Ornithological Monographs No.42 Speciation and Geographic Variation in Black-tailed Gnatcatchers Jonathan L Atwood SPECIATION AND GEOGRAPHIC IN BLACK-TAILED VARIATION GNATCATCHERS ORNITHOLOGICAL MONOGRAPHS This series,publishedby the American Ornithologists'Union, has been established for major papers too long for inclusion in the Union's journal, The Auk Publication has been made possiblethrough the generosityof the late Mrs Carll Tucker and the Marcia Brady Tucker Foundation, Inc Correspondenceconcerningmanuscriptsfor publication in the seriesshouldbe addressed to the Editor, Dr David W Johnston, 5219 Concordia St., Fairfax, Virginia 22032 Copies of OrnithologicalMonographs may be ordered from the Assistant to the Treasurer of the AOU, Frank R Moore, Department of Biology, University of Southern Mississippi, Southern Station Box 5018, Hattiesburg, Mississippi 39406 (See price list on back cover.) Ornithological Monographs, No 42, viii + 74 pp Editor, David W Johnston Special Reviewers for this issue, Lloyd F Kiff, Western Foundation of VertebrateZoology, 1100 Glendon Ave., Los Angeles,California 90024; Robert M Zink, Museum of Zoology, Louisiana State University, Baton Rouge, Louisiana 70803 Author, Jonathan L Atwood, Department of Biology, University of California, Los Angeles, California 90024 and Mahomet Bird Observatory, Box 936, Mahomet, Massachusetts02345 First received, 12 December 1986; final revision completed, 13 October 1987 Issued March 29, 1988 Price $10.00 prepaid ($8.00 to AOU members) Library of CongressCatalogueCard Number 88-70367 Printed by the Allen Press,Inc., Lawrence, Kansas 66044 Copyright ¸ by the American Ornithologists'Union, 1988 ISBN: 0-943610-53-2 SPECIATION AND GEOGRAPHIC IN BLACK-TAILED VARIATION GNATCATCHERS BY JONATHAN L ATWOOD Department of Biology University of California Los Angeles, California 90024 and Manomet Bird Observatory Box 936 Manomet, Massachusetts 02345 ORNITHOLOGICAL MONOGRAPHS PUBLISHED THE AMERICAN BY ORNITHOLOGISTS' WASHINGTON, 1988 NO D.C UNION 42 TABLE LIST OF FIGURES LIST OF TABLES LIST OF APPENDICES INTRODUCTION OF CONTENTS vi vii vii GOALS OF THE STUDY NOMENCLATURAL HISTORY MATERIALS AND METHODS MORPHOLOGICAL ANALYSES BIOMETRIC ANALYSES VOCAL ANALYSES RESULTS 10 GEOGRAPHIC AND ECOLOGICAL DISTRIBUTION 10 Polioptila californica 10 Polioptila melanura 13 Polioptila nigriceps 17 AREAS OF SYMPATRY BETWEEN POLIOPTIL•I MELANURA AND P CAL1- FORNICA 17 Palm Springs, Riverside Co., California 18 Valle de Trinidad, Baja California, Mexico 18 Bahia San Luls Gonzaga(and vicinity), Baja California,Mexico 18 "San Felipe Canyon," San Diego Co., California 19 BREEDING BIOLOGY 21 VOCAL Dilqq•RENCES AND REPRODUCTIVE ISOLATION 22 Comparison of major vocalizations 22 Vocal playback experiments 30 MORPHOLOGICAL VARIATION 31 Secondary sexual variation 31 Relative variability 33 Character correlations and redundancy 38 Univariate character analyses 38 Geographicheterogeneityamong characters 38 Geographic patterns of character variation 39 Multivariate analyses 50 Inter- and intraspecific variation 50 Potential hybrid specimens 56 DISCUSSION 59 SPECIESLIMITS AND TAXONOMY 59 COMMON NAMES 63 HISTORICAL BIOGEOGRAPHY 63 ACKNOWLEDGMENTS 66 SUMMARY 67 LITERATURE CITED 68 LIST Figurel OF FIGURES Approximate geographicdistributions of Polioptila melanura, P californica, and P nigriceps Groupings of specimensofPolioptila melanura into sample areas for analysis Groupings of specimensof Polioptda californica and P nigriceps into sample areas for analysis Categoriesused in analysisof variation in tail spot shape Polioptila caHfornica habitat in coastal southern California and Baja California north of 30øN latitude 11 Polioptila californica habitat in the Vizcaino desertof central Baja California 12 Polioptila californica habitat in the Cape region of Baja California 12 Polioptda melanura habitat in the Sonoran desert of southeastern California and northwestern Baja California 14 Polioptila melanura habitat in the Sonoran desert of southern 10 Polioptila melanurahabitat in the Chihuahuandesertof Durango, 11 Polioptila melanura habitat in the Chihuahuan desert of Guanajuato, Mexico 16 12 Polioptila nigricepshabitat in the arid thorn scrubof the Mexican Arizona and the northwestern Mexico 14 15 16 17 18 19 20 21 23 24 western coast 17 Distribution of Polioptila melanura and P californica in central Baja California 20 Habitat in the region of sympatry between Polioptila melanura and P californica 21 Vocalization Type I in Polioptila californica,P nigriceps,P caerulea, P plumbea, and P albiloris 23 Vocalization Type I in Polioptila melanura 24 Vocalization Type II in Polioptila melanura 25 Vocalization Type II in Polioptda californica 26 Vocalization Types III and IV in Polioptila melanura and P californica 28 Vocalization Types V and VI in Polioptila melanura, P californica and P nigriceps 29 UPGMA phenogram of morphological characters in Polioptila melanura 22 15 15 mainland's 13 Mexican mainland 39 Geographic variation in P6LEN in males 40 Geographic variation in P6LEN in females 41 Geographic variation in MASS in Polioptila melanura and P californica 42 25 26 27 28 29 30 31 32 33 34 35 36 37 Geographicvariation in TLEN 43 Geographicvariation in R5PCT 46 Geographicvariation in BRSTB 47 Geographicvariation in R5SPCT 48 Geographicvariation in R6SSH 49 Principal componentsanalysisfor males 51 Principal componentsanalysisfor females 52 UPGMA phenogramfor males 53 UPGMA phenogramfor females 54 Geographicvariation in PC1 scores 56 Geographic variation in PC2 scores 57 Geographicvariation in PC3 scores 58 Canonicaldiscriminant analysisof allopatric and sympatricsamples of Polioptila melanura and P californica 58 38 Phylogeneticrelationshipsand distribution in three taxa of North American xeric-adaptedbirds 64 LIST Table OF TABLES Specimensusedin analysesof morphologicaland vocal variations Abundanceof Polioptila melanura and P californicain areasof sympatry 21 Speciesrecognitionduring playback tests 31 Secondarysexual dimorphism in linear measurements 32 Secondarysexual dimorphism in color analyses 33 Variability and geographicheterogeneityin morphologicalcharacters of Polioptila melanura 34 Variability and geographicheterogeneityin morphologicalcharactersof Polioptila californica 35 Variability and geographicheterogeneityin morphologicalcharacters of Polioptila nigriceps 36 Geographic differencesin relative variability 37 10 Correlation coefficients between correlations of variation and lat- itude and longitude 38 11 Correlation coefficientsbetween MASS and morphological char- acters 42 12 13 14 15 Geographicpatternsof charactervariation in Polioptila melanura 44 Geographicpatternsof charactervariation in Polioptila californica 45 Factor loadingson PCs 1-3 (all speciescombined) 50 Factor loadingson PCs 1-3 (by species) 55 16 ANOVA for PCs 1-3 LIST I 55 OF APPENDICES Localitiesof Specimensusedin Morphologicaland Vocal Analyses 71 vii INTRODUCTION The gnatcatchers,genusPolioptila, representa well defined yet poorly studied taxon of New World birds Comprised of approximately 10 species(Paynter 1964, A.O.U 1983), the genushas generally been classifiednear the gnatwrensMicrobatesand Ramphocaenus(Mayr 1946; Mayr and Amadon 1951; Paynter 1964), although differencesin external morphology, nests,and eggsindicate that Polioptila might not be closelyrelated to thesegenera(Rand and Traylor 1953; Paynter 1964; Kiff 1977) At higher taxonomic levels the position of Polioptila is even lesscertain, with most authors consideringthe genusto be a member of the Old World insect eaters (family Muscicapidae, subfamily Sylviinae; Ridgway 1904; Mayr and Amadon 1951; A.O.U 1983) However, the resultsofDNA x DNA hybridization comparisonssuggestthat the gnatcatchers,alongwith the gnatwrens, Verdin (Auriparusfiaviceps), true creepers(Certhia, Salpornis), and wrens, are not closely allied to the sylviine warblers but rather represent a monophyletic New World clade (Sibley and Ahlquist 1985) These authors (pers comm.) place the gnatcatchers,gnatwrens,and Verdin in the subfamily Polioptilinae of the family Certhiidae, with other subfamiliesof the Certhiidae being the Certhiinae (true creepers)and the Troglodytinae (wrens) Regardlessof its relationship to other groups of birds, Polioptila itself is a distinctive, easily recognizedgenuswith little pheneticdivergenceamong its component species.This morphological uniformity has led to some confusion with regardto specieslimits within the genus.As an extreme example of the taxonomic difficultiespresentedby Polioptila, Phillips et al (1973) noted that museum specimens now consideredto representthree distinct species(P plumbea, P albiloris, and P nigriceps)had earlier been ascribed to a single subspeciesby Griscom (1930) GOALS OF THE STUDY In this study I considera long-standingquestionconcerningspecieslimits within the "black-tailed" gnatcatchersthat are presently referred to as Polioptila melanura (A.O.U 1983) Although this systematicproblem is addressedmainly on the basisof behavioral and ecologicaldata, I also provide information on morphologicalcharactervariation in three (as defined here) sibling species:P melanura, P californica, and P nigriceps(Fig 1) This study is intended primarily as a taxonomic revision; however, the resultsare presentedin the context of other recent studies of geographic variation in birds and thereby contribute to the growingbody of literature dealing with the evolutionary significanceof intraspecific variability (Zink and Remsen 1986) Also, I compare the distributional and cladistic patterns of these three speciesof Polioptila with other genera occurring in the arid regions of western North America, and use these comparisonsas the basis for discussingthe possible history and mode of speciation in the "blacktailed" gnatcatchers(Cracraft 1982, 1983) NOMENCLATURAL HISTORY The first formal referenceto a gnatcatcherof the "black-tailed" group was the descriptionby Baird (1854) of Culicivoraplumbea from Arizona The absenceof a black cap in the type specimen,a male in basicplumage,was consideredto be ORNITHOLOGICAL STATUTE MONOGRAPHS NO 42 MILES FIO.1 Approximate geographic distributions of Polioptilamelanura, P californica, andP nigriceps characteristicof the species.Subsequently,Lawrence (1855) describedfrom Texas a specimenof a male gnatcatcherwith a black cap and mostly black outer rectrices, believing it to be an exampleof the black-cappedCulicivoraatricapilla of Middle America [= present day White-lored Gnatcatcher, Polioptila albiloris (Ridgway 1904; A.O.U 1983)] Lawrence(1857) later recognizedthat his black-tailed gnat- catcherand atricapilla were actuallydifferentspecies.However, either beingunaware of the existenceof Baird's plurnbea or of the similarity ofplurnbea to his putative species,Lawrence describedthis black-tailed, black-cappedgnatcatcher from Texas as a new species,naming it Polioptila rnelanurafollowing Sclater's (1855) changeof the genericname of Culicivorato Polioptila During theseearly years, Polioptila plurnbea (Baird) was referred to by the English names LeadcoloredGnatcatcheror Arizona Gnatcatcher,and Polioptilarnelanura(Lawrence) as the Black-cappedGnatcatcher (Baird et al 1875; Coues 1878) Later, Brewster (1881) reported that the presenceor absenceof a black cap in these birds, rather than being a valid character separatingP plurnbea and P rnelanura,instead was merely dependenton age, season,and sex Consequently, he consideredrnelanurato be synonymouswith the older name plurnbea However,Penard(1923) recognizedthat Pallasin 1769 had incorrectlyassigned another speciesof gnatcatcherfrom Middle and SouthAmerica (the present-dayTropical Gnatcatcher,Polioptila plurnbea) to the genusTodus Correct placement of this species(Todusplurnbea) in the genusPolioptila thus resultedin a nomenclatural conflictwith Baird'sPolioptilaplurnbeaand, becauseof priority, requiredthat Polioptilaplurnbea(Baird) be changedto Polioptila rnelanura(Lawrence).During the early 1900s the previous English names for Polioptila rnelanura had been GNATCATCHER SPECIES LIMITS replacedin generalusageby the name PlumbeousGnatcatcher(A.O.U 1886; Ridgway 1904; Grinnell 1915; Bailey 1927) Polioptilacalifornicawasalsooriginallydescribedasa distinctspecies(Brewster 1881), with the first associatedEnglish name being the California Black-capped Gnatcatcher.During the early 1900s, P californicawas generallyknown as the Black-tailed Gnatcatcher (A.O.U 1886; Ridgway 1904; Willett 1912; Grinnell 1915; Bailey 1927) The presentlyacceptedview of specieslimits in "black-tailed" gnatcatcherswas proposedby Grinnell (1926) Despite recognizingthat the two forms could be easily distinguishedvocally (Grinnell 1904), Grinnell believed that P melanura and P californicawere nonethelessconspecific.The primary basisfor this conclusion was the morphological similarity of populations in the Cape region of Baja California to populations in the Sonoran desert of southeastern California and Arizona Accordingto Grinnell (1926), "even thoughcalifornica[of coastal southernCalifornia and northern Baja California] is to melanura of southeastern California and Arizona asa full species,variation geographicallyto the southward, throughthe [centralBaja California population], to the Cape form, and intergradation thencewith melanura through individual variation, warrants considering [californica]just the extreme in a continuousseriesof subspecies."The English names Grinnell gave to theseforms were the PlumbeousBlack-tailed Gnatcatcher for the nominatesubspecies (includingpopulationslater describedasP.m lucida) and the California Black-tailedGnatcatcherfor P.m californica(Grinnell 1926) Theserevisionswere acceptedin the 4th edition of the A.O.U Check-list(A.O.U 1931) Recent field ornithologistsin the United Statesand Mexico have generally ascribedthe English name Black-tailed Gnatcatcher to all populations of gnatcatcherswith outer rectricesthat are mostly black (Peterson1941; A.O.U 1957) Someinvestigatorshave distinguishedthe "coastal"Black-tailedGnatcatcher(P m californica)from populationsof P.m lucida occurringin the desertsof California (McCaskie and Pugh 1964; Atwood 1980) Rea (1983) and Unitt (1984) both alluded to the possibility of P melanura and P californica being distinct specieson the basisof their vocal differences,but provided no in depth analysis of specieslimits Curiously, the most recent contribution to the nomenclatural confusion surrounding these birds has come from the A.O.U Check-list Committee itself, which reversed the more traditional associations of common and scientificnames by applying the English name Plumbeous Gnatcatcher to Polioptila (melanura) californica and Black-tailed Gnatcatcher to Polioptila melanura (A.O.U 1985) Here, I use species names and limits in the "black-tailed" gnatcatchersas follows:(a) Polioptila californica,composedof the currentlyrecognized(A.O.U 1957) subspeciesP melanura californica, P melanura pontills, and P melanura margaritae, and (b) Polioptila melanura, composedof the subspecies P melanura melanura, P melanura lucida, and P melanura curtara Support and discussion of these conclusionswill be provided in subsequentsections The third speciesincludedin this study,the mostly white-tailed Black-capped Gnatcatcher(Polioptila nigriceps),also has had a rather confusednomenclatural history becauseof uncertaintiesregardingits relationshipto the White-lored Gnatcatcher(Polioptila albiloris).I follow here the treatment usedby the 6th edition GNATCATCHER SPECIES LIMITS 61 the use of subspeciesis generallyavoided becausethe basictaxonomic (and evolutionary) entity is consideredto be the species.As discussedabove, P.m melanura and P.m lucida might legitimately be consideredphylogeneticspecies based on the presenceof morphological differencesbetween the two populations Although available specimenmaterial from the isolatedpopulationof P melanura on Isla Tibur6n is limited, at the present time I recommend tentative retention of P.m curtata (van Rossem 1932) I know of no specimensof this species fromIslaSanEsteban, andonlya single specimen exists fromIsla•ngel de la Guarda All of theseisland populationsofP melanurawarrant further study Clinal variation in multiple morphologicalcharacterswas found in populations of P californica north of the Cape region of Baja California; at approximately 25øN latitude, many charactersshoweda relatively sharptransition or step.Under thebiologicalspecies concept,two subspecies ofP californicawouldbe reasonably recognized:P c californicain the northernportionsof the species'range,and P c margaritae in the Cape region I see no compelling reason to maintain P c pontdisfor populationsin central Baja California, which in mostcharactersmerely seem to exhibit smooth, clinal variation with populations farther north; in fact, this subspecieswas originally defined in terms of these"transitional" characters (van Rossem 1931a) Strict applicationof the phylogeneticspeciesconceptmight considerP c californica and P c margaritae to be "species" on the basis of diagnosticmorphologicaldifferences(P c margaritaehaving relatively shorttails and more extensive white on the outer rectrices) However, Cracraft (1983) conceded that "under some circumstances subspecificnames could be applied to populations showing clinal variation, and subspecies boundariescould then be determinedby sharp gradientsin charactervariation." The pheneticdifferences betweenP c californica and P c margaritae are not as profound as thosebetweenP.m melanura and P.m lucida Certainly P c californica and P c margaritae should not be consideredbiological species(no distributional disjunctionor evidenceof reproductive isolation exis0, and becauseof the relatively slight differentiation between them I considerthe two forms only marginalspeciesby the phylogeneticdefinition Assumingconcordancebetweengeneticand pheneticdivergenceaswell as similar rates of evolution within thesegroups(both somewhatquestionableassumptions; Zink 1988), P c californicaand P c margaritae divergedmore recentlyfrom one another than did P.m melanura and P.m lucida A study of geneticdifferencesbetweenthesefour major populationsof"black-tailed" gnatcatcherswould be most instructive Also, additional specimen material from populations of P californicaon islandsin the Gulf of California would be desirable A third member of the genusPolioptila, the mostly "white-tailed" P nigriceps, was included in this study for comparative purposes Although several of the important vocalizationsofP nigricepsresemblethoseofP californica,the overall vocal repertoiresof P nigricepsand P californicaare more different from each other than are the vocalizationsofP californicaand P melanura from eachother An electrophoretic analysis of these three speciesindicated that P nigricepswas sharplydifferentfrom Sonorandesertsamplesof P melanura and specimensof P californicafrom northern Baja California and California; the "black-tailed" gnatcatcherswere indistinguishablebasedon the limited number of loci that were studied(R Matson, pers comm.) Thus, morphological,behavioral, and genetic 62 ORNITHOLOGICAL MONOGRAPHS NO 42 analysesall supportthe contentionthat Polioptila melanura (Sonorandesertgroup) and Polioptila californica are more closely related to each other than either is to Polioptila nigriceps In summary, two or three speciesof Polioptila with mostly black outer rectrices exist, dependingon whether "species"is definedas a reproductivelyisolatedunit (sensuMayr) or asa distinguishableevolutionarygroup(sensuCracraft).Following Amadon's (1966) use of bracketsin designatingsuperspecies and allospecies,my arrangementof the "black-tailed" gnatcatchercomplex is as follows: Superspecies: Allospecies: Allospecies: Polioptila [melanura] Polioptila [melanura] melanura Polioptila melanura melanura Polioptila melanura lucida Polioptila melanura curtata (tentative) P olioptila [melanura] californica Polioptila californica californica Polioptila californica margaritae The following key to these species,based on the data set used in this study (islandpopulationsexcluded)and usingonly morphologicalcharacters(asopposed to distributional information), correctly identified 97 percent of the specimens that were examined with 90 percent accuracy Most misidentifications involved females ofP m lucida that were identified as P.m melanura, and specimensof P californicafrom the Cape regionof Baja California (P c margaritae) that were confused with P.m lucida Incorporation of distributional data into such a key would result in virtually 100 percent accuracy;populations occurringin the only regionsof overlap betweenP californicaand P.m lucida are easilydistinguished morphologically,and distributional characterscould be used to separatepopulations that are phenetically convergent KEY TO THE "BLACK-TAILED" GNATCATCHER COMPLEX a 1A Outer web of rectrix not completelywhite P californica lB Outer web ofrectrix completelywhite and tail length < 47.5 mm and rectrix tail spot shapeType III or IV P californica 1C Not as above 2A Formula A > 0.95 or length of tail spot on rectrix < mm 2B Not as above 3A 3B 4A 4B Length of tail spot on rectrix < 6.5 mm P californica Lengthof tail spoton rectrix > 6.5 mm P.m lucida Formula B < 102 and rectrix tail spot shapeType I or II P.m lucida Formula B < 102 and rectrix tail spot shapeType III P californica 4C Formula B > 112 4D Not as above P.m melanura 5A Formula C > 2,420 or length of tail spot on rectrix > mm 5B Length of tail spot on rectrix _-