j. Field Ornithol., 63(4):436-444

TERRITORIALITY, USE AND ECOLOGICAL DISTINCTNESS OF AN ENDANGERED PACIFIC ISLAND REED-WARBLER

ROBERTJ. CRAIG• Department of Natural ResourceManagement and Engineering, U-87 Universityof Connecticut Storrs, Connecticut 06269-d087 USA and Divisionof Fish and Wildlife Commonwealth No. ,MP 96950 USA Abstract.--NightingaleReed-Warblers ( luscinia) were studiedon Saipan,Mar- iana Islands from 1988 to 1991 to assessthe territoriality, habitat use and morphologyof this endangered,endemic, island .Data were comparedto thosefor closelyrelated mainland forms to gain insightsinto differencesevolved in an insular environment.Peak territorial activitywas reachedin January-February, the dry season.Males exhibitedhigh sitefidelity, but femaleshad low mate fidelity betweenyears. inhabitedmostly upland thicket-meadowvegetation, but they also used reed marshes,wetland thickets,and forest openings.Territory size in thicket-meadowcomplexes averaged 9338 _ 3433 (SD) m2. Males were larger than females;mass, wing, tail, and tarsuslength reliably distinguished sexes.Compared to the mainland A. orientalisand A. arundinaceus,A. lusciniahad (1) polygyny largely or entirely absent,(2) much larger territories, (3) upland rather than marshesas the principal breedinghabitat and (4) body size, particularly bill size, greatly increased. The shift to upland habitatsmay accountfor the differencein socialbehavior and territory size. Greater size is typical of island forms that may confrontcompetitively impoverished or food-limited environments.

TERRITORIALISMO, USO DE HABITAT Y SlNGULARIDADES ECOL•GICAS DEL AMENAZADO ACROCEPHAL US L USCINIA Sinopsis.--De 1988 a 1991 se estudi6en Saipan (Islas Marianas) el territorialismo,uso de habitat y morfologladel end6micoy amenazadocon desaparecerAcrocephalus luscinia. Se compararon los datos tomadoscon los de especiescontinentales similares, para hacer notar diferenciasevolutivas que se dan en ambientesinsulates. E1 pico de la conductaterritorial se alcanz6en los mesesde enero y febrero que correspondena la temporadaseca. Los machosexhibieron gran fidelidada localidades,pero las hembrasmostrar6n entre aftospoca fidelidad hacia sus parejas. Las aves habitan mayormente en matorrales de alturas, pero ademfisutilizan anegadosde junquillosy matorralesde pantanosy aperturasen bosques. E1 tamafiodel territorio en complejode matorralespromedi6 9338 _+ 3433 (DE) m2. Los machosresultaron de mayor tamafio que las hembras;el peso,largo del ala, cola y tarsoes confiablepara distinguir entre los sexos.Comparado con A. orientalisy A. arundinaceus,A. lusciniatiene: (1) territorios de mayor tamafio, (2) areas altas o montafias como habitat principal de anidamientoen vez de anegadoscostaneros, (3) tamafio, y (4) poliginia vir- tualmenteo totalmenteausente. E1 cambioa habitatsde alturas puededeberse a diferencia en la conductasocial y tamafiodel territorio. E1 mayortamafio es tlpicode especiesde islas que puedanconfrontar empobrecimiento competitivo en ambienteslimitados en alimentos. Reed-warblers(Sylviinae: Acrocephalus) are widespreadfrom through Australasia, where many speciesinhabit marshesand are po- lygynous(review in Urano 1985). Members of the genusare strongisland • Currentaddress: Northern Marianas College, Box 1250,Commonwealth No. Mariana Islands, Saipan,MP 96950 USA.

436 Vol.63, No. 4 Reed-Warbler Ecology [437 colonizers,presently represented in the tropical Pacificby as many as 10 species(Pratt et al. 1987). The relativelylarge, long-billedNightingale Reed-Warbler(A. luscirtia)of the Mariana Islands,Micronesia, is among the most distinctive of these. Baker (1951) believed it to be descended from the EasternGreat Reed-Warbler(A. orientalis),a speciesthat has populationsin the Bonin Islands, about 750 km N of the 812-km long Mariana chain (Baker 1951). These two species,along with the Euro- pean-African A. arurtdirtaceus,the AustralasianA. sterttoreusand all other Pacificisland Acrocephalus form a superspeciesthat occursover much of the Old World (Mayr et al. 1986). Three subspeciesof Nightingale Reed-Warbler are currently recog- nized(Mayr et al. 1986). A. l. luscirtia,originally found on Guam, Saipan and Alamagen,has been extinct on Guam sincethe late 1960s(Engbring et al. 1986). A. l. rtijoi of Aguijan, thoughtto be extinct,was rediscovered in May 1992, and A. I. yamashirtaeof becameextinct ca. 1981 (Glass 1987). Steadman's(1989) findingsof numerous,human-related prehistoricextinctions of Polynesianbirds suggestthat the speciesmay have once inhabited additional islands as well. On Saipan,Nightingale Reed-Warblers are presentlywidespread, with populationsestimated at 4867 in 1982 (Engbringet al. 1986). To provide basicdata for future conservationand managementof this largely un- known, endangeredspecies, I documentthe social behavior, habitat use and morphologyof the Saipanpopulation. I alsocompare my data with thosefor relatedmainland formsto gain insightsinto differencesevolved in an insular environment.

STUDY AREAS AND METHODS I studied Nightingale Reed-Warblers on Saipan in 1988 and 1989 duringJanuary-February, dry seasonmonths that correspondto the time of peak singingand presumablypeak breeding(Glass 1987). Additional observationswere made from November 1990 to September1991. I ob- servedbirds throughoutthe island (Fig. 1) but primarily in the Marpi region. Detailed investigationsof habitat use and territoriality were conducted on about 26 ha in the flat lowlands (about 10-40 m elevation) below the cliffs in the Marpi regionand on the plateau (about 130-200 m elevation) near the top of thesecliffs. These flat areasare vegetatedby tangantangan (Leucaenaleucocephala) thickets and elephant grass(Permiseturn purpu- reum) meadows(Fig. 2). The introducedtangantangan, aerially seeded after World War II to prevent erosion,often occupiesland formerly cultivatedfor sugarcane(Fosberg 1960). I studiedterritoriality by mist-netting birds in ATX (36 mm mesh) 12-m netsand colorbanding them. Birds were attractedto netsby playing tape recordedconspecific songs. Once banded,I confirmedthe identity of territory holdersvisually with binoculars,and recordedtheir movements on field maps. Territory boundarieswere determinedwith a variant of the minimum polygonmethod (Hill and Lein 1989) in which a curved 438] R.J. Craig j. FieldOrnithol. Autumn 1992

I I p146000 Pagan'

oAlamagan I oGuguan I

'•. • O ßG ._(2 16ø00

•Salpan •Tinlan Aguijan

•fi• ,14o00 ' I •Guam

Mariana Islands I -N- Saipan I •.1 I I I

FIGURE1. Principalstudy sites on Saipan,and the geographicrelationship of Saipanto the neighboringMariana Islands,Micronesia. Adapted in part fromJenkins (1983). perimeterwas fitted aroundoutermost perches used by singingor in- traspecificallyaggressive males undisturbed by playbackof tapedcalls. I mapped seven1988 territorieson 1:2000 base maps made from 1:10,000 color aerial photographsusing a Bauschand Lomb Zoom- transferscope. Precise map scalewas determinedfrom taped measure- ments made in the field between distinct landmarks. Field data were transferredto thesebase maps, and territory area was measuredwith a compensatingpolar planimeter. Data on habitat structure from these same seven territories were ob- tainedby dividingthem into eight "blocks"of approximatelyequal size and placinga 30-m transectin eachblock (stratified random sample). On each transect,habitat type (elephantgrass meadow, tangantangan thicket,intergradation between these two types)and plant speciescom- positionwere recordedat 1-m intervals(240 total samplepoints/terri- tory).At 3-m intervals(80 points/territory)I recordedvegetation height in the followingclasses: <2, >2-3, >3-6 and >6 m. Vol.63, No. 4 Reed-WarblerEcology [439

- -...... ' "'"h" $•'- ' '"':

ß ,- ; •'t' $'.. "•J• 2, '•" ' '•- • '•,• -•,7.•.••-g'• -,• •'•-',2 •lr• I" •'•,•. • •,

;• • ' •'• ' r'' a• '-' ' .... • /' •-•" I'•' '- '$ -' - . ;: t.- • , ,'"•Fx,,' ,' 'a, • •.•..x ,.' • ß

FIGURE2. NightingaleReed-Warbler habitat at Marpi, Saipan,sho•ng elephantgrass meadows(foreground), tanganta•ga• thickets (midground), a•d the Marpi cliffsveg- etatedby native forest.

To assessthe morphologyof this population,I made the following measurementson mist-nettedindividuals: mass, wing chord,tail length, tarsuslength, and bill length(from the distal end of thenostrils). Breeding maleNightingale Reed-Warblers were distinguished by their cloacalpro- tuberances,whereas breeding females were sexed by their broodpatches. After colorbanding, sexes were verifiedby observingpairs on their ter- ritories. For comparison,I also made morphologicalmeasurements on museum skins of summeringA. orientalisfrom China.

RESULTS Socialbehavior and territoriality.--Male Nightingale Reed-Warblers defendedterritories by singing,and usedeither exposed treetops, interior thicketsor elephantgrass stems as songperches. Males respondedvig- orouslyto playbackof tape recordedsongs, and couldbe inducedto leave their territories in pursuit of taped songs.I never observedfemales to sing,but they usuallyaccompanied their matesduring aggressiveen- counters,such as during response to playback.On no occasiondid I observe any male on the 50 territoriesstudied over three peak breedingseasons (28 individual males) to associatewith more than one female. Of 11 malesbanded in 1988,9 (82%)were present on the same territory in 1989. One movedto a new territory 1900 m from its originalone, and one disappeared.Hence, all malesbut one bandedin 1988 (91%) were relocated.In 1991,I located9 of 16 (56%)males banded by 1989.Taking 440] R.J. Craig j. FieldOrnithol. Autumn 1992 into accountthe missedJanuary-February 1990 season,average annual male turnoverfor 3 yr was 18%. This turnoverwas not necessarilythe same as male mortality, becausemales might have left the study area. Six of the 9 banded males presentin 1991 were on the territory in which I originally bandedthem. One movedafter at least 2 yr at one territory, and two moved to new territories after 1989. Of eight femalesbanded in 1988, three (38%) were relocatedin 1989, and two of thesewere mated to the samemale on the sameterritory. The third female, the mate of the only male to disappear,remained on its 1988 territory with a new mate. Of 10 femalesseen well during 1990- 1991, one was banded in 1988 and one was banded in 1989, both on territories other than thosethey initially occupied. In January-February 1988 the study area contained 14 territories, whereas in January-February 1989 there were 19 territories, a 36% increase.Territorial activity changedseasonally, such that in November- December1990 I couldlocate only 13 territories,but by January-Feb- ruary 1991 the number increasedto 17. In May 1991, all 17 territories were still active.By July, however,only 15 territorieswere active,and in SeptemberI found only 14 territories,indicating that breedingactivity was subsiding.Two territories active in 1989 but vacant in 1990-1991 were on land that had been partly clearedfor grazing following 1989. .--Most birds I foundon Saipaninhabited areas with a mosaic of tangantanganand elephant grass. Birds also inhabited native reed (Phragmiteskarka) marshesnear San Rocque, and Lake Su- supe. Contrary to Marshall's (1949) and H. D. Pratt's (pers. comm.) 1976 findings, however, the specieswas not abundant in the marshes,and birds locatedwere mostly around the marsh edges.One reed-warbler singingnear the shoreof Lake Susupeseemed principally associatedwith a nativepago (Hibiscus tiliaceus) thicket bordering a band of bulrushes(Scirpus littoralis) edging the lake. Birds did not inhabit the grassyupper slopesof Mt. Tapotchau (elev. 465 m). They were first encounteredat 400 m inhabitingnative forest edge adjacent to a vegetable farm, and at 350 m in a protectedvalley vegetatedby elephant grass, tangantanganthickets and Pandanusclumps. At Gualo Rai, I found reed- warblers in a vine-coveredvalley with bamboo(Bambusa vulgaris) and open woodsof predominantlynative tree species.At Talufofo, I found birds singing from open woodsof bamboo,pago, tangantangan,sosuge (Acacia confusa),elephant grass patches and vines. Elsewhere on the island, including at Navy Hill, Naftan Peninsula, Peninsula and Fadang, tangantanganand elephant grasswere the principal vege- tation in reed-warbler habitat. I found no territorial reed-warblers in interior native forestanywhere on the island, althoughon two occasions I foundindividuals foraging in relativelyscrubby interior forest.In 1976, H. D. Pratt (pers.comm.) found severalbirds in native foreston the east coastof Marpi. At Marpi, where I gathered data on the structure of habitats chosen by Nightingale Reed-Warblers, I found that habitat choicewas variable, Vol.63, No. 4 Reed-WarblerEcology [441

TABLE1. Percentvegetation cover in the territoriesof sevenNightingale Reed-Warblers.

Plant species Habitat Elephant Tangan- Meadow Thicket Mixed grass tangan Vines Other

Mean 26.3 50.1 23.7 49.9 70.8 9.0 15.7 SD 13.7 20.0 9.3 20.0 13.7 8.0 13.1 with neither elephant grassmeadows nor tangantanganthickets consis- tentlypredominating in territories.Thickets averaged highest in coverage, however(Table 1). Asidefrom tangantanganand elephantgrass, no plants other than vines had appreciable cover in reed-warbler territories. The predominantvegetation height in territorieswas > 3-6 m, althoughvari- ation among territories was again great (Table 2). In tangantanganthe predominantheight was >3-6 m, whereasin elephantgrass it was <2 m. Size of the seventerritories mapped in 1988 averaged9338 + 3433 (SD) m2. Regressionof territory sizewith principal habitat parameters showedlittle relationshipbetween area and coverby elephantgrass mead- ow (r2 = 0.02, P > 0.5) or tangantanganthicket (r 2 = 0.14, P > 0.5). In 1989, five of theseterritories appearedunaltered. Parts of the other two territorieswere usurpedby adjacentterritorial birds. One had a slight boundarychange in 1989, but the other had substantiallyaltered bound- aries.As I spentless time determiningterritory boundariesafter 1988, I was uncertain whether the malespresent in thesetwo territories,the same onesthat were presentin 1988, compensatedfor their lossof someterritory by expandingtheir territory elsewhere.Both of thesemales used the same preferred songperches as in 1988, however. Morphology.--Male Nightingale Reed-Warblers were significantly larger than femalesin mass (equal variancemodel t = 4.1, 26 dr, P < 0.001), wing length (t = 7.7, 27 dr, P < 0.0001), tail length (t = 5.0, 27 dr, P < 0.0001) and tarsuslength (t = 4.5, 27 dr, P < 0.0001), but did not significantlydiffer from femalesin bill length (t = 1.2, 27 dr, P > 0.25). For the 95% confidenceinterval, mass,and wing, tail and tarsus length had non-overlappingranges, and were therefore useful in sexing

TABLE2. Percent coverof vegetationheights (m) in the territoriesof sevenNightingale Reed-Warblers.Coverage for plant specieswas computedas observations/heightcat- egorydivided by total samplingstations (80).

Plant species Elephant grass Tangantangan Overall <2 >2-3 <2 >2-3 >3-6 >6 <2 >2-3 >3-6 >6

Mean 33.6 15.0 11.1 19.6 41.3 3.2 26.1 29.6 41.3 2.9 SD 15.0 11.8 7.7 6.3 13.5 2.0 9.5 11.3 12.5 1.7 442] R. J. Craig J. Field Ornithol. Autumn 1992

TABLE 3. Morphology of Nightingale Reed-Warblers.

Wing chord Tail length Tarsus Bill length Mass (g) (mm) (mm) length (mm) (mm)

Male Mean 35.9 87 83 34.8 23.0 SD 1.9 2 2 1.0 1.0 CV 5.4 1.9 2.6 2.8 4.4 95% CI 0.9 0.8 1.0 0.5 0.5 n 19 19 19 19 19

Female Mean 32.0 82 78 33.2 22.6 SD 3.1 2 3 0.7 0.9 CV 9.8 2.7 3.8 2.0 3.9 95% CI 2.4 1.6 2.1 0.5 0.6 n 9 10 10 10 10 nonbreedingindividuals (Table 3). Hence, birds with a mass>35.0 g, wing length >86.4 mm, tail length >81.5 mm or tarsuslength >34.3 mm were males,and birds with a mass<34.4 g, wing length <83.2 mm, tail length <79.8 mm or tarsus length <33.7 mm were females.

DISCUSSION Like mostterritorial (Morse 1980), NightingaleReed-War- bler males defendedterritories against conspecificswith song and by pursuit of intruders. High site fidelity by males and low mate fidelity by femalesare also typical of passerines(e.g., Nolan 1978), althoughmale fidelity, and perhaps survivorship,appeared higher (82%) in this sed- entary speciesthan in Nolan's (1978) migratory population of Prairie Warblers (Dendroicadiscolor) (65%). Low female mate fidelity may be particularlyfavored in this smallisland population if geneticreassortment leadsto higher averageoffspring survivorship. Unlike A. orientalis,in which 17-43% of malesare polygynous(Urano 1985), and A. arundinaceus,in which 12% are polygynous(Dyrcz 1977), A. lusciniaappeared entirely monogamous.A. orientalis(Saitou 1976) and A. arundinaceus(Dyrcz 1977) are principally breedersof reed marshes, and marsh passerinesare believedto developpolygyny becausehabitat is limited and territories differ strongly in quality. Females maximize their reproductiveoutput by mating polygynouslywith a male on a su- perior territory (Verner and Willson 1966). This situationmay not hold in the upland habitats used by A. luscinia on Saipan, hence, polygyny doesnot appear to be favored in this population. Also unlike the marsh nesting A. orientalisand A. arundinaceus,A. lusciniaon Saipan has a wide range of breedinghabitats (the historically rare Guam [Jenkins 1983] and Pagan [Glass 1987] populationsmight have been restrictedto wetlands, although H. D. Pratt [pers. comm.] suspectsthis perceptionto be artifactual). The elephant grass-tangan- Vol.63, No. 4 Reed-WarblerEcology [443 tangan mosaicsnow most frequently used by A. lusciniaare, however, comprisedof introducedspecies (More and McMakin 1979). The range of habitats I observedbirds use, from reed marshesto forest openings, suggeststhat A. luscinia initially expanded its habitat use from reed marshesto adjacentHibiscus thickets and structurally similar disturbed sites, such as typhoon-createdforest openingsof meadow and thicket vegetation.Moreover, reed-warblersmay have usednative tangantangan (Leucaenainsularurn) that now growsmixed with L. leucocephalaalong the coast. Territories of A. lusciniawere 10 times larger than territories of A. orientalis,which Saitou (1976), using methodssimilar to mine, found to havereed marshterritories averaging 830 m2 at the peak breedingseason. Larger birdstypically have larger territoriesthan smallerbirds (Schoener 1968), but the differencein territory size of A. orientaliscompared to the 50% larger (by mass)A. luscinia(Nisbet and Medway 1972) seemsdis- proportionatelygreat, and is likely due to the habitat shift to uplandsby A. luscinia.Marshes are frequently highly productive,and in such pro- ductivehabitats smaller territories are sufficientto provideadequate food (Orians 1971). The populationincrease in Nightingale Reed-Warblersafter 1988 may have been a result of population recoveryfrom the effectsof a 1987 supertyphoon,whose >300 km/h winds destroyedmuch of the island's thicket habitats (P. Glass, pers. comm.). In contrast, the loss of two territories after 1989 appeareddue to agricultural activity. Compared to mainland membersof the A. arundinaceuscomplex, such as A. orientalis,A. lusciniais a large, very long-billedspecies. Mean male A. orientalisfall massis 24.1 g (Nisbet and Medway 1972), whereasmale bill length measuredon museumspecimens is 12.8 q- 0.7 mm (n = 8). Both thesedifferences are observedfrequently among island populations when comparedto mainland relatives(Blondel 1985). Larger bill size is related to a greater range of food size selection(Hespenheide 1973), a potentialadvantage in a competitivelyimpoverished or foodlimited island. Indeed, unlike its mainland relatives,which are insectivorous,the Night- ingale Reed-Warbler has extendedits prey selectionto includeinsects of at least3 cm in length (pers.obs.) and lizards (Marshall 1949). The Nightingale Reed-Warbler on Saipan is, therefore, a sedentary, sexuallysize dimorphic species of thicket-meadowcomplexes. Unlike close mainland relatives,it is largely or entirely monogamous,it defendscom- paratively large territoriesprincipally in uplandsrather than in marshes, and it is larger and longer billed.

ACKNOWLEDGMENTS

I thank A. Palacios,P. Glass and J. Reichel for extendingtheir hospitalityduring my stay on Saipan. D. Civco permitteduse of the RemoteSensing Laboratory, University of Connecticut,and A. Andors,American Museum of , permitteduse of the researchcollection. G. Clark, Jr., D. Power, H. D. Pratt and D. Steadmanprovided thoughtfulreviews of the manuscript,and B. Lussierprepared the map. The map insetwas reproducedwith permissionof the AmericanOrnithologists' Union. Financial supportwas 444] R.J. Craig J.Field Ornithol. Autumn 1992 providedby the University of ConnecticutResearch Foundation and Pittman-Robertson Federal Aid to Wildlife. Scientificcontribution No. 1321 of the Storrs Agricultural Exper- iment Station.

LITERATURE CITED

BAKER,R. H. 1951. The avifauna of Micronesia, its origin, evolution, and distribution. Univ. Kansas Mus. Nat. Hist. Publ. 3:1-359. BLONDEL,J. 1985. Habitat selectionin islandversus mainland birds.Pp. 477-516, in M. L. Cody, ed. Habitat selectionin birds. AcademicPress, New York, New York. DYRCZ,A. 1977. Polygamyand breedingsuccess among Great ReedWarblers Acrocephalus arundinaceusat Milicz, . Ibis 119:73-77. ENGBRING,J., F. L. RAMSEY,AND V. J. WILDMAN. 1986. Micronesianforest bird survey, 1982: Saipan, , Aguijan, and Rota. U.S. Fish Wildl. Serv. Rep. FOSBERG,F. R. 1960. The vegetationof Micronesia. Am. Mus. Nat. Hist. Bull. 119: 1-76. GLASS,P.O. 1987. Nightingale Reed-Warbler surveysand inventories.Pp. 154-157, in Commonwealthof the , Div. Fish Wildl. ProgressReport Oct. 1, 1982-Sep. 30, 1987. HESPENHEIDE,H. A. 1973. Ecologicalinferences from morphologicaldata. Ann. Rev. Ecol. Syst. 4:213-229. HILL, B. G., AND M. R. LEIN. 1989. Territory overlap and habitat use of sympatric chickadees. Auk 106:259-268. JENKINS,J. M. 1983. The native forestbirds of Guam. Ornithol. Monogr. 31. MARSHALL,J.T. 1949. The endemicavifaua of Saipan,Tinian, Guam and Palau. Condor 51:200-221. MAYR, E., M. A. TRAYnOR,JR., AND G. E. WATSON. 1986. Checklistof birds of the world, Vol. 11. Harvard Univ. Press,Cambridge, Massachusetts. 637 pp. MORE, P., AND P. MCMAKIN. 1979. Plants of Guam. Coil. Agr. Life Sci., Univ. Guam. Mangilao. 186 pp. MORSE,D.H. 1980. Behavioralmechanisms in ecology.Harvard Univ. Press,Cambridge, Massachusetts.383 pp. NISBET,I. C. T., AND L. MEDWAY. 1972. Dispersion, population ecologyand migration of Eastern Great Reed-WarblersAcrocephalus orientalis wintering in Malaysia. Ibis 114:451-494. NOLAN,V., JR. 1978. Ecologyand behaviorof the Prairie Warbler Dendroicadiscolor. Ornithol. Monogr. 26. ORIANS,G. H. 1971. Ecologicalaspects of behavior.Pp. 513-546, in D. S. Farner, and J. R. King, eds.Avian Biology, Vol. 1. AcademicPress, New York, New York. PRATT,H. D., P. L. BRUNER,AND D. G. BERRETT. 1987. A field guideto birdsof Hawaii and the tropical Pacific. PrincetonUniv. Press,Princeton, New Jersey. 409 pp. SAITOU,T. 1976. Territory and breeding density in the Eastern Acrocephalusarundinaceus orientalis. Misc. Reports, Yamashina Inst. Ornithol. 8:157- 173. SCHOENER,T. 1968. Sizesof feedingterritories among birds. Ecology49:123-141. STEADMAN,D. 1989. of birds in easternPolynesia: a review of the record,and comparisonswith other Pacific island groups.Archaeol. Sci. 16:177-205. URANO,E. 1985. Polygynyand the breedingsuccess of the Great Reed Warbler Acro- cephalusarundinaceus. Res. Pop. Ecol. 27:393-412. VERNER,J., AND M. F. WILLSON. 1966. The influenceof habitatson mating systemsof North American passefinebirds. Ecology47:143-147. Received17 Jun. 1991; accepted16 Dec. 1991.