horucos, Dominican Republic. Journal of Carib- bean Ornithology 19:108-109. The Wilson Journal of Ornithology 120(4);87 1-874, 2008 Nest, Nestling Care, and Breeding Season of the Spangled Cotinga {Cotinga cayana) in French Guiana Johan Ingels' ABSTRACT. — 1 report a nest, nestling care, and the breeding season of the Spangled Cotinga {Cotinga cayana) near Saul, French Guiana. The open cup nest was found on 20 October 2007 with a 1 5-20 days old nestling, ~12 m above ground in a yellow mombin tree (Spondias mombin). Nest activities were followed during 4 days, from dawn to dark. The nestling’s plum- age had a scaled color pattern like that of the female, but was paler gray. The female left the nest between 0600 and 0630 hrs in the morning, and provisioned the ne.stling 6-8 times a day. Food items delivered to the nestling were mainly bluish to blackish fruits, ~6-7 ‘ Galgenberglaan 9, B-9070 Destelbergen, Belgium; e-mail: johan.ingels@skynet.be mm in diameter, but also included a ~6 cm cricket (Orthoptera) and a ~10 cm lizard. The female arrived for a last provisioning between 1630 and 1710 hrs, after which she stayed with the nestling until morning. This and other records of breeding Spangled Cotingas suggest that nesting coincides with the dry .season. Re- ceived 24 January 200H. Accepted 26 May 2008. The seven speeies of blue cotingas in the genus Cotinga form a distinct group within the Cotingidae; little is known about their diet, foraging behavior, vocalizations, displays, re- production, and timing of breeding (Snow 872 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 1982, 2004; Brooks et al. 1999; Chaves 2001; Sanchez et al. 2007). Cotinga species have wide gapes and consume figs, mistletoe ber- ries, palm and lauraceous fruits, and have also been recorded taking animal prey, including flying termites and ants, and small amphibians (Snow 1982, 2004). The Spangled Cotinga {Cotinga cayana) is the most widespread Co- tinga species, occurring throughout the Ama- zon Basin and Guianan shield (Snow 2004). Three Spangled Cotinga nests have been re- ported: a nest containing one egg on 21 Au- gust 1924 near Belem, Brazil (Pinto 1953); a female with a nestling collected on 24 Feb- ruary 1946 near Villavicencio, in the eastern Andes of Colombia (Niceforo 1947); and a female constructing a nest on 2 October in the Reserve Naturelle Les Nouragues in French Guiana (Tostain et al. 1992). I describe a nest of a Spangled Cotinga in French Guiana, re- port on the behavior of the female and nest- ling at the nest, describe food items brought to the nestling, and discuss timing of the breeding season. OBSERVATIONS I observed a single Spangled Cotinga nest between 20 and 23 October 2007 at Saiil (3° 37' N, 53° 12' W, -200 m ASL) in the interior of French Guiana. This small village of —70 inhabitants occupies a clearing of —75 ha sur- rounded by primary lowland forest. The — 1- ha garden on the village outskirts where the nest was found had cottages for ecotourists with continuous forest on two sides and ad- jacent gardens on the other. It was planted with flowering shrubs (e.g.. Hibiscus and Poinciana spp.) and fruiting trees (e.g.. Cocos nucifera, Mangifera indica, Persea ameri- cana, Musa spp., and Citrus spp.). Human im- pact on the forest around the village is re- stricted to an airstrip, selective logging for lo- cal construction needs, and a few isolated in- habited clearings and agricultural lands, scattered away from the village. Events at the nest were observed from a distance of —20 m with 7 X 35 binoculars. I recorded the behav- ior of the female caring for the nestling, food items brought to the nestling, and nestling be- havior from dawn to dark during 4 days. On the morning of 20 October 2007, a fe- male Spangled Cotinga with a —6 cm cricket (Orthoptera) in her bill was perched on top of a — 1 5 m dead tree at the edge between forest and garden. Minutes later, I observed the fe- male perching on a —25 cm thick branch, in- side the crown of a yellow mombin tree {Spondias mombin, Anacardiaceae), —25 m from the dead tree. After sitting there motion- less for several minutes, she hopped slowly down the branch to arrive at a nest, where she fed the cricket to a nestling. I estimated the nestling to be 15-20 days of age based on a presumed nestling period of —30 days for Co- tinga species (Snow 1982). The nestling had pin feathers over its entire body, half grown remiges, and a —25 mm long tail. A few tufts of grayish down still adhered to the head feathers. The nestling’s plumage had a scaled color pattern like that of the female, but was paler gray. It was characterized by grayish tips on primaries, secondaries, greater wing-co- verts, and rectrices. The outer edges of the secondaries were also grayish. The dorsal feathers had broader pale grayish edges and the ventral plumage was paler than in the adult female. The nest tree was — 18 m tall and isolated in the garden among a few low fruiting trees, mostly Citrus spp. The nest site was the flatter part of a crotch at a height of —12 m between two branches with a diameter of —25 cm, not surrounded by foliage, and was illuminated and visible from all sides. The nest was a tiny cup with an outer diameter of —5 cm and a height of —1.5 cm, finished with grayish li- chens or a grayish-white fungal mycelium on the outside. The overall grayish appearance of nest and nestling blended well with the gray- ish bark of the yellow mombin tree. The nestling was fed and cared for by the female only. I previously observed a male near the nest site on 6 October 2007, but did not see a male near or at the nest, nor any interaction between the female and a male. The female left the nest between 0600 and 0630 hrs in the mornings of 21-23 October to return with food 25, 40, and 15 min later, re- spectively. Normally, the female arrived with a dark bluish to blackish berry in her bill, 6- 7 mm in diameter, which she fed to the nest- ling, followed by 5-7 more berries by regur- gitation. On 23 October 2007, the female ar- rived at the nest at 1710 hrs with a —10 cm long lizard in her bill, which she presented head first to the nestling and which it managed SHORT COMMUNICATIONS 873 to swallow. Mean (± SD) intervals between feedings during the day were 107 ± 12 min (n = \2) and ranged between 85 and 120 min for a total of 6-8 feedings a day. All provi- sioning of the young was done in silence. The female did not fly directly to the nest when arriving with food, often first perching in the dead tree for several minutes, before flying to the yellow mombin tree. Once in this tree, it took another 5-10 min to approach the nest. Normally she arrived at the nest by hopping down in —0.5 m intervals along one of the thick branches of the fork. To be fed, the nest- ling raised its head and opened its bill without any begging calls. Minutes after being fed, the nestling turned its posterior up to the female, producing one, rarely two fecal sacs, which the female swallowed. Each feeding session was followed by the female brooding the nest- ling for 25-80 min (mean ± SD = 51 ± 21 min, n = 7). The female arrived in the evening at the nest for a last feeding between 1630 and 1710 hrs (mean ± SD = 1653 hrs ± 17 min, n = 3). After the last feeding, the female stayed with the nestling until morning. During daytime brooding sessions, the nest- ling usually sat in front of the female, occa- sionally partly covered by her breast feathers. In bright sun, the female spread her wings slightly shading the nestling. During brooding sessions, as well as during periods when the female was away from the nest, the nestling alternated long bouts of prone roosting upon the nest with short bouts of feather preening and wing exercising. When the nestling was on its own in the nest and in full sunlight, it stretched its head upward with bill wide open and spread its wings slightly, perhaps in an effort to thermoregulate. At night, the nestling was entirely covered by the female. DISCUSSION The nest described is only the fourth re- ported for the Spangled Cotinga (Niceforo 1947, Pinto 1953, Tostain et al. 1992). The nest was a tiny structure for a cotinga with a length of 20 cm and a mean mass of —65 g (Snow 1982). The gray material on the outside of the nest and the gray color pattern of the nestling made both inconspicuous against the gray bark of the nest tree. My observations ot an apparent single egg clutch si/e, and cam- ouflaged nest and nestling were consistent with those of others (Niceforo 1947, Pinto 1953). Movements of the female near the nest were reduced to a minimum. When approach- ing the nest, she often stopped, perching mo- tionless for minutes before finally arriving at the nest. The nestling begged for food by sim- ply raising its head and gaping. At the nest, the female and nestling Spangled Cotinga were silent. It is evident the nest, nestling, and behavior of both female and nestling are all highly adapted to avoid predation (Snow 2004). Snow (2004) estimated nest construction to last —20 days, an incubation period of —20 days, and a nestling period of —30 days for Cotinga species. Assuming this is accurate, construction of the nest I found would have started at the end of August, a single egg would have been laid in the first half of Sep- tember and the nestling would fledge at the beginning of November if it survived. Tostain et al. (1992) observed a female building a nest on 2 October. The breeding season of the Spangled Cotinga in French Guiana appears to coincide with the drier period of the year from the end of July until the end of Novem- ber. Nesting records in Brazil and Colombia also occurred during local dry seasons (Ni- ceforo 1947, Pinto 1953). ACKNOWLEDGMENTS I thank Frederik Brammer, Daniel M. Brooks, Pierre- Yves Henry and Cesar Sanchez M. for help with rel- evant literature, Scott Mori for identifying the nest tree, Olivier Claessens for information about the abun- dance of Cotinga species in French Guiana, and Des Jackson and two anonymous referees for helpful com- ments on the manuscript. I am grateful to Lucien Ti- mane for allowing me to work freely on his property. LITERATURE CITED Brooks, D. M., L. Pando-Vasquez, and A. Ocmin- Petit. 1999. Comparative life history of cotingas in the northern Peruvian Amazon. Ornitologla Neotropical 1 0; 1 93-206. Chave:s, L. 2001. Observations on diet, foraging be- haviour, vocalisations and displays of Spangled Cotinga Cotinga cayana. Cotinga 16:103—104. Niceioro, M. H. 1947. Notas sobre aves de Colombia 11. Caldasia 4:317-377. Pinto, O. M. de O. 1953. Sobre a cok\-ao Carlos Hs- tevao de pelos, ninhos e ovos das aves de Belem (Para). Papeis Avulsos do Departamento de Zoo- logia Sao Paulo 11:11 1-222. 874 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 Sanchez, C., V. Ruiz-Gutierrez, and D. Martinez. 2007. Description of male vocalizations of the Turquoise Cotinga {Cotinga ridgwayi). Wilson Journal of Ornithology 119:455-458. Snow, D. 1982. The cotingas. Cornell University Press, Ithaca, New York, USA. Snow, D. 2004. Lamily Cotingidae (cotingas). Pages 32-108 in Handbook of the birds of the world. Volume 9. Cotingas to pipits and wagtails (J. del Hoyo, A. Elliott, and D. A. Christie, Editors). Lynx Edicions, Barcelona, Spain. Tostain, O., J.-L. Dujardin, C. Erard, and J.-M. Thiollay. 1992. Oiseaux de Guyane (The birds of Erench Guiana). Societe d’Etudes Ornitholo- giques, Brunoy, Erance. The Wilson Journal of Ornithology 120(4):874-878, 2008 Nests, Eggs, an(d Incubation Behavior of Grey-heatied Bullfinch (Pyrrhula erythaca) Jia Chenxi^’2 and Sun Yuehua^ ABSTRACT. — We describe characteristics of nest sites, nests, and the first report of the eggs and incu- bation behavior of the Grey-headed Bullfinch {Pyrrhu- la erythaca). We found nine nests in coniferous forest during June and July 2003 at Lianhuashan Natural Re- serve in central China. Only the female built the nest, but her mate remained in close attendance. Nests were cup shaped and built on horizontal branches of conif- erous trees, 1.3-16.0 m above ground. Clutch size was three eggs {n = 3); the eggs were white in color and spotted with reddish brown. Only the female incubated and averaged 85% nest attendance. On-nest and off- nest bouts (x ± SD) were 31 ± 17 min and 4 ± 5 min, respectively. Egg color, egg size, and clutch size were similar to those reported for other bullfinches, but nest materials differed slightly. Received 24 September 2007. Accepted 30 January 2008. The genus Pyrrhula consists of six species of bud-eating birds of which five species oc- cur in China. Bullfinches have a Palearctic distribution, including the Azores and Japan. The breeding biology of the Eurasian Bull- finch {P. pyrrhula), which has the widest dis- tribution, has been well studied, but relatively little is known of the other species. The Grey- headed Bullfinch (P. erythaca) is distributed in the Himalayan Mountains, central China, and Taiwan, and breeds in mixed coniferous and deciduous forests (Cheng 1987). The nat- ural history of the Grey-headed Bullfinch is ' Key Laboratory of Animal Ecology and Conser- vation Biology, Institute of Zoology, Chinese Acade- my of Sciences, Beijing 100101, China. 2 Corresponding author; e-mail: Jiacx@ioz.ac.cn poorly documented and little is known about its breeding. We previously published a note on the nestlings and nestling behavior of this species based on observation of one nest (Jia et al. 2003). Ali and Ripley (1996: 210) de- scribed one nest as “the usual frail bullfinch type about twelve feet (3.6 m) from the ground on top of a young pine tree. It con- tained young birds about a week old on 19 July”. There are no published descriptions of the eggs of this species. The objective of this paper is to provide descriptions of the nest, eggs, and incubation behavior of the Grey- headed Bullfinch. METHODS Study Area. — This study was conducted in the Lianhuashan Natural Reserve in Gansu Province, central China (34° 40' 67" N, 103° 30' 84" E). Forest habitat occurs in the reserve on north-facing slopes, and some northeast and northwest-facing slopes between 2,600 and 3,600 m. Only shrubs and grasses grow on south-facing slopes. The main cover types in our study area are: (1) coniferous forest dominated by fir {Abies fargesii) and spruce {Picea asperata)-, (2) coniferous-deciduous forest including spruce, fir, birch (Betula uti- lis), and willows {Salix spp.); and (3) shrub- lands including willows, sea buckthorn {Hip- pophae rhamnoides), and grasses. Mean an- nual temperatures range from 5.1 to 6.0° C with a maximum of 34.0° C and minimum of — 27.1° C. The climate is semiarid and, in the SHORT COMMUNICATIONS 875 TABLE 1. Grey-headed Bullfinch nests central China during June and July 2003. observed in the Lianhuashan Natural Reserve, Gansu Province, Nest Date found Nest tree Nest Species DBH (cm) Stage Height above ground (m) Distance to the trunk (m) Fate 1 17 Jun Fir 36 Nest-building 4.2 5 Abandoned 2 18 Jun Fir 24 Incubation 9.0 1.4 Successful 3 20 Jun Spruce — Nest-building 8.0 Abandoned 4 28 Jun Fir 34 Incubation 1.3 2.2 Abandoned 5 1 Jul Spruce 22 Incubation 16.0 0.8 Successful 6 2 Jul Spruce 34 Nest-building 7.0 3 Successful 7 3 Jul Spruce 34 Egg-laying 2.8 2.4 Abandoned 8 3 Jul Fir 28 Nestling 2.2 2.4 Successful 9 13 Jul Fir 34 Incubation 7.0 3 Successful higher elevations, the annual precipitation is —65 cm. The study area has been described in more detail by Sun et al. (2003). Field Procedures. — We systematically searched for nests from June to July 2003. Nests were located by visually following fe- males that were either carrying nesting mate- rials or returning to the nest following an in- cubation break, or by following adults provi- sioning young with food. We also used the location of males as they called to incubating females. Egg length and breadth were mea- sured to the nearest 0.02 mm with vernier cal- ipers, while egg mass was weighed to the nearest 0.1 g using an electronic scale. All measurements were obtained during incuba- tion recesses. Incubation behaviors of females were ob- served at two nests. We monitored one nest 7 m above ground using a digital video cam- corder at a distance of 10 m. We videotaped the nest five times for 1 to 3 hrs each time on 13 and 14 July. The second nest was 1.3 m above ground and was continuously moni- tored for 44 hrs and 21 min through use of a Gemini Temperature Data Logger (Gemini Data Loggers UK Ltd, Chichester, UK) from 1639 hrs (Beijing time) on 29 June to 1300 hrs on 1 July when the nest was abandoned. The data logger probe was inserted into a dummy egg that contained paraffin wax and arranged among the eggs of the nest. Temper- ature of the dummy egg was recorded by the data logger and indicated when the incubating female was on or off nest. We also visually observed the nest after installing the data log- ger for 102 min on 29 June to verify the va- lidity of this method. All observers were in camouflage and all video equipment was cam- ouflaged using branches and leaves. We measured canopy cover, shrub cover, and ground cover on 10 X 10 m square plots at each nest to describe the nest microhabitat. Nests and microhabitat measurements were taken after nests were completed or aban- doned. All data are presented as means ± SD. RESULTS Nine nests were found in the spruce-fir for- est during June and July 2003 (Table 1) at an altitude of —2,850 m. Three nests were found during early nest-building stages on 17 June, 20 June, and 2 July. Two of these nests were later abandoned. During nest building, the fe- male brought materials to the nest accompa- nied by the male on five observations at the three nests. We recorded three nest-building visits during one 17-min bout at one nest. Live nests were found in laying or incuba- tion periods. One female was flushed from a nest on 3 July as we checked whether the nest was active. This nest contained one egg and was abandoned, probably because of our dis- turbance. Lour nests were found from 18 June to 13 July during incubation. One nest was found on 3 July with three nestlings that fledged 2 days later. The male called the in- cubating female from the nest and provided her with food at a nearby tree during the in- cubation period. Nests were on horizontal branches of con- iferous trees (Lig. 1) with mean diameter at breast height (DBH) of 31 ±5 cm (// = 8). Mean height of the nests was 6.2 ± 4.8 m 876 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 FIG. 1. Nest site of the Grey-headed Bullfinch. (range = 1.3-16.0 m, n = 8) above ground level. Mean distances of the nests to the trunk and to the end of the branch were 2.5 ±1.3 m (range = 0. 8-5.0 m) and 0.9 ± 0.4 m (range = 0.4-1. 4 m) (;? = 8), respectively. Nests were loosely constructed of twigs and lined with fine rootlets. Nests were an open cup shape (Fig. 2); measurements for four nests were 6.5 ± 0.5 cm inside diameter, 13.2 ± 2.3 cm outside diameter, 3.5 ± 0.3 cm in- side depth, and 7.5 ± 2.4 cm outside height. Nest trees were on slopes with a mean slope of 12° (/z = 8, range — 2 - 25°). Mean ground cover within 10 X 10 m sampling plots around eight nests was visually estimated to be 85% (70-90%) grass. Mean shrub cover within the plots was 48% (10-80%). Mean canopy cover above the nests was 54% (30- 80%) and tree density in the sampling plots around the nests was 3.1 (1-7) trees per 100 m“ plot. All complete clutches consisted of three eggs (/? = 3). The eggs were pale white with small reddish brown spots, mostly at the large end (Fig. 2). Ten eggs from four nests mea- sured 21.33 ± 0.67 mm long (range = 20.40 -22.24 mm) by 14.86 ± 0.48 mm wide (range = 14.06-15.60 mm), and weighed 2.3 ± 0.2 g (range = 2.0 - 2.7 g). Thirty-four on-bouts and 35 off-bouts were recorded by the data logger during the period monitored (excluding one abnormal 1 1 8 min off-bout and one incomplete 85 min on-bout from all analysis which were the last off-bout and on-bout before the nest was abandoned) (Fig. 3). We recorded activity for one com- plete diurnal cycle on 30 June from 0610 hrs when the female made an initial departure to 2003 hrs (her last arrival). During this period the female had 25 on/off bouts. Ten on-bouts and 12 off-bouts were record- ed in 567 min at the nest monitored by the video camera. Videos showed the female in- cubated alone. The mean on-bout duration, combining the nest with the data logger and the nest with the video camera, was 31 ± 17 SHORT COMMUNICATIONS 877 FIG. 2. Nest and eggs of the Grey-headed Bullfinch. min (range = 1-73 min), 89% of which were between 10 and 60 min {n = 44 on-bouts). The mean off-bout duration (both monitored nests) was 4 ± 5 min (range = 1-24 min). 77% of which was within 5 min {n = 47 off- bouts). All of six off-bouts of more than 10 min were between 12 and 24 min. The two females had similar nest attentiveness patterns 1700 1800 1900 2000 2100 Time (hrs) FIG. 3. Egg temperature recordings of the on/off nest differences at one Grey-headed Bulltinch nest during 1700—2100 hrs on 29 June 2003 in Lianhuashan Natural Reserve, Gansu Province, central China. Temperatures were recorded every 30 sec during the incubation period with a Gemini Temperature Data Logger. 878 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 (percent total monitored time females were in- cubating, excluding night, 1,818 min total ob- served time); the female at the data logger nest incubated 85% of the time while the fe- male at the nest with the video camera incu- bated 86% of the time (mean for both nests = 85%). DISCUSSION Grey-headed Bullfinches appeared to be easily disturbed and prone to abandon their nests during the course of nest construction and egg laying. Three nests were deserted not long after being discovered. One nest deser- tion in the egg laying period was most likely because of our disturbance; the reasons for the other two cases are unknown as they received little disturbance from researchers during field work. All observed nests were on branches of larger conifer trees (DBH > 22 cm). This dif- fers from Ali and Ripley (1996)’s description of one nest that was on top of a young pine {Pinus spp.) tree. Only the female, accompa- nied by the male, built the nest during five nest-building attempts observed. The male called the incubating female from the nest to provide her with food at a nearby tree during the incubation period, consistent with obser- vations of the Orange Bullfinch {P. aurantia- ca) (Roberts 1992). Nest materials of the Grey-headed Bullfinch differed slightly from those reported in nests of the Red-headed Bullfinch (P. erythrocepha- la). Orange Bullfinch, and Eurasian Bullfinch, which contained moss besides twigs and root- lets, and even a few hairs (Orange Bullfinch) and grass (Eurasian Bullfinch) (Ali and Ripley 1996, Snow and Perrins 1998). Egg color and size were similar to those of the Orange Bull- finch and Red-headed Bullfinch (Ali and Rip- ley 1996). Our observed clutch size of the Grey-headed Bullfinch was three eggs, similar to that reported for the Orange Bullfinch (3- 4 eggs) (Roberts 1992). The Orange Bullfinch was reported to prob- ably lay two clutches from June to early Au- gust (Roberts 1992), and the Eurasian Bull- finch at times laid three clutches from late April to late August (Snow and Perrins 1998). Grey-headed Bullfinch nests were found in a variety of nesting stages in early July with one nest fledging and another in the building stage. This suggests that renesting after a failed nest is possible and, although we have no evidence on double clutches, some birds may be double brooded. ACKNOWLEDGMENTS We thank Fang Yun, Zhao Shiqing, Li Jinlin, and staff of the Lianhuashan Natural Reserve for assistance in the field. This study was funded by the National Natural Science Foundation of China (30370204, 30670284, and 30620130110). LITERATURE CITED Ali, S. and S. D. Ripley. 1996. Handbook of the birds of India and Pakistan. Second Edition. Volume 10. Oxford University Press, New Delhi, India. Cheng, Tso-hsin. 1987. A synopsis of the avifauna of China. Science Press, Beijing, China. JiA, C. X., Zh. Wang, and Y. H. Sun. 2003. A short note on nestling behaviour of Grey-headed Bull- finch at Lianhuashan Natural Reserve, Gansu, China. Sichuan Journal of Zoology 22:17. Roberts, T. J. 1992. The birds of Pakistan. Volume 2. Oxford University Press, Karachi, Pakistan. Snow, D. W. and C. M. Perrins. 1998. The birds of the Western Palearctic. Volume 2. Oxford Univer- sity Press, Oxford, United Kingdom. Sun, Y. H., J. E. Swenson, Y. Fang, S. Klaus, and W. SCHERZINGER. 2003. Population ecology of the Chinese Grouse, Bonasa sewerzowi, in a frag- mented landscape. Biological Conservation 110: 177-184. SHORT COMMUNICATIONS 879 The Wilson Journal of Ornithology 120(4):879-883, 2008 Parental Care in Tawny-bellied (Sporophila hypoxantha) and Rusty-collared {S. collaris) Seedeaters Carolina Facchinetti,' Alejandro G. Di Giacomo,^ and Juan C. Reboreda^ "* ABSTRACT. — The genus Sporophila (Emberizidae) comprises species of small finches characterized by marked sexual dichromatism, which in birds is posi- tively associated with extent of female bias in parental care. We analyzed differences in parental care in Taw- ny-bellied (5. hypoxantha) and Rusty-collared (5. col- laris) seedeaters. We video-recorded nest activity dur- ing incubation and when young were 2-4 and 7-9 days of age. Females of both species built the nest and incubated the eggs alone. Female Tawny-bellied Seed- eaters: (1) incubated 59% of the time, (2) had a higher frequency of nest visits than males when chicks were 2-4 days of age, and (3) their visits were longer be- cause after feeding they remained in the nest brooding the chicks. There were no gender differences in fre- quency of nest visits when chicks were 7-9 days of age, but visits of females were longer than those of males. Female Rusty-collared Seedeaters: (1) incubat- ed 51% of the time and (2) had a higher frequency of nest visits when chicks were 7-9 days of age. Both males and females brooded chicks and there were no gender differences in frequency and length of nest vis- its when chicks were 2-4 days of age. Parental care in both species is female biased, but the extent of male care is slightly higher in Rusty-collared than in Tawny- bellied seedeaters. Received 4 February 2008. Accept- ed 29 April 2008. The most common mating pattern among birds is social monogamy with biparental care (Lack 1968, Silver et al. 1985, Neudorf 2004). However, there is a considerable variation in type and extent of male care within monoga- mous species, as males may or may not de- fend territories, build the nest, incubate the eggs, brood and deliver food to the nestlings, and assist fledglings after they leave the nest (Clutton-Brock 1991). ' Departamento de Ecologia, Genetica y Evolucidn, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, Pabellon II, Ciudad Universitaria, C1428EGA, Buenos Aires, Argentina. ^ Departamento de Conservacion, Aves Argentinas/ Asociacidn Ornitoldgica del Plata, Matheu 1246, C1249AAB Buenos Aires, Argentina. ■"'Corresponding author; e-mail: reboreda@ege.fcen.uba.ar. The genus Sporophila (Emberizidae) com- prises 31 species of small finches that inhabit grassy open and semi-open habitats from Mexico to central Argentina (Meyer de Schauensee 1952, Ridgely and Tudor 1989). All Sporophila species are characterized by a marked plumage-color dimorphism with males strongly patterned and colorful, and fe- males duller and similar between species (Ridgely and Tudor 1989). Variation in mel- anin-based dichromatism in birds is positively associated with extent of gender bias in pas- sive brood defense (Owens and Hartley 1998). Males of dichromatic species are less likely to participate in nest building (Soler et al. 1998) and to share incubation with females (Vemer and Willson 1969). The reproductive biology of Sporophila seedeaters is little studied and most research was in Central America (i.e.. Gross 1952, Skutch 1954, Alderton 1961, Stutchbury et al. 1996, but see Marcondes-Machado 1997, Francisco 2006). These studies indicate that most Sporophila seedeaters are socially mo- nogamous and have biparental care, but do not report details of sexual differences in parental care. The reproductive biology of Tawny-bellied Seedeater {S. hypoxantha) and Rusty collared Seedeater {S. collaris) is virtually unknown, except for the work of Di Giacomo (2005). Tawny-bellied Seedeaters are included within the “capuchinos”, a monophyletic group of 1 1 species (Lijtmaer et al. 2004) that share a similar male plumage coloration pattern (Ridgely and Tudor 1989). Rusty-collared Seedeaters are included in a different clade than Tawny-bellied Seedeaters (Lijtmaer et al. 2004) allowing us to compare gender differ- ences in parental care between species that are relatively distant within the genus. The objec- tives of our study were to: ( 1 ) analyze the ex- tent of gender bias in parental care in S. hypo- xantha and S. collaris, and (2) compare gen- 880 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 der differences in parental care between these species. METHODS Study Area. — The study was conducted at Reserve El Bagual (26° 10' S; 58° 56' W), in the Province of Formosa, Argentina, during breeding seasons 2004—2005 and 2005—2006. This 3,300-ha reserve is in the region of the eastern or humid Chaco. Average annual rain- fall in the area is 1 ,500 mm with mean month- ly temperature varying from 16.9° C in July to 26.7° C in January. Study Species. — Tawny-bellied Seedeaters are sexually dichromatic but genders do not differ in body size (~8-9 g). They are resi- dents at our study site and nest in dry grass- lands of Eliomiriis miiticus and Imperata brasiliensis or Andropogon latei^alis, and in wet grasslands of Paspalum intermedium and Sorghastrum setosum. Laying starts during the second half of October and continues until the second half of March. The nest is a deep semi- spherical open cup built at a height of —70 cm above ground level. Clutch size is 2—3 eggs; incubation starts with the laying of the penultimate egg and lasts 11-12 days. Young fledge when they are 9-10 days of age (Di Giacomo 2005). Rusty-collared Seedeaters are sexually di- chromatic and males and females do not differ in body size (-12-13 g.). They are residents at our study site and nest in wetlands of Cv- perus giganteus from early November to early April. The nest is a semispherical open cup built at a height of 1-1.5 m above ground lev- el. Clutch size is 2-3 eggs; incubation starts with laying of the penultimate egg and lasts 12-13 days. Young leave the nest when they are 9-12 days of age (Di Giacomo 2005). Data Collection. — We searched for nests during the breeding season and visited nests daily or every other day to identify start of laying and date of hatching. We recorded nest- building activity in 18 nests of Tawny-bellied Seedeaters and 10 nests of Rusty-collared Seedeaters. We recorded the gender of the in- dividual that carried nest material during 3-4 consecutive trips for each nest (—15-25 min of observation/nest). We video-recorded nest activity using a Sony Hi8 CCD video camera placed 1.5-2 m from the nest. Video records lasted between 2 and 4 hrs. We video recorded six nests during incubation (12 hrs of video records) for Tawny-bellied Seedeaters, 13 nests with chicks 2-4 days of age (36 hrs), and five nests with chicks 7-9 days of age (12 hrs). We video recorded four nests of Rusty- collared Seedeaters during incubation (8 hrs), four nests with chicks 2-4 days of age (8 hrs), and six nests with chicks 7-9 days of age (20 hrs). We observed both members of the pair at all of these nests. We video recorded each nest throughout incubation during morning (between 0700 and 1100 hrs) and afternoon (between 1500 and 1900 hrs). We video re- corded nests only during the morning for the chick stages. Data Analysis. — We watched the videotapes in the laboratory and identified the gender of the parent, number of times (to the nearest sec) it landed at, entered, exited, and departed from the nest, and activity at the nest (incu- bation, feeding, or brooding). We measured the length of each activity in those visits where there were two different behaviors (i.e., feeding and brooding). We calculated the fre- quency of visits to the nest (nest visits/hr) for each parent, average length of the visits (sec), time elapsed between visits (sec), and propor- tion of time spent incubating eggs, and feed- ing and brooding chicks. We used nonparametric statistics for anal- yses due to lack of normality of the data and small sample sizes of the groups. Statistical analyses were performed using StatView Ver- sion 5.0 statistical software (SAS Institute 1998). All P-values are two-tailed with alpha (a) set at 0.05. Data are presented as means ± standard errors. We provide confidence in- tervals when comparisons yield a nonsignifi- cant result because of small sample sizes (Colegrave and Ruxton 2003). RESULTS Tawny-bellied Seedeaters.— TcmdAQS built the nest and incubated the eggs alone. We did not find differences in frequency and length of incubation bouts between morning and af- ternoon (Wilcoxon Signed Rank tests, fre- quency: z = -0.94, P = 0.35, 11 = 6; length: 2 = -0.10, P = 0.92, n = 6). Females spent 59 ± 5% of the time at the nest including observations of morning and afternoon to- gether; frequency and length of incubation bouts were 2.3 ± 0.2 bouts/hr and 16.5 min SHORT COMMUNICATIONS 881 □ Females ■ Males 2-4 7-9 Age of chicks (days) FIG. I. (A) Frequency and (B) length of nest visits by female and male Tawny-bellied Seedeaters {Spo- rophila hypoxantha) at Reserve El Bagual, Formosa Province, Argentina. Values are mean ± standard er- rors. The number of nests analyzed was 13 and hve for chicks 2-4 and 7-9 days of age, respectively. ± 27 sec, respectively. Males fed females dur- ing incubation at three of six nests with a fre- quency of 0.9 ± 0.5 feedings/hr. Clutch size in nests monitored during incubation was 2.5 ± 0.5 (/2 = 6 nests). Females had a higher frequency of nest vis- its than males when chicks were 2-4 days of age {z = -2.2, P = 0.028, n - 13; Fig. 1). Visits of females were longer than those of males (z = -3.2, P = 0.002, n = 13; Fig. 1); the female remained in the nest brooding the chicks after feeding them on 86% of the visits. Females spent 40 ± 3% of the time brooding chicks while this behavior was not observed in males. The frequency of female’s visits with brooding was negatively associated with air temperature (Spearman Rank correlation, p = -0.59, z = -2.04, P = 0.042, n = 13), but there was no association between length of brooding visits and air temperature (p = 0.19, z = 0.67, P = 0.51, /? = 13). The number of chicks in nests monitored at this stage was 1.9 ± 0.5 {n = 13). There were no gender differences in fre- quency of nest visits when chicks were 7-9 days of age (z - -0.7, P = 0.50, n = 5, mean ± SD females: 4.2 ± 1.4, males: 4.8 ± 2.9, 95% confidence interval: -3.8 - 2.6; Fig. 1), but females still had significantly longer visits than males (z = -2.0, P - 0.043, n = 5\ Fig. 1). Females continued brooding chicks in 21% of the visits and spent 5 ± 2% of the time in this activity; this behavior was not observed in males. The number of chicks in the nests monitored at this stage was 1.6 ± 0.5 {n = 5). Both males and females removed fecal sacs during the chick stage. Rusty-collared Seedeaters. — Females built the nest and incubated the eggs alone. We did not find differences between morning and af- ternoon in frequency (z = —0.37, P = 0.72, n - 4) and length (z = -0.0, P - 0.99, n = 4) of incubation bouts. We combined morning and afternoon observations; females spent 51 ± 8% of the time at the nest with a frequency and length of the incubation bouts of 3.0 ± 0.6 bouts/hr and 12.5 ± 3.2 min, respectively. We observed that in one of four nests the male fed the female during incubation (frequency = 0.65/hr). Clutch size in the nests monitored during incubation was 2.7 ± 0.5 {n — 4 nests). The frequency and length of nest visits when chicks were 2-4 days of age did not differ between genders (z = -0.73, P = 0.46, n = 4, mean ± SD females: 3.1 ± 1.7, males: 1.5 ± 2.1, 95% confidence interval: —2.4 - 5.5; and z — —1.46, P = 0.14, n = 4, mean ± SD females: 263.8 ± 330.8 sec, males: 56.1 ± 75.2 sec, 95% confidence interval: -228.1 - 644.1, respectively. Fig. 2). Females spent 19 ± 15% of the time brooding chicks. The male brooded the chicks in two of four nests. The number of chicks in the nests monitored at this stage was 2.2 ± 0.5 (/? = 4). The frequency of nest visits was higher for females when chicks were 7-9 days of age (z = -2.0, P = 0.046, u = 6; Fig. 2). Females brooded the chicks 13 ± 8% of the time at this stage, while brooding by males was not observed; we did not detect gender differences in length of nest visits (z = — 1.4, P = 0.17, u = 6, mean ± SD females: 364.8 ± 617.8 sec, males: 16.7 ± 2.76 sec, 95% confidence interval: —299.3 - 995.5; Fig. 2). The number 882 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 □ Females ■ Males 2-4 7-9 Age of chicks (days) FIG. 2. (A) Frequency and (B) length of nest visits by female and male Rusty-collared Seedeaters {Spo- rophila coUaris) at Reserve El Bagual, Formosa Prov- ince, Argentina. Values are mean ± standard errors. The number of nests analyzed was four and six for chicks 2-4 and 7-9 days of age, respectively. of chicks in the nests monitored at this stage was 1.7 ± 0.8 {n = 6). Both males and fe- males removed fecal sacs during the chick stage. DISCUSSION Parental care is female biased in Tawny- bellied and Rusty-collared seedeaters. The fe- male built the nest and incubated the eggs without assistance of the male in both species. The pattern of parental care after hatching was also similar between species with females spending more time at the nest than males. The main difference between species was that male Rusty-collared Seedeaters brooded the chicks while this was not observed in Tawny- bellied Seedeaters. Other differences between species (i.e., lack of gender differences in fre- quency and length of visits in the period of 2-4 days of age in Rusty-collared Seedeaters) could be attributed to low power because of small sample sizes. There are previous studies in other species of seedeaters that provide some qualitative data on type and extent of parental care in- cluding: White-collared Seedeater (5. morel- leti\ Skutch 1954), Slate-colored Seedeater {S. schistacea; Stutchbury et al. 1996), Variable Seedeater (S. corvine'. Gross 1952, Skutch 1954), Lined Seedeater (S. lineola', Marcon- des-Machado 1997), Yellow-bellied Seedeater {S. nigricollis', Alderton 1961), and Double- collared Seedeater {S. caerulescens', Francisco 2006). The female incubates the eggs and broods the chicks alone in all of these species and males did not participate in nest building, except for Yellow-bellied Seedeaters. Both parents provision the young in most of these species but there is no information on gender bias in provisioning rates (except for Double- collared Seedeater where females have a high- er frequency of nest visits than males, Fran- cisco 2006). The only species where males do not participate in chick feeding is Slate-col- ored Seedeater (Stutchbury et al. 1996). None of the previous studies reported mate feeding by males during incubation (as we observed in Tawny-bellied and Rusty-collared seedeat- ers) or male brooding (as we observed in Rus- ty-collared Seedeater). The results for species of seedeaters studied to date indicate this group has female biased biparental care with noticeable differences in the type and extent of male care. Plumage coloration often has an important role in conspecific interactions (Andersson 1994); Owens and Hartley (1998) reported gender differences in parental care are asso- ciated with melanin-based dimorphism. Eu- melanin pigments often signal competitive ability and social dominance (Senar 1999), and act as honest signals of male parental quality in a few species (Siefferman and Hill 2003). We did not study the pigments respon- sible for plumage coloration, but it is likely the rufous rump and underparts of Tawny-bel- lied Seedeaters, and the black head, black pec- toral band, and cinnamon rump and underparts of Rusty-collared Seedeaters are produced by eumelanins and pheomelanins (Gill 1995). Further studies analyzing preference of fe- males for coloration of males, and the asso- ciation between coloration of males and pa- rental quality may help in understanding the function of sexual dichromatism in this group. SHORT COMMUNICATIONS 883 ACKNOWLEDGMENTS We thank Alparamis SA and Aves Argentinas/Aso- ciacion Ornitologica del Plata for allowing us to con- duct this study at Reserve El Bagual. We also thank two anonymous reviewers for helpful comments on a previous version of this manuscript. CF was supported by a fellowship from the Consejo Nacional de Inves- tigaciones Cientificas y Tecnicas (CONICET). JCR is a Research Fellow of CONICET. LITERATURE CITED Alderton, C. C. 1961. The breeding cycle of the Yel- low-bellied Seedeater in Panama. Condor 63:390- 398. Andersson, M. 1994. Sexual selection. Princeton Uni- versity Press, Princeton, New Jersey, USA. Clutton-Brock, T. H. 1991. The evolution of parental care. Princeton University Press, Princeton, New Jersey, USA. COLEGRAVE, N. AND G. D. RuxTON. 2003. Confidence intervals are a more useful complement to nonsig- nificant tests than are power calculations. Behav- ioral Ecology 14:446-450. Di Giacomo, A. G. 2005. Aves de la Reserva El Ba- gual. Pages 203-465 in Historia Natural y Paisaje de la Reserva El Bagual. Temas de Naturaleza y Conservacion 4 (A. G. Di Giacomo and S. E Kra- povickas. Editors). Aves Argentinas/Asociacion Ornitologica del Plata, Buenos Aires, Argentina. Francisco, M. R. 2006. Breeding biology of the Dou- ble-collared Seedeater {Sporophila caerulescens). Wilson Journal of Ornithology 118:85-90. Gill, E B. 1995. Ornithology. W. H. Freeman and Company, New York, USA. Gross, A. O. 1952. Nesting of Hicks’ Seedeater at Barro Colorado Island, Canal Zone. Auk 69:433- 446. Lack, D. 1968. Ecological adaptations for breeding in birds. Methuen Press, London, United Kingdom. Lijtmaer, D. a., N. M. M. Sharpe, P. L. Tubaro, and S. C. Lougheed. 2004. Molecular phylogenetics and diversification of the genus Sporophila (Aves: Passeriformes). Molecular Phylogenetics and Evo- lution 33:562-579. Marcondes-Machado, L. O. 1997. Reproductive be- havior of Sporophila lineola (Linnaeus) (Passeri- formes, Emberizidae). Revista Brasileira de Zoo- logia 14:517-522. Meyer de Schauensee, R. 1952. A review of the ge- nus Sporophila. Proceedings of the Academy of Natural Sciences of Philadelphia 104:153-196. Neudore, D. L. H. 2004. Extrapair paternity in birds: understanding variation among species. Auk 121: 302-307. Owens, I. P. F. and I. R. Hartley. 1998. Sexual di- morphism in birds: why are there so many differ- ent forms of dimorphism? Proceedings of the Royal Society of London, Series B 265:397-407. Ridgely, R. S. and G. Tudor. 1989. The birds of South America. Volume 1. Oxford University Press, Oxford, United Kingdom. SAS Institute. 1998. StatView user’s guide. Version 5.0. SAS Institute Inc., Cary, North Carolina, USA. Senar, J. C. 1999. Plumage coloration as a signal of social status. Proceedings of the International Or- nithological Congress 22:1669-1686. SiEEEERMAN, L. AND G. E. HiLL. 2003. Structural and melanin coloration indicate parental effort and re- productive success in male Eastern Bluebirds. Be- havioral Ecology 14:855-861. Silver, R., H. Andrews, and G. E Fall. 1985. Paren- tal care in an ecological perspective: quantitative analysis of avian subfamilies. American Zoologist 25:823-840. Skutch, a. E 1954. Life histories of Central American birds. Families Fringillidae, Thraupidae, Icteridae, Parulidae and Coerebidae. Pacific Coast Avifauna 31:19-49. SOLER, J. J., A. P. M0LLER, AND M. SOLER. 1998. NeSt building, sexual selection and parental investment. Evolutionary Ecology 12:427-441. Stutchbury, B. j. M., P. R. Martin, and E. S. Mor- ton. 1996. Nesting behavior of the Slate-Colored Seedeater {Sporophila schistacea) in Panama. Or- nitologia Neotropical 7:63-65. Verner, j. and M. E Willson. 1969. Mating systems, sexual dimorphism and the role of male North American passerine birds in the nesting cycle. Or- nithological Monographs 9:1-76. 884 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 The Wilson Journal of Ornithology 120(4):884-887, 2008 Postnatal Growth Rates of Hummingbirds: Review and New Records Bemd P. Freymann* ^^ and Karl-Ludwig Schuchmann^ ABSTRACT — We review the published informa- tion on postnatal growth rates of hummingbirds (13 species), and report previously unpublished records for nine additional trochilid species. The allometric rela- tionship based on the logio-transformed data of K (lo- gistic growth rate constant) and body mass has a slope of -0.313 and an intercept of -0.346 (n = 22; F = 0.18; P = 0.049). The allometric relationship has a slope of -0.366 and an intercept of -0.327 (n = 20; F = 0.30; P = 0.013) if the two Nearctic records are excluded. Visual inspection suggests that higher /^-val- ues occur in Nearctic hummingbirds (x = 0.422; n = 2) compared to Neotropic species (x = 0.269; n = 20). We suggest a revival of studies collecting basic life history information such as postnatal growth rates of birds, especially of tropical taxa. Received 12 October 2007. Accepted 18 February 2008. The Metabolic Theory of Ecology (MTE) is a controversially debated theory related to search for nomological explanations in ecol- ogy (Brown et al. 2004). It represents a revival of allometric scaling studies linking organis- mal metabolic rates to body mass — across several orders of magnitude — via temperature as the central factor (Gillooly et al. 2001). Gil- looly et al. (2002) in the framework of MTE, also developed a fundamental model, based on first principles of allometry and biochemical kinetics, that predicts the time of ontogenetic development as a function of body mass (bm) and temperature. Their model focuses on em- bryonic growth but is, to a certain extent, also applicable to postnatal growth (Gillooly et al. 2002). The discussion of MTE indicates that raw data on life history aspects such as post- natal growth rates of single and, in particular. * Alexander Koenig Research Institute and Museum of Zoology (ZFMK), Leibniz Institute for Terrestrial Biodiversity, Research Group: Biology and Phylogeny of Tropical Birds, Adenauerallee 160, 53113 Bonn, Germany. 2 Current address: Community and Conservation Ecology Group, Centre for Ecological and Evolution- ary Studies, University of Groningen, P. O. Box 14, 9750 AA Haren, The Netherlands. ^Corresponding author; e-mail: b.freymann@rug.nl tropical taxa, are still sparse and urgently needed. Hummingbirds (Trochilidae), which are predominantly nectarivorous, lay invari- ably two eggs in nests that are mostly of cup- like, occasionally also domed or semi-domed, architecture (Schuchmann 1999). After an in- cubation period of 16-19 days the altricial chicks stay for another 23-26 days, in case of the high Andean trochilids 30-40 days, in the nest (Schuchmann 1999). The nestlings are fed by regurgitation of fluid and arthropods, often spiders (e.g., Thomas 1994, Marin 2001). We report previously unpublished re- cords of postnatal growth rates of humming- birds and review the published records of this family. We show how this information applies to important ecological questions, e.g., tem- perate versus tropical growth rates. METHODS New Records. — The previously unpublished data on postnatal growth rates of humming- birds presented here were measured by the se- nior author and Julia Beintmann (1998) during 1984-2004 at a variety of localities (Table 1). We obtained growth constants {K) for nest- lings using the logistic equation provided by Ricklefs (1976): Wj = A X (1+e-^ («■>)-! where Wj = mass at time t; A = maximum mass; e = base of the natural logarithms; K = growth rate constant; and t, = age at the inflection point of the growth curve. An in- verse measure of growth, the time required to grow from 10 to 90% of the asymptote (Rick- lefs 1976) was calculated using the equation: tio-90 (days) = 4.4/K Literature Review. — We searched the liter- ature for published records starting with the avian growth rate data set from Starck and Ricklefs (1998). K- and tio_9o-values were ei- ther from the original publications or com- puted by us using the formulas from Ricklefs SHORT COMMUNICATIONS 885 TABLE 1 . Postnatal growth rates of hummingbirds (Trochilidae). K = growth rate, n = individuals mea- sured, t,o_9o = inverse growth rate. Subfamily Species Body mass [g] K [d-'j 4 0-90 [d] n Locality (altitude [m]) Reference Phaethornithinae Ramphodon naeviiis 6.9^ 0.206 21.4 2 Brazil (650) This study Glands hirsuta 6.8^ 0.251 17.5 1 Trinidad (500) This study Phaethornis ruber 2.4 0.469 9.4 1 Surinam (50) Schuchmann 1986 Trochilinae Campylopterus hemileucurus 10.3 0.148‘’ 14.6 7 Costa Rica (1,900) Marm 2001 Colibri coruscans 7.7 0.29 15.2 5 Ecuador (3,000) Beintmann 1998 0.300 14.7 8 Venezuela (4,000) Ztichner 1998 Eiilampis jugularis 9.5^ 0.273 16.1 3 St. Lucia (900) This study Orthorhyncus cristatus 3.8^ 0.270 16.3 2 Guadeloupe (150) This study Chlorostilbon mellisugus 2.8 0.278 15.8 1 Venezuela (700) Thomas 1994 Thalurania colombica 4.2^ 0.322 13.7 1 Colombia (500) This study Amazilia tzacatl 5.0 0.362 12.1 2 Panama (50) Ricklefs 1976 0.334 13.2 3 Colombia (500) Schuchmann and Tdgel in Starck and Ricklefs 1998 Polyerata fimbriata 5.2 0.256 17.2 1 Surinam (s.l.) Haverschmidt 1952 Saucerottia tobaci 4.4^ 0.264'’ 16.7 2 Trinidad (s.l.) Muir 1925 Lampornis clemenciae 7.5 0.463 9.5 2 USA (1,650) Schuchmann 1985 Sternoclyta cyanopectiis 6.8 0.28 15.7 10 Venezuela Fierro-Calderon and (1,350-2,000) Martin 2007 Oreotrochilus estella 8.0 0.190'’ 23.2 2 Peru (3,900) Dorst 1962 Haplophaedia lugens 5.9 0.172 25.6 2 Colombia (1,800) This study Ocreatus imderwoodii 2.9^ 0.195 22.6 1 Colombia (1,800) This study Lesbia victoriae 5.1 0.25 17.6 5 Ecuador (3,100) Beintmann 1998 Oxypogon guerinii 5.5 0.233 18.9 6 Venezuela (4,000) Ziichner 1998 Metallura tyrianthina 3.8 0.27 16.3 6 Ecuador (3,100) Beintmann 1998 Calliphlox evelynae 3.3 0.408 10.8 2 Bahamas (s.l.) Schuchmann in Starck and Ricklefs 1998 Selasphorus riifus 3.5 0.380 11.6 2 USA (600) Constantz 1980 ^ Source not original publication but Schuchmann (1999); if available mean of males and females given. ^ Data newly computed. (1976). Missing (data on atiult bo(dy mass not reporteid in the original publications are from Schuchmann (1999). Statistical Analysis. — We log,o-transformed the raw (data for body mass (g) and K prior to analysis. We used least-squares regression (GLM) in Statistica 6.1 (StatSoft 2003) to test whether body mass predicts K (assuming in- dependence among the data). RESULTS We obtained published information on K- and t|,)_9()-va*ues for 13 trochilid species from the literature (Table 1). We also present pre- viously unpublished data on the postnatal growth parameters of nine additional hum- mingbird species (Table 1). This increases the number of hummingbird species (/? = 20) with published information on K- and t,o_9o-values by 69%: the number of published records rises from one to three for the subfamily Phaethor- nithinae. We also present an additional record for one previously reported species (Table 1 ). The allometric relationship based on the log,()-transformed data of K and body mass (g) has a slope of —0.313 and an intercept of -0.346 in = 22; r~ = 0.18; P = 0.049) for all species included. The allometric relation- ship has a slope of —0.366 and an intercept of -0.327 in = 20; r- = 0.30; P = 0.013) if the two Nearctic records (Lanipornis clenicn- ciae and Selaspliorus rufiis) are excluded. A statistical comparison of the means of the K- values of the Nearctic versus Neotropic hum- 886 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 mingbird species is not meaningful due to small sample sizes, but visual inspection sug- gests that higher AT-values occur in Nearctic hummingbirds (x = 0.422; n = 2) compared to the neotropical trochilids (x = 0.269; n = 20). DISCUSSION Our synthesis of previously unreported as well as published records represents a signif- icant quantitative and qualitative improvement of the information available on postnatal growth development of juvenile humming- birds. The quantitative improvement is appar- ent at first glance, given the increase of rec- ords by 69%. The qualitative improvement is more obscure. The most complete review of postnatal growth rates was published by Starck and Ricklefs (1998). The newer work by Fierro-Calderon and Martin (2007) on the Violet-chested Hummingbird (Sternoclyta cy- anopectus) is a detailed and informative study on the reproductive biology of this species. It also includes a review and discussion of sev- eral reproductive parameters of hummingbirds comparing tropical with temperate species. A reanalysis of Starck and Ricklef’s (1998) data set on trochilids yields the following al- lometric relationships between K (y) and body mass (x): y = -0.154x-0.382 (n = 10; r~ = 0.06; P = 0.51 for all species) and y = -0.360x-0.281 (n = S; P = 0.33; P - 0.14 excluding the Nearctic species). A slight in- crease in data quality due to increased sample size is apparent in case of the joint analysis of all species together (P — 0.18 vs. = 0.06). Our value for the slope of the allometric relationship of —0.366 (excluding Nearctic species) closely resembles the overall value reported for 557 bird species of —0.316 (Starck and Ricklefs 1998), but is consider- ably lower than those reported by the same authors for other altricial taxa (Columbifor- mes = —0.26, Psittaciformes = —0.18, Pici- formes = —0.17, Passeriformes = —0.15). A separate analytical treatment of the Neotropic versus Nearctic hummingbird species resulted in a significant improvement of the fit of the regression lines. Visual inspection suggests that higher A'-values occur in Nearctic hum- m.ingbirds compared to the Neotropic species. These findings support the hypothesis by Ricklefs (1968, 1976) that tropical passerines grow more slowly than passerines from tem- perate regions is also valid for hummingbirds. Questions concerning the ultimate mecha- nisms underlying this observation remain un- answered (e.g., different predation pressures on temperate vs. tropical hummingbirds). Ricklefs (1968, 1976) discussed the factors of different day length and basal metabolic rate. While the first model holds for tropical versus temperate taxa in Ricklefs’ data set, it could not be extended to the comparison with arctic bird species. However, Ricklefs found that basal metabolism was 25% lower in tropical taxa compared to the studied temperate spe- cies. We note that our mean ^-value for Ne- arctic hummingbirds (x = 0.422; n = 2) close- ly resembles that for some nectarivorous sun- birds (Nectariniidae) (x = 0.417; n = 3) (Earle 1982, Goldstein and Yom-Tov 1988, Maher 1991, reviewed in Starck and Ricklefs 1998), while that of the neotropical trochilids (x = 0.269; n — 20) does not. This study provides important, new infor- mation on postnatal growth rates of hum- mingbirds. There is an obvious lack of data on this topic and we suggest not only a revival of allometric scaling studies, but a revival of studies (e.g., Fierro-Calderdn and Martin 2007) collecting basic life history information on postnatal growth rates of birds, especially of tropical taxa. ACKNOWLEDGMENTS We are indebted to C. E. Braun, Roland Prinzinger, and two anonymous referees for their insightful criti- cism and careful reviews of an earlier version of the manuscript. BPF is grateful to Sara Nelly de Visser for encouragement, to Stefanie Rick, ZFMK, Ornitholog- ical Library, for assistance with acquisition of litera- ture, and to the German National Merit Foundation (Studienstiftung des Deutschen Volkes) and e. fellows. net for financial support. KLS acknowledges the Brehm Fund for International Bird Conservation for funding Julia Beintmann’s thesis and a grant from the German National Council to conduct field work in Ecuador (Schu/766/5-3, 5-4, 5-5). LITERATURE CITED Beintmann, J. 1998. Vergleichende Reproduktions- biologie dreier Kolibriarten (Trochilidae) Lesbia victoriae, Metallura tyrianthina und Colibri co- ruscans in den Anden Ecuadors. Thesis. Rheinis- che Friedrich-Wilhelms-Universitat, Bonn, Ger- many. Brown, J. H., J. F. Gillooly, A. P. Allen, V. M. Sav- SHORT COMMUNICATIONS 887 AGE, AND G. B. West. 2004. Toward a metabolic theory of ecology. Ecology 85:1771-1789. CoNSTANTZ, G. D. 1980. Growth of nestling Rufous Hummingbirds. Auk 97:622-624. Dorst, J. 1962. Nouvelles recherches biologiques sur le trochilides des hautes andes Peruviennes {Or- eotrochilus estella). L’Oiseau 32:95-126. Earle, R. A. 1982. Aspects of the breeding biology of the Whitebellied Sunbird. Ostrich 53:65-73. Fierro-Calderon, K. and T. E. Martin. 2007. Repro- ductive biology of the Violet-chested Humming- bird in Venezuela and comparisons with other tropical and temperate hummingbirds. Condor 109:680-685. Gillooly, J. E, J. H. Brown, G. B. West, V. M. Sav- age, AND E. L. Charnov. 2001. Effects of size and temperature on metabolic rate. Science 293: 2248-2251. Gillooly, J. E, E. L. Charnov, G. B. West, V. M. Savage, and J. H. Brown. 2002. Effects of size and temperature on developmental time. Nature 417:70-73. Goldstein, H. and Y. Yom-Tov. 1988. Breeding bi- ology of the Orange-tufted Sunbird in Israel. Ar- dea 76:169-174. Haverschmidt, F. 1952. Notes on the life history of Amazilia fimbriata in Surinam. Wilson Bulletin 64:69-79. Maher, W. J. 1991. Growth and development of the Yellowbellied Sunbird, Nectarinia jiigularis, in North Queensland. Emu 91:58-61. Marin, M. 2001. Postnatal development of the Violet Sabrewing in Costa Rica. Wilson Bulletin 113: 110-114. Muir, A. 1925. The nesting of the Emerald Hum- mingbird (Saucerottia tobaci erythronota) in Trin- idad. Ibis 67:648-654. Ricklefs, R. E. 1968. Patterns of growth in birds. Ibis 110:419-451. Ricklefs, R. E. 1976. Growth rates of birds in the humid New World tropics. Ibis 118:179-207. ScHUCHMANN, K.-L. 1985. Morpho- und Thermoge- nese nestjunger Blaukehlkolibris {Lampornis cle- menciae). Journal fiir Ornithologie 126:305-308. ScHUCHMANN, K.-L. 1986. Natal care and growth in a nestling Reddish Hermit {Phaethornis ruber) in Surinam. Ardea 74:101-104. SCHUCHMANN, K.-L. 1999. Family Trochilidae (Hum- mingbirds). Pages 468-680 in Handbook of the birds of the world. Volume 5. Barn-owls to hum- mingbirds (J. del Hoyo, A. Elliot, and J. Sargatal, Editors). Lynx Edicions, Barcelona, Spain. Starck, j. M. and R. E. Ricklefs. 1998. Avian growth rate data set. Pages 381-423 in Avian growth and development (J. M. Starck and R. E. Ricklefs, Ed- itors). Oxford University Press, New York, USA. StatSoft Inc. 2003. STATISTICA. Version 6.1. StatSoft Inc., Tulsa, Oklahoma, USA. Thomas, B. T. 1994. Blue-tailed Emerald Humming- bird {Chlorostilbon mellisugus) nesting and nest- ling development. Ornitologia Neotropical 5:57- 60. ZucHNER, T. 1998. Reproductive patterns of two hum- mingbird species at high elevation in the Vene- zuelan Andes. Ostrich 69:341. The Wilson Journal of Ornithology 120(4):887-890, 2008 Begging Behavior of Fledgling Rusty-breasted Cuckoo (Cacomantis sepulcralis) Tomas Grim^ ABSTRACT. — I describe previously unknown beg- ging calls and displays of a fledgling Rusty-breasted Cuckoo (Cacomantis sepulcralis) fed by a Pied Fantail (Rhipidura javanica) in Singapore. The cuckoo emit- ted two types of begging calls: (1) ‘host-absent beg- ging call’ (loud ‘tsi’ repeated at 1-sec intervals) and (2) ‘standard’ begging call in the presence of the Pied Fantail (wheezy ‘seeee’ repeated 1-2 times/sec). The fledgling also performed the ‘wing-shake begging’ dis- play, i.e., it raised one of its wings at a time towards the approaching Pied Fantail. This display was similar ' Department of Zoology, Palacky University, tr. Svobody 26, CZ-771 46 Olomouc, Czech Republic; e-mail: tomas.grim@upol.cz to that of the best studied brood parasite, the Common Cuckoo (Cuculus canorus). The structure of both types of begging calls of the Rusty-breasted Cuckoo was dif- ferent in comparison to the Common Cuckoo and rel- atively more similar to some other closely related spe- cies of the genus Cacomantis. Received 2 October 2007. Accepted 9 February 2008. The biology and ecology of brood parasitic cuckoos has received major attention during the last two decades (Davies 2()()0). However, most data are from only one species, the Com- mon Cuckoo (Cuculii.s canoni.s). The knowl- 888 THE WILSON JOURNAL OF ORNITHOLOGY • VoL 120. No. 4. December 2008 (A) (B) FIG. 1 . Sonagrams of begging calls of fledgling Rusty-breasted Cuckoo in the absence of the Pied Fantail (A) and during its presence (B). The sonagrams were created using AVISOFT software and the background noise (cicada calls) was cleaned. edge of even the general biology of other spe- cies is extremely poor, especially at the chick and fledgling stages (Grim 2006, 2007). This is true for most tropical brood parasites and in agreement with the poorly known biology of tropical birds in general (Martin 1996). Therefore, I report an observation of previ- ously undescribed begging calls and begging behavior of a tropical cuckoo. OBSERVATIONS I observed a Pied Fantail (Rhipidiira javan- ica) feeding a fledgling Rusty-breasted Cuck- oo {Cacomantis sepulcralis) on 11 August 2007 in Sungei Buloh Wetland Reserve, Sin- gapore. I localized the fledgling in the man- grove {Rhizophora sp., Avicemiia sp.) habitat by its loud begging call. When the Pied Fan- tail was not present the cuckoo emitted the ‘host-absent begging call’. This was a rela- tively loud ‘tsi’, about 0.1 sec in duration, 4- 7 kH, repeated at 1/sec intervals (Fig. 1 A left). At times the fledgling produced a doubled ‘tsi’ call (Fig. 1 A right). The fledgling dramatically changed the structure of calling when the Pied Fantail approached and it produced the ‘stan- dard begging call’. This was a wheezy ‘seeee’, about 0.3-0. 5 sec in duration at 6-7 kH. The cuckoo repeated the call at 1/sec intervals and increased in both rate (to 2 calls/sec) and fre- quency (to 7-8 kH) when the Pied Fantail was at close range (several cm from the chick) (Fig. IB). The Rusty-breasted Cuckoo fledgling sat on a branch 2 m above ground level and changed the perch only once during my observations. I observed the cuckoo chick from a distance of ~7 m for a period of —10 min during which it was fed five times by the Pied Fan- tail. The cuckoo started to beg at a faster rate when it observed the approaching fantail. It also showed the ‘wing-shake begging’. Dur- ing this display the chick raised its wing at an angle of —90° above horizontal and towards an approaching fantail. The display of the Rusty-breasted Cuckoo was asymmetric, i.e., the chick raised only one wing at the time. The fledgling in all five cases of feedings raised vertically the wing towards the ap- proaching fantail (right wing 3 times, left SHORT COMMUNICATIONS 889 wing 2 times). In one case the fantail started to approach from the right side of the chick but finally arrived at the left side; the cuckoo synchronously changed the raised right wing for the raised left one. The chick stopped the wing-shake begging display when the fantail left the area. I saw only one fantail at the time but cannot exclude the possibility that two birds actually fed the chick. I was unable to identify the prey fed by the fantail to the cuck- oo. DISCUSSION Cuckoo fledglings are occasionally fed by adults that did not raise them or even by spe- cies that do not raise cuckoos (Sealy and Lor- enzana 1997), but the Pied Fantail was pre- viously reported as a regular host of the Rus- ty-breasted Cuckoo in Java (Payne 2005). However, the hosts of this cuckoo in Malay Peninsula were ‘not identified’ (Wells 1999: 387). Some 60 passerine species have been recorded as hosts of the Rusty-breasted Cuck- oo (Payne 2005). The information on Rusty-breasted Cuckoo breeding biology is sparse. Johnsgard (1997: 215) reports ‘No information’ and Payne (1997) provides no data on cuckoo behavior in the postfledging period. Wells (1999: 387) reports ‘No information’ on the breeding be- havior of the Rusty-breasted Cuckoo while Brooker and Brooker (1989) mention the fledging period is —19 days. Payne (2005: 447) provides more details; he reports that one of three egg morphs of the Rusty-breasted Cuckoo mimics eggs of fantails (Rhipidura spp.), the nestling cuckoo evicts host eggs and young, the nestling period is 17-19 days, and the length of the post-fledging care is 1 month. The compendium by Payne (2005) de- scribes begging behavior by other cuckoo spe- cies, but mentions only a few records of Rus- ty-breasted Cuckoo fledglings and virtually no information on their calls or behavior. Higgins (1999: 682) reports ‘no information on calls of nestlings’ in the closely related Brush Cuckoo (Ccicomantis variolosus). Brooker and Brooker (1989), and Payne (1997, 2005) treat- ed the Rusty-breasted Cuckoo as a subspecies of the Brush Cuckoo. There is some information on begging calls of two closely related cuckoo congeneric spe- cies. The begging call of a juvenile Chestnut- breasted Cuckoo (C. castaneiventris) is ‘a re- peated high-pitched thin wheezy siiiaar-swee- sweep" (Higgins 1999: 690). Fully fledged young Fan-tailed Cuckoos (C. flabelliformis) begged from a host with ‘plaintive almost ci- cada-like zeep-zeep-zeep’ (Higgins 1999: 698). These descriptions seem similar to the ‘standard begging call’ of the Rusty-breasted Cuckoo fledgling (wheezy ‘seeee’). However, I have no information on age, gender or hun- ger level of the cuckoo chicks and any com- parisons are only preliminary. It would be in- teresting to compare the fledgling’s call with that of its putative fosterer’s own chicks but I was unable to find any information on beg- ging calls of Pied Fantail nestlings (Boles 2006, Wells 2007). The Rusty-breasted Cuckoo fledgling ut- tered a different call when the Pied Fantail was absent (‘tsi’; Fig. lA). I found only one description of a ‘host-absent begging call’ in the literature for cuckoos. Sicha et al. (2007) described the ‘host-absent vocalization’ of the Common Cuckoo as distinct ‘si’ sounds re- peated at intervals of 0.5-5 sec (Sicha et al. 2007: Fig. 1). The host-absent begging call of the Common Cuckoo has a higher frequency (7-8 kH) than that of the Rusty-breasted Cuckoo (4-7 kH) but sounds relatively similar to the human ear (pers. obs. of both cuckoo species). In contrast, the ‘standard begging call’ (in the presence of fosterers) of the Com- mon Cuckoo chick spans a much wider range of frequencies (5-10 kH) than that of the Rus- ty-breasted Cuckoo (6-7 kH) and has a strik- ingly higher rate (10-20 vs. 1-2 calls/sec). There may be important species differences in the call structure between the two taxa and I acknowledge that differences may arise from factors unrelated to species identity (e.g., age, developmental stage, actual hunger level). The Rusty-breasted Cuckoo fledgling also performed a ‘wing-shake begging’ display. This begging strategy seems to be universal in birds (Grim 2008) and the unusual aspect of this behavior in the observed cuckoo spe- cies was that the display was asymmetric (i.e., only one wing was raised at a time). Asym- metric wing shaking has also been observed in other cuckoo species (Tanaka and Ueda 2005, Tanaka et al. 2005, Grim 2008). In con- trast, passerine chicks raise both wings at a time as a rule (Grim 2008). 890 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 Observations on behavior are valuable and indispensable for understanding the basic breeding biology of cuckoos and brood para- site-host interactions. Additional information will be valuable to facilitate cross-species comparisons. I encourage ornithologists and birdwatchers that may have observations on both chicks and fledglings of non-European cuckoos to publish their observations. ACKNOWLEDGMENTS I am grateful to Vaclav Pavel for preparing the sono- grams. I thank Human Frontier Science Program Or- ganization for supporting my brood parasitism research (grant HFSP RGY69/2007) and to Dana Campbell for correcting the language. LITERATURE CITED Boles, W. E. 2006. Family Rhipiduridae (fantails). Pag- es 200-242 in Handbook of the birds of the world. Volume 1 1 . Old World flycatchers to Old World warblers (J. del Hoyo, A. Elliott, and D. A. Chris- tie, Editors). Lynx Edicions, Barcelona, Spain. Brooker, M. G. and L. C. Brooker. 1989. Cuckoo hosts in Australia. Australian Zoological Reviews 2:1-67. Davies, N. B. 2000. Cuckoos, cowbirds and other cheats. T & A. D. Poyser, London, United King- dom. Grim, T. 2006. The evolution of nestling discrimina- tion by hosts of parasitic birds: why is rejection so rare? Evolutionary Ecology Research 8:785- 802. Grim, T. 2007. Equal rights for chick brood parasites. Annales Zoologici Fennici 44:1-7. Grim, T. 2008. Wing-shaking and wing-patch as nest- ling begging strategies: their importance and evo- lutionary origins. Journal of Ethology 26:9-15. Higgins, P. J. (Editor). 1999. Handbook of Australian, New Zealand and Antarctic birds. Volume 4 (Par- rots to Dollarbird). Oxford University Press, Mel- bourne, Australia. JoHNSGARD, P. A. 1997. The avian brood parasites. Ox- ford University Press, New York, USA. Martin, T. E. 1996. Life history evolution in tropical and south temperate birds: what do we really know? Journal of Avian Biology 27:263-272. Payne, R. B. 1997. Family Cuculidae (cuckoos). Pages 508-607 in Handbook of the birds of the world. Volume 4. Sandgrouse to cuckoos (J. del Hoyo, A. Elliott, and J. Sargatal, Editors). Lynx Edi- cions, Barcelona, Spain. Payne, R. B. 2005. The cuckoos. Oxford University Press, New York, USA. Sealy, S. G. and j. C. Lorenzana. 1997. Feeding of nestling and fledgling brood parasites by individ- uals other than the foster parents: a review. Ca- nadian Journal of Zoology 75:1739-1752. SiCHA, V, P. Prochazka, and M. Honza. 2007. Hope- less solicitation? Host-absent vocalization in the Common Cuckoo has no effect on feeding rate of Reed Warblers. Journal of Ethology 25:147-152. Tanaka, K. D. and K. Ueda. 2005. Horsfield’s Hawk- Cuckoo nestlings simulate multiple gapes for beg- ging. Science 308:653-653. Tanaka, K. D., G. Morimoto, and K. Ueda. 2005. Yellow wing-patch of a nestling Horsfield’s Hawk Cuckoo Cuciilus fugax induces miscognition by hosts: mimicking a gape? Journal of Avian Biol- ogy 36:461-464. Wells, D. R. 1999. The birds of the Thai-Malay Pen- insula. Volume I. Non-passerines. Christopher Helm, London, United Kingdom. Wells, D. R. 2007. The birds of the Thai-Malay Pen- insula. Volume II. Passerines. Christopher Helm, London, United Kingdom. SHORT COMMUNICATIONS 891 The Wilson Journal of Ornithology 120(4):89 1-897, 2008 Natural History of the Red Owl {Tyto soumagnei) in Dry Deciduous Tropical Forest in Madagascar Scott G. Cardiff -*'^ and Steven M. Goodman^ ABSTRACT — Recent observations of the Red Owl {Tyto soumagnei) in Madagascar demonstrated that it inhabits dry deciduous forest, and roosts on rock ledg- es and in cave entrances in the extreme north of the island. We observed a Red Owl at a sinkhole site in the Reserve Speciale d’Ankarana, found evidence of its use of an additional cave, and collected its pellets in three separate dry seasons between 2000 and 2003. Tsingy tufted-tailed rats (Eliurus antsingy) constituted almost 50% of the total prey mass of Red Owls at Ankarana. Their diet at Ankarana differed from that of Red Owls from Masoala in the humid northeast of Madagascar, as the Ankarana pellets contained insects, frogs, and numerous geckos. Red Owls appear to con- sume more native than introduced rodents and do not appear to prey upon birds or bats like other large owls on the island. Forest degradation could reduce densi- ties of tufted-tailed rats and could be a conservation threat to this owl. Received 30 May 2006. Accepted 7 February 2008. The Red Owl {Tyto soumagnei) is rare, en- demic to Madagascar and was not observed by biologists for several decades (Thorstrom and Rene de Roland 2003). It is considered “Endangered” (lUCN 2006) and was thought to occur only in humid forests in the east (Grandidier 1878, Halleux and Goodman 1994, Goodman et al. 1996, Thorstrom et al. 1997, ZICOMA 1999, Irwin and Samonds 2002). Limited diet studies suggest the owl feeds on frogs, geckos, afrosoricidans {Micro- gale spp., Oryzorictes hova), rodents {Eliurus ' Columbia University, Department of Ecology, Evolution, and Environmental Biology, 1200 Amster- dam Avenue, New York, NY 10027, USA, and Divi- sion of Vertebrate Zoology (Mammalogy), American Museum of Natural History, Central Park West at 79th Street, New York, NY 10024, USA. 2 Field Museum of Natural History, 1400 South Lake Shore Drive, Chicago, IL 60605, USA, and Va- hatra, BP 3972, Antananarivo (101), Madagascar. ^Current address: 1331 Pinewood Drive, Pittsburgh, PA 15243, USA. Corresponding author; e-mail; sgc2 102@columbia.edu spp., Rattiis rattus), and the eastern rufous mouse lemur {Microcebus rufus) (Lavauden 1932, Halleux and Goodman 1994, Goodman and Thorstrom 1998). These different prey an- imals are native to Madagascar with the ex- ception of the roof rat {Rattus rattus). Until recently, observers have noted owls roosting only in trees in intact and degraded humid for- est (Thorstrom and Rene de Roland 1997, Thorstrom et al. 1997, Irwin and Samonds 2002). Cardiff and Befourouack (2003) and Van Esbroeck (2006) extended the known range of the owl to Ankarana in the north of the island. Van Esbroeck (2006) observed Red Owls at Ankarana in 1999; that observation represents the first record of the occurrence of this owl in dry deciduous forest and the first record of it roosting in a cave entrance or rocky crevice. Our observation of a Red Owl at Ankarana in 2000 (Cardiff and Befourouack 2003) prompt- ed us to obtain further information on the dis- tribution, diet, and roosting habits of this poorly known species and its congener, the Barn Owl {T. alba) at Ankarana. METHODS The Reserve Speciale d’Ankarana, —60 km south of Antsiranana in northern Madagascar (Fig. 1) (12° 49' to 13°01'S and 49° 00' to 49° 16' E), encompasses a massif of Jurassic limestone with dry deciduous forest partially surrounded by and intruded upon by semi-de- ciduous forest (Cardiff and Befourouack 2003). Precipitation near the reserve averages 180-200 cm per year (Rossi 1976, Hawkins et al. 1990). The local limestone karst features many caves and several river sinkholes (Ra- dofilao 1977). These include, on the east side of the reserve, a large collapse sinkhole (col- lapse doline), where Van Esbroeck (2006) ob- served a Red Owl in 1999 at the Perte des Riveres, which is also known as the Perte de la Besaboba Number 1 (Radofilao 1977; J. 892 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 FIG. 1. Recorded distribution of Tyto soiimagnei on Madagascar and in Ankarana, and distribution of Tyto alba at Ankarana. SHORT COMMUNICATIONS 893 Radofilao, unpubl. data) and “Perte des trois rivieres” (Van Esbroeck 2006). Savannah oc- curs within —200 m and semi-deciduous for- est on basalt occurs within 500 m of this sink- hole site. We recorded the presence or absence of all owl species seen or heard at 28 caves from 2003 to 2006 in and around Ankarana during cave biological inventory and monitoring work (Cardiff 2006). We made additional ob- servations at the Perte des Rivieres site in the 2000, 2001, and 2003 dry seasons between June and November. We noted the presence of pellets and checked potential roost positions for owls on several occasions to understand roost characteristics and periodicity of roost occupation in 2000 and 2001, and also col- lected pellets in 2003. We collected >45 pel- lets and pellet fragments at this site over the three dry seasons. We assumed that all pellets found at the site were from Red Owls based on the concurrent observation of an individual present or feathers attributable to the species on two of three collection dates in 2000 and on two of five dates in 2001. We also collected broken pellets at the entrances of the Grotte d’ Antsironandoha on the western side of An- karana. We attributed pellets collected at Grot- te d’ Antsironandoha to Barn Owls because we observed a Barn Owl at this location. We did not conduct formal inventories for owls in for- ests away from cave entrances. We measured the dimensions of several intact pellets and identihed prey items in all collected material. We washed the pellet contents through a sieve after soaking pellets in soapy water and identified recovered hard items to species or species group using collections material at the Field Museum of Natural History in Chicago. We counted the minimum number of individuals (MNI) by the number of unique bilateral ele- ments for each species or species group (Good- man et al. 1991), and calculated relative contri- bution to total number of individuals and rela- tive contribution to total prey mass based on reference prey masses (Marti 1987). RESULTS Distribution. — We observed Tyto spp. at live ot 28 caves. Three of those observations were of a Barn Owl and one was not identified to species. We only observed Red Owls at the Perte des Rivieres, but we also observed and photographed a feather of a Red Owl in 2006 at the entrance to the Grotte des Chauves- souris (“Andavampanihy”), a large cave 1 km west of the Perte des Rivieres. Roosting Habits. — On 31 July 2000 at 0800 hrs, SGC observed one Red Owl on a ledge in a crack in the limestone below the overhang of the edge of the Perte des Rivieres sinkhole. SGC again observed one individual at the same location on 6 August 2000 and again on 3 September 2000. SGC subsequently ob- served an owl at the site on 1 July 2001 and 5 August 2001 but on different rock ledges. Feathers of Red Owls were also present with collected pellets on 12 November 2000. We did not detect the owl at this site during 16 additional visits in 2000-2001. The Red Owl was roosting in the general area of the opening to the lower passage in the sinkhole in all observations. The owl was in a fissure about 0.5 X 0.5 m in the ceiling (—8 m from the floor) in front of the entrance to the horizontal passage under the overhang when observed in 2000. Guides had report- edly previously seen two Red Owls together at this location. The owl was roosting on the first occasion in 2001 on a ledge about 3 m above the cave floor and 5 m into the hori- zontal passage. On another occasion, it was roosting in a small opening on the side of the sinkhole at the entrance to the horizontal pas- sage, — 1.5 m above the floor. Food Habits. — Four intact Red Owl pellets measured 35 X 24, 40 X 24, 41 X 32, and 58 X 29 mm (length X width). Pellets contained a minimum number of individuals (MNI) of 172 insects, frogs, geckos, and mammals (Ta- ble 1 ). Mammals constituted the greatest pro- portion of the prey items by individual and by mass for all years with geckos being important in 2()()0. Frogs, a primate, and insects were present in the pellets in some years but not in others (Table 1 ). The Madagascan pygmy shrew (Simciis inadagascariensis) contributed most to the MNI in 2001 and the roof rat con- tributed up to 27% of the total prey biomass in a given year. Tsingy tufted-tailed rats (Fli- urus an tsingy) contributed the most of any single prey item to prey mass in all years and to prey total MNI in 2000 and 2003 (Table 1). Barn Owl Pellets from Grotte d' Antsironandoha contained frogs (MNI = 4), birds (Turni.x nigricollis, MNI — I; Foitdia 894 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 TABLE 1. Relative importance of tsingy tufte(J-tailed rats as prey items of one or more Red Owls at Ankarana, Madagascar based on regurgitated pellet samples from three different years. Possible frog and gecko prey species identification is based on records of herpetological collections (A. Raselimanana and M. Razafim- pahanana, unpubl. data; C. Raxworthy, unpubl. data). PN = number of entire pellets (* additional fragments also examined). MNI = Minimum Number of Individuals. Prey item Mass (g) Collection year (and months), collection totals, and percent of totals for prey items 2000 (Sep-Nov) PN = 8* MNI = 30 2001 (Jun-Sep) PN = 25 MNI = 90 2003 (Jun- PN = MNI ^ Jul, Nov) 12* = 52 All years combined PN = 45* MNI = 172 MNI (%) Mass (%) MNI (%) Mass (%) MNI (%) Mass (%) MNI (%) Mass (%) Insecta 8.9 0.8 5.8 0.8 6.4 0.6 Coleoptera ^4h 2.2 0.3 3.8 0.4 2.3 0.2 Dictyoptera: Blaberidae^* 7.5' 1.9 0.4 0.6 0.1 Orthoptera -2.5^ 6.7 0.5 3.5 0.2 Amphibia: Anura 17.3 2.8 5.2 0.8 Large^ -9.5J 5.8 1.5 1.7 0.4 Medium‘s -4.5^^ 7.7 1.0 2.3 0.3 SmalP ~3.0) 3.8 0.3 1.2 0.1 Squamata: Gekkonidae 19.8 24.4 24.3 28.8 26.7 27.9 23.8 Large‘S ~50' 10,0 8.4 7.8 11.7 7.7 10.6 8.1 10.5 Medium*^ -29- 20.0 9.7 12.2 10.6 19.2 15.3 15.7 1 1.7 SmalP ~15' 6.7 1.7 4.4 2.0 1.9 0.8 4.1 1.6 Mammalia 63.3 80.2 66.7 74.9 48.1 69.7 60.5 74.9 Soricomorpha 10.0 0.4 36.7 2.5 17.3 1.5 26.2 1.7 Simcus madagascariensis 2.3*^ 10.0 0.4 36.7 2.5 15.4 1.0 25.6 1.5 Afrosoricida 1.9 0.6 0.6 0.2 Microgale brevicaudata 10.7' 1.9 0.6 0.6 0.2 Rodentia 53.3 79.9 30.0 72.4 25.0 58.5 32.6 70.5 Eliurus antsingy 87.5' 46.7 68.4 16.7 43.7 17.3 41.7 22.1 49.8 Mus muscidiis 9.7^ 4.4 1.3 1.9 0.5 2.9 0.7 Rattus rattus 102.7^ 6.7 1 1.5 8.9 27.4 5.8 16.3 7.6 20.0 Primata 5.8 9.7 1.7 2.7 Microcebus tavaratra 61. U 5.8 9.7 1.7 2.7 ^ Gromphadorlu?ia sp. cf. Laliostoma sp./Mantidactylus sppJAglyptodactylus sp. cf. Boophis sp./Mantidactylus spp./ Ptychadena madagascariensis. ManteUa spp./Anodontohyla sp./Cop/iyla sp./Stumpffta sp. cf. Homopholis boivini/Uroplatus fimbriatuslU. henkeli. Uf. Uroplatus sikoraelParoedura homalorhina/P. stumpffi. 8 cf. U. ebenaui/P. karstophila/P. oviceps/Geckolepis spp. /Hemidactylus spp. Estimated. ' Clarke and Moore ( 1995). J Estimated (Goodman et al. 1993b). ^ Goodman et al. (1993b). ' SMG, unpubl. data. "’Goodman et al. (1991). " Rasoloarison et al. (2000). madagascariensis, MNI = 1; Merops super- ciliosus, MNI = 1 ), afrosoricidans and sori- comorphans {Setifer setosus, MNI = I; Sun- cus madagascariensis, MNI = 1), rodents (Mus nmscidus, MNI = 8; Rattiis rattus, MNI = 28), and a northern rufous mouse lemur (Microcebiis tavaratra, MNI = 1 ). DISCUSSION This work provides new information on the habits of Red Owls in a forest biome in which it was only recently found to occur. Our End- ings add to knowledge of the distribution, roosting habits, and diet of this rare owl, which has implications for their conservation at Ankarana. The Ankarana record is the only published account of Red Owls using a cave (Van Es- broek 2006). However, the occurrence of the owl at a different nearby cave suggests it is not an unusual roost site. The continued pres- ence of at least one individual at the site in the dry season of several successive years (1999, 2000, 2001, 2003) suggests the use of SHORT COMMUNICATIONS 895 the first site is not incidental. An additional unpublished record of a Red Owl using a cave entrance for roosting elsewhere on the island suggests these owls use caves for roosting more often than previously thought. In 2003, A. Raselimanana and SMG found a single Red Owl roosting in a shallow cave in the Mt. Papango portion of the Parc National de Mi- dongy-Sud (Fig. 1) in an extensive area of hu- mid forest in the southeastern portion of the island. This record is important because it ex- tends the known range of the species to the south (Irwin and Samonds 2002) and shows that it uses caves for roosting in the eastern portion of its range. Use of cave entrances suggests a large number of potential roost sites could be available for Red Owls at An- karana, a zone of extensive karst, and that Red Owls could occur in other areas with caves or rock crevices. However, Barn Owls appear to use cave entrances more commonly than Red Owls at Ankarana based on our observations. Our data also expands knowledge of food habits of Red Owls. Early observations sug- gested this species fed extensively on frogs (Lavauden 1932) and a captive individual fed on frogs (Halleux and Goodman 1994). No anurans occurred in pellets from the Masoala Peninsula (Goodman and Thorstrom 1998). Frogs were present in the Ankarana pellets even though frogs are more abundant on the Masoala Peninsula than at the notably drier Ankarana. Goodman and Thorstrom (1998) found no insects, no Siinciis spp., and few geckos in pellets from the Masoala Peninsula. We found some insects and many Madagascan pygmy shrews and geckos in pellets from An- karana. Mean prey weight was lower at An- karana than on the Masoala Peninsula due to the presence of insects, frogs, and smaller- bodied afrosoricidans in the pellets. The lack of bats or birds, and the importance of tufted- tailed rats and native mammals in the diet of Red Owls remained consistent between An- karana and Masoala, and appear to distinguish the dietary regime of Red Owls from that of Barn Owls, including at Ankarana, and the Madagascar Owl {Asia nuulasfciscariensis) on the island (Goodman and Langrand 1993; Goodman and Thorstrom 1998; Goodman et al. 1993a,b; Goodman and Griffiths 2006; Ra- soma and Goodman 2007; this study). The ab- sence of birds or bats in the diet of Red Owls is notable given the presence of at least one species of bat in the cave where the owl roost- ed at Ankarana (Cardiff 2006), and the pres- ence of birds and bats (Hipposideros com- mersoni, Coleiira afra) (Goodman et al. 2008) in the diet of Barn Owls at Ankarana. Unlike diet, pellet dimensions do not seem to distin- guish Red Owls at Ankarana (35-58 X 24- 32 mm, n = 4) from Barn Owls at other sites on Madagascar (37-62 X 21-33 mm, n = 23) (Goodman et al. 1993b) and identification of pellet species origin by size alone is not ap- propriate. CONSERVATION IMPLICATIONS Habitat degradation may threaten Red Owls at Ankarana. This species roosts and feeds in degraded areas on the Masoala Peninsula; the roost at Ankarana is a rocky area surrounded by forest edge not far from degraded forest and savannah. The owls at both sites appear to feed primarily on tufted-tailed rats. Tufted- tailed rats often climb vegetation and trees, do not seem to occur in savannahs, and rarely occur in areas with severely degraded vege- tation (Carleton 2003). Ongoing degradation of forests at Ankarana may reduce availability of tsingy tufted-tailed rats as a food source for Red Owls. Forest degradation could also re- duce availability of large trees with cavities that remain the only confirmed nest site for this owl (Thorstrom and Rene de Roland 1997), although it is likely that Red Owls nest in rock cavities since it is now known to roost in such cavities. Additional conservation concerns exist for the Perte des Rivieres site at Ankarana. Sed- iments and debris borne by floodwaters during the wet season, most likely resulting from cy- clones and upstream deforestation, have pre- viously blocked most of the length of hori- zontal cave passage at the roost site (S. G. Cardiff, unpubl. data). The potential effect of blocking of the cave by sediment and debris on the suitability of the site for roosting by Red Owls is unknown. The site is also acces- sible to tourists and, although descending into the sinkhole is dangerous without proper equipment and expertise, some tourist visits to the bottom occur. Repeated disturbances by tourists or others could possibly cause Red Owls to abandon use of the site. We encour- age other studies to examine potential threats 896 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 to conservation of Red Owls and interactions between Red Owls and Bam Owls. ACKNOWLEDGMENTS We thank the Association National pour la Gestion des Aires Protegees (ANGAP) and the Ministre de I’Environnement, des Eaux et Forets (MENVEF) for permission to conduct research at Ankarana. We also extend our gratitude to ANGAP Ankarana for support with logistics. This manuscript benefited from com- ments by several anonymous reviewers. Mohamed, Sa- bry Mamilaza, and other Ankarana guides initially in- formed us of the presence of Red Owls and we could not have conducted this study without them. This study is dedicated to the memory of Sabry Mamilaza, who died in 2001. We are also grateful to Lisa Stano, who assisted with collection of owl pellets. We conducted this study under the protocol of collaboration between the Departement de Biologie Animale at the Universite d’ Antananarivo, WWF-Madagascar, and the Field Mu- seum of Natural History. The Volkswagen Foundation and the National Geographic Society (6637-99 and 7402-03) provided funding for inventories in 2001- 2003. The U.S. Peace Corps provided support to SGC in 2000-2002. This material is based upon work sup- ported under a National Science Foundation Graduate Research Fellowship to SGC. We thank Russell Thor- strom, Lily Arison Rene de Roland, and Lucienne Wil- me for providing additional information on owl and prey species. LITERATURE CITED Cardiff, S. G. 2006. Bat cave selection and conser- vation in Ankarana, northern Madagascar. Thesis. Columbia University, New York, USA. Cardiff, S. G. and J. Befourouack. 2003. The Re- serve Speciale d’ Ankarana. Pages 1501-1507 in The natural history of Madagascar (S. M. Good- man and J. P. Benstead, Editors). University of Chicago Press, Chicago, Illinois, USA. Carleton, M. D. 2003. Eliunis, tufted-tailed rats. Pages 1373-1380 in The natural history of Madagascar (S. M. Goodman and J. P. Benstead. Editors). University of Chicago Press, Chicago, Illinois, USA. Clarke, D. C. and A. J. Moore. 1995. Variation and repeatability of male agonistic hiss characteristics and their relationship to social rank in Grompha- dorhina portentosa. Animal Behaviour 50:719— 729. Goodman, S. M. and O. Grieeiths. 2006. A case of exceptionally high predation levels of Rousettiis madagascariensis by Tyto alba (Aves: Tytonidae) in western Madagascar. Acta Chiropterologica 8: 553-556. Goodman, S. M. and O. Langrand. 1993. Food habits of the Bam Owl Tyto alba and the Madagascar Long-eai’ed Owl Asia madaga.scariensis on Mada- gascar: adaptation to a changing environment. Mu- see Royal de I’Afrique Centrale Tervuren, Belgique, Annales Sciences Zoologiques 268:147-153. Goodman, S. M. and R. Thorstrom. 1998. The diet of the Madagascar Red Owl {Tyto soumagnei) on the Masoala Peninsula, Madagascar. Wilson Bul- letin 110:417-421. Goodman, S. M., S. G. Cardiee, and F. H. Ratrimo- MANARivo. 2008. First record of Coleiira (Embal- lonuridae) on Madagascar and identification and diagnosis of members of the genus. Systematics and Biodiversity 6:283-292. Goodman, S. M., G. K. Creighton, and C. Raxwor- THY. 1991. The food habits of the Madagascar Long-eared Owl Asio madagascariensis in south- eastern Madagascar. Bonner zoologische Beitrage 42:21-26. Goodman, S. M., O. Langrand, and C. J. Raxwor- THY. 1993a. Food habits of the Madagascar Long- eared Owl Asio madagascariensis in two habitats in southern Madagascar. Ostrich 64:79-85. Goodman, S. M., O. Langrand, and C. J. Raxworthy. 1993b. The food habits of the Bam Owl Tyto alba at three sites on Madagascar. Ostrich 64:160-171. Goodman, S. M., A. Andrianarimisa, L. E. Olson, AND V. SoARiMALALA. 1996. Patterns of elevation- al distribution of birds and small mammals in the humid forests of Montague d’Ambre, Madagascar. Ecotropica 2:87-98. Grandidier, a. 1878. Note sur un nouveau Strigide de Madagascar. Bulletin de la Societe philomathique de Paris 7(2):65-66. Halleux, D. and S. M. Goodman. 1994. The redis- covery of the Madagascar Red Owl Tyto soumag- nei (Grandidier 1878) in north-eastern Madagas- car. Bird Conservation International 4:305-311. Hawkins, A. F. A., P. Chapman, J. U. Ganzhorn, Q. C. M. Bloxam, S. C. Barlow, and S. J. Tonge. 1990. Vertebrate conservation in Ankarana Spe- cial Reserve, northern Madagascar. Biological Conservation 54:83-110. Irwin, M. T. and K. E. Samonds. 2002. Range exten- sion of the Madagascar Red Owl Tyto soumagnei in Madagascar: the case of a rare, widespread spe- cies? Ibis 144:680—683. International Union eor the Conservation oe Na- ture (lUCN). 2006. lUCN Red List of threatened species. International Union for Conservation of Nature and Natural Resources Species Survival Commission, Cambridge, United Kingdom, http:/ /WWW. iucnredlist.org (accessed 17 May 2006). Lavauden, M. L. 1932. Etude d’une petite collection d’oiseaux de Madagascar. Bulletin du Museum national d’Histoire naturelle (2) 4:629-640. Marti, C. D. 1987. Raptor food habits studies. Pages 67-80 in Raptor management techniques manual (B. A. Giron Pendleton, B. A. Millsap, K. W. Cline, and D. M. Bird, Editors). National Wildlife Federation, Washington, D.C., USA. Radoeilao, j. 1977. Bilan des explorations speleolo- giques dans T Ankarana. Annales de I’Universite de Madagascar, serie Sciences de la Nature et Mathematiques 14:195-204. Rasoloarison, R., S. M. Goodman, and J. U. Gan- SHORT COMMUNICATIONS 897 ZHORN. 2000. Taxonomic revision of mouse le- murs {Microcehus) in the western portions of Madagascar. International Journal of Primatology 21:963-1019. Rasoma, J. and S. M. Goodman. 2007. Food habits of the Barn Owl (Tyto alba) in spiny bush habitat of arid southwestern Madagascar. Journal of Arid Environments 69:537-543. Rossi, G. 1976. Karst et dissolution des calcaires en milieu tropical. Zeitschrift fiir Geomorphologie, Supplementbande 26:124-152. Thorstrom, R. and L.-A. Rene de Roland. 1997. First nest record and nesting behaviour of the Madagascar Red Owl Tyto soiimagnei. Ostrich 68: 42-43. Thorstrom, R. and L.-A. Rene de Roland. 2003. Tyto soLimagnei, Madagascar Red Owl. Pages 1110-1113 in The natural history of Madagascar (S. M. Goodman and J. P. Benstead, Editors). Uni- versity of Chicago Press, Chicago, Illinois, USA. Thorstrom, R., J. Hart, and R. T. Watson. 1997. New record, ranging behaviour, vocalization and food of the Madagascar Red Owl Tyto soumagnei. Ibis 139:477-481. Van Esbroeck, J. 2006. Madagascar Red Owl Tyto sou- magnei in Ankarana Special Reserve, Madagascar. Bulletin of the African Bird Club 13:205-206. ZICOMA. 1999. Les Zones d’lmportance pour la Con- servation des Oiseaux a Madagascar. Projet ZI- COMA, Antananarivo, Madagascar. The Wilson Journal of Ornithology 120(4):897-900, 2008 Bird Responses to a Managed Forested Landscape Richard H. Yahner* ABSTRACT — I examined bird population respons- es to a managed forested landscape resulting from management for Ruffed Grouse (Bonasus umhellus) habitat in central Pennsylvania during three consecu- tive springs, 2005-2007. The number of bird species increased from 2001-2002 {n = 40) to 2005-2007 {n — 46). Abundance of all species combined declined (0.10 > P > 0.05), perhaps because the area was more heterogenous in 2001-2002 than in 2005-2007. Red- eyed Vireo {Vireo olivaceus) was the most common species in both 2001-2002 and 2005-2007. Six of the 20 common species were detected only in the treated sector in both periods; none was specific to the refer- ence sector. Despite increased forest maturation, no populations of early successional bird species declined {P < 0.05) between periods, but populations of three other species did. Management of the Barrens Grouse Habitat Management Area for Ruffed Grouse habitat did not have a profound effect on bird populations from 2001-2002 to 2005-2007 subsequent to the last cutting cycle. Received 16 January 2008. Accepted I May 2008. Long-term studies of the respon.ses of birds to managed landscapes are crucial to under- standing regional population trends (Gullion 1990; Yahner 2000, 2003). Thus, an important question is not necessarily can habitat be man- ' School of Forest Resources, Pennsylvania State University, 1 19 Forest Resources Building, University Park, PA, 16802, USA; e-mail: rhy@psu.edu aged for early successional bird species (Yah- ner 2003), but how soon after management must habitat be manipulated further to ensure adequate early successional habitat is avail- able due to loss via plant succession. The Barrens Grouse Habitat Management Area (GHMA) was created in 1976 for Ruffed Grouse {Bonasus umbellus) habitat via even- aged management (Yahner 1993). Habitat management consisted of the establishment of 4-ha “activity centers” for adult grouse (Gul- lion 1977), which resulted in a series of 1-ha contiguous plots of different ages in mixed- oak (Quercus spp.) and aspen (Populus spp.) cover types (Yahner 1993). Grouse habitat management provided valuable breeding hab- itat for bird species dependent on young for- ests. The objective of the present study was to compare community structure and composi- tion of breeding birds in the present (2005- 2007) 5 years after a fourth cutting cycle (in winters 1999-2000) at the Barrens GHMA to that just after the fourth cycle in 2001-2002) (Yahner 2003). METHODS Field work was conducted during three con- secutive springs 2005, 2006, and 2007 (late May-mid-Jun) at the Barrens GHMA, which consists of 1 , 1 20 ha on wState Game Lands 898 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 176, Centre County, in the Valley and Ridge Province of Pennsylvania (Yahner 1993) (40° 47' N, 78° 58' W). The Barrens GHMA is di- vided into a treated (managed) sector and a reference (unmanaged) sector of approximate- ly equal area. The treated sector was 136 contiguous blocks (60 in aspen and 76 in mixed-oak cov- er type); each block was 4 ha and was sub- divided into 1-ha plots (A-D, each —100 X 100 m) (Yahner 2003). The 60 blocks in aspen cover type were cut in four winter cutting cy- cles: 1976-77, plot A; 1980-81, plot B; 1986-87, plot C; and 1999-2000, plot D, for a 100% cut. Only plots A-C were cut in the 76 blocks of the mixed-oak type: 1976-77, plot A; 1980-81, plot B; and 1999-2000, plot C, for a 75% cut. Fifteen-20 residual oversto- ry trees (woody stem >7.5 cm dbh, >1.5-m tall) were retained per plot in plots C and D of the aspen type, and in plots B and C of the mixed-oak type. Overstory trees in the uncut plots (D in mixed oak) of the treated sector and in the reference (unmanaged) sector were —90-100 years old and consisted mainly of red maple {Acer ruhrum), white oak (Quercus alba), chestnut oak (Q. rnontana), northern red oak (Q. rubra), black cherry {Primus serotina), quaking aspen {Popiihis tremuloides), big- tooth aspen {P. grandidenta), and pitch pine {Pinus rigida) (Yahner 1993, 2003). Major un- derstory trees (woody stem 1—7.5 cm dbh, >1.5-m tall) were red maple, oak, black cher- ry, and aspen. I used the same 90 representative plots as in previous studies (Yahner 1993, 1997, 2003). Briefly, 10 plots (A-D) in both cover types of the treated sector {n = 80 plots) and 10 plots in the control sector were selected. Plots were separated by 200 m and plot cen- ters were >50 m from disturbances (e.g., log- ging roads) to minimize edge effects (Strelke and Dickson 1980, Yahner 1987). Area sam- pled was 25.4 ha or 2.3% of the total study area. I visited each plot in late May to early June each year (2005-2007) in a randomized order (Yahner 2003). All birds seen or heard within 30 m of each plot center during a 5-min period after a 1-min equilibrium period were count- ed. Point counts were made between sunrise and 0900 hrs EST. A 30-m radius point count was used to minimize edge effects at interfac- es of plots of different age post-cutting in the treated sector (Repenning and Labisky 1985). Birds flying over a plot were not counted, and movements of individual birds were moni- tored carefully to reduce the likelihood of counting the same bird more than once in the same morning. I calculated species richness and total abun- dance of each species in 2005-2007 combined in each of the three sectors (aspen = 100% clearcut; mixed-oak = 75% clearcut, and ref- erence = 0% clearcut). I compared observed versus expected abundances of all bird species combined and of common species (>10 con- tacts in either 2001-2002 or 2005-2007), us- ing Chi-square goodness-of-fit tests (Sokal and Rohlf 1995). I calculated expected num- ber of contacts in 2001-2002 or in 2005-2007 as the total number of contacts observed per period divided by 2 or 3 years, respectively (after Yahner 2003). RESULTS The number of species increased from 2001-2002 {n = 40) to 2005-2007 {n = 46) (Table 1). Abundance of all species combined declined (0.10 > P > 0.05) between periods. The most common species in both periods was Red-eyed Vireo {Vireo olivaceus). Six of the 20 common species were found only in the treated sector in both periods; none was spe- cific to the reference sector. Species present on the area in 2001-2002 but not in 2005-2007 were Cooper’s Hawk {Accipiter cooperii). Ruby-throated Hum- mingbird {Archilochus colubris), Olive-sided Flycatcher {Contopus cooperi). Willow Fly- catcher {Empidonax traillii), and Yellow- breasted Chat {Icteria virens). Species absent on the area in 2001—2002 but observed in 2005-2007 were Red-tailed Hawk {Buteo ja- maicensis), Baltimore Oriole {Icterus galbu- la). Worm-eating Warbler {Helmitheros ver- mivorus). Black-throated Green Warbler {Den- droica virens). Blue-winged Warbler {Vermi- vora pinus). Mourning Warbler {Oporornis Philadelphia), Hooded Warbler {Wilsonia ci- trina). Least Flycatcher {Empidonax mini- mus), Brown Thrasher {Toxostorna rufum). Wild Turkey {Meleagris gallopavo), and Vee- ry {Catharus fuscescens). Population densities of Gray Catbird {Du- SHORT COMMUNICATIONS 899 TABLE 1. Population abundance (contacts/100 ha/year) of common species (>10 contacts during springs (late May-mid-Jun) in either 2001-2002 or 2005-2007) at the Barrens Grouse Habitat Manage- ment Area (actual number of contacts in parentheses). Contacts of uncommon species are included in values of species richness and total population abundances of all species combined. 2001-2002 2005-2007 Blue Jay, Cyanocitta cristata^ 26 (13) 10 (8) Tufted Titmouse, Baeolophus bicoIoC 10 (5) 22 (17) Black-capped Chickadee, Poecile articapillus 20 (10) 16 (12) Blue-gray Gnatcatcher, Polioptila caenilea 18 (9) 14 (11) Wood Thrush, Hylocichla mustelina"^ 20 (10) 9 (7) Gray Catbird, Dumetella carolinensis^'^ 35 (18) 83 (63) Eastern Wood-Pewee, Contopus virens 14 (7) 13 (10) Red-eyed Vireo, Vireo olivaceus 162 (82) 161 (123) Chestnut-sided Warbler, Dendroica pensylvanica^"^ 32 (16) 92 (70) Ovenbird, Seiurus aurocapilla"^ 122 (62) 69 (53) Common Yellowthroat, Geothylpis trichas"" 57 (29) 39 (30) Hooded Warbler, Wilsonia citrina^ 0 (0) 13 (10) American Redstart, Setophaga ruticilla^ 65 (33) 34 (26) Rose-breasted Grosbeak, Pheiicticus ludovicianus 33 (17) 39 (30) Indigo Bunting, Passerina cyanea'^ 39 (20) 33 (25) Eastern Towhee, Pi p ilo e ryth roph tha h n us 75 (38) 72 (55) Field Sparrow, Spizella pusilkf 30 (15) 21 (16) Chipping Sparrow, S. passerina"" 20 (10) 20 (15) Scarlet Tanager, Piranga olivaceus 30 (15) 28 (21) Abundance, all species combined‘s 946 (480) 845 (644) Total number of species 40 46 “ Population abundance significantly different between periods (x^ ^ 3.84. df = I, P < 0.05). Population abundance slightly different between periods (x^ 2: 2.71, df = I, 0.10 > P > 0.05). -Species observed only in the treated sector. metella carolinensis). Chestnut-sided Warbler (Dendroica pensylvanica), and Hooded War- bler dramatically increased (P < 0.05) from 2001-2002 to 2005-2007 at the Barrens GHMA (Table 1). Populations of Tufted Tit- mice (Baeolophus bicolor) also had significant increases (0.10 > P > 0.05) from 2001-2002 to 2005-2007. No populations of early suc- cessional bird species declined {P > 0.05) over time despite increased forest maturation from 2001-2002 to 2005-2007. Populations of three species declined (P < 0.05) from 2001-2002 to 2005-2007: a woodland spe- cies, Ovenbird {Seiurus aurocapilla), a sec- ondary forest species, American Redstart (5'^- tophaga ruticilla), and a corvid. Blue Jay (Cyanocitta cristata). DISCUSSION I attributed a higher number of species in 2005-2007 on the Barrens GHMA, in part, to different levels of sampling between periods (e.g., 2 vs. 3 consecutive years, respectively). Perhaps the decline in abundance of all spe- cies combined was a result of the site being more heterogeneous in 2001-2002 than in 2005-2007. Populations of Blue-winged Warbler may be increasing in central Pennsylvania at the expense of populations of Golden-winged Warbler (Vermivora chrysoptera) (Kubel 2005), but the Blue-winged Warbler was ab- sent earlier on the area (Yahner 1997, 2003). Population declines of Ovenbird at the Bar- rens GHMA in 2005-2007 paralleled signifi- cant Pennsylvania decreases in populations of this species according to the Breeding Bird Survey (BBS) (USDI 2007) during the period from 1980 to 2006. According to the BBS, American Redstart and Blue Jay populations in Pennsylvania have not declined significant- ly during this period. Blue Jays, however, may be experiencing localized population decreas- es due to West Nile virus (Male 2003, Rohnke 2005). Gray Catbird populations increased in 2005-2007, but no similar trend was found in BBS data for Pennsylvania (USDI 2007). In contrast, populations of both Chestnut-sided and Hooded warblers increased significantly based on BBS data for Pennsylvania. I suspect that increases in populations on the area of these latter two species were due to regional 900 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 trends and were affected positively by pres- ence of older habitat, especially in aspen D and in oak C plots on my study site. Chestnut- sided Warblers were especially abundant in these types of plots at the Barrens GHMA (Schill 2007). My prediction (Yahner 2003) that early suc- cessional species would decline in 2005-2007 did not occur at the Barrens GHMA. It is probable that plant succession has not pro- gressed adequately to affect population den- sities of these species, despite 5-7 years hav- ing elapsed since the last cutting cycle on the area. Populations of Field Sparrows {Spizella piisilla) were quite low at the Barrens GHMA prior to the fourth cutting cycle in 1999-2000; subsequent to this cycle, populations in- creased significantly (Yahner 2003) and have not changed since, based on results obtained in the present study. Habitat management for Ruffed Grouse on the Barrens GHMA gen- erally did not have a profound effect on the sympatric bird community after the last cut- ting cycle. It is probable that more time must elapse before significant responses by bird populations will occur as a result of even-aged management at the Barrens GHMA. ACKNOWLEDGMENTS This study was funded by the Pennsylvania Agri- cultural Experiment Station. I thank H. L. Besecker for clerical assistance. LITERATURE CITED Gullion, G. W. 1977. Forest manipulation for Ruffed Grouse. Transactions of the North American Wild- life and Natural Resources Conference 42:449- 458. Gullion, G. W. 1990. Fore.st-wildlife interactions. Pages 349-383 in Introduction to forest science (R. A. Young and R. L. Griese, Editors). Second Edition. John Wiley and Sons, New York, USA. Kubel, J. E. 2005. Breeding ecology of Golden- winged Warblers in managed habitats in central Pennsylvania. Thesis. Pennsylvania State Univer- sity, University Park, USA. Male, T. 2003. Potential impact of West Nile virus on American avifaunas. Conservation Biology 17: 928-930. Repenning, R. W. and R. E Labisky. 1985. Effects of even-age management on bird communities of the longleaf pine forest in northern Florida. Journal of Wildlife Management 49:1088-1098. Rohnke, a. T. 2005. Possible ecological effects of hab- itat alteration and West Nile virus on songbird populations in central Pennsylvania. Thesis. Penn- sylvania State University, University Park, USA. Schill, K. L. 2007. Habitat use and nesting ecology of Chestnut-sided Warblers. Thesis. Pennsylvania State University, University Park, USA. SoKAL, R. R. AND F. J. Rohlf. 1995. Biometry. Third Edition. W. H. Freeman and Company, New York, USA. Strelke, W. K. and j. G. Dickson. 1980. Effect of forest clear-cut edge on breeding birds in east Tex- as. Journal of Wildlife Management 44:559-567. U. S. Department of Interior (USDI). 2007. Breed- ing Bird Surveys. USDI, Geological Survey, Pa- tuxent Wildlife Research Center, Laurel, Mary- land, USA. www.pwrc.usgs.gov/BBS/ (accessed 20 December 2007). Yahner, R. H. 1987. Use of even-aged stands by win- ter and spring bird communities. Wilson Bulletin 99:218-232. Yahner, R. H. 1993. Effects of long-term forest clear- cutting on wintering and breeding birds. Wilson Bulletin 105:239-255. Yahner, R. H. 1997. Long-term dynamics of bird communities in a managed forested landscape. Wilson Bulletin 109:595-613. Yahner, R. H. 2000. Long-term effects of even-aged management on bird communities in central Penn- sylvania. Wildlife Society Bulletin 28:1 102-1 1 10. Yahner, R. H. 2003. Responses of bird communities to early successional habitat in a managed land- scape. Wilson Bulletin 115:292-298. SHORT COMMUNICATIONS 901 The Wilson Journal of Ornithology 1 20(4):901-905, 2008 Freeze-Frame Fruit Selection by Birds Mercedes S. Foster' ABSTRACT — The choice of fruits by an avian fru- givore is affected by choices it makes at multiple hi- erarchical levels (e.g., species of fruit, individual tree, individual fruit). Factors that influence those choices vary among levels in the hierarchy and include char- acteristics of the environment, the tree, and the fruit itself. Feeding experiments with wild-caught birds were conducted at El Tirol, Departamento de Itapua, Paraguay to test whether birds were selecting among individual fruits based on fruit size. Feeding on larger fruits, which have proportionally more pulp, is gener- ally more efficient than feeding on small fruits. In trials {n = 56) with seven species of birds in four families, birds selected larger fruits 86% of the time. However, in only six instances were size differences significant, which is likely a reflection of small sample sizes. Re- ceived 15 August 2007. Accepted 26 February 2008. Frugivorous birds have the potential to choose among food items at multiple hierar- chical levels (Sallabanks 1993, Wheelwright 1993). They can choose among different spe- cies of fruits, among individual trees of a pre- ferred species of fruit, and among individual fruits on their chosen tree. Once a fruit is plucked and perhaps ‘tested’ in the bill, a bird can choose to discard the fruit or continue to feed on it. Factors influencing choice may be external to the plant, characteristics of the plant unrelated to the fruit, or characteristics of the fruit. External factors include nature of the habitat and presence nearby of fruiting in- dividuals of the same or other species (e.g., Foster 1990, Saracco et al. 2005). Relevant characteristics of the plant include crown size, foliage density,, and fruit abundance (Foster 1990, Sallabanks 1992, Laska and Stiles 1994, Ortiz-Pulido and Rico-Gray 2000). Characters of fruits can be either extrinsic (e.g., accessi- bility, Moermond and Denslow 1983, Levey and Moermond 1984) or intrinsic (e.g., pulp: seed ratio, nutrient content, amount of pulp, fruit and seed size, parasite infestation, color. ' uses, Patuxent Wildlife Research Center, Nation- al Museum of Natural History, P. O. Box 37012, Wash- ington, D.C. 20013, USA; e-mail: fosterm(?\si.edu taste, ripeness, etc.). Potentially, characters in all these categories can influence choice of a fruit species or of an individual tree in which to feed; only a subset of the characters can influence the selection of a particular fruit and the decision to eat it or not. Considerable effort has been invested in ex- amining the basis of choice at all levels, and nearly every external, tree, and fruit feature has been correlated with species preference, visitor frequency, or fruit removal rate in one plant species or another (e.g., Sorenson 1983, 1984; Johnson et al. 1985; Piper 1986; Hedge et al. 1991; White and Stiles 1991; Sallabanks 1992, 1993; Willson and Comet 1993; Willson and Whelan 1993; Whelan and Willson 1994; Stanley and Fill 2002; Stanley et al. 2002). However, clear patterns of correlations have not emerged, although visitor frequency, ab- solute number of fruits removed, and propor- tion of fruits removed are often correlated with size of the initial fruit crop (e.g., Foster 1990, Laska and Stiles 1994, Ortiz-Pulido and Rico-Gray 2000). The absence of strong, widespread correlations between visitation rates and fruit characteristics is counter intu- itive. Frugivores should maximize their feed- ing efficiencies (benefit obtained from the fruit/cost of fruit handling) among conspecific plants by preferentially feeding in trees with the largest or most nutritious fruits to maxi- mize benefits; trees whose fruits have the smallest seeds (ballast) to minimize costs; or trees whose fruits have the greatest pulp:seed ratio (e.g., Martin 1985, Wheelwright 1993). The absence of a correlation between rate of fruit removal and any fruit trait could reflect the variability of those traits among fruits within trees and the hierarchical nature of the points of choice. A bird could use a non-fruit criterion (e.g., foliage density or fruit abun- dance) to select a tree for feeding and then use a fruit character to select among the fruits in that tree based on traits that maximize feed- ing efficiency. 902 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. J20, No. 4, December 2008 I examined this possibility in Cocu (Allo- p/iylus eclulis, Sapindaceae), whose fruits (6. 7-9. 4 X 6. 6-9. 6 mm) are eaten by at least 26 species of birds (Foster 1987). The birds feed with noticeably different frequencies in individual Cocu trees (Foster 1990). The only tree/fruit characteristic with which visitation rates are broadly correlated is size of the fruit crop. Abundance of fruits is highly correlated with tree crown size and foliage density, which may be the operant characteristics. General correlations with fruit characteristics have not been noted (Foster 1990). The only Allophylus fruit characteristics visually avail- able to birds are color, size, and shape al- though birds that manipulate the fruit with the bill may be able to detect pulp:seed ratio and/ or seed size or larval infestation. Taste or smell may allow birds to detect nutrient con- tent or presence of toxic substances (Sorenson 1983). I have no knowledge of either of the latter possibilities for A. eclulis and focused on visual cues. Birds ate only obviously ripe fruits of A. eclulis in 216 hrs of observation. Ripe fruits are a uniform bright, shiny red, and shape is equally uniform. Therefore, I only considered fruit size. The objectives of the study were to: (1) establish which fruit char- acters with a potential to influence feeding ef- ficiency directly (i.e., amount of pulp, size of seed, pulpiseed ratio) were correlated with fruit size, and (2) conduct cage experiments to investigate whether birds were choosing fruits on the basis of size. METHODS Allophylus eclulis occurs in southern South America east of the Andes (Lopez et al. 1987). I studied it in the vicinity of El Tirol (27° 1 1 ' S, 55° 46.5' W), -19.5 km by road NNE of Encarnacion, Departamento de Itapua, Para- guay, in a remnant of Atlantic Forest. The area is described in Smith and Foster (1984). Tree and bird characteristics, and interactions have been reported elsewhere (Foster 1987, 1990). I examined pulp weight/fruit, seed weight/ fruit, and pulp:seed ratio with random samples of fruits (/? = 10-31) from 14 trees (6 trees in 1981 and 1983, plus 2 additional trees in 1983) to see if they were correlated with fruit size. Because Allophylus fruits are spherical with somewhat flattened ends, I calculated size as the volume (mm^) of a cylinder (V = TTr^h; r = radius, h = height), even though differences between fruit height and diameter rarely exceeded 1 mm and were all <5 mm. Whole fruits were weighed to 0.01 g on a field balance; seeds were removed from the fruits, cleaned, and weighed; and pulp weight was calculated by subtraction. These values were used to calculate the pulp:seed ratio. I have no data on the nutrient and energy contents of individual fruits although composite samples from individual trees often differed (Foster 1990). However, increased pulp at any size translates to an absolute increase in energy and nutrients. Birds known to eat Cocu were mist-netted in the forest in 1978 and 1980-1982, placed in inverted mosquito head nets (28 X 38 cm) with wire frames that kept them open and a stick perch, and taken to the test area. Birds were held 1-2 hrs in the dark prior to testing and then introduced into a cage (100 X 75 X 100 cm high) of gray fiberglass screen with one lengthwise and one crosswise horizontal perch. I introduced 20 freshly collected and mea- sured (height and diameter with calipers) fruits into the cage once a bird was accus- tomed to it and had rested calmly for 15 min. Each fruit was hooked to the tip of a fine wire attached to a frame of 2-mm-diameter dowels suspended from the roof of the cage. Ten fruits were positioned on either side of and parallel to but above the lengthwise perch with distances from the perch depending on the size of the bird. Birds were able to reach and pluck a fruit but also could move freely along the perch without touching the fruits. Fruit positions were numbered consecutively. Fruits were selected to cover a large range of sizes (volumes) and randomly attached to the wires. The volume of the fruit at each position was recorded. I recorded the position number of each fruit from a blind (—2 m distant) as it was eaten. I terminated a trial after a bird had eaten about half the fruits, had not fed after 30 min in the cage, or if an initial period of feeding was followed by a 15-min period of non-feeding. Each bird was held in the head net for 1-2 hrs at the end of its first trial and then used for a second trial. The right outer tail feather of each test bird was clipped so that it could be recognized if netted subse- SHORT COMMUNICATIONS 903 TABLE 1. Average sizes of fruits eaten (E) or not eaten (NE) in fruit selection trials, by bird species. Trials Differences in X fruit sizes (E - NE), all trials Birds n Differences in X fruit sizes (E - NE), all birds^ Species mm^ pb mm^ ph Blue Manakin (Chiroxiphia caudata) 5 17 0,320 3 12 0.612 Pale-breasted Thrush (Turdus leucomelas) 4 (12)^ 0.688 2 (52)^ 0.549 Black-goggled Tanager {Trichothraupis melanops) 23 31 0.003 13 33 0.016 Ruby-crowned Tanager (Tachyphoniis coronatus) 11 39 0.006 6 28 0.157 Violaceous Euphonia {Euphonia violacea) 6 63 <0.001 3 44 0.101 Guira Tanager {Hemithraiipis guira) 5 35 0.059 4 40 0.030 Red Pileated Pinch {Coryphospingus cucullatus) 2 39 0.214 1 42 0.478 All species combined 56 512/608^ 31 <0.001 32 271/368^ 28 0.001 “ Only the first trial is included for birds with two trials. ^ P = probability (t test) that average sizes (mm^) of eaten and not eaten fruits are equal. Parentheses = average size of fruits not eaten > average size of fruits eaten. Total fruits at 20/trial, E/NE. quently; no bird was used for more than two trials. I tested the significance of correlations be- tween fruit size and other fruit characteristics with linear regression analysis. To learn if in- dividual birds were discriminating among available fruits by size, I compared the aver- age volume of fruits selected in each trial with that of the fruits that were not selected. I also compared the combined data for all trials for each species as well as the combined data for all trials of all species. Finally, I repeated the latter two comparisons including data from only a single trial (the first) for each bird, even though the 1-2-hr period that elapsed between trials was likely sufficient to make them in- dependent. I tested for homogeneity of vari- ances {F test) and compared means with either a two-sample Student’s r-test (equal variances) or Welch’s r-test (unequal variances). I used a sign test to examine if differences between av- erage volumes of eaten and uneaten fruits in each trial were in the same direction, i.e., if eaten fruits were the larger, even if differences were not significant. I used PAST Version 1.78 (Hammer et al. 2001) for all analyses. RESULTS Pulp weight per fruit was positively corre- lated (/- values = 0.6915-0.9814; all P < ().()()2) with fruit volume in all 14 samples. Results for both seed weight and pulp:seed ra- tio were mixed. Both traits were positively correlated with fruit volume, but only six of the seed weight correlations were significant (r = 0.3591-0.8845; P = 0.0001-0.047) and only seven of the pulpiseed ratio correlations (r = 0.5785-0.8027; P = 0.001-0.024). The slopes of the relationships of pulp weight, seed weight, and pulp:seed ratio to fruit vol- ume were 0.00 13X, 0.0006X, and 0.009X, re- spectively. Thirty-two individuals of seven species in four families selected and ate fruits in 56 trials (Table 1). Some birds plucked fruits while on a main perch; others removed them by hover- plucking or perching on the wooden fruit dis- play frame and reaching downward. The av- erage volumes of the eaten and uneaten fruits differed in 55 of 56 trials; the eaten fruits were larger (sign test, r = 48, P < 0.001) more often {n = 48) than were uneaten fruits {n — 7). However, in only six (11%) of all trials, spread among five bird species, were differ- ences in average volumes of the eaten and un- eaten fruits significant, likely because of small sample sizes (20 fruits/trial); in each case the fruits selected by birds were larger than those not selected. Average volumes of eaten and uneaten fruits differed significantly when val- ues for all tests were combined (Table 1). The same was true when only a single trial was included for each bird. In the comparisons by species, size differences were significant only for Black-goggled Tanager, Ruby-crowned Tanager, and Violaceous Euphonia (scientific names are in Table 1 ), although a was 0.059 for Guira Tanager. Differences were signifi- cant for all species combined and Black-gog- gled and Guira tanagers when tests were con- 904 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 ducted using only a single sample for each bird (Table 1). DISCUSSION Frugivorous birds are particularly suited for studies of foraging and feeding behavior be- cause locations of their feeding areas (fruiting plants) are fixed and their ‘prey items’ (fruits) do not attempt to escape. Thus, feeding sites used and fruits eaten generally represent choice rather than opportunistic consumption of a suitable food item that serendipitously ap- pears. I hypothesized that birds feeding on A. edulis would select the largest fruits available, a practice that has been demonstrated previ- ously in other bird species (e.g., Johnson et al. 1985, Piper 1986, White and Stiles 1991, Sal- labanks 1993, Wheelwright 1993, Stanley et al. 2002). Seed size (ballast) increases with increasing fruit size, but amount of pulp in- creases at a faster rate. Consequently, larger fruits supply greater amounts of pulp and have a greater pulp: seed ratio. Both of these con- ditions enhance feeding efficiency, especially among Type III (push and bite) feeders (e.g., Eiiphonia spp., Foster 1987). Birds in cage tests generally (86% of the trials and 86% of the species) selected fruits that were larger than those not selected, regardless of feeding type. The lack of significance among the dif- ferences likely reflects the small sample sizes. The probability that sizes of eaten and uneaten fruits were statistically the same, in most in- stances, increased when n decreased, when only a single trial for each bird was included in the analysis. The seeming anomalous result from the Pale-breasted Thrush in which un- eaten fruits averaged larger than those eaten (although not significantly) likely reflects small sample size. Birds selected larger fruits on average, even though differences were only occasionally sta- tistically significant. This may reflect a ‘freeze-frame’ foraging strategy in which a bird selects a fruit only from among those within its immediate field of vision, where siz- es can be directly compared. In this scenario, the bird selects one of the larger fruits within the frame even though the frame does not nec- essarily include the largest fruits on the tree or even any fruits in the upper half of the size range of fruits on the tree. After eating 1-2 fruits, the bird changes perches and views the fruits in another frame. Martin (1985) suggested that birds should prefer fruits that are within an optimal size range defined by handling time (including search time) and reward size. Handling costs also increase as fruit size increases until they outweigh the benefits derived from eating a larger fruit. The increased search time re- quired to re\'iew and compare sizes of some proportion of fruits before selecting one on a tree such as Cocu could increase costs beyond the increase in benefits from selecting a larger fruit. The cost increase might be even greater if ascertaining the direction of size differences among separated fruits and retaining the lo- cations of previously viewed fruits were dif- ficult, and required moving between fruits more than once. Avery et al. (1993) demon- strated for Cedar Waxwings {Bombycilla ced- roriim) that average diameters of test fruits had to differ by some threshold amount before the birds exhibited size preferences, and that birds did not exhibit a preference between fruits of adjacent size classes. Their work sug- gested that birds were less able to recognize threshold differences when fruit sizes, rather than comprising discrete categories, formed a continuum as in Allophyliis. As long as fruit sizes are within the optimal size range for the species, the increased time required by a bird to compare a large sample of fruits before se- lecting one might reduce the extra benefits from feeding on larger fruits to insignificance. This could be particularly applicable when numbers of fruits that birds must eat to meet daily existence energy needs are small and feeding time minimal, as with Allophyliis, es- pecially for Type I (pluck and swallow) and Type II (cut or mash) feeders (Foster 1987). Increased activity in the tree might also attract predators and increase vulnerability of the birds. Birds appear to exert choice at two levels when feeding on Allophyliis edulis fruits, first by selecting a feeding tree based on fruit abundance and second by selecting among fruits within that tree based on size. Testing to learn if they accomplish the latter by freeze- frame foraging will require additional cage experiments and observations of feeding birds in the wild, using a variety of plant species whose fruits differ in size. SHORT COMMUNICATIONS 905 ACKNOWLEDGMENTS I thank the Paraguayan Ministry of Agriculture, Hernando Bertoni, and Hilario Moreno for facilitating research permits; Miguela and Armando Reynaers for permission to work at El Tirol, hospitality, and logistic support; E. W. Schupp for assistance in the field; and S. W. Droege and L.-A. Hayek for advice on statistical analyses. S. W. Droege, D. C. Hahn, and two anony- mous reviewers provided useful comments on an early draft of the manuscript. Financial support for field work was provided by the Museum of Vertebrate Zo- ology, University of California, Berkeley, and the U. S. Fish and Wildlife Service. LITERATURE CITED Avery, M. L., K. J. Goocher, and M. A. Cone. 1993. Handling efficiency and berry size preferences of Cedar Waxwings. Wilson Bulletin 105:604-611. Foster, M. S. 1987. Feeding methods and efficiencies of selected frugivorous birds. Condor 89:566-580. Foster, M. S. 1990. Factors influencing bird foraging preferences among conspecific fruit trees. Condor 92:844-854. Hammer, 0., D. A. T. Harper, and P. D. Ryan. 2001. PAST: paleontological statistics software package for education and data analysis. Palaeontologia Electronica 4(1): 1-9. Hedge, S. G., K. N. Ganeshaiah, and R. U. Shaank- ER. 1991. Fruit preference criteria by avian frugi- vores: their implication for the evolution of clutch size in Solatium pubescens. Oikos 60:20-26. Johnson, R. A., M. F. Willson, J. N. Thompson, and R. I. Bertin. 1985. Nutritional values of wild fruits and consumption by migrant frugivorous birds. Ecology 66:819-827. Laska, M. S. and E. W. Stiles. 1994. Effects of crop size on intensity of fruit removal in Viburnum pnmifolium (Caprifoliaceae). Oikos 69:199-202. Levey, D. J. and T. C. Moermond. 1984. Fruit choice in neotropical birds: the effect of distance between fruits on preference patterns. Ecology 65:844- 850. Lopez, J. A., E. L. Little Jr., G. F. Ritz, J. S. Rom- bold, AND W. J. Hahn. 1987. Arboles Comunes del Paraguay. Peace Corps, Washington, D.C., USA. Martin, T. E. 1985. Resource selection by tropical fru- givorous birds: integrating multiple interactions. Oecologia 66:553-573. Moermond, T. C. and J. S. Denslow. 1983. Fruit choice in neotropical birds: effects of fruit type and accessibility on selectivity. Journal of Animal Ecology 52:407-419. Ortiz-Pulido, R. and V. Rico-Gray. 2000. The effect of spatio-temporal variation in understanding the fruit crop size hypothesis. Oikos 91:523-527. Piper, J. K. 1986. Seasonality of fruit characters and seed removal by birds. Oikos 46:303-310. Sallabanks, R. 1992. Fruit fate, frugivory, and fruit characteristics: a study of the hawthorn. Crate gus monogyna (Rosaceae). Oecologia 1992:296-304. Sallabanks, R. 1993. Hierarchical mechanisms of fruit selection by an avian frugivore. Ecology 74: 1326-1336. Saracco, j. F, j. a. Collazo, M. J. Groom, and T. A. Carlo. 2005. Crop size and fruit neighborhood ef- fects on bird visitation to fruiting Schefflera moro- totoni trees in Puerto Rico. Biotropica 37:81-87. Smith, A. R. and M. S. Foster. 1984. Chromosome numbers and ecological observations of ferns from El Tirol, Paraguay. Fern Gazette 12:321-329. Sorensen, A. E. 1983. Taste aversion and frugivore preference. Oecologia 56:117-120. Sorensen, A. E. 1984. Nutrition, energy and passage time: experiments with fruit preference in Euro- pean Blackbirds (Turdus merula). Journal of An- imal Ecology 53:545-557. Stanley, M. C. and A. Lill. 2002. Importance of seed ingestion to an avian frugivore: an experimental approach to fruit choice based on seed load. Auk 119:175-184. Stanley, M. C., E. Smallwood, and A. Lill. 2002. The response of captive Silvereyes (Zosterops la- teralis) to the colour and size of fruit. Australian Journal of Zoology 50:205-213. Wheelwright, N. T. 1993. Fruit size in a tropical tree species: variation, preference by birds, and heri- tability. Vegetatio 107/108:163-174. Whelan, C. J. and M. F. Willson. 1994. Fruit choice in migrating North American birds: field and avi- ary experiments. Oikos 71:137-151. White, D. W. and E. W. Stiles. 1991 . Fruit harvesting by American Robins: influence of fruit size. Wil- son Bulletin 103:690-692. Willson, M. F. and T. A. Comet. 1993. Food choices by Northwestern Crows: experiments with cap- tive, free ranging and hand-raised birds. Condor 95:596-615. Willson, M. F. and C. J. Whelan. 1993. Variation of dispersal phenology in a bird-dispersed shrub, Corniis drummomlii. Ecological Monographs 63: 151-172. 906 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 The Wilson Journal of Ornithology 1 20(4): 906-9 1 0, 2008 Predation of Rio Grande Wild Turkey Nests on the Edwards Plateau, Texas Justin Z. Dreibelbis,’ Kyle B. Melton, ^ Ray Aguirre,^ Bret A. Collier,' '^ Jason Hardin, ^ Nova J. Silvy,' and Markus J. Peterson' ABSTRACT — We followed the fate of nests of Rio Grande Wild Turkeys {Meleagris gallopavo interme- dia) on the Edwards Plateau of Texas during 2006 and 2007 using motion-activated digital cameras on a sub- set of nests to evaluate the frequency of nest predation and to identify nest predators. Predation was the pri- mary cause of loss for nests with cameras, accounting for 57 and 65% in 2006 and 2007, respectively. Pre- dation for nests without cameras also was high (69 and 65% for 2006 and 2007, respectively) suggesting the cameras did not increase the probability of nest failure. We documented partial-and multiple-predator events that could result in misidentification of nest predators. Our results provide insight into nest predator com- munities and confirm that multiple predator events oc- cur with regularity in the wild. Received 3 December 2007. Accepted 29 April 2008. Natality is one of the primary biological processes influencing dynamics of wildlife populations (Everett et al. 1980). Understand- ing which factors cause changes in individual and group natality is important for managing bird populations. Methods to estimate and un- derstand components of nest survival have re- ceived recent attention, particularly for species of ground nesting birds (Dinsmore et al. 2002, Shaffer 2004, Grant et al. 2005). A variety of factors can influence nest survival, but for ground nesting birds, nest predation appears most influential (Ricklefs 1969, Farnsworth and Simons 2000, Rollins and Carroll 2001, Stephens et al. 2005). Given the vulnerability of ground nesting species, predation will af- ' Department of Wildlife and Fisheries Sciences, Texas A&M University, College Station, TX 77843, USA. ^ Texas Parks and Wildlife Department, Waco, TX 76706, USA. ^ Texas Parks and Wildlife Department, Comfort, TX 78013, USA. Texas Parks and Wildlife Department, Palestine, TX 75803, USA. ^ Corresponding author; e-mail: bret@tamu.edu feet nest survival and population productivity (Baker 1978, Rollins and Carroll 2001). Accurate identification of nest predators of ground nesting birds is important in under- standing effects of predation on population parameters (Lariviere 1999, Rader et al. 2007). Nest predation studies often rely on physical evidence at the nest, such as tracks, hair, and eggshell fragments to identify pred- ators (Major 1991, Lariviere 1999). Use of physical evidence can be highly subjective (Trevor et al. 1991, Lariviere 1999), and may fail to account for multiple-predator and par- tial-predation events (Leimgruber et al. 1994). Predation events may be difficult to identify if eggshells are removed by the incubating hen following partial nest predation (Lariviere and Walton 1998), or if predation is by reptilian or avian species, as snakes consume whole eggs in the nest (Staller 2001) and avian spe- cies often remove eggs from the nest before consumption (Montevecchi 1976). Abundance of Rio Grande Wild Turkeys {Meleagris gallopavo intermedia) on the southeastern Edwards Plateau, Texas has de- clined since the late 1970s (Randel et al. 2005, Collier et al. 2007a). Recent work has focused on evaluating factors contributing to this de- cline (Collier et al. 2007b), including variation in reproductive potential and nest survival (Melton 2007). Predation is the primary cause of nest failure in the region (Cook 1972, Mel- ton 2007), and nest loss can adversely influ- ence Wild Turkey populations (Davis 1959, Baker 1978). Our objectives were to: (1) iden- tify predators of Rio Grande Wild Turkey nests and (2) examine the frequency of total nest loss, partial predation events, and multi- ple-predator predation events. METHODS Study Area. — Our study area was within the Edwards Plateau region of Texas; we studied SHORT COMMUNICATIONS 907 Wild Turkey nesting from January through July 2006 and 2007 on four sites in Kerr, Real, Bandera, and Medina counties. All study sites were rangelands with flat to rolling divides, shallow soils, and limestone bedrock (Gould 1975), and included private ranches and the Kerr Wildlife Management Area (Texas Parks and Wildlife Department). Study sites ranged in size from 984 to 8,858 ha and all were man- aged for hunting of native and exotic wildlife; livestock grazing occurred on three of the sites (Kerr, Medina, and Bandera counties). Field Procedures. — We trapped Wild Tur- key hens during January-March, 2006 and 2007. We attached radio transmitters (69.0- 95.0 g; Advanced Telemetry Systems, Isanti, MN, USA) to 39 and 22 hens in 2006 and 2007. We located individual hens three times weekly (White and Garrott 1990) during the breeding season until behaviors indicated a hen had initiated a nest (Ransom et al. 1987). We located nests within 1 day after we sus- pected hens had begun incubation. Once lo- cated, we ascertained initiation date, clutch size, and approximate age for each nest. We estimated nest age and initiation date by back- dating from the day we found the nest to the day we first located the hen in the nest area. We defined the active nesting period as 39 days; the sum of the average number of eggs in a clutch (11) and a 28-day incubation pe- riod (Bailey and Rinnell 1967, Melton 2007). We floated eggs to estimate age of nests found during incubation (Westerskov 1950), and monitored nests three times weekly from a distance of >100 m to prevent further distur- bance to the hen. We assumed the nest was active if hen locations remained constant. One week before estimated hatch date, we visited nests daily to ensure accurate identification of hatch date. We used motion-activated trail cameras (Game Spy 100 and Outfitter Cam, Moultrie Feeders, Alabaster, AL, USA) at a sample of nests. Each camera was equipped with 16 MB of internal memory (we added a 256 MB memory card to each camera in 2007), a 10.2 mm lens, and a 9.14 m flash. We learned through a pilot study in 2005 that cameras set within 5 m of a nest require flash reduction, otherwise night photographs were over-ex- posed. To reduce flash, we covered 100% of the flash surface with one to three layers of TABLE 1 . Nest predators documented via remote- ly-triggered cameras at active Rio Grande Wild Tur- key nests in the Edwards Plateau, Texas, 2006-2007 {n = number of nests with photographed predation events). Species (/; 2006 = 7 nests) (n 2007 = 1 1 nests) Nine-banded armadillo {Dasypus novemcinctiis) 0 1 Bobcat {Lynx rufiis) 0 1 Feral hog (Sus scrofa) 2 1 Gray fox ( Urocyon cinereoargenteiis) 4 2 Common raccoon {Procyon lotor) 2 7 Common Raven (Corviis corax) 0 3 Striped skunk {Mephitis mephitis) 2 0 Texas rat snake {Elaphe obsolete lindheimeri) 1 0 Total multiple predator events 3 4 masking tape, dependent upon nest distance (most often one layer/m from the nest under 5 m). We attached the camera, based on veg- etation surrounding the nest area, to a tree near the nest or to a post. We programmed cameras to take two pictures ~5 sec apart, fol- lowed by a 5 or 10-min delay. After the delay period, the next event in the nest area would trigger the camera. We checked cameras after initial setup, only when the bird was located out of the nesting area for more than 1 day. Nests receiving camera surveillance were cho- sen randomly across study sites depending on camera availability and nest initiation timing. RESULTS We placed cameras at 21 of 47 active turkey nests in 2006, with 12 (57%) nests depredated and eight (38%) nests abandoned. These rates are comparable to 69% depredation and 15% abandonment for those nests in our study without cameras. Three of 12 depredated nests with cameras involved more than one preda- tor, four involved a single predator, and live had no photographs of the nest predator (Table 1 ). We placed cameras at 31 of 71 active nests in 2007. Twenty of 31 (65%) nests with cam- eras were depredated and 6 of 31 ( 19%) were abandoned. Four of the depredated nests in- 908 THE W ILSON JOURNAL OF ORNITHOLOGY • VoL 120, So. 4. December 2008 vohed more than one predator. se\en in- volved a single predator, and nine had no predator photographs. We observed 68^ (27/ -40) predation and 18*T (7/40) abandonment at nests without cameras in 2007. Nests survived on average 12.5 and 13.0 days with and with- out eameras in 2006. and 18.4 and 18.7 days with and without cameras in 2007. \\*e were able to examine timing of preda- tion events in greater detail on approximately half the nests with cameras. For example, a multiple predator event occurred at a nest of a yearling hen found incubating her first nest containing 11 eggs on 17 May 2006. We flushed the hen. estimated nest age at 6 days of incubation, and placed a camera at the nest. \\'e recorded a remarkable series of predation events on 19 May at this nest. At 1818 hrs. a common raccoon ( scientific names of predator species are in Table 1 ) was recorded leaving the nest area and subsequent photographs showed a raccoon consuming an egg ~2 m from the nest. Later that evening (2212 hrs). two photographs (<10 sec apart) were taken of a raccoon predating the nest. Shortly there- after (2242 hrs). a gray fox visited the nest. Less than 1 hr later (2328 hrs). a striped skunk depredated the nest followed by a gray fox that visited the nest at 2344 hrs. We docu- mented additional predator visits on subse- quent days. Raccoons were obser\ed at the nest on 20 May at both 0111 and 0705 hrs as well as on 22 May at 0005 hrs. We photo- graphed feral hogs at the nest on 22 May at 0534 hrs and 2315 hrs. removing the remain- ing shell fragments from earlier predation events. The last recorded nest visitor was a raccoon on 24 May at 0409 hrs. The hen re- mained in the general v icinitv of the nest until 24 May when we examined the nest site at 1 126 hrs. finding no eggshell remains and lit- tle disturbance to the leaf litter. Given there was no evidence (egg shells, tracks, scat. hair, etc.) at the nest site when researchers arrived, we initially believ ed a reptilian or avian pred- ator was responsible. W e also documented an instance of panial nest predation. We located the nest on 17 April 2006. during incubation by an adult hen of her first nest of the season, which contained 16 eggs. We monitored the nest for 28 days, which was successful, and located the hen with eight poults on 14 May. W hen we re- turned to the nest area to collect eggshells, we found remnants of only nine hatched eggs. U pon checking the photographs, we found the nest had been paniallv depredated by a Texas rat snake 11 days earlier. On 3 May at 2118 hrs. we photographed the snake in the nest. The hen hatched the remaining nine eggs on 13 May 2006. There was no physical ev idence at the nest, and we initially believed the nest was predated by either a reptilian or avian predator. DISCUSSION Our observations indicate that nest preda- tion was the proximate factor affecting ov erall nest surv iv al of Rio Grande Wild Turkeys dur- ing our study, although our sample of nests was fairly small. Additionally, our results sug- gest that nest predation events involv ing mul- tiple predators were common. There is a di- verse predator communitv on the Edwards Plateau (Davis and Schmidly 1994) and key predators can change from year to year. The method of depredation used and the ev idence left at the nest site after depredation events (e.g.. eggshell fragments) may overlap among species. Gray fox were documented in 2006 at 57^ of the predation ev ents but were pho- tographed at only 2 ( 18^ ) predation ev ents in 2007 (both of which involved multiple pred- ators). Three of 11 (27*T) camera nests in 2007 identified Common Ravens removing eggs; however, no ravens were photographed in 2006. Nests depredated by ravens were similar to those depredated by snakes as they contained no shell fragments and had little dismrbance around the nest. Staller (2001) correctly identified 61^ of predators at Nonhem Bobwhite {CoUnus vir- ginianiis) nests using physical ev idence at the nest site as compared to data from rniniamre V ideo cameras: how ev er. div ersitv of predators on his smdv area was small. Only 12^ of pre- dation events from Staller (2001) involved multiple predators compared to Leimgruber et al. ( 1994) who observ ed multiple predator vis- its (2—5 species) in 43^ of predation events, a rate similar to ours. Hernandez et al. ( 1997) attempted to construct a dichotomous key for identification of ground-nest predators in vv est Texas but were not successful because of in- sufficient physical evidence and overlap of nest predation habits among species. Incubat- SHORT COMMUNICATIONS 909 ing Blue-winged Teal {A?ias discors) and Mal- lard {A. platyrhynchos) hens are known to re- move damaged eggs and shell fragments from the nest area following partial predation e\ents by striped skunks (Lariviere and Wal- ton 1998). The relationship between ground nesting birds and nest predators is complicated, and we caution researchers to understand the lim- itations of using physical evidence to predict nest predator species. Our results provide in- sight into nest predator communities and con- firm that multiple predator events are frequent (39% of the predation events recorded with cameras in our study) in the wild. Multiple predation events can greatly alter physical ev- idence left at the depredated nest site: thus, it is crucial that researchers test and apply any method which is used to assess nest predator communities before mitigation strategies are developed. ACKNOWLEDGMENTS Funding was provided by the Texas Turkey Stamp Fund through the Texas Parks and Wildlife Depart- ment. the National Wild Turkey Federation Texas State Superfund, and the Department of Wildlife and Fish- eries Sciences. Texas A&M University. We thank D. B. Frels Jr. and Kerr Wildlife Management Area per- sonnel. and Max Traweek and Region II Texas Parks and Wildlife personnel for assistance with trapping and field work. We thank all cooperating landowners and land managers for help and willingness to participate with our research. This research was conducted under Texas A&M University Animal Use Permit 2005-005. LITERATURE CITED B.xiley. R. W. .xnd K. T. Rixnell. 1967. Events in the turkey year. Pages 73-91 in The Wild Turkey and its management (O. H. Hewitt. Editor). The Wild- life Society. Washington. D.C.. USA. B.aker. B. W. 1978. Ecological factors affecting Wild Turkey nest predation on south Texas rangelands. Proceedings of the Annual Conference of the Southeastern Association of Fish and Wildlife Agencies 32:126-136. Collier. B. A.. D. A. Jones. J. N. Schaap. C. J. Ran- DEL III. B. J. WiLLSEY. R. AGLIRRE. T. W. SCHWERTNER. N. J. SiLVY. AND .M. J. PeTERSON. 2007a. Survival of Rio Grande Wild Turkeys on the Edwards Plateau of Texas. Journal of Wildlife Management 71:82-86. Collier. B. A.. K. B. Melton. J. Z. Dreibelbis. W. R Klvlesky. G. a. Prol'dfoot. R. Aglirre. D. G. Hewitt. T. W. Schwertner. S. J. DeMaso. N. J. SiLVY. AND .M. J. Peterson. 2(X)7b. Sex ratio var- iation in Texas Rio Grande Wild Turkeys. Journal of Wildlife Management 71:1793-1799. Cook. R. L. 1972. A study of nesting turkeys in the Edward’s Plateau of Texas. Proceedings of the An- nual Conference of the Southeastern Association of Game and Fish Commissioners 26:236—244. D.avis, j. R. 1959. A preliminary progress report on nest predation as a limiting factor in Wild Turkey populations. Proceedings of the National Wild Turkey Symposium 1:138-145. D.avis, W. B. .and D. j. ScmnDLY. 1994. The mammals of Texas. University of Texas Press, Austin, USA. Dinsmore, S. j., G. C. White, .and F. L. Knopf. 2002. Advanced techniques for modeling avian nest sur- vival. Ecology 83:3476-3488. Entrett, D. D., D. W. SPE.AKE. .ANT) W. K. Maddox. 1980. Natality and mortality of a north Alabama Wild Turkey population. Proceedings of the Na- tional Wild Turkey Symposium 4:117-126. F.arnsworth, G. L. .ANT) T. R. Simons. 2000. Obser- vations of Wood Thrush nest predation in a large contiguous forest. Wilson Bulletin 112:82-87. Goltd, F. W! 1975. Texas plants: a checklist and ecolog- ical summar} . Texas Agricultural Experiment Station Bulletin MP-5 85/Revised. College Station. USA. Gr-vnt, T. a., T. L. Silaffer. E. M. M.adden, ant) P. J. PiETZ. 2005. Time-specific variation in passerine nest survival: new insights into old questions. Auk 122:1-12. Hern.antdez. F, D. Rollins, .vnt) R. C.antl’. 1997. Evaluating evidence to identify ground-nest pred- ators in west Texas. Wildlife Society Bulletin 25: 826-831. L ariviere. S. 1999. Reasons why predators cannot be inferred from nest remains. Condor 101:718-721. L. ARIVIERE, S. .and L. R. W.alton. 1998. Eggshell re- moval by duck hens following partial nest dep- redation by striped skunk. Prairie Naturalist 30: 183-185. ^ Leimgruber. R. W. j. McShea, .a.nd J. H. Rappole. 1994. Predation on artificial nests in large forest blocks. Journal of Wildlife Management 58:254- 260. M. ajor. R. E. 1991. Identification of nest predators by photography, dummy eggs, and adhesive tape. Auk 108:190-195. Melton, K. B. 2007. Reproducti\e ecolog> of Rio Grande Wild Turke\ s on the Edwards Plateau. The- sis. Texas A&M Universin. College Station. USA. -Montevecchi. W. .a. 1976. Egg size and the egg pred- atory behavior of crows. Behaviour 57:307-320. Rader. .M. J., T. W. Teinert. L. A. Brennan, F. Her- nandez. N. J. SiLVY. AND X. B. Wl. 2007. Iden- tifying predators and nest fates of bobwhites in southern Texas. Journal of Wildlife Management 71:1626-1630. Randel. C. j.. R. Aglirre. D. A. Jones. J. N. Schaap. B. J. WiLLSEY. M. J. Peterson, and N. J. Silvy. 2005. Nesting ecology of Rio Grande Wild Turkey in the Edwards Plateau of Texas. PrcK'eedings of 910 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4. December 2008 the National Wild Turkey Federation Symposium 9:212-218. R.\nsom Jr., D.. O. J. Ro.ngst.ad, .\nd D. H. Rusch. 1987. Nesting ecology of Rio Grande turkeys. Journal of Wildlife Management 51:435 — 139. Ricklefs. R. 1969. .An analysis of nesting mortalit\ in birds. Smithsonian Conoibutions to Zoolog> 9:1-48. Rollins, D. and J. R Carroll. 2001. Impacts of pre- dation on Northern Bobwhite and Scaled Quail. Wildlife Society Bulletin 29:39-51. Staller, E. L. 2001. Identify ing predators and fates of Northern Bobw hite nests using miniature video cam- eras. Thesis. Uni\ersit>- of Georgia, Athens, USA. Shaffer, T. L. 2004. A unified approach to analyzing nest success. Auk 121:526-540. Stephens, S. E., J. J. Rotella, M. S. Lindberg, M. L. Taper, .v.nd J. K. Ringelm.an. 2005. Duck nest survival in the Missouri Coteau of North Dakota: landscape effects at multiple spatial scales. Eco- logical Applications 15:2137-2149. TrE\ OR, j. T. R. W. SE.ABLOOM. AND R. D. Sa^T-ER. 1991. Identification of mammalian predators at artificial w aterfowl nests. Prairie Namralist 23:93-99. Westerskov, K. 1950. Methods for determining the age of game bird eggs. Journal of Wildlife Man- agement 14:56-67. White, G. C. .and R. A. Garrott. 1990. Analysis of wildlife radio-tracking data. Academic Press. San Diego. California. USA. The Wilson Journal of Ornithology 120(4):910-913. 2008 No Evidence for Spring Re-introduction of an Arbovirus by Cliff Swallows Valerie A. O'Brien.* Amy T. Moore.* Kathryn P. Huyvaert.- and Charles R. Brown*-’ .\BSTRACT. — We sampled 100 Cliff Swallows (Petrochelidon pyrrhonota). just after arrival in Ne- braska breeding areas, to ascertain if migrating birds re-introduce Buggy Creek virus (BCRV; Togaviridae) to north-temperate localities in spring. Most birds sam- pled were pre^■iously banded and w ere known to have used parasite-free nesting colonies in past summers and/or were seronegative to BCRV: thus, they were unlikely to have been previously exposed to the vims in their breeding areas. None of the birds had evidence of viral RNA in blood, as measured by RT-PCR. These results are consistent with other studies that have shown little evidence that migrator\’ birds re-introduce arbovimses to temperate localities between years. Re- ceived 14 February 2008. Accepted 5 May 2008. Whether arthropod-home vimses (arbovi- mses) are re-introduced in spring by migra- tory birds in temperate latitudes is a major question in the study of bird-associated vims- es (Reeves 1974. Scott and Weaver 1989. ' Department of Biological Sciences. University of Tulsa. Tulsa. OK 74104 USA. - Department of Fish. Wildlife, and Conservation Biology. Colorado State University, Fort Collins. CO 80523^'USA. ^ Corresponding author: e-mail: charles-brown @ utulsa.edu Crans et al. 1994). Arbovimses are rarely found in over-wintering insect vectors such as mosquitoes (Rosen 1987. Reeves 1990. Day 2001). and the conventional wisdom is that infected birds from the tropics — that are fed upon by insect vectors (e.g.. mosquitoes) after arrival in breeding areas — may provide a mechanism for annual recurrence of vims in temperate latitudes of central and northern North America (Cilnis et al. 1996. Unnasch et al. 2006). Empirical evidence for this scenario is limited, however, and consists mostly of a few records of birds (bound for unknown des- tinations) with eastern equine encephalomy- elitis vims when captured in spring after crossing the Gulf of Mexico (Calisher et al. 1971). Demonstrating vims re-introduction re- quires sampling birds upon their arrival at breeding sites sufficiently early in the nesting season that re-infection by local vectors can be excluded if positive birds are found. No studies have systematically surveyed newly arrived migratory birds for arbovimses. Buggy Creek vims (Togaviridae. Alphavi- rus) is an unusual arbovims vectored primar- ily by the swallow bug (Hemiptera: Cimici- dae; Oeciacus vicarius). an ectoparasite of the SHORT COMMUNICATIONS 911 colonially nesting Cliff Swallow (Petrocheli- don pyrrhonota). Vertebrate hosts for Buggy Creek virus (BCRV) are Cliff Swallows and House Sparrows {Passer domesticiis) that oc- cupy nests in swallow colonies (Hayes et al. 1977, Scott et al. 1984). A related alphavirus. Fort Morgan virus, is a strain of BCRV (Pfef- fer et al. 2006). BCRV, although not docu- mented to affect humans, is phylogenetically and serologically related to western equine en- cephalomyelitis virus (WEEV) (Calisher et al. 1988, Powers et al. 2001), and WEEV affects both people and livestock (Reisen and Monath 1989). Birds have been suggested to move WEEV between North and South America (Weaver et al. 1997). We sampled Cliff Swallows for virus im- mediately after the birds’ arrival at breeding sites in southwestern Nebraska, USA as part of our efforts to understand the population dy- namics of BCRV and its association with Cliff Swallows (Brown et al. 2001; Moore et al. 2007; Brown et al. 2007, 2008). , Our objective was to examine whether these birds were in- fected with BCRV upon their return and could potentially re-introduce the virus to their breeding areas. Cliff Swallows breed throughout much of North America, nesting in large colonies un- derneath cliff overhangs and bridges, and win- ter in southern Brazil, Uruguay, and northern Argentina (Brown and Brown 1995). BCRV occurs annually in our Nebraska study area and is commonly isolated from the insect vec- tors (Brown et al. 2001, 2007; Moore et al. 2007). Its predictable annual occurrence sug- gests the virus either over-winters in swallow bugs and/or in resident House Sparrows, or is re-introduced each season by Cliff Swallows when they return from their winter range in South America. METHODS Long-term work in our study area (in Keith, Garden, Lincoln, and Morrill counties, Nebra.s- ka) indicates the first Cliff Swallows appear on about 18 April each year with numbers slowly increasing during the following 10 days (Brown and Brown 1996: 443). The first arrivals tend to concentrate at the same 2-3 colony sites in the study area (C. R. Brown, pers. obs.). We mist-netted Cliff Swallows between 23 and 29 April 2(X)6 and 2(X)7 at two colony sites (41° 15' N, 101° 37' W; 41° 13' N, 101° 37' W) that contained most birds present in the study area at that time. The early sampling dates ensured that birds at both sites were newly arrived. Both colony sites sampled had been fumigated mul- tiple times per summer during the previous 10 seasons to remove swallow bugs, suggesting that few bugs were present in April and the like- lihood of any bird being infected by a bug after arrival and before sampling was low. The in- secticide used is highly effective against swal- low bugs (Brown and Brown 2004). Fumigation procedures are described by Brown and Brown (1996). Birds caught were bled by Jugular or brachial venipuncture, in which 0.1 ml of blood was col- lected and placed in 0.4 ml of BA-1 diluent (Moore et al. 2007). Samples were centrifuged and 25 |jl1 of supernatant was added to 100 pi of a guanidine thiocyanate-based lysis buffer. RNA was extracted after the addition of 100 pi of 100% ethanol using the QIAmp Viral RNA Mini Kit (Qiagen), following the manufacturer’s protocol. A positive BCRV control was includ- ed in each extraction. Reverse-transcription PCR (RT-PCR) was performed using the OneStep RT-PCR Kit (Qiagen) following the manufacturer’s proto- col. We used BCRV-specific primers that yielded a 208-bp fragment from the E2 region of the viral genome. Primer sequences and thermocycler conditions are described in Moore et al. (2007). Product (6.5 pi) was elec- trophoresed on a 4% Nusieve/agarose gel to identify any positive pools, using at least one BCRV positive control on each gel and a 100- bp ladder. This protocol was used for detect- ing BCRV in both swallow bug pools and sera of nestling House Sparrows, which are com- monly infected (about 25% of bug pools and >20% of nestling sparrows; Moore et al. 2007; V. A. O’Brien and C. R. Brown, unpubl. data). Our RT-PCR methods have also detect- ed BCRV in sera of Cliff Swallows during the summer nesting season, including samples confirmed by both RT-PCR and plaque assay on Vero cells (V. A. O’Brien and C. R. Brown, unpubl. data). RESULTS We captured 1(X) Cliff Swallows during the sampling pericxls in the 2 years. None of the 1(X) birds had evidence of circulating BCRV 912 THE WILSON JOURNAL OF ORNITHOLOGY • VoL 120, So. 4, December 2008 RNA in blood, as judged from RT-PCR. All but 14 birds had been banded in the smd\ area in a pre\ ious breeding season. Eight) -one were at least 2 years of age and their histors of breed- ing-colony use was known for at least one pre- \ ious year. Fi\ e had been banded the pre\ ious year as recently fledged juveniles. Fift\ -nine of the 86 birds ith past histories were know n to have been resident at onl\ fumigated colonies in the past (the same sites sampled in this smdy). 13 had used only non-fumigated sites in the past, and 9 had used both fumigated and non-fumigated sites in previous seasons. The 5 juveniles had been capmred at fumigated colo- nies a few days after fledging. DISCUSSION Birds that had used parasite-free sites in past seasons w ere unlikeh to have been e.xposed to BCR\’ in a pre\ ious summer and therefore not likely to show latent, chronic infections (as seen for some arboviruses; Reisen et al. 2003). The 47 birds sampled in 2007 were tested for BCRV'-specific antibodies using a plaque reduc- tion neutralization test (Huyaven et al. 2008); none of these birds was seropositive (G. R. Young and N. Komar, pers. comm.). Thus, the individuals sampled in this smd\ were well suit- ed to smdying whether virus could be intro- duced bv migrants that were infected prior to arrival in breeding areas. Hayes et al. (1977) sampled 52 adult Cliff Swallows for the Fort Morgan strain of BCRV on 30 -May 1974 in northeastern Colorado. —215 km from our study area. That study used plaque assay and found no evidence of BCRV in sw allow s. Hayes et al. ( 1977) concluded that no evidence existed for spring re-introduction of V irus bv returning birds, although their samples were taken sufficiently late in the season that birds had begun egg-lav ing at time of sampling. Using a more sensitive assay (RT-PCR). we also found no evidence of circulating BCR\^ (i.e.. viral RN.A) in blood of adult Cliff Swallows, and our birds had arrived at most onlv a few days before sampling. Most birds we sampled had probably not been exposed to BCR\' in breeding areas, by v irtue of their use of fumigated colony sites in past years ( and. for some, their lack of antibod- ies to BCRV). Thus, they were prime candidates for transporting virus from wintering areas or from stopov er sites en route. Surv eys for BCRV hav e not been conducted in South .America, and whether it occurs in wintering areas is unknow n. BCRV is found at Cliff Swallow colony sites south of our studv area, for example in west central Oklahoma, about 750 km from the Ne- braska smdy area (Hopla et al. 1993; C. R. Brown. V. A. O'Brien, and A. T. -Moore, unpubl. data). Migrating Cliff Swallows conceivably could be infected there and mov e the v irus north to Nebraska. Our results are consistent with the absence of direct ev idence that migrating birds, re-in- troduce arbov iruses to temperate localities. It is more likely these viruses persist annually by over-wintering in insect vectors or alter- native resident hosts. Over-wintering of BCRV in swallow bugs is known to occur (Hayes et al. 1977. Rush et al. 1980. Strickler 2006). House Sparrows may be more suitable hosts for BCRV than Cliff Swallows, at least in summer (V. -A. O'Brien and C. R. Brown, unpubl. data), and may prov ide another mech- anism for annual persistence of virus. This is especially true if BCRV is maintained v ia la- tent. chronic infections in venebrate tissue over long periods of time (Huyvaert et al. 2008). The role of migratory birds in re-intro- ducing arboviruses to temperate latitudes in spring is unclear, and we urge all studies find- ing ev en negative ev idence for re-introduction be reported. -ACKNOWLEDG-MENTS We thank the School of Biological Sciences of the University of Nebraska for use of the Cedar Point Bi- ological Station; Ginger Young. Nicholas Komar, and the Centers for Disease Control for serological results: and C. E. Braun. R. E. Carleton. and an anonymous reviewer for helpful suggestions on the manuscript. This work was supported by the National Science Foundation 1 051 4824 » and the National Instimtes of Health .AI057569). LITER-ATURE CITED Brown. C. R. and .M. B. Brow^. 1995. Cliff Swallow t Hinmdo pyrrhonota i. The birds of North .Amer- ica. Number 149. Brow>:. C. R. .vnt) M. B. Brown. 1996. Coloniality in the Cliff Swallow : the effect of group size on so- cial behavior. University of Chicago Press. Chi- cago. Illinois. US.A. Brow^. C. R. ANT) M. B. Browt;. 2004. Empirical measurement of parasite transmission between groups in a colonial bird. Ecology 85:1619—1626. Browt^. C. R.. -M. B. Browtc. .a.. T. .Moore. .\nd .N. SHORT COMMUNICATIONS 913 Komar. 2007. Bird movement predicts Buggy Creek virus infection in insect vectors. Vector- Borne and Zoonotic Diseases 7:304-314. Brown, C. R., N. Komar, S. B. Quick, R. A. Sethi, N. A. Panella, M. B. Brown, and M. Pfeefer. 2001. Arbovirus infection increases with group size. Proceedings of the Royal Society of London, Series B 268:1833-1840. Brown, C. R., M. B. Brown, A. Padhi, J. E. Foster, A. T. Moore, M. Pfeefer, and N. Komar. 2008. Host and vector movement affects genetic diver- sity and spatial structure of Buggy Creek virus (Togaviridae). Molecular Ecology 17:2164-2173. Calisher, C. H., K. S. C. Maness, R. D. Lord, and P. H. Coleman. 1971. Identification of two South American strains of eastern equine encephalomy- elitis virus from migrant birds captured on the Mississippi delta. American Journal of Epidemi- ology 94:172-178. Calisher, C. H., N. Karabatsos, J. S. Lazuick, T. P. Monath, and K. L. Wolff. 1988. Reevaluation of the western equine encephalitis antigenic com- plex of alphaviruses (family Togaviridae) as de- termined by neutralization tests. American Journal of Tropical Medicine and Hygiene 38:447-452. CiLNis, M. J., W. Kang, and S. C. Weaver. 1996. Genetic conservation of Highlands J viruses. Vi- rology 218:343-351. Crans, W. j., D. E Caccamise, and J. R. McNelly. 1994. Eastern equine encephalomyelitis virus in relation to the avian community of a coastal cedar swamp. Journal of Medical Entomology 31:711- 728. Day, j. E 2001. Predicting St. Louis encephalitis virus epidemics: lessons from recent, and not so recent, outbreaks. Annual Review of Entomology 46: 111-138. Hayes, R. O., D. B. Francy, J. S. Lazuick, G. C. Smith, AND E. P J. Gibbs. 1977. Role of the Cliff Swallow bug {Oeciacus vicarius) in the namral cycle of a western equine encephalitis-related alphavirus. Jour- nal of Medical Entomology 14:257-262. Hopla, C. E., D. B. Francy, C. H. Calisher, and J. S. Lazuick. 1993. Relationship of Cliff Swallows, ectoparasites, and an alphavirus in west-central Oklahoma. Journal of Medical Entomology 30: 267-272. Huyvaert, K. P, a. T. Moore, N. A. Panella, E. A. Edwards, M. B. Brown, N. Komar, and C. R. Brown. 2008. Experimental inoculation of House Sparrows {Passer domesticus) with Buggy Creek virus. Journal of Wildlife Diseases 44:331-340. Moore, A. T, E. A. Edwards, M. B. Brown, N. Ko- mar, AND C. R. Brown. 2007. Ecological corre- lates of Buggy Creek virus infection in cimicid swallow bugs Oeciacus vicarius, southwestern Nebraska, 2004. Journal of Medical Entomology 44:42-49. Pfeefer, M., J. E. Foster, E. A. Edwards, M. B. Brown, N. Komar, and C. R. Brown. 2006. Phy- logenetic analysis of Buggy Creek virus: evidence for multiple clades in the western Great Plains, United States of America. Applied and Environ- mental Microbiology 72:6886-6893. Powers, A. M., A. C. Brault, Y. Shirako, E. G. Strauss, W. Kang, J. H. Strauss, and S. C. Weaver. 2001. Evolutionary relationships and systematics of the alphaviruses. Journal of Virol- ogy 75:10118-10131. Reeves, W. C. 1974. Overwintering of arboviruses. Progress in Medical Virology 17:193-220. Reeves, W. C. 1990. Epidemiology and control of mosquito-bome arboviruses in California, 1943- 1987. California Mosquito Vector Control Asso- ciation, Sacramento, USA. Reisen, W. K. and T. P. Monath. 1989. Western equine encephalomyelitis. Pages 89-137 m The arbovi- ruses: epidemiology and ecology. Volume 5 (T. P. Monath, Editor). CRC Press, Boca Raton, Florida, USA. Reisen, W. K., R. E. Chiles, V. M. Martinez, Y. Fang, AND E. N. Green. 2003. Experimental infection of California birds with western equine encephalo- myelitis and St. Louis encephalitis viruses. Jour- nal of Medical Entomology 40:968-982. Rosen, L. 1987. Overwintering mechanisms of mos- quito-borne arboviruses in temperate climates. American Journal of Tropical Medicine and Hy- giene 37(Supplement):69S-76S. Rush, W. A., D. B. Francy, G. C. Smith, and C. B. Cropp. 1980. Transmission of an arbovirus by a member of the Family Cimicidae. Annals of the Entomological Society of America 73:315-318. Scott, T. W. and S. C. Weaver. 1989. Eastern equine encephalomyelitis virus: epidemiology and evo- lution of mosquito transmission. Advances in Vi- rus Research 37:277-328. Scott, T. W, G. S. Bowen, and T. P. Monath. 1984. A field study on the effects of Fort Morgan virus, an arbovirus transmitted by swallow bugs, on the reproductive success of Cliff Swallows and sym- biotic House Sparrows in Morgan County, Colo- rado, 1976. American Journal of Tropical Medi- cine and Hygiene 33:981-91. Strickler, S. a. 2006. Winter ecology of a bird-as- sociated arbovirus. Thesis. University of Tulsa, Tulsa, Oklahoma, USA. Unnasch, R. S., E. W. Cupp, and T. R. Unnasch. 2006. Host selection and its role in transmission of ar- boviral encephalitides. Pages 73-89 in Disease ecology: community structure and pathogen dy- namics (S. K. Collinge and C. Ray, Editors). Ox- ford University Press, Oxford, United Kingdom. Weaver, S. C., W. Kang, Y. Shirako, T. Rumenapf, E. G. Strauss, and J. H. Strauss. 1997. Recom- binational history and molecular evolution of western equine encephalomyelitis complex al- phaviruses. Journal of Virology 71:613-623. 914 THE \MLSON JOURNAL OF ORNITHOLOGY • Vol. 120. So. 4, December 2008 The Wilson Journal of Ornithology 1204 »:9 14— 915. 2008 Marbled Godwit Collides with Aircraft at 3.700 m Carla J. Dove'-^ and Court Goodroe- ABSTR.\CT. — On 25 August. 2007. a Marbled Godwit (Umosafedoa) was struck by a Southwest .Air- lines Boeing 737 at 3.700 m. The bird was identified b> examination of feather remains recovered from the aircraft and represents an altimde record for this spe- cies. Received 13 December 2007. Accepted 12 April 2008. Bird/aircraft collisions (birdstrikes) are an increasing hazard to aviation safetx and cause millions of dollars of damage to both civil and military aviation each year (Dolbeer et al. 2000). Height distributions of major groups of birds that collide with aircraft have been re- cently investigated using data from civil air- craft birdstrikes. but anempts to make species comparisons with these data were not possible because only —23 of the birdstrikes above 152 m were identified to the species level (Dolbeer 2006). Increasing the percentage of birds identified to species that are involved in birdstrikes will improve techniques used to re- duce birdstrikes as well as aid in our under- standing of flight pauems and altimdes of dif- ferent species of birds. We describe the identification of a high al- timde birdstrike involving a Marbled Godwit (Limosa fedoa ) with an aircraft that had taken off from El Paso. Texas. US.A. OBSERVATIONS On 25 .August 2007. a Southwest .Airlines Boeing 737 aircraft struck a bird at 1051 hrs MDT. The bird was struck just east of El Paso. Texas during the *climb‘ phase of flight on deparmre. The altimde was recorded by the flight crew at the time of the strike at about 3.700 m above ground level. The plane re- mmed to El Paso, landing 24 min after takeoff with damage to the left horizontal stabilizer. ■ National Museum of Natural History. MRC 116. P. O. Box 37012. Washington. D.C. 20013. US.A. - Southwest .Airlines Company. 2702 Lovefield Drive. Dallas. TX 75235. US.A. -Corresponding author: e-mail: dovec@si.edu fuselage, and radome. Sufficient feather re- mains from the wings and body of the bird were salvaged and submitted to the Smithson- ian Instimtion. The bird was identified as a Marbled Godwit based on color patterns and size of the feathers in comparison with mu- seum specimens. It is unknown how many birds were struck or if additional species were involved in this event. DISCUSSIO.N This is the highest known altimde record for a godwit. The next highest record was a Bar-tailed Godwit {Limosa lapponica) found dead on Mt. Ruapehu (New Zealand) by hik- ers at 2.610 m in January 2006 (Battles and Horn 2006). Previous published high altimde birdstrike records identified to the species lev- el include RuppelEs Vulmre (Gyps nieppelUi) at 11.300 m (Layboume 1974). and Mallard {Anas plaryrhynchos) at 6.400 m (Manville 1963). The Marbled Godwit involved in this birdstrike was within the range of the southern fall migration for this species ( Grano-Trever 2000) and documents that a fairly short-dis- tance migrant can attain high flight altimdes during migration. .Accurate birdstrike repon- ing detailing such data as in this event, to- gether with molecular techniques that now aid in species level identifications from blood and tissue samples, will improve ornithological knowledge of flight altimdes of birds. .ACKLNOWXEDG.MENTS We thank the Southwest .Airlines crew for accurately reporting this birdstrike. M. .A. Heacker and N. C. Rotzel assisted with the bird identification. LITER.ATU RE CITED Battley. P. F. A.VD C. Horn. 2006. .A high-altimde Bar- tailed Godwit {Limosa lapponica \ on Mt Ruapehu. North Island. New Zealand. Notomis 53:381-382. Dolbeer R. .A. 2006. Height distribution of birds re- corded by collisions with civil aircraft. Journal of Wildlife Management 70:1345-1350. Dolbeer. R. .A.. S. E. Wrjght. ast) E. C. Cle-arv. SHORT COMMUNICATIONS 915 2000. Ranking the hazard level of wildlife species to aviation. Wildlife Society Bulletin, 28:372-378. Gratto-Trever, C. L. 2000. Marbled Godwit (Limosa fedoa). The birds of North America. Number 492. Laybourne, R. C. 1974. Collision between a vulture and an aircraft at an altitude of 37,000 feet. Wil- son Bulletin 86:461-462. Manville, R. H. 1963. Altitude record for Mallard. Wilson Bulletin 75:462. The Wilson Journal of Ornithology 120(4):916, 2008 The 2008 William and Nancy Klamm Service Award In 1953, a young man from Steubenville, Ohio, attended his first Wilson Ornithological Club meeting with his parents. This was the beginning of a more than fifty-year relation- ship between Richard C. Banks and the Wil- son Ornithological Society. Dick joined the Wilson Ornithological Society in 1959 as a graduate student, and since then his service has been varied and extensive. He has been active on numerous committees including the Wilson Prize Committee, selection committee for the Fuertes, Nice, and Stewart Awards, Publications Committee, Scientific Program Committee, and the Resolutions Committee, and in many instances he served as Chair of the committee. He has served as an elected Council member (1979-1982), Second Vice- President (1987-1989), First Vice-President (1989-1991), and President of the Society (1991-1993). During his presidency, Dick es- tablished the Committee on Undergraduate Outreach to help stimulate an interest in or- nithology among undergraduate students, and to help maintain and focus that interest to stimulate students to continue studies in or- nithology. He was the first editor of the Or- nithological Newsletter, which was estab- lished as a quarterly publication in 1976 and led to the formation of the Ornithological So- cieties of North America (OSNA). He contin- ued as editor of the newsletter until 1989. He co-chaired the Local Committee for the 2005 annual meeting in Beltsville, Maryland. He has given his time reviewing many manu- scripts for the Wilson Bulletin, and has also made more tangible donations, both to the Van Tyne library and the Centennial Fund. Dick continues to be very active on Wilson Coun- cil, and has modeled professional behavior and responsibility for many Wilson members at various stages in their careers. He is a staunch supporter of ornithology, ornitholo- gists, and the Wilson Ornithological Society. It is with the utmost respect and gratitude that we present the 2008 William and Nancy Klamm Service Award to Richard C. Banks. — Sara R. Morris, W. E. Davis Jr., Je- rome A. Jackson, John Kricher, and Doris Watt (Klamm Service Award Committee). 916 The Wilson Journal of Ornithology 120(4):917— 926, 2008 Ornithological Literature Compiled by Mary Gustafson and Robert B. Payne THE MEINERTZHAGEN MYSTERY: THE LIFE AND LEGEND OF A COLOS- SAL FRAUD. By Brian Garfield. Potomac Books Inc., Washington, D.C., USA. 2007: xiv + 353 pp, 14 black and white photo- graphs. ISBN 978-1-59797-041-9. $19.25 (hardcover). ISBN 978-1-59797-160-7. $12.57 (paper). — Richard Meinertzhagen (1878-1967) spumed a career in the family banking business to join the British military, attaining the rank of colonel, and in 1921 was transferred to the Foreign Office, where he fancied himself to have been engaged mainly in espionage. Most of his postings in both ca- pacities were overseas in Africa and Asia, and he boasted and wrote expansively on his ex- ploits as a warrior and spy. Along the way he nurtured a passion for natural history, amassed a large personal collection of birds, and pub- lished books and papers in ornithological jour- nals. He also stole specimens of birds from museum collections, fabricated label data for them, and published deliberately falsified in- formation about birds in scientific journals — facts that have only been brought to light in recent years through painstaking research and scientific investigation. Many of us in the mu- seum community who had long been aware of Meinertzhagen’s ornithological perfidies had come to wonder if the rest of his history might prove to be just as grand a prevarication. Brian Garfield set out to answer exactly that question because, in his words: “The preci- sion and energy that scientists invested in re- examining his zoological cons were not rep- licated in other circles. There was no effort to verify and collate his separate parallel paths.” In short, as the subtitle makes clear, Meiner- tzhagen’s entire legacy proves to be a gigantic lie. Although he was “a noted figure in sci- ence, war, and espionage ... in all three fields he also was a fantasist who perpetrated colos- sal deceptions.” Garfield, according to his rather robustious web site, is mainly a prolific American writer of crime and mystery fiction whose consid- erable success hinges on his insistence on en- tertaining the reader. That is not necessarily the best qualification for a historian, even though Meinertzhagen’s egregious outrages make his history more entertaining than most. One will find attributions for which no source is cited and conclusions based on evidence that at times seem less than satisfactory. But this is partly due to the nature of lies them- selves, as it is much more difficult to prove that someone didn’t commit a particular act than to prove that he did. Thus, much of Gar- field’s case against Meinertzhagen rests on the fact that no corroboration can be found for practically any of his alleged major deeds as a warrior and spy. As might be expected, Garfield has little understanding of ornithologists or what they do in museums. He confuses taxonomy and taxidermy. Hugh Whistler (whose bird skins numbered some 17,000) and other scientific ornithologists are referred to as “birders.” Meinertzhagen’s activity in museums in Lon- don, Tring, and Berlin is said to have been “cataloging birds.” It should be Erwin, not Irwin Stresemann, etc. Meinertzhagen was a compulsive chronicler of himself. But his in-law Malcolm Mugger- idge considered that Meinertzhagen’s “copi- ous diaries ... in elegant leather [that] adorned the walls of his study . . . would prove to be a monument to his fantasy self.” And so it appears. There seem to have been no original journals kept in the field, at least none that survived, and only the fictitious fan- tasies remain. One very interesting document that Garfield cites is a 66-page typescript entitled “Con- quisitio Mea. A Record of my Ornithological Activities and Collections.” In 1942, Meiner- tzhagen sent this to Alexander Wetmore, then Assistant Secretary of the Smithsonian Insti- tution, with a view towards having it pub- lished in the United States. “Conquisitio” is Latin for “a bringing together” but can also mean “collecting.” The first 1 1 pages consist of a brief history of Meinertzhagen's collect- ing, his methods of preparing and storing 917 918 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 specimens, his philosophy of nomenclature, etc. The rest treats specimens that he consid- ered to be the highlights of his collection, some of them now known to have been stolen. Before returning the manuscript in 1944, Wet- more had a copy made to insure its preser- vation in the event of the sinking of a mail ship or similar calamity. In all likelihood, however, he would also have harbored suspi- cions about Meinertzhagen's veracity and probably wanted to insure that a copy was preserved for posterity. I examined this doc- ument in the Smithsonian Archives and found that it is not a “photocopy” as Garfield states, but a laborious retyping of the entire manu- script by an extremely talented stenographer. Its importance lies in the fact that this “is one of the few documents of length from the pe- riod that [Meinertzhagen] did not have an op- portunity to destroy or revise beyond recog- nition” (page 260. note 19). Garfield cites the Conquisitio (hereafter CqM) in the following passage (page 47): In the 1940s he wrote, intending this for publication, that in June 1897 he went trekking into Lapland on a camping-bird- ing expedition. In the same document he wrote that his brother Dan had gone into Lapland a year later, in 1898. Perhaps we are meant to infer that brother Dan was copying Richard's trail-blazing pioneer work. In fact, it was Dan who went north through Scandinavia into Lapland in 1897. Richard did not go there until many years later; in 1897 he was at work in the bank in London. On the same page Garfield also refers to Meinertzhagen's “clumsy later attempt to pur- loin credit for the Lapland expedition” as sug- gesting that the relationship with his brother “may have been shaded with envy.” Unfor- tunately for the truth of Garfield's claim, there is absolutely no mention whatever in CqM of Richard going to Lapland in 1897. His first foreign collecting after escaping the “miser- able slavery” of the bank appears to have been as a “junior subaltern of a British Reg- iment in the deserts of Rajputana'" (CqM. page 4). Leading the list of “Collecting Are- as” represented in his collection is: “1898 My brother Dan made a trip to Arctic Lapland. returning with a large collection of eggs and several skins” (CqM. page 6). 1898 is erro- neous for the date of collection, but being the year of Dan’s death it could easily be viewed as the date obtained by Richard. Therefore, if Meinertzhagen actually asserted that he had been to Lapland in 1897. we are left to guess where he may have done so. Dan's death was a traumatic event in Meinertzhagen's life so that “purloining” credit for the Lapland ex- pedition is a potent accusation that for now remains undocumented. Was this the result of fault} memory, bad organization of notes, or some other deficiency in Garfield’s scholar- ship? Garfield tells us that Meinertzhagen wrote “that both Thomas Huxley and Charles Dar- win were quite fond of him. and that Darwin dandled Richard on his knee” (page 41). He rightly questions the dandling as being highly unlikely. But he fails to document where Mei- nertzhagen made such a statement, which a careful historian would surely have footnoted. The assertion happens to come from CqM and Garfield could have had a lot more fun with it than he did. as the passage epitomizes Mei- nertzhagen's infinite capacit}' for lying. “The passion for birds was bom in my brother Dan and myself. At an early age we pored over Gould's Birds of Great Britain and. later on. over Lilford’s Birds of the British Islands, un- til we knew every bird by name and sight. It was about then that we sat in the laps of two venerable white-haired old gentlemen. Profes- sors Huxley and Darwin, and at the feet of Henry Seebohm. fresh from his discovery of the breeding grounds of the Grey Plover” (CqM. pages 1-2). Let's apply a little chro- nology to this farrago of nonsense: RM was bom in 1878. the reclusive Darwin died four years later in 1882 after years of debilitating illness, the first volume of Lilford's work ap- peared in 1885 to be finished posthumously in 1897. and Seebohm discovered the Grey Plo- ver's nest in 1875, three years before Meiner- tzhagen was bom. In addition to being a thief and habimal liar, there is a cloud of plagiarism, attributed to an anonymous source, over Meinertzhagen’s book on birds of Egypt and. as had been doc- umented before, over parts of his book on Arabian birds as well. Far worse, however, is the suspicion, well-justified in my opinion. ORNITHOLOGICAL LITERATURE 919 that Meinertzhagen was a wife-killer. His sec- ond wife, Annie Jackson, was an accom- plished ornithologist in her own right before marrying Richard, and was active in museum and field, being a crack shot. But she evi- dently caught on to Richard’s lying ways and was aware that he was publishing false infor- mation about Asian birds and stood to expose him. She changed her will to exclude Richard in favor of their children and shortly thereaf- ter, on her remote estate in Scotland, was killed by a single pistol bullet that penetrated her head and spine. Richard was the only oth- er person present. Garfield obtained the police reports (which had wrongly been rumored to have been or- dered to be kept sealed for some long period of time) and reviewed the evidence and opin- ions of this shocking episode, ultimately con- cluding that (pages 171-172): “For the re- cord, RM is innocent until and unless he should be proven guilty. . . . The case for RM’s innocence derives in part, and with iro- ny, from his habit of lying.” That is, because all of his tales about killing people in wars were shown by Garfield to be unsubstantiated. Thus, the argument begins to take on the as- pect of a school logic problem — given that a person always tells a lie, determine which of the following of his statements must therefore be true. There was never any suggestion that Annie was suicidal. The local papers, presum- ably reflecting what Meinertzhagen told the police, reported the case as an accident and Meinertzhagen wrote in his diaries that he and Annie had been at target practice, he had gone to retrieve a target and heard a shot behind him, turned and saw Annie dead on the ground. Later he evidently told more than one person that they had had a duel. If the as- sumption is that Meinertzhagen always lied, then his wife did not die in an accident or a duel. Ergo, he killed her. So much for the ly- ing defense. Meinertzhagen was long infatuated with his much younger cousin Theresa Clay, who, fol- lowing Annie’s death in 1928, became Rich- ard’s collaborator and companion. Theresa went on to become an authority on bird ec- toparasites, although she was evidently com- plicit in fabricating data with Meinertzhagen, which bears more comprehensive investiga- tion. Meinertzhagen had first been caught steal- ing bird specimens from the British Museum in 1919 and was banned from the collection until Walter Rothschild interceded on his be- half and he was readmitted in 1921, whereaf- ter he continued to steal from that and other collections, although he was long held in sus- picion by British Museum administrators. He was also caught stealing books and valuable color plates and Garfield relates that the police were called in on more than one occasion but charges were never filed. Some of Meiner- tzhagen’s bogus war exploits were incorporat- ed into military histories and are still perpet- uated. On and on he went, unchecked, pro- tected by his status in society, his intimidating personality, and the phlegmatic British aver- sion to creating any sort of unpleasant distur- bance. Garfield (page 10) alludes to a class “cover- up” and this cover-up appears still be in ef- fect. It is a disturbing aspect of his book that so much is attributed to anonymous sources. Many of those now willing to cast Meiner- tzhagen in a negative light still do not wish to be identified. A long quote about how Mei- nertzhagen perpetrated his Asian frauds and Annie realizing what he had done is from a “key ornithological source, who has asked for anonymity” (page 169). A “lady ornitholo- gist” reports that Meinertzhagen asked her on a mountain climbing date in a “deadly voice” and opined that he was a mad man and had murdered his wife (page 171). On the same page, the rumor of a duel is reported in “a recorded interview [with] one of Meinertzhag- en’s relatives.” An endnote in this passage does not identify the relative nor say where the recording may be housed but merely re- ports more suspicions from “a well known science journalist” and “another of his rela- tions” (page 303, note 37). Two notes down, a motive for Annie’s killing is “offered by an ornithologist who wrote to me on 17 June 2006.” Why is the date important if the name is omitted? A long note about the late South African ornithologist P. A. Clancey's relationship with Meinertzhagen “is related third-hand by var- ious sources” (page 327, note 12), which is certainly how Garfield must have received the report that Clancey “traveled with RM on sa- fari in 1948-49, and later reported RM's noc- 920 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 turnal sprees to several fellow ornithologists, apparently including Storrs Olson” (page 311, note 7). Nowhere is there an indication of the identity or significance of Storrs Olson. A well-known curmudgeonly ornithologist, who shall go unnamed, reports to the readers of this review that Storrs Olson has only a distant memory of meeting Clancey and no recollec- tion whatever of anything that he may have said. Near the end of his text, Garfield char- acterizes the whole saga in a pithy summa- tion: “From a sociological angle the Mei- nertzhagen story makes for a fascinating study of ruling-class dynamics. From a psy- chiatric angle it makes for an eye-popping case study in narcissistic pathology” (page 246). I trust his explanation that his book was “not written to settle scores or to right personal wrongs. It is an examination of a life that was lived and then revised and then believed” (page 251, note 7). Its various shortcomings aside, Garfield’s book is a sterling contribution towards unraveling the Meinertzhagen mystery and it provides powerful testimony in support of the fact that we cannot believe a word that Richard Meinertzhagen ever wrote or said about any- thing. But we also need to be wary of what has been written about him by others as well. — STORRS L. OLSON, Curator, Divi- sion of Birds, National Museum of Natural History, Smithsonian Institution, Washington, D.C. 20560, USA; e-mail: olsons@si.edu LOST WORLDS: ADVENTURES IN THE TROPICAL RAINEOREST. By Bruce M. Beehler. Yale University Press, New Ha- ven, Connecticut, USA. 2008: 258 pages, 11 black-and white wash paintings, 10 maps, and 19 black-and-white photographs. ISBN: 978-0-300-12228-2. $28.00 (cloth).— This book is the story of Bruce Beehler’s more than 30 years of research and conservation work in rainforests around the world. It is a book of adventure in wild places, and the sci- entific and socio-political lessons learned in the turmoil of conservation efforts in devel- oping nations. Beehler is an ornithologist and much of what he presents has an ornitholog- ical focus. The book consists of a Preface, an Introduction, 10 chapters, an Epilogue, a Bib- liography, and an Index. The Preface begins in a rainforest camp setting in Papua (western New Guinea, formally Irian Jaya), Indonesia, where he is escaping a desk job working for the U.S. Department of State. He explores his reasons for being there, and the explanation sets the tenor for the rest of the book: “. . . it was the absolute necessity of determining how the best of these rainforests could be preserved for generations to come — so that our great-grandchildren could hear the rain- forest whisper its secrets.”. . . “This book, then, is a story of the rainforest and its over- looked importance to humankind’s long-term well-being.” Chapter 1, In the Rainforest, traces Beeh- ler’s initial encounter with the rainforests of Papua New Guinea (eastern New Guinea) in 1975, brought there by the allure of birds of paradise, bowerbirds, honeyeaters, and cas- sowaries. He recounts his adventures in field camps, conversing with natives in Neomela- nesian Pidgin, and in general learning the ways of the rainforest, its birds and people. Chapter 2, In the Zone and on the Plantation, takes the reader half a world away from New Guinea to Panama and the famous Pipeline road near which his group from Princeton University camped. He compares the similar- ities with New Guinea rainforests (e.g., dom- ination by a canopy of hardwoods, lots of vines and Hanes) and contrasts the differences (e.g., the presence of monkeys in Panama, and the presence of an indigenous forest-dwelling people in New Guinea). The scene then shifts to the rainforests of India in Chapter 3, On the Trail of Ripley and Ali, and Chapter 5, Bio- diversity and Intrigue across the Inner Line. In 1981 Beehler had joined S. Dillon Ripley, Secretary of the Smithsonian Institution as Ripley’s scientific assistant, and for a decade set up Ripley’s bird surveys in India. He re- counts his difficulties with the Indian bureau- cracy, habitat alteration on a grand scale, and his interactions with Ripley and Salim Ali, the dean of Indian ornithologists with whom Rip- ley had worked for four decades. The “Inner Line” is the politically sensitive frontier where foreigners are not usually allowed. Here Beehler encountered hostility, due in part to a negative reaction to perceived “sci- ORNITHOLOGICAL LITERATURE 921 entific imperialism” practiced by first world scientists. In Chapter 4, Wallace’s Promised Island, a pleasant interlude on the Atherton Tableland of Queensland, Australia, is followed by a di- sastrous visit to a Freeport mining camp high in the mountains of Papua (Irian Jaya), where they were kept under virtual house arrest by the Indonesian military and not allowed to conduct the bird surveys they were there to do. By 1999 Beehler was working with Coun- terpart International, a non-govemmental or- ganization, on a project called Forest Gardens, an experimental agroforestry program that en- couraged tropical rural farmers to conserve standing forest through a system akin to grow- ing shade coffee. Beehler visited Ivory Coast in West Africa (Chapter 6), and reports on the conservation difficulties posed by the contin- ual movements of landless displaced people and the resultant clearing of forests by squat- ters. Ivory Coast experienced the strife and political disintegration that threaten conser- vation projects throughout the developing world. In Chapter 7, Local People Really Do Count, Beehler recounts an expedition to Pa- pua New Guinea in 1993 (he has visited New Guinea 40 times) to a field camp on the Na- gore River. He describes the camp life in de- tail, and then analyzes the conservation les- sons that he learned there, which include the mistake of underestimating the local people and their commitment to conserving their for- est (they have been doing it for centuries). The message to take away is that local people are critical for successful conservation initiatives. Chapter 8, Pitiful Scraps of Forest, deals with a Forest Gardens restoration project on Cebu Island in the Philippines, a place plagued by problems of landless migrants that had squat- ted and destroyed much of central Cebu Na- tional Park, due in part to lack of any decisive government oversight. Beehler comments: “No question, the single greatest threat to the world’s rainforests is population. ... If the world cannot manage its population in a ratio- nal way through smart family planning and strong economic incentives, problems galore are ahead for the earth’s remaining rain- forests.” Chapter 9, Lemurs, Vangas, Chameleons, and Poverty, describes the conservation prob- lems in another impoverished island, Mada- gascar. We learn of lemurs (endemic primates) and endemic families of birds in the rich rem- nant fragments of rainforest, but are reminded that these remnants are in a sea of human- devastated landscape, deforested to provide charcoal, together with plantations and rice paddies, the products of a burgeoning human population. The final chapter. The Lost World, describes the adventures of Beehler and his group in a high mountain camp in a pristine area of the Foja Mountains in Paupa (Irian Jaya) in 2005. After arriving by helicopter, they surveyed the plants and animals and re- solved long-standing mysteries regarding a bird of paradise, a bowerbird, and a tree-kan- garoo. The book is about adventure, marvelous birds and mammals, and the trials and tribu- lations of conservation of rainforest in devel- oping countries. In the Epilogue, Beehler re- iterates what he learned about conservation in each of his forays into the rainforests of the planet and concludes with the statement: “We conservationists have to insure that all threads tie together: good governance, sensible eco- nomics, strong planning, enforcement, en- gaged local stewardship and, yes, creation and management of protected areas to preserve the most precious places on earth.” Despite the bewildering conservation and bureaucratic problems the author has encountered, the book has a tone of guarded optimism. The book is attractive with each chapter be- ginning with a black-and-white painting. The maps place you in the various rainforests and the photographs illustrate important people, places, and habitats. The book is directed at a general readership: only common names are given for organisms — a single-page appendix with scientific names would have been use- ful— and the bibliography (no in-text cita- tions) contains only 35 entries. Nevertheless, the book contains much for the professional ornithologist and conservationist. The stories are well-written and sparkle with adventure. I recommend this delightful well-written book to anyone with an interest in rainforest birds or the conservation of rainforest ecosys- tems.—WILLI AM E. DAVIS JR., Professor Emeritus, Boston University, 23 Knoll wood Drive, East Falmouth, MA 02536, USA: e- mail: wedavis@bu.edu 922 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 BIRDS OF THE DOMINICAN REPUB- LIC AND HAITI. By Steven Latta, Christo- pher Rimmer, Allan Keith, James Wiley, Her- bert Raffaele, Kent McFarland, and Eladio Fernandez. Barry Kent MacKay, Tracy Ped- ersen, and Kristin Williams, principal illustra- tors. Princeton University Press, Princeton, New Jersey, USA. 2006: vii + 258 pages, 57 color plates, 2 detailed maps, 1 line drawing, and numerous range maps. ISBN: 978-0-691- 11890-1, $75.00 (cloth). ISBN 978-0-691- 11891-8 $35 (paper). — Mention the Domini- can Republic (DR) to most Americans and they think of Albert Pujols or any number of excellent baseball players from that tiny coun- try. To the knowledgeable tropical birder, though, thoughts of the DR immediately bring to mind 31 endemic species and literally doz- ens of endemic subspecies found on that is- land. Among the endemics are such unusual forms as Palmchat {Dulus dommiciis). Least Parauque (Siphonorhis brewsteri), and Bay- breasted Cuckoo (CoccyzKS rufigularis), plus an assortment of ground-tanagers, palm-tana- gers, chat-tanagers, todies, and crows. Fortu- nately, most of these species are still reason- ably common and can be found in the many new national parks and preserves that have been developed in the DR in recent years. Al- though a budding ecotourism industry cater- ing to birders has developed in recent years, the Dominican Republic still seems fairly safe for an independent traveler who can speak a little Spanish (albeit Spanish that seems to function without the letter s). With this new field guide and a good map, he or she has a chance to see a majority of these island en- demics in a short period of time. This field guide begins with an introduction to the island of Hispaniola, which supports the countries of the Dominican Republic and Hai- ti. We learn this is an interesting place with what is now a single island having been split in two at various times in the past. This helps explain some of the rather unusual distribu- tional patterns of some of the endemic birds, where new species evolved on the separate is- lands and either stayed in place, with ecolog- ically equivalent species having ranges on what effectively were the two separate islands, or spread across the island, allowing two sim- ilar species to occupy most of the island but separate ecologically in some fashion (as in the two todies). The many endemic species and subspecies are listed and discussed. A section covers the advances made in conser- vation on this island in recent years with a list of the mostly new national parks. The prob- lems facing these parks and birds outside of protected areas are also presented. A series of 57 color plates fills the second section of the book. Most of these plates were copied from those used in A Guide to the Birds of the West Indies by Herbert Raffaele and others. Because Raffaele developed his book as part of his job in the federal govern- ment, he has been able to let ornithologists from other West Indian islands use parts of this material for developing regional bird books. Latta et al. took good advantage of this fact, although they also added 105 new im- ages of Hispaniolan birds in this section. These plates do a wonderful job of illustrating the West Indian endemics and provide some images of the many winter residents on this island; one may still want to bring along a field guide from home to deal with more un- usual plumages among these winter residents. The bulk of the book (—200 pages) covers species descriptions, with usually about two species per page. Most species have a map of the island that shows their distribution, al- though that is also described in the text in more detail. Much of the distributional mate- rial comes from the annotated checklist of Al- lan Keith and collaborators: The Birds of His- paniola: Haiti and the Dominican Republic. These species accounts are quite complete, and often provide some natural history of the species involved. Common names for each species in both the DR and Haiti are provided. Appendix A has nine pages of detailed de- scriptions of birding sites in both countries. Appendix B provides a checklist. Two indices are provided, one with local names and one with scientific and common names. A list of references ends the book. Although I reviewed only the English ver- sion of Birds of the Dominican Republic and Haiti, both Erench and Spanish versions of this book were produced at the same time as the English version. This was possible, in part, because of the reduced investment in color plates. The Dominican Republic has made ad- mirable strides in recent years with regard to conserving natural resources such as birds ORNITHOLOGICAL LITERATURE 923 through developing reserves and changing laws. Having a book that is written in your language and about your birds can be an im- portant rallying point for local biologists, as they deal with the local poverty and lack of federal support for their efforts. For this rea- son alone, the authors of this book are to be commended for producing such a fine regional work, as it undoubtedly will help preserve birds in the Dominican Republic in the future. Unfortunately, conditions are so environmen- tally devastated in Haiti that even a wonderful field guide written in French will probably do little to save birds there. That certainly is not the fault of this well written field guide. — JOHN FAABORG, University of Missouri, Columbia, MO 65211, USA. e-mail: faaborgj@ missouri.edu AVES BRASILEIRAS E PLANTAS QUE AS ATRAEM [BRAZILIAN BIRDS AND PLANTS THAT ATTRACT THEM]. By Jo- han Dalgas Erisch and Christian Dalgas Frisch. Dalgas Ecoltec, Sao Paulo, Brazil. 2005: 480 pages, 98 color photographs, 193 color plates, and 1254 maps. ISBN 85-85015- 07-1. (cloth). AVES BRASILEIRAS: MINHA PAIXAO (BRAZILIAN BIRDS: MY PAS- SION). By Johan Dalgas Frisch. Dalgas Ecol- tec, Sao Paulo, Brazil. 2005: 186 pages, many color and black-and-white photographs. ISBN 85-850 15-08-X. (cloth). $49.00 for both books plus postage (from www.avesbrasileiras.com. br). — The first of these two books, both of which are virtually entirely in Portuguese and come as part of a single-price package, rep- resents the third edition of a guide first launched in 1981. In its current incarnation it purports to provide an accurate field guide to Brazilian birds (the bulk of the book), but also to provide an illustrated manual to what would be termed in the U.K. as “gardening for wild- life”. To this end, the father and son Frisch team devote just fewer than 60 pages to ex- plaining how to attract wild birds to your Bra- zilian garden, should you be fortunate to pos- sess one (most Brazilians don’t, even amongst the expanding middle classes), by planting certain trees and shrubs. Other chapters of this unconventional field guide are devoted to the national bird concept (Frisch senior was in- strumental in Brazil’s adoption of the Rufous- bellied Thrush {Turdus rufiventris) as the country’s national bird), unraveling the mys- teries of migration, nature’s influence on poets and writers, the history of the present work, and the importance of creating habitat for birds and, by extension, other wildlife. Natu- rally, most readers of this review will be in- terested in how this book might serve them as a guide to identifying Brazilian birds. It is claimed that the field guide follows the latest thinking on avian classification, al- though the systematics used are, in fact, those of Sibley and Monroe, which were certainly ground-breaking at the time (1990), but in to- day’s fast-moving world of “molecular orni- thology” look somewhat dated. The authors would have done better to have followed the list published online by the Brazilian Comite Brasileiro de Registros Ornitologicos (http:// www.cbro.org.br), which is constantly updat- ed to reflect novel developments in nomencla- ture and taxonomy, as well as on the basis of new records, or if they preferred a book for- mat list, that presented in the third edition of Howard and Moore (E. C. Dickinson. [Editor]. 2003. The Howard and Moore Complete Checklist of the Birds of the World. Christo- pher Helm, London, UK). The material is arranged with text and maps facing the illustrations of the relevant species in “true” field guide format. The text, how- ever, is minimal, providing merely the Brazil- ian, English, scientific, and Spanish names, as well as each species’ total length with, at the foot of each page, an explanation of the ety- mology of the scientific name (in Portuguese). Identification features, habitat, vocalizations, etc., are all left to the imagination. It might also be added that in some particularly “busy” double-page spreads, for instance the 21 species of Tyrannidae packed onto pages 234-235, the relevant etymology section ap- pears elsewhere, in this example on page 62! The maps are clear and well produced, albeit at a scale that will never permit detailed elu- cidation of range, and seem comparatively free of dramatic errors, although needless to say in a work of this scope there are some. They also endeavor to portray range through- out South America, not just in Brazil. This book contains >3, ()()() individual im- ages. Despite this number, a great many dis- 924 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 tinctive female (and other) plumages are not illustrated and, even worse (perhaps especially from the point of view' of the visiting birder), well over 500 species that have occurred in Brazil, including several distinctive or excit- ing endemics, are not even mentioned, much less depicted. (Although the authors mention the approximate number of species (1,800) to have occurred in Brazil in their introduction, the number omitted from this book seems to be nowhere admitted.) As a result, keen visi- tors eager to add such sought-after species as Black-hooded Antwren {Fonnicivora ery- throfiotos), Restinga Antwren {F. lirtoralis), or Marsh Antwren {Stymphalornis aciitiros- tris), to name just three Thamnophilidae, will be unable to identify them using this book. As further examples, most species of Laridae, and a great many shorebirds and pelagic birds are all left undepicted. Although the introduction to the new edition does not clearly state at whom the book is directed, the emphasis is presumably on the amateur Brazilian (of which it might be stated there are still very few who specifically travel to watch birds purely as a hobby), but even then the choice of species is still confusing. One might just understand the rationale for including Spix’s Macaw' {Cyanopstitta spixii), although it should have been coupled with a note con- cerning its extinction in the w ild, but why il- lustrate Eskimo Curlew {Nwneniiis borealis). which judging from the map might still be ex- pected just about anyw'here in southern and western South America! It goes without say- ing that in the unlikely event the person for- tunate enough to re-encounter the near-myth- ical Kinglet Calyptura (Calyptiira cristata) is reliant on this book; they will not find it here- in. The illustrations themselves sadly leave much to be desired. The same attenuated pro- file appears to have been used as a basis for every raptor, and a similar “color by num- bers” charge can be leveled at other groups too. Bill shapes are often at significant odds with reality, colors frequently range from far too gaudy (most tinamous and pigeons, for ex- ample) to the opposite extreme, and all spe- cies are shown perched, even seabirds, swifts, and hirundines. It should be remarked that the observer familiar with Brazilian, or South American birds, will still be able to identify a great many of the more boldly marked species (e.g., tanagers) using this book, but the beau- tifully intricate plumages displayed by many woodcreepers, nightjars, and owls appear al- most monstrous as depicted here. Some im- ages do stand out as being much better than the remainder, including all those associated with the “gardening for wildlife” section, and some of the toucanets, for example, within the field guide section, presumably because they were painted by a different artist, apparently Tomas Sigrist, based on the style, although his name does not appear to be credited. With the recent publication of the long- awaited Birds of Peru (Schulenberg et al. 2007. Princeton University Press. Princeton, NJ, USA), birders traveling to South America particularly await modem field guides for Bra- zil and Bolivia. Fortunately, several for the former country are in production, so that just like the “from rags to riches” tale that has unfolded for those birding the Indian Subcon- tinent in recent years, we may in the future witness a situation where the birder is faced not with the challenge of finding a reliable book to carry in the field, but deciding which book to take. That time is not yet come, but in the interim, there are two better field guides to Brazilian birds than the one reviewed here. All the Birds of Brazil (Todas as Aves do Brasil) by Deodato Souza (2006. Subbuteo Natural History Books, Shrewsbury, UK) has been on the market for some years, in both English and Portuguese editions. The illustra- tions are small and frequently poor, but the brief text frequently covers vocalizations, mainly cribbed from Sick. Ridgely and Tudor, or other reliable sources. More recently pub- lished is Tomas Sigrist’s dual-language field guide (2007. Birds of Eastern Brazil! Aves do Brazil Oriental, Sao Paulo, Brazil), which boasts superior illustrations to any other Bra- zilian field guide. Indeed, some of the plates, including many raptors and antbirds, by Ed- uardo Brettas are of the same standard as those adorning major North American or Eu- ropean field guides. Unfortunately, Sigrist's work solely relies on a key to impart infor- mation concerning habitat and distribution, and there is nothing on vocals or hints on identification, despite their being quite consid- erable white space left unused on most pages facing the plates. Both these latter works are ORNITHOLOGICAL LITERATURE 925 less bulky and more accurate than that by the Frisch team. The second part of the present package is a book, entirely in Portuguese, subtitled the life and work of Johan Dalgas Frisch, and this re- counts his efforts at recording Brazilian bird song, campaigning for the implementation and protection of Tumucumaque National Park on the Brazilian frontier with French Guiana and Suriname, and documenting the migration of U.S. -breeding Purple Martins {Progne subis) to and from the State of Sao Paulo, in south- east Brazil, amongst other subjects. Frisch has plainly understood the imperative of interest- ing and gaining the support of the rich and powerful, if one is to succeed in conservation work, as numerous photographs in the book prove. Given the large number of environ- mental challenges facing Brazil, more char- acters like Frisch with the capacity to popu- larize birds and other wildlife, will be needed. Also of importance, as has been demonstrated elsewhere in the world, is the production of an accurate, cheaply priced, and appropriately sized field guide to the country’s birds in the native language, to engender a large-scale in- terest in birds, and their conservation, among the general public.— GUY M. KIRWAN, 74 Waddington Street, Norwich NR2 4JS, UK; e- mail: GMKirwan@aol.com CODING AND REDUNDANCY: MAN- MADE AND ANIMAL-EVOLVED SIG- NALS. By Jack P. Hailman. Harvard Univer- sity Press, Cambridge, Massachusetts, USA. 2008: xiii and 257 pages, 25 figures, 333 ref- erences. ISBN 978-0-674-02795-4. $39.95 (cloth). — The introduction reviews the history of ideas in animal behavior, communication, and its coding. The other chapters describe the principles of behavior codes. Analogies are used to match cases of human signals in man- made machines and animal signals with many examples for birds, several from studies by the author. The chapters describe many levels of com- plexity of signals with examples for each lev- el. How are headlight-tail light colors of cars like gender plumage markers in woodpeckers; weather radio alarm calls like bird nest-de- parture calls; indicator lights on waffle irons like wing-flagging in copulating gulls; or a smoke-detector test like a Herring Gull {Lams argentatiis) parent and chick in begging for food? Answers: these are cases of binary cod- ing, where one signal has two states: the first, simple one-bit signals (flicker male and fe- male); the second, event markers when the signal is on briefly when the status changes; the third, state indicators (signal is on while a state is maintained); the fourth, status replies where a check by one individual either elicits a signal in response or gets no response (if chick pecks at red spot on parent’s bill, then parent regurgitates food). Other classes of signals involve more than two states (“multi-valued coding’’) and more than two kinds of signals (“multivariate cod- ing’’). For a man-made signal, the author de- scribes the Morse code. Samuel Morse in- vented telegraphy and the code that made it a useful means of communication in the early 1800s. A sender presses a key to close an electric circuit, releases the key to open the circuit; this sends a signal each time the key is pressed and each time it is released. Each letter is composed of clicks with short inter- vals, or long intervals, or a mix of short and long intervals between them. In English, the most frequently used letter was E (13% of the letters), next T, next A and O, then I, N, R; and the least used letters were Q and Z (less than 1%). Morse assigned shorter signals for frequently used letters and longer signals for less frequent letters. His code included the time between clicks within a letter (short or long), between letters (a longer time), and the frequency of use of the letters in his signal design. A single short click (., or dot) was used for E, a long click (_, or dash) was used for T, three short clicks for S (...). a short click and then a long click for “A" (._), and so on. The code is designed for efficiency in time of signal transmission, as it takes less time to code and transmit a common letter than a less common letter. In a bird signal, the Black-capped Chicka- dee {Poecilc attricapillns) “chick-a-dee" phrase in the presence of a predator has more “dee" notes when the predator is small than when it is large. This makes sense in terms of “design", as small predators (such as a small owl) are more likely to attack a chickadee than large predators (a large hawk), so the higher 926 THE W ILSON JOURN.\L OF ORNITHOLOGY • Vol. 120, So. 4. December 2008 number of "dee” notes per call is more likely to signal imminent danger to the social part- ners of the calling chickadee. Both Morse code and chickadee alarm calls are cases of multi-valued coding. Other signals are multivariate, either dis- crete or composite. A man-made example is Braille, a discrete linguistic code used by the blind and detected by touch. The code in- volves the presence or absence of raised dots in three rows of dots in tw o columns, where the combination of raised dots codes a letter. The code consists of six on/off variables, so a composite signal can have two (exp 6) = 64 different values, which represent letters and numerals and other things. A bird example is the pattern and color of duck wings. In eastern North America, each species of dabbling duck has a unique com- posite visual signal. The author recognizes eight colors that occur in different combina- tions on seven parts of the wing. The seven spatial elements are like seven code words, and each patch can have one of the eight col- ors. These signals may let the ducks son into single-species flocks when a mixed group takes flight. We do not know that ducks sort this \\ ay. or that there is an adaptive value to be in several small flocks rather than a large mixed flock, but the potential information is there. A novel approach of the book is the appli- cation of information theory to a wide range of signaling systems in humans and animals such as birds. Hailman explains how to esti- mate the infonnatiou in signals. The text and an appendix use the terms and equations to estimate the parameters of information {H = log2 n): entropy: surprisal (S = log’ p). or "news”, as in "when a dog bites a man that is not news, but when a man bites a dog that is news”: and redundancy of simple and com- plex signal coding. However, not much use is made of these parameters, and the "informa- tion” from the signal sender is not always used by the receiver. Not all states are equally likely, and this can be taken into account in describing the information in binary coding. The estimates are not used to predict differ- ences between species. Nor are the informa- tion-theory parameters compared across dif- ferent classes of signals (such as kinds of be- havior). The author explains the coding of the ISBN number of the book as an error-check- ing case of "designed redundancy”. He ad- vises the math can be skipped when reading the book. The book is clearly written and a fascinat- ing read, as well as informati\ e and interest- ing. It shows ordinary readers as well as bi- ologists how to recognize connections be- tween the design of codes and behavior sig- nals. and to see the great variety in animal signals, in many behaviors and many kinds of birds.— ROBERT B. PAYNE. Professor Emeritus. University of Michigan. 1306 Granger Avenue. Ann Arbor. MI 48109. USA; e-mail: rbpayne@umich.edu The Wilson Journal of Ornithology 120(4):927— 940. 2008 PROCEEDINGS OE THE EIGHTY-NINTH ANNUAL MEETING JOHN A. SMALLWOOD, SECRETARY The eighty-ninth annual meeting of the Wil- son Ornithological Society (WOS) was held Thursday 17 April, through Sunday, 20 April 2008 at the Mobile Convention Center, Mo- bile, Alabama in joint session with the Asso- ciation of Field Ornithologists (AFO), at the invitation of the University of Southern Mis- sissippi (USM) Migratory Bird Research Group and sponsored by the Gulf Coast Bird Observatory, Mississippi Coast Audubon So- ciety. University of Southern Mississippi, Dauphin Island Sea Laboratory, University of California Press, USM Coastal Research and Extension Center, and the Grand Bay National Estuarine Research Reserve (NERR). Frank R. Moore, Director of the USM Migratory Bird Research Group, chaired the local com- mittee, which also included Robert Diehl, John Dindo, Jennifer Owen, and Mark Wood- rey. The WOS Council met from 1307 to 1802 hrs on Thursday, 17 April, in Room 105-B of the Mobile Convention Center. That evening there was a social at the Center for the con- ferees and guests. The Council met again from 1121 to 1306 hrs the following day. The opening session on Friday convened in a large space at the Center known as Room 210-BCD at 0806 hrs with welcoming re- marks from Frank R. Moore. WOS President James D. Rising, and AFO President Cecilia M. Riley. Timothy J. O'Connell. Chair of the WOS Student Travel Awards Committee, of- fered comments on those awards, and Mark Woodrey of the local committee provided in- formation concerning field trips. The opening ceremony continued with a presentation by Edward H. Burtt Jr. on Margaret Morse Nice, which led to his introduction of Jerome A. Jackson, who concluded the opening cere- mony by delivering the .Margaret Morse Nice Plenary Lecture. “Thinking like a mountain, seeing like a woodpecker: behavioral ecology and conservation of woodpeckers." The scientific program included 58 contrib- uted papers and 47 contributed posters, which were organized into eight paper sessions, a poster session, and a symposium on migration and coastal ecology. In addition, WOS hosted the Margaret Morse Nice Lecture, and the As- sociation of Field Ornithologists hosted the AFO Plenary Lecture. “How* phylogenies can guide research in other fields: examples from hummingbirds.” presented by J. Van Remsen. On Friday evening the poster session was held in conjunction with a reception at the Center. In addition to self-guided bird-watching op- ponunities in the vicinity of the Mobile Con- vention Center, the local committee organized several longer field trips. These included two trips each to Dauphin Island and the Pasca- goula River, and a post-conference trip to Grand River National Estuarine Research Re- serve. The conferees enjoyed a 60-min reception prior to the annual banquet, which was held in a very large space at the Center known as Room 201-ABCD. After an enjoyable dinner, WOS President James D. Rising introduced the new AFO president. David N. Bonter. President Bonter led those assembled in an ex- pression of gratitude to Frank R. Moore. Chair of the local committee, and the many persons who had worked hard to make the conference a success: this expression took the form of a collective bout of exuberant hand clapping. After presenting AFO awards. President Bon- ter returned the podium to WOS President Rising, who in turn thanked the three elected members of the WOS Council who had com- pleted their terms of office. Robert L. Curry, Daniel Klein Jr., and Douglas W. White, and welcomed the three newly elected members of Council, Jameson F. Chace. Sara R. Morris, and Margaret A. Voss. The following WOS awards and commendations also were pre- sented: MAR(;ARKT MORSE MCE MEDAL (for the W'O.S plenary lecture) Jerome A. Jackson, “Thinking like a moun- tain. seeing like a woodpecker: behavioral ecology and conservation of woodpeckers.” 927 928 THE W ILSON JOURNAL OF ORNITHOLOGY • Vol. 120. No. 4. December 2008 EDWARD'S PRIZE (for the best major anicle in volume 119 of The Wilson Journal of Ornithology) James A. Cox and Gary L. Slater. "Coop- erative breeding in the Brown-headed Nut- hatch.'* WILLIAM AND NANCY KLAMM SER- VICE AWARD (for distinguished service to the Wilson Or- nithological Society) Richard C. Banks. LOUIS AGASSIZ FUERTES AWARD Andres Cuervo. "The evolutionary assem- bly of a species rich avifauna; avian speciation and differentiation in the Andean cloud for- est.” PAUL A. STEWART AWARDS Anya Hies. "Do female Stripe-headed Spar- rows outsing their mates during territorial in- cursions?” Irene Liu. "Female eavesdropping and male song type matching in a songbird.” Trina Schneider-Bayard. "Quantifying cues: testing the influence of social cues on habitat selection behavior.” Andrea Townsend. "Optimal dynamics of sociality, relatedness, and parasites in Ameri- can Crows: linking theoretical predictions with empirical data.” ALEXANDER WILSON PRIZE (for best student paper) Cunis W. Budney. "Comparative phylo- geography of neotropical birds: ecology pre- dicts levels of genetic differentiation.” LYNDS JONES PRIZE (for best student poster presentation) Zoltan Nemeth. "Phenotypic organ flexibil- ity around the annual cycle in two Nearctic- neotropical migratory thrush species.” NANCY KLAMM BEST UNDERGIU\D- UATE STUDENT OIU\L PAPER AWARD Kelly Hallinger. "Lifetime fitness of Tree Swallows e.xposed to aquatic mercury.” NANCY KLAMM BEST UNDERGRAD- UATE STUDENT POSTER AWARD Jack M. Stenger. "The bacterial degradation of phaeomelanic and eumelanic feathers.” WILSON ORNITHOLOGICAL SOCIETY TIUWEL AWARDS Eric Beck. Cameron University. "Oklahoma marshbird monitoring project.” Bethany Belock. Saint Mary's College. "An example of the role of weight, plumage color, cere color, and total area of UV reflec- tant plumage on mate selection in Budgerigars (Melopsittacus undulates)." Peggy Buckley. Canisius College. "A com- parison of migrant species composition at two western New York television towers.” Ryan Burdge. College of William and Mary. "Avian pesticide exposure and food in- take on golf courses.” Curtis Burney. Louisiana State University. "Comparative phylogeography of neotropical birds; ecology predicts levels of genetic dif- ferentiation.” Vince Cavalieri. Oklahoma State Universi- ty. "Scale effects on occurrence and relative abundance of forest songbirds in eastern Oklahoma.” Katie Chmielowiec. Canisius College. "A comparison of fall migration and stopover by Northern Saw-whet Owls during irruptive and nonirruptive years in coastal Maryland.” Charles Clarkson. University of Virginia. "Food supplementation, territory establish- ment. and song in the Prothonotary Warbler.” Jason Counter. Eastern Kentucky Universi- ty. "Responses of flocks of Tufted Titmice to different-sized raptors.” Katherina Eorgues. Trent University. "Making the connection between shorebirds and off-road vehicles.” Maura Hanna. Canisius College. "Morpho- logical differences in Gray Catbirds across age classes.” Bruce Hitch. Auburn University. "Do avian abundances differ between habitat types? Im- plications for assessment of habitat quality.” Ian Horn. Eastern Kentucky University. "Eastern Phoebes use different strategies to provision young.” Mikaela Howie. College of William and Mary. "The infiltration of aquatic mercury into a terrestrial ecosystem.” Jason Jacobs. Canisius College. "Seasonal differences in energetic condition of Blackpoll Warblers on Appledore Island.” Erik Johnson. Louisiana State University. ANNUAL REPORT 929 “Ectoparasites affect bird condition in neo- tropical forest fragments.” Christine Lattin, Eastern Kentucky Univer- sity, “Is song length an important signal of aggression for Blue Grosbeaks? A playback experiment.” Kristen Lear, Ohio Wesleyan University, “Bacterial degradation of flight and body con- tour feathers by B. lichenifonnis.'' Kim Martinezack, Saint Mary’s College, “Relationships between dominance measures and physical characteristics of the Budgerigar (Melopsittacus undulates)." Damion Marx (deeeased), Florida Atlantic University, “Relationships among peak an- nual wading bird nest effort, lake stage, and hydrologie reeession at Lake Okeechobee, Florida, USA.” Emily Runnells, Hobart and William Smith, “Fat stores and energetie eondition of Catha- rus thrushes during spring and autumn migra- tion at a Great Lakes stopover site.” Staey Stefan, College of Charleston, “Sea- sonal variations in recoveries of South Caro- lina-banded Brown Pelieans and Royal Terns: a eomparison of two seabird speeies experi- encing a decline in local nesting numbers.” Jack Stenger, Ohio Wesleyan University, “The bacterial degradation of phaeomelanic and eumelanie feathers.” Bethany Stephan, Canisius College, “Com- parison of avifauna in three riparian environ- ments in western New York: the impact of anthropogenic habitat alteration.” Amanda Stockwell, Canisius College, “Mi- gration and stopover ecology of the Tennessee Warbler at an inland stopover site near Kala- mazoo, Michigan.” Ryanne Sullivan, Canisius College, “An- nual variation in the stopover eeology of the Tennessee Warbler in Michigan.” Sean Williams, Ohio Wesleyan University, “Structure of albino feathers: why so weak?” Meredith Wilson, Ohio Wesleyan Univer- sity, “Dynamies of Staphylococcus aureus on bird feathers.” Seleetion committee for the Nice Medal: Doris J. Watt (chair), Charles R. Blem, Wil- liam E. Davis Jr.; for the Edwards Prize: Clait E. Braun (chair), J. Daniel Lambert, Cynthia A. Stacier; for the Klamm Service Award: Sara R. Morris (chair), William E. Davis Jr., Jerome A. Jackson, John C. Kricher, Doris J. Watt; for the Fuertes and Stewart Awards: Robert B. Payne (chair), Laura Payne; for the Alexander Wilson Prize, the Lynds Jones Prize, and the Naney Klamm undergraduate presentation awards: E. Dale Kennedy (chair), Mary Bomberger Brown, L. Scott Johnson, Daniel Klem Jr., Alan Macearone, John A. Smallwood, Margaret A. Voss, Douglas W. White; and for the WOS Travel Awards: Tim- othy O’Connell (chair), David Bonter, Mia R. Revels. BUSINESS MEETING President James D. Rising ealled the annual WOS Business Meeting to order at 1731 hrs, Friday, 18 April 2008, in Room 201 -BCD in the Convention Center and recognized that a quorum was present. Secretary John A. Small- wood presented a synopsis of the previous day’s Couneil meeting, noting that at the end of March 2008, the society’s membership stood at 1,690 individuals, including 211 stu- dents and 84 new members of all membership eategories; the renewal rate for previous mem- bers was 82% for the current year. In addition, 319 libraries and institutions subseribed to The Wilson Journal of Ornithology. The Sec- retary then asked those gathered to stand in reeognition of the following members who had died sinee the last annual meeting: W. Marvin Davis (University, MS), George A. Hall (Morgantown, WV), Frederick E. Lud- wig (Barbeau, MI), Franklin McCamey (De- catur, GA), Val Nolin (Bloomington, IN), Kenneth C. Parkes (Pittsburgh, PA), Henry W. Pelzel (Wiehita, KS), Darrell W. Pogue (Tyler, TX), R. M. Schramm (Tucson, AZ), Henry T. Wiggin (Brookline, MA), Glen E. Woolfenden (Venus, EL), and Addison Young (Larchmont, NY). Secretary John A. Smallwood then re- marked to those assembled that encouraging students to participate in our meetings and to make presentations in the scientihc programs is a high priority of the Wilson Ornithological Society. Because this was a joint meeting with AEG, the two societies pooled their funds al- located for student travel awards. All 39 stu- dents who applied for travel funds submitted applications that were deemed worthy of some level of linancial support. Because there were so many students and the cost of travel had become so high, the amount of aid requested 930 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 was significantly greater than the $9,000 al- located for travel awards. Secretary Small- wood was pleased to report that WOS Council had allocated an additional $3,600 in travel support for those students participating in this meeting. Secretary John A. Smallwood informed those assembled that Council had unanimous- ly re-elected Clait E. Braun as Editor of The Wilson Journal of Ornithology for 2009, and noted that under Editor Braun’s care the Jour- nal sets a high standard for peer-reviewed publication. Einally, the Secretary expressed WOS Council’s gratitude to the Investing Trustees for their superb management of the society’s investments, especially during the current pe- riod of widespread economic uncertainty. Having completed his synopsis of the Council meeting. Secretary Smallwood first introduced Treasurer Melinda M. Clark, who presented the Treasurer’s Report, and then Editor Clait E. Braun, who presented the Editor’s report. John A. Smallwood, as Chair of the Nom- inating Committee, which also included Mary Bomberger Brown, Edward H. Burtt Jr., and Douglas W. White, presented that committee’s report: Members of Council for 2008-201 1 (3 nominees for 3 positions), Jameson E Chase, Salve Regina University; Sara R. Morris, Can- isius College; and Margaret A. Voss, Penn- sylvania State University-Erie. Those assem- bled were invited to make additional nomi- nations, but no one did so. Thus, President James D. Rising closed the nominations as a result of a motion by Dale Kennedy, seconded by Reed Bowman; that motion was passed unanimously by voice vote. Jerry Jackson then moved and Dan Klem seconded that the Secretary cast a single unanimous vote for the slate of nominees, and by acclamation, it hap- pened that way. Having come to the end of the WOS Busi- ness Meeting agenda. President James D. Ris- ing inquired if any of those assembled had additional items of business. Because no one offered any, Sara Morris moved and Bob Bea- son seconded that the meeting stand ad- journed. This came to pass at 1643 hrs. REPORT OF THE TREASURER Statement of Revenues and Expenses Eor The Year Ending 31 December 2007 2007 2007 2008 12 Months Annual Annual Actual Budget Budget Revenues Contributions $ 1,836 $ 1,200 $ 1,200 Student Travel Research Fund 315 — — Van Tyne Library Book Fund 86 — — Sales — back issues 239 518 518 Sales — books — Van Tyne Libr. 609 500 500 Subscriptions 15,070 17,317 17,317 Page charges 18,390 15,506 15,506 Royalties 4,230 3,409 3,409 BioOne Electronic Licensing 13,469 10,760 10,760 Mailing list rental income 222 660 660 Memberships 26,767 31,332 31,332 Other income 3,625 — — Total revenues from operations $ 84,858 $ 81,202 $ 81,202 Expenses Journal publication costs Editorial honorarium $ 8,000 $ 4,000 $ 4,000 Editor travel/supplies 6,474 1,000 1,000 Editorial assistance 17,291 25,000 25,000 Copyright expense 48 48 Printing — bulletin 7 1 ,645 64,400 64,400 Artwork 900 — — Printing color plates 1,125 2,400 2,400 $ 105,435 $ 96,848 $ 96,848 ANNUAL REPORT 931 Operating expenses Postage & mailing — back issue $ — $ 440 $ 440 Storage back issues 301 680 680 Van Tyne Library — expenses 710 1,500 1,500 OSNA management services 24,221 21,000 21,000 Credit card fees 1,150 1,100 1,100 Travel expenses — OSNA rep 1,500 1,500 Travel expenses — general 788 450 450 Travel expenses — Ornith Council 242 200 200 Meeting expenses 3,747 1,000 1,000 Membership expenses 1,377 2,000 Accounting fees 2,283 4,500 4,500 Insurance — D&O 1,650 1,425 1,425 Office supplies 498 570 570 Postage — general 169 260 260 Other expenses 154 — 35 Filing fees 5 5 5 Discretionary expenses 3,000 3,000 $ 37,295 $ 37,630 $ 39,665 Awards Membership awards $ $ $ 600 Hall/Mayfield — 1,000 1,000 Stewart 2,500 3,000 3,000 Fuertes 2,500 2,500 2,500 Wilson, Lynds Jones, Klamm 1,400 1,200 1,200 Student travel grants 5,125 5,000 5,000 Nice award expenses 2,307 3,000 3,000 $ 13,832 $ 15,700 $ 16,300 Contributions Support — Ornith Council $ 9,000 $ 9,000 $ 9,000 Support — Ornith Council, restricted to revision costs 7,500 7,500 — Am Bird Conservancy Dues — 250 250 AAZN dues — 250 250 $ 16,500 $ 17,000 $ 9,500 Total Expenses $ 173,062 $ 167,178 $ 162,313 Expenses in excess of revenues before investment $ (88,204) $ (85,976) $ (81,111) Investment activity Revenues Investment earnings — budgeted $ — $ — $ — Realized gain/loss — ML 35,219 23,612 23,612 Realized gains/losses — Howland 84,951 18,968 18,968 Realized gains/losses — Sutton 5,329 5,812 5,812 Unrealized gain/loss — ML 21,330 36,722 36,722 Unrealized gain/loss — Howland (40,050) 29,887 29,887 Unrealized gain/loss — Sutton 9,761 4,794 4,794 Investment earnings — ML 26,981 20,()()() 2(),()()() Investment earnings — Howland 63,028 25,000 25,000 Investment earnings — Sutton 3,914 4,200 4.200 Total revenues from investment activity $ 210,463 $ 168,995 $ 168,995 Investment fees $ 28,461 $ 25,091 $ 25,091 Investment revenues in excess of expenses $ 182,002 $ 143,904 $ 143,904 Total revenues in excess of expenses $ 93,798 $ 57,928 $ 62,793 932 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 STATEMENT OF FINANCIAL POSITION 3 1 December 2007 Assets Cash investments Merrill Lynch — cash Coamerica — Van Tyne checking — Van Tyne Univ Michigan account . Sutton Fund — cash equivalents Howland Mgmt — cash equivalent .. Total cash and cash equivalents Other Investments Merrill Lynch — common stocks — Merrill Lynch — corp bonds Merrill Lynch — mutual funds Sutton Fund — equities Sutton Fund — corp bonds Howland Mgmt — equities Howland Mgmt — fixed income Total Other Investments Total Assets Fund Balances Restricted funds — Sutton Fund Unrestricted funds Net Income Fund balance-Klamm Total Fund Balances $ 46,791 1,543 434 12,947 229,775 $ 291,490 766,559 37,990 47,711 155,437 9,303 1,326,209 246,918 2,590,127 $ 2,881,617 $ 177,686 807,231 93,798 901,029 1,802,902 $ 2,881,617 Melinda M. Clark, Treasurer EDITOR’S REPORT This report is for the period 1 January through 31 December 2007. We received 190 new manuscripts in 2007. Only one of these manuscripts was published in Volume 119 (2007), illustrating the backlog of manuscripts from 2006 and earlier. However, 9 received in 2007 are in the March 2008 (Volume 120) Is- sue and 23 will be in the June 2008 Issue (al- ready sent to Allen Press). All four Issues of Volume 1 19 were published on schedule in 2007 and we expect this pattern will continue. We are slowly trying to increase the rejected and withdrawal rates but need stronger re- views to help encourage authors to send their manuscripts to other possible outlets. We have made up the September 2008 Issue (31 manuscripts) and presently have 38 addi- tional accepted manuscripts awaiting publi- cation. This number will more than fill the December 2008 Issue and it is unlikely that any manuscripts received in 2008 will be pub- lished in 2008. Thus, we continue to have a backlog even though we published 826 pages (119 papers) in 2007. I do note that both Con- dor (1,000) and Auk (1,500) publish more pages each year. There is a cost for publishing more pages as cost per page is ~$100. The number of manuscripts received from outside of North America (north of Mexico) is increasing and is expected to increase mark- edly in the future. The Wilson Journal of Or- nithology is seen as a desirable publication outlet, especially by authors trained in North America and now working in Mexico and South America. We published 24 papers in 2007 by authors outside of North America (14 from Mexico and South America). Locations of the senior authors of non-Central or South America manuscripts ranged from China to Is- rael to New Zealand. We published an addi- tional 13 manuscripts by authors in North America on studies completed on species out- ANNUAL REPORT 933 side of North America. We truly are attracting an international audience as we publish more papers from outside of North America. We have added one person to our editorial board to help with manuscripts from Mexico and South America. All manuscripts by authors whose first language is not English take con- siderable effort from the editorial office. We continue to use electronic and paper mail, and all reviewers are contacted by e- mail. Authors in North America are requested to submit one paper copy and an electronic file. Authors from outside of North America can use electronic submission at their discre- tion. The process is functional and can be speedy, depending largely on referees and au- thors promptly returning manuscripts to us. The past 18+ months have convinced me that editing a major Journal is not a part-time “job” as one could spend most of his or her time working on editorial matters. Both the Editor and Editorial Assistant could be full time positions. I look forward to continuing to serve the Wilson Ornithological Society. Clait E. Braun, Editor The reports of the standing committees are as follows; REPORT OF THE UNDERGRADUATE OUTREACH COMMITTEE The Wilson Ornithological Society is wel- coming and generous to undergraduate stu- dents. We hope our efforts will through time sustain an active society membership. Pre- senting a poster or paper at a national orni- thological meeting is a landmark experience for an undergraduate student. In addition to discounted membership dues, we also offer support for participation in meetings, includ- ing travel funds, free one-year memberships, banquet tickets, and awards for paper and poster presentations. Young students are now a major feature at our annual meetings; more than half of the presentations on the program for the Mobile meeting are by undergraduate and graduate students. To maintain our attractiveness in the face of rising meeting costs and competition from sis- ter societies, we should monitor our support levels. On-line discussions prior to the meet- ing have considered offering additional dollars for student travel to meetings. Expanded membership discounts to students are another possibility. To assess how our undergraduate outreach is being received by the students themselves, I am working with the IT Department at Al- bion College to create an on-line forum to so- licit suggestions from undergraduate and grad- uate student members on how the society could be of more use to them. My goal is to have this site up as soon as possible after the annual meeting. I welcome suggestions on topics for discussion threads. Eaculty mentors are the basis for our suc- cess in generating the level and quality of un- dergraduate participation in our meetings that we now enjoy. Consequently, encouraging mentoring faculty to be active in WOS could be our most effective outreach endeavor. The work of the Nominating Committee in iden- tifying and recruiting mentoring professionals is important. Perhaps it would be useful to feature symposia, workshops, or panel discus- sions on mentoring undergraduates or teach- ing ornithology as part of future meetings. Douglas W. White, Chair REPORT OF THE JOSSELYN VAN TYNE MEMORIAL LIBRARY COMMITTEE I am pleased to submit this report of the activities at the Josselyn Van Tyne Memorial Library for the period 1 January through 31 December 2007. Loans Loans of library materials included 25 transactions to 21 members. These loans in- cluded books loaned, and copied and scanned articles. Acquisitions Exchanges: A total of 140 publications was received by exchange from 109 organizations or individuals. Gifts: We received 29 publications from 25 organizations. Subscriptions: We also received 30 publi- cations from 22 subscriptions. We spent a total of $667 on subscriptions in 2007. Donations: we received items from 6 mem- bers and friends: John Farmer, Lloyd Kiff (for 934 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 The Peregrine Fund) Phil Mattocks, Mike McLeish, Bob Payne, and Victor Zhukov, for a total of 257 items, including 27 books and 190 journal issues. Purchases: 19 journal issues and 1 book, $110.50. Dispersals Gifts to Other Institutions: 163 issues of American Birds to the University of Windsor, 130 journal issues to the Peregrine Fund li- brary, 5 books and 1 journal issue to the Arch- bold Biological Station library. Back Issues: We sent out 33 back issues of The Wilson Bulletin for only the cost of post- age. Duplicates: We sold 16 books and 5 journal issues for $863. Accessibility on the Web Web Site: The web site (http://www. ummz.umich.edu/birds/wos/) continues to provide access to the library. Journals cur- rently received are listed on the site as well as how to access the University of Michigan’s on-line catalogue, which interested people can use to check holdings. Books for Sale: We have our duplicate books for sale listed on the web site. Journals for Trade: Also listed on the web site are the journals we have available for sale or trade. Thank You’s Many thanks to our secretary, Janet Bell, for keeping the library loan records, and to our work-study student Rebecca Carter for copying and scanning articles, keeping the li- brary running, and mailing out back issues of The Wilson Bulletin/Journal. Janet Hinshaw has done a wonderful job as the Wilson Or- nithological Society Librarian. Thanks to our student library worker, Olivia Pennebaker, for copying and scanning articles, keeping the li- brary running, and mailing out back issues of The Wilson Bulletin/Journal. Robert B. Payne, Chair REPORT OF THE SCIENTIFIC PROGRAM COMMITTEE The Committee on the Scientific Program was chaired by WOS Second Vice-President Robert C. Beason and AFO First Vice-Presi- dent David N. Bonter, who were assisted by session moderators Mark Deutschlander, Ste- phen Hager, Mikaela Howie, Austin Hughes, Karl Miller, Frank Moore, Scott Rush, Jean- Pierre Savard, and James Tucker. PAPER SESSIONS Migration and Coastal Ecology Symposium Jeffrey Buler, University of Delaware and the uses Wetlands Research Center, “Migra- tory birds and the changing coastal land- scape.’’ Robb Diehl, University of Southern Missis- sippi, “Weather, migratory birds and the coastal setting.’’ Brent Hales, University of Southern Mis- sissippi, “Coastal economic and social devel- opment.’’ Kim Hall, Michigan State University, “Global climate change impacts on coastal landscapes.’’ Paul Hamel, Hardwoods Experiment Sta- tion, USDA Forest Service, “Endangered/ threatened migratory birds and the Gulf Coast.” Eben Paxton, USGS Southwest Biological Science Center, “Population regulation of mi- gratory birds.” Mark Woodrey, Grand Bay NERR and Wy- lie Barrow, USGS Wetlands Research Center, and Bill Vermillion and Mark Parr, Gulf Coast Joint Venture, “Conservation of migratory birds along the Gulf Coast.” General Sessions Arlene Arnold, Bart Ballard, and Thomas Langsheid, Texas A&M University, “Terres- trial habitat use and chronology of migrating birds through southern Texas.” John C. Arvin, Gulf Coast Bird Observa- tory, “Evidence of widespread northward range extensions along a 1000-km gradient from tropical to warm-temperate climates in northeastern Mexico and southern Texas.” Lisa Gardner Barillas and Yong Wang, Al- abama A&M University, “Temporal and spa- tial variations of habitat associations of fall migrating songbirds at an inland site in north- ern Alabama, USA.” Charles R. Blem, Flathead Lake Biological Station, “Nest boxes and conservation of Pro- thonotary Warblers: a 21 -year study.” David Bonter, Cornell Laboratory of Orni- thology, “Eurasian Collared-doves in North ANNUAL REPORT 935 America: colonization dynamics and implica- tions for native doves.” Peggy E. Buckley, Arthur R. Clark, and Sara R. Morris, Canisius College, “Compari- son of migrant species composition at two western New York television towers.” Ryan B. Burdge and Daniel A. Cristol, The College of William and Mary, “Avian pesti- cide exposure and food intake on golf cours- es.” Curtis W. Burney and Robb T. Brumfield, Louisiana State University, “Comparative phylogeography of neotropical birds: ecology predicts levels of genetic differentiation.” John P. Carpenter, Yong Wang, and Hugh Metcalfe, Alabama A&M University, “Ceru- lean Warbler home range estimates and roost site selection in northeast Alabama.” Vincent S. Cavalieri, Timothy J. O’Connell, and David M. Leslie Jr., Oklahoma State Uni- versity, “Scale effects on occurrence and rel- ative abundance of forest songbirds in eastern Oklahoma.” Katie A. Chmielowiec, Canisius College, David Brinker, Maryland Department of Nat- ural Resources, and H. David Sheets and Sara R. Morris, Canisius College, “A comparison of fall migration and stopover by Northern Saw-whet Owls during irruptive and nonirrup- tive years in coastal Maryland.” Charles Clarkson, University of Virginia, “Pood supplementation, territory establish- ment, and song in the Prothonotary Warbler.” Robert L. Curry, Villanova University, “Contributions of George L Gaumer to orni- thology of the Yucatan and Cozumel: throw- ing the baby out with the bathwater?” Anthony C. Dalisio, Sterling College, Wil- liam E. Jensen, Emporia State University, and Timothy H. Parker, Whitman College. “The response of male Dickcissels to geographic song variation.” Mark Deutschlander, Hobart and William Smith Colleges, and Rachel Muheim, Virginia Polytechnic Institute, “Fuel reserves affect migratory orientation of thrushes and spar- rows both before and after crossing an eco- logical barrier near their breeding grounds.” Robert C. Dobbs and Paul R. Martin, Queen’s University, and Wylie C. Barrow Jr., and Clinton W. Jeske, USGS National Wet- lands Research Center, “Hurricane-related de- clines in food availability and migrant land bird abundance at autumn stopover sites.” Ryan M. Dziedzic and Michael J. Hamas, Central Michigan University, “Spring stop- over of forest-dwelling and shrub-dwelling migrants at a Great Lakes coastal wetland: the roles of habitat, arthropods, and phenology.” Megan J. Fitzpatrick, Douglas W. White, and E. Dale Kennedy, Albion College, “The effects of thermal environment on incubation behavior in House Wrens {Troglodytes ae- don)."' Sarah E. Goodwin and W. Gregory Shriver, University of Delaware, “Declines of sound sensitive species in response to increased traf- fic volume.” Stephen B. Hager, Augustana College, Hei- di Trudell, Principia College, Kelly J. McKay, BioEco Research and Monitoring Center, Ste- phanie M. Crandall, University of Illinois Ex- tension, and Lance Mayer, Iowa City, lA, “Bird density and mortality at windows.” Heath M. Hagy, Richard M. Kaminski, Samuel K. Riffell, and Francisco J. Vilella, Mississippi State University, and Kenneth J. Reinecke, USGS Patuxent Wildlife Research Center, “Winter waterfowl dynamics in man- aged moist-soil wetlands in the Mississippi Alluvial Valley.” Kelly K. Hallinger, Rebecka L. Brasso, and Daniel A. Cristol, College of William and Mary, “Lifetime fitness of Tree Swallows ex- posed to aquatic mercury.” Maura E Hanna, Canisius College, Richard Keith and Brenda Keith, Kalamazoo Nature Center, and Sara R. Morris, Canisius College, “Morphological differences in Gray Catbirds across age classes.” Alan “Bruce” Hitch and J. B. Grand, Au- burn University, “Do avian abundances differ between habitat types? Implications for as- sessment of habitat quality.” Mikaela G. Howie and Dan A. Cristol, Col- lege of William and Mary, “The infiltration of aquatic mercury into a terrestrial ecosystem." Austin L. Hughes, University of South Car- olina, “Temporal pattern of vocalization type usage in singing sessions of male tyrant fly- catchers.” Douglas A. James and Ragupathy Kannan, University of Arkansas, “Nesting habitat of the Great Hornbill {Buceros hocornis) in the Anaimalai Hills of southern India." 936 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 L. Scott Johnson. Susan L. Balenger, and Brian S. Masters. Towson University. '‘Do ex- tra-marital affairs make Mountain Bluebirds blue?” Daniel Klem Jr.. Muhlenberg College. “Preventing bird-glass collisions.” Christine Lattin. Eastern Kentucky Univer- sity. '‘Is song length an important signal of aggression for Blue Grosbeaks? A playback experiment.’* Timothy J. Ludwick and Alan M. Fedynich. Caesar Kleberg Wildlife Research Institute. Glenn H. Perrigo. Texas A&M University- Kingsville. and T. Wayne Swertner. Texas Parks and Wildlife Department. “Nesting ecology of urban columbids in South Texas.” Alan D. Maccarone. Friends University. John N. Brzorad. Lenoir-Rhyne College, and Heather M. Stone. Friends University. “Char- acteristics and energetics of Great Egret and Snowy Egret foraging flights.’* Andrew J. McGann and Robert F. Curry. Villanova University. “Songs of the critically endangered Cozumel Thrasher: what would a needle in a haystack sound like?** Bailey D. McKay. University of Minnesota. “Phylogeography of the Yellow-throated War- bler (Dendroica dominica)." Karl E. Miller. Florida Fish and Wildlife Conservation Commission. “Impacts of Hur- ricane Charley on a Florida Scrub- Jay popu- lation.’* Timothy J. 0*Connell. Oklahoma State University. ‘'Investigating the use of Partners in Flight scores for ecological assessment.** Laura M. Palasz and Philip C. Stouffer. Louisiana State University, '‘Assessment of Henslow’s Sparrow abundance and habitat se- lection across Louisiana.** Emily Pifer, Purple Martin Conservation Association. John R. Sauer and Jane Fallon. USGS Patuxent Wildlife Research Center, and John Tautin. Purple Martin Conservation As- sociation, “Hurricane Katrina: did it affect Purple Martins and Purple Martin roosts?*’ Matthew J. Reetz and Kathryn E. Sieving. University of Florida, and Scott K. Robinson. Florida Museum of Natural History. “Re- sponses of four songbird species to experi- mental cowbird parasitism in a recently in- vaded area.” Paul G. Rodewald, Ashley A. Buchanan, and Stephen N. Matthews. The Ohio State University. “Stopover behavior of migrant landbirds in two fragmented landscapes: lake- shore and inland regions of Ohio.*’ Scott A. Rush. L^niversity of Georgia. E. C. Soehren. Alabama Department of Conserva- tion and Natural Resources. A. T. Fisk. Uni- versity of Windsor. M. S. Woodrey. Mississip- pi State University, and R. J. Cooper. Univer- sity of Georgia. '‘Habitat use by marsh birds along the northern Gulf of Mexico with focus on Clapper Rails.*’ Jean-Pierre L. Savard, Environment Cana- da. and Bruno Drolet and Melanie L. Cousi- neau. Canadian Wildlife Service. “Evidence of a five-year population cycle in Rusty Blackbirds {Eiiphagiis caroUniis)." John A. Smallwood. Montclair State Uni- versity. Mark F. Causey. Damascus. MD. Da- vid Mossop, Yukon College. James R. Kluc- sarits, Alvemia College. Bob and Sue Rob- ertson. Kempton. PA. Richard J. Melvin. American Kestrel Foundation. Joey Mason. Middleboro. MA. Michael J. Maurer, Marion, MA. John W. Parrish Jr., and Timothy F. Breen. Georgia Southern University. Kenneth Boyd. Fort Gordon. GA. Russell D. Dawson. University of Northern British Columbia, and Gary R. Bortolotti. University of Saskatche- wan. “Why are American Kestrel (Falco span'ehiis) populations declining in North America? Evidence from nest box programs.*’ Amanda B. Stockwell and Ryanne Sullivan. Canisius College, Richard Keith and Brenda Keith. Kalamazoo Nature Center, and Sara R. Morris. Canisius College. ‘'Migration and stopover ecology of the Tennessee Warbler at an inland stopover site near Kalamazoo. Michigan.** Antoinette Taylor. Philip C. Stouffer. and Michael J. Chamberlain. Louisiana State Uni- versity. “Effects of site preparation on breed- ing birds in early successional Louisiana pine plantations.** Sarah E. Warner. W. Gregory Shriver. and Mamie A. Pepper. University of Delaware. ‘'Tidal marsh breeding birds as bioindicators of mercury contamination along the Delaware Bay.” Jill Wick and Yong Wang. Alabama A&M University. “Home range size and habitat use of two songbird species in forest stands treat- ed with prescribed fire and thinning.’* Michael Wierda. Clemson University, and ANNUAL REPORT 937 Jacqueline Bird, Alen Rebertus, and Alec Lin- say, Northern Michigan University, “Foraging ecology of Pileated Woodpeckers in Duke’s Experimental Forest in the Upper Peninsula of Michigan.” Amanda Jo Williams, Jennifer C. Owen, and Frank R. Moore, University of Southern Mississippi, and Mary Garvin, Oberlin Col- lege, “Changes in immunocompetence of the Gray Catbird during an experimental West Nile virus infection.” Stefan Woltmann and Thomas W. Sherry, Tulane University, “Territory fidelity and age structure in a tropical understory bird. Chest- nut-backed Antbird {Myrmeciza exsul)." POSTERS Lyndell M. Bade, University of Missouri- St. Fouis, Colleen Crank, Missouri Botanical Garden & Fitzsinger Road Ecology Center, Kathleen Beilsmith, Parkway North High School, St. Fouis, MO, and Patricia G. Parker, University of Missouri-St. Louis, “Reproduc- tive ecology of the Eurasian Tree Sparrow in two suburban environments in St. Louis.” Eric Beck, Michael Husak, and Michael A. Patten, Cameron University and University of Oklahoma, “Oklahoma marshbird monitoring project.” Bethany Belock and Doris Watt, Saint Mary’s College, “An examination of the role of weight, plumage color, cere color and total area of UV reflectant plumage on mate selec- tion in Budgerigars {Melopsitticiis undii- lates)'" Mary Bomberger Brown, University of Ne- braska, “The Tern and Plover Conservation Partnership: a model for interior Least Tern and Piping Plover conservation.” Ashley A. Buchanan and Paul G. Rode- wald. The Ohio State University, “Move- ments, habitat selection, and stopover duration of migrant songbirds in the western Lake Erie Basin of Ohio.” John P. Carpenter, Yong Wang, and Hugh Metcalfe, Alabama A&M University, “Ceru- lean Warbler home range estimates and roost site selection in northeast Alabama.” Jason Courter and Gary Ritchison, Eastern Kentucky University, “Responses of fiocks of Tufted Titmice to different-sized raptors.” Mark Deutschlander. Hobart and William Smith Colleges, and Rachel Muheim and John Phillips, Virginia Polytechnic Institute, “White-throated Sparrows use polarization cues on the horizon to calibrate their magnetic compass at sunrise and sunset.” Kathryn Dirks, Molly Grove, Angela Roles, and Mary Garvin, Oberlin College, “DNA se- quence-based identification of rootlets used in Gray Catbird (Diimetella carolinensis) nest linings.” Ryan M. Dziedzic and Michael J. Hamas, Central Michigan University, “Seasonal use of dynamic Lake Huron coastal habitats by migrating shorebirds.” Rodney K. Felix Jr., and Robb Diehl, Uni- versity of Southern Mississippi, and Janet M. Ruth, U.S. Geological Survey, “Passerine mi- gratory movements aloft in the southwestern United States.” Katherina Forgues, Trent University, Can- ada, “Making the connection between shore- birds and off-road vehicles.” liana Garcia-Grossman, Lydia Moore, Harden Wisebram, Alice Manos, and Mary Garvin, Oberlin College, “Feeding preference of Cule.x pipiens between two potential West Nile virus reservoir host species.” Andrew D. George, Timothy J. O’Connell, Karen R. Hickman, and David M. Leslie Jr., Oklahoma State University, “Avian response to Old World bluestem Bothriochloa ischae- mum monocultures in mixed grass prairie.” Lewis Grove, Emma DeLeon, Ben Coulter, Michael Lanzone, and Andrew L. Mack, Car- negie Museum of Natural History, “Automat- ic detection of recorded nocturnal flight calls: a comparison of methods.” Stephen B. Hager, Christopher R. Bertram, and Katie R. Demer. Augustana College, “Breeding birds and nest productivity at Green Wing Environmental Laboratory, northcentral Illinois.” Christy Hand and Patrick Jodice. Clemson University, and Felicia Sanders, Santee Coast- al Reserve, “Research on the half shell: diet composition of American Oystercatchers dur- ing the non-breeding season.” R. Ian Horn and Gary K. Ritchison. Eastern Kentucky University, “Eastern Phoebes use different strategies to provision young.” Jason D. Jacobs, Canisius College. Kristen M. Covino. University of Maine, and Sara R. Morris. Canisius College, “Seasonal differ- 938 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 ences in energetic condition of Blackpoll War- blers on Appledore Island.” Jodie M. Jawor and Jocole Green, Univer- sity of Southern Mississippi, “Steroid hor- mones and immune function in Northern Car- dinals {Cardinalis cardinalis).'" Erik I. Johnson and Philip C. Stouffer, Lou- isiana State University, “Ectoparasites affect bird condition in neotropical forest frag- ments.” Kelly Kussmaul and Doris Watt, Saint Mary’s College, “Heterophil to lymphocyte ratios as indicators of stress in Budgerigars {Melopsittacus undulates) under different housing arrangements.” Kristen M. Lear and Edward H. Burtt Jr., Ohio Wesleyan University, “Bacterial degra- dation of flight and body contour feathers by B. licheniformis.'' Katherine Leith, Michael Wierda, and Wil- liam Bowerman, Clemson University, James Sikarskie, Michigan State University, Dave Best, U.S. Fish and Wildlife Service, and Ter- yl Grubb, U.S. Forest Service, “Using factor analysis as a tool for the improvement of field sampling strategies.” William B. Lewis, E. Dale Kennedy, and Douglas W. White, Albion College, “Analysis of the vocalizations of the Northern Cardinal {Cardinalis cardinalis).'' Timothy Ludwick, Autumn Smith, and Alan Fedynich, Caesar Kleberg Wildlife Re- search Institute, “A comparative study of feeding habits and helminth diversity in south Texas doves.” Kimberly Martinczak and Doris Watt, Saint Mary’s College, “Relationships between dom- inance measures and physical characteristics of the Budgerigar {Melopsittacus undulates)." Kelly J. McKay, BioEco Research and Monitoring Center, and Stephen B. Hager, Au- gustana College, “The birds of Green Wing Environmental Laboratory in northcentral Il- linois.” Marks McWhorter, Mary Brown, and Jen- nifer C. Owen, University of Southern Mis- sissippi, “Spleen size and activity as it relates to migratory disposition in the Gray Catbird {Dumetella carolinensis)." Zoltan Nemeth, Michael J. Sellers, Jennifer C. Owen, and Frank R. Moore, University of Southern Mississippi, “Phenotypic organ flex- ibility around the annual cycle in two Nearc- tic-neotropical migratory thrush species.” Meredith Palmer and Edward H. Burtt Jr., Ohio Wesleyan University, “Resistance of tu- raco feathers to bacterial degradation.” Heidi L. Puckett, L. Wes Burger Jr., and Samuel K. Riffell, Mississippi State Univer- sity, “Mid-contract management effects on breeding grassland songbirds in CP33 habitat buffers in eastern Mississippi.” Orin Robinson, John Dindo, and Lauren Showalter, Dauphin Island Sea Laboratory, “Post Katrina monitoring of nesting birds in coastal Alabama.” Emily Runnells, Hobart and William Smith Colleges, David Bonter, Cornell Laboratory of Ornithology, and Mark Deutschlander, Hobart and William Smith Colleges, “Fat stores and energetic condition of Catharus thrushes dur- ing spring and autumn migration at a Great Lakes stopover site.” Chad Soard and Gary Ritchison, Eastern Kentucky University, “Carolina Chickadee calls encode information about predator threat.” S. J. Stefan, F. J. Sanders, B. C. Doyle, M. Hughes, and P. G. R. Jodice, College of Charleston, “Seasonal variations in recoveries of South Carolina-banded Brown Pelicans and Royal Terns: a comparison of two seabird spe- cies experiencing a decline in local nesting numbers.” Jack M. Stenger and Edward H. Burtt Jr., Ohio Wesleyan University, “The bacterial degradation of phaeomelanic and eumelanic feathers.” Bethany K. Stephan and Anna Marie Pari- se, Canisius College, Michael Hamilton and Robert L. DeLeon, Buffalo Ornithological So- ciety, and H. David Sheets and Sara R. Mor- ris, Canisius College, “Comparison of avifau- na in three riparian environments in western New York: the impact of anthropogenic hab- itat alteration.” Ryanne Sullivan and Amanda B. Stockwell, Canisius College, Richard Keith and Brenda Keith, Kalamazoo Nature Center, and Sara R. Morris, Canisius College, “Annual variation in the stopover ecology of the Tennessee War- bler in Michigan.” James W. Tucker Jr., Gregory R. Schrott, and Reed Bowman, Archbold Biological Sta- tion, “Habitat selection in the endangered ANNUAL REPORT 939 Florida Grasshopper Sparrow {Ammodf^amus savannarum flohdanus)A James W. Tucker Jr., Gregory R. Schrott, and Reed Bowman, Archbold Biological Sta- tion, “Occupancy modeling to examine detec- tion probabilities and population trends of the endangered Florida Grasshopper Sparrow.” John M. Waud, Rochester Institute of Tech- nology, Omar Gordillo, Comision Nacional de Areas Naturales Protegidas, Tuxtla Gutierrez, Mexico, David Mathiason, Rochester Institute of Technology, and Mark Deutschlander, Ho- bart and William Smith Colleges, “Avian spe- cies as indicators of riparian function in Chia- pas, Mexico.” Melinda J. Welton, Gulf Coast Bird Obser- vatory, David L. Anderson, Louisiana State University, Gabriel J. Colorado, Universidad Nacional de Colombia, Colombia, and Tiffany A. Beachy, University of Tennessee, “Migra- tion stopover of the Cerulean Warbler {Den- droica cerulea) in northern Middle America.” Rebecca Whelan, Tera Levin, and Mary Garvin, Oberlin College, “Detection of vola- tiles in the uropygial gland secretions of Gray Catbirds (Dumetella carolinensis) through solid-phase microextraction head space sam- pling and gas-chromatograph-mass spectrom- etry.” Michael Wierda, William Bowerman, Amy Roe, Kathryn Parmentier, William Bridges, and Katherine Leith, Clemson University, James Sikarskie, Michigan State University, David Best, U.S. Fish and Wildlife Service, Teryl Grubb, U.S. Forest Service, and Dennis Bush, Michigan DEQ, “Michigan Bald Eagle Biosentinel Program, monitoring trends of persistent organic pollutants in Great Lakes ecosystems.” Sean M. Williams and Edward H. Burtt Jr., Ohio Wesleyan University, and Ralph W. Schreiber and Elizabeth A. Schreiber, Smith- sonian Institution, “Structure of albino feath- ers: why so weak?” Meredith P. Wilson and Edward H. Burtt Jr., Ohio Wesleyan University, “Dynamics of Staphylococcus aureus on bird feathers.” ATTENDANCE Alabama: Auburn, Alan T. Hitch, Brian W. Rolek; Montgomery, Lauren M. Showalter, John A. Trent; Normal, John P. Carpenter, Lisa M. Gardner Barillas, Leela Pahl, Jill M. Wick, Yong Wang. Arkansas: Fayetteville, Elizabeth M. Adam, Douglas A. James. Arizona: Flagstaff, Eben H. Paxton; Tuc- son, Clait E. Braun. California: Areata, John E. Hunter; Berke- ley Jennifer M. Wang. Colorado: Denver, Dennis B. Williams. Delaware: Newark, Jeffrey Buler, Sarah E. Goodwin, Gregory W. Shriver, Sarah Warner. Florida: Ft. Myers, Bette J. S. Jackson, Je- rome A. Jackson; Gainesville, Karl E. Miller, Matthew Reetz; Lake Placid, Reed Bowman, Fred Lohrer, Gregory R. Schrott, James W. Tucker; Orlando, Reed F. Noss; Tallahassee, Jim Cox. Georgia: Athens, Scott A. Rush; Com- merce, Constance Head. Illinois: Rock Island, Stephen B. Hager. Indiana: Notre Dame, Beth Belock, Kelly Kussmaul, Kim Martinezak, Doris Watt; South Bend, Melinda M. Clark. Kansas: Sterling, Anthony C. Dalisio; Wichita, Alan D. Maccarone, Heather M. Stone. Kentucky: Richmond, Jason R. Courter, Richard I. Horn, Christine R. Lattin, Gary Ritchison. Louisiana: Baton Rouge, David L. Ander- son, Curtis W. Burney, Erik I. Johnson, Laura M. Palasz, Antoinette Taylor; Lafayette, Mag- gie Luent; New Orleans, Mollie Cashner, Katherine E. Law, Sedge Woltmann, Stefan Woltmann; Shreveport, James L. Ingold. Maryland: Chevy Chase, Ellen Paul; Tow- son, Scott L. Johnson. Michigan: Albion, Megan Fitzpatrick, E. Dale Kennedy, Will Lewis, Douglas White; Ann Arbor, Robert Payne, Laura Payne; East Lansing, Kimberly R. Hall; Mount Pleasant, Ryan M. Dziedzic, Michael J. Hamas. Minnesota: St. Paul, Bailey McKay. Missouri: St. Louis, Lyndell M. Bade. Mississippi: Bilo.xi, Ian Woodrey, Lorie Woodrey, Mark Woodrey; Hattiesburg. Mary (Francie) Brown, M. Susan Devries, Robb Diehl, Rodney K. Felix Jr., Rodney K. Felix Sr., Sheri Glowinski-Matamoros, Jodie Jawor, Marks McWhorter, Frank R. Moore, Zoltan Nemeth, Jen C. Owen, Michael Sellers, Jaclyn Smolinsky, Sarah Wheeless, Amanda J. Wil- 940 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 Hams; Mississippi State, Heath M. Hagy, Hei- di L. Puckett; Stoneville, Paul B. Hamel. Montana: Poison, Charles R. Blem. North Carolina: Hickory, John N. Brzo- rad. Nebraska: Lincoln, Mary Bomberger Brown. New Jersey: New Brunswick, Bertram G. Murray Jr.; Randolph, John A. Smallwood. New York: Bowmansville, Bethany Ste- phan; Bujfalo, Peggy E. Buckley, Katie A. Chmielowiec, Maura Hanna, Jason Jacobs, Amanda Stockwell, Ryanne Sullivan; Geneva, Mark E. Deutschlander, Emily S. Runnells; Grand Island, Sara R. Morris, Elizabeth Mor- ris; Rochester, John M. Waud. Ohio: Columbus, Sandra L. L. Gaunt, Paul G. Rodewald; Delaware, Edward H. Bum Jr., Kristen M. Lear, Meredith Palmer, Jack M. Stenger, Sean M. Williams, Meredith P. Wil- son; Oberlin, Kathryn Dirks, liana Garcia- Grossman, Mary C. Garvin, Tera Levin, Eliz- abeth Wisebram; Sandusky, Robert C. Beason; Wilmington, David N. Bonter, Bonnie Keller, Bob Powell. Oklahoma: Lawton, Eric J. Beck; Stillwa- ter, Vincent S. Cavalieri, Andrew D. George, Timothy J. O’Connell, Paul van Els; Tahle- quah, Mia R. Revels. Pennsylvania: Allentown, Daniel Klem Jr., Peter G. Saenger; Erie, Emily Pifer, Margaret A. Voss; Rector, Adrienne Jo Leppold, An- drew L. Mack; Villanova, Robert L. Curry, Andrew J. McGann. South Carolina: Charleston, Stacy J. Ste- fan; Clemson, Latice Puentes, Christy Hand, Katherine P. Leith, Lindsay J. Moore, Peggy Shrum, Michael R. Wierda; Columbia, Austin L. Hughes. Texas: Kingsville, Arlene J. Arnold, Tim- othy Ludwick; Lake Jackson, John C. Arvin, Cecilia M. Riley. Virginia: Cambridge Springs, Eugene S. Morton; Charlottesville, Charles E. Clarkson; Dillwyn, Mike Stinson; Hampden-Sydney, Mi- chael D. Collins; Shipman, Allen Hale; Wil- liamsburg, Jacob Armiger, Ryan Burdge, Kel- ly K. Hallinger, Mikaela G. Howie. Vermont: Dorset, Elizabeth P. Gilbert; Northfield, William H. Barnard. Washington: Bainbridge Island, Lee H. Robinson. Washington D.C.: Richard C. Banks, Carla Dove, Joseph R. Jehl, Storrs L. Olson. Canada, Ontario: Kingston, Robert C. Dobbs; Toronto, James D. Rising, Trudy L. Rising; Quebec: Sainte-Foy, Jean-Pierre L. Savard. The Wilson Journal of Ornithology 120(4):941— 942, 2008 REVIEWERS FOR VOLUME 120 Reviewers are the lifeblood of a journal as editors depend on them to help identify manuscripts with merit and offer suggestions to improve the data analysis, overall science, and writing. These individuals receive little recognition, but are extremely important in the process of improving the science and quality of what is published. We thank all of those listed below who served as referees for manuscripts processed (accepted and published, withdrawn, or rejected) after 1 July 2007 through completion of the December 2008 issue of Volume 120. Those shown in boldface reviewed more than one manuscript. The Wilson Ornithological Society and the editorial staff are indebted to and thank each person who served as a reviewer. — Clait E. Braun, Editor J. T. Ackerman, D. G. Ainley, A. R Aleixo, M. W. Allard, E A. Amidon, V. Amrhein, R. G. Anthony, R Arcese, W. Arendt, A. Arjun, R. A. Askins, A. S. Aspbury, S. Austin-By- thell, F. Bairlein, M. C. Baker, G. A. Balda- sarre, B. R. Barber, C. Barber, R E Battley, R. C. Beason, G. Beauchamp, M. J. Bechard, B. M. Beehler, C. D. Benesh, C. W. Benkman, R. E. Bennetts, R. O. Bierregaard, D. M. Bird, J. G. Blake, R Blancher, W. J. Bock, D. N. Bonter, S. Boyd, J. T. Boylan, J. D. Brawn, L. A. Brennan, M. Brigham, D. M. Brooks, C. R. Brown, D. E. Brown, M. B. Brown, R A. Buckley, T. M. Burg, A. Burger, L. W. Burger Jr., D. Burhans, R. Burnett, S. L. Burson, E. H. Burtt, R. W. Butler, R Byholm, E. G. Campbell, R. A. Canterbury, R. E. Carleton, J. Carpenter, J. H. Carter III, C. Cassaday, J. N. Caudell, D. Cerasale, J. Chaves, D. Cim- prich, R. Clapp, C. J. Clark, E. D. Clotfelter, R. R. Cohen, R. A. Cole, N. J. Collar, J. A. Collazo, C. T. Collins, M. W. Collopy, M. A. Colwell, R. N. Conner, C. J. Conway, S. J. Cooper, S. M. Correa, M. C. Coulter, D. A. Cristol, J. R Croxall, S. Dale, N. B. Davies, S. K. Davis, S. R. DeKort, R. S. DeLotelle, J. L. Deppe, K. C. Derrickson, A. A. Dhondt, J. G. Dickson, D. R. Diefenbach, R. Diehl, J. J. Dinsmore, R E Doherty Jr., D. D. Dolton, J. J. Dosch, C. J. Dove, V. J. Dreitz, S. Droege, H. Drummond, B. D. Dugger, K. Dugger, E. H. Dunn, R. O. Dunn, J. B. Dunning Jr., E. M. Dzialowski, J. M. Eadie, W. R. Eddleman, J. C. Eitniear, R. M. Erwin, R. Escalante, M. Evans, T. D. Evans, J. M. Eair, S. G. Fancy, G. L. Farnsworth, M. J. Fernandez, L. Fish- pool, J. A. Fitzsimons, J. Fjeldsa, C. H. Flath- er, T. F. Fondell, J. Foster, M. S. Foster, C. M. Francis, K. E. Franzreb, T. M. Freeberg, B. R. Freymann, S. W. Gabrey, I). Pk Ciammon, J. M. Garcia-C, S. L. L. Gaunt, M. Genovart, R. R. George, S. S. Germaine, J. Gilardi, T. W. Gillespie, M. D. Giovanni, C. B. Goguen, I). A. Granfors, T. A. Grant, M. Green, M. C. Green, H. F. Greeney, J. S. Greenlaw, M. Griesser, J. A. Grzybowski, C. G. Guglielmo, R. J. Gutierrez, M. T. Guzy, J. Haffer, T. M. Haggerty, M. L. Hall, S. Hallager, D. G. Harp- er, M. Hau, F. E. Hayes, M. A. Hayes, W. M. Healy, R. Hendricks, J. Herkert, G. Herring, S. K. Herzog, C. Hill, H. L. Hinam, R. T. Holmes, T. R. Huels, J. M. Hull, C. Hunter, R. L. Hutto, L. D. Igl, J. L. Ingold, K. Islam, E M. Jaksic, D. Jakubas, R. L. Jarvis, J. R. Jehl Jr., W. Jetz, R. H. Johnsgard, D. H. John- son, L. Johnson, L. S. Johnson, M. D. John- son, C. D. Jones, H. L. Jones, J. A. Jones, S. L. Jones, J. G. Jorgensen, J. O. Karubian, E. D. Kennedy, K. R. Kenow, R. E. Kenward, D. M. Keppie, E. L. Kershner, D. King, D. Klem Jr., S. Kleven, W. D. Koenig, O. Komar, D. N. Koons, N. K. Krabbe, A. W. Kratter, M. Krist, J. A. Kushlan, W. R. Kuvlesky, C. Kwit, S. K. Kyle, D. F. Lane, E. K. Latch, S. C. Latta, D. J. Lebbin, M. R. Lein, D. L. Leo- nard, G. Leonardi, J. Lepson, S. B. Lerman, D. J. Levey, C. A. Lindell, C. D. Littlefield, C. Loiseau, G. A. Londono, P. E. Lowther, G. A. Lozano, R. Luttrell, R. S. Lutz, T. J. Mabee, J. R. Madden, M. L. Mallory, R. W. Mannan, M. Marm, J. S. Marks, K. Martin, L. B. Martin, T. E. Martin, S. Matsuoka, R. Mauck, I. McAllan, A. E. McAndrews, C. McCaffrey, J. R. McCarty, K. McGraw, M. L. McKinney, M. K. McNicholl, S. R. Mc- Williams, E. Mellink, B. Mila, E. H. Miller, A. R. M0ller, D. Monk, E R. Moore, D. H. Morse, E. S. Morton, H. C. Mueller, D. Mul- cahy, R. Nagarajan, N. L. Newfield, V. Ny- man, S. H. Oppel, G. H. Orians, J. F. Ornelas, S. Oyler-McCance, K. C. Parsons, R. B. Rayne, wS. F. Rearson, B. S. Redersen, D. W. Rerkins, C. M. Rerrins, M. C. Rerry, vS. E. Re- ters, B. Rinshow, M. Riorkowski, T. D. Ritts, J. H. Rlissner, W. F. Rorter, W. Rost, R. A. Rowell, H. Pdysii, V. V. Rravosudov, D. R. C. 941 942 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 Prescott, R. Prinzinger, R. O. Prum, C. Ran- dier, J. L. Rasmussen, J. T Ratti, J. C. Re- boreda, L. R. Reitsma, K. Renton, M. Restani, M. H. Reynolds, R. T. Reynolds, J. M. Rhym- er, N. H. Rice, C. P. Riehl, W. J. Ripple, J. D. Rising, J. W. Rivers, C. S. Robbins, R. J. Robertson, J. A. Robinson, S. A. Rohwer, J. J. Roper, R. N. Rosenfield, S. S. Rosenstock, J. T. Rotenberry, R. R. Roth, A. Roulin, D. Renison, D. C. Rudolph, T. B. Ryder, J. Sal- gado Ortiz, D. W. Sample, C. Sanchez, B. K. Sandercock, J. R. Sauer, L. Savoy, R. R. Schaefer, J. A. Schmutz, J. A. Schnack, T. S. Schulenberg, P. L. Schwagmeyer, S. H. Schweitzer, J. M. Scott, S. G. Sealy, N. E. Seavy, J. A. Sedgwick, C. L. Seewagen, A. N. Setiawan, J. Shafer, S. Sharbaugh, D. Shu- tler, L. M. Siefferman, C. H. Sieg, L. F. Sil- veira, M. F. Small, D. W. Smith, R. J. Smith, J. A. Soha, R. Sojda, C. Somers, T. A. Sor- dahl, D. A. Spector, C. N. Spottiswoode, M. T. Stanback, R. D. Stark, D. F Stauffer, D. W. Steadman, R. B. Stiehl, D. F Stotz, P. C. Stouffer, B. J. M. Stutchbury, J. N. Styrsky, D. F. Swanson, P. B. Taylor, R. C. Telfair, R. Thorstrom, K. M. C. Tjorve, B. W. Tobalske, E. J. Tramer, J. Traylor, F Valera, S. Van Bal- en, A. N. Van Buren, M. F Vasconelos, C. Vaughan, J. H. Vega Rivera, N. A. M. Ver- beek, K. T Vierling, S. C. Votier, G. E. Wal- lace, J. R. Walters, Y. Wang, D. J. Watt, K. A. Weakland, R J. Weatherhead, J. M. Wells, S. West, C. M. White, M. Wikelski, J. W. Wiley, N. Wilkins, M. Williams, S. O. Williams, M. J. Willis, K. S. Winker, H. Winkler, K. Win- nett-Murray, M. Winter, M. C. Witmer, S. Wolf, M. S. Woodrey, C. Woods, F. G. Woo- laver, J. M. Wunderle Jr., R. H. Yahner, T. Yerkes, Y Yom-Tov, J. R. Young, T Yuri, R. Zambrano, K. Zyskowski. The Wilson Journal of Ornithology 1 20(4);943-962, 2008 Index to Volume 120, 2008 Compiled by Kathleen G. Beal This index includes references to genera, species, authors, and key words or terms. In addition to avian species, references are made to the scientific names of all vertebrates mentioned within the volume and other taxa mentioned prominently in the text. Nomenclature follows the AOU Check-list of North American Birds (Seventh Edition) and F. Gill and M. Wright (Birds of the World, Recommended English names. Princeton University Press, Princeton, New Jersey and Oxford, United Kingdom). Reference is made to books reviewed and an- nouncements as they appear in the volume. A abundance of Cypseloides niger in southern Rocky Mountains Colorado and New Mexico, 331-338 Accentor, Alpine, see Prunella collaris Accipiter gentilis, 239-247 gularis, 96 striatus, 464 Ackerman, Joshua T, John Y. Takekawa, Jill D. Bluso. Julie L. Yee, and Collin A. Eagles-Smith, Gen- der identification of Caspian Terns using exter- nal morphology and discriminant function anal- ysis, 378-383 Acrocephalus arundinaceus, 398 orientalis, 269, 272 Aegolius acadicus, 390 Aegotheles cristatus, 510 Afton, Alan D., see Jonsson, Jon Einar and Agelaius phoeniceus, 342, 470, 820-829 Aguirre, Ray, see Dreibelbis, Justin Z., Kyle B. Mel- ton, , Bret A. Collier, Jason Hardin, Nova J. Silvy, and Markus J. Peterson Aguirre, Ray, see Melton, Kyle B., Justin Z. Dreibel- bis, , Bret A. Collier, T. Wayne Schwert- ner, Markus J. Peterson, and Nova J. Silvy Aimophila aestivalis, 625-627 Akepa, see Loxops coccineus Aleixo, Alexandre, see Tobias, Joseph A., Daniel J. Lebbin, , Michael J. Andersen, Edson Guilherme, Peter A. Hosner, and Nathalie Sed- don A I soph is portoricensis, 464 Amakihi, Hawaii, see Hemignathus virens Amazon, Puerto Rican, see Amazona vittata Red-lored, see Amazona autumnalis White-fronted, see Amazona alhifrons Yellow-naped, see Amazona auropalliata Amazona alhifrons, 353-365 auropalliata, 353-365 autumnalis, 354, 355 fnschi, 176, 357, 361, 363 vittata, 357 Ammer, Frank K.. Petra Bohall Wood, and Roger J. McPherson. Gender identification of Cirasshop- per Sparrows comparing behavioral, morpho- logical, and molecular techniques, 221-225 Ammodramus hairdii, 667—673 henslow'ii, 111 — 119 maritimus, 394 savannarum, 111 — 119, 221—225 Amphispiza bilineata, 569 Anairetes alipinus, 535 parulus, 535, 539, 541 Anas cyanoptera, 390-392 platyrhynchos, 3 1 1 wyvil liana, 311-319 Anderson, Michael J., see Tobias, Joseph A., Daniel J. Lebbin, Alexandre Aleixo, , Edson Guilherme, Peter A. Hosner, and Nathalie Sed- don Andropadus latirostris, 399 Virens, 387, 398-401 Ani, Groove-billed, see Crotophaga sulcirostris Smooth-billed, see Crotophaga ani Anodorhynchus leari, 176 Anoiis spp., 378 Ant-Tanager, Red-crowned, see Habia rubica Antbird, Chestnut-backed, see Myrmeciza exsul Dusky, see Cercomacra tyrannina Spotted, see Hylophylax naevioides Warbling, see Hypocnemis cantator White-bellied, see Myrmeciza longipes Anthus cen'inus, 85 spragueii, 667—673 Antpitta, Rusty-breasted, see Grallaricula ferruginei- pectus Antshrike, Western Slaty, see Thanmophilus atrinucha Antunes, Andre Pinassi, see Rohe, Fabio and Aploderma aequatoriale, 249 narina, 249 Aphelocoma ultramarina, 342 Ara macao, 353-365 militaris, 176 ruhrogenys, 176 Aratinga canicularis, 353—365 strenua, 357. 361 Arcese, Peter, .see Janssen, Michael H.. , Mark S. Sloan, and Kelly J. Jewell Arcos-Torres. Agustina, see Solano-Ugalde. Alejandro, and Ardea cinerea, 631—632 herodias, 1 73 Areta, Juan I., see Nikli.son. Alina M.. . Roman A. Ruggera, Karie E. Decker. Carlos Bosque, and Thomas E. Martin Asia of us, 641—645 943 944 THE W ILSON JOURNAL OF ORNITHOLOGY • VoL 120. \o. 4. December 2008 Athene cunicularia. T2 .Atlapetes semirnfus. 856-S62 Auklei- Cassin's- see Prychoramphus aleuricits Autoniolus infnscatits. 10. 11. 18. 20. 21 leiicophrhalamiis. 10—25. 45 melanopezus. 1 1 ochrolaemus. 1 1 paraensis. 10. 11. 20 ntfipileatus. 21 rubiginosiis. 1 1 ntfipileatus. 1 1 B Baeolophus bicolor. 183 Balasubramaniam. Shandiya. and Patrick-Jean Guay. Purple Swamphens xPorphyrio porphyrio) ai- lempiing to pre\ upon Black Swan (Cygnus atratus} eggs and preying upon a c\ gnet on an urban lake in Melbourne. Australia. 633-635 Baldwin. Michael J.. W ylie C. Barrow Jr. Clinton Jes- ke. and Frank C. Rohwer. Metabolizable ener- g> in Chinese taUow fruit for Yellow -rumped Warblers. Northern Cardinals, and .American Robins. 525-530 Ballard. Grant, see Nur. Nadav. . and Geoffrey R. Geupel Banko. Paul C.. see Farmer. Chris. Bridget .A. Fred- erick. . Roben M. Stephens, and Caner W. Snow Barbtail. Sponed. see Premnoplex bntnnescens Barrantes. Gilben. Cesar Sanchez. Branko Hilje. and Rodolfo Jaffe. Male song \ ariation of Green \ ioletear xColibri thalassinus) in the Talaman- ca Mountain Range. Costa Rica. 519-524 Barreto. Guillermo R.. see Bensch. Carolina and Barrow Jr. Wylie C.. see Baldwin. Michael J.. . Clinton Jeske. and Frank C. Rohwer Banramia longicauda. 129 bea\ er. .American, see Castor canadensis Beckmann. Christa. .An intraspecitic killing in adult Pacific Reef Egrets xEgrerta sacra t. 422 — 124 behavior age influence on territory size, habitat selection, re- producti\ e success in male Wilsonia cana- densis. 446 — 154 breeding, adoption then reproduction w ith father b\ female Sialia sialis. 419 — 122 courtship behavior of And ropadus virens. 398 — 101 display of male Phaethomis niber. 201-204 ecologN of tropical birds. 26-3“ feeding double-scratching b\ Xanthocephalus xantho- cephalus. 65 “-659 factors influencing fidelin of Carpodacus mexi- caniis. 371-3^^ male Hemignathits virens fed Loxioides bailleui fledgling. 416 — 118 of Tyrannus dominicensis. 655— 65“ fledging of Py rodents scutatits. 413 — 116 flocking of Spizella wonheni, 569-574 foraging comparison of juvenile and adult Turdits migratorius. 209-2 1 3 habitat use by Mergits serrator. 743-754 nest defense b\ Diryothorus litdovicianits. 46“ — 1“2 nocmmal of Momonts momota. 653-654 of Chen caentlescens and Chen rossii. 725-731 of Cnipodectes sitpemtfits. 38 — 19 of Dryolirnnas {citvieri\ aldabramts. 50-61 of Oreoniystis bairdi. 195-199 of Phalaenoptilus nuttallii. 505-5 1 2 of Scytalopits simonsi. 4 “3 — 1““ pol\ andry in Melospiza melodia. 345 polygyn> in Onts flammeohts. 645—648 post-fledging movement of Cathanis iistidanis. 62-73 raising of Molothnts ater by Aimophila aestivalis. ^ 625-627 range extension of Melospiza georgiana nigrescens. 393-395 reproducti\ e success of Dendroica centlea in Indi- ana. 105-110 roosting during winter of Seiunts aitrocapilla in core foraging area. 455 — 150 simultaneous incubation by two females and nest- ling provisioning b\ four adults at a Passer- citlits sandivichensis nest, 628—630 Bellbird. Three- w attled, see Procnias tricanmculatits Benham. Phred M.. see Reitsma. Leonard R.. Michael T. Hallw orth. and Benitez-Diaz. Hesiquio. see Honey-Escandon. Magali. Blanca E. Hernandez- Banos. Adolfo G. Nava- rro-Siguenza. . and .A. Tow nsend Peter- son Berr>. Laiifie and Estanislao Taisacan. Nest success and nest predation of the endangered Rota White-e\ e xZosterops rotensisx. 618-619 Bensch. Carolina and Guillermo R. Barreta. Diet of the Yellow -knobbed Curassow in the central \ enezuelan Llanos. 767-“““ Bezx. M. Bemadene. see Mamzak. Greg D.. . and Donald J. Brightsmith Biamonte. Esteban, see Sandoval. Luis. Esteban Bia- monte. and .Alejandro Solana-Ugalde Biancucci. Luis and Thomas E. Manin. First descrif>- tion of the breeding biologx and namral histors of the Ochre-breasted Brush Finch xAtlapetes semintfits t in \ enezuela. 856-862 Bingham. Ralph L.. see Breeden. Jeffrey B.. Fidel Her- nandez. . Nova J. Sil\y. and Gar\ L. Waggerman Bildstein. Keith L.. review. 430 — 131 Birgits latro. 59 Binem. .American, see Botaunts lentiginosits Least, see Ixobrychus exilis Blackbird. Common, see Turdits mentia Red- w inged, see Agelaiits phoeniceits A'ellow -headed, see Xanthocephalus xanthocephalus Blackcap. Eurasian, see Sylvia atricapilla Bluebird. Eastern, see Sialia sialis BluethroaL see Litscinia svecica INDEX TO VOLUME 120, 2008 945 Bluso, Jill D., see Ackerman, Joshua T, John Y. Take- kawa, , Julie L. Yee, and Collin A. Ea- gles-Smith Bobolink, see Dolichonyx oryzivonis Bombycilla cedrorum, 149, 151, 342 Bonasa umbellus, 239-247 (Erontispiece), 897-900 Bond, Jeanine and Daniel Esler, Bill entanglement in subcutaneously-anchored radio transmitters on Harlequin Ducks, 599-602 Bondo, Kristin J., Lauren N. Gilson, and Reed Bow- man, Anvil use by the Red-cockaded Wood- pecker, 217-221 Booby, Blue-footed, see Sida nebouxii Bosque, Carlos, see Niklison, Alina M., Juan I. Areta, Roman A. Ruggera, Karie L. Decker, , and Thomas E. Martin Botaurus lentiginosiis, 5 1 3-5 1 8 Bowman, Reed, see Bondo, Kristin J., Lauren N. Gil- son, and Brant, Black, see Branta bernicla nigricans Branta bernicla nigricans, 755-766 Breeden, Jeffrey B., Eidel Hernandez, Ralph L. Bing- ham, Nova J. Silvy, and Gary L. Waggerman, Effects of traffic noise on auditory surveys of urban White-winged Doves, 384-389 breeding biology adoption then reproduction with father by female Sialia sialis, 419-422 differences in growth of Branta bernicila nigricans between a major breeding colony and out- lying breeding aggregations, 755-766 Grallaricida ferrugineipectus, 345-352 habitat and other features of Ficedula narcissina, 92-98 long-term trends in breeding in old-growth Adiron- dack forest and surrounding regions, 153- 158 nests, eggs, and incubation of Pyrrhida erythaca, 874-878 of Atlapetes semirufiis in Venezuela, 856-862 of Pyroderus scutatiis granadensis, 862-867 parental care in Sporophila hypoxantha and Sporo- phila collaris, 879-883 reproductive success of Troglodytes aedon in sub- urban and rural landscapes, 99-104 reproductive success of Vireo latimeri in a montane habitat, 460-466 use of legs and feet for control by scoters during aerial courtship, 594-599 Brennan, Leonard A., Texas quails; ecology and man- agement, reviewed, 236-237 Brigham, R. Mark, see Woods, Christopher R and Brightsmith, Donald J., see Matuzak, Greg D., M. Ber- nadette Bezy, and Brotogeris jiigularis, 353-365 Brown, Charles R., see O’Brien, Valerie A., Amy T Moore, Kathryn P. Huyvaert, and Brown. David R. and Thomas W. Sherry. Solitary win- ter roosting of Ovenbirds in core foraging area. 455-459 Bubo virginianus, 173, 239—247, 505, 507 Bucephala clangula, 320-330, 732-742 islandica, 320—330 Buehler, David A., see Giocomo, James L., E. Daniel Moss, , and William G. Minser Bullfinch. Grey-headed, see Pyrrhida erythaca, bullfrog, see Rana catsbeiana Bunting, Godlewski’s, see Emberiza godlewskii Indigo, see Passerina cyanea Lark, see Calaniospiza melanocorys Meadow, see Emberiza cioides Painted, see Passerina ciris Buren, A. N. Van and P. Dee Boersma, Humboldt Pen- guins (Sphenisciis hwnboldti) in the northern hemisphere. Burger Jr., L. Wesley, see Wood, Douglas R., Erancis- co J. Vilella, and Bustos, Oscar Francisco Reyna, see Rosas-Espinoza, Veronica Carolina, Elisa Maya-Elizarraras, , and Francisco Martin Huerta-Martinez Butcher, Jerrod A. see Campomizzi, Andrew J., Shan- non L. Ferrell, and , Nest site selection by a male Black-capped Vireo Buteo albicaudatiis, 229 jamaicensis, 229, 422, 708-716 regalis, 708-716 swainsoni, 708-716 Byrd, G. Vernon, see Gibson, Daniel D. and c Cacomantis sepulcralis, 887-890 Cahill, J. R. A., E. Matthysen, and N. E. Huanca, Nest- ing biology of the Giant Conebill {Oreomanes fraseri) in the High Andes of Bolivia, 545-549 Calaniospiza melanocorys, 62 Calcariiis ornatiis, 667-673 pictiis, 398 Calypte anna, 522, 523 Calyptophiliis friigivoriis, 1 90 te fills, 190—195 Campomizzi, Andrew J., Shannon L. Ferrell, and Jer- rod A. Butcher, Nest site selection by a male Black-capped Vireo, 407-409 Campephihis principalis, 500 Canales-Delgadillo. Julio C., Laura M. Scott-Morales, Mauricio Cotera Correa, and Marisela Pando Moreno, Observations of flocking behavior of Worthen’s Sparrows (Spizella wortheni) and occurrence in mixed-species flocks, 569-574 Caprimulgus carolinensis, 509 climacurus, 510 tristigma, 5 1 0 vociferus, 509, 778-783 Cardiff. Scott G. and Steven M. Goodman. Natural history of the Red Owl ( I'yto soumagnei) in dry deciduous tropical forest in Madagascar. 891- 897 Cardinal. Northern, see Cardinalis cardinalis Cardinalis cardinalis, 99. 103. 149. 150 Cardiielis crassiroslris, 534 t list is, 149. 150. 151 946 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 Carleton, Renee E., Ectoparasites affect hemoglobin and percentages of immature erythrocytes but not hematocrit in nestling Eastern Bluebirds, 565-568 Carpodacus mexicanus, 342, 371—377 Carrasco, Luis, see Karubian, Jordan and Castor canadensis, 320 cat domestic, see Felis catus [domesticus] feral, see Felis cams [domesticus] Catbird, Gray, see Dumetella carolinensis Cathariis bicknelli, 458 fuscescens, 277—285, 528 guttatus, 458 ustidatus, 62-73, 74-84, 157, 277-285, 296-303, 467, 468 Cauchard, Laure, see Overington, Sarah E., , and Kinberly-Ann Cote Centropus toidou, 59 Cercomacra tyrannina, 31, 33 Certhia americana, 155 Cervus elaphus, 830 Cettia annae, 268, 273 borealis, 268 carol inae, 273 cetti, 268 diphone, 268-276 fortipes, 268-276 haddeni, 268-278 ruficapilla, 268, 273 seebohrni, 268—276 Chaetiira pelagica, 154 Chandler, C. Ray, see Wetzel, Daniel R and Charadrius melodus, 404—407 montanus, 100—107 Chat-tanager, Eastern, see Calyptophiliis frugivorus Western, see Calyptophilus terius Chaumis ictericus, 228—230 Chavez-Ramirex, Felipe, see Kim, Daniel H., Wesley E. Newton, Gary R. Lingle, and Chen caerulescens, 422, 725—731 rossii, 725-731 Chenxi, Jia and Sun Yuehua, Nests, eggs, and incu- bation behavior of Grey-headed Bullfinch (Pyr- rhida erythaca), 874—878 Chickadee, Black-capped, see Poecile atricapillus Boreal, see Poecile hudsonica Gray-headed, see Poecile cincta chicken, see Callus gallus Childers, Theresa M. and Stephen J. Dinsmore, Den- sity and abundance of Mountain Plovers in northeastern Montana, 700—707 Chlidonias niger, 167—175, 381 spp., 376 Chondestes grammacus, 342 Christie, David, see del Hoyo, Jesop, Andrew Elliott, and Chuck-will’s-widow, see Caprimulgus carolinensis Cinclodes, Bar-winged, see Cinclodes fiiscus Cinclodes fuscus, 475 Circus cyaneus, 173 Cnipodectes subbrunneus, 38-49 superriifus, 38-49 Coccothraustes vespertinus, 342 Coccyzus americanus, 287 vieilloti, 464 Cohen, Jonathan B., Elizabeth H. Wunker, and James D. Fraser, Substrate and vegetation selection by nesting Piping Plovers, 404-407 Colaptes auratus, 495 Colibri coruscans, 522, 523 delphinae, 523 serrirostris, 523 thalassinus, 5 1 9-524 Collier, Bret A., see Dreibelbis, Justin Z., Kyle B. Mel- ton, Ray Aguirre, , Jason Hardin, Nova J. Silvy, and Markus J. Peterson Collier, Bret A., see Melton, Kyle B., Justin Z. Drei- belbis, Ray Aguirre, , T. Wayne Schwertner, Markus J. Peterson, and Nova J. Silvy Collister, Douglas M., see Wilson, Scott, Keith A. Hobson, , and Amy G. Wilson Compospiza garleppi, 476 Conebill, Chestnut-vented, see Conirostrum speciosurn Cinereous, see Conirostrum cinereurn Giant, see Oreomanes fraseri Tamurugo, see Conirostrum tamarugense White-browed, see Conirostrum ferrugineiventre Conirostrum cinereurn, 531—544 ferrugineiventre, 531-544 speciosurn, 548 tamarugense, 548 Contopus sordidulus, 342 Virens, 146-152, 155. 291, 292 Contreras, Thomas A., see Reetz, Matthew J., Eliza- beth Farley, and Conway, Courtney J., see Nadeau, Christopher R, , Bradley S. Smith, and Thomas E. Lewis Coot, American, see Fulica americana Coracias spatulatus, 479 Correa, Mauricio Cotera, see Canales-Delgadillo, Julio C., Laura M. Scott-Morales, , and Mar- isela Pando Moreno Corx’us albus, 59 brachyrhynchos, 109, 149, 151 caurinus, 422 corax, 72, 173 Cotinga cayana, 871-874 Cotinga, Spangled, see Cotinga cayana Cote, Kimberly-Ann, see Overington, Sarah E., Jaure Cauchard, and Coturnicops noveboracensis, 513, 606—610 Coucal, Malagasy, see Centropus toulou Cowbird, Bronzed, see Molothrus aeneus Brown-headed, see Molothrus ater Shiny, see Molothrus bonariensis crab, robber, see Birgus latro Craik, Shawn R. and Rodger D. Titman, Movements and habitat use by Red-breasted Merganser broods in eastern New Brunswick, 743-754 INDEX TO VOLUME 120, 2008 947 Crandall, Stephanie M., see Hager, Stephen B., Heidi Trudell, Kelly J. McKay, , and Lance Mayer Crane, Red-crowned, see Grus japanensis Cranioleuca albicapilla, 535, 538, 541 Crax daubentoni Creeper, Brown, see Certhia americana Kauai, see Oreomystis bairdi Crotophaga ani, 422 sulcirostris, 214-216 Crow, American, see Corxnis brachyrhynchos Northwestern, see Corx'us caurimis Pied, see Corvus albus Crypturellus obsoletus, 228-230 Cuckoo, Banded Ground, see Neomorphus radiolosus Puerto Rican Lizard, see Coccyzus vieilloti Rusty-breasted, see Cacomantis sepidcralis Yellow-billed, see Coccyzus americmms Curassow, Yellow-knobbed, see Crax daubentoni Cyanocitta cristata, 109, 114, 149, 467-472 Cygnus atratus, 633-635 Cypseloides niger, 331-338 D Dacnis, Tit-like, see Xenodacnis parina Davis, Andrew K., Factors influencing fidelity of House Finches to a feeding station, 371-377 Davis Jr., William E., reviews, 231-232, 232-233 Debinski, Diane M., see Olechnowski, Brian E M. and Decker, Karie L., see Niklison, Alina M., Juan I. Areta, Roman A. Ruggera, , Carlos Bosque, and Thomas E. Martin del Hoyo, Jesop, Andrew Elliott, and David Christie, Handbook of the birds of the world. Volume 1 1 : old world fly-catchers to old world war- blers, reviewed, 235-236 Dendrocopus leucotos, 219 [Picoides] major, 96, 215, 217, 219 medius, 219 minor, 219 syriacus, 219 Dendroica caerulescens, 157, 296-303, 446, 450, 451 cerulea, 105-110 coronata, 155, 290, 296-303, 525-530 discolor, 455, 458 dominica, 401-403 magnolia, 157, 277-285, 296-303 petechia, 289, 291, 292, 451, 470, 830-839 pensylvanica, 155 pinus, 217 tigrina, 455, 458 Virens, 157,409-412 de Queiroz, Alan, Double-scratching by Yellow-head- ed Blackbirds, 657-659 Desrochers, Andre, see Hadley, Adam and Dessecker, Daniel R., .see Zimmerman, Guthrie S., Rick R. Horton, , and R. J. Gutierrez Dickcissel, see Spiza americana Dicrurus hottentottus, 97 macrocercus, 97 Diem, Kenneth L., see Pugesek, Bruce H. and diet anvil use by Picoides borealis, 217-221 foraging ecology of High Andean insectivorous birds in remnant Polylepis forest patches, 531-544 freeze-frame fruit selection by birds, 901-905 metabolizable energy in Chinese tallow fruit for Cardinalis cardinalis, Turdus migratorius, Dendroica coronata, 525-530 of Asia otus in northern and central Negev Desert, Israel, 641-645 of Crax daubentoni, 767-777 of Melanerpes formicivorus at forest in Mexico, 494-498 of nesting Charadrius melodus, 404-407 of nestling Black-crowned Night-herons, 637-640 of Patagona gigas ingesting calcium-rich minerals, 651-653 of parrots in a modified landscape, 353-356 of six neotropical bird species, 214-216 use of clay licks by Maroon-fronted Parrots, 176- 180 Di Giocomo, Alejandro G., see Facchinetti, Carolina, , and Juan C. Reboreda Dinsmore, Stephen J., see Childers, Theresa M. and Dionne, Mark, Celine Maurice, Jean Gauthier, and Francois Shaffer, Impact of hurricane Wilma on migrating birds: the case of the Chimney Swift, 784-792 Dipsochelys dussumieri, 59 disease avian cell-mediated immune response to drought, 813-819 no evidence for spring re-introduction of an arbo- virus by Petrochelidon pyrrhonota, 910—913 trichomoniasis in Band-tailed Pigeons, 603-606 distribution effects of landscape configuration for Caprimulgus vociferous, 778-783 of Cnipodectes superrufus, 38—49 of Cypseloides niger in southern Rocky Mountains Colorado and New Mexico, 331-338 of Patagona gigas in Argentina. 648-65 1 Diuca speculifera, 613-617 Doerr, Joseph G., see Levad, Richard G., Kim M. Pot- ter, Christopher W. Shultz, Carolyn Gunn, and Dolichonyx oryzivoriis, 820-829 D’Orazio, Kelly A. and Diane L. H. Neudorf. Nest defense by Carolina Wrens. 467-472 Dove. Carla and Court Goodroe. Marbled Godwit col- lides with aircraft at 3.700 m. 914-915 Dove. Mourning, see Zenaida macroura Dreibelbis. Justin Z.. see Mellon. Kyle B.. . Ray Aguirre. Bret A. Collier. T. Wayne Schwcrtner. Markus J. Peterson, and Nova J. Silvy Dreibelbis. Justin Z.. Kyle B. Melton. Ray Aguirre. 948 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 Bret A. Collier. Jason Hardin, Nova J. Silvy, and Markus J. Peterson, Predation of Rio Grande Wild Turkey nests on the Edwards Pla- teau, Texas. 906-910 Droege, Sam. see McNulty, Stacy A.. . and Raymond D. Masters Drolet, Bruno, see Robert, Michel, . and Jean- Pierre L. Savard Drongo, Black, see Dicninis macrocercus Hair-crested, see Dicrunis hottentottus Di-yocopus pileatus, 148. 149, 499—504 Dryolimnas [cuvieri} aldabramis, 50—61 cuvieri, 50, 631—632 Duck. Harlequin, see Histrionicus histrionicus Dugger. Bruce D.. see Uyehara. Kimberly J., Andrew Engilis Jr., and Dwnetella carolinensis, 146—152, 146-152. 277—285. 287. 291, 292. 297 Dunnock. see Prunella modiilaris E Eagle, Bald, see Haliaeetus leucocephaliis Eagles-Smith, Collin A., see Ackerman, Joshua T, John Y. Takekawa. Jill D. Bluso. Julie L. Yee, and Eason. Perri K.. see Vanderhoff. E. Natasha and ecology behavior of tropical birds, 26-37 cavity number and use by other species as correlates of group size in Picoides borealis, 181 — 189 Sitta spp.. 692-699 winter for Cotunncops noveboracensis, 606—610 eggs abnormal eggs of Rio Grande Wild Turkey. Melea- gris gallopavo, 226-228 clutch variations of egg size in Rhea americana, 674-682 description for Calyptophilus tertiiis and Xanoligea montana, 190-195 of Scytalopus simonsi, 473 — 177 Egret, Pacific Reef, see Egretta sacra Egretta sacra, 422 — 124 Elaenia. Lesser, see Elaenia chiriquensis Yellow-bellied, see Elaenia flavogaster Elaenia chiriquensis, 31, 32 flavogaster, 31, 32 Elaphe alleghaniensis, 420 anomala, 96 obsoleta, 109 spp.. 182 Elliot, Andrew’, see del Hoyo. Jesop, , and Da- vid Christie Ellison, Kevin S., Nest reuse by Vermilion Elycatchers in Texas, 339-344 Emberiza cioides, 97 godlewskii, 97 Empidonax alnorum, 256. 259, 278 flavescens, 278 fulvifrons, 256-267 minimus, 278. 342 occidentalis, 332 oberholseri, 259 traillii, 259, 278, 342, 830-839 traillii extimus, 256 virescens, 256. 278. 342 Engilis Jr., Andrew, see Uyehara. Kimberly J., . and Bruce D. Dugger Epicrates inornatus, 464 Erithacus rubecula, 209 Esler, Daniel, see Bond, Jeanine and , Bill en- tanglement in subcutaneously-anchored radio transmitters on Harlequin Ducks, 599-602 Estades, Cristian, M. Angelica Vukasovic, and Jorge A. Tomasevic, Giant Hummingbirds (Patagona gigas) ingest calcium-rich minerals. 651-653 Eudyptes spp., 575-581 Eugenes fidgens, 522 Euphonia gouldi, 200-201 Euphonia. Olive-backed, see Euphonia gouldi Euptilotis neoxenus, 249 Evers. Erin M., see Linkhart, Brian D., , Julie D. Megler, Eric C. Palm. Catherine M. Sali- pante, and Scott W. Yanco F Eaaborg, John, see White. Jennifer D. and Eacchinetti. Carolina. Alejandro G. Di Giacomo, and Juan C. Reboreda. Parental care in Tawny-bel- lied {Sporophila hypoxantha) and Rusty-col- lared (5. collaris) seedeaters, 879-883 Pair. J. M. and S. J. Whitaker, Avian cell-mediated immune response to drought. 813-819 Pairy-Wren, Red-backed, see Malurus melanocephalus Ealco newToni, 59 Earley. Elizabeth, see Reetz, Matthew J., , and Thomas A. Contreras Parmer, Chris, Bridget A. Prederick. Paul C. Banko, Robert M. Stephens, and Carter W. Snow’, Pal- lia (Loxioides bailleui) fledgling fed by Hawaii Amakihi (Hemignathus virens), 416-418 Eelis cams, {domesticus\ 60, 190, 194 Pelton, Adam. Annika M. Pelton, and David B. Lin- denmayer. The display of a Reddish Hermit (Phaethornis ruber) in a lowland rainforest. Bolivia. 201-204 Pelton. Annika M., see Pelton. Adam. , and Da- vid B. Lindenmayer Pemandez. Eladio M.. see Rimmer, Christopher C., Lance G. Woolaver, Rina K. Nichols, , Steven C. Latta. and Esteban Garrido Pernandez, Gustavo J. and Juan C. Reboreda. Between and within clutch variation of egg size in Greater Rheas, 674—682 Perrell, Shannon L., see Campomizzi. Andrew J., . and Jerrod A. Butcher. Nest site selec- tion by a male Black-capped Vireo Eicedula albicollis, 96. 97 hypoleuca, 96, 97. 446 pan a, 96, 97 zanthopygia, 92—98 INDEX TO VOLUME 120. 2008 949 Einch. Cochabamba Mountain, see Compospiza gar- leppi House, see Carpodaciis mexicamis Ochre-breasted Brush, see Atlapetes semirufiis White-winged Diuca. see Diiica speculifera Elamingo, Greater, see Phoenicoptenis roseus Elicker. Northern, see Colaptes auratiis Elycatcher. Acadian, see Empidonax virescens Alder. See Empidonax alnorwn Ash-throated, see Myiarchus cinerascens Collared, see Eicedida albicoUis Cordilleran. see Empidonax occidentalis Dusky, see Empidonax oberholseri European Pied, see Eicedida hypoleiica Great Crested, see Myiarchus crinitus Narcissus, see Eicedida narcissina Red-breasted, see Eicedida parx a Southwestern Willow, see Empidonax traillii exti- mus Vermilion, see Pyrocephaliis rubiniis Willow, see Empidonax traillii Yellow-rumped. see Eicedida zanthopygia Eoliage-gleaner. Buff-throated, see Aiitomohis ochro- laemiis Chestnut-crowned, see Aiitomolus rufipileatiis Olive-backed, see Automoliis infuscaiitus Para, see Automoliis paraensis Ruddy, see Automoliis rubiginosus White-eyed, see Automoliis leiicophthalamiis Eorest-Ealcon. Barred, see Micrastiir ruficollis Collared, see Micrastiir semitorqiiatiis Poster, Mercedes S.. Preeze-frame fruit selection by birds. 901-905 fox. red, see Vulpes viilpes Prancisco. Mercival R.. Paulo R. R. Oliveira Jr., and Vitor O. Lunardi. Nest and fledglings of the Red-ruffed Pruitcrow {Pyroderiis sciitatiis), 413-416 Eraser, James D.. see Cohen, Jonathan B., Elizabeth H. Wunker, and Eraser, M. J.. see Robertson. C. J. R.. P. Hyvonen. . and C. R. Pickard Erederick. Bridget A., see Parmer. Chris. . Paul C. Banko. Robert M. Stephens, and Carter W. Snow Freymann. Bernd P and Karl-Ludwig Schuchmann. Postnatal growth rates of hummingbirds; re- view and new records. 884-887 Fruitcrow. Red-ruffed, see Pyroderiis sciitatiis Fry, W. Roger, see Oehler. David A.. Steve Pelikan. . Leonard Weakley Jr.. Alejandro Kusch, and Manuel Marin Eiilica americana, 5 1 3—5 1 8 Fulmar. Northern, see Eidmariis glacialis Eulmarus glacialis, 635-636 G Gallinida cldoropiis, 5 1 3-5 1 8 Gallinule, Purple, see Porphyria martiniciis Callus galliis, 399. 474. 479 Garrido. Esteban, see Rimmer. Christopher C.. Lance G. Woolaver. Rina K. Nichols. Eladio M. Fer- nandez. Steven C. Latta. and Garrido, Esteban, see Townsend. Jason M.. . and Danilo A. Mejia Garridiis glandariiis, 72. 96 Gauthier. Gilles. see Senechal, Helene, , and Jean-Pierre L. Savard gecko, house, see Hemidactyliis frenatus Genetics of breeding and wintering population of Limnoth- lypis swainsonii, 433-445 of martin species. 683-691 Geococcyx spp., 206 George. T. Luke, see Sperry. Jinelle H., . and Steve Zack Geospiza scandens, 31 Geothlypis trichas, 80. 146-152, 277-285. 286-295. 288. 291. 292. 296-303. 830-839 Geupel. Geoffrey R.. see Nur, Nadav. Grant Ballard, and Gibson, Daniel D. and G. Vernon Byrd. Birds of the Aleutian Islands, Alaska, reviewed. 425-426 Gilson. Lauren N., see Bondo. Kristin J.. . and Reed Bowman Giocomo. James L.. E. Daniel Moss. David A. Bueh- ler. and William G. Minser. Nesting biology of grassland birds at Fort Campbell. Kentucky and Tennessee. 111-119 Glaiicidiiim bolivianum. 548 Glaiicomys volans, 181-189 Gnatcatcher. Black-tailed, see Polioptila melanura Blue-gray, see Polioptila caeridea Godwit. Marbled, see Limosa fedoa Goldeneye. Barrow 's, see Encephala islandica Common, see Encephala clangida Goldfinch. American, see Cardiielis tristis Goodman. Steven M.. see Cardiff. Scott G. and Goodroe, Court, see Dove. Carla and Goose. Lesser Snow, see Chen caeridescens Ross's, see Chen rossi Goshaw k. Northern, see Accipiter gentilis Grackle, Boat-tailed, see Qiiiscaliis major Common, see Qiiiscaliis quiscula Great-tailed, see Qiiiscaliis mexicamis Grallaricida ferriigineipectiis, 345-352 flavirostris, 348. 349 nana, 348. 349 peruviana, 348. 349 Grassquit. Blue-black, see Volatinia jacarina Graves, Gary R., review. 233-235 Graves, Gary R.. see Winker, Kevin and Grebe. Pied-billed, see Podilymhiis podiceps Greenbul. Little, see Andropadiis virens Yellow-whiskered, see Andropadiis latirostris Grim. Tomas. Begging behavior of fledgling Rusty- breasted Cuckoo (Cacomantis sepulcralis). 887-890 Grosbeak. Evening, see Coccothraustes vespertiniis Rose-breasted, see Pheiicticus ludovicianiis Grouse. Ruffed, see Eonasa iimbelliis 950 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 Grus japonensis, 610—613 Guay, Patrick-lean, see Balasubramaniam, Shandiya and Guilherme, Edson, see Tobias, Joseph A., Daniel J. Lebbin, Alexandre Aleixo, Michael J. Ander- sen, , Peter A. Hosner, and Nathalie Seddon Gull, Black-backed, see Larus mariniis California, see Larus californicus Lranklin’s, see Larus pipixcan Western, see Larus occidentalis Gunn, Carolyn, see Levad, Richard G., Kim M. Potter, Christopher W. Shultz, , and Joseph G. Doerr Gustafson, Mary, review, 235-236 Gutierrez, R. J., see Zimmerman, Guthrie S., Rick R. Horton, Daniel R. Dessecker, and Gymnorhimis cyanocephalus, 217 H Habia rubica, 12 habitat age influence on selection by male Wilsonia cana- densis, 446-454 bird responses to managed forested landscape, 897- 900 comparison of effect of burned and unbumed habi- tats on Jiinco hymenalis, 131 features of breeding Bucephala islandica in eastern Canada. 320-330 influence of grazing and available moisture on breeding densities of grassland birds in cen- tral Platte River Valley, Nebraska. 820-829 long-term effects of wastewater irrigation on habitat and a bird community in central Pennsylva- nia, 146-152 of nesting Charadrius melodus, 404-407 of Tyto soumagnei in Madagascar, 891—897 preference of Neomorphus radiolosus, 205—209 quality of anthropogenic habitats for Vermivora chry- soptera in central Pennsylvania. 801-812 riparian bird species response to vegetation and lo- cal habitat features, 840-855 riparian willow habitat of songbirds in Greater Yel- lowstone Ecosystem, 830-839 selection by Dendroica cerulea in Indiana, 105-1 10 selection by Picoides borealis, 793-800 use by fledged Catharus ustulatus, 62—73 wetland features that influence occupancy by Anas wyvilliana, 311-319 winter use by Poecile hudsonica, 139-145 Hadley, Adam and Andre Desrochers, Winter habitat use by Boreal Chickadee flocks in a managed forest, 139-145 Hager, Stephen B., Heidi Trudell, Kelly J. McKay, Ste- phanie M. Crandall, and Lance Mayer, Bird density and mortality at windows, 550-564 Haley, Katherin L., see Mojica, Elizabeth K., J. Mi- chael Meyers, Brian A. Millsap, and Haliaeetus leucocephalus, 304—330 Hall, C. Scott and Stephen W Kress, Diet of nestling Black-crowned Night-herons in a mixed spe- cies colony: implications for tern conservation, 637-640 Hallworth, Michael T, see Reitsma, Leonard R., , and Phred M. Benham Hamao, Shoji, Maria J. S. Veluz, Takema Saitoh, and Isao Nishiumi, Phylogenetic relationship and song differences between closely related Bush Warblers, (Cettia seebohmi and C. diphone), 268-276 Hanson, Thor, Lirst observation of duetting in the Ol- ive-backed Euphonia {Euphonia gouldi), 200- 201 Hardin, Jason, see Dreibelbis, Justin Z., Kyle B. Mel- ton, Ray Aguirre, Bret A. Collier, , Nova J. Silvy, and Markus J. Peterson Hardy, Douglas R. and Spencer P. Hardy, White- winged Diuca Pinch {Diuca speculifera) nest- ing on Quelccaya Ice Cap, Peru, 613-617 hare, snowshoe, see Lepus americanus Harrier, Northern, see Circus cyaneus Hawk, Red-tailed, see Buteo jamaicensis Sharp-shinned, see Accipiter striatus White-tailed, see Buteo albicaudatus Hawk-Owl, Togian, see Ninox burhani Heath, Shane R. and Prederick A. Servello, Effects of predation and food provisioning on Black Tern chick survival, 167-175 Helmitheros vertnivorurn [yermivorus), 280 Hemidactylus frenatus, 215 Hemignathus virens, 416—418 Hermit, Reddish, see Phaethornis ruber Hernandez, Pidel, see Breeden, Jeffrey B., , Ralph L. Bingham, Nova J. Silvy, and Gary L. Waggerman Hemandez-Banos, Blanca E., see Honey-Escandon, Magali, , Adolfo G. Navarro-Siguenza, Hesiquio Benitez-Diaz and A. Townsend Pe- terson Heron, Great Blue, see Ardea herodias Grey, see Ardea cinerea Highland-tanager, Hispaniolan, see Xenoligea montana Hilje, Branko, see Barrantes, Gilbert, Cesar Sanchez, , and Rodolfo Jaffe Hill, Geoffrey E., Ivorybill hunters: the search for proof in a flooded wilderness, reviewed, 232- 233 Hirundo rustica, 214 Histrionicus histrionicus, 599-602 Hobson, Keith A., see Wilson, Scott, , Douglas M. Collister, and Amy G. Wilson Hockey, Philip A. R., see Wanless, Ross M. and Hoeflich, Ernesto C. Enkerlin, see Valdes-Pena, Rene A., Sonia Gabriela Ortiz-Maciel, Simon O. Val- dez Juarez, , and Noel E R. Snyder home range for post-fledging grassland birds, 120-130 Honey-Escandon, Magali, Blanca E. Hernandez-Ba- nos, Adolfo G. Navarro-Siguenza, Hesiquio Benitez-Diaz, and A. Townsend Peterson, Phy- INDEX TO VOLUME 120, 2008 951 logeographic patterns of differentiation in the Acorn Woodpecker, 478-493 Horton, Rick R., see Zimmerman, Guthrie S., , Daniel R. Dessecker, and R. J. Gutierrez Hosner, Peter A. and Noemi E. Huanca, Nest, eggs, and parental care of the Puna Tapaculo {Scy- talopus simonsi), 473-477 Hosner, Peter A., see Tobias, Joseph A., Daniel J. Leb- bin, Alexandre Aleixo, Michael J. Andersen, Edson Guilherme, , and Nathalie Sed- don Huanca, N. E., see Cahill, J. R. A., E. Matthysen, and Huanca, Noemi E., see Hosner, Peter A. and Hubbard, Michael W., see Wells, Kimberly M. Sued- kamp, Joshua J. Millspaugh, Mark R. Ryan, and Huerta-Martinez, Erancisco Martin, see Rosas-Espi- noza, Veronica Carolina, Elisa Maya-Elizarrar- as, Oscar Erancisco Reyna Bustos, and Hummingbird, Amethyst-throated, see Lampornis amethystinus Anna’s, see Calypte anna Blue-throated, see Lampornis clemenciae Giant, see Patagona gigas Magnificent, see Eugenes ful gens Huyvaert, Kathryn R, see O’Brien, Valerie A., Amy T Moore, , and Charles R. Brown Hylocichla mustelina, 62, 64, 70, 127, 146-152, 155, 277-285, 296-303, 342 Hylophylax naevioides, 27, 29 Hypocnemis cantator, 33 Hyvonen, P, see Robertson, C. J. R., , M. J. Eraser, and C. R. Pickard I Icterus cucullatus, 342 dominicensis, 462 galhula, 297 parisorum, 342 Ictinia mississippiensis, 287 Indrawan, Mochamad, Pamela C. Rasmussen, and Sunarto, A new White-Eye (Zosterops) from the Togian Islands, Sulawesi, Indonesia, 1-9 Ingels, Johan, Nest, nestling care, and breeding season of the Spangled Cotinga (Cotinga cayana) in French Guiana, 871-874 Islam, Kamal, see Roth, Kirk L. and Islam, Zafar-Ul, see Menon, Shady, , Jorge Soberon, and A. Townsend Peterson Ixohrychus exilis, 5 1 3-5 1 8 J Jackson. Jerome A., George Miksch Sutton: artist, sci- entist, and teacher, reviewed, 231-232 jaeger, see Stercorarius spp. Jaeger, Long-tailed, see Stercorarius longicaudus Jaffe, Rodolfo, see Barrantes. Gilbert, Cesar Sanchez, Branko Hilje, and Janssen, Michael H., Peter Arcese, Mark S. Sloan, and Kelly J. Jewell, Polyandry and sex ratio in the Song Sparrow, 395-398 Jay, Blue, see Cyanocitta cristata Eurasian, see Garrulus glandarius Mexican, see Aphelocoma ultramarina Pinyon, see Gymnorhinus cyanocephalus Jeske, Clinton, see Baldwin, Michael J., Wylie C. Bar- row Jr., , and Frank C. Rohwer Jewell, Kelly J., see Janssen, Michael H., Peter Arcese, Mark S. Sloan, and Johnson, Andrew B. and Kevin Winker, Autumn stop- over near the Gulf of Honduras by Nearctic- Neotropic migrants, 277-285 Johnson, L. Scott, see Newhouse, Michael J., Peter P. Marra, and Jonsson, Jon Einar and Alan D. Afton, Lesser Snow Geese and Ross’s Geese form mixed flocks during winter but differ in family maintenance and social status, 725-731 Juarez, Simon O. Valdez, see Valdes-Pena, Rene A., Sonia Gabriela Ortiz-Maciel, , Ernesto C. Enkerlin Hoeflich, and Noel E R. Snyder Junco, Dark-eyed, see Junco hyemalis Junco hyemalis, 131-138, 283, 342 Junglefowl, Red, see Gallus gallus K Kappes Jr., John J., Cavity number and use by other species as correlates of group size in Red-cock- aded Woodpeckers, 181-189 Karubian, Jordan and Luis Carrasco, Home range and habitat preferences of the Banded Ground- cuckoo (Neomorphus radiolosus), 205—209 Kestrel, Madagascar see Falco newtoni Kim, Daniel H., Wesley E. Newton, Gary R. Lingle, and Felipe Chavez-Ramirez, Influence of graz- ing and available moisture on breeding densi- ties of grassland birds in the central Platte Riv- er Valley, Nebraska, 820-829 Kingbird, Eastern, see Tyrannus tyrannus Grey, see Tyrannus dominicensis Kinglet, Golden-crowned, see Regulus satrapa Kirschel, Alexander N. G., Novel courtship behavior in the Little Greenbul {Andropadus virens), 398-401 Kite, Mississippi, see Ictinia mississippiensis kittiwake, see Rissa spp. Kittiwake, Black-legged, see Rissa tridactyla Klippenstine, Dwight R. and Spencer G. Sealy, Dif- ferential ejection of cowbird eggs and non-mi- metic eggs by grassland pas.serines, 667-673 Koenig, Walter D., see Stromberg, Mark R., . Eric L. Walters, and John Schweisinger Kress, Stephen W. see Hall, C. Scott and Kubel, Jacob E. and Richard H. Yahner, Quality of anthropc^genic habitats for Golden- winged Warblers in central Pennsylvania, 801-812 Kusch, Alejandro, see Oehler, David A., Steve Pelikan. W. Roger Fry. Leonard Weakley Jr., . and Manuel Marin 952 THE W ILSON journal OL ORNITHOLOGY • VoL 120. So. 4. December 2008 L Lampomis amerhysrinus. 522 clemenciae. 522 viridipallens. 522 Lanins ludovicianus. >42 Larus califomicus. 1 59- 1 66 marinus. 505 occidentalis. 505 pipLxcan. 422 LareraUus jamaicensis. 513. 514 Latm. Steven C.. see Rimmer. Christopher C.. Lance G. Woolaver. Rina K. Nichols. Eladio M. Ler- nandez. . and Esteban Garrido Leader. Zohar. Yoram Yom-Tov. and Uzi Motro. Diet of the Long-eared Ow l in the northern and cen- tral Negev Desen, Israel. 641-645 Lebbin. Daniel J.. see Tobias. Joseph A.. . Al- exandre Aleixo. Michael J. Andersen. Edson Guilherme. Peter A. Hosner. and Nathalie Sed- don Lehman. Paul, review. 425 — 126 Lein. M. Ross. Song \ ariation in Buff-breasted Lly- catchers { Empidonax fidvifrons k 256-26~ Leptasrhenura adicola. 53“ xenorhorax. 534. 53“ yanacensis. 531-5-f4 Lepus americanns. 245 Leslie Jr.. David M.. see McConnell. Scon, Tunothy J. 0‘ Connell, and Levad. Richard G.. Kim M. Poner. Christopher W. Shultz. Carolyn Gunn, and Joseph G. Doerr. Distribution, abundance, and nest-site charac- teristics of Black Swifts in the southern Rocky Mountains of Colorado and New Mexico. 331- 338 Lewis. Thomas E.. see Nadeau. Christopher R. Coun- ney J. Conwav. Bradley S. Smith, and Li. Leng. see Wang. Qiang and Licaia. Diane, see Muir. James A.. . and Thom- as E. Martin Limnothlypis suainsonii. 433 — 145 Limosa fedoa, 914—915 Lindberg. Mark S.. see Saline. David E. and Lindenmayer. David B.. see Lelton. Adam, Annika M. Lelion. and Lingle. Garx R.. see Kim. Daniel H.. Wesley E. New- ton. . and Lehpe Cha\ es-Ramires Linkhan. Brian D.. Erin M. Evers. Julie D. Megler. Eric C. Palm. Catherine M. Salipante. and Scott W . Yanco. Eirst observed instance of polygyny in Flammulated Ow Is. 645— 648 Lloyd. Huw . Eoraging ecology of High .Andean insec- tivorous birds in remnant Polylepis forest patches. 531-5-44 Longspun Che smut-collared, see Calcarius omarus Smith's, see Calcarius pictus Lophodytes cucullams. 732-742 Lorenz. Stephan and Sampath Sene\eratne. Northern Eulmar predation of Common Murre. 635-636 Loxioides baiUeui. 31. 416 — 118 Loxops cocci ne us. 28 Lundari. \'itor O.. see Erancisco. Mercival R.. Paulo R. R. Oliveira Jr., and Luscinia megarhynchos. 254. 387 .M Macaw . Lears, see Anodorhynchus leari Mihtarx. see Ara militaris Red-fronted, see Ara rubrogenys Scarlet, see Ara rnacao Macrodipteryx longipennis. 510 Magpie. Red-billed Blue, see Urocissa erythrorhyncha Mallard, see Anas plaryrhynchos Malurus rnelanocephalus. 398 Manacus vitellinus. 32 Manakin. Golden-collared, see Manacus vitellinus Margarops fuscatus. -164 Marin. Manuel, see Oehler. David A.. Ste\e Pelikan. W. Roger Er>. Leonard Weakley Jr. Alejandro Kusch. and Marra. Peter R. see Newhouse. Michael J.. . and L. Scott Johnson manin. see Progne spp. Martin. Sand, see Riparia riparia Martin. Thomas E.. see Biancucci. Luis and Martin. Thomas E.. see Muir. James A.. Diane Licata, and Manin. Thomas E.. see Niklison. .Alina M.. Juan I. Areta. Roman .A. Ruggera. Karie L. Decker. Carlos Bosque, and Masters. Raymond D.. see McNuln. Stacy .A.. Sam Droege. and Matthysen. E.. see Cahill. J. R. .A.. . and N. E. Huanca Matuzak. Greg D.. M. Bemadene Bez>. and Donald J. Brightsmith. Eoraging ecologx of parrots in a modified landscape: seasonal trends and intro- duced species. 353-365 Maurice. Celine, see Dionne. Mark. . Jean Gau- thier. and Erancois Shaffer Maya-Elizarraras. Elisa, see Rosas-Espinoza, \ eronica Carolina. . Oscar Erancisco Re\ na Bus- tos. and Francisco Manin Huerta-Maninez Ma>er. I.ance. see Hager. Stephen B.. Heidi Trudell. Kelly J. McKay. Stephanie M. CrandaU. and McCaffery. Brian J.. see Renner. Heather M. and McCarthy. Eugene M.. Handbook of avian hybrids of the world, reviewed. 233-235 McConnell. Scott. Timothy J. O'Connell, and David M. Leslie Jr.. Land cover associations of nest- ing territories of three s>mpatric buteos in shortgrass prairie. 708-716 McCracken. Kevin G.. see Wilson. Roben E. and McKa\. Bailey D.. A recording of a tx pe B song of the Yellow -throated W arbler. -fOl — W)3 McKay. Kelly J.. see Hager. Stephen B.. Heidi Trudell. . Stephanie M. Crandall, and Lance Mayer INDEX TO VOLUME 120, 2008 953 McKinnon, Emily Anne and Raleigh J. Robertson, The signal function of a melanin-based plumage or- nament in Golden-winged Warblers, 366-370. McNulty, Stacy A., Sam Droege, and Raymond D. Masters, Long-term trends in breeding birds in an old-growth Adirondack forest and the sur- rounding region, 153-158 McPherson, Roger J., see Ammer, Frank K., Petra Bo- hall Wood, and Meadowlark, Eastern, see Stiirnella magna Meadowlark, Western, see Sturnella neglecta Megadyptes antipodes, 378 Megler, Julie D., see Linkhart, Brian D., Erin M. Ev- ers, , Eric C. Palm, Catherine M. Sali- pante, and Scott W. Yanco Meijia, Danilo A., see Townsend, Jason M., Esteban Garrido, and Melanerpes albeolus, 486 angiistifrons, 486 aurifrons, 479, 495 carolimis, 149, 181-189, 219, 220 erythrocephalus, 219 flavifrons, 219 flavigida, 486 formicivonis, 219, 478—493, 494—498 hojftnatmi, 214-216 hypopoliiis, 219 lewis, 219, 479 lineatus, 486 pucherani, 479 pygmaeus, 479 striatipectus, 486 uropygialis, 479, 495 Melanitta fusca, 582-593 spp., 594-599 Meleagris gallopavo, 226-228, 906-910 ocellata, 229 Melospiza georgiana, 257 georgiana nigrescens, 393-395 melodia, 146-152, 297, 342, 395-398, 470 Melozone leucotis, 214 sp., 214 Melton, Kyle B., Justin Z. Dreibelbis, Ray Aguirre, Bret A. Collier, T Wayne Schwertner, Markus J. Peterson, and Nova J. Silvy, Abnormal eggs of Rio Grande Wild Turkeys on the Edwards Plateau, Texas, 226-228. Melton, Kyle B., see Dreibelbis, Justin Z., , Ray Aguirre, Bret A. Collier, Jason Hardin. Nova J. Silvy, and Markus J. Peterson Mennill, Daniel J., see Tremain, Sarah B., Kyle A. Swiston, and Menon, Shaily, Zafar-Ul, Islam, Jorge Soberon, and A. Townsend Peterson, Preliminary analysis of the ecology and geography of the Asian nuthatches (Aves; Sittidae). 692-699 Merganser, Hooded, see Lophodytes cncullatus Red-breasted, see Mergus senator Mergus serrator, 743—754 methods effect of traffic noise on auditory surveys of urban Zenaida asiatica, 384-389 gender identification of Ammodraimis savannarwn comparing behavioral, morphological, and molecular techniques, 221-225 gender identification of Sterna caspia using external morphology, 378-383 maximizing detection probability of wetland-depen- dent birds during point-count surveys in northwestern Florida, 513-518 Meyers, J. Michale, see Mojica, Elizabeth K., , Brian A. Millsap, and Katherin L. Ha- ley Micrastiir riificollis, 228-230 seniitorquatiis, 229 Microligea palustris, 194 migration autumn stopover of Nearctic and Neotropic migrants near the Gulf of Honduras, 277-285 changes of migratory landbirds during stopovers in a New York City park, 296-303 effect of hurricane Wilma on migrating Chaetura pelagica, 784-792 of Florida subadult Haliaeetiis leucocephalus, 304- 310 stopover of migrating birds in New Orleans, Loui- siana in relation to weather, 286-295 stopover of migrating Cathariis ustulatus in south- west Costa Rica, 74-84 Millsap, Brian A., see Mojica. Elizabeth K., J. Michael Meyers, , and Katherin L. Haley Millspaugh, Joshua J., see Wells, Kimberly M. Sued- kamp, , Mark R. Ryan, and Michael W. Hubbard Miinus polyglottos, 528, 717-724 mink, see Mustela vison Minser, William G., see Giocomo, James L., E. Daniel Moss, David A. Buehler, and Mockingbird, Northern, see Minnis polyglottos Mojica, Elizabeth K., J. Michael Meyers. Brian A. Millsap, and Katherin L. Haley, Migration of Florida subadult Bald Eagles, 304-310 Molothrus aeneiis, 287 ater, 113, 114, 149, 340, 467-472. 625-627, 667- 673, 820-829 bonariensis, 460-466 Momotns nioniota, 214—216, 653—654 Monticola gularis, 97 soli tar ins, 97 Moore. Amy T. see O'Brien. Valerie A.. . Kathryn P. Huyvaert, and Charles R. Brown Moorhen. Common, see Gallinula chloropns Moreno, Marisela Pando, see Canales-Delgadillo. Julio C.. Laura M. Scott-Morales. Mauricio Cotera Correa, and morphology gender identification Sterna caspia using external morphology. 378-383 specimen shrinkage in Cinnamon Teal. 390-392 mortality t)r injury at windows. 550-564 954 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 Limosa fedoa collides with aircraft at 3,700 m, 914- 915 of Band-tailed Pigeons, 603-606 of fledgling, subadult, and adult Lams californicus, 159-166 Porphyria porphyria attempting to prey upon Cyg- nus atratus, 633—635 predation by Ardea cinerea on Dryolimnas cuvieri aldabranus, 631—632 predation of Uria aalge by Fidmariis glacialis, 635— 636 Morton, Eugene S., see Stutchbury, Bridget J. M. and Moss, E. Daniel, see Giocomo, James L., , Da- vid A. Buehler, and William G. Minser Motacilla alba, 85 flava, 89 tschutschensis, 85—91 Motmot, Blue-crowned, see Momotus momota Motro, Uzi, see Leader, Zohar, Yoram Yom-Tov, and Mountain-gem, Green-throated, see Lampornis viridi- pallens Moyle, Robert G., Beth Slikas, Linda A. Whittingham, David W. Winkler, and Frederick H. Sheldon, DNA sequence assessment of phylogenetic re- lationships among New World martins (Hirun- dinidae: Progne), 683-691 Muir, James A., Diane Licata, and Thomas E. Martin, Reproductive biology the Red-ruffed Fruitcrow (Pyroderus scutatus granadensis), 862-867 Murre, Common, see Uria aalge Mustela gularis, 96 spp., 88 vison, 173 Myadestes townsendi, 332, 342 Myiarchus cinerascens, 813—819 'crinitus, 148, 149, 183, 287 Myrmeciza exsid, 33 longipes, 29, 30, 33 Myrmotherid snowi, 20 unicolor, 20 N Nadeau, Christopher R, Courtney J. Conway, Bradley S. Smith, and Thomas E. Lewis, Maximizing detection probability of wetland-dependent birds during point-count surveys in northwest- ern Florida, 513-518 natural history Dryolimnas [cuvieri^ aldabranus, 50-61 Grallaricula ferrugineipectus, 345-352 Navarro-Siguenza, Adolfo G., see Honey-Escandon, Magali, Blanca E. Hernandez-Banos, , Hesiquio Benitez-Diaz, and A. Townsend Pe- terson Neomorphus radiolosus, 205—209 nest and nesting behavior of Tachycineta euchrysea, 867-87 1 and nestling care of Cotinga cayana, 871-874 defense by Thryothorus ludovicianus, 467-472 description for Calyptophilus tertius and Xenoligea montana, 190-195 habitat selection by Melanitta fusca, 582-593 of Pyroderus scutatus, 413—416 of Scytalopus simonsi, A12-A11 raising by Grus japonensis, 610-613 reuse by Pyrocephalus rubinus in Texas, 339—344 site characteristics of Cypseloides niger in southern Rocky Mountains Colorado and New Mexi- co, 331-338 success and predation of Zosterops rotensis, 618- 619 nesting ecology of Bucephala clangula and Lophodytes cu- cullatus, 732-742 of Buteo spp., 708-716 of grassland birds at Fort Campbell, Kentucky and Tennessee, 1 1 1-1 19 of Oreomanes fraseri, 545—549 of Oreomystis bairdi, 195-199 nestling postnatal growth rates of hummingbirds, 884-887 survival of Chlidonias niger, 167-175 Neudorf, Diane L. H., see D’Orazio, Kelly A. and Newhouse, Michael J., Peter P. Marra, and L. Scott Johnson, Reproductive success of House Wrens in suburban and rural landscapes, 99-104 Newton, Wesley E., see Kim, Daniel H., , Gary R. Lingle, and Felipe Chavez-Ramirez Nichols, Rina K., see Rimmer, Christopher C., Lance G. Woolaver, , Eladio M. Fernandez, Steven C. Latta, and Esteban Garrido Nicolai, Christopher A., James S. Sedinger, and Mi- chael L. Wege, Differences in growth of Black Brant goslings between a major breeding col- ony and outlying breeding aggregations, 755- 766 Night-heron, Black-crowned, see Nycticorax nycticor- ax Nightingale, Common, see Luscinia megarhynchos Nightjar, Freckled, see Caprimulgus tristigma Long-tailed, see Caprimulgus climacurus Standard-winged, see Macrodipteryx longipennis Niklison, Alina M., Juan I. Areta, Roman A. Ruggera, Karie L. Decker, Carlos Bosque, and Thomas E. Martin, Natural history and breeding biology of the Rusty-breasted Antpitta {Grallaricula ferrugineipectus), 345-352 Ninox burhani, 8 Nishiumi, Isao, see Hamao, Shoji, Maria J. S. Veluz, Takema Saitoh, and noddy, see Anous spp. Nucifraga columbiana, 217 Nur, Nadav, Grant Ballard, and Geoffrey R. Geupel, Regional analysis of riparian bird species re- sponse to vegetation and local habitat features, 840-855 Nutcracker, Clark’s, see Nucifraga columbiana Nuthatch, Brown-headed, see Sitta pusilla INDEX TO VOLUME 120, 2008 955 Pygmy, see Sitta pygmaea Red-breasted, see Sitta canadensis White-breasted, see Sitta carolinensis Nycticorax nycticorax, 173, 637—640 o O’Brien, Valerie A., Amy T. Moore, Kathryn P. Huy- vaert, and Charles R. Brown, No evidence for spring re-introduction of an arbovirus by Cliff Swallows, 910-913 O’Connell, Timothy J., see McConnell, Scott, , and David M. Leslie Jr. Odontophorus hyperythrus, 207 Oehler, David A., Steve Pelikan, W. Roger Fry, Leo- nard Weakley Jr., Alejandro Kusch, and Ma- nuel Marin, Status of Crested Penguin {Eudyp- tes spp.) populations on three islands in south- ern Chile, 575-581 Oenanthe oenanthe, 80, 342 Olechnowski, Brian E M. and Diane M. Debinski, Re- sponse of songbirds to riparian willow habitat structure in the Greater Yellowstone Ecosys- tem, 830-839 Oliveira Jr., Paulo R. R., see Francisco, Mercival R., , and Vitor O. Lunardi Oporornis formosus, 277-285 tolmiei, 478 Oreomanes fraseri, 531-544, 545-549 Oreomystis bairdi, 195-199 Oriole, Baltimore, see Icterus galbula Greater Antillean, see Icterus dominicensis Hooded, see Icterus cucullatus Scott’s, see Icterus parisorum Ortiz-Maciel, Sonia Gabriela, see Valdes-Pena, Rene A., , Simon O. Valdez Juarez, Ernesto C. Enkerlin Hoeflich, and Noel F. R. Snyder Otus ftammeolus, 645—648 Ovenbird, see Seiurus aurocapilla [auricapillus] Overington, Sarah E., Laure Cauchard, and Kimberly- Ann Cote, Kleptoparasitism by Grey Kingbirds {Tyrannus dominicensis) in Barbados, 655-657 Owl, Black-and-white, see Strix nigrolineata Burrowing, see Athene cunicularia Great Gray, see Strix nebulosa Great Horned, see Bubo virginianus Long-eared, see Asio otus Madagascar Red, see Tyto soumagnei Northern Saw-whet, see Aegolius acadicus Owlet-Nightjar, see Aegotheles cristatus P Pali la, see Loxioides bailleui Palm, Eric C., .see Linkhart, Brian D., Erin M. Evers, Julie D. Megler, , Catherine M. Salipan- te, and Scott W. Yanco Parakeet, Orange-chinned, see Brotogeris jugularis Orange-fronted, see Aratinga canicularis Pacific, see Aratinga strenua Reddish-bellied, see Pyrrhura frontalis Parrot, Grey, see Psittacus erithacus Lilac-crowned, see Amazona finschi Maroon-fronted, see Rhynchopsitta terrisi Scaly-headed, see Pionus maximiliani Thick-billed, see Rhynchopsitta pachyrhyncha Parula, Northern, see Parula americana Parula americana, 155, 157, 455, 458 Parus major, 387, 446 Passer domesticus, 30, 121, 368, 391, 422 Passerculus sandwichensis, 394, 628-630, 667-673 Passerella iliaca, 830-839 Passerina ciris, 342 cyanea, 146-152, 277-285, 286-295 Patagioenas fasciata, 603-606 Patagioenas plumbea, 228-230 Patagona gigas, 648-651, 651-653 Paxton, Barton J., see Watts, Bryan D., Michael D. Wilson, Fletcher M. Smith, , and J. Bill Williams Pelikan, Steve, see Oehler, David A., , W. Rog- er Fry, Leonard Weakley Jr., Alejandro Kusch, and Manuel Marin Penguin, Yellow-eyed, see Megadyptes antipodes Perlut, Noah G., see Zalik, Nathan J. and Peterson, A. Townsend, see Honey-Escandon, Magali, Blanca E. Hernandez-Banos, Adolfo G. Navar- ro-Siguenza, Hesiquio Benitez-Diaz, and Peterson, A. Townsend, see Menon, Shaily, Zafar-Ul, Islam, Jorge Soberon, and Peterson, Markus J., see Dreibelbis, Justin Z., Kyle B. Melton, Ray Aguirre, Bret A. Collier, Jason Hardin, Nova J. Silvy, and Peterson, Markus J., see Melton, Kyle B., Justin Z. Dreibelbis, Ray Aguirre, Bret A. Collier, T. Wayne Schwertner, , and Nova J. Silvy Petrochelidon pyrrhonota, 910-913 Phaethornis ruber, 201-204 Phalacrocorax spp., 376 Phalaenoptilus nuttallii, 505—512 Pheucticus ludovicianus, 149, 151 Philydor atricapillus, 20 novaesi, 20 Phoenicopterus roseus, 121 Phylloscartes ceciliae, 20 Phylloscopus borealis, 85 trochilus, 388 Pickard, C. R., see Robertson, C. J. R., P. Hyvonen, M. J. Fraser, and Picoides albolarvatus, 217, 219 arizonae, 495 borealis, 1 8 1 - 1 89, 2 1 7-22 1 , 793-800 pubescens, 150, 219 scalaris, 495 villosus, 148, 149, 155, 217, 219 Pic us viridis, 219 Pigeon, Band-tailed, see Patagioenas fasciata Plumbeous, see Patagioenas plumbea Pionus maximiliani, 357 Pipilo crissalis, 342 erythrophthalmus, 149. 150, 151 maculatus, 342 956 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 Pipit, Red-throated, see Anthus cervinus Sprague’s, see Anthus spragueii Pistorius, Pierre A., Grey Heron {Ardea cinerea) pre- dation on the Aldabra White-throated Rail {Dryolimnas cuvieri aldabraniis), 631-632 Plover, Mountain, see Charcidrius rnontaniis Piping, see Charadrius melodus plumage of Vermivora chrysoptera, 366-370 Podilymbus podiceps, 5 1 4 Poecile atricapillus, 139-145, 150, 155, 367 cincta, 139, 143 hudsonica, 139-145 montana, 97 paliistris, 97 Polioptila caerulea, 287 melanura, 342 Pooecetes gramineus, 667-673 Poorwill, Common, see Phalaenoptilus mittallii population cycles in Bonasa iimbellus, 239-247 of Charadrius rnontanus in northeastern Montana, 700-707 of Eudyptes spp. in southern Chile, 575-581 sex ratio in Melospiza melodia, 395-398 Porphyria martinicus, 513-518 Porphyria porphyria, 633-635 Porzana Carolina, 514 Post, William, Winter ecology of Yellow Rails based on South Carolina specimens, 606-610 Potter, Kim M., see Levad, Richard G., , Chris- topher W. Shultz, Carolyn Gunn, and Joseph G. Doerr Powell, Shawn C., see Yaukey, Peter H. and Prairie-Chicken, Greater, see Tympanuchus cupido Predation at nests of Meleagris gallapavo, 906-910 at nests of Tharnnophilus atrinucha, 620-624 by Micrastur ruficollis on relatively large prey, 228— 230 effects on Chlidonias niger, 167-175 intraspecific killing by Egretta sacra, 422-424 Premnoplex brunnescens, AlO Price, J. Jordan, see Reichard, Dustin G. and Priotelus roseigaster, 249 proceedings eighty-ninth annual meeting, 927-940 Procnias tricarunculatus, 522 Procyon lotor, 114, 173 Progne spp., 683-691 subis, 287 Protonotaria citrea, 455, 458 Prunella collaris, 397, 398 modularis, 397 Psittacus erithacus, 176 Ptychoramphus aleuticus, 505 Puffinus puffinus, 505 spp., 378 Pugesek, Bruce H. and Kenneth L. Diem, Timing and location of mortality of fledgling, subadult, and adult California Gulls, 159-166 Pygmy-Owl, Yungas, see Glaucidium bolivianum Pyrocephalus rubinus, 339-344 Pyroderus scutatus, 413-416, 862-867 Pyrrhula erythaca, 874-878 Pyrrhura frontalis, 357 Q Quail, Mountain Wood, see Odontophorus hyperythrus Quetzal, Eared, see Euptilotis neoxenus Quiscalus major, 342 mexicanus, 342 quiscula, 149 R raccoon, see Procyon lotor Rail, Aldabra, see Dryolimnas [cuvieri] aldabranus Black, see Laterallus jamaicensis Clapper, see Rallus longirostris King, see Rallus elegans Virginia, see Rallus limicola White-throated, see Dryolimnas cuvieri Yellow, see Coturnicops noveboracensis Rallus elegans, 514, 515 limicola, 513 longirostris, 513—518 Rana catsbeiana, 312 range annual and seasonal use by Picoides borealis, 793- 800 of breeding Melospiza georgiana migrescens, 393- 395 of Neomorphus radiolosus, 205—209 phylogeographic patterns of Melanerpes formicivo- rus, 478-493 Rasmussen, Pamela C., see Indrawan, Mochamad, , and Sunarto rat, see Rattus spp. black, see Rattus rattus Norway, see Rattus norvegicus Rattus norvegicus, 60, 194 rattus, 59, 194 spp., 190, 458 Raven, Common, see Conms corax Reboreda, Juan C., see Lacchinetti, Carolina, Alejan- dro G. Di Giacomo, and Reboreda, Juan C., see Lernandez, Gustavo J., and Redstart, American, see Setophaga ruticilla Reetz, Matthew J., Elizabeth Farley, and Thomas A. Contreras, Evidence for Bachman’s Sparrow raising Brown-headed Cowbirds to fledging, 625-627 Regains satrapa, 155 Reichard, Dustin G. and J. Jordan Price, Species rec- ognition in a vocal mimic: repetition pattern not the only cue used by Northern Mocking- birds in discriminating songs of conspecifics and Brown Thrashers, 717-724 Reitsma, Leonard R., Michael T. Hallworth, and Phred M. Benham, Does age influence territory size. INDEX TO VOLUME 120, 2008 957 habitat selection, and reproductive success of male Canada Warblers in central New Hamp- shire? 446-454 Renner, Heather M. and Brian J. McCaffery, Demog- raphy of Eastern Yellow Wagtails at Cape Ro- manzof, Alaska, 85-91 report editor comments, 237 reproduction does age influence reproductive success of male Wilsonia canadensis, 446—454 Rhea, Greater, see Rhea americana Rhea americana, 674-682 Rhynchopsitta pachyrhyncha, 176 terrisi, 176—180 Riehl, Christina, Communal calling and prospecting by Black-headed Trogons (Trogon melanocephal- us), 248—255 Rimmer, Christopher C., Lance G. Woolaver, Rina K. Nichols, Eladio M. Fernandez, Steven C. Latta, and Esteban Garrido, First description of nests and eggs of two Hispaniolan endemic species; Western Chat-tanager {Calyptophilus tertius) and Hispaniolan Highland-tanager (Xenoligea montana), 190-195 Riparia riparia, 399 Rissa spp., 378 tridactyla, 399 Robert, Michel, Bruno Drolet, and Jean-Pierre L. Sa- vard. Habitat features associated with Barrow’s Goldeneye breeding in eastern Canada, 320- 330 Roberts, Pauline K., see VanderWerf, Eric A. and Robertson, C. J. R., P Hyvonen, M. J. Fraser, and C. R. Pickard, Atlas of bird distribution in New Zealand 1999-2004, reviewed, 426-430 Robertson, Raleigh J., see McKinnon, Emily Anne and Robin, American, see Turdus migratorius Clay-colored [Thrush], see Turdus grayi European, see Erithacus rubecula White-throated [Thrush], see Turdus assimilis Rodgiguez-Ferraro, Adriana, and Virginia Sanz, Nat- ural history and population status of the Yel- low-shouldered Parrot on La Blanquilla Island, Venezuela, Rohe, Fabio and Andre Pinassi Antunes, Barred Forest Falcon {Micrastur ruficollis) predation on rel- atively large prey, 228-230 Rohnke, Adam T and Ricahrd H. Yahner, Long-term effects of wastewater irrigation on habitat and a bird community in central Pennsylvania, 146-152 Rohwer, Frank C., see Baldwin, Michael J., Wylie C. Barrow Jr., Clinton Jeske, and Rosas-Espinoza, Veronica Carolina, Elisa Maya-Eli- zarraras, Oscar Francisco Reyna Bustos, and Francisco Martin Huerta-Martinez, Diet of Acorn Woodpeckers at La Primavera Forest, Jalisco, Mexico, 494-498 Roth, Kirk L. and Kamal Islam, Habitat selection and reproductive success of Cerulean Warblers in Indiana, 105-110 Ruggera, Roman A., see Niklison, Alina M., Juan I. Areta, , Karie L. Decker, Carlos Bosque, and Thomas E. Martin Ryan, Mark R., see Wells, Kimberly M. Seudkamp, Joshua J. Millspaugh, , and Michael W. Hubbard s Safine, David E. and Mark S. Lindberg, Nest habitat selection of White-winged Scoters on Yukon Flats, Alaska, 582-593 Saitoh, Takema, see Hamao, Shoji, Maria J. S. Veluz, , and Isao Nishiumi Salipante, Catherine M., see Linkhart, Brian D., Erin M. Evers, Julie D. Megler, Eric C. Palm, , and Scott W. Yanco Salvelinus fontinalis, 320, 322 Sanchez, Cesar, see Barrantes, Gilbert, , Branko Hilje, and Rodolfo Jaffe Sandoval, Luis, Esteban Biamonte, and Alejandro So- lana-Ugalde, Previously unknown food items in the diet of six neotropical bird species, 214- 216 Sandpiper, Upland, see Bartramia longicauda Sapsucker, Red-naped, see Sphyrapicus nuchalis Yellow-bellied, see Sphyrapicus varius Savard, Jean-Pierre L., see Robert, Michel, Bruno Dro- let, and Savard, Jean-Pierre L., see Senechal, Helene, Gilles Gauthier, and Saxicola torquatus, 28, 29, 30 Schuchmann, Karl-Ludwig, see Freymann, Bernd P. and Schweisinger, John, see Stromberg, Mark R., Walter D. Koenig, Eric L. Walters, and Schwertner, T. Wayne, see Melton, Kyle B., Justin Z. Dreibelbis, Ray Aguirre, Bret A. Collier, , Markus J. Peterson, and Nova J. Silvy Sciurotamias davidianus, 96 Scoter, species, see Melanitta spp. White-winged, see Melanitta fusca Scott-Morales, Laura M., see Canales-Delgadillo, Julio C., , Mauricio Cotera Correa, and Mar- isela Pando Moreno Scytalopus magellanicus, 473 micropterus, 473, 476 parkeri, 476 schulenhergi, 475 simonsi, 473-477 superciliaris, 473, 475 Sealy, Spencer G., see Klippenstine, Dwight R. and Seddon, Nathalie, see Tobias, Joseph A., Daniel J. Lebbin, Alexandre Aleixo, Michael J. Ander- .sen, Edson Guilherme, Peter A. Hosner, and Sedinger, James S., see Nicolai, Christopher A., , and Michael L. Wege 958 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 Seedeater, Rusty-collared, see Sporophila collaris Tawny-bellied, see Sporophila hypoxantha Seewagen, Chad L. and Eric J. Slayton, Mass changes of migratory landbirds during stopovers in a New York City park, 296-303 Seiurus aurocapilla, [auricapillus] 62, 149, 150, 151, 157, 277-285, 296-303, 446, 455-450 noveboracensis, 277-285, 296-303, 455, 458 Senechal, Helene, Gilles Gauthier, and Jean-Pierre L. Savard, Nesting ecology of Common Golden- eyes and Hooded Mergansers in a boreal river system, 732-742 Seneveratne, Sampath, see Lorenz, Stephan and Servello, Frederick A., see Heath, Shane R. and Setophaga ruticilla, 80, 149, 150, 151, 155, 279, 289, 291, 292, 446 Shaffer, Francois, see Dionne, Mark, Celine Maurice, Jean Gauthier, and shag, see Phalacrocorax spp. shearwater, see Pufftnus spp. Shearwater, Manx, see Pufftnus pufftnus Sheldon, Frederick H., see Moyle, Robert G., Beth Sli- kas, Linda A. Whittingham, David W. Winkler, and Sherry, Thomas W, see Brown, David R. and Shrike, Loggerhead, see Lanius ludovicianus Shultz, Christopher W, see Levad, Richard G., Kim M. Potter, , Carolyn Gunn, and Joseph G. Doerr Sialia mexicana, 569, 813-819 sialis, 183, 419-422, 565-568 Silcock, Ross, review, 426-430 Silvereye, see Zosterops lateralis Silvy, Nova J., see Breeden, Jeffrey B., Fidel Hernan- dez, Ralph L. Bingham, , and Gary L. Waggerman Silvy, Nova J., see Dreibelbis, Justin Z., Kyle B. Mel- ton, Ray Aguirre, Bret A. Collier, Jason Hardin, , and Markus J. Peterson Silvy, Nova J., see Melton, Kyle B., Justin Z. Drei- belbis, Ray Aguirre, Bret A. Collier, T Wayne Schwertner, Markus J. Peterson, and Siskin, Thick-billed, see Carduelis crassirostris Sitta canadensis, 142 carolinensis, 217 pusilla, 217 pygmaea, 217 spp., 692-699 Slayton, Eric J., see Seewagen, Chad L. and Slikas, Beth, see Moyle, Robert G., , Linda A. Whittingham, David W. Winkler, and Frederick H. Sheldon Sloan, Mark S., see Janssen, Michael H., Peter Arcese, , and Kelly J. Jewell Smith, Bradley S., see Nadeau, Christopher R, Court- ney J. Conway, , and Thomas E. Lewis Smith, Fletcher M., see Watts, Bryan D., Michael D. Wilson, , Barton J. Paxton, and J. Bill Williams snake, Chifeng beauty, see Elaphe anomala eastern rat, see Elaphe alleghaniensis Puerto Rican boa, see Epicrates inomatus Puerto Rican racer, see Alsophis portoricensis rat, see Elaphe obsoleta Snow, Carter W, see Farmer, Chris, Bridget A. Fred- erick, Paul C. Banko, Robert M. Stephens, and Snyder, Noel F. R., see Valdes-Pena, Rene A., Sonia Gabriela Ortiz-Maciel, Simon O. Valdez Jua- rez, Ernesto C. Enkerlin Hoeflich, and Soberon, Jorge, see Menon, Shaily, Zafar-Ul, Islam, , and A. Townsend Peterson Solano-Ugalde, Alejandro and Agustina Arcos-Torres, Nocturnal foraging observations of the Blue- crowned Motmot (Momotus momota) in San Jose, Costa Rica, 653-654 Solano-Ugalde, Alejandro, see Sandoval, Luis, Este- ban Biamonte, and Solitaire, Townsend’s, see Myadestes townsendi Sora, see Porzana Carolina Sparrow, Baird’s, see Ammodramus bairdii Black-throated, see Amphispiza bilineata Coastal Plain Swamp, see Melospiza georgiana ni- grescens Field, see Spizella pusilla Fox, see Passerella iliaca Grasshopper, see Ammodramus savannarum ground, see Melozone spp. House, see Passer domesticus Lark, see Chondestes grammacus Rufous-collared, see Zonotrichia capensis Savannah, see Passerculus sandwichensis Seaside, see Ammodramus maritimus Song, see Melospiza melodia Swamp, see Melospiza georgiana Vesper, see Pooecetes gramineus White-crowned, see Zonotrichia leucophrys White-eared Ground, see Melozone leucotis Worthen’s, see Spizella wortheni species nova from genus Automalus, 10-25 Togian White-eye, Zosterops somadikartai, 1-9 Sperry, Jinelle H., T. Luke George, and Steve Zack, Ecological factors affecting response of Dark- eyed Juncos to prescribed burning, 131-138 Sphyrapicus nuchalis, 215, 479 varius, 219, 495 Spinetail, Creamy-crested, see Cranioleuca albicapilla Spiza americana, 62, 72, 111-119, 120-130, 820-829 Spizella pusilla, 111-119, 342 wortheni, 569-574 Sporophila collaris, 879-883 hypoxantha, 879-883 squirrel, flying, see Glaucomys volans rock, see Sciurotamias davidianus Swinhoe’s striped, see Tamiops swinhoei Stephens, Robert M., see Farmer, Chris, Bridget A. INDEX TO VOLUME 120, 2008 959 Frederick, Paul C. Banko, , and Carter W. Snow Stercorarius longicaudus, 88 spp., 378 Sterna caspia, 378-383 forsteri, 381 hirundo, 381 paradisaea, 381 Stonechat, see Saxicola torquatus Strix nebulosa, 422 nigrolineata, 214—216 Stromberg, Mark R., Walter D. Koenig, Eric L. Wal- ters, and John Schweisinger, Estimate of Trich- omonas gallinae-'xndncQd mortality in Band- tailed Pigeons, upper Carmel Valley, Califor- nia, winter 2006-2007, 603-606 Stubtail, Asian, see Urosphena squameiceps Sturnella magna, 62, 111-119, 120-130 neglecta, 667-673 Stutchbury, Bridget J. M. and Eugene S. Morton, Re- cent advances in the behavioral ecology of tropical birds, 26-37 Sula nebouxii, 399 Sunarto, see Indrawan, Mochamad, Pamela C. Ras- mussen, and survival of Motacilla tschutschensis at Cape Romanzof, Alaska, 85-91 Swallow, Barn, see Hirundo rustica Cliff, see Petrochelidon pyrrhonota Golden, see Tachycineta euchrysea Tree, see Tachycineta bicolor Violet-green, see Tachycineta thalassina Swamphen, Purple, see Porphyrio porphyrio Swan, Black, see Cygnus atratus Swift, Black, see Cypseloides niger Chimney, see Chaetura pelagica Swiston, Kyle A., see Tremain, Sarah B., , and Daniel J. Mennill Sykes Jr., Paul W., Lyn S. Atherton, and Rebecca L. Payne, Yellow-throated and Red-eyed vireos foraging on green anoles during migration, Sylvia atricapilla, 210, 269, 272, 279, 433 communis, 212 Synallaxis riitilans, 1 8 T Tachycineta [Tachycinita] albilinea, 31 bicolor, 422 euchrysea, 867-87 1 thalassina, 8 1 3—8 1 9 Taisacan, Estanislao, see Berry, Lainie and Takekawa, John Y., see Ackerman, Joshua T, — , Jill D. Bluso, Julie L. Yee, and Collin A. Eagles-Smith Tamiops swinhoei, 96 Tapacula, Long-tailed, see Scytalopus micropterus Magellanic, see Scytalopus magellanicus Puna, see Scytalopus simonsi White-browed, see Scytalopus superciliaris Tarwater, Corey E., Predators at nests of the Western Slaty Antshrike, {Thamnophilus atriniicha), 620-624 taxonomy of Cettia seebohrni and Cettia diphone, 26?>—216 Teal, Chestnut, see Anas castanea Cinnamon, see Anas cyanoptera Tegu, Black-white, see Tupinambis merianae Terenura maculata, 20 sicki, 20 Tern, Arctic, see Sterna paradisaea Black, see Chlidonias niger Caspian, see Sterna caspia Common, see Sterna hirundo Forster’s, see Sterna forsteri Thamnophilus atrinucha, 620—624 Thrasher, Bendire’s, see Toxostoma bendirei Brown, see Toxostoma rufum Curve-billed, see Toxostoma curvirostra LeConte’s, see Toxostoma lecontei Pearly-eyed, see Margarops fuscatus Thrush, Bicknell’s, see Catharus bicknelli Blue Rock, see Monticola solitarius Clay-colored, see Turdus grayi Great, see Turdus fuscater Hermit, see Catharus guttatus Pale-breasted, Turdus leucornelas Red-legged, see Turdus plumbeus Rufous-bellied, see Turdus rufiventris Song, see Turdus philomelos Sooty, see Turdus nigrescens Swainson’s, see Catharus iistulatus White-throated, see Monticola gularis Wood, see Hylocichla rnustelina Thryothorus euophrys, 34 fasciatoventris, 34 leucotis, 31, 33, 34 ludovicianus, 467—472 rufalbus, 34 Tinamou, Brown, see Crypturellus obsoletus Tit, Great, see Parus major Marsh, see Poecile palustris Willow, see Poecile montana Tit-spinetail, Andean, see Leptasthenura andicola Tawny, see Leptasthenura yanacensis White-browed, see Leptasthenura xenothorax Tit-Tyrant, Ash-breasted, see Anairetes alpinus Tufted, see Anairetes parulus Titanohierax gloveralleni, Titman, Rodger D., see Craik, Shawn R. and Titmouse, Tufted, see Baeolophus bicolor Tobias, Joseph A., Daniel J. Lebbin, Alexandre Aleixo. Michael J. Andersen, Edson Guilherme, Peter A. Hosner and Nathalie Seddon, Distribution, behavior, and conservation status of the Rufous Twistwing {Cnipodectes superrufus), 38-49 Tomasevic, Jorge A., PNtades, Cristian. M. Angelica Vukasovic. and tortoise, Aldabra giant, sec Dipsochelys dussumieri Tossas. Adrianne G., Reproductive success of the Puerto Rican Vireo in montane habitat. 460- 466 960 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 4, December 2008 Towhee, California, see Pipilo crissalis Eastern, see Pipilo erythrophthalmiis Spotted, see Pipilo maculatiis Townsend, Jason M., Esteban Garrido, and Danilo A. Mejia, Nests and nesting behavior of Golden Swallow (Tachycineta eiichrysea) in aban- doned bauxite mines in the Dominican Repub- lic, 867-871 To.xostoma bendirei, 342 cun'irostre, 342 lecontei, 342 rufum, 343 Tozer, Douglas C., Nests of Black-throated Green War- blers in tree cavities, 409-412 Tremain, Sarah B., Kyle A. Swiston, and Daniel J. Mennill, Seasonal variation in acoustic signals of Pileated Woodpeckers, 499-504 Troglodytes aedon, 99-104, 216 troglodytes, 155, 287 Trogon, Amazon White-tailed, see Trogon viridis Bare-cheeked, see Apaloderma aequatoriale Black-headed, see Trogon melanocephalus Choco, see Trogon coniptus Citreoline, see Trogon citreolus Hispaniolan, see Priotelus roseigster Narina, see Apaloderma narina Slaty-tailed, see Trogon massena Violaceous, see Trogon violaceus Trogon citreolus, 249 coniptus, 249 massena, 249 melanocephalus, 248-255 violaceus, 249 viridis, 249 trout, brook, see Salvelinus fontinalis Trudell, Heidi, see Hager, Stephen B., , Kelly J. McKay, Stephanie M. Crandall, and Lance Mayer Tupinambis merianae, 229 Turdus assimilis, 62 fuscater, 28 grayi, 28, 31, 32, 214-216 merula, 209 migratorius, 146-152, 209-213, 297, 525-530 nigrescens, 214-216 philomelos, 350 plumbeus, 464 rufiventris, 27 Turkey, Ocellated, see Meleagris ocellata Wild, see Meleagris gallopavo Twistwing, Brownish, see Cnipodectes subbruneus Rufous, see Cnipodectes superrufus Tympanuchus cupido, 129 Tyrannus dominicensis, 655-657 tyrannus, 287 Tyto soumagnei, 891-897 u Uria aalge, 635-636 Urocissa erythrorhyncha, 96 Urosphena squameiceps, 269, 272 Uyehara, Kimberly J., Andrew Engilis Jr., and Bruce D. Dugger, Wetland features that influence oc- cupancy by the endangered Hawaiian Duck, 311-319 V Valdes-Pena, Rene A., Sonia Gabriela Ortiz-Maciel, Simon O. Valdez Juarez, Ernesto C. Enkerlin Hoeflich, and Noel F. R. Snyder, Use of clay licks by Maroon-fronted Parrots (Rhynchopsitta terrisi) in northern Mexico, 176-180 Vanderhoff, E. Natasha and Perri K. Eason, Compari- sons between juvenile and adult American Robins foraging for mulberry fruit, 209-213 VanderWerf, Eric A. and Pauline K. Roberts, Foraging and nesting of the Akikiki or Kauai Creeper (Oreomystis bairdi), 195—199 Veery, see Cathariis fusee scens Veluz, Maria J. S., see Hamao, Shoji, , Takema Saitoh, and Isao Nishiumi Vermivora chrysoptera, 366-370, 801-812 peregrina, 278 pinus, 366-370 Vilella, Francisco J., see Wood, Douglas R., , and L. Wesley Burger Jr. Violetear, Green, see Colibri thalassinus Sparkling, see Colibri coriiscans White-vented, see Colibri serrirostris Vireo altiloquiis, 460 atricapilla, 407—409 gilvLis, 297 griseus, 286-295 huttoni, 478 latirneri, 460-466 olivaceus, 146-152, 157, 277-285, 286-295, 897- 900 solitarius, 155, 287 Vireo, Black-capped see Vireo atricapilla Black-whiskered, see Vireo altiloquus Blue-headed, see Vireo solitarius Hutton’s, see Vireo huttoni Red-eyed, see Vireo olivaceus Warbling, see Vireo gilvus White-eyed, see Vireo griseus vocalization begging behavior of Cacomantis sepulcralis, 887- 890 communal calling and prospecting by Trogon me- lanocephalus, 248-255 difference between Cettia seebohmi and Cettia di- phone, 268—276 male song variation of Colibri thalassinus in the Talmanca Mountain Range, Costa Rica, 519-524 of Mimus polyglottos, 717—724 seasonal variation in Dryocopus pileatus, 499-504 song variation in Empidonax fulvifrons, 256—267 type B song of Dendroica dominica, 401-403 Volatinia jacarina, 31, 32, 398 von Wehrden, Henrik, The Giant Hummingbird (Pata- gona gigas) in the mountains of central Argen- INDEX TO VOLUME 120, 2008 961 tina and a climatic envelope model for its dis- tribution, 648-651 Vukasovic, M. Angelica, see Estades, Cristian, , and Jorge A. Tomasevic Vulpes vulpes, 88 w Waggerman, Gary L., see Breeden, Jeffrey B., Fidel Hernandez, Ralph L. Bingham, Nova J. Silvy, and Wagtail, Eastern Yellow, see Motacilla tschutschensis Western Yellow, see Motacilla flava White, see Motacilla alba Walters, Eric L., see Stromberg, Mark R., Walter D. Koenig, , and John Schweisinger Wang, Ning, Yanyun Zhang, and Guangmei Zheng, Breeding ecology of the Narcissus Flycatcher in north China, 92-98 Wang, Qiang and Feng Li, Nest raising by Red- crowned Cranes in response to human-mediat- ed flooding at Zhalong Nature Reserve, China, 610-613 Wanless, Ross M. and Philip A. R. Hockey, Natural history and behavior of the Aldabra Rail {Dry- olimnas [cuvieri} aldabranus), 50-61 Warbler, Arctic, see Phylloscopus borealis Black-throated Blue, see Dendroica caendescens Black-throated Green, see Dendroica virens Blue-winged, see Vermivora pinus Canada, see Wilsonia canadensis Cape May, see Dendroica tigrina Cerulean, see Dendroica cerulea Cetti’s, see Cettia cetti Chestnut-sided, see Dendroica pensylvanica Golden-winged, see Vermivora chrysoptera Great Reed, see Acrocephalus arundinaceus Green-tailed, see Microligea palustris Hooded, see Wilsonia citrina Japanese Bush, see Cettia diphone Kentucky, see Oporornis forrnosus MacGillivray’s, see Oporornis tobniei Magnolia, see Dendroica magnolia Oriental Reed, see Acrocephalus arientalis Philippine Bush, see Cettia seebohmi Pine, see Dendroica pinus Prairie, see Dendroica discolor Prothonotary, see Protonotaria citrea Swainson’s, see Limnothlypis swainsonii Tennessee, see Vermivora peregrina Willow, see Phylloscopus trochilus Wilson’s, see Wilsonia pusilla Worm-eating, see Helmitheros vermivorum \vermi- vorus] Yellow, see Dendroica petechia Yellow-rumped, see Dendroica coronata Yellow-throated, see Dendroica dominica Waterthrush, Northern, see Seiurus noveboracensis Watts, Bryan D., Michael D. Wilson, Fletcher M. Smith, Barton J. Paxton, and J. Bill Williams, Breeding range extension of the Coastal Plain Swamp Sparrow, 393-395 Watts, Bryan D., see Wilson, Michael D. and Waxwing, Cedar, see Bombycilla cedrorum Weakley Jr., Leonard, see Oehler, David A., Steve Pel- ikan, W Roger Fry, , Alejandro Kusch, and Manuel Marin weasel, see Mustela spp. Siberian, see Mustela gularis Wege, Michael L., see Nicolai, Christopher A., James S. Sedinger, and Wells, Kimberly M. Suedkamp, Joshua J. Millspaugh, Mark R. Ryan, and Michael W. Hubbard, Fac- tors affecting home range size and movements of post-fledging grassland birds, 120-130 Wetzel, Daniel P. and C. Ray Chandler, Adoption; ad- aptation or reproductive error in Eastern Blue- birds, 419-422 Wheatear, Northern, see Oenanthe oenanthe Wheeler, Brian K„ Raptors of eastern [western] North America, reviewed, 430-431 White, Jennifer D. and John Faaborg, Post-fledging movement and spatial habitat-use patterns of juvenile Swainson’s Thrushes, 62-73 Whip-poor-will, see Caprimulgus vociferus Whitaker, S. J., see Fair, J. M. and White-eye, Black-crowned, see Zosterops atrifrons Everett’s, see Zosterops everetti Lemon-bellied, see Zosterops chloris Mountain, see Zosterops montanus Pale-bellied, see Zosterops consobrinorum Rota, see Zosterops rotensis Sanghie, see Zosterops nehrkorni Seram, see Zosterops stalkeri Togian, see Zosterops sornadikartai Whitethroat, Common, see Sylvia communis Whittingham, Linda A., see Moyle, Robert G., Beth Slikas, , David W. Winkler, and Fred- erick H. Sheldon Wilson, Amy G., see Wilson, Scott, Keith A. Hobson, Douglas M. Collister, and Wilson, J. Bill, see Watts, Bryan D., Michael D. Wil- son, Fletcher M. Smith, Barton J. Paxton, and Wilson, Michael D. and Bryan D. Watts, Landscape configuration effects on distribution and abun- dance of Whip-poor-wills, 778-783 Wilson, Michael D., see Watts, Bryan D., , Fletcher M. Smith, Barton J. Paxton, and J. Bill Williams Wilson, Robert E. and Kevin G. McCracken. Specimen shrinkage in Cinnamon Teal, 390-392 Wilson, Scott, Keith A. Hobson. Douglas M. Collister, and Amy G. Wilson. Spring migratory stopover of Swainson’s Thrush along the Pacific coast of southwest Costa Rica. 74-84 Wilson, William J., Use of legs and feet for control by scoters during aerial courtship. 594-599 Wilsonia canadensis, 1 55 citrina. 62, 277-285. 291. 342. 370 pusilla. 288 Winker, Kevin and Gary R. Graves, Genetic structure 962 THE WILSON JOURNAL OL ORNITHOLOGY • Vol. 120, No. 4, December 2008 of breeding and wintering populations of Swainson’s Warbler, 433-445 Winker, Kevin, see Johnson, Andrew B. and Winkler, David W., see Moyle, Robert G., Beth Slikas, Linda A. Whittingham, , and Lrederick H. Sheldon Wood, Douglas R., Lrancisco J. Vilella, and L. Wesley Burger Jr., Red-cockaded Woodpecker home range use and the macrohabitat selection in a loblolly-shortleaf pine forest, 793-800 Wood, Petra Bohall, see Ammer, Prank K., , and Roger J. McPherson Wood-Pewee, Eastern, see Contopus virens Western, see Contopus sordidulus Woodpecker, Acorn, see Melanerpes formicivoriis Arizona, see Picoides arizonae Cuban Green, see Xiphidiopicus percussus Downy, see Picoides pubescens Gray-breasted, see Melanerpes hypopolius Great Spotted, see Dendrocopus [Picoides] major Hairy, see Picoides villosus Hoffmann’s, see Melanerpes hojfmanni Ladder-backed, see Picoides scalaris Lesser Spotted, see Dendrocopos minor Lewis’s, see Melanerpes lewis Middle Spotted, see Dendrocopos medius Pileated, see Dryocopus pileatus Plain-tailed, see Thryothorus euophrys Red-bellied, see Melanerpes carolinus Red-cockaded, see Picoides borealis Red-headed, see Melanerpes erythrocephalus Syrian, see Dendrocopus syriacus White-backed, see Dendrocopus leucotos White-headed, see Picoides albolarvatus Yellow-fronted, see Melanerpes flavifrons Woods, Christopher P. and R. Mark Brigham, Common Poorwill activity and calling behavior in rela- tion to moonlight and predation, 505-5 1 2 Woolaver, Lance G., see Rimmer, Christopher C., , Rina K. Nichols, Eladio M. Pernandez, Steven C. Latta, and Esteban Garrido Wren, Black-bellied, see Thryothorus fasciatoventris Buff-breasted, see Thryothorus leucotis Carolina, see Thryothorus ludovicianus House, see Troglodytes aedon Rufous-and-white, see Thryothorus rufalbus Winter, see Troglodytes troglodytes Wunker, Elizabeth H., see Cohen, Jonathan B., , and James D. Eraser X Xanthocephalus xanthocephalus, 342, 657-659 Xenodacnis parina, 531—544 Xenoligea montana, 190-195 Xiphidiopicus percussus, 215 Y Yahner, Richard H., Bird responses to a managed for- ested landscape, 897-900 Yahner, Richard H., see Kubel, Jacob E. and Yahner, Richard H., see Rohnke, Adam T. and Yanco, Scott W., see Linkhart, Brian D., Erin M. Ev- ers, Julie D. Megler, Eric C. Palm, Catherine M. Salipante, and Yaukey, Peter H. and Shawn C. Powell, Numbers of migratory birds stopping over in New Orleans, Louisiana, USA in relation to weather, 286-295 Yee, Julie L., see Ackerman, Joshua T, John Y. Ta- kekawa, Jill D. Bluso, , and Collin A. Eagles-Smith Yellowthroat, Common, see Geothlypis trichas Yom-Tov, Yoram, see Leader, Zohar, , and Uzi Motro York, Matthew W., review, 236-237 Yuehua, Sun, see Chenxi, Jia and z Zack, Steve, see Sperry, Jinelle H., T Luke George, and Zalik, Nathan J. and Noah G. Perlut, Simultaneous in- cubation by two females and nestling provi- sioning by four adults at a Savanna Sparrow nest, 628-630 Zenaida asiatica, 384-389 rnacroura, 149 Zhang, Yanyun, see Wang, Ning, , and Guang- mei Zheng Zheng, Guangmei, see Wang, Ning, Yanyun Zhang, and Zimmer, Kevin J., The White-Eyed Foliage-gleaner (Funariidae: Automolus) is two species, 10-25 Zimmerman, Guthrie S., Rick R. Horton, Daniel R. Dessecker, and R. J. Gutierrez, New insight to old hypotheses: Ruffed Grouse population cy- cles, 239-247 Zonotrichia capensis, 30, 3 1 leucophrys, 30 Zosterops anomalus, 1, 3, 7 atrifrons, 1-9 chloris, 1 consobrinorum, 1, 2 everetti, 1 lateralis, 3 1 montanus, 1 nehrkorni, 1, 3, 5 rotensis, 618—619 somakidartai, 1-9, (Frontispiece) stalkeri, 1 Wilson Journal of Ornithology Published by the Wilson Ornithological Society Volume 120 2008 Quarterly EDITOR; EDITORIAL BOARD: REVIEW EDITOR INDEX EDITOR EDITORIAL ASSISTANT CLAIT E. BRAUN RICHARD C. BANKS KATHY G. BEAL JACK CLINTON EITNIEAR SARA J. OYLER-McCANCE MARY GUSTAFSON KATHY G. BEAL NANCY J. K. BRAUN The Wilson Ornithological Society Founded 3 December 1888 Named after ALEXANDER WILSON, the first American Ornithologist President — James D. Rising, Department of Zoology, University of Toronto, Toronto, ON M5S 3G5, Canada; e-mail: rising@zoo.utoronto.ca First Vice-President — E. Dale Kennedy, Biology Department, Albion College, Albion, MI 49224, USA; e-mail: dkennedy@albion.edu Second Vice-President — Robert C. Beason, USDA, Wildlife Services, 6100 Columbus Avenue, Sandusky, OH 44870, USA; e-mail: beason@netzero.com Editor — Clait E. Braun, 5572 North Ventana Vista Rd., Tucson, AZ 85750, USA; e- mail: TWILSONJO@comcast.net Secretary — John A. Smallwood, Department of Biology and Molecular Biology, Mont- clair State University, Montclair, NJ 07043, USA; e-mail: smallwood@montclair. edu Treasurer — Melinda M. Clark, 52684 Highland Drive, South Bend, IN 46635, USA; e- mail: MClark@tcservices.biz Elected Council Members — Carla J. Dove, Greg H. Farley, and Mia R. Revels (terms expire 2009); Rebecca Holberton, Robert S. Mulvihill, and Timothy O’Connell (terms expire 2010); Jameson F. Chace, Sara R. Morris, and Margaret A. Voss (terms expire 2011). DATES OF ISSUE OF VOLUME 120 OF THE WILSON JOURNAL OF ORNITHOLOGY NO. 1 — 12 March 2008 NO. 2 — 28 May 2008 NO. 3 — 5 September 2008 NO. 4 — 3 1 December 2008 1 10 26 38 50 62 74 85 92 99 105 111 120 131 139 146 153 159 167 176 181 190 CONTENTS OF VOLUME 120 NUMBER 1 Major Articles A new White-eye {Zosterops) from the Togian Islands, Sulawesi, Indonesia Mochamad Indrawan, Pamela C. Rasmussen, and Sunarto The White-eyed Foliage-gleaner (Furnariidae: Automolus) is two species Kevin J. Zimmer Recent advances in the behavioral ecology of tropical birds Bridget J. M. Stutchbury and Eugene S. Morton Distribution, behavior, and conservation status of the Rufous Twistwing {Cnipodectes superrufus) Joseph A. Tobias, Daniel J. Lebbin, Alexandre Aleixo, Michael J. Andersen, Edson Guilherme, Peter A. Hosner, and Nathalie Seddon Natural history and behavior of the Aldabra Rail {Dryolimnas \cuvieri\ aldabranus) Ross M. Wanless and Philip A. R. Hockey Post-fledging movement and spatial habitat-use patterns of juvenile Swainson’s Thrushes Jennifer D. White and John Paaborg Spring migratory stopover of Swainson’s Thrush along the Pacific Coast of southwest Costa Rica Scott Wilson, Keith A. Hobson, Douglas M. Collister, and Amy G. Wilson Demography of Eastern Yellow Wagtails at Cape Romanzof, Alaska Heather M. Renner and Brian J. McCaffery Breeding ecology of the Narcissus Flycatcher in north China Ning Wang, Yanyun Zhang, and Guangmei Zheng Reproductive success of House Wrens in suburban and rural landscapes Michael J. Newhouse, Peter P Marra, and L. Scott Johnson Habitat selection and reproductive success of Cerulean Warblers in Indiana Kirk L. Roth and Kamal Elam Nesting biology of grassland birds at Fort Campbell, Kentucky and Tennessee James J. Giocomo, E. Daniel Moss, David A. Buehler, and William G. Minser Factors affecting home range size and movements of post-fledging grassland birds Kimberly M. Suedkamp Wells, Joshua J. Millspaugh, Mark R. Ryan, and Michael W. Hubbard Ecological factors affecting response of Dark-eyed Juncos to prescribed burning Jinelle H. Sperry, T. Luke George, and Steve Zack Winter habitat use by Boreal Chickadee flocks in a managed forest Adam Hadley and Andre Desrochers Long-term effects of wastewater irrigation on habitat and a bird community in central Pennsylvania Adam T. Rohnke and Richard H. Yahner Long-term trends in breeding birds in an old-growth Adirondack forest and the surrounding region Stacy A. McNulty, Sam Droege, and Raymond D. Masters Timing and location of mortality of fledgling, subadult, and adult California Gulls Bruce H. Pugesek and Kenneth L. Diem Effects of predation and food provisioning on Black Tern chick survival Shane R. Heath and Erederick A. Servello Use of clay licks by Maroon-fronted Parrots {Rhynchopsitta terrisi) in northern Mexico Rene A. Valdes-Pena, Sonia Gabriela Ortiz-Maciel, Simon O. Valdez Juarez, Ernesto C. Enkerlin Hoeflich, and Noel E R. Snyder Cavity number and use by other species as correlates of group size in Red-cockaded Woodpeckers John J. Kappes Jr. Short Communications First description of nests and eggs of two Hispaniolan endemic species: Western Chat-tanager {Calyptophilus tertius) and Hispaniolan Highland-tanager {Xenoligea montana) Christopher C. Rimmer, Lance G. Woolaver, Rina K. Nichols, ELidio M. Eerndndez, Steven C. Latta, and Esteban Garrido 195 200 201 205 209 214 217 221 226 228 231 237 239 248 256 268 277 286 296 304 311 320 331 Foraging and nesting of the ‘Akikiki or Kaua‘i Creeper {Oreomystis bairdi) Eric A. VanderWerf and Pauline K. Roberts First observation of duetting in the Olive-backed Euphonia {Euphonia gouldi) Thor Hanson The display of a Reddish Hermit {Phaethornis ruber) in a lowland rainforest, Bolivia Adam Eelton, Annika M. Pelton, and David B. Lindenmayer Home range and habitat preferences of the Banded Ground-cuckoo {Neomorphus radiolosus) Jordan Karubian and Luis Carrasco Comparisons between juvenile and adult American Robins foraging for mulberry fruit E. Natasha Vanderhojf and Perri K Eason Previously unknown food items in the diet of six neotropical bird species Luis Sandoval, Esteban Biamonte, and Alejandro Solano-Ugalde Anvil use by the Red-cockaded Woodpecker Kristin J. Bondo, Lauren N. Gilson, and Reed Bowman Gender identification of Grasshopper Sparrows comparing behavioral, morphological, and molecular techniques Erank K. Ammer, Petra Bohall Wood, and Roger J. McPherson Abnormal eggs of Rio Grande Wild Turkeys on the Edwards Plateau, Texas Kyle B. Melton, Justin Z. Dreibelbis, Ray Aguirre, Bret A. Collier, T. Wayne Schwertner, Markus J. Peterson, and Nova J. Silvy Barred Forest Falcon {Micrastur ruficollis) predation on relatively large prey Edbio Rohe and Andre Pinassi Antunes Ornithological Literature Compiled by Mary Gustafson Editor’s Comments NUMBER 2 Major Articles New insight to old hypotheses: Ruffed Grouse population cycles Guthrie S. Zimmerman, Rick R. Horton, Daniel R. Dessecker, and R. J. Gutierrez Communal calling and prospecting by Black-headed Trogons ( Trogon melanocephalus) Christina Riehl Song variation in Buff-breasted Flycatchers {Empidonax JulviJrons) M. Ross Lein Phylogenetic relationship and song differences between closely related Bush Warblers {Cettia seebohmi and C. diphone) Shoji Hamao, Maria J. S. Veluz, Takema Saitoh, and Lsao Nishiumi Autumn stopover near the Gulf of Honduras by Nearctic-Neotropic migrants Andrew B. Johnson and Kevin Winker Numbers of migratory birds stopping over in New Orleans, Louisiana, USA in relation to weather Peter H. Yaukey and Shawn C. Powell Mass changes of migratory landbirds during stopovers in a New York City park Chad L. Seewagen and Eric J. Slayton Migration of Florida sub-adult Bald Eagles Elizabeth K Mojica, J. Michael Meyers, Brian A. Millsap, and Katherin L. Haley Wetland features that influence occupancy by the endangered Hawaiian Duck Kimberly J. Uyehara, Andrew Engilis Jr., and Bruce D. Dugger Habitat features associated with Barrow’s Goldeneye breeding in eastern Canada Michel Robert, Bruno Drolet, and Jean-Pierre L. Savard Distribution, abundance, and nest-site characteristics of Black Swifts in the southern Rocky Mountains of Colorado and New Mexico Richard G. Levad, Kim M. Potter, Christopher W Shultz, Carolyn Gunn, and Joseph G. Doerr 339 345 353 366 371 378 384 390 393 395 398 401 404 407 409 413 416 419 422 425 433 446 455 Nest reuse by Vermilion Flycatchers in Texas Kevin S. Ellison Natural history and breeding biology of the Rusty-breasted Antpitta {Grallaricula ferrugineipectus) Alina M. Niklison, Juan I. Areta, Roman A. Ruggera, Karie L. Decker, Carlos Bosque, and Thomas E. Martin Foraging ecology of parrots in a modihed landscape: seasonal trends and introduced species Greg D. Matuzak, M. Bernadette Bezy, and Donald ]. Brightsmith The signal function of a melanin-based plumage ornament in Golden-winged Warblers Emily Anne McKinnon and Raleigh ]. Robertson Factors influencing fidelity of House Finches to a feeding station Andrew K Davis Gender identification of Gaspian Terns using external morphology and discriminant function analysis Joshua T Ackerman, John Y. Takekawa, Jill D. Bluso, Julie L. Yee, and Collin A. Eagles-Smith Effects of traffic noise on auditory surveys of urban White-winged Doves Jeffrey B. Breeden, Eidel Hernandez, Ralph L. Bingham, Nova J. Silvy, and Gary L. Waggerman Short Communications Specimen shrinkage in Cinnamon Teal Robert E. Wilson and Kevin G. McCracken Breeding range extension of the Coastal Plain Swamp Sparrow Bryan D. Watts, Michael D. Wilson, Eletcher M. Smith, Barton J. Paxton, and J. Bill Williams Polyandry and sex ratio in the Song Sparrow Michael H. Janssen, Peter Arcese, Mark S. Sloan, and Kelly J. Jewell Novel courtship behavior in the Little Greenbul {Andropadus virens) Alexander N. G. Kirschel A recording of a Type B song of the Yellow- throated Warbler Bailey D. McKay Substrate and vegetation selection by nesting Piping Plovers Jonathan B. Cohen, Elizabeth H. Wunker, and James D. Eraser Nest site selection by a male Black-capped Vireo Andrew J. Campomizzi, Shannon E Earrell, and Jerrod A. Butcher Nests of Black-throated Green Warblers in tree cavities Douglas C. Tozer Nests and fledglings of the Red-ruffed Fruitcrow {Pyroderus scutatus) Mercival R. Erancisco, Paulo R. R. Oliveira Jr., and Vitor O. Lunardi Palila {Loxioides bailleui) fledgling fed by Hawai'i ‘Amakihi {Hemignathus virens) Chris Parmer, Bridget A. Prederick, Paul C. Banko, Robert M. Stephens, and Carter W Snow Adoption: adaptation or reproductive error in Eastern Bluebirds? Daniel P. Wetzel and C. Ray Chandler An intraspecific killing in adult Pacific Reef Egrets {Egretta sacra) Christa Beckmann Ornithological Literature Compiled by Mary Gustafson NUMFIHR 3 Major Articles Genetic structure of breeding and wintering populations of Swainson’s Warbler Kevin Winker and Cary R. Graves Does age influence territory size, habitat selection, and reproductive success of male Canada Warblers in central New Hampshire? Leonard R. Reitsma, Michael T. Hallworth, and Phred M. Benharn Solitary winter roosting of Ovenbirds in core foraging area David R. Brown and Ihomas W Sherry 460 467 473 478 494 499 505 513 519 525 531 545 550 565 569 575 582 594 599 603 606 610 613 Reproductive success of the Puerto Rican Vireo in a montane habitat Adrianne G. Tossas Nest defense by Carolina Wrens Kelly A. D’Orazio and Diane L. H. Neudorf Nests, eggs, and parental care of the Puna Tapaculo {Scytalopus simonsi) Peter A. Hosner and Noeml E. Huanca Phylogeographic patterns of differentiation in the Acorn Woodpecker MagaJi Honey-Escanddn, Blanca E. Hernandez- Banos, Adolfo G. Navarro-Siguenza, Hesiquio Benitez-Dlaz, and A. Townsend Peterson Diet of Acorn Woodpeckers at La Primavera Forest, Jalisco, Mexico Veronica Carolina Rosas-Espinoza, Elisa Maya-Elizarraras, Oscar Erancisco Reyna Bustos, and Erancisco Martin Huerta-Martinez Seasonal variation in acoustic signals of Pileated Woodpeckers Sarah B. Tremain, Kyle A. Swiston, and Daniel J. Mennill Common Poorwill activity and calling behavior in relation to moonlight and predation Christopher P. Woods and R. Mark Brigham Maximizing detection probability of wetland-dependent birds during point-count surveys in northwestern Florida Christopher P. Nadeau, Courtney J. Conway, Bradley S. Smith, and Thomas E. Lewis Male song variation of Green Violetear {Qolibri thalassinus) in the Talamanca Mountain Range, Costa Rica Gilbert Barrantes, Char Sanchez, Branko Hilje, and Rodolfo Jaffe Metabolizable energy in Chinese Tallow fruit for Yellow-rumped Warblers, Northern Cardinals, and American Robins Michael J. Baldwin, Wylie C. Barrow Jr., Clinton Jeske, and Erank C. Rohwer Foraging ecology of High Andean insectivorous birds in remnant Polylepis forest patches Huw Lloyd Nesting biolo^ of the Giant Conebill {Oreomanes fraseri) in the High Andes of Bolivia J. R. A. Cahill, E. Matthysen, and N. E. Huanca Bird density and mortality at windows Stephen B. Hagar, Heidi Trudell, Kelly ]. McKay, Stephanie M. Crandall, and Lance Mayer Ectoparasites affect hemoglobin and percentages of immature erythrocytes but not hematocrit in nestling Eastern Bluebirds Renee E. Carleton Observations on flocking behavior of Worthen’s Sparrow {Spizella wortheni) and occurrence in mixed-species flocks Julio C. Canales-Delgadillo, Laura M. Scott-Morales, Mauricio Cotera Correa, and Marisela Pando Moreno Status of Crested Penguin {Eudyptes spp.) populations on three islands in southern Chile David A. Oehler, Steve Pelikan, W. Roger Pry, Leonard Weakley Jr., Alejandro Kusch, and Manuel Marin Nest habitat selection of White-winged Scoters on Yukon Flats, Alaska David E. Safine and Mark S. Lindberg Short Communications Use of legs and feet for control by scoters during aerial courtship William J. Wilson Bill entanglement in subcutaneously-anchored radio transmitters on Harlequin Ducks Jeanine Bond and Daniel Esler Estimate of Trichomonas gallinae-mAxxcQd mortality in Band-tailed Pigeons, upper Carmel Valley, California, winter 2006-2007 Mark R. Stromberg, Walter D. Koenig, Eric L. Walters, and John Schweisinger Winter ecology of Yellow Rails based on South Carolina specimens William Post Nest raising by Red-crowned Cranes in response to human-mediated flooding at Zhalong Nature Reserve, China Qiang Wang and Peng Li White-winged Diuca Finch {Diuca speculifera) nesting on Quelccaya Ice Cap, Peru Douglas R. Hardy and Spencer P. Hardy 618 620 625 628 631 633 635 637 641 645 648 651 654 655 657 660 667 674 683 692 700 708 717 Nest success and nest predation of the endangered Rota White-eye {Zosterops rotensis) Lainie Berry and Estanislao Taisacan Predators at nests of the Western Slaty Antshrike ( Thamnophilus atrinucha) Corey E. Tarwater Evidence for Bachman’s Sparrow raising Brown-headed Cowbirds to fledging Matthew J. Reetz, Elizabeth Earley, and Thomas A. Contreras Simultaneous incubation by two females and nestling provisioning by four adults at a Savannah Sparrow nest Nathan J. Zalik and Noah G. Perlut Grey Heron {Ardea cinerea) predation on the Aldabra White-throated Rail {Dryolimnas cuvieri aldabranus) Pierre A. Pistorius Purple Swamphens {Porphyrio porphyrio) attempting to prey upon Black Swan {Cygnus atratus) eggs and preying upon a cygnet on an urban lake in Melbourne, Australia Shandiya Balasubramaniam and Patrick-Jean Guay Northern Fulmar predation of Common Murre Stephan Lorenz and Sampath Seneveratne Diet of nestling Black-crowned Night-herons in a mixed species colony: implications for Tern conservation C. Scott Liall and Stephen W. Kress Diet of the Long-eared Owl in the northern and central Negev Desert, Israel Zohar Leader, Yoram Yom- Tov, and Uzi Motro First observed instance of polygyny in Flammulated Owls Brian D. Linkhart, Erin M. Evers, Julie D. Megler, Eric C Palm, Catherine M. Salipante, and Scott W. Yanco The Giant Hummingbird {Patagona gigas) in the mountains of central Argentina and a climatic envelope model for its distribution LLenrik von Wehrden Giant Hummingbirds {Patagona gigas) ingest calcium-rich minerals Cristidn E Estades, M. Angelica Vukasovic, and Jorge A. Tomasevic Nocturnal foraging observations of the Blue-crowned Motmot {Momotus momota) in San Jose, Costa Rica Alejandro Solano- Ugalde and Agustina Arcos- Torres Kleptoparasitism by Grey Kingbirds ( Tyrannus dominicensis) in Barbados Sarah E. Overington, Laure Cauchard, and Kimberly-Ann Cote Double-scratching by Yellow-headed Blackbirds Alan de Queiroz Ornithological Literature Compiled by Mary Gustafson and Clait E. Braun NUMBER 4 Major Articles Differential ejection of cowbird eggs and non-mimetic eggs by grassland Passerines Dwight R. Klippenstine and Spencer G. Sealy Between and within clutch variation of egg size in Greater Rheas Gustavo J. Eerndndez and Juan C. Reboreda DNA sequence assessment of phylogenetic relationships among New World martins (Hirundinidae: Progne) Robert G. Moyle, Beth Slikas, Linda A. Whittingham, David W. Winkler, and Frederick El. Sheldon Preliminary analysis of the ecology and geography of the Asian nuthatches (Aves: Sittidae) Shady Menon, Zafar-ul Islam, Jorge Soberon, and A. Townsend Peterson Density and abundance of Mountain Plovers in northeastern Montana Theresa M. Childers and Stephen J. Dinsmore Land cover associations of nesting territories of three sympatric Buteos in shortgrass prairie Scott McConnell, Timothy J. O’Connell, and David M. Leslie Jr. Species recognition in a vocal mimic: repetition pattern not the only cue used by Northern Mockingbirds in discriminating songs of conspecifics and Brown d'hrashers Dustin G. Reichara and J. Jordan Price 725 732 743 755 767 778 784 793 801 813 820 830 840 856 862 867 871 874 879 884 887 891 897 Lesser Snow Geese and Ross’s Geese form mixed flocks during winter but differ in family maintenance and social status Jon Einar Jonsson and Alan D. AJton Nesting ecology of Gommon Goldeneyes and Hooded Mergansers in a boreal river system Hdene Senechal, Gilles Gauthier, and Jean-Pierre L. Savard Movements and habitat use by Red-breasted Merganser broods in eastern New Brunswick Shawn R. Craik and Rodger D. Titman Differences in growth of Black Brant goslings between a major breeding colony and outlying breeding aggregations Christopher A. Nicolai, James S. Sedinger, and Michael L. Wege Diet of the Yellow-knobbed Gurassow in the Gentral Venezuelan Llanos Carolina Bertsch and Guillermo R. Barreto Landscape configuration effects on distribution and abundance of Whip-poor-wills Michael D. Wilson and Bryan D. Watts Impact of Hurricane Wilma on migrating birds: the case of the Ghimney Swift Mark Dionne, Celine Maurice, Jean Gauthier, and Frangois Shajfer Red-cockaded Woodpecker home range use and macrohabitat selection in a loblolly-shortleaf pine forest Douglas R. Wood, Francisco J. Vilella, and L. Wesley Burger Jr. Quality of anthropogenic habitats for Golden-winged Warblers in central Pennsylvania Jacob E. Kubel and Richard H. Yahner Avian cell-mediated immune response to drought J. M. Fair and S. J. Whitaker Influence of grazing and available moisture on breeding densities of grassland birds in the central Platte River Valley, Nebraska Daniel H. Kim, Wesley E. Newton, Gary R. Lingle, and Felipe Chavez- Ramirez Response of songbirds to riparian willow habitat structure in the Greater Yellowstone Ecosystem Brian F M. Olechnowski and Diane M. Debinski Regional analysis of riparian bird species response to vegetation and local habitat features Nadav Nur, Grant Ballard, and GeoJJrey R. Geupel Short Communications First description of the breeding biology and natural history of the Ochre-breasted Brush Finch {Atlapetes semirufus) in Venezuela Luis Biancucci and Thomas E. Martin Reproductive biology of the Red-ruffed Fruitcrow {Pyroderus scutatus granadensis) James A. Muir, Diane Licata, and Thomas E. Martin Nests and nesting behavior of Golden Swallow ( Tachycineta euchrysea) in abandoned bauxite mines in the Dominican Republic Jason M. Townsend, Esteban Garrido, and Danilo A. Mejia Nest, nestling care, and breeding season of the Spangled Gotinga {Cotinga cay and) in French Guiana Johan Ingels Nests, eggs, and incubation behavior of the Grey-headed Bullfinch {Pyrrhula erythacd) Jia Chenxi and Sun Yuehua Parental care in Tawny-bellied {Sporophila hypoxantha) and Rusty-collared (5. collaris) Seedeaters Carolina Facchinetti, Alejandro G. Di Giacomo, and Juan C. Reboreda Postnatal growth rates of hummingbirds: review and new records Bernd P Freymann and Karl-Ludwig Schuchmann Begging behavior of fledgling Rusty-breasted Guckoo {Cacomantis sepulcralis) Tomas Grim Natural history of the Red Owl ( Tyto soumagnei) in dry deciduous tropical forest in Madagascar Scott G. Cardiff and Steven M. Goodman Bird responses to a managed forested landscape Richard H. Yahner 901 Freeze-frame fruit selection by birds Mercedes S. Foster 906 Predation of Rio Grande Wild Turkey nests on the Edwards Plateau, Texas Justin Z. Dreibelbis, Kyle B. Melton, Ray Aguirre, Bret A. Collier, Jason Hardin, Nova J. Silvy, and Markus J. Peterson 910 No evidence for spring re-introduction of an arbovirus by Cliff Swallows Valerie A. O'Brien, Amy T. Moore, Kathryn P Huyvaert, and Charles R. Brown 914 Marbled Godwit collides with aircraft at 3,700 m Carla J. Dove and Court Goodroe 9 1 6 The 2008 William and Nancy Klamm Service Award 917 Ornithological Literature Compiled by Mary Gustafson and Robert B. Payne 927 Proceedings of the Eight-ninth Annual Meeting 941 Reviewers for Volume 120 943 Index to Volume 120 Contents of Volume 120 THE WILSON JOURNAL OF ORNITHOLOGY Editor CLAIT E. BRAUN 5572 North Ventana Vista Road Tucson, AZ 85750-7204 E-mail: TWILSONJO@comcast.net Editorial Board RICHARD C. BANKS KATHY G. BEAL JACK CLINTON EITNIEAR SARA J. OYLER-McCANCE Editorial Assistant NANCY J. K. BRAUN Review Editor MARY GUSTAFSON Texas Parks and Wildlife Department 2800 South Bentsen Palm Drive Mission, TX 78572, USA E-mail: WilsonBookReview@aolcom GUIDELINES EOR AUTHORS Please consult the detailed “Guidelines for Authors” found on the Wilson Ornithological Society web site (http://www.wilsonsociety.org). All manuscript submissions and revisions should be sent to Clait E. Braun, Editor, The Wilson Journal of Ornithology, 5572 North Ventana Vista Road, Tucson, AZ 85750-7204, USA. The Wilson Journal of Ornithology office and fax telephone number are (520) 529-0365. The e-mail address is TWilsonJO@comcast.net. NOTICE OF CHANGE OF ADDRESS Notify the Society immediately if your address changes. Send your complete new address to Ornithological Societies of North America, 5400 Bosque Boulevard, Suite 680, Waco, TX 76710. The permanent mailing address of the Wilson Ornithological Society is: %The Museum of Zoology, The University of Michigan, Ann Arbor, MI 48109. Persons having business with any of the officers may address them at their various addresses given on the inside of the front cover, and all matters pertaining to the journal should be sent directly to the Editor. MEMBERSHIP INQUIRIES Membership inquiries should be sent to Timothy J. O’Connell, Department of Natural Resource Ecology and Management, Oklahoma State University, 205 Life Sciences West, Stillwater, OK 74078; e-mail: oconnet@ okstate.edu THE JOSSELYN VAN TYNE MEMORIAL LIBRARY The Josselyn Van Tyne Memorial Library of the Wilson Ornithological Society, housed in the University of Michigan Museum of Zoology, was established in concurrence with the University of Michigan in 1930. Until 1 947 the Library was maintained entirely by gifts and bequests of books, reprints, and ornithological magazines from members and friends of the Society. Two members have generously established a fund for the purchase of new books; members and friends are invited to maintain the fund by regular contribution. The fund will be administered by the Library Committee. Robert Payne, University of Michigan, is Chairman of the Committee. The Library currently receives over 200 periodicals as gifts and in exchange for The Wilson Journal of Orni- thology. For information on the Library and our holdings, see the Society’s web page at http:// www.wilsonsociety.org. With the usual exception of rare books, any item in the Library may be borrowed by members of the Society and will be sent prepaid (by the University of Michigan) to any address in the United States, its possessions, or Canada. Return postage is paid by the borrower. Inquiries and requests by borrowers, as well as gifts of books, pamphlets, reprints, and magazines, should be addressed to: Josselyn Van Tyne Memorial Library, Museum of Zoology, The University of Michigan, 1109 Geddes Avenue, Ann Arbor, MI 48109-1079, USA. Contributions to the New Book Fund should be sent to the Treasurer. This issue of The Wilson Journal of Ornithology was published on 31 December 2008. Continued from outside back cover 820 Influence of grazing and available moisture on breeding densities of grassland birds in the central Platte River Valley, Nebraska Daniel H. Kim, Wesley E. Newton, Gary R. Lingle, and Felipe Chavez-Ramirez 830 Response of songbirds to riparian willow habitat structure in the Greater Yellowstone Ecosystem Brian E M. Olechnowski and Diane M. Debinski 840 Regional analysis of riparian bird species response to vegetation and local habitat features Nadav Nur, Grant Ballard, and Geoffrey R. Geupel Short Communications 856 First description of the breeding biology and natural history of the Ochre-breasted Brush Finch {Atlapetes semirufus) in Venezuela Luis Biancucci and Thomas E. Martin . ^ 862 Reproductive biology of the Red-ruffed Fruitcrow {Pyroderus scutatus granadensis) James A. Muir, Diane Licata, and Thomas E. Martin 867 Nests and nesting behavior of Golden Swallow {Tachycineta euchrysea) in abandoned bauxite mines in the Dominican Republic Jason M. Townsend, Esteban Garrido, and Danilo A. Mejia 871 Nest, nestling care, and breeding season of the Spangled Cotinga {Cotinga cayana) in French Guiana Johan Ingels 874 Nests, eggs, and incubation behavior of the Grey-headed Bullfinch {Pyrrhula erythaca) Jia Chenxi and Sun Yuehua ^ 879 Parental care in Tawny-bellied {Sporophila hypoxantha) and Rusty-collared (5. collaris) Seedeaters Garolina Facchinetti, Alejandro G. Di Giacomo, and Juan C. Reboreda 884 Postnatal growth rates of hummingbirds: review and new records Bernd P Freymann and Karl-Ludwig Schuchmann 887 Begging behavior of fledgling Rusty-breasted Guckoo {Cacomantis sepulcralis) Tomas Grim 891 Natural history of the Red Owl {Tyto soumagnei) in dry deciduous tropical forest in Madagascar Scott G. Gardijfand Steven M. Goodman 897 Bird responses to a managed forested landscape Richard H. Yahner 90 1 Freeze-frame fruit selection by birds Mercedes S. Foster 906 Predation of Rio Grande Wild Turkey nests on the Edwards Plateau, Texas Justin Z. Dreibelbis, Kyle B. Melton, Ray Aguirre, Bret A. Collier, Jason Hardin, Nova J. Silvy, and Markus J. Peterson 910 No evidence for spring re-introduction of an arbovirus by Gliff Swallows Valerie A. OBrien, Amy T Moore, Kathryn P. Huyvaert, and Charles R. Brown 914 Marbled Godwit collides with aircraft at 3,700 m Carla J. Dove and Court Goodroe 916 The 2008 William and Nancy Klamm Service Award 917 Ornithological Literature Compiled by Mary Gustafson and Robert B. Payne 927 Proceedings of the Eighty-ninth Annual Meeting • ' f I | | 941 Reviewers FOR Volume 120 943 Index to Volume 120 Contents of Volume 120 The Wilson Journal of Ornithology (formerly The Wilson Bulletin) Volume 120, Number 4 CONTENTS December 2008 Major Articles 667 Differential ejection of cowbird eggs and non-mimetic eggs by grassland Passerines Dwight R. Klippenstine and Spencer G. Sealy 674 Between and within clutch variation of egg size in Greater Rheas Gustavo J. Fernandez and Juan C. Reboreda 683 DNA sequence assessment of phylogenetic relationships among New World martins (Hirundinidae: Progne) Robert G. Moyle, Beth Slikas, Linda A. Whittingham, David W. Winkler, and Frederick H. Sheldon 692 Preliminary analysis of the ecology and geography of the Asian nuthatches (Aves: Sittidae) Shady Menon, Zafar-ul Islam, Jorge Soberon, and A. Townsend Peterson 700 Density and abundance of Mountain Plovers in northeastern Montana Theresa M. Childers and Stephen J. Dinsmore 708 Land cover associations of nesting territories of three sympatric Buteos in shortgrass prairie Scott McConnell, Timothy J. O'Connell, and David M. Leslie Jr. 717 Species recognition in a vocal mimic: repetition pattern not the only cue used by Northern Mockingbirds in discriminating songs of conspecifics and Brown Thrashers Dustin G. Reichard and J. Jordan Price 1Tb Lesser Snow Geese and Ross’s Geese form mixed flocks during winter but differ in family maintenance and social status Jon Einar Jonsson and Alan D. AJion 732 Nesting ecology of Common Goldeneyes and Hooded Mergansers in a boreal river system Hdene Senechal, Gilles Gauthier, and Jean-Pierre L. Savard 743 Movements and habitat use by Red-breasted Merganser broods in eastern New Brunswick Shawn R. Craik and Rodger D. Titman Ibb Differences in growth of Black Brant goslings between a major breeding colony and outlying breeding aggregations Christopher A. Nicolai, James S. Sedinger, and Michael L. Wege 767 Diet of the Yellow-knobbed Curassow in the Central Venezuelan Llanos Carolina Bertsch and Guillermo R. Barreto 778 Landscape configuration effects on distribution and abundance of Whip-poor-wills Michael D. Wilson and Bryan D. Watts 784 Impact of Hurricane Wilma on migrating birds: the case of the Chimney Swift Mark Dionne, Cdine Maurice, Jean Gauthier, and Francois Shajfer 793 Red-cockaded Woodpecker home range use and macrohabitat selection in a loblolly-shortleaf pine forest Douglas R. Wood, Francisco J. Vilella, and L. Wesley Burger Jr. 801 Quality of anthropogenic habitats for Golden-winged Warblers in central Pennsylvania Jacob E. Kubel and Richard H. Yahner 813 Avian cell-mediated immune response to drought J. M. Fair and S. J. Whitaker Continued on inside back cover