Behavior, Mimetic Songs and Song Dialects, and Relationships of the Parasitic Indigobirds (Vidua) of Africa

BEHAVIOR, MIMETIC SONGS AND SONG DIALECTS, AND RELATIONSHIPS OF THE PARASITIC INDIGOBIRDS (VIDUA) OF AFRICA

ROBERT B. PAYNE

ORNITHOLOGICAL MONOGRAPHS NO. 11

PUBLISHED BY THE AMERICAN ORNITHOLOGISTS' UNION

1973 BEHAVIOR, MIMETIC SONGS AND SONG DIALECTS, AND RELATIONSHIPS OF THE PARASITIC INDIGOBIRDS ('?IDU..4,) OF AFRICA ORNITHOLOGICAL MONOGRAPHS

This series,published by the American Ornithologists'Union, has been establishedfor major paperstoo long for inclusionin the Union's journal, The Auk. Publicationhas been made possiblethrough the generosityof Mrs. Carll Tucker and the Marcia Brady Tucker Foundation, Inc. Correspondenceconcerning manuscripts for publicationin the series shouldbe addressedto the incomingEditor, Dr. John W. Hardy, Moore Laboratoryof Zoology,Occidental College, Los Angeles,California 90041. Copiesof OrnithologicalMorrographs may be orderedfrom the Treasurer of the AOU, Burt L. Monroe, Jr., Box 23447, Anchorage,Kentucky 40223. (See price list on inside back cover.)

OrnithologicalMonographs, No. 11, vi + 333 pp. Editor, Robert M. Mengel Editorial Assistant, James W. Parker

IssuedJanuary 31, 1973 Price $8.00 prepaid($6.40 to AOU Members) Library of CongressCatalogue Card Number 73-75355 Printedby the Allen Press,Inc., Lawrence,Kansas 66044

Erratum. For a time it was hoped that publicationof OrnithologicalMonographs No. 11 would occur in 1972, the year that appearson the running headsthroughout. Publica- tion in 1972 was not realized, however,and through one of thosenoxious oversights that plague the publication processthe entire work had been printed before the failure to change the running heads was detected.--Ed. G ' I H J

Breeding plumages of the males of ten kinds of indigobirds. A, Vidua chalybeata chalybeata, form "aenea" (FMNH 278472, Richard-To!l, Senegal); B, V. c. amauropteryx ( RBP 4469, Sabi Valley, Rhodesia); C, V. c. ultramarina (FMN H 204118, Mojjio, Ethiopia): D. l/. ]unerea codringtoni (RBP 4437. Penhalonga-Umtali. Rhodesia): E, l/. 1. !usitensisssp. nov. (RBP 4575, Lusitu River, Rhodesia); F. V. 1. [unerea (RBP 4424, 6 mi. E. Tzaneen, Transvaal); G, V. wilsoni, form "nigeriae" (RBP 4937, Panshanu Pass, Nigeria); H, l/. wilsonL form "camerunensis" (RBP 4855, Zaria, Nigeria); I, l/. wilsoni, form "wilsoni" (RBP 4960, Zaria, Nigeria); J, V. purpurascens (RBP 4419, Merensky Reserve, Transvaal). BEHAVIOR, MIMETIC SONGS AND SONG DIALECTS, AND RELATIONSHIPS OF THE PARASITIC INDIGOBIRDS (VIDUA) OF AFRICA

ROBERT B. PAYNE

ORNITHOLOGICAL MONOGRAPHS NO. 11

PUBLISHED BY

THE AMERICAN ORNITHOLOGISTS' UNION

1973 TABLE OF CONTENTS

INTRODUCTION ......

ACKNOWLEDGMENTS ...... 4

ITINERARY...... z...... 7

BREEDING BEHAVIOR OF INDmOBIm)S ...... 9 Social Behavior and Mating Systems...... 10 Courtship Behavior ...... 29

BROOD PARASITISM AND THE INTERACTIONS OF INDIGOBIRDS AND FIREFINCHES ...... 39 Evidenceof Parasitismand Host Specificityin Indigobirds ...... 39 Ecology of Indigobird Parasitism ...... 42 Behavioral Interactions Between Indigobirds and Firefinches ...... 44 Mimicry of Eggs and Young ...... 52

VOCALIZATIONS OF FIREFINCHES AND THEIR MIMICRY BY THE INDIGOBIRDS ...... 61 Firefinch Vocalizations ...... 62 Vocal Mimicry in Indigobird Song ...... 94 Discussion ...... 101

NONMIMETIC VOCALIZATIONS AND SoNG DIALECTS ...... 111 Chatters ...... 129 Complex Nonmimetic Songs ...... 131 Repertoire Size, Mating Systems,and Individual Recognition ...... 146 Origin of Nonmimetic Songs and Song Dialects ...... 147

SoNG DIALECTS AND POPULATION STRUCTURE ...... 155 Estimatesof Population Structure from Song Dialects in the Indigobirds ...... 156 Significanceof Population Structure ...... 158

BEHAVIORALCONTEXT OF INDmOmRD VOCALIZATIONS AND Tm• RESPONSESOF BIRDS TO SoNGS ...... 160 Analysis of Song Recordings ...... 160 Behavioral Context of Songs ...... 161 Responsesof Wild Males to Recorded Songs ...... 164 Responsesof Captive Female Indigobirdsto Species-and Population-specific Songs ...... 165

ASSORTATIVEMATING AND ISOLATING MECHANISMS ...... 173 Assortative Mating ...... 173 Breeding Biology and Potential Isolating Mechanisms ...... 175

DISTRIBUTION OF FIREFINCHES AND INDIGOBIRDs ...... 186 Southern Africa ...... 186 East Africa ...... 194 North and West Africa ...... 198 Central Africa ...... 205 Discussion ...... 207

VARIATION AND RELATIONSHIPS IN THE INDIGOBIRDS ...... 209 Basis of Identification and Taxonomic Allocation of Specimens ...... 209 Analysis of Variation ...... 211 Vidua chalybeata ...... 218 Vidua purpurascens ...... 235 Vidua /unerea ...... 240 Vidua wilsoni ...... 250 The V. [unerea--V. purpurascensComplex ...... 263 Intergradation of V. chalybeata and Other Species ...... 273 Discussion ...... 277

THE IMPORTANCE OF SONG BEHAVIOR AND GEOGRAPHIC ISOLATION IN DIFFEREN- TIATION AND SPECIATION OF THE INDIGOBIRDS ...... 280 Determination of Mating Groups by Imprinting to Host Songs ...... 280 GeographicVariation in Behaviorand the Breakdownof ReproductiveIsolation Between Indigobirds ...... 282 Two Models of Speciation...... 284 Discussion ...... 292

SUMMARY ...... 298

LITE•a•JRE CITED ...... 302

AP2ENDIXA (Locality and Other Data for Audiospectrographs) ...... 313 APPENDIXB (Distribution of Firefinches and Indigobirds) ...... 322

vi INTRODUCTION Studiesof behaviorhave advancedour understandingof the relationships of birds of the world, both amonggenera and familiesand amongclosely related species.While comparativestudies of the behaviorof closelyrelated birds have generallyconfirmed the ideasof speciesrelationships based on the appearanceof birds, in a few instancesbehavior study has been importantin revealingspecies whose existence was previouslyunsuspected. The eminent pioneerBritish naturalistGilbert White recognizedthe Chiffchaff (Phyllos- copuscollybita) as a speciesdistinct from the Willow Warbler (P. trochilus) by differencesin songsand call notesmany yearsbefore taxonomistsbecame aware of any morphologicaldifferences (Mayr, 1963: 52). Speciesdiffer- encesbetween extremely similar birds have been first recognizedthrough behavioraldifferences, especially song, in severalgenera including flycatchers (Empidonax), warblers (Cisticola), and grackles(Cassidix) (Stein, 1958, 1963; Traylor, 1967; Selanderand Giller, 1961). Not only have cognate speciesbeen detected;behavior studies have also led to the recognitionof morphologicallydissimilar populations as membersof the samespecies (e.g., Lanyon, 1969; Thielcke, 1969b). The indigobirds(subgenus Hypochera, genus Vidua, subfamilyViduinae) are small parasiticfinches differing slighfiyin the color of the glossof the blackishbreeding plumage of the male and in bill and foot color; the non- breedingmales and the femalesare small brownishbirds. They are also known as indigofinches, widow-finches, steel finches, combassous and Atlas- witwen. Museumworkers have reachedno agreementon the relationships among the indigobirdsthrough studiesof the plumage charactersalone; Mackworth-Praedand Grant (1949) recognizedeight species,whereas White (1962, 1963a) has consideredthe indigobirdsto be a singlepolymorphic,polytypic species. Only a few previouslycollected female museum specimens(both V. chalybeata,the Village Indigobird)have been taken with malesof knownform and havehad notedon the label the charactersby which the femalesmay sometimesbe identified,namely the colorsof the bill and feet. In the absenceof field reportsfrom Africa of differencesin behavior it was most surprisingand excitingwhen Nicolai (1961) discoveredthat V. chalybeatain captivitymimics the song of its host, the SenegalFirefinch (Lagonostictasenegala). This observationled him to a discoveryof host- specificvocal mimicry in most other speciesof viduines (Nicolai, 1964). Stimulatedby his observationsand by Traylor's (1966) study of variation in this complexof forms, I visitedAfrica to observebehavior and record the songsof mostforms of indigobirdsand to collectthe singingmales of each kind of indigobird as well as the females mating with them. The songmodels and fosterersof the indigobirds,the firefinches(Lagoho- 2 ORNITHOLOGICAL MONOGRAPHS NO. 11 stictaspp.), are membersof the subfamilyEstrildinae (the waxbills) in the familyEstrildidae. Mayr et al, (1968) recognizeeight species of firefinches, buttwo of theseare members of superspecies groups, and as they are knownto be specificallydistinct I am regardinglandanae as conspecificwith L. rubricam(as did Chapin,1954: 523) and nitidulaas conspecificwith L. rufopicta(following White, 1963b:202). The firefinchspecies that the indigobirdsmimic and parasitizeare L. senegala,L. rhodopareia,L. rubricata,L. larvata,and L. rara. The remainingfirefinch, L. rufopicta(including nitidula), is not knownto be a songmodel of the indigobirds. The indigobirdsand other viduinesare usuallyregarded taxonomically as a subfamilyof the widespreadOld World weaverfinches, Ploceidae. Earlier workersoften regardedthe viduinesas closelyrelated to the Estrildidae,but Friedmann(1960) emphasizedthat many similaritiesbetween the viduines and estrildid.sare adaptive,inasmuch as the viduinesmimic their estrildid hostsin eggcolor and the markingsof the young.Friedmann and others have reasonedthat the viduinesare most likely derivedfrom the Ploceinae, the weaverfinches, and shouldbe placedin the samefamily with them. On the other hand Sibley (1970) has found that in somebiochemical features the viduinesare moresimilar to the Old World sparrows(Passer and related genera)than eitherof thesegroups are to the Ploceinaeor the Estrildidae, andhe suggeststhat the closestrelatives of the viduinesmay be the Old World sparrows,which he regardsas a separatefamily, Passeridae.The present studyis concernedwith the relationshipswithin the indigobirdspecies complex rather than with family relationshipsamong the Old World finchesand sparrows,although some comparisons of the behaviorof viduineswith the weaver finches,sparrows, and grassfinchesare included. For convenienceI have followedhere the systematicarrangement (except for the viduinespe- cies themselves)of Check-listof Birds of the World, volumesXIV and XV, in which the Estrildidaeare recognizedas a family and in which Passerinae, Ploceinae,and Viduinae (with the indigobirds) are subfamiliesof Ploceidae. The viduinescomprise about 12 speciesoften unitedin a singlegenus, Vidua. All forms are African. Traylot (1968) recognizesthree subgenera-- Hypochera(the indigobirds),Vidua (four speciesin whichthe breedingmale has a long, slendertail), and Steganura(the paradisewhydahs, in which the tail is large and ornate). In the presentstudy four speciesof indigobirdsare recognizedtaxonomically as a resultof the field work, analysisof museumspecimens, and long contemplation.The forms "nigeriae,""camerunensis," and "wilsoni" are regardedas conspecific,and the name V. 'wilsoniis usedfor all of these,but the form names,in quotationmarks, are usedto describethe appearanceof the birds. The forms codringtoni,nigerrima, and [unerea are regardedas subspeciesof a singlespecies, V. [unerea. The othertwo speciesare V. pur- 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 3 purascensand V. chalybeata. The rationale behind this taxonomic schemeis discussedon pp. 209-210 and beyondin the systematicsection. The relationshipsamong the indigobirdsare inadequatelydescribed simply in traditionalterms of eitherbiological or typologicalspecies. As notedin earlierreports (Nicolai, 1967, 1968; Payne,1967, 1968a, 1'968b)the indigobirdsthat look alike generallymimic the songsof a singlespecies of firefinch, and differentforms living in the samearea mimic differentfirefinches. For example,in all of the localitiesthat I visited, V. chalybeataalmost always mimickedthe songsof L. senegala.In my field work nearly three miles of recordingtape were exposedto indigobirdsongs. Most recordedbirds were subsequentiycollected, often with their females.By comparingthe songswith the male and female study specimensassociated with each behavioralobser- vation,it waspossible to establishthat the indigobirdsmimic differentsongs andbehave as goodbiological species in someareas. At otherlocalities some of thesespecies apparentiy mimic the samesongs and interbreedwith each other. The presentpaper attempts in the first placeto describethe behavioral and evolutionaryrelationships among the indigobirds,and secondarilyto help in identificationof the indigobirdsin museumspecimens and in the field. Knowledgeof the behaviorof the indigobirdsallows not only evaluationof their speciesrelationships but alsoestimates of their populationstructure, in- cludingpopulation size and the degreeof isolationbetween neighboring populations. The number of individualsthat are likely to interbreedwith each otheris probablyin generalmuch smaller than the total numberof individuals in a species.In the indigobirdsthe local restrictionof certainnonmimetic song dialectsto very small regions,together with estimatesof populationdensity, make it possibleto estimatethe "neighborhoodsize" or "effectivepopulation size"(in the senseof Wright, 1969) of thesebirds. By comparingthe manner in whichthese song dialects vary it waspossible also to gainan idea of whether populationsare isolatedfrom eachother and to seein what mannernew song dialectsmay arise. Songdialects seem to be ideal markersfor studiesof population biology in certain birds. The complexityof the relationshipsamong the indigobirdsmay resultfrom imprintingby the youngupon the songsand calls of their host species,and oneof the mainpurposes of the presentstudy was to testthe hypothesispro- posedby Nicolai (1964, 1967) that speciationin the viduinesis a directcon- sequenceof host-specificimprinting rather than of geographicalisolation, a much more commoncondition in differentiationof birds (Mayr, 1963: 481- 515). Relationshipsamong the indigibirdswere further re-examinedin a studyof all available-museummaterial, including 302 birds taken in the field work, and comparingthe patternsof variationin morphologyas well as in song. Field observationsprovided information about the behavior,mating sys- 4 ORNITHOLOGICAL MONOGRAPHS NO. 11

tems,and techniquesof broodparasitism in the indigobirds.The behavioral contextof the differentkinds of songswas studiedto find the possiblefunc- tionsof song,and some playback experiments were made in the field andin captivity.Additional observations on the behaviorof indigobirdsand their firefinchhosts were madein captivityat Norman,Oklahoma, and at Ann Arbor, Michigan,where caged and aviaryfirefinches have bred. I watched the behavioralinteractions between the broodparasites and their firefinch hostsin the captivebirds flying free in our housefor two years.

ACKNOWLEDGMENTS This studywas made possible through the supportof the NationalScience Foundation(Fellowship 45166; GB-7720; GB-29017X). The Frank M. ChapmanMemorial Fund providedfunds for examinationof specimensat the AmericanMuseum of Natural History. My wife Karen accompaniedme in the field and was especiallyhelpful in locatingbirds and in recordingtheir songs. She also prepared the frontispiece from specimens,photographs, and living models. Many peopleprovided facilities, assistance with permits, and local distributionalinformation which aided the fieldwork. I am gratefulto V. Ame, A. Bauchi,C. W. Benson,D. H. Eccles,A. Forbes-Watson,H. Friedmann, C. H. Fry, W. J. Hamilton,G. Harrison,D. Jackson,J. G. E. Lewis,G. L. Lombard,I. MacFadyen,A. P. Mead, C. K. Niven, D. H. C. Plowes,A. C. Raines,G. Ranger,J. C. Ras, G. T. Roux, N.J. Skinner,M. Slogrove,R. H. N. Smithers,J. G. Williams, and J. M. Winterbottom. The Director of Na- ture Conservationpermitted field studiesat Hans MerenskyNature Reserve, Transvaal.P. Britton,J. H. Elgood,C. H. Fry, andD. Wellsmade available their sightobservations of Zambianand Nigerianbirds. Informationon Afri- canlocalities was provided by R. C. Button,B. P. Hall, R. C. Long,K. C. Parkes, G. T. Roux, H. Schouteden,W. Serle, R. Van Gelder, and C. J. Vernon. J. M. Winterbottom and R. K. Brooke made available the nest recordsof the SouthAfrican OrnithologicalSociety and the RhodesianOr- nithologicalSociety. K. Immelmann,M.-Y. Morel, J. Nicolai, and H. E. Wolters have discussed with me some of their observations on the behavior of African finches. For many reasonsI am especiallygrateful to M.P. S. Irwin, M. A. Traylor, and J. M. Winterbottomfor sharingtheir infomarion on the distributionand systematicsof African birds. Throughthe kindnessof the officialsand curatorsof manymuseums it was possibleto examine on loan the specimensunder their care or to visit the collectionsconcerned. I am indebtedto the followingfor their cooperation: Academyof NaturalSciences of Philadelphia(R. M. de Schauensee),Albany Museum,Grahamstown (H. Deacon),American Museum of NaturalHistory (D. Amadon), Berlin Museum: Universit•ityon A. Humboldt (G. Mauers- 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 5 berger, B. Stephan), BraunschweigUniversity (K. Immelmarm,O. von Frisch), BritishMuseum (Natural History) (I. C. J. Galbraith,D. A. Good- win), CarnegieMuseum (K. C. Parkes), Durban Museumand Art Gallery (P. A. Clancy,W. J. Lawson), East LondonMuseum (M. Courtenay-Lati- ruer,C. D. Quickelberge),Field Museumof Natural History(M. A. Traylor), Institut Royal ScienceNaturelies Belgium (R. Verheyen), Kaffrarian Mu- seum,King William's To•wn,South Africa (C. J. Skead,D. M. Comins), Kenya National Museum,Nairobi (A. Forbes-Watson),Los AngelesCounty Museum(H. Friedmann),MacGregor Museum, Kimberley (R. Liversidge), Museo Civico di Storia Naturale di Milano (E. Moltoni), Mus6um National d'HistoireNaturelie (C. Voisin), Museumof ComparativeZoology (R. A. Paynter,Jr.), Natal Museum (T. Cashion), PeabodyMuseum (N. P. Ash- mole), Port ElizabethMuseum (J. Spence,B. G. Donally), National Mu- seumsof Rhodesiaat Bulawayo,Salisbury, and Umtali (R. H. N. Smithers, M.P. S. Irwin, D. H. C. Plowes), National Museum of Zambia (C. Cro.ss), Royal Ontario Museum (J. C. Barlow), Natur-Museum und Forschnngs- Institut Senckenberg,Frankfurt am Main (J. Steinbacher),Mus6e Royal de l'Afrique Centrale (R. Schouteden,A. de Roo), South African Museum and South West African Museum (J. M. Winterbottom), Transvaal Museum (O. P.M. Prozesky),Ahmadu Bello University(A. P. Mead), Cornell Uni- versity (A. R. Weisbrod), LouisianaState University (G. H. Lowery, Jr.), IbadanUniversity (R. H. Parker), Universityof Malaya (D. West), Univer- sityof Michigan(R. W. Storer), United StatesNational Museum (R. L. Zusi), and ZoologischeMuseum Alexander Koenig, Bonn (H. E. Wolters); D. A. Zimmermanmade availablehis collection. G. L. Guy (Buiawayo) and P. le S. Milsrein (Pretoria) identifiedthe seedsand insectsfound in the stomachs of the birds. For coloranalysis of plumageR. F. Johnstonkindly madeavail- ablethe spectrophotometerat the Museumof Natural History,University of Kansas. Facilitiesfor keepinglive birds and specimensand equipmentfor song analysiswere provided by the Departmentof Zoology,the Animal Behavior Laboratory,and the StovallMuseum of the Universityof Oklahoma,and the MatthalBotanical Gardens and the Museumof Zoologyat the Universityof Michigan.C. C. Carpenterand J. F. Sassamanaided in the careof captive birds,and Gloria Sullivanand D. R. Miller assistedin the preparationof audiospectrographs.Martha Lackey, C. S. Adkisson, and J. R. Purdue addedtheir illustrativeand photographictalents. For theirhelpful suggestions and assistance on themanuscript I thank R. D. Alexander,Erica Dunn, D.C. Smith,R. W. Storer,J. G. Strauch,Jr., and Gloria Sullivan. All specimenscollected have been deposited in the Universityof Michigan Museum of Zoology. 6 ORNITHOLOGICALMONOGRAPHS NO. 11

FigureI. Locationof the main studyareas of indigobirdsin Africa. Letters indicatethe names of countrieswhere indigobirds may be observed, and numbers indicatethe localities of taperecording or collectionsin the present study. A ----Angola, Bo _--Botswana, B •_ Burundi,Ca = Cameroon,Ch = Chad,CB m Congo (Brazzaville),C = Congo(Kinshasa), CI: C6ted'Ivoire, D: Dahomey,E: Ethiopia,G = Gambia,Gh = Ghana,Gu : Guinea,HV = HauteVolta, K --- Kenya,L: Lesotho,M -- Malawi,Ma -• Mall,Mo = Mozambique,Ni _-- Niger, N _--Nigeria, PG = PortugueseGuinea, CAR _-- Republique Centrale Africaine, Rh= Rhodesia,R: Rwanda,S: Swaziland,SA = SouthAfrica, Se = Senegal, SL : SierraLeone, SO = Somalia,SWA = SouthWest Africa, T = Togo, TA = Tanzania,U = Uganda,Z : Zambia.Localities: I Hluhluwe,2 Ndumu, 3 Louw'sCreek, 4 MarbleHall, 5 Tzaneen,6 MerenskyReserve, 7 Kondowe, 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 7

ITINERARY My wife and I drovea coveredFord pick-uptruck whichwas our mobilehome as we campedwith the birds, and we also stayedwith friendswe met along the way. Two yearsof field observationsand 55,000 miles of travel were completed in the field study. The first year, 1965-1966, we spentin SouthAfrica during the breedingseason of the indigobirds,and the secondyear, 1966-1967, we followed the breedingseason northwards from South Africa through Rhodesia, Botswana,Mozambique, and Malawi to Kenya. I made observationsfor two additionalmonths in northernNigeria from July to September,1968. The study localitiesare shownon the map of Africa in Figure 1. The firstindigobird appeared on 14 January1966 at Tzaneen,Transvaal (23o52 ' S lat., 30ø16'E long.). We made observationsand collectionsof females from singingmales at Tzaneen, the Downs (24 ø10'S, 29 ø19'E), Hans Merensky Nature Reserve (23ø39'S, 30ø40'E), Kondowe (23ø45'S, 30ø48'E), the Lowveld Fisheries ResearchStation near Marble Hall (25ø00'S, 29 ø19'E), and Louw's Creek (25o40' S, 31ø20'E) in Transvaal through 3 April. Observationswere continued in the Zululand game reservesat Ndumu (27ø56'S, 32ø16'E) and Hluhluwe (28ø10'S, 32ø04'E) until 23 April, when we returned to the Cape Province for museum studies and field work with other birds. In the secondseason we beganobservations at Merenskyfrom 21 Decemberto 2 January,before the breedingseason had begun. From January 16 to 25, birds were observed,tape-recorded, and collectedat the Lowveld Fisheries,and from 29 Januaryto 12 Februaryobservations were continuedat Merenskyand Tzaneen. Finding three distinct speciesof indigobirdsassociated with three firefinchesin eastern Transvaal, we then drove to eastern Rhodesia, where three firefinches were known to occur, and found the expectedthree kinds of indigobirdsalong the Penhalonga-Umtali road at 18ø5YS, 32ø40'E. On 2 and 3 March we looked for indigobirdsbut found none at Mt. Selinda. From 3 to 9 March we stayedat the Sabi Valley ExperimentalStation (20ø20'S, 32ø18'E) and recordedand collected more indigobirdswithin three miles of this spot. Roadsideobservations of indigobirds in Mozambiquewere made from 50 to 65 miles northeastof Tete (Tete ---- 16ø10'S,33ø35'E) on routeto Malawi. A heavyrain (11 inchesin 24 hours) in Malawi the next day made it impossibleto drive to the Chididi regionof southern Malawi, where R. C. Long had collectedthree kinds of indigobirds,so on 16 March we drove to Monkey Bay (14ø06'S, 34ø55'E) on Lake Malawi (---- Lake Nyasa) where we worked until 21 March. On 22 March we saw a greenishindigobird along the roadsideat 15ø32'S, 35 ø18'E, besidea creek nine miles south of the

8 Maun, 9 Shorobe, 10 Lusitu River, 11 Chipinga, 12 Sabi Valley, 13 Penhalonga- Umtaliroad, 14 Salisbury, 15 Tete-Mwanza road, 16 Chileka, 17 Zomba, 18 Monkey Bay, 19 Lilongwe, 20 Salima, 21 Malindi, 22 Olorgesailie,23 Nairobi, 24 Kisumu- Kericho road, 25 Kakamega, 26 Sigor, 27 Sokoto, 28 Gusau, 29 Zaria, 30 Panshanu, 31 Bauchi 25 mi W, 32 Numan, 33 Kiri. 8 ORNITHOLOGICAL MONOGRAPHS NO. 11 market at Zomba, and we campedhere for the next few days. On 25 March we drove to Lilongwe (13ø59'S, 33ø44'E) where we recordedand collectedfor two days. We then droveover the rift escarpmentdown to Lake Malawi again and workednear Grand Beachand Salima alongthe roadside,returning from there to Blantyre. By this time the roadswere dry only as far southas Chikwawa,where we found no indigobirds.We returnedto Sabi Valley, Rhodesia,from 4-7 April and then to the Lusitu River area where we camped at 3,700 feet elevation at Hayfield B campand worked the river valleybelow at 1,200 feet (20ø01'S,32059 ' E) duringthe daysof 8-9 April. From therewe drove via Francistownto Maun, Botswana (19ø59'S, 23ø2YE) on 14 April and camped at the edge of town by the Thamalakane River. We recorded and collected in Maun with excursions to Boro, "Leomarin" safari camp, and Moremi Game Reserveuntil 24 April. After returningto Transvaalwe flew to Nairobi, Kenya, on 4 May and there rented a Volkswagen camper. After working in the museum and searching Nairobi for indigobirdswe droveto the coastat Malindi (3 ølYS, 40ø07'E), where red-billedindigobirds had been reported, and remained there from 9-12 May. We drovethen to Voi and alongthe Taveta road on 13-14 May without finding male indigobirds.On 15 May we drovesouth from Nairobi to Lake Magadi and found indigobirdsat Olorgesailie(Iø3YS, 36ø28'E) on a gravellyhill at the edge of the plain. We returned to Nairobi on 22 May and on 24 May left for the Kisumu area. For the following five days we camped along the Kericho-Kisumu road near Muhoroni, working with indigobirdsat mile 34 (0ø15'S, 38ø00'E) east of Kisumu. No indigobirdswere seen around Kakamega Forest. On 1 June we reached Sigor, West Pokot (lø30'N, 35ø28'E), and here Karen found the second speciesof indigobird V. purpurascensas well as the widespreadV. chalybeata. On 5 June we left Sigor and later recordeda bird at 0ø39'N, 34ø45'E, about 22 miles north of Kakamega. The followingday more birdswere recordedon the Kisumu- Kericho road. A trip to Namanga turned up no indigobirds;we returned to Lake Magadi and Olorgesailiefrom 10-14 June. On 15 June we recordedthe songof a V. chalybeatain Nairobi and flew to the Transvaal on the next day. From 20-24 Junewe birded at Merenskyand Tzaneen and found the breeding seasonto be nearly completedfor Transvaalindigobirds, though a few malesof eachspecies here were still in breedingplumage, and in earlyJuly we looked for the birds in Natal and the easternCape Provincebut found none. Fieldstudies in 1968were based at AhmaduBello University (11øi0'N, 7o40 ' E), Zaria, Nigeria. All four forms of Nigerian indigobirds (V. chalybeataneumanni and also the "nigeriae," "camerunensis,"and "wilsoni" forms of V. wilsoni) were observedhere from 10 July to 10 August and later, all within walking distance. Trips were taken to Dumbi Woods (10ø52'N, 7ø34'E) on 17 August,to Gusau (12ø09'N, 6ø39'E) and Sokoto(13ø04'N, 5ø15'E) on 12-14 Augustand to Kogum (9ø17'N, 8ø13'E) on 19-21 August. A longertrip was made to Numan (9ø30'N, 12ø0YE), Kiri (9ø41YN,12ø00'E), Ganye (8ø20'N, 12ø05'E), Bauchi, and PanshanuPass (10ø06'N, 9ø12'E), 30 mileseast of Jos,where "nigeriae"again was found, from 23-30 August. Field studies at Zaria were concluded on 3 September. 1972 PAYNE: PARASITIC INDIGOBIRDS Ol • AFRICA

Figure 2. Mopane woodland. Hans Mererisky Nature Reserve.Transvaal. The top left of the dead tree was a call-site of Vidtta purF.rasccns in 1966 and 1967. Habitat of Lagonostictase,egala, L. rltodopareia, V. cltaly'beata,and V. pttrpttraacetts.

BREEDING BEHAVIOR OF INDIGOBIRDS

In most kinds of birds the social behavior centers around the nest and the territory and care of the young. As is the case with other brood parasites, the parasiticfinches do not nestnor do they feed their young. Field observationsprovide some information about the socialstructure and mating systems of thesebrood parasitesas well as the behavioralrelationships among different kinds of the indigobirdsliving in the same area. The main study area for behavioralobservations was Hans Merensky Nature Reserve, Transvaal, South Africa; the habitat and local birds have beendescribed previously (Gilliland, 1962; Payne, 1968c). Here two species of indigobirds(Vidua chalybeataand V. purpurascens)live in brushy,grassy vegetation(Figure 2) along a river with their firefinch hosts. They were regularlycensused for two years in this habitat. Birds in the reservewere un- disturbedby humanapproach to asclose as 60 feet. In other areasthe indigobirdsof the speciesV. chalybeatalive in villagesand when feeding or singing oftenpermit one to approachwithin 20 feet. In the field work singingmales were observed at distances of 60 feet or more to avoid disturbance. More than 500 hoursof field observationswere madeduring the breedingseason of the indigobirds. We foundthe indigobirdsby drivingor walkingin bushyhabitat until we heardor sawthe tiny black finchesperching on the topsof treesor bushes. 10 ORNITHOLOGICAL MONOGRAPHS NO. 11

Figure 3. Singing Vidua chalybeata at Ivlaun, Botswana, 19 April 1967.

The confidingbehavior of singingmales made possible prolonged observations at their singingperches or call-sites. The term "call-site"is an appropriate one for the songperches of the indigobirds. Ranger (1955: 70) earlier used it in describingthe specialplace where a honey-guide(Indicator sp.) sings "day after day of certainmonths year after year." Rangernoted that a call- site "may form a centre point round which the world of the honey-guide revolves. It has been shown that not only are the females attracted by the site-call, but at least in the case of Indicator indicator are other males, sub- adults and still youngerbirds attracted by it as well" (Ranger, 1955: 80). Just as a honey-guidehas a certain tree on which it perchesand calls or sings, and at which all mating occurs,the indigobirdsuse a specialperch as the point of focus of their breedingbehavior. Both the honey-guidesand the indigobirdsare brood parasites,and in both the call-sitetakes the place of a nest as the center of most breeding activity.

SOCIAL BEHAVIOR AND MATING SYSTEMS Male activity at the call-site.•In the breeding seasonthe indigobirdssing on their call-sites. The call-sitesare usually dead, leaflesstwigs on the tops of trees in open woodland,often at the edge of a clearing,fiver, or road. Each breedingmale spendsmore than half of each day singingon the same twig. Hour after hour, day after day, visitsto indigobirdcountry show the malessinging in the full sunlight,even when temperaturesexceed 100 ø F. The routine is broken occasionallywhen the birds fly to the groundand feed 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 11

TABLE 1

ACTIVITY OF MALE INDIGOBIRDS IN THE BREEDING SEASON

Activity of male at first encountereach day chalybeata purpurascens funerea Unidentified Singing on call-site, c• 90 38 19 11 Feeding 4 3 2 2 Intruding at active call-site, B 30 18 8 6 Chasing,c• -B 10 6 5 8 Perched quietly in tree, alone 5 1 3 3 Flew by 6 2 2 1 Drinking 3 0 0 0 on the fallen grassseeds near the call-site,when they chaseoff other males,or when they court and mate with the females at the site. Singingmales assume an erect postureduring song (Figure 3). Plumage is held closeto the body althoughthe head feathersare often erected. The bill is directedforward and is openedonly slightly,making it necessaryto look carefullyto seethat the bird is singing.During songthe headis turned from sideto sidewith every few phrases.The birds often shift positionand face in differentdirections on the perch. Occasionallyin mid-afternoona singingmale movesinto the shadeof the lower branchesand singsor rests. Althoughbirds in the studywere not individuallymarked, a few singing maleswere individuallyrecognizable; each was the sole occupantof its call- site over a period of days. A male Vidua chalybeataat Maun, Botswana,was recognizableby a patternof brown feathersevidently retained from the sparrow-likenon-breeding plumage. In the temporaryabsence of the mottledmale other malesin full breedingplumage occasionally perched on the site but did not sing. Each time, within a few minutesthe mottled bird returned,chased off the intruder,and then resumedits songon the site. A male V. chalybeata at Merenskyhad a protruding,bent, long wing covert that wasentirely resistant to effortsof the bird to preen the featherinto position. On all my visitsto the tree this bird was the singingoccupant. Another male at Merenskywas recognizablebecause it mimickedLagonosticta rhodopareia although the bird was a Vidua chalybeata,the only chalybeataheard to mimic this firefinch. The bird was recordedcontinuously over three days on the site; no other malesappeared. The constancyof eachof thesebirds at its call-sitesuggests that an individualmale generallyremains as the singingbird at a call-sitefor a period of severaldays or longer. The way in which males spendtheir time is evident in Table 1, which recordswhat each individual male was doing the first time I found it each day in the breedingseason. Birds chasingor flying over a certain site were countedonly once on each day; an effort was made not to count the same 12 ORNITHOLOGICAL MONOGRAPHS NO. 11 bird more than once. More than half of the encounters were with the con- spicuoussinging males on the call-sites. Nearly a fourth of the males met each day were birds other than the alpha (a) males or stud males at the sites;these intruders were repeatedlychased from the sitesby the dominant a-malesbut were sometimessuccessful in the end in establishingthemselves at the sites. Males spentrelatively little time feeding,drinking, or consorting with femalesaway from the call-sites. Aggressivebehavior at the call-site.--The singingmales vigorously defend their call-sitesand the immediatearea around the sites. Aggressiveencount- ers betweendefending a-males and the intruderswere commonearly in the breedingseason and at siteswhere dominant males had beencollected earlier. In the breedingseason in southernAfrica intrudersinvaded the call-siteareas 59 times during 159 site-hoursof observation;this figure includesonly in- trusionsby apparentlydifferent individual males at each site. Sometimes threesomesand foursomesof males in breedingplumage visited the sites together. A challengeto a singingmale usuallytakes the form of the intruderflying directly into the lower branchesof a call tree whether or not the a-male is present. An intruder bill-wipesand hops a few inchesat a time towardsthe dominant maleswhen present in the top of the tree. As an intruder ap- proacheswithin a few feet, the a-male flies toward the intruder, and the intruder then flies to the temporarilyunguarded call-site twig. The a-male then supplantsthe intruder,who either flies off or attemptsagain to gain the top twig after a seriesof supplantingattacks. As many as 60 supplanting attacksmay occur in a five-minuteperiod; seriesof supplantingsmay continue as long as 12 minutes. The intruder often faces away as he continues to approachthe a-male on the songtwig (Figure 4). By concealingits bill the intrudermay appeasethe a-male. Supplantingattacks are usuallysilent; sometimes,however, they are accompaniedby harshsongs or chattersusually givenby the a-male. The intrudersobserved never directly attacked or sup- plantedthe dominantmale at his call-site. Supplantingattacks are usually performedwith the dominant bird in a crouchedposition; both birds may bill-wipe. The a-male often extendsits head towards the intruder. Only once in my field observationswas the tail raised, althoughthis posturedoes appear often in captiveindigobirds, and it is a commoncomponent of agonisticbehavior in variouspasserines including the sparrows,Passerinae (Andrew, 1961: 337). Aerial chases are often involved in male-male encounters at the call-site. Chasingmales sometimes fluff out the body plumageas in the "bumblebee" flightdisplay of Euplectesa[er (Emlen, 1957). The chaseis sometimesslow. The two males usually were less than three feet apart in the chasesseen at Merensky,and oncethe a-malehit the intruder. Often after a seriesof chases 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFP, ICA 13

Figure 4. Two male Vidua cbalvbeata contesting an agave stalk used as a call-site, 34 miles east of Kisumu, Kenya. The dominant bird (above) faces the intruder. the intruder stays lower and faces away between supplanting attacks. the two birds land togetheron thc groundand feed with no apparentaggres- sion for severalminutes before they resume their chase. Singingmales also chase off other maleson nearbybushes upon sight. The a-males fly directly at the intruders. Usually the intruding bird flies and is pursued by the a-male; at other times the intruder remains on the perch as the a-male perchesand peersat him. On only two ob•-ervedencounters did one bird make physical contact, knocking the other from the perch and tumblingwith him to the ground. At the end of the chase,when the intruder finally retreats, the a-male chatters in flight while returning to the call-site, perches,and sings. The chasesat the call-site involve a considerableexpenditure of time and effort especiallyduring the periodof call-siteestablishment in early summer and at sites where the dominant bird had been removed and two males at- temptedat the sametime to establishthemselves. At Marble Hall, Transvaal, two males at a site where the a-male had been shot had prolongedseries of chasesand were in flight during most of the daylight minutesfor two full days. In one 80-minute period the two made 200 passesat the call tree. Chaseswere ovoid in shapewith the narrow end of the ovoid focusinglike a yo-yo at the call-sitetwig. When one bird broke off in the chasethe other perchedon the twig and assumedcontrol of the call-site. As the average 14 ORNITHOLOGICAL MONOGRAPHS NO. 11 diameterof the figure traced around the call-siteby the chasingmales was about200 feet, the seriesof chasescumulatively involved as muchas 120,000 feet in litfie over an hour, or more than 20 miles! Establishmento[ call-sitesin summer.--I madeeight censuses between 22 December 1966 and 2 January 1967 before any indigobirdshad laid their eggsat Merensky (as determinedfrom the ovariesof femalescollected). The two species(Vidua chalybeataand V. purpurascens)coexisting at Merensky eachestablished occupancy at their traditionalcall-sites in early summer.Of the 19 sitesused in the previousyear, 10 were occupiedduring late Decem- ber. Males in early summerspent only the morningson the sitesand often perchedquietly for many minutesrather than singingcontinuously as they do in the breedingseason. Groups of non-breedingmales were seentogether at times;Irwin (1952) has alsonoted flocks of malesearly in the season. Much supplantingand chasingbetween males contending for a site occurs in early summer. At a site observedfor five hours on 25 Decembertwo male V. chalybeataspent virtually all but 10 minutes(when they fed) in chases and supplantingattacks. Chasesessions lasted as long as 15 minutes. At the end of eachperiod of chasing,both birdsflew to the call-sitetree whereone male repeatedlysupplanted the other. During one period of 47 minutesthe birds made fully 333 supplantingattacks. At the end of the day the two malesflew to a nearbygrassy field and fed for 20 minutes,and on the followingday conflictcontinued. Other call-sitesremained unoccupied during thesecontests. The contestedsites were perhapsoptimal in somefeature, but field observationslater in the breedingseason showed no differencein the frequencyof matingsat sitesestablished in Decemberand at sitesoccupied in early January. The sametrees were usedas call-sitesin successiveyears, and at Merensky the same speciesgenerally came to occupycertain call-sitesfrom year to year. Figure 5 showsthat only four of the sitesoccupied in Decemberwere occupiedby the samespecies using the sametrees in the previousbreeding season,but a local changein site occupancyhad taken place by late January, sinceby that time all but two siteswere occupiedby the specieswhich had usedthe samesites in the previousyear (floodsprevented censushag the north bank in 1967). One site was shiftedwhen the earlier one was flooded, and anotherwas shiftedwhen the originalcall-site tree 200 feet away was felled. Two new sitesfound in 1967 were not usedin the previousyear. A tree in theapparenfiy unoccupied area in the middleof the censusstrip was occupied in 1966 only on Sundays,when the loud water pumpby it was silentand the singingbird couldbe heard. The other call-siteswere the samein both years, and a highdegree of traditionin the useof certaintrees as singingplaces and matingsites is evident. The treesused as call-sitesare not obviouslydifferent in structureor surroundinghabitat from many other treesin the samearea. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 15

TABLE 2

TIME SPENT BY BREEDING MALES AT THE CALL-SITE

Number of minutesper hour Status Number Species of male of hours Max. Min. Mean t.95• V. chalybeata alpha 32 59 3 42.7 4.78 replacement 30 48 0 20.9 6.77 V. purpurascens alpha 40 59 14 42.1 3.78 replacement 9 52 16 30.3 10.27 V. [unerea alpha 23 55 12 43.7 5.00 replacement 5 56 7 32.8 --

In the breedingseason indigobirds occasionally called from other trees,but thesewere abandonedwithin a few days, and I never saw a female visit a male excepton the traditionalcall-sites. Dominancerelations at the call-sites.--Defenseof the call-sitesis important in the socialbehavior of indigobirdsas all matingtakes place exclusively by the a-malesand exclusivelyat the call-sites.In all copulationsobserved the a-male,not a peripheralmale, was the one that mated. Dominantmales spendmuch time singingon the call-site,advertising their presenceto other indigobirds.Continuous periods of an hour or more of observationin southern and eastAfrica were madeto determinethe time spenton the siteby the singinga-males; the results are givenin Table2. Dominantmales of eachspecies spendan averageof 40 to 45 minuteseach hour on the call-sitesat all times,morning, midday, and afternoon. Males fed intensivelyfor a few minutes by the site beforethe beginningof singingat sunrise,and at sunsetthe malesdeparted from the sitesto feed and roost. The defenseby a-malesof the traditionalsinging and mating sitesprevent othermales from breedingthere, and thesenon-breeding birds quicklyassume occupancyof the siteswhen the a-malesare removed. The many intruding males which the a-males chase from the call-sites indicate the existence of many adult maleswithout sitesof their own. Four intrudingmale Vidua chalybeatacollected had testesof the same size (4 to 5 mm) as the a-males. To find the effectof the exclusivebehavior of the a-malesin keepingother malesaway from the site, and consequentlyupon restrictingthe number of malesbreeding in the area, I observedthe sequenceof eventsat more than 60 call-siteswhere I had shot the a-male. The importanceof the dominant male in excludingother males was readily evident, as in almost all sites anothermale took its placewithin a few hours. In one instanceas soonas eight minutes after the a-male V. chalybeatawas removed another male flew to the 16 ORNITHOLOGICAL MONOGRAPHS NO. II unoccupiedsite and sangfor nearlyan hour;he matedwith a visitingfemale on the sameday. Removalof a-malesat mostsites was followed by the establishmentof replacementmales usually within two hours. Continuedremoval of malespermitted a seriesof previouslynon-breeding males to occupythe call-sites.At Marble Hall nine male V. chalybeatain turn held one call-site and were collectedin eightdays. Throughoutthis time breedingmales re- mainedon other nearbycall-sites and themselveschased off other potential replacementmales. Smithers(1961) alsohas observed replacement of males taken from their call-sites. In theabsence of theoriginal a-male a seriesof potentialreplacement males often conflictedas each attemptedto establishitself at the site and the con- tendersspent much time chasing. Table 2 indicatesthat suchreplacement maleswere presenton the call-sitesa significantlysmaller proportion of the time than were the a-males. Considerablesinging time (about 30 percent) was lost in the establishmentof a new dominancesystem at a call-site. The proportionof time andenergy spent in establishingsocial order is dearly high. Throughthe aggressiveneglect of their call-sitesthe newly established maleswere lessoften availablefor mating. (The term "aggressiveneglect" describesthe decreasedbreeding success observed when a bird spendsits time in aggressivebehavior directed towards other birds rather than in be- haviordirectly related to its reproductivesuccess. Presumably a bird could avoid interferencewith its breedingif it toleratedother birds nearby. The term has usuallybeen appliedto interspecificaggressive behavior (Hutchin- son and MacArthur, 1959; Ripley, 1962), but the conceptof reproductive interferenceseems equally appropriatefor the result of aggressivebehavior directedtowards birds of the same species.) On severalinstances a female indigobirdvisiting a replacementmale at the site was ignoredas he flew out and chasedan intrudingmale. The averagenumber of visitsby femalesto the call-sites was less at sites where the a-male had been removed and was replacedby a new singingmale than at siteswhere the a-maleremained (Ta- ble 3). The differencewas statistically significant for V. chalybeata(p < .05). Evidentlythe femalescease coming to siteswhere the male is too oftenabsent or perhapsthe femalesare lessattracted to siteswhere there is lesssinging, althoughfemales do alsovisit activesites when malesare not immediately present.The significantdecrease in femalesvisiting the siteswhere the a- maleshad been collectedprovides direct evidenceof the functionof a stable socialsystem based on local dominancein maintainingthe breedingsystem of a population. Dispersionand territorialbehavior.--The call-sitesof breedingindigobirds are spacedout a few hundredyards apart, on the average,although locally the densityof singingmales may be considerablygreater. At Merenskythe averagedistance between each site and the nearestneighboring site was 485 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 17

TABLE 3

EFFECT OF THE STABILITY OF SOCIAL STRUCTURE UPON MATING OPPORTUNITIES IN INDIGOBIRDS

Number Numbero[ [emalevisits to o[ call-siteper hour Status of hours Species male observed Max. Min. Mean t.95•

V. chalybeata •pha 30 3 0 1.10 .22 replacement 30 2 0 .80 .04 V. purpurascens alpha 37 3 0 .68 .14 replacement 12 2 0 .58 .19 V. funerea •pha 23 8 0 1.17 .36 replacement 5 1 0 .60 - feet. Populationdensity was evidentlylower in areas such as Kisumu, Malindi, and Olorgesailie,Kenya, as only a singlecall-site was found in each of theseareas. Indigobirdsat Marble Hall usedone of the densestaggrega- tions of call-sitesthat I found. There four male Vidua chalybeatasang simul- taneouslyon adjacenttrees no morethan 300 feet from eachother. The adja- centmales had frequentchases back and forth midwaybetween their call-sites, but they spentabout the sameamount of time each day on their call-sitesas the birds at Mererisky;laying femaleswere taken at three of the sites and copulationwas seenat the fourth. I found densepopulations with more than four birds each within 200 feet of another and each on its call-site also on the Umtali-Penhalongaroad in easternRhodesia and a mile southof Monkey Bay, Malawi. In generalthe densityof singingmales appeared to parallelthe local abundanceof the firefinch hosts, but no good data are available on populationdensities of the firefinches. In areaswhere two or morespecies of indigobirdscoexist, the singingmales defendedtheir call-sitesagainst all maleindigobirds in breedingplumage regardlessof their species. The majority of intrusionsinto a call-site area, however,were by birds of the samespecies as the singingmale. During 40 hoursof observationat Mereriskyin 1967 I notednine conspecificintrusions at call-sitesof Vidua chalybeatawhereas V. purpurascenschallenged these birds threetimes. At the sitesof V. purpurascenssix conspecificchallenges and six challengesby male V. chalybeatawere observed. In all instances whenthe intrudingmales were of a speciesother than the singingmale they were chasedoff just as were conspecificmales. The high proportion(9 out of 24) of heterospecificchallenges at Merensky,where the two indigobirds are equallyabundant and havemany opportunitiesto interact,shows clearly 18 ORNITHOLOGICAL MONOGRAPHS NO. 11

that indigobird territorial behavior is not restrictedto the exclusionof males of a singlespecies, even thoughthe malesmay differ in songand in the color of gloss in thc black plumage. The censusdata for 1966 and 1967 at Merensky (Figure 5) permit com- 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 19 parisonof call-sitedispersion within and betweenthe two species,V. chalybeataand V. purpurascens.The mapsindicate a uniformpattern of disper- sionmuch as if theseindigobirds were all a singlespecies. In 1966 the males on 9 siteshad conspecificneighbors at the nearestcall-sites. Males at the other 10 siteshad nearestneighbors of the other species.Distances between conspecificneighbors ranged from 280 to 820 feet and averaged490 feet, while distancesbetween heterospecific nearest neighbors ranged from 280 to 780 feet and averaged479 feet; both the rangesand the meansof thesedis- tanceswere nearly identical. In 1967 the dispersionof singingmales in the breedingseason was similar. Male indigobirdsshot from the call-sitesfrom 5 through10 February1967 werequickly replaced on the sitesby othermales; in 16 instancesthe replacementmales were of the samespecies and in 2 instancesthe replacementswere of the otherspecies, suggesting a tendency for eachsite to be the traditional focusof breedingactivity of a singlespecies. A tendencyfor faithfulnessof speciesto a site was evidentalso in the similarityof the 1966 and 1967 patternsof site occupancy.Male V. chalybeataand male V. purpurascenseach usedprecisely the sametrees in 1967 that the samespecies did in 1966. No differencesin habitat between the sites of the specieswere evident at Merensky. At Zaria, Nigeria, in west Africa, the males of different forms of indigobirdswere uniformly spaced along Bee-eater Creek (Figures 6 and 7). Around the universityand in the nearbysurrounding villages and cultivatedfields only Vidua chalybeataoccurred (Figure 7). The differencein distributionof the indigobirdsevidently resulted from the occurrenceof only one firefinchin the populatedareas, L. senegala,host of V. chalybeata,whereas four speciesof firefincheslived by the creek (Fry, 1965; my field observation). Along the creek,as at Merensky,the indigobirds(Vidua chalybeataand the "camerunensis"and "wilsoni"forms of V. wilsoni) were spacedas if they were single species.Males collected from their sitesalong the creekwere usually replaced by conspecifics.Of 15 replacementsall but two were conspecifics(a shot "camerunensis"was replacedby a "nigeriae";this bird was replacedby a V. chalybeata;its own subsequentreplacement was another "camerunensis)." Challengesand chasesat the call-siteswere also seenbetween V. chalybeata and "camerunensis"(three instances) and between"camerunensis" and "wilson?' (one instance). The west African indigobirdsthus showedfidelity of the species(and of the "form" in the V. wilsonicomplex) to eachcall-site alongthe creekalthough no differencesin the habitatof the siteswere apparent and three foster speciesof firefincheswere caught along the creek in the same net. The dispersionof indigobirdcall-sites is similarto the dispersionresulting from the territorialityof many birds of northerntemperate regions. The 20 ORNITHOLOGICAL MONOGRAPHS NO 11

Fibre 6. Disturbed Guinea woodland along Bee-eater Creek near Zarla, Nigeria. Yearly burning clears the area in the dry season; a few scattered trees remain unfelled. Four species of timfinches and two species of indigobirds live here. behavioral mechanismsresulting in this pattern of use of spacemay differ, as somenorth temperatespecies exclude conspecifics by patrolling a beat over an expanse of suitable habitat (e.g., Johnson, 1963), but the uniform male dispersionresulting from the radial defenseof the call-site in the indigobirds is similar, and it does restrict the number of breeding males in a given area. Call-site behavior seemsto have no effect in restricting the number of breeding females (nor of the total number of eggs laid) in an area inasmuch as many femalesvisit a singlemale at his call-siteand as nearly all femalessam- pled during the breedingseason were laying (see p. 26). Interspecific dispersionin the indigobirds, evident in all localities where two or more speciesof indigobirdscoexist (e.g., Figures5, 7), appearsto be directly causedby a common set of nearly identical behavioral communication signalsand responsesto them shared among all of the indigobird taxa. Different kinds of indigobirdsare morphologicallyvery similar in size, shape, and color; at a distance all indigobird males look alike. The nonmimetic vocalizationsof different speciesof indigobirdsalso are similar in many features, and breedingmales respondin an agonisticmanner to playbacksof the recorded nonmimetic songsof other species,as is describedlater. Sharing of behavioralsignals which elicit interspecificterritorial behaviorin the indigo- 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 21

K•no

A.B.U, 0%

/70

• Kufena

I MILE Figure7. Locationof indigobirdcall-sites near AhmaduBello University(A.B.U.), Zaria, Nigeria: n = Vidua chalybeataneumanni, c = "camerunensis,"w • "wilsom•' (theselast two are regardedas formsof V. wilsoni). Small squaresindicate buildings or villages. birdsvery likely results from the fact that the different species are very closely related. Variousgroups of birdswhich exclude other species from their territories are similarto the indigobirdsin the exclusionof thosecognate species that 22 ORNITHOLOGICAL MONOGRAPHS NO. 11 share with themselvesa commonset of behavioralreleasers. Examples of mutual interspecificterritoriality of suchcognate species include species of tits (Parus: Dixon, 1961; Thielcke, 1969a), wrens (Thryothorus: Grant, 1966), grackles(Cassidix: Selander and Giller, 1961), meadowlarks(Stur- nella: Lanyon, 1957; Szijj, 1966), flycatchers(Empidonax: Stein, 1963; Johnson,1963), and earlier studiesreviewed by Johnson(1963). The behavioralmechanism effective in the maintenanceof interspecificterritoriality in all of theseappears to be the commonresponse to the commonbehavioral releasersof these species. The exclusionof other speciesof male indigobirdsalong with conspecifics from a matingsite by the studmale indigobirdmay increasethe probability of assortativemating; a female visitinga site where one male has replaced anotheris not likely to find a male of anotherspecies waiting for her, as both malesand femalesof a speciesare mostpersistent in hauntinga call-site. In this manner,mutual spacingmay have a minor effect on mating structurein indigobirdpopulations. Exclusion of males of other speciesof indigobirds may also decreasethe chancesof interferenceat the mating site. The importancealso of interspecificspacing as an adaptationrelated to ecologicalcompetition (Orians and Willson, 1964) appearsto be minor in the indigobirds,especially when it is appreciatedthat the defendedsite is the matingsite, not the feedingarea. Ecologicaloverlap in food resourcesoccurs --the indigobirdsand also their hostsall feed on the same speciesof grass seedat MereriskyReserve, at Lusitu River, and at Monkey Bay as determinedby field observationsand analysisof stomachcontents. However, grass seed(e.g., Eleusinaindica, one of the identifiedshared foods at Lusitu River and at MonkeyBay) doesnot appearto be in shortsupply during the breeding season,and it is abundantin the weedy native cultivatedfields. At MerenskyReserve both Vidua chalybeataand V. purpurascenshad in their cropsseeds of Setariasp., Urochloasp., andDigitaria sp., commongrasses in tropical and subtropicalgrassland and brush country (Walter, 1964: 566- 571). Fallen grassseed was readily evident in summertimewhen I casually inspectedthe groundin the grassy,brushy areas where indigobirdsand firefincheswere feedingalone and in mixed speciesflocks. The abundanceof this food is probablyrelated to the occurrencetogether of severalspecies of firefinchesall eatingthe samefood at areassuch as Zaria (Fry, 1965, 1966; also my field observations).Indeed, in his review of ecologicaloverlap of birdsin eastAfrica, Moreau (1948) found the great majority of speciesto be clearly different in their feeding stationsor food exceptfor small finches, which all fed on the seasonallyvery abundantgrass seeds. In my observations,both the indigobirdspecies and theirhosts also fed togetheron emerg- ingtermite swarms during the earlysummer rains. In theirbehavior, the indigobirdsseem to be under no shortageof food inasmuchas breedingmales 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 23

TABLE 4 FREQUENCYDISTRIBUTION OF ENCOUNTERSWITH FEMALEINDIGOBIRDS IN THE BREEDING SEASON

Away [rom call-site At call-sites alone with • with i• group V. chalybeata 131 22 7 1 1 V. [unerea 54 1 2 - - V. purpurascens 63 8 6 - - unidentified 18 7 1 - - Totals 266 59 are able to advertizeand defendtheir call-sitesand to mate repeatedlywith no morethan 10 percentof theirtime each day spent in searchingfor food. Furthermore,ecological overlap in termsof theirhost exploitation is insignifi- cant;as indicated by speciesdifferences in songmimicry, the coexisting indigobirdsdo not competefor hostnests but parasitizedifferent species of firefinches.The onlylocal resource which is at all bothobviously restricted and in mutualdemand is the traditionalcall-site, one tree or bushout of hundreds in the cruisingrange of an individualindigobird. In summarythe dispersion pattern of differentspecies of indigobirdsseems to be a directresult of theirsharing of a commonset of homologousvisual andauditory communication signals within the species complex. The lack of evolutionarydivergence may existbecause the presenceof numbersof one indigobirdspecies does not affectthe numbersor chanceof survivalof another. The mutuallyexclusive areas around the call-sitesare highlyvariable andare compressed into a smallarea in regionsof highfirefinch numbers, as aroundMonkey Bay and Marble Hall, andprobably habitat and the density of the hostspecies are moreimportant than territorialityin determiningthe numberof indigobirdswhich can live in an area. Female activity.-•Females were not often seen away from the call-sites, thoughthey spendlittle time at thesesites. The numbersof first encounters eachday in the breedingseason with eachfemale (I tried not to countthe samebird twice on any day) are listed from the observationsin southern Africa, in Table 4. Of the total 325 encounterswith females,266 weremade at the call-sites;this is not becausethe femalesspend much time at the call- sites,but becauseI did myself. The femalesspend most of their own time awayfrom the call-site,feeding and lookingfor host nests.Females away from the call-siteusually were alone, less often with firefinchesor other femaleindigobirds. Only oncedid I seea femalewith a male awayfrom a call-site. Femalesand malesoccasionally occurred together in groups,sometimes with other finches,when they were feeding. 24 ORNITHOLOGICAL MONOGRAPHS NO. 11

TABLE $

TIME OF LAYING OF FEMALE INDIGOBRRDS1

Time of day Stage of laying cycle • 0800 0800-1000 1000--1400 1400-1600 H 1600

Ovulation 0 1 5 3 13 Yolk in duct 0 4 8 0 5 Soft-shelledegg 0 0 1 0 12 Hard-shelled egg 3 3 3 1 0 x Includes all V. chalybeata, V. purpurascens, and V. tunerea taken in breeding condition in southern Africa.

Femalesvisit the call-siteprecisely at the momentof ovulation. These femaleshave a large yolky ovarianfollicle partially engulfedby the infundib- ulum of the oviduct (Table 5). All of the females that solicited and mated at the call-site,and that were collected,were ovulating. Femalesalso visit the call-sitesduring other stagesof the laying cycle; at these times they are not receptiveto the courtshipdisplays of the males. Non-ovulatingfemales occasionallyfeed with the males after display,but more often fly from the site. Collectionsof femalesof all three speciesin southernAfrica showed that layingand ovulationmay occurthroughout the day, thoughby late afternoon all femaleswith hard-shelledeggs have laid them. The persistenceof an a-male in singingand attendanceat the call-sitemakes him availableto the females,which may likewiselay and ovulateat nearly any time of day; I observedcopulations from 08:00 through 17:40 hours or shortlybefore sunset. Breedingactivity o/ young males and /emales.-•Male viduinesgenerally have beenthought to remainin a brown, sparrowyplumage until their second year (Lynes, 1924: 678; Friedmann, 1960: 155). However, few sparrowy indigobirdsin any populationsampled in the presentstudy were males,and nearly all males collectedwere in breedingplumage. Since indigobirdsare small (12-14 g), probably their life expectancyis short and a large proportion of birds (probably more than 30 percent) in the breedingseason may be no more than a year old. Yet lessthan 10 pereentof all malesobserved in this studywere in fact in sparrowyplumage during the breedingseason. While someyoung males apparently do not assumebreeding plumage in their first year, my observationsindicate that many others do attain complete breedingplumage and someof thesemay sing and establishthemselves on call-sites. Males in sparrowyplumage and without black feathersin the breeding seasonall had small testesranging from 1 x 1 to 2 x 1 mm; no off-season malescollected had testeslarger than 1 x 1 mm. Four sparrowymales gave 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 25

TABLE 6

RELATION OF AGE TO BREEDING ACTIVITY IN MALES

Percent o/pneumatization Status ofskull Species of male 100-80 80-60 60-40 40 V. chalybeata alpha 6 12 4 7 replacement 3 2 9 13 V. funerea alpha 2 4 2 0 replacement 1 3 1 5 V. purpurascens alpha 4 12 2 5 replacement 5 6 5 3 long, completesongs at the call-siteswhere malesin breedingplumage had earlier been shot. No sparrowymales sang at undisturbedsites. Males with only a few sparrowyfeathers in their otherwisecomplete adult breedingplumage had large testesin the size range of the breedingmales (4.5 x 3 to 6 x 4.5 mm). All of the sparrowymales evidentlywere young reared in the previousbreeding season as the pneumatizedarea of the skull in dorsalaspect was no more than 25 percent, less than in most singingmales in breeding plumage.Viduines generally retain unpneumatized areas in the skull throughout life (Chapin, 1954: 580; Payne, 1967: 367); the degreeof pneumatization is presumablysomewhat greater in the older birds. In an adult female Vidua purpurascenskept in captivityfor 23 monthsbefore it died, the skull was 85 percentpneumatized. Thus in an indigobirdof a minimum age of two and onehalf years,the skullwas still incompletelypneumatized, though more completelyso than in mostbirds taken in the field. Two malesof this species that died after nearly three yearsin captivityalso had skulls80-90 percent pneumatized.It seemsmost likely that the indigobirdswith a relativelylarge proportionof the skull unpneumatizedare the youngerbirds and that some pneumatizationoccurs through the successiveyears of life. I would guess that the birdswith lessthan half of the skull pneumatizedare individualsno more than one or two years of age. Breedingmales at the call-siteswere, on the average,judging from their skulls,the older birds. The a-malesin southernand easternAfrica generally had skulls more than 60 percent pneumatized,whereas replacement males had skullsless completelypneumatized (Table 6); the differenceis significant (p < .01 ). The occurrenceof somea-males with little skull pneumatization suggeststhat somemales not only assumebreeding plumage but also 26 ORNITHOLOGICAL MONOGRAPHS NO. 11 may breed in their first year. The presencein indigobirdpopulations (see pp. 15-16) of largenumbers of nonbreedingmales in breedingcondition, both yearlingsand older males,provides a surplusof adultswhich may breed at the traditionalcall-sites only when the a-malesare killed or are otherwise relieved of their dominance. Female indigobirdsappear to breed regularlyin the first year, as no cor- relationwas evidentbetween laying activity and the degreeof pneumatization of the skull. Of 302 indigobirdsI collectedin the field only one had a completely pneumatizedskull. This bird (a female V. chalybeata)was laying, and thusthe oldestindividuals appear to breed. Two breedingfemales had lessthan 10 percentof the skull pneumatized,indicating that femalesbreed in their first year. To find whetherthe youngerfemales tend to beginlaying at a later date in the breedingseason than the older females(those with more completely pneumatizedskulls) I tabulatedthe percentageof skull pneumatizationand the presenceor absenceof visiblepost-ovulatory follicles and of large,yolky, growingyellow ovarian follicles in birdstaken in the first monthof breeding in Transvaal,Rhodesia, and Malawi. Therewas no suggestionof a difference in the time of the first layingof the seasonin the youngerand olderfemales. Reproductivematurity in femaleindigobirds therefore appears to bereached in the first year of life. Mating systerns.--Atevery call-site where I spentmuch time watchingthe indigobirds,I saw severalvisits by female indigobirdsto the singingmale eachday. The male courtednearly all of thesefemales. An individualfemale may perhaps visit a call-site more than once a day, but it was clear that several different females used the services of one male. From one active call- site of Vidua chalybeataat Marble Hall I shot four females on one day (6 March 1966). I shot each female just after she was courted by the resident male. All four femaleswere laying and had an egg in the oviduct. Thus, at leastfour breedingfemales were usinga singlecall-site (and its male). For all of the speciesof indigobirdsI collectedtwo or more breedingfemales from the call-sitesof singlemales, and at all call-siteswhere I spentsufficient time watchingI saw more than one female visiting. Clearly, the indigobirds are polygynous. The only placewhere females associate with malesregularly appears to be at the call-site. The femalesvisit the call-sitesfor mating and also sometimes when they are not ovulating,perhaps to maintaintheir familiaritywith the call-site. Each visit lastsonly a minute or so (unlessthe female remainsto feed). No pairing behaviormore permanentthan the proximity of female and male at the time of courtshipand copulationwas evidentin the indigobirds. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 27

The numberof indigobirdsusing each call-site may be on the orderof 10 or 12 (including6 or morefemales), judging from the continuingreplacement of singingmales and the continuingvisits of femalesafter otherswere collected at each site. At a call-site at Marble Hall six female V. chalybeata were shot in one week and all were laying. I regularlyexamined, at least once each day, the other call-sitesof this speciesat Marble Hall to find whetherthe birds showing up at thecall-site where I wascollecting were birds drawnaway from the alreadyestablished call-sites. The malesremained on theseother call-sitesand were themselvesharrassed at timesby other intruding adult malesattempting to take over their singingperches. Females also continuedto visit theseother call-sitesregularly. The call-sitewhere I was collectinghad at leastsix females,and ninemales were shot here duringnine days,the last eighteach having replaced an earliersinging male. As far as I know, indigobirdscontinued to use the call-siteafter thesefirst 15 were shot. I havenot collectedany sitesexhaustively until malesand femalesno longercame, but at severalsites I removedthree or four males,and at all call-siteswhere I shotthree or more females(this was done for all species), more females continued to come. At an isolated call-site of V. purpurascens near Sigor,Kenya, where there were no other call-sitesof this specieswithin at least a half mile, severalmales and femalesgathered near the call-site tree each day in early June, 1967. On 1 June I saw 12 birds feedingtogether. A male flew from the flock to the call-site,sang, and was shot;later the same day two moremales and a femalewere shotfrom the samespot. The following day I sawnine birds in the flock by the tree. The flock was seenby the site each afternoonthrough 4 June, when only two birds (a male and a female) were left, as I had collectedfour additional males and four more females(two of themlaying). The flock sizeon eachday wasclosely accounted for by the numberof birdsseen on the previousday minusthe numbershot. The number of birds seen indicated that few additional birds were recruited to this site, and about six males and six females were using the call-site, thoughat any one time only one male was the stud. It would be of interestto know whetherindividual females regularly visit and mate with a singlemale on one call-siteor if more often they mate with severalmales during a breedingseason. No quantitativedata are available on the constancyof a female to a call-site,and no birds were color-banded in this study. A degreeof fidelity of a female to a call-siteis suggestedby the appearanceof a female red-billedV. chalybeataamauropteryx at a call- site of a male V. purpurascens(both male and female of this secondform are white-billed) at MereriskyReserve. In two yearsof field work involving 160 observationsof female V. chalybeatathe only times I saw a female of this indigobirdvisiting the call-siteof anotherspecies were three observations of femaleV. c. amauropteryxat thissite on 18 and 19 March 1966. Probably 28 ORNITHOLOGICAL MONOGRAPHS NO. 11 thesethree involveda singlefemale, as on 19 March the femaleflew to the site in early afternoon,only 10 minutesfrom the time of the visit on the previousday. I disturbedit, and the femaleflew. Again in 15 minutesa femaleamauropteryx flew to the site and was immediatelycollected; it was laying. I watchedthe call-sitefor the next four days,but no more red-billed femaleswere seento visit it. The timing of the visits at the samehour on bothdays and the absenceof furthervisits after the femalewas shot suggest that a singlefemale made all threevisits, and this in turn suggestsa degree of site (and mate) constancyof female indigobirds. Whetheror not they do so regularly,females do sometimesvisit more than one call-site and mate with more than one male. At Marble Hall on 22 Janu- ary 1967 I sawa femalevisiting two sites. The bird, a femaleV. chalybeata, flew to an Acacia giraffaetree usedas a call-siteby one male, but the stud was feedingin the grass. The female then flew 200 feet from this tree to anothercall-site where anothermale was singing. She mated at the second site. Perhapsfemales regularly visit more than one call-site. The matingsystem of indigobirdsmay be characterizedas both polygynous and promiscuous.Polygyny characterizes the behaviorof birds in which some malesmate with two or more femalesin a short time (such as a breeding season) (Lack, 1968; Verner and Willson, 1969). More than half of the call-siteswhere observationsand collectionswere made over a few days were the commonmating sitesof two or more females,and thereforethe indigobirds are highly polygynouscompared with many speciesregarded as polygynous,where fewer than half of the maleshave more than one mate (Verner and Willson, 1969: 6, Table 1). The term promiscuityhas been usedin two senses.The dictionarydefinition of promiscuity(in Webster'sNew Inter- nationalDictionary, Second Edition) as "indiscriminatemating," however, is probablynot an appropriatedescription of the matingsystem of any bird, as somedegree of mate or site preferenceis likely in all birds. In the senseof bird ecologists,promiscuity occurs when "copulationby both sexes[is] likely to involve more than one member of the other" (Verner and Willson, 1969: 6); that is, in a promiscuouspopulation males are polygynousand at the sametime femalesare polyandrous.In general,birds with this sort of prom- iscuitylack a lastingpair-bond between the male and female (Lack, 1968; Verner and Willson, 1969). The matingsystem of the indigobirds(all species) is promiscuousin the secondusage of the term; the male spendsnearly all his time on the call-site and mates with several females, while other femalesare searchingalone for the host nests. A female seemsto come to the call-siterather than to any one male. The term "promiscuouspolygyny" best describesthe mating systemof the indigobirds,where each breeding male hasmore matesthan doeseach female, and wherethe only functionof the male is in mating. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 29

COURTSHIP BEHAVIOR Courtship behavior and displaysof the indigobirdx.-•Severalobservers have describedbriefly the courtshipdisplay of the indigobirdsas an aerial danceby the male abovethe female,a displaysimilar in mostspecies of the Viduinae (Friedmann,1960; Nicolai, 1964). No speciesdifferences in display have been known amongthe differentindigobirds. In my field work I observedmore than 200 courtshipdisplays (including 21 copulationsin all speciesof indigobirdsrecognized here), and the displaysof courtingmales and femalesappeared to be identicalin all species.Although the visualcom- ponentsof displaywere nearly identical, some differences in the songof the courtingmale were recorded;these are discussedin the sectionon vocalizafions. The sequencesof events associatedwith mating were noted in greatest detail at Merensky Reserve,Transvaal, in the forms Vidua chalybeataand V. purpuraxcens. Becausethese and all other forms of indigobirds seen in the field had the samecomponents of displayas well as very similarsequences of thesecomponents, the followingdescription of courtshipbehavior applies to indigobirdsin general. More than half of the courtshipsequences observed were terminatedbefore the birds had an opportunityto completecopulation and post-copulatorybehavior, as the femaleswere usuallyshot for morpho- logicalcomparison and for microscopicexamination of their ovariesfor signs of breeding activity. The componentsof courtshipbehavior include both stereotypedpostures and motion patterns(i.e., displays) not usually seen in other contextsand also events (such as perching) which are not stereotypednor restrictedto the time and place of courtship. The most common sequenceof these behavioral componentsis the one in which they are listed, as follows: 1. Female llies to the call-site. No calls or obviousdisplays are givenby the female as she first appearsin flight headingdirectly to the call-siteor as she approachesit. 2. Greetingflight. The male, when he seesthe female approaching,leaves the call-site,flies rapidly towardsher, and then accompaniesher in flight back to the call-site. The male flies to the female when she is within about 100 feet of the call-site,although this varies with the distanceof the female at the time the male first seesher. The male often chattersrapidly in the greeting flight. Sometimes a femaleapproaches the call-sitefrom behindthe male and the male doesnot see her or displayuntil shehas perchedon the site. When this happensthe greeting flight is omitted from the courtshipsequence. 3. Female perchexon the call-site. In almost all of my observations,when a female indigobirdwas first seenwithin 100 feet of a call-siteshe was flying to the call-site,where she perchedon the bare twigs near the singingperch of the male. Perching was not accompaniedby any ritualized display, as the female 30 ORNITHOLOGICAL MONOGRAPHS NO 11

Figure 8. Courtship display of Vidtta [unerea at Tzaneen, Transvaal, 11 February 1967. often remained in an upright posture and was quiet, though sometimesshe crouched and solicited copulation at once. 4. Solicitation. Receptivefemales crouch at the call-site,droop and flutter the wings, hold the tail at a slightly depressedlevel with no conspicuousmovement, retract the neck, and tilt the head upwards. The flattened solicitation posture of the indigobirdsis very similar to the female Pin-tailed Whydah (Vidlta macroura) illustrated by Nicolai (1964: 179). Solicitation may occur at the moment the female first lands on the call-site, with no display given first by the male. More often, however, the female does not solicit until the male hovers in front of her. 5. Hover. The most conspicuousdisplay in indigobirdcourtship is a hovering display given by the male to the female on the call-site. The hover display consistsof the male dancing in the air in front of, or sometimesover, the female. He alternately flaps his wings and folds them at his sides,and as he does this he rises and falls in the air by the female. The male bobs up and down, bouncingas if on a string, for about one to four seconds. His posture as determinedfrom photographs (Figures 8, 9) is nearly vertical, his head is bowed toward the breast, and his brightly-coloredfeet are exposedagainst the blackish plumage. The feathers of his head are somewhat fluffed. In the display the bill and feet of the male are set off in sharp contrast against the glossyblack plumage, as seen from the view- point of the female. The male usually singsa complex nonmimetic song during the hovering flight. The male indigobird usually gives this display in flight without first perching on the call-site at the end of his race to the site with the female in the greeting 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 31

Figure 9. Vidua chalybeatacourting a female at his call-site.Marble Hall, Transvaal, 21 January 1967. The female. one of two present.is unreceptiveand faces away. flight. When he does not give a greeting flight. he hovers as soon as he sees the female perched on the call-site. Audiospectrographsof displaying males show the sounds of the wingbeats of hovering males. A number of seriesof wingbeats,each separatedby a period of about 140 msecof silence(Audiospectrograph 3½)), shows that during hovering the wings are moved several times in a period of 100 msec; the periods of silence are the times when the male folds his wings. The temporal componentsof the wingbeat displayas shown on the audiospectrographswere comparedfor the different species of indigobirds and no species differences were evident, 6. Copulation. If the female solicits, the male mounts within a second or two. Copulation may occur when the female solicits immediately at the call-site, with no intervening greeting flight or hover display, as on one occasionwhen I saw a female fly to the site and solicit before the singing male had even seen her. As soon as he did see her, he flew to her and mounted immediately. More often the male hoversin front of the female and then flies over her, hovering,and drops onto her back and mounts. Copulation is brief, no longer than three seconds. Usually the copulating birds are silent: no vocalizations were heard on 13 occasionswhen I was closeenough to hear the wingbeatsof the displayingmale and the song he gave while hovering. When the female does not solicit, the male does not usually attempt to mount, though at times he may. If the female does not solicit,the male may perchon a twig a few feet from her and then resumesinging, or he may perch and hover again, or he may fly to the ground. 7. Flight to the ground. After courtship(with or without copulation) the male flies to the ground below the ca!l-site. Here he mimics the songsand calls 32 ORNITHOLOGICAL MONOGRAPHS NO. 11

TABLE 7

COURTSHIP BEHAVIOR IN THE INDIGOBIRDS

V. V. V. Behavioral sequenceto termination• chalybeata purpurascens funerea a. male ignores female 7 4 2 b. male chases female 11 8 0 c. male hovers and perches 7 5 9 d. male greets and hovers 4 2 4 e. male hovers, pair flies to grass 15 8 7 f. male greets and hovers, pair flies to grass 13 2 1 g. no hover, male flies to grassand mimics 4 2 2 h. male hovers, male flies to grass (female shot) 30 8 11 i. no courtshipbehavior (disturbedby other birds) 6 2 1 Sequenceis terminated when one bird is shot or when one flies away. of a firefinch, and the femalejoins him. The two feed togetherin the grasswhile the male continueshis mimetic songfor as long as 14 minutes. At the end of this ritual the female flies off alone and the male resumessinging on the call-site. The flight to the ground and mimetic song were noted in 96 visits. Mimetic songwas a consistentaspect of this behavior,as on everyoccasion when the male flew to the groundafter courtshiphe sang. Occasionallythe male begansinging while he was flying from the call-site to the ground; this happenedat Mann, Botswana,where there appearedto be no grassseeds below the call-siteof a male V. chalybeata,and he flew to a grassypatch about 100 feet from the call-site. On all 53 occasionswhen the female was able to do so (when she was not collected), she flew down from the call-site and joined the calling male on the ground. On 5 other instancesthe female was wary of me and she flew to the ground first; then the male joined her and gave the mimetic songs. Most pairs remained on the ground for about six minutes,the male singingand feeding and the female feeding. I saw no further matingbehavior in birds on the ground,although Fried- mann (1960: 83) reportsone instanceof courtshipdisplay on the ground. Ritualized mimetic singing on the ground followed all undisturbedcopulations, and it also occurredafter the unsuccessfulhover display,so it is better regarded as a post-courtshipdisplay than a post-copulatorydisplay. The male indigobirds may also go to the grassand sing when a female avoidslanding on the call-site but rather goesat onceto the groundbelow it, or when the mating of the two on the call-site itself is disturbedby other birds. This basic sequenceof eventsin the courtshipbehavior of the indigobirds may be altered,and I have listed the main variationsin Table 7. On 13 visitsof femalesto the call-sitethe male ignoredthe female. Severalof these instancesinvolved males which had just been chasingintruding males, and all of them involvedeither disturbanceby other birds (someof them other kinds of viduines) or a low sexualresponsiveness of the malesin the first week of the breedingseason. Another 19 instancesended with male and femaleflying from the site,usually after the male approachedthe femalein greeting 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 33 flight but, insteadof stopping,the femalepassed on over the call-site. One male-chases-female encounter was seen when a second male arrived at the call-site a secondafter the female did; the a-male chasedoff both of the visitors.An interferingbird of anotherspecies broke up a courtshipsequence whena malePin-tailed Whydah (Vidua macroura)was hovering by a female indigobirdat a call-sitetree that he sharedwith a male V. chalybeata. The femaleflew as the male whydahhovered, and the male indigobirdfollowed the femaleto the grassand sangthe mimeticsong. On two otherinstances two female indigobirdsvisited the call-site of this same male at the same moment,and the male displayedto one of them and ignoredthe second. On a few occasionsin the field and (more regularly) in aviary birds the male indigobird did not hover when the female flew to the call-site, but rather he displayedtowards the femalewhile remainingon his singingperch. My wife Karenrecorded one of thesehappenings at MarbleHall on 23 Janu- ary 1967 at a call-siteof Vidua chalybeataby a fish pond. The male had beenharried on his perchjust a minuteearlier by a divingdrongo (Dicrurus adsimilis),and when a femaleindigobird flew to his call-site,he faced her, bent his head down with the bill nearly touchingthe breast, droopedhis wingsand quiveredthem slightly,and gavea nonmimeticsong. The posture was similarto that of a hoveringmale exceptthat the male seemedmuch lessemotionally tense and he did not bob up and down as hoveringmales do. Possiblyin this instancehis drive to court the femalewas diminishedby therecent attack by the drongo.In the aviarythis behavior of perchedmales occurredmuch more commonly,perhaps on a tenth of the times when the femaleflew to a call-site.Inasmuch as the aviaryindigobirds were not breed~ ing, the absenceof the aerial componentof hoveringsuggests, as doesthe displayin the harrassedwild male,that thisperching display is simplya low- intensityversion of the hoverdisplay. The behaviorof the Marble Hall bird wasfollowed in a few secondsby themale hopping to anotherperch and then hoveringin the air, but ashe bouncedin frontof her the femaleflew away. The eventsleading up to matingoccur very rapidlyin the indigobirds; scarcelyfive secondsmay passbetween the time of the flight of the female towardthe call-site and the moment of copulation.In sucha rapidsequence of events there is little time for the male and female to balance out or resolve anyopposing drives associated with mating, such as attacking or fleeing.The mannerof behaviorof the maleis mostassertive, and the aerialcomponents of display(greeting flight, hover) may havesome significance as agonistic actions,as the male indigobirdsdrive away intrudingmales from the call-site with aerial flight. In one pair of our captiveV. chalybeata(as told to me by GloriaSullivan, my assistantin Oklahoma)a physicallyrather weak femaleflew to a call-sitein the aviaryjust as the malewas flying towards it; thefemale perched, the male flew above her and hovered, then as he dropped 34 ORNITHOLOGICAL MONOGRAPHS NO. 11 ontoher and touchedher the femaledropped, dead, to the ground. The behavior of the wild femaleindigobird in matingis essentiallysubmissive; even with no evidentmale displayshe may crouchand solicitmating at the call- siteat once,if the male is on the site. The eventsfollowing courtship (flight to the ground,song mimicry, feeding) contrastsharply with the precourtship behavior,as theselater eventsoccupy several minutes and the male and female seemto be amicableand relaxedwhile they are on the ground. It seems to me that the indigobirdsmay resolvethe conflictingdrives associated with the hecticprecopulatory behavior during the leisurelytime they spendoff the main site of activity and togetherin the grass. The prolongedmimetic song in the grassmay also causethe female to ovulateon the spotif shehas not just alreadydone so; this may ensurethat her next eggwill be fertilizedby that male. The ideathat hearingthe song of the foster species(whether given by the foster male itself or by the viduinemimic) may promoteovulation was suggestedto me by my only two observationsof viduineovulation in captivity.First, a femaleV. chalybeata ovulatedwhen a male L. senegalafirefinch was building a nest in our house,even thoughno male indigobirdswere in the houseat the time (p. 49). Second,in a seriesof experimentsin whichI playedback the recorded songsof the Melba Finch (Pytilia melba), its host,to a captivefemale Para- diseWhydah (Vidua paradisaea),the femalewhydah laid an egg after a few weeksof hearingthe songsof its fosterspecies. The femalewhydah had been completelyisolated visually and acousticallyfrom all other birdsfor six weeks beforeshe laid an eggon the floor of her aviary on 9 January1971. Shehad heard only the songsof Pytilia melba (includingsongs mimicked by male ParadiseWhydahs) and alsoof P. phoenicopterafor a total of eightminutes a day (at intervalsno more frequentthan alternatedays) for 11 days since 13 December 1970. The only circumstancein commonin these two ovula- tionsof captivefemales was that the femaleviduine was exposed to the songs of its foster species,so it seemslikely that hearing the mimetic song itself may affectthe ovariandevelopment and ovulationin the femaleviduines in the wild, includingthe time when the femalelistens to severalminutes of the male mimickingits foster speciesin the grassby the call-site. The sequenceof the eventsof courtshipbehavior are very similaramong all of the differentkinds of indigobirds.I recordedsequences for 175 visits by femalesto the call-sites,at timeswhen their maleswere on the sites,in the threespecies in southernAfrica (Vidua chalybeata,V. purpurascens,V. lunerea); a smallernumber of observationsof the other kinds of indigobirds observedin east and west Africa showedprecisely the samesequences and variations. The sequencesof eventsin courtshipbehavior are recordedin Table 7. The variationsof the behavioralsequences are numericallyabout the same in the different kinds of indigobirds. In all species,any one of the 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 35 elementsof courtshipbehavior may be omitted from a sequence,and this variability (in additionto the similarityof the eventsin the more common sequencein all species)indicates a lack of speciesspecificity in the courtship displaysof the indigobirds.That is to say, differentspecies of indigobirdsall have the same displaysand give them in the same sequence. Behavior o[ male indigobirds directed towards other kinds o[ birds.-- Breedingmale indigobirdscourt not only female indigobirds,but they also fly to and hover over a variety of other smallbirds that may use the call-site as a casual perch, both females of other kinds of viduines and also many unrelatedsmall birds. On five occasionsI saw male indigobirdshover to sparrowy-plumagedVidua macrouraand then fly to the grass,and on four of thesethe male gavemimetic songs. Male indigobirdsalso displayed to female ParadiseWhydahs (Vidua paradisaea). I saw no visitsby other species of non-indigobirdviduine females, and the scarcityof visitsby non-indigobird viduinefemales underscores the specialimportance of the indigobirdcall-site as a specificcenter of activityof the indigobirds.Occasionally male viduines other than indigobirdsapproached an indigobirdcall-site, and on eachoccasion thesemales of the long-tailedforms (V. paradisaea,V. macroura,and alsothe ResplendentWhydah, V. hypocherina)were ignoredor were chased from the call-siteby the male indigobird. Finchesand sparrowsother than the viduinesoccasionally perch on a call- site,and when they do theymay be courtedby the studmale indigobird. On 14 of 16 occasionswhen I recordeda singleCut-throat Finch (Amadina iasciata) at a call-sitethe intruder was courted,and on 8 of thesethe male indigobirdgreeted, hovered, flew to the grass,and also mimickedfirefinch sounds. Other speciescourted at the site were White-wingedWidows, Euplectesalbonotatus (female and juvenileonly, 2 times), Red Bishops,E. orix (female and juvenileonly, 4 times), Grey-headedSparrows, Passer griseus (7 times in 22 visits), Red-billed Quelea, Quelea quelea (2 in 12 visits), Yellow-eye Canaries,$erinus mozambicus(4 in 12 visits), Bronze Mannikin,Lonchura cucullata (juvenile, 2 in 3 visits), a Scaly-feathered Finch, $poropipessquamiirons (1, 1 visit), and other unidentifiedbrownish ploceids(2 in 6 visits). To noneof theselast birds did the maleindigobird givea mimeticcall. The actof flyingto the smallbirds and hovering in front of themmay have an elementof attackor threat;some birds flew, and males oftensupplant other small birds at the call-site,thereby reducing the chances of interferencein the mating of the indigobirds. Birdsother than finches and sparrows are also displayed to by maleindigobirds. Brownish-streakedfemale and youngCinnyricinclus leucogaster, a starlingof 40-50 g, were courtedon 14 of 17 encountersrecorded; once the maleindigobird afterwards flew to the grass.No instancesof courtshipdisplay directedto purple-and-whitemale starlingsof this specieswere seen. 36 ORNITHOLOGICAL MONOGRAPHS NO. 11

Male indigobirdsdisplayed to all of the followingspecies, some of them not at all colored,marked, or proportionedlike female indigobirds:bee-eaters (Merops apiaster, M. boehmi), barbets (Pogoniuluschrysoconus), woodpeckers ( Dendropicosfuscescens ), babblers ( Turdoides}ardinei) , drongos (Dicrurus adsimilis), starlings (Lamprocolius chalybeus), and sunbirds (Nectariniasenegalensis, N. talatala). All membersof thesespecies courted were lone birds;groups of visitorswere not courted. The blackishmale of the sunbird,drongo, and starling were perhaps superficially like maleindigo- birdsin plumage,but it shouldbe notedhere that no real male indigobird intruderswere ever greetedwith courtshipdisplay. The total dissimilarityof the green bee-eatersto female indigobirdsemphasizes the total lack of any speciesspecificity in mate selectionin courtshipbehavior by the male indigobirds. Behavioralisolation of the indigobirdsof differentspecies coexisting in the sameareas does not seemto involve male discrimination.In many populations femalesof sympatricspecies are barely recognizablein the field to the humaneye. Male indigobirdstend to court,in additionto their own females, the other birds most similar to female indigobirds;the Amadina fasciata re- semblesfemale indigobirdsin being both brown and short-tailed,and this finch was the mostfrequently courted of all birds other than femaleindigo- birdsthemselves. The starlingCinnyricinclus leucogaster also is notablyshort- tailed, althoughit is muchlarger. No displaysgiven to birds other than female indigobirdswere followed by mating. Behavior of other male viduines directed towards female indigobirds.-•In about 100 hours of observationof malesof other speciesof viduinesI never saw a female indigobirdvisit their call-sites,but malesof other speciesdid occasionallycourt female indigobirdsat indigobirdcall-sites when the male indigobirdswere absent. Theseoff-site malesmay have been without a site of their own. Specieswhich courted female indigobirds were Vidua macroura and V. paradisaea.In additionan adult male Shaft-tailedWhydah (V. regia) courteda female V. chalybeatawhen the indigobirdfed on the ground20 feet from the V. regiacall-site. The femaleindigobirds were unreceptive to these males. The behaviorof all male viduinesobserved in courtingindiscriminately a wide variety of small brown birds, both viduine and non-viduine,suggests that female behaviorplays a major role in mate selectionand speciesdiscrimination. Discussionof courtshipdisplays.--A brief comparisonof the displaysof the indigobirdsand other viduineswith the displaysof other Old World finchesand sparrowsmay be helpfulin suggestingthe possibleorigin of these behaviorpatterns from the behaviorof the non-parasiticrelatives and their adaptivemodification in the indigobirds.Displays have been describedfor 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 37 a numberof speciesof Ploceinaeby Crook (1964) and Emlen (1957); sparrows of the genusPasser were reviewedby Andrew (1961); and the waxbills and grassfinches of the family Estrilididaeare well known from the work of Moynihan and Hall (1954), Morris (1958), Kunkel (1959), Harrison (1962b), Immelmann (1962b), and Immelmann et al. (1965). I have ob- servedcourtship displays in most speciesof viduines;many of theseare well describedby Nicolai (1964, 1969). The hover displayis sharedby most viduines. I have seen hoveringin Vidua macroura,V. regia, V. [ischeri,and the ExclamatoryParadise Why- dah (V. interjecta), as well as in the indigobirds. In several features the hoveris similarto the displaysof many Ploceusand Queleaspecies (Ploce- inae). The most conspicuousdisplays of nestingPloceinae are the nest ad- vertisementdisplays given by malesat the neststhey have built (Crook,1963, 1964; Colliasand Collias,1970). The generalpattern of displayconsists of the male raisingits wings over the back and vibrating or waving them. In this displaythe malehangs onto the nestand oftenpoints with his bill to the entrance. The displayattracts females to the nest, and when a female visits, the male may intensifyits wing quiveringor give song bows (Crook, 1964: 64). In somePloceus species it functionsas a nestadvertisement display, in othersas a precopulatorydisplay, and in somespecies both (Crook, 1969: 283). Severalspecies of Passerhave a similar form of displayat the nest (Harrison, 1965a; Payne, 1969), and Harrison (1965a) has suggestedthat Ploceinaeand Passerinaemay thereforebe closelyrelated. The courtship hover of the indigobirdsand other viduinesis similar in form to the wing vibrationof eachof thesegroups, and it was likely derivedfrom the displays of one of them. The similarityof the viduinehovering is especiallyevident in the Straw-tailedWhydab (Vidua fischeri). In this speciesthe male often holds onto its perch at the call-sitewith its feet while it beatsits wings (Nicolai, 1969; my own field observations),much as a male Ploceusor Passer percheson its nest when it displays. Many Estrildidaebob up and downon a perchduring their courtshipdisplays,but they do not raise and wave their wingsin either nestadvertisement or precopulatorysituations (Andrew, 1961: 560; other referencesabove). The wing movementsare probablybasically more importantthan the bounc- ing body in the indigobirdhover, as male indigobirdsin low-intensitydis- play do not bob when they remain on a perch. As the male indigobirdor otherviduine hovers, it usuallybows its head towardsthe female,emphasizing the contrastin the colorsof its bill and head. Andrew(1961: 329-330) suggeststhat in manysong-birds the loweringof the headin courtshipis derivedor ritualizedfrom the nest-buildingmotions of the male;in all of the Old World finchesand sparrowsthat build nests, themale does the building.Bill loweringis commonin malecourtship dis- 38 ORNITHOLOGICAL MONOGRAPHS NO. 11 play in Ploceinaeand in Passer,whereas among the Estrildidaesome species lower the bill, someraise it, and some do both (Andrew, 1961 ). The lower- ing of the bill in the displayof the viduinessuggests a closerrelationship to Ploceinaeor Passerinaethan to the grassfinches. Songis givenin male courtshipdisplay by the Estrildidaeand the Ploceinae but not the Passerinae,according to Andrew (1961: 577-578). However, Passererninibey is vocal during its display (Payne, 1969: 302), and Passer rnoabiticusmay be, as the bill of a displayingmale is held open (Harrison, 1965a: 27). As the male indigobirdusually gives a nonmimeticsong while hoveringover the female, the viduinesmay share this behaviorwith species in all groupsof the Old World finchesand sparrows. The greetingflight of the indigobirdsmay be derivedfrom an unritualized flight approachof a male of the Passerinaeor the Ploceinaeto a femalethat is attractedto its nest or territory. Male Euplecteslack a specialdisplay flight directedtowards the female, and the absenceof any specialfeather elevationin the indigobirdgreeting flight further suggeststhat this flight is not derivedfrom the Euplectesform of undirectedflight display,as Euplectes regularlyfluffs out its bright red or yellow feathers (Emlen, 1957). No comparableflight displayis known in Estrildidae. Someestrildids do have a conspicuous"whirring" flight by the male to the nest or to the female before its nest advertisementor precopulatorydisplay (Morris, 1958; Harrison, 1962a), but the over-all similarityof this behaviorto the indigobirdgreeting flight seemssmall. The ritualizedpost-courtship flight to the ground,where the male sings,is apparentlyrestricted to the viduines.I have seenthis both in the indigobirds and in other Vidua species.Although the birds may feed at this time the displayis probablynot derivedfrom courtshipfeeding, because courtship feedingis rare in nestingfinches and has been noted in only one speciesof Ploceine (Andrew, 1961: 577-578; Crook, 1964: 86). The female indigobird regularly does feed at this time while the male sings. As mentioned above,this behaviormay relieveaggression or fear of the pair which may have resultedfrom the rapidity of eventsand especiallythe male "attack" of thefemale before copulation. Perhaps also the feeding reinforces the response of the femaleto the associatedsong. This may increaseher responsiveness to firefinchsong at other times,including times when shemay find the nest of a firefinch as well as timeswhen she may return on a regular basisto the samemale singingon his call-site. This last aspectseems likely inasmuchas in a species(Vidua rnacroura)whose male doesnot havemimetic songs but neverthelessdoes fly to the groundand call after courtship,the femalejoins him and feedsas the indigobirdsdo; this I saw repeatedly. Solicitationof the indigobirdsresembles the behavior of Ploceinae and Passerinaemore closelythan it doesthis displayin Estrildidae.In thesefirst 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 39 groupsthe femalequivers the wings,whereas female Estrildidae are unusual amongsongbirds as they do not quiverthe wingsbut rather hold them mo- tionlessagainst the body (Andrew, 1961: 560; Nicolai, 1964: 179; Payne, 1969; the presentstudy). FemaleEstrildidae quiver the tail (Andrew,1961; Nicolai, 1964). Tail vibration occurs generallyin the Ploceinae (except Quelea) accordingto Andrew (1961: 561), but Nicolai (1964: 180) regards tail quiveringas uncommonin Ploceinaeand describesthe solicitationof the Yellow-backedWhydah, Euplectes (Coliuspasser) macrourus, as typical of the groupin havingno tail quiver. Crook (1964: 36) notesthat somefemale ploceinestail quiver but that the wingsare quiveredmore vigorously.Prob- ably the tail quiver is inconspicuousbecause of the vigorouswing quivering. In Passer the female is said not to quiver the tail (Andrew, 1961: 346). Nicolai (1964: 180) has suggestedthe Euplectinae(currently included in the Ploceinae) as the group ancestralto the Viduinae on the basisof the female solicitationposture. Solicitationof femalePasser, however, appears to be as similarto that of Vidua as is the displayof Euplectes. Courtshipdisplays of the indigobirdsare more similar in generalto those of the Ploceinae and the Passerinae than to those of the Estrildidae. Because many of the courtshipdisplays of the Ploceinaeand Passerinaeare similar,it is not possibleat the presenttime to claim either as the related group from which the Viduinae are more likely evolved. The displaysare modifiedfrom those nest-buildingforms in the orientationof the viduine displaysto the call-siterather than to a nest,but theydo not appeargreatly modified in their form, sincethe hoveringdisplay and other behaviorpatterns are similar to displaysin the nestingspecies. Only one of the displays,the post-courtship flight to the groundwhen the male singsand the femalefeeds, appears to be stronglymodified in form in relationto the broodparasitism of the viduines. This displayas well as the othermale displaysis adaptedto the call-sitebehavior of the indigobirds,as it permitsthe male to consortand feed with the female,perhaps stimulating her to ovulate,and as the polygynousmale can remain near the site for matingswith other females.

BROOD PARASITISM AND THE INTERACTIONS OF INDIGOBIRDS AND FIREFINCHES

EVIDENCE OF PARASITISM AND HOST SPECIFICITY IN INDIGOBIRDS Indigobirdswere long thought to lay theireggs in thenests of manyestrildid finchesincluding the firefinches,but not to be restrictedto a singlespecies of host (Friedmann, 1960). In captivity the indigobirdshave parasitized Lagonostictasenegala (Poulsen, 1956); other aviculturistscited in Friedmann (1960) have reportedthat indigobirdssometimes build their own nests. The first clearevidence of host-specificparasitism of L. senegalaby Vidua chalybeatain the field cameonly in 1955 from G. and M.-Y. Morel in Senegal. 40 ORNITHOLOGICAL MONOGRAPHS NO. 11

Otherworkers have suspected this host-parasite relationship from the resemblancein mouth markingsof the young (Neunzig, 1929a, 1929b) and the similardistribution and habitatof the two species(Bannerman, 1948; Fried- mann, 1950; Chapin, 1954; Mackworth-Praedand Grant, 1960). The specific vocal mimicryof V. chalybeataand other indigobirdssuggests host- specificityin parasitismas well; neverthelessthis mimicry is no more than an indirect meansof determiningthe host. It is thereforeof interestto examinethe directevidence available for hostspecificity in all regionsof Africa. Few early recordsof indigobirdparasitism were basedon adequateidenti- ficationof the birds. No speciesdifferences had been describedin the eggs, juvenal plumage,or markingsof identifiedyoung, and prior to the present studyno speciesdifferences were known in femalemorphology. Many of the records listed by Friedmann (1960) of little white eggs found in various estrildidnests and identifiedas thoseof indigobirdswere in somecases more likely laid by otherkinds of viduines,or possiblyeven by the nestingestrildids themselves. Critical evidenceof a one host-•one parasite relationshipcomes from Senegalwhere the Morels have for many years studiedparasitism of the Senegalfirefinch, Lagonostictasenegala, by the Village Indigobird, Vidua chalybeata (Morel and Morel, 1955; Morel, 1959, 1964, 1967, 1969). The Morels have seen female indigobirdsenter a nest of L. senegalaand then depart, leavingthe nestwith one egg more than it had before, and they have bandedand photographed(see Friedmann,1960: plate 6) youngindigobirds in many nestsof L. senegala.They found indigobirdeggs or youngin more than 300 nestsof L. senegala.Identification of the indigobirdsis no problem there,because along the SenegalRiver area at the southernedge of the Sahara no indigobirdsoccur other than V. chalybeata,and no firefinchesother than L. senegalaare knownwithin 160 miles. Indigobirdshave not beenfound to parasitizeany other speciesof birds in Senegal. Supportingevidence of parasitismof L. senegalaby V. chalybeatais available from other areas. In northernNigeria, at Zaria, C. H. Fry (1965) ex- amineda nestof L. senegalacontaining two youngfirefinches and a young indigobird. He noted that the young of host and parasitewere of the same ageand resembledeach other closelyin the form of the head and bill and in the mouth markings. The nestwas in a residentialarea where V. chalybeata is the only commonindigobird. I saw malesof this form on 24 occasions but neverany other indigobirdin the sameuniversity residential area in 1968, and hence it is most likely that Fry's young indigobirdwas V. chalybeata. In the Ngong area of Kenya, van Someren(in Friedmann, 1960: 85) saw a juvenal-plumagedindigobird with a flock of L. senegala;the youngindigobird beggedfor food from the firefinches. No indigobirdsother than V. chalybeatacentralis are known in the Kenya highlands. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 41

R. K. Brooke saw a youngindigobird with L. senegalaat Hot Springs, lower Odzi River, Rhodesia,on 3 March 1965. He quotesfrom his journal (letter, 13 March 1969): "A juvenileindigo finch in companywith little ruddy waxbills[= L. senegala]: it looked like a very stout juvenile bronze mannikinbut with streakywings and very palepink feet andlegs." The pink feet and legs togetherwith the streaky appearancesuggest V. chalybeata amauropteryx. I saw three juvenal-plumagedindigobirds each feedingin a family party of L. senegalaat Maun, Botswana. On one occasiona young bird begged from an adult female L. senegalaas did two young firefinches. The only indigobird known from the Maun area and elsewhere in northwestern Botswana is V. chalybeata. D. N. Mansfield (in Traylor, 1966: 67; and letter 10 April 1967) found parasitismof firefinchesidentified as L. senegalaand L. rubricataat Lilongwe, Malawi. Mansfield independentlydiscovered the vocal mimicry of indigobirds and he ascribedthe observedparasitism of thesefirefinches to V. chalybeata and V. funerearespectively. However, in the absenceof any known methodof identifyingthe eggsor youngof the differentindigobirds, other than by raising them for a year or two, their identificationhere is open to some question,though very likely these were indeed the indigobirdspecies involved. One additionalobservation of a juvenile indigobirdwith a family group probablyrefers to V. funerea;C. J. Vernon (1967) saw a young indigobird as well as four youngfirefinches in a family grouprespond to the alarm call of an adult Lagonostictarubricata near Pietermaritzburg,Natal. BecauseV. [unereais the only indigobirdknown for southernNatal in the well-collected Pietermaritzburg-Durbanarea, the young is regarded on distributional groundsas very probably V. [unerea. Althoughthe list of directobservations of parasitismby reliablyidentified indigobirdsis short (with the exception of the detailed studies of M.-Y. Morel), it is noteworthythat no valid observationscontradict the notion of species-specificparasitism of a firefinch by each kind of indigobird. Success- ful parasitismby indigobirdsof speciesother than firefinchesis unknownin nature. Various recordsof suchparasitism are unsupportedby the collection of the birds which hence are of questionableidentification. These records includevisits of indigobirdsto trees containingnests, indigobirds seen in an area whereunidentified eggs are found in nests,and observationsof indigobirds buldingtheir own nests(see Jourdainand Shuel, 1935; Roberts, 1939; Bannerman, 1948; J. Vincent, 1936; A. W. Vincent, 1949; van Someten, 1916; Friedmann, 1960). Indigobirdsare generallyraised only by firefinchesand in everyknown instance the host speciesis alsothe songmodel of the indigobird. 42 ORNITHOLOGICAL MONOGRAPHS NO I 1

Figure 10. Nesting habitat of Lagonosticta senegala at Kondowe, Transvaal. Fire- finches use feathers of domestic chickens for display and nesting, cracked grain for food, and the roofs of traditional African housesfor nestingsites.

ECOLOGY OF INDIGOBIRD PARASITISM The ecologicalinteraction between Vidua chalybeataand its firefinch host Lagonostictasenegala has been studied in Senegalby M.-Y. Morel (1967, 1969). The following commentsmainly are taken from her studieson these two species. Firefinchesnest not only in the hush but also in the traditionalthatched- roof African housesin villages,and they feed largely on grains and weed seedsof humancultivation (Figures 10 and 11 ). L. senegalahas a short life expectancy(seven months) and a high rate of reproduction.The survivalof thesebirds has bcen studied by Morel (1969) in great detail from more than 7,000 bandedbirds. Young birds may breed when they are four monthsof age, and a pair of firefinchesmay nest four times in a singleyear. The breedingseason coincides with the time when grass seedsare most readily available;even so, the firefinchesraise an averageof only 2.7 young per successfulnest from a clutch averaging3.4 eggs. Many nestsare abandoned, especiallyduring incubation,and nestswith youngmay also be deserted. Indigobirdslay eggsin 35 percentof all nestsof L. senegala.From one to four indigobirdeggs are laid in each parasitizednest; the averageis 1.7 eggs. Usuallythe indigobirddoes not removea host eggfrom the nest. The young indigobirdshatch in a shortertime than do the firefinches(10 days versus 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 43

Figure l I. Nest of La.eonosticta sene.eala in thatched roof of house at Kondowe, Transvaal. 10 February 1967.

13 days). Young indigobirdsgrow up with their host nest matesand do not destroythem. The youngall fledge at about the sametime. Fledglingsin a family party are not aggressive;young firefinches sometimes preen the young indigobirdsmuch as they preen each other. The young respond to family- specificcalls of the adult firefinches. After two to three weeks the young are independentand move into the surroundingsavanna while the firefinches remain in the villages. The effect of parasitismupon nestingsuccess of the firefinchesis slight, in contrastto that resultingfrom the removal of eggsor killing of young hosts by many other parasiticbirds. The number of eggslaid by firefinchesis not affectedby the presenceof indigobirdeggs in the mixed clutch. Fewer parasitized nests are desertedduring incubation than are unparasitizednests. M.-Y. Morel suggeststhat the extra number of eggs in the nest functions as a super-normalincubation stimulus to the adult firefinchesand thereby in- creasesthe probability of hatching successof parasitizedfirefinches. After hatching, a slight adverseeffect of parasitismoccurs in the firefinch nests. Adult firefinchesmay raise one more young bird (counting indigobirds) in parasitizednests than in unparasitizedones, but since some nestshave more than one indigobird,the averagenumber of youngfirefinches fledging from the successfulparasitized nests is less than from unparasitizcdnests. Calcu- lating survivalfrom the time of egg laying, in both kinds of nestsan average of 2.7 firefinchyoung are rearedto fledging,and the over-allbreeding success 44 ORNITHOLOGICAL MONOGRAPHS NO. 11 of firefinchesis not modifiedby indigobirdparasitism (Morel, 1969: 68). The net effect of the brood parasiteon the successof a firefinch nest is zero: parasitismincreases the chancesof hatchingbut this effect may be counterbalancedin the nestlings,possibly through competitionfor food brought to the nest. Exhaustionfrom breedingeffort is indicatedin nestingfemale firefinches: layingin a secondnest once a malehas begun to buildit maybe delayedin fe- malesaccording to the time elapsedsince an earlierbrood (Morel, 1969: 92). Males showno exhaustion,but of coursetheir own breedingrate is slowed alongwith the mates'.The malesgenerally build the nests,but sometimesthe pairmay re-use an old firefinch nest, and when they do, the nests are less often parasitizedby any indigobirdyoung. Over her 10-yearstudy Morel notedno net changein firefinch numbersin the town of Richard-Toll. From Morel'sdata on the delayedtime of breedingagain when large broods are rearedand on the likelihoodof dyingin that time, it seemslikely that sea- sonaland lifetime successin fledgingyoung firefinches is lower in adultsthat rear young indigobirdsas well. Probably any unparasitizedbreeding fire- fincheswould enjoymore firefinchoffspring in a season,and at least in some circumstances,natural selectionfor avoidanceof parasitismwould be expected.Morel's (1969: 92) suggestionof precocialbreeding as an adaptation to escapeparasitism, however, is improbable---breedingat an early age (as in unparasitizedspecies of estrildids;Immelmann, 1962b) may be selectedfor regardlessof high mortality, not becauseof it.

BEHAVIORAL INTERACTIONS BETWEEN INDIGOBIRDS AND FIREFINCHES I seldomsaw indigobirdsand firefinchestogether and I never saw an indigobirdin the act of laying. Someinformation on the strategyof female indi- gobkds in finding host nests, and on the possiblecommunication between host and parasiteassociated with mimicry of firefinch songby the indigobirds was providedby field observations;these were supplementedby observations of free-flying captivesin our house and in aviaries. Behavior o[ indigobirdsand [irefincheson common [eeding grounds.--In the field, male indigobirdsoften fed with firefinches on fallen grassseeds, thoughother speciesof finchesdo this more often and indigobirdsalso fed with other grassfinches.In all I recorded13 instancesof male indigobirds feedingwith firefinchesand each time the indigobirdwas with its host firefinch species.Once a male indigobirdchattered and anothergave a single mimetic call note but usuallythe birds fed quietly a few inchesfrom each other with no apparent direct interaction. Feedingsessions lasted from 3 to 10 or more minutes. No agonisticor appeasementbehavior between host and parasitewas evident. Indigobirds seemto lack specialdisplays directed to the firefinches,nor do the adults 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 45

Figure 12. Contact perching and allopreening Ior pecking) of two L. senegala at Maun, Botswana. Note the sleeked posture of the aggressivebird and the ruffled head feathers of the bird submitting to the blows to the head. A third firefinch watches. use appeasementdisplays of the firefinches themselvesßThe firefinches and related estrildinefinches have severalbehavior patternswhich function so as to inhibit or to redirect any aggressivemotivation of another individual. Thesedisplays include curtseying, nodding, bowing, fluffed singing,and allopreening(Figure 12) or contactperching and have been describedby Kunkel (1967), Harrison (1965b), and Sparks (1963, 1965). Responses of indigobirds and firefinches to each others' vocalizations.-- Behavioral responsesbetween male indigobirdsand firefinchesoccur mainly at call-sites of the indigobirds when the males mimic their hosts. On five occasionsI saw firefinchesapproach the mimickingindigobirds at the moment of mimicry. In six otherinstances the songtype was not noted. Responseto the mimicryby firefincheswas clearlyevident as firefinchesusually stayed near the ground;they sometimeshopped on the groundwhen the indigobird was silentor singingnonmimetic phrases but flew up to him whenhe gave a mimetic song. On the call-sitefirefinches usually perched and peeredat the mimics. Their posturealternated between upright perchingand crouch- ing; in the crouch positionthe firefinch twistedthe tail and body towards the indigobirdand bill-wiped. In the field thesemovements did not appear ritualized, although they did somewhatresemble the curtseyingdisplay (Kunkel, 1967:241 ) described for Lagonosticta senegala and L. rubrk'ata. 46 ORNITHOLOGICAL MONOGRAPHS NO. 11

On only two occasionsdid a firefinchfly to a call-sitewhere the male indigobird was not its songmimic. One of theseinstances involved a pair of L. senegalawhere they were shakingand dryingthemselves after a bath, in a tree which alsoheld a singingVidua wilsoni. On the other occasion,a male L. senegalajust happenedto perchat a call-siteof a male V. purpurascens. Male indigobirdsusually either ignored visiting firefinches or supplanted them. Only oncedid a malevocalize towards the firefinch. A maleL. senegalagave a contactcall which may be rendered"twee" and the male V. chalybeata answeredit with the same note; the two called back and forth for a few seconds.No firefinchattacked an indigobird.All of the firefincheswhich flew to mimickingindigobirds were eitherlone males (on 9 occasions)or pairs (2). The sessionsterminated when the firefinchesflew off; the indigobirds did not follow them. Occasionallymale indigobirdsresponded to the songsof firefinches;when a firefinch sangthe male indigobirdbecame silent or countersangwith it. Sometimesthe malesflew from the call-siteto the singingfirefinches. Most observationsof respondingindigobirds were made at Mereriskyat a Vidua purpurascenssite near which a male Lagonostictarhodopareia gave clear, loud seriesof whistledtrills. Usually the indigobirdremained silent for a half minute after the firefinchsong; then it chatteredand sangmainly nonmimetic phrases. On three occasionsI heard countersingingbetween host and parasiteand this was muchlike the countersingingof two L. rhodopareia males to each other that I heard once in the field at Mererisky. In these sessionsthe male indigobirdusually sangback the same phrasewhich the firefinch had just sung (10 of 14 times in one countersingingseries). I noticedcountersinging between indigobird and firefinch also in our captives of thesetwo species.On four occasionsin the field a male indigobirdflew from the siteto a singingfirefinch--V. purpurascensflew twice to L. rhodopareia, V. chalybeataflew once to L. senegala,and V. f. codringtoniflew onceto L. rubricata. Each time the indigobirdreturned alone to the call-site in a minute or two, sanga harsh, nonmimeticsong, and showedno further interest in the firefinch. Is vocalsong mimicry directed toward the firefinchhost?•The field observations contain little evidenceof mimicry in the indigobird songsas an adaptationfor interspecificcommunication in the indigobirds.Although indigobirdssometimes imitated the phrasesas the firefinchessang them, they generallydid not direct their singingtowards visiting firefinches at the call- site, nor did they sing a high proportionof mimeticphrases when the hosts visitedthe site. Firefinchesapproaching indigobirds were usually ignored or supplanted.The existenceof countersingingbetween host and parasitesug- gestsa possiblerole of mimicryin the nest-findingbehavior of indigobirds,by soundingout the locationof host nests. This singingbehavior is limited to 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 47 the area near the call-site, however, and I was unable to find any nests my- selfby searchingnear the singingfirefinches. No doubtthe indigobirdsmight be more successfulthan I wasin findinga nestin this manner. However,I sawnothing to suggestthat countersingingwas used by indigobirdsin finding host nestseither in the field or in captivebirds. Only the male indigobirds mimic,and vocal mimicryis evidentlynot usedas an interspecificsignal to appeasethe host nor to find its nest. Gloria Sullivan (1970) has watchedaviaries of finchesfor many hours to find whetherindigobirds direct their mimeticsongs toward the firefinches. The behavioralresponses of captive,nesting Lagonosticta senegala to the vocal mimicry of Vidua chalybeatain an aviary includedmainly flying towardsthe singingmimic, but the indigobirdsdid not appearto respondto the firefinchesnor to exploittheir behavioralresponses. The indigobirds werenot attractedto the respondingfirefinches, nor did they everapproach the nests. Nicolai (1964:153 ) oncesaw a femalefirefinch (L. senegala)in captivity enter an old, unusednest when a male Vidua chalybeatasang a firefinch- mimetic nest call near the nest. Nicolai noted that the female was ill and was not breeding,and her entry into the nestwas in no way an inspectionof a possiblelaying site. I know of no otherobservation that might suggestthat firefinchesmay respondto indigobirdmimicry in a natural situation. Althoughvocal mimicry by the adultindigobirds is apparentlynot usedto appeasethe hosts,the brownish,streaked plumage of femaleindigobirds may appeasethem, as the plumageis notablylacking in resemblanceto that of adult firefinches.Nesting firefinches chase off intrudingfirefinches of their own.species from the nestsin the wild and in captivity,but in both situations they may toleratefemale indigobirds near their nests. The plumageof the female indigobirdsis not only inconspicuous,it also is neutral in lacking the bright marks which may function as intraspecificaggressive signals in the host species(Payne, 1967: 364-366). The role of songin estrildidethology differs from that in most passerine birds, as notedin nearly all studiesof captiveand wild birds. Estrildid finches include many highly social birds. Their songsappear not to be territorial advertisementsbut rather are sexual and social in function (Morris, 1958; Immelmann, 1959, 1962b; Kunkel, 1959, 1967; Harrison, 1962; Hall, 1962a). This interpretationof estrildidsong is basedmainly on variousstudies of speciesof the more sociablegroups Erythrurinae and Amadinaerather than on the more solitaryspecies of Estrildinae,the group that includesthe firefinches. The generalizationapplies also, at least in part, to the firefinches. Firefinchesare dispersedin pairs in the field during the breedingseason (Chapin, 1954: 519, 521, 529; McLachlanand Liversidge,1957: 449, 452; Kunkel, 1967: 239; Morel, 1969: 99-100; my observations). Male Lagono- 48 ORNITHOLOGICAL MONOGRAPHS NO. 11 sticta senegala are intolerant of other males within a few feet of their nestsand may attack intruderswith sharpblows to the head (Morel, 1969: 100). Both in the field in Nigeria and in our house,where one pair of L. senegalalived upstairs and another pair downstairs,breeding firefinches chased off other firefinches from their nest areas. Territorial behavior was usuallysilent. The countersingingof wild L. rhodopareia,a more vocal species, however,suggests a role of songin its territorial advertisement.Also the responseof firefinchesto singingindigobirds suggests some agonistic behavior, sincemost firefinchesresponding to vocal mimicry were males and only male firefinchessing; and while curtseyingmay be sexual in origin (Kunkel, 1967: 245), crouchingclearly indicatesan agonisticcomponent in the responseof firefinchesto mimicry,as a greatmany passerinescrouch in threat (Andrew, 1961:339-341 ). In general,firefinch song seemsnot to be a signalof highintensity of agonisticbehavior, and vocalmimicry by the indigobirdstherefore does not resultin an active territorialresponse of the hostsand a possibleexclusion of the indigobirdfrom its nestarea. Nor is the behavioralsignificance of firefinch vocalizationso stronglyappeasing that the breedingindigobirds are usingmimicry to pose as the small firefinches' innocentcompanions. The communicatorysignificance of vocal mimicry in indigobirdsapparently does not dependupon a firefinch respondingto the mimicry;rather the mimicryis an intraspecificsignal used in matingbehavior among the indigobirdsthemselves. How theindigobird finds a nest. Femaleindigobirds that I sawaway from the call-sitewere usuallyperched in a tree, peeringaround. The lone females closely watched the movementsof feeding or courting firefincheson the groundand followedthem when the firefinchesmoved out of sight. Several of thesefemales were collected;some had hard-shelledeggs in the oviduct. Femalesalso peeredinto nooksand crannieswhere nestsmight be when no firefincheswere in sight. A female Vidua chalybeatawith a shelledegg in the oviductsearched while I watchedit for at least 20 minutesby hopping througha row of aloesplanted around a villagenear Kisumu,Kenya; its host Lagonostictasenegala nests in aloesin Kenya (Mackworth-Praedand Grant, 1960: photo on plate 18). On three occasionsfemale V. chalybeatawere seenfeeding with their host species.Only twicewere femalesseen with males away from the call-sitein associationwith firefinches. At Maun, Botswana, I sawon two occasionsa pair of indigobirdsperched in a bushquietly watch- inga familygroup of firefinchesfeeding on the groi•nd; the male indigobird terminatedthe watch and returnedto his call-site140 feet away. Sinceob- servationsat the call-sites show that a breedingmale'•pends nearly all of his timeat the call-site,it appearsthat a femalefinds th• nestsof the firefinches entirelyon her own by watchingthe behaviorof the firefinchesand by searching. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 49

In Senegal,M.-Y. Morel (1959: 158-159; 1969: 82) likewiseobserved that femaleindigobirds travel alone when searchingfor hostnests and laying; occasionallytwo femalesnear the samehost nest may squabble,but malesdo not visit the nests. Morel has seen a lone female enter the nest of a L. senegala while the femalefirefinch was incubatingand lay an egg withoutdis- turbing the firefinch. Captivefemale indigobirdsin our housevisited the nestsof firefinches alone,following the firefinches(especially males with grassin the bill) behind bookcasesand peeringinto closetswhere the nestswere built. Males never followedour firefinches,searched through the bookcases,or accompaniedthe femaleindigobird. Firefinches appeared to be obliviousof the indigobirdsexcept when nesting;female indigobirdssometimes then were chasedfrom the nestitself by the nestingfirefinches but more often they were tolerated. We noted this interest of a female indigobird in nesting Lagonostictasenegala only for Vidua chalybeata;a pair of captiveV. pur- purascensliving with two nestingpairs of L. senegalafor nine monthsin our house showed no interest in the firefinches or in their nests. Female V. chalybeatain my aviaries,where both L. senegalaand L. rhodopareiawere nesting, followedand frequentlyinvestigated nests only of L. senegala. The stimulusfor ovariandevelopment and laying in indigobirdsis apparently the sameas for femalefirefinches, namely the sightand soundof the malefirefinches as they build the nests.The femalefirefinch stands by and watchesor follows while the male flies to the nest with grass,feathers, bits of paper,and other nestingmaterial; these items he alsoholds in the bill as he bobs up and down to his mate in a courtshipdisplay known as the straw display (Harrison, 1962a: 263); Kunkel (1959: 337) believesthis display to be derivedfrom the nest-buildingmotions of the male. Female V. chalybeata in our houseregularly visited nestswhen they saw male L. senegala carryingmaterial to the nests,and they also watchedclosely when the male firefinchgave the strawdisplay to his mate. One nest in the housewas built by a pair of L. senegalain early July, 1969, in a Christmastree left for the birds in the living room. A female V. chalybeatabegan to visit the nest regularlyfor five daysafter shefirst found it. On 26 July the femalesuddenly died; in her oviductwas an egg,and clearlythe bird had becomeeggbound. Ovarian developmentand ovulationhad occurredin the absenceof a male V. chalybeata as none had been in the house for two months. No firefinch eggshad beenlaid by 29 July, so the sightof firefincheggs themselves could not have stimulatedthe female indigobirdhere. This observationsuggested to me the idea of stimulationof the reproductivecondition of female indigobirds by observingand hearing the nestingbehavior of the male firefinch. Nicolai (1964: 192-193) has reacheda similarconclusion by reasoningthat the synchronizationof layingof the viduineswith theirhosts implies a psycho- 50 ORNITHOLOGICAL MONOGRAPHS NO. 11 physiologicalresponsiveness of the layingfemale viduine, though he described no examplesof this. Morel (1969: 82) notes that at firefinch nestswhich are re-used,female indigobirdsusually do not lay at the sametime as the firefinches.When they lay in a re-usednest their eggsare laid too late for the largerclutch size to be a super-stimulusto the firefinches,and eggswhich are incubatedusually do not hatchby the time of firefinchegg hatching. When firefinchesre-use a nest rather than build a new one they may add some nestingmaterial, but lessnest-building behavior is seenand the femaleindigobirds are apparently insufficientlystimulated and thus not ready to lay by the time the firefinches do. In contrastto the abovedescriptions of the way lone female indigobirds find the host nests,Nicolai (1964: 151) has reportedthat in captivitymale viduinesof certain speciesmay lead the female to the nest of the host by mimickingthe nest call, a vocalizationgiven by the host male at the nest which it builds. Most of Nicolai's detailed observations of this behavior were for the Shaft-tailedWhydah, Vidua regia,which mimics and parasitizesthe Violet-earedWaxbill, Uraeginthus(Granatina) granatina,but he reportsit alsofor indigobirds.The significanceof this report is worth consideringbe- causeit suggestsa functionfor vocal mimicry rather different from its sig- nificanceas an intraspecificsexual signal. Nicolai (1964:151-153) described a chanceaviary observation of male mimicryof the nestcall apparentlyused to lead the female to the host nest. A translationof his observationsfollows; I have altered the text to agree with the terms used for the vocalizationsin the presentpaper, and I have includedmy own commentin brackets: A male V. regia was singing mimetic and nonmimetic phrases in calm alternation, and suddenly it broke off a series of nonmimetic phrasesand passeddirectly into a hearty granatina nest call and stared straight ahead in a horizontal posturein a fixed direction. There, halfway up in a juniper tree a male granatina was beginningto build a nest. As long as the waxbill was busy at its nest, the whydah male sang a series of granatina nest calls repeatedly and became silent at once each time the estrildid flew off to get new material. Then the whydah began again each time as the returning waxbill flew back with his load and landed at the nest site. [The male whydah had very seldom before given a mimetic nest call, making all the more remarkable the specific call when its host specieswas nest building.] Observationsat later times showed that regia males in breeding condition interrupt all other activities at the sight of a nest-building host bird and sing at once the granatina nest call in a posture pointing towards the site of the nest. As also other viduine males (Steganura, Hypochera, Vidua) accompany their females in their search for host nests and to the egg laying, I think it is nearly certain that the regia male, through the utterance of the nest call, directs the attention of his own female to the host nests, which otherwise would possibly have escaped his mate. He thengoes on to commentupon the adaptivesignificance of thisbehavior --even thougha femalemight find somenests on her own, it is mostimpor- 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 51 tant that she find the host nestsat an early stagebecause a femaleviduine ovulatesonly throughseeing the breedingactivity of a host pair, and if a mimeticmale attractsthe attentionof his mateto any neststhat mightother- wisepass unnoticed, she may lay more eggsand the male may mate with her more often, and leave more offspring. The greatimportance of chanceobservations of captivebirds and the in- sightthey give to the ethologisthave been codifiedby Lorenz (1970:371 ) into what he calls "a very simplerule: if the captiveanimal does show a complicatedbehavior pattern of recognizablesurvival value, the observeris fully justifiedin the conclusionthat the behaviourpattern observed is a con- stantproperty of the speciesin question."I agreethat a flashof insightmay sometimesbrighten the mind of a biologistwhile he is watchingcaptive animals.But the flashshould illuminate the way to testthe adaptivesignificance of behaviorby experimentsor continuedobservation; it should not blind the investigatorinto complacentacceptance of a perhaps misin- terpreted "revelation." At the least, if the behavior is not seen by others, especiallyin the wild, it will be difficult to confirm "Lorenz'srule." In my field work in Transvaal and Botswana I found Vidua regia to be like the indigobirdsin its call-sitebehavior and its mating system,so it is possibleto commentupon Nicolai'sfindings, especially since he statesthat the indigobirds(Hypochera in the passageabove) sharethe situation-specific mimicryof the nestcall. As notedin the previoussection, singing male indigobirdsat the call-sitemay answerthe vocalizationsof their hostswith the samecall, and the two birdsoccasionally may countersing.The nestcall of the estrildinesis givenby the male buildingthe nest. It seemslikely that Nicolai'swaxbill male may have been givingits nest call at the time it was nestbuilding, and the male viduine,which was right besideit (both were confined in the samesmall aviary) answeredit with the samecalls as long as the waxbill was at the nest and callingitself. The female viduinewas likely attractedby the nestcall givenby eitherbird. I considerit unlikelythat a wild male, not confinedin an aviary, would leave its call-siteand look for nestsand then call whenit mightfind them. The nestcall of the Estrildinae is a very soft call, hard to hear at any distance(Harrison, 1962a; Goodwin, 1964; Morel, 1969; my own observationson firefinches,pp. 87, 90), and it is unlikelythat a male indigobird,singing actively at its call-site,would hear the soft nest call of a firefinch at a distance of more than a few hundred feet. Nor do the indigobirdsrespond regularly to thiscall whenthey do hear it. I haveplayed back tape recordings of vocalmimicry of nestcalls of their hoststo singingmale V. chalybeataand V. purpurascenswithin a hundred feet of .theircall-sites (with a volumelouder than that ever given by a firefinchitself), andthe malesnever responded (p. 164). More importantly, the matingsystem of theseviduines is a polygynousone, not a monogamous 52 ORNITHOLOGICAL MONOGRAPHS NO. 11 onewith a strongpair bondas assumedby Nicolai,and the breedingmales are not free to roam about the bush and leave their call-sites unattended. It wouldbe strangeif wild maleindigobirds do accompanythe femaleto the nestof the host,because this behavior would be completelyunexpected and evenmaladaptive in lightof theirmating system. Any timethat a malemight spendwith one female would be time when he would be absentfrom his call- site. Highlypolygynous birds in whichthe malehas no pair bondwith the femalesafter they are inseminatedhave muchto lose (potentialmates and offspring) if they neglecttheir sexual self-advertizement.It would seem a betterstrategy for a polygynousa-male to remainon his call-sitenearly continuously,in orderto attractfemales and to be on handfor matingwith them. There are somenon-breeding adult males without a call-site,but it is unlikely that one of thesewould assista breedingfemale, because the eggsshe lays would have been fertilized by anothermale at its own call-site. Such an altruisticmale would be increasingthe chancesof successof the a-male's genesat the expenseof his own--a better strategyfor him would be to at- temptto displacea breedingmale himselfat the call-site. In my field obser- vationsthe male indigobirdnever accompaniedthe layingfemale; rather he sangall day on the call-site. Nor have any of our captivemales visited the nestsof firefinchesor shownany interestin nestbuildingbehavior. Morel's (1969) observationslikewise have shownthat female indigobirdsare alone when they are searchingfor nests and when they lay. Hence, both from theoreticalconsiderations and direct observationsdescribed earlier, it seems likely that male indigobirdsdo not normally accompanythe female and advertize nests to her.

MIMICRY OF EGGS AND YOUNG Viduine finchesresemble their estrildidhost speciesin egg color and size, mouthmarkings, and juvenalplumage. As youngviduines are rearedtogether with the host young the resemblanceof the young viduinesto the young estrildidsis thoughtto be of adaptivesignificance in increasingtheir chances of beingreared successfullyby the fosterparents. The mimetic resemblance of indigobirdeggs and youngto their presumedhosts has been discussedin some detail (Neunzig, 1929a, 1929b; Friedmann, 1960; Steiner, 1960; Nicolai, 1964, 1965a; Payne, 1967; Wickler, 1968). However, all of these studiesof indigobirdswere of a singlespecies, Vidua chalybeata. Examina- tionof additionalmaterial permits a comparisonof theseresemblances among the other indigobirdsand firefinchesas well. Eggs.--The eggs of all of the estrildids and of the viduines are white (Chapin, 1954; Friedmann, 1960: 22). Most of the many presumptive viduineeggs described from variousnests and arbitrarilysaid to be indigobird eggsof one form or another (recordsin Friedmann,1960: 72-73, 77- 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 53

79, 83-86) are of little valuesince neither the layingfemale nor the hatching youngwere identified. One valid recordis that of Olsen (in Friedmann, 1960: 83) whobred Viduachalybeata in captivityand found it to havewhite eggs "slightlylarger and more roundedthan those" of Lagonostictasenegala. In Senegalthe eggs of L. senegalaaverage 13.6 (12.1-15.5) x 10.8 (9.6-12.0) mm, and the eggsof V. chalybeataare 15.1 (13.6-17.0) x 11.8 (11.1- 13.0) mm (Morel, 1969: table23). The averageegg size of the two species is significantlydifferent, but the largerange in sizeand overlapsuggests that not all eggscan be identifiedto species. In my study,most eggs in the oviductsof femaleindigobirds were incom- pletelyformed. The sizesof matureovarian follicles and yolks in the oviduct were similarin all species.I recoveredundamaged shelled eggs from the oviductsof femaleV. chalybeatain Transvaal;the eggsmeasured 13.7 x 11.1, 13.8 x 11.1, 13.4 x 10.7, and 13.2 x 10.4 min. Egg sizeof V. chaly- beatathere averages about 13.5 x 10.8 ram,and its eggsare then very similar to the eggsof its hostL. senegalawhich average 13.4 x 10.5 mm (McLachlan andLiversidge, 1957: 452). White,shelled eggs of the samesize were taken from femaleV. chalybeatain Rhodesia,Malawi, and Botswanaand from V. purpurascensin Rhodesia,but thesewere broken and measurementsare only approximate.Membranous-shelled oviducal eggs with little or no calciumin the shellwere found in femaleV. chalybeata,V. purpurascens,V. [unerea, andV. wilsoniand all of thesewere from 12-14 x 10-11 min. The similarity in bodysize of all indigobirdsand in thesize of eggstaken from their oviducts indicatesno greatspecies differences in eggsize. Nor are theirhosts very different;all speciesof Lagonostictahave white eggs with ranges of 13.4-16.0 x 10.5-11.7 mm (McLachlanand Liversidge,1957: 448-452; Mackworth- Praed and Grant, 1960: 643-649; James,1970: 223). Mimicry of the mouthmarkings of the.young.---Young estrildid finches typicallyhave a distinctivepattern of blackspots on thepalate and sometimes on thetongue and lower mouth. They also have reflectile papillae or tubercules,elaborations of the oral flange, at theouter corner of thegape. These are especiallyconspicuous in a dark area suchas the insideof the covered nests,and theypresumably help to directthe parentbirds to the mouthsof their youngwhen feedingthem. Theyare thought also to stimulatethe parents to feedyoung with this pattern and perhapsto discriminateagainst alien youngsuch as the parasitic viduines(Neunzig, 1929b: 2; Nicolai, 1964: 172-175; Payne, 1967: 371- 372). Estrildidsin Australia,where the viduinesare absent,have equally conspicuousmarkings (Steiner, 1960), andthere the markingsmay aid nesting finchesin feedingexclusively their own youngin colonieswhere two or morespecies may occur(Immelmann, 1962a, 1962b). The mouthspots persistin adultestrildids, where in somespecies including Lagonosticta sene- 54 ORNITHOLOGICAL MONOGRAPHS NO. 11 gala they are directedtowards the mate in display (Goodwin, 1965: 296; Morel, 1969: 111; Giittinger, 1970: 1028). The pattern of black mouth spotsin pickled museumspecimens of the young estrildidsand viduinesled Neunzig (1929a, 1929b) to suggestthat each speciesof viduinemimics and parasitizesa singlespecies of estrildid host. In the seriesof nestlingsavailable to him, all viduineswere spotted,but generallyeach speciesdiffered from all the others;and each speciesmatched the mouthpattern of its presumedhost. The only firefinchesand indigobirds describedby Neunzigwere Lagonostictasenegala and Vidua c. chalybeata. The patternsof mouthspots and gapepapillae in the firefinchesand indigobirdsare shownin Figure 13. Data for the figuresof nestlingfirefinches are taken from the followingmaterials: (a) five Lagonostictas. senegala, four L. s. ruberrirna,and eight L. s. rendallihatched and reared in my aviaries, and alsoone pickled specimen of L. s. senegalafrom Richard-Toll,Senegal, now in AMNH; (b) six L. rhodopareiajamesoni reared in my aviariesin Ann Arbor in 1971; (c) L. rubricata and L. larvata, from Immelmann et al. (1965: 175, 207); and (d) L. tara, specimenin AMNH taken by Chapin at Faradje, Congo, probably the specimenused as the model for Chapin's drawing(1954: plate 24). I collectedjuvenile L. senegalaat Maun, Botswana, while they were still young and had large whitish reflectile granuleswith a blue mark between them. Juvenile L. s. senegala,L. s. ruberrirna, and L. s. rendallithat I rearedin captivitywere similar. I took the juvenileL. rubricata near Lilongwe,Malawi, from a family party. The adult firefinchmouth patterns were taken from specimensI collectedin the 1965-68 field studies; adult firefinchesretain the mouth spotsbut the refractile granulesare lost beforethe postjuvenalmolt. Accordingto Chapin (1954: 523) the youngof L. [ru.] landanae has markings like those of L. ru. congica. Figures of the indigobirdswere taken from a nestlingV. chalybeataabout eight daysof age, a pickledspecimen in AMNH collectedby G. Morel at Richard-Toll,Senegal. The juvenileV. chalybeata(RBP 4635) I shot from a flock of L. senegala at Maun, Botswana,where no other speciesof indigobirdsoccur; it had per- sistent reflectile granules and was completely in juvenal plumage. Other Maun birds as well as a juvenile,pink-footed, pink-billed V. chalybeatafrom Merensky had similar mouth markingsexcept that the tip of the tongue was black in older birds in which the gape tuberculeshad regressed.A young male V. purpurascenstaken at Merensky,identified by its whitish feet and bill, was in postjuvenalmolt. The spotson the palatepersist in adult indigobirdsbut are generallyfainter in the birds that have a greaterdegree of skull pneumatization;in seriesof adultmales of all indigobirdspecies I collected,at least one bird in each had the full complementof five spotson the palate. Other maleslacked the two posteriorsmall spots but thesewould probablybe visiblein youngerbirds. All adult femalesshowed, if faintly, sometrace of 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 55 56 ORNITHOLOGICAL MONOGRAPHS NO. 11 the three anterior,upper mouth spots,but the two posteriorspots were not visible in most. The patternof mouthspots is generallysimilar in all speciesof firefinches, althoughthe two smallerspots on the posteriorportion of the upper palate are not alwayspresent. These are lackingin someL. senegalabut occurin others. I have seenthese spots in five of my captive-raisedL. s. ruberrima nestlings,but they were lackingin four L. s. senegalayoung. The posterior spotsare not distinctlypresent in the juvenileL. rubricata. The dark mark- ingson the tonguebecome more distinctin older birds. Most juvenilesalso have a dark area at the tip of the upper mandible. All five speciesof indigobirdshave the samepattern of black spots;species differencesare evidentneither in adultsnor in juvenilesin the materialI have seen. Each indigobirdmore or lessresembles every speciesof firefinchin thesemarkings. The absenceof conspicuousdifferences in mouth spot pat- ternsamong the firefinchspecies suggests that the youngindigobirds of all specieshave a generalizedfirefinch-like mimetic mouth pattern. The colorsof the mouthsof young,however, differ considerablyamong the firefinches. L. senegalafrom East Africa and L. rhodopareiaimported from Rhodesiahave bred in my aviariesrepeatedly. Plate 1 showsthe mouthlinings of a youngL. senegalaaged 7 daysand a youngL. rhodopareia aged4 daysthat werephotographed at the sametime. To standardizecolor comparisons,the color platesof Kornerupand Wanscher(1967) were set beside each bird and the color names and numbers used here refer to those of the book. The youngL. senegalahad white gape papillae with deepblue basal areas extendingalong the commissurefrom the upper to the lower papillae,and the mouthlining was pastelpink on the lower mandible,the uppermandible posterior to the blackspots, and the liningof the commissural flangesbetween them. The spottedanterior half of the upper jaw and the tonguetip were yellow. In contrast,the L. rhodopareianestlings had gape papillae of bluish white, darkeningto blue at their base, and the edge and inner lining of the oral flangeswere bright purplishred-violet red. They had no hint of yellow in the mouth,which was pink excepton the front of the pastelpink uppermandible and the tip of the tongue. To standardize color terms, the color names above match color numbers 20 A 1 (white), 20 E 8 (deep blue), 11 A 4 (pastel pink), and 4 A 6 (reddishyellow) in L. senegala,and colors20 A 2 (bluish white), 20 D 8 (deep blue), 13 B 7 (purplishred-violet red), 12 A 4 (pink = rose), and 13 A 2 (reddish[or pinkish]) in L. rhodopareia. A similar color differencefor these two species'nestlings was noted by Nicolai (1970: 429), except that Nicolai's figures show no yellow in L. senegalaand no blue in the tuberculesof L. rhodopareia. Nicolai's birds werepainted by H. Kacherfrom Nicolai'sfield notestaken in Tanzania(J. ^ B

C D

Plate 1. Color photographsof the mouths of two young firefinches: a, b, Lagon- sticta senegala,age sevendays; c, d, L. rhodopareia,age four days. Figures a and c were taken by C. S. Adkisson. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 57

Nicolai, pers. comm.). On distributionalgrounds the firefinchesinvolved were L. s. ruberrimaand L. rh. jamesoni,and so were probablythe sameas the formsshown in Figure 14. A color photographon p. 413 accompanying the sectionon estrildid finches (Wolters and Immelmann, 1970) in the same book showsa youngL. senegalawith yellow areasin the mouth. Perhaps somewidespread or at leastlocal differencesin mouth colorsof youngfirefinches do in fact exist. The two kinds of firefinchesreared in my aviarieshad their distinctive mouthpatterns and colorson the day of hatching,and the colorswere stable throughthe nestingperiod. The bluesof L. senegalabecame blackish shortly afterfledging; the gapetubercules were gone by day 46. L. rhodopareiahad slightlysmaller tubercules on the day of fledging,and the red spotswere barely visibleas small red flecks two weeksafter fledging,when the young were feeding themselves. I have not seenthe mouth colorsof live youngof other firefinches,but eachof thesespecies appears to be unique,as far as publisheddescriptions permitcomparisons of colors. Chapin (1954: 519) describeda nestlingL. rara rara as follows: "Bill blackishexternally, skin of gapeswollen, purplish red, with two little white balls at each side. Tongueand palate cream color, throatflesh color .... "The mouthof youngL. larvata (ssp.vinacea ?) has bluish-whitepapillae, between them is a violet-bluepapilla at the cornerof the mouth, and betweenthis and the bluish-whitepapilla of the lower mandible the skin is dark (Immelmannet al., 1965: 207). A fledglingL. rubricam congicawas describedby Chapin (p. 520) as havingthe "skin of gape with two pinkish white swellingsat each side, a black dot on the inner side of eachswelling." An L. [rubricam]landanae fledgling described by Ansorge (in Chapin,p. 523 ) was "in generalagreement" with this last form; "the gape has two lumps at each side,whitish at the top, black at base." Immelmann et al. (1965: 175) describethe nestlingof L. rubricata(subspecies not given) as havingthe oral flangeswhite with the papillaeblue, and the spacebetween themblack. Furtherstandardized color comparisons in the field or in aviaries are requiredto find justhow similarthe colorsof the livingbirds might be; they disappearcompletely a few hours after death. The youngindigobirds found in the nestsof L. senegalaand of L. rhodopareia have mouth patternsand colorslike those of their foster siblings, accordingto Nicolai'sillustration (1970: 429). Presumablythe birds illustrated are V. chalybeataand V. purpurascens,respectively. I failed to take any noteson the colorsof independentjuvenile indigobirds collected in the field. It wouldbe interestingto examinethe mouthsof otherspecies of live youngindigobirds and to comparetheir colors with thoseof live youngfirefinches.If the adult firefinchesdo discriminateagainst feeding young with mouthsof colorsunlike those of their own young,then only the parasitic 58 ORNITHOLOGICAL MONOGRAPHS NO 11

Figure 14. Resemblancein female and juvenal plumage of indigobirdsand firefinches: RBP 4609, l'idua chalybeata, adult female, Maun, Botswana; RBP 4764, V. purpurascens,juvenile, molting, Mererisky, Transvaal; RBP 4633, V. chalybeata, juvenile, no molt, Maun, Botswana; RBP 4637, Lagonosticttt senegala, juvenile. Maun, Botswana; RBP 4893, L. sene.•,ala.adult female, Gusau, Nigeria. indigobirdshaving the matchingcolors would be likely to be raisedsuccess- fully by the foster firefinches. Mimicry o! juvenalplumage.--The juvenal plumageof viduinefinches resemblesthat of their estrildidfoster speciesin color and in suppressionof the heavystreaking characteristic of the adult females.The resemblanceis thoughtto be mimetic(Nicolai, 1964: 185-187; 1965a;Payne, 1967: 371- 372). The similarityin juvenalplumage between indigobirds and firefinches is not very close,however. Figure 14 showsa juvenile Vidua chalybeata and a juvenileLagonosticta senegala that I took at the samelocality. The youngindigobird had all of its feathersgrown and wasfeeding itself, but it still retainedthe mouth granules,had a large bursa, and had not begunthe postjuvenalmolt. Its plumageis burlierand lessheavily streaked than that of adult female V. chalybeatafrom Maun. The upperpartsof the young indigobird are grayishwith buffy edgesto the feathers;the firefinchis unstreaked gray aboveand has a red rump. In contrastto the paler, nearly unstreaked juvenalplumage of V. chalybeatain museumspecimens from Senegaland Sudan,the streakedplumage of the young southernAfrican indigobirdis 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 59

$0-

ol0-

50O 6OO WAVE LENGTH (m/.z.) Figure 15. Similarity of plumagecolor in juvenile Lagonostictasenegala and Vidua chalybeataplotted from spectrophotometricdata: a, breast; b, back. more like that of the adult femaleindigobird than like that of the young firefinch. The underpartsof the youngof both the southernindigobirds and their fosterbrood mates are unstreaked,but casualexamination and spectro- photometricanalysis of the two youngbirds shows that the indigobirdis paler (Figure 15, for techniquesee p. 213). The only other speciesof indigobirdrepresented by a juvenilespecimen (RBP 4767) of certainidentification is a V. purpurascenstaken at Merensky, Transvaal(see p. 238). Thebird is in earlypostjuvenal molt; the old juvenal feathersare gray,edged with buff, andare streakedto the samedegree as in threeV. chalybeatajuveniles from southernAfrica. No speciesdifferences are apparentin juvenalplumage between these indigobirds. Mimicry in beggingbehavior of the young.--As describedby Kunkel (1959: 340), Immelmann(1962b: 148-149), andNicolai (1964: 171-172), the estrildidshave a specialtechnique of transferringfood from parentto 60 ORNITHOLOGICAL MONOGRAPHS NO. 11

Figure 16. Young Lagonosticta sene•ala being fed by its parent. The young bird begs in this posture with the head held low and turned around; the parent inserts its own bill at the cornersof the gape betweenthe papillae of the young. young. The parent regurgitatesundigested food from the crop by pumping, and the young begs with a distinctiveside-to-side swaying motion of the head exposingthe mouth markingsand in somespecies (including the firefinches) with a waving displayof the tongue. When the parent is perched besidethe young the young twistsits head upsidedown in beggingto the parent. The parentinserts its bill into the mouth of the youngand feedsit in this position(Figure 16). This sameposture and behavioris used by begging youngviduines (Vidua rnacrouraand the paradisewhydabs V. obtusaand V. paradisaea aucupum) (Nicolai, 1964: 171-172; for use of these names seePayne, 1971). This specializedfeeding behavior restricts the viduinesto parasitismof the estrildid finchesalone; other kinds of birds would fail as potential hostsas they are unable to feed and raise the young viduines. As in the caseof eggcolor, the presence(though not the pattern)of mouthmarkings, and the juvenal plumage, the beggingtechniques of the young are con- sideredmimetic becausethe resemblanceof parasitesto host has been main- rainedby naturalselection, even though the commonancestors of the estrildids and viduines may already have had this character (Payne, 1967: 372). 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 61

Comparisonof the indigobirdswith their firefinchhosts indicates a close resemblancein the eggstage, a clearcase of mimicryin the nestlings,and a poorly definedsimilarity in the plumageof the juveniles. Thus the early stagesmore closelymimic the hoststhan do the later stages.Resemblance of the juvenileindigobirds to the firefinchyoung may be relativelyunimpor- tant for severalreasons. Whereas breeding birds (includingL. senegala)are oftenlikely to deserttheir nestsin the early stagesthey lessoften desertwell developedyoung, in the feedingof which they have investeda considerable amount of time and effort, and selectionfor mimetic charactersis apparently greatestearly in the nestingcycle when the chancesof removalor desertion by the fosterparents are greatest.Second, the plumageis not readilyvisible in the domednests of the firefinches,especially inside the dark African houses where L. senegalabuilds its nest; the foster parents are more likely to see the mouthmarkings. Finally the youngcan feed themselvesafter a few days out of the nestand are not long dependentupon the parentsto providethe abundant,readily available grassseeds. The speciesdifferences in signalsprovided by the mouth colors of the estrildidsmay accountfor the restrictionof successfulfostering of young indigobirdsto a singlehost firefinch. The specializedfeeding technique of the youngviduines matches the peculiarbegging technique of many kinds of young estrildid finches. The white eggslikewise match those of all estrildids. The pattern of black spotsand tuberculesin the mouth may excludethe indigobirdsfrom successfullyparasitizing some kinds of estrildidswith other marks, at least from thosein which the parentsmay discriminateby failing to feed any young unlike their own. The only conspicuousvisual character that is known among different speciesof firefinchesthat would make one kind of indigobirda mismatchto somefoster brood-matesis the color com- binationsin the mouthsof the young. The species-specificcolor patternsof the youngfirefinches probably have providedan importantselective basis for the evolutionof species-specificbrood parasitismamong the indigobirds.

VOCALIZATIONS OF FIREFINCHES AND THEIR MIMICRY BY THE INDIGOBIRDS The vocalizationsof firefinchesand indigobirdswere recordedwith a Uher 4000 Report-L tape recorder at a tape speedof 7• ips. The Uher M-514 microphonewas mountedat the focal point of a 30-inch fiberglassparabolic reflector. In the field it was usuallypossible to record indigobirdsong ses- sionsof three minutesor longer with the microphonewithin 60 feet of the male on the call-site. Most of the indigobirdswere collected immediately after their songshad beenrecorded, and the studyskins prepared from these were comparedwith other museumspecimens for positiveidentification. We made field recordingsof firefinchesonly in a few circumstanceswhen we saw 62 ORNITHOLOGICAL MONOGRAPHS NO. 11 the birdssinging in the bushor grass.A few recordingswere alsomade of captivefirefinches, including nestlings and fledglingsof two species. Tape recordingswere analyzed with a Kay ElectricCompany Sound Spec- trograph(7029A); wide-bandfrequency analysis was used for maximaltem- poral resolutionand for the distinctive,contrasting visual shapes in which it portrayedthe elementsof song. The audiospectrographs(or sonagrams) were then copiedwith high-contrastphotographic film, and the final prints are here directlyreproduced. The audiospectrographsgive a graphicrepre- sentationof the time (msec) and frequency(kilocycles per second= kilo- Hertz, or kHz) sound charactersof a vocalization. The data for time and placeof eachrecording shown in audiospectrographicform in the figuresin this paper and the locationof eachrecording in my field tapesare listedin Appendix A. FIREFINCH VOCALIZATIONS Firefinch vocalizationshave been studied by a number of ethologists (Kunkel, 1959, 1967; Harrison, 1956, 1962a, 1962b; Goodwin, 1964, 1969; Nicolai, 1964; Immelmann et al., 1965; Morel, 1969), and Nicolai (1965b) haspublished a 45 rpm recordingof captiveL. rubricam,L. rhodopareia, and L. larvata vinacea. However, for none of the firefinchesis the completevocal repertoirewell known, in part becauseof variationsof the samecall givenby singlebirds and in part becausesome calls are very soft and unlikelyto be easilyheard, for examplethe nestcall (see belowfor details). Even for L. senegala,the best known firefinch, the vocalizationsare not known in full detail; Harrison (1962a) describesfour calls, Nicolai (1964) recognizesfive and Morel (1969) givesseven; I have tape-recordedseven. A few non-vocal sounds of firefinches are also described below. Problems of the correct speciesidentification of firefinchesin behavioralstudies have arisen(Harrison, 1963a), and it is advisableto savespecimens of each bird recorded. A descriptionof firefinchsounds permits comparison of the mimeticsongs of the indigobirdswith the songsof the firefinchspecies. Where the functions are known,these are discussed.The sourcesof the followinginformation are field recordingsand recordingsof captivesas well as the publisheddescriptions of other workers. The inventoryavailable is still incomplete.The existenceof variousadditional firefinch vocalizations is indicatedby certain songsof the indigobirdsthat resemblethe whistledor twitteringquality of other mimeticsounds and that are sungin sdquencewith them. Until more completeinventories of firefinchvocalizations have been tape-recordedand describedin termsof theiraudiospectrographic characters, it will not be possibleto draw many firm conclusionsabout the homologiesor derivationsof the vocalizationsamong different species of firefinches.The firefinchvocaliza- tionsare of interestin the presentpaper mainly in providingan inventoryof 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 63 soundsas a backgroundto documentthe behavioralmimicry of the indigobirds. In the followingdescriptions of firefinchsounds I have attemptedto bring togetherthe variousterms used by other studentsof finch behaviorfor the samesounds, but in many casesthe lack of audiospectrographicdescrip- tionshave made it impossibleto know what soundswere described.The synonymousterms are listed in parentheses;synonyms in English are the namesof Harrison (1962a) or, if so designated,those of Goodwin (1964, 1969); names in German are those of Nicolai (1964, 1965b), and names in French are from Morel (1969). For convenienceof comparison,the mimeticvocalizations of indigobirds are shown near their firefinch song models rather than later in the audiospectrographs. Lagonosticta senegala 1. Alarm note (excitement note, Erregungsruf,cri d'alarme ou d'excita- tion): both sexesgive an abrupt alarm note when disturbed,a "chuk" or "zeck" (Audiospectrograph1 a-c). The call is givenin severalcircumstances includingdisturbance at the nest by man or indigobird. It is also given in agonisticencounters between adult male firefinches before a chase. The call is often given singly, sometimesin two's or three's, althoughI noted that a rapidsuccession was given when a male joinedin a groupof finchesmobbing a snakethat had apparentlyrobbed a nestin Nigeria(Audiospectrograph 1 a). 9.. Contact call (Distanzruf, cri de cohesion): the contact call is a note givenby malesor femalesin flocksor separatedfrom eachother. Firefinches in a groupcall back andforth with thisnote, and matestemporarily separated call until one flies to the other. As indicatedby the term Distanzru[the call maintainssocial contactbetween birds over distances,which may be more than 40 feet. Sometimesseveral of thesenotes are given in a sequence.A sharpbut weak versionof the contactcall is describedby Morel (1969: 104) as the cri de bataille. Structureof the contactcall note is variable, and it rangesin durationfrom 100 to morethan 300 msec. Contactcalls are clear,slurred whistles, a pleas- ing "twee" or "dwee"to the ear. The pitch usuallyrises smoothly, sometimes at the beginningand sometimesnear the end of the call, and the call may endby risingabruptly and then dropping gradually. Examples of contactcalls are seenin the representations.of the songsin Audiospectrographs1 b, c. fl. Song (Gesang,chant): songis composedentirely of an alarm note followedby two to sevencontact call notes;occasionally several alarm notes mayintroduce a song.The songis differentiatedfrom the contactcall mainly in the temporalpattern of a relativelystereotyped, repeated sequence of notes, and in the postureof singingmales (body upright,the bill tilted up, and the headfeathers fluffed), a singingposture shared by many estrildids(Moynihan and Hall, 1954; Kunkel, 1959; Gtittinger, 1970). 64 ORNITHOLOGICAL MONOGRAPHS NO. 11

7' Q 6-

4.-

b 6'

õ-

4.'

?- c 6'

õ-

?- d 6'

I 0 Sec .5 1.0 Audiospectrograph1. Vocalizationsof Lagonostictasenegala: a, alarm call, Zaria; b, song,L. s. senegala,Zaria; c, song,captive L. s. senegala;d, distresscall, 12-day nestling;e, wing whirr, captiveL. s. senegala. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 65 kHz 8.• 7'

4' $-

7' b 6-

I 8' 7' c 6'

7' d 6'

$- 2'

Sec. 1.0 1.5 Audiospectrograph2. Mimetic vocalizationsof Vidua chalybeata.a, alarm call, Olorgesailie;b, song, 20 miles east of Kisumu, centralis;c, song, Merensky, amauropteryx; d, song, Zaria, neumanni. 66 ORNITHOLOGICAL MONOGRAPHS NO. 11

kHz'

6' •3

ii ;.0 Sec Audiospectrograph3. Variations in mimicry of Lagonostictasenegala song by an individual male Vidua chalybeatarecorded four miles south of Penhalonga,Rhodesia. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 67

7' (l b 6'

$- 2' I,- 7'

2' I,- 7' 6'

2ß r(-'?•;.•l. ,,.:...-..ti•-½: t f!• - ½'•';''ii,!.J. : •.•.. ' ."" '•'•.'r½.." ,'•'!"' ;..•'h•'9 •."u'. 't '•*','•:•: -...' ' . ,.,,.'" ' ß

4 ,' .•t ." •,

0 See .• 1.0 1.5 Audiospectrograph4. Vocalizafions of Lagonostictasenegala: a, bcggiagcall of 5-6 daynesffing; b, beggingcall of 7-8 daynesding; c, beggingcall of 11day nesting; d, beggingcall of 13 day nesdings(two bbdsin nest); e, beggingcall of youngm familyparty; f, roost•gcall, flock call, or juvenilelocation call of inde•ndentyoung; g, nestcall of nest-build•gmale L. senegala. 68 ORNITHOLOGICAL MONOGRAPHS NO. 11

•.Hz

8' 7' 6'

I 8'

7' o

1• ,•:••.;i .... :.i:' •'•-,.•. ' ,.:•i•:;•:,.i•o,•i,•,•.•,=...... :. •. ,i ; ii'.•!i'

o &o;''' :• .... •.o .... i.•' ' ' Audiospectrograph5. Mimetic vocalizations of l/iduachalybeata: a, beggingcall, MonkeyBay, amauropteryx;b, alarm call and juvenilecontact call, Merensky, amauropteryx;c, beggingcall or nestcall, Sabi Valley, amauropteryx; d, beggingcall or nest call, Zaria, neumanni. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 69

kHz 8-

7' C

6' d

g h

6' i j k

I 0 See .5' 1.0 1.5 Audiospectrograph6. Vocalizations of Lagonostictarhodopareia: a, alarm call, Merensky; b, alarm call, nesting female at nest, Monkey Bay; c, alarm call, Sigor, Kenya; d, "ti-ti-ti," Marble Hall; e, "wee-et,"Merensky; f, "weee-eee,"captive male from Rhodesia; g, "sisisisi,"Merensky; h, seriesof whistles,captive Rhodesian female; i, prolongedwhistle, female, captive,Rhodesia; j, "tu-tu-tu,"Merensky; k, L-shaped whistles, Merensky. 70 ORNITHOLOGICAL MONOGRAPHS NO. 11

kHz 8'

6' b C•.

4' $-

2' I,: ?- k

0 Sec

Audiospectrograph7. Mimetic vocalizations of Vidua purpurascens: a, alarm call, Merensky;b, alarmcall, SabiValley; c, alarmcall, Sigor; d, "sisisi,"Merensky; e, "ti-ti-ti,"Merensky; g, "wee-et,"Merensky; h, "weee-eee,"Merensky; i, seriesof whistles,Sigor; j, prolongedwhistle, Sabi Valley; k, "wee-et"leading into "tu-tu-tu," Merensky;1, buzz-whistle, Merensky; m, L-shapedwhistle, Merensky. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 71

kHz'

! i 1.0 Sec

Audiospectrograph 8. Mimetic songsof one male Vidua purpurascensrecorded at Sabi Valley, Rhodesia. 72 ORNITHOLOGICAL MONOGRAPHS NO. 11

kHz

a 6-

b 6-

c d 6-

I I 0.1 1.0 Sec Audiospectrograph9. Mimetic songsof Vidua purpurascensat Penhalonga,Rhodesia. Songs a and b are from two birds, c and d are from a third. Several variations of these songs were recorded from each bird. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 73

kHz

o d

e f g

I.o Sec Audiospectrograph10. Mimeticsongs of V. purpurascensrecorded at MonkeyBay, Malawi. Note the resemblanceof h and i to the mimetic songsof other species(Audio- spectrographs15e, g; 1; 3). 74 ORNITHOLOGICALMONOGRAPHS NO. 11

kHz 8

I 8' 7' b 6'

4 $ /.//

I 8' 7

I 8'

,•- . ;

0 $.½ .• 1.0 •.• ' ' Audiospectrograph11. Mimicry of L. rhodopareiagiven by a maleV. chalybeata amauropteryxat Mererisky. a, alarm call; b, "sisisi;"c, "weee-et;"d, seriesof whistles;e, prolongedwhistle; f, "tu-tu-tu;"g, buzz-whistle;h, L-shaped whistle. Compareaudiospectrographs with Figure 22. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 75 kHz 8 •

7- o b C d 6'

I 8'

7' e f g 6*

6' 5'

I 8'

6' i j k I

I 0 Sec ,5 1.0 1.5 Audiospectrograph12. Vocalizationsof Lagonostictarubricata: a, alarm call, Tzaneen; b, alarm call, Nicolai, 1965b, 1965; c, trilling call, Zomba; d, courting screech,Ni½olai,• 1965b; e, "wee-ee,"slurred whistles, Zomba; f, "wee-ee,"Nicolai, 1965b; g, slurredwhistle, Nicolai, 1965b; h, descendingtrill, Zomba; i, j, rapidly descendingwhistles, Nicolai, 1965b; k, prolongedwhistle, Nicolai, 1965b; 1, warbled whistle, Nicolai, 1965b. 76 ORNITHOLOGICAL MONOGRAPHS NO. 11

c d e

f • •Ft* g •F'

i I O.I 1.0 Sec Audiospectrograph13. Vocalizations of a male Lagonostictarubricata recorded two miles from Lilongwe on the Lilongwe-Lukuni road, Malawi: a, alarm calls given in flight as bird flew over microphone into tree; b, trill; c, abrupt clink notes; d, e, wavering short whistles; f, g, down-slurred whistles; h, broken short whistles. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 77

kHz 8'

?- 6- O

I 8'

0 Sec .5 .0 1.5 Audiospectrograph 14. Mimicry of Lagonosticta rubricata vocalizations by three forms of indigobirds: a, alarm call, Panshanu "nigeriae"; b, trilling call, Tzaneen funerea; c, trilling call, Panshanu "nigeriae"; d, "ti-ti-ti," Tzaneen j•unerea; e, slurred whistles, Panshanu "nigeriae"; f, descendingtrill, Tzaneen funerea; g, descendingtrill, Penhalongacodringtoni; h, descendingtrill, Panshanu"nigeriae"; i, prolongedwhistle, Tzaneen funerea; j, whistle-warble, Tzaneen j•unerea; k, rapid warble, Panshanu " nigeriae." 78 ORNITHOLOGICAL MONOGRAPHS NO. 11

kHz

g h

0.1 1.0 Sec Audiospectrograph15. Mimetic vocalizationsof two Vidua [unerea nigerrimaon the Lilongwe-Likuniroad, Malawi. Note the similarityof the alarm calls or trills a, b, to the callsof localLagonosticta rubricata (Audiospectrograph 13 a, b) and to the alarmcalls of L. rhodopareia80 milesdistant at MonkeyBay (Audiospectrograph22 b). Also comparethe short waveringwhistles, d, and descendingwhistles, c, e, g, to thoseof local L. rubricata(Audiospectrograph 14) and to the songsof L. senegala and its mimics (Audiospectrographs1, 2, 3). 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 79

kHz

Q 6-

0.1 1.0 S ec Audiospectro•aph 16. Mimetic trills o• Yid•; f•nere• codringtoni: a is recorded •rom a male at Zomba, Malawi, b and d are from two males at Penhalonga, Rhodesia, and c and e are from another green Penhalonga bird. 80 ORNITHOLOGICAL MONOGRAPHS NO. 11

kHz 8' b c d 7'

I 8-

7 -

3- 2-1•

I 8-

7- f g

0 Sec .5 1.0 1.5 Audiospectrograph 17. Vocalizations of Lagonosticta larvata: a, alarm call, L. l. togoensis,Zaria; b, alarm call, Zaria, captive; c, "beri," Zaria; d, "beri," captive L. l. vinacea,Nicolai, 1965 b; e, plaintive whistle and "whee-hew," L. l. togoensis,Zaria; f, plaintive whistle, L. l. vinacea, Nicolai, 1965b; g, "hew, hew," L. l. vinacea, Nicolai, 1965b. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 81

kHz 8' (] 7' b r' c 6'

2 I 8

0 Sec 1.0 1.5

Audiospectrograph18. Mimicry of Lagonostictalarvata by indigobirdsin Nigeria: a, alarm call, Zaria; b, "beri," Zaria; c, plaintive whistle, Zaria; d, "whee, whee," Zaria; e, "whee-hew," Zaria; f, "whee-hew" variant, Zaria. 82 ORNITHOLOGICAL MONOGRAPHS NO. 11

kHz 8'

8' ?' c

?- d

I 8'

0 Sec .5 1.0 1.5 Audiospectrograph19. Lagonostrictarara and its indigobird mimics in Nigeria: a, alarm call, L. rara, Zaria; b, c, alarm calls, possiblyother calls, "wilsoni," Zaria; d, song, "wilsoni," Zaria; e, song, "camerunensis,"Zaria. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 83

kHz 8'

I 8' 7'

8' 7' d

0 Sec .5 1.0 1.5 Audiospectrograph20. Mimicry of nest calls or begging calls of firefinches by indigobird males: a, alarm call, whistle, and begging calls mimicked by Vidua purpurascens,Sabi Valley; b, V. [unerea codringtoni, Penhalonga;c, V. [. [unerea, Tzaneen; d, V. [. nigerrirna, Lilongwe; e, f, "camerunensis,"Zaria. 84 ORNITHOLOGICAL MONOGRAPHS NO. 11

kHz 8-

0 Sec .5 1.0 1.5 Audiosp½ctrogrsph21. Vocallzatioasof Lagonostictarufopi½ta: a, alarm cabs, Zsris; b, c, song, captive from Zaria. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 85

Songin L. senegalais primarilysexual in motivation(Kunkel, 1959, 1967; Morel, 1969; my observations)rather than aggressive.It is given most fre- quentlyby lone malesin the absenceof their matesand is suppressedwhen a conspecificbird is in sight of the male. Besidesthe song (= solitary song, UngerichteteGesang, chant solitaire), L. senegalaalso may have a sexual song (chant sexuelle) given by the male in courtship. This sexual songis describedby Morel (1969: 106) as very soft, simple,and tinklingor clinking in sound;it has apparentlynot been tape-recorded.Occasionally the female may sing. Song was tape-recordedin a captivefemale when its mate was removedfrom the cage,and in the pattern of notesthe songmatched that earlier givenby the male (Sullivan,1970: 21). Usually amongthe firefinches only the male sings. We recordedsong in SouthAfrica, Botswana,and Nigeria and in eachof theseareas the songwas quite variable in the numberof syllablesand in their duration and inflection. Most songs were 1.0 to 1.4 secondsin duration; songswith fewer syllableswere usuallycomposed of syllablesof longerduration. No consistentinterpopulation differences in this feature of song are apparent in my recordingsof L. senegalaor of their indigobird mimics. Audiospectrograph1 showstwo songsof L. s. senegala,a songwith three contactcall elementsrecorded at Zaria, Nigeria, and a six-syllabledsong from a captiveL. s. senegala.Morel (1969: 106) describessong in L. s. senegala at Richard-Toll, Senegal,as variable in syllablenumber (two to six). The songsof L. senegalaas mimickedby Vidua chalybeata(discussed under indigobirdsong dialects) vary from one populationto another,and in somecases thesesong variants are rather local. One differenceamong races of L. senegala recordedin captivitywas that manyL. s. senegalagave a songwith six contactcall notes,the first four alternatelylow and high in pitch (Audiospec- trograph1 c). This versionof the songhas not beennoted in captiveL. s. rendallior L. s. ruberrima. Any songdifference among subspecies of L. senegala is probablya complexone, however, since some songs of wild L. s. senegala lack the alternatingnotes (Audiospectrograph1 b). 4. "Stip" note: (undirecteddisplay note): a call that soundslike "sfip" is sometimesgiven during displayby a male with a feather or straw in his bill (Harrison,1962a: 268), thoughmy displayingmales usually did not vocalize. Harrison also noted a clickingsound when the feet of the displayingmale leave and return to the displayperch. 5. Distresscall (fear screech): three young 12 days of age removed from the nestgave a loud screech(Audiospectrograph 1 d) characterizedby a wide frequencyband of soundenergy. The siblingsremained silent in the nest;the parentsflew to a nearbyperch and utteredalarm notes. I notedthis behaviorseveral times in three nestingsin our house. The call was first heard on day 6, the day whenthe eyesopened. Adults occasionally gave a similar 86 ORNITHOLOGICAL MONOGRAPHS NO. 11 distresscall when handled. The distresscall was not heard in the field; it is presumablygiven when a predatorseizes a bird. Another "distresscall" (Cri de ddstresse)was statedby Morel (1969:104) to be givenby adultsafter their younghave disappearedfrom the nest. The call is plaintive,long, and sharp,apparently similar to a contactcall, which would be appropriatein callingtogether other membersof a social group. 6. Wing whir: when the male flies to the female with a feather in the bill at the time of the strawdisplay (Harrison, 1962a: 268) he makesa loud whirring noise with the wings (Audiospectrograph1 e). This mechanical soundmay be an importantsignal in courtship.On severaloccasions I have heard a short note much like the contactcall given during the wing whir. 7. Beggingcall (Bettelrufedes Jungvogels):the earliesttime at which the youngraised in captivitygave beggingcalls was on day 5 or 6 after hatching. In young of this age the beggingcall noteswere short (less than 50 msec)and well spaced(more than 400 msecapart in 10 syllablesrecorded) (Audiospectrograph4 a). By day 7-8 the samebird gave a call note of 80 msecduration and the beggingcall note developeda two-pitchcharacter with two non-harmonicallyrelated bands of soundenergy rising at differentpitches (Audiospectrograph4 b); this two-tonecharacter was lessclearly evidentin the 5 to 6 day-oldyoung. By day 11 in anotheryoung L. s. senegala(a male) the beggingbecame more rapid and the individualbegging notes were longer (more than 100 msec) and distinctly two-toned (see Audiospectrograph 4 c). This bird gave 66 of these beggingcall notes in 14 secondsof the recording;no other noteswere given. Young in the nest usuallybeg audibly only when the parentsvisit the nestor when the younghave not beenfed for hours. 8. Juvenilelocation call: in my captiveyoung L. s. senegalaon day 13 the simplerepetition of the beggingcall note gaveway to the irregularalternation of two kinds of notes;one was two-toned and longer than the earlier beggingnotes while the other was shorterand resembledthe alarm call of an adult (Audiospectrograph4 d). My recordingsof one bird at day 22, six days after fledging,were nearly identicalto a field recordingwhich I made of a juvenilein a family at Maun, Botswana,on 18 April 1967 (shown in Audiospectrograph4 e). The elementresembling an adult's alarm call appearsto be givenwith greaterfrequency when the youngbirds are disturbed. The dependentjuveniles have two distinctcalls given in alternation;the short call, whichis now recognizableas an alarm call and the longer,plaintive call. Well-leatheredjuveniles which I recordedin Nigeria as they beggedundis- turbedfrom an adult male gave only the two-tonejuvenile location call and no alarm calls;the sequenceof thesetwo callsmay dependon the amountof disturbanceof the young firefinches. Captive young birds once out of the nest called almost continuouslyfor two weeks. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 87

The juvenilelocation call of the independentyoung firefinchesrecorded beggingin captivityand in the field in Nigeria (Audiospectrograph4 I) continuedto developto a stagewhere it resembledthe contactcall of the adults. For severalweeks after the youngfed themselvesbut while they remainedin the family group, the call retainedthe two slurredwhistles rising in pitch at different rates. The structureof the call in youngL. senegalaof variousages clearly indicatesthat the contactcall notesand the songof adultsare ontogenicallyde- rived directlyfrom the beggingcall of the nestlings.It is of interestto note that both beggingcalls and contact calls are functionally similar in their socialfunction; both bring togethermembers of a socialgroup---a family, a pair, or a flock. A youngmale raisedin our housegave a full songat the ageof 14 weeks; at this time he courted his sister. Morel (1969) noted that wild Lagonosticta senegalain Senegalsometimes breed at the age of four months. 9. Nest call (Nestlocken,cri au nid): a very soft call inaudible at more than 10 to 20 feet is given by the male at the nest for a few days before laying, usuallywhen the female is near the nest. The nest call (Audiospectro- graph 4 g), recordedfrom a captive male L. s. senegalain our bedroom, when only one pair of L. senegalalived in our house,consisted of two notes irregularly altered in sequence,a whistled note resemblinga short contact call and an abrupt note similar to the alarm call or perhapsto the begging call of the youngestnestlings. The element resemblingthe contact note was variable in temporalpattern and pitch. Nicolai (1964: 144), however,has figured as the nest call a seriesof singlenotes, each of which resemblesthe notesof the beggingcall of young firefinches,rather than the more complex pattern recordedhere. The call picturedby Nicolai was a mimetic call given by an indigobirdand not by a nestingfirefinch itself. Evidentlythis firefinch specieshas two kindsof nestcall, one of themnearly identical to the begging call of the young. The similarityof this matingcall with the callsof the young suggeststhat a female may derive from the soundan associationwith her own life as a nestling,and this may attracther to the nest;at any rate, this versionof the nest call appearsto be derivedfrom the beggingcall of the young. The whistledelement of the nest call of L. senegalawas once tape- recordedduring a wing-whirdisplay of one of my captivemales, confirming the associationof this call with courtshipbehavior. Other casesof what Eibl-Eibesfeldt (1970: 196) has called "infantile behavior" are known in the courtshipof estrildidfinches. In the Blue Waxbill (Uraeginthusangolensis) the nestcall resemblesthe beggingcall (Goodwin, 1965: 288). In the Diamond Finch (Emblemaguttata) the courtingmale givesa displayin the postureof a beggingyoung finch (Nicolai, 1965c: 35); a similarposture is usedby the courtingmale Uraeginthusspecies (Goodwin, 88 ORNITHOLOGICAL MONOGRAPHS NO. 11

1965: 288). In severalother estrildids the male useshead-swaying or tongue- waggingmotions like thoseof the beggingyoung when he executeshis courtshipdisplays (Giittinger, 1970: 1053-1057).

Lagonosticta rhodopareia 1. Alarm call: when L. rhodopareiais flushed, disturbed at its nest, or placedin a smallcage the call givenis a rapid seriesof harsh,stuttering trills (Audiospectrograph6 a, b, c). The call may be represented"trrrrr-trrt- trr-trrt" (Goodwin, 1969: 92). As many as 34 notesmay be givenin two seconds,with the notesusually divided into shorterseries, slightly separated. Variation in alarm calls both within a stutter and between stutters occurs, especiallyin loudnessand pitch of the harshnotes. Someof thesenotes appear to gradeinto the trills of the contactcalls. The alarm call is characteristic of the specieseverywhere and recordingsof L. rhodopareia]amesoni in Transvaal and Malawi are indistinguishablefrom the alarm call of the male L. r. rhodopareiaI recordedin northernKenya. 2. Contact calls and song: the same seriesof noteswere used for maintaining contactbetween separated birds and in song when lone males were held in isolation. The only differencenoted between contact calls and song is the greaterpersistence of lone malesin singing. The two sexesoften have distinctlydifferent vocalizations,though I have heard a female at Mererisky give the calls of her mate when he was shot. Contactcalls are quite varied and includeboth slurredwhistles and trills, includingthe followingversions which I have recorded. (a) A rapid trill of notesresembling the alarm call stutter but charac- terized by an inverseS-shape on the sonagrams;the trill may be given as "sisisisisisi."This call was recordedin Transvaaland Rhodesia;Audiospec- trograph 6 g representsthe call of a bird recordedat Merensky. (b) Slow trills of simple ascendingwhistled notes; the notes are often slurredon the high ending;"ti-ti-ti-ti-ti" (Audiospectrograph6 d). The call shown was recorded at Marble Hall, Transvaal. (c) Rapid trills of the same syllables,given about 14 per second; "titititi." (d) "Weee-eee";clear, ascendingsimple whistles. Audiospectrograph6 [ givesthe call of a lone captivemale from Rhodesia. (e) "Wee-et,wee-et"; two-parted slurred whistles. Audiospectrograph 6 e givesthe call of a bird at Merensky. (f) A seriesof whistlesof 50 to 80 mtec duration each at a pitch of about 3 kHz. The call shown (Audiospectrograph6 h) is from a captive Rhodesian female. (g) A prolongedwhistle, a noteon onepitch or rising•,lightly and up to a secondin duration. This was heard in the field in Africa. It was given 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 89 repeatedlyalso by a femaleperched in a tree after escapingfrom a cageand by captivefemales in my aviaries(Audiospectrograph 6 i). The call is inter- pretedby Goodwin (1969: 92) as "feeee." (h) Slowtrills of simpledescending whistled notes, "tu-tu-tu-tu." Audio- spectrograph6 ] showsa slow trill recordedat Merensky. (i) A seriesof noteswhich beginwith a slurredwhistle and end with a buzz. This was heard from several males in Transvaal and Rhodesia but has not beenrecorded; the mimetic(indigobird) version is shownin Audiospec- trograph 7 I. ( j ) A whistleof about80 msecfollowed immediately by a risingsharp note;series of thesephrases sound like a slowtrill with eachnote ending abruptly.The audiospectrograph(Audiospectrograph 6 k) showsa reversed L-•hape;the bird was recordedat Merensky. In songseveral motifs are given;each one is usuallyrepeated several times. No regularlyrepeated sequence of motifsis evident. Goodwin (1964: 102- 103) haspointed out that one cannotdraw a distinctline betweenintra-pair callingand male singingin this firefinchsimply on the basisof the vocalizations involved;rather the contextand posturalbehavior of the bird as it vocalizesdetermine whether or not it is singing. 3. Chip: a high (8-9 kHz) note given singlyis evidenton Nicolai's (1965b) record. 4. "Squelch"note: givenwith the strawdisplay by male ]arnesoniis a notethat soundslike "fwit," a soft,splashing sound, covering 0-2 kHz and terminatedby a soft click. Harrison(1962a: 265) describesthe "squelch" as"similar to thesound made when air is suddenlysqueezed from wet sponge- rubberwith a rushof tiny bubbles.... "I haveheard this only in courting males.The noteappears to resembleclosely the squelchnote of "L. rubricata" (Harrison,1962a: 268); probablyHarrison's birds were in factL. rhodopareia (Harrison,1963a). An abrupt,soft "stip" call on Nicolai's(1965b) record resemblesthe "stip"note described for L. senegalaand maybe a versionof the "squelch." 5. Distresscall: my captivebirds held in the handgave a loudsquawk similar to that of L. senegala;the call was not heard in the field. 6. Beggingcall: the beggingcalls of two youngL. rh. ]arnesonithat were rearedsuccessfully in my aviariesin September,1971, were recorded fromday 12 to day33 (the lastday that they were heard to beg). The calls wereaudiospectrographed too late for illustrationin this report;they were very similarto the mimeticbegging calls of the indigobirdsshown in Audio- spectrograph20 b-f andthe secondhalf of 20 a. The lower-pitchedcalls of the 12-daynestlings resembled closely Audiospectrograph 20 e, and the highercalls of older]arnesoni nearly matched Audiospectrograph 29 a andb. 90 ORNITHOLOGICAL MONOGRAPHS NO. 11

7. Soft contactcalls: a pair of L. rh. ]amesoniin captivitygave a series of short notesof low volumewhen disturbedat the roostingsite. The male gave a high note which droppedabruptly from 7 to 5 kHz with the total duration about 40 msec. The female had a lower note (3.5 to 1.8 kHz) which was prolongedat the end, producinga low "chu" sound. I recorded and audiospectrographedthese calls but becausethey were very softthey were somewhatindistinct and are not reproducedhere. 8. Nest call: Goodwin (1969: 92) describedthe nest call of a male L. rh. ]amesonias a soft, husky "tchu-tchu-tchu-tchu."I have heard this in one nestingpair of ]amesoni,but the call was too weak to be tape-recorded. 9. Bill click: when a female was released in our house for the first time the residentmale flew to her and gave a greetingdisplay (Goodwin, 1964: 89-91, describesthe display) on the ground besideher with his tail pointedtowards her, hoppedaround her, and gave an audibleseries of bill snapshere designatedas a bill click. The bill click was also given by the male as he accompaniedthe femaleat her feedingarea; he often bobbedin a courtshipdisplay (but withoutany featheror straw)while he was bill clicking. Vocalizationsof L. rhodopareiawhich can be heard on Nicolai's (1965b) recordingare the alarm call, "ti-ti-ti-ti," "titititi," and the "stip" note. J. Nicolai hasinformed me that thesewere recordedfrom a captivemale of the form L. rh. rhodopareia.A few other variant calls (includingthe nest call) not recordedfor the firefinchitself, but which are sungby its mimic V. purpurascensare describedin the sectionon indigobird vocal mimicry.

Lagonosticta rubricata 1. Alarm call: The alarm call is a somewhatharsh, nasal sound "tchit" givenusually as a singlenote (Audiospectrograph12 a). As describedby Goodwin (1964: 97) and Immelmann et al. (1965: 182) L. rubricam gives thisnote singly, in pairs,or in longerseries depending on whetherfear (single "tchits") or distress(series of calls) is the dominantmotivation. As recorded by usin the field in Malawi thisnote corresponds to a pairednote on Nicolai's record (Audiospectrograph12 b) which showsmore clearly the restriction of most soundenergy to the 3-5 kHz bands. 2. Trilling call: Goodwin (1964: 98) describeda prolonged,harsh trill as an aggressiveor threateningvocalization, referring apparentlyto the same kind of trill I recordedfrom a wild bird near Zomba, Malawi (Audiospectro~ graph 12 c). I recordedit also from a male at Lilongwe,Malawi, when it flew to a tree next to an indigobirdmimicking this species(Audiospectro- graph13 a, b). As notedby Goodwinthe trilling call resemblesthe alarm call of L. rhodopareiain soundthough it is givenmore slowlyby L. rubricam. 3. Contact calls and song: as in some other estrildidsseveral calls are givenbetween mates and functionas contactcalls, and the samemotifs are 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 91 usedby singingmales as elementsof song (Goodwin, 1964: 102). These are describedbelow on the basisof their structure;some may have other signal value. (a) "Ti-ti-ti-ti," a trilled seriesof ascendingnotes with a terminalslur, as in L. rhodopareia,but higher in pitch. The trill was not tape-recorded from L. rubricata, but I heard a male firefinch of this speciesgive it at Tzaneen and at Zomba. (b) "Wee-ee,wee-ee"; a slurredwhistle given repeatedly. In a seriesI recordedfrom a male at Zomba,a seriesof simpleup-slurred plaintive whis- ties was followed by a seriesof the samenotes each drawn at the end into a down-slur(Audiospectrograph 12 e). Nicolai (1965b) has recordeda similar seriesof two kinds of slurredwhistles in a captiveL. rubricatahaernato- cephala, a faster note of 180 msec which descendsat the end and a slower- rising S-shapednote of about 300 msec endingon a high pitch (Audio- spectrograph12 f, g). The slurredwhistles vary betweensongs. Some of the variantsof thesecalls may correspondto Goodwin's(1964: 101) "chub" and "chwe"calls and to McLachlanand Liversidge's(1957: 449) "wink" notes. (c) Descendingtrill: I recordeda descendingtwittering trill from a singingmale L. rubricatacollected near Zomba (Audiospectrograph12 h). This is one of the more distinctivevocalizations of the species.I have heard it alsoin Transvaal,Rhodesia, and Nigeria. Harrison (1962a: 264) described a phraseresembling the first half of the songof a Willow Warbler (Phyllo- scopustrochilus), which has, however, a muchslower song than L. rubricata's descendingtrill, but Harrison'sfirefinches may have been a differentspecies (Harrison, 1963). (d) Shortwhistle. An abrupt,hard note sometimes is precededby a clear monotonicwhistle (Audiospectrograph12 ]) and at other times is followed by a short,low whistlednote (Audiospectrograph12 i). Theserecordings are taken from Nicolai's (1965b) record. (e) Prolongedwhistle. A prolongedclear whistleof up to a secondin durationis sometimesgiven; this variesin durationand in pitch. This is probablythe samecall describedas "fee6" by Goodwin(1964: 101). N,icolai (1965b) hasrecorded a burryversion of thiscall (Audiospectrograph12 k). (f) Rapid warble. A modulatedor warbledcall heard in the field resembles a recordingof Nicolai (Audiospectrograph12 l). An individualbird may giveseveral variants of this call (Audiospectrograph13 d, e, h). Some callsrecorded bridge the gap betweenthe prolongedwhistle and the rapid warble. zg. Nest call: this call is indicatedat the end of Nicolai'srecording; it was not heard in the field. It sounds like the vocalizations of other firefinch speciesas shownin their mimickedforms in Audiospectrograph20. 92 ORNITHOLOGICAL MONOGRAPHS NO. 11

5. Courtingscreech: a harshscreech is givenby courtingmales to females (Goodwin, 1964: 101); Nicolai (1965b) recordedthe call shown in Audiospectrograph12 d. 6. Squelchnote: a soft but harsh "fwit" note followed by a metallic "clink" is givenby malesin the straw display (Goodwin, 1964: 100). 7. Distresscall: an "unpleasanthusky screech" is givenby youngbirds seizedin the hand, and adultshave a deepernote. This call was describedby Goodwin (1964: 100).

Lagonosticta larvata 1. Alarm call: the alarm call is givensingly or repeatedlywhen birds are disturbedin the field or are held in small cages. It is a sharp, abrupt note whichrises rapidly in pitch (Audiospectrograph17 a, b). Most of the sound energyis at about 5 kHz. Harrison (1962a: 267) has describeda seriesof these notes as "dwit-it-it." 9•. Contact calls and song: no clear distinctioncan be made between contactcalling and singing.Most vocalizingL. larvatathat I heard in Nigeria appearedto be singing. The following vocalizationswere recorded both whenlone malessang for more than 10 minutesand whenpairs were perched together. A lone male singingin a tree near Zaria maintainedan erect song posture much as in L. senegala. Harrison (1962a: 267) recognizedtwo phrasesin L. I. vinacea,a clear whistled"beri-beri" and a lower "hew-hew- hew." The "beri-beri"call is nearlyidentical in L. I. togoensisthat I recorded in the field in Nigeria (Audiospectrograph17 c) and in L. I. vinacearecorded by Nicolai (1965b) in captivity (Audiospectrograph17d). The motif is similarthough generally of longerduration than an apparentlyhomologous motif of both L. rubricataand L. rhodopareia.A secondphrase noted in Nigerian birds (Audiospectrograph17 e) is a "whee-hew,whee-hew" motif. This resemblesin its descendingelements the "hew-hew" of L. I. vinacea (Nicolai, 1965b) (Audiospectrograph17 g). A third phrasein this firefinch is a plaintivewhistle which first descends,then ascends;this gradesinto the ascendingportion of the "beri-beri" motif and is apparentlythe same in vinaceaand togoensis(Audiospectrograph 17 e, l). The songof the nominate form L. I. farrata of northeasternAfrica, a subspecieswhich is morphologically intermediatebetween vinacea and togoensis,appears not to have been recorded. 3. Squelchnote: malesperforming a straw displaygive a squelchnote similarto that of L. rhodopareia,according to Harrison (1962a: 267; 1963). 4. Distress call: adults that I removed from a mist net and held in the hand uttered a squawkingcall soundingjust like the distresscall of other fire finches. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 93

Lagonosticta rara 1. Alarm call: a sharp,nasal "chek" note is givensingly or repeatedly. Flushedwild birds(Audiospectrograph 19 a) and a captivemale, recorded by me in Nigeria,had a harshcall which quicklyrose and fell, and which differedfrom the simpleascending alarm call of mostother firefinches. The recordingagrees with descriptionsof Harrison's(1962a: 266) "keeyh"excitement note and Nicolai's (in Immelmann et al., 1965: 200) Alarmru[. 2. Contactcalls and song: Harrison (1962a: 266) notesthe songto be variable,usually consisting of three parts---"afew separate,low-pitched notes,"notes "rather higher in pitch and a little hoarse,"and finally "long trillingrepetitions of ratherhigh-pitched harsh notes." Nicolai's descriptions (in Immelmannet al., 1965: 200) seemto corresponddosely. In Nigeria birds in the field sanga similar song,and the harsh chipperingtrills were mostconspicuous. A dose approximationof the songis apparent(in Audio- spectrograph19 d) of the mimeticvocalization of someNigerian indigobirds; however, the song as given by the firefinch was not recorded. J. Nicolaihas playedfor me his tape recordingof a captiveL. rara with somelong whistlesin its song. The whistlesare longer in durationthan the vocalizationsI heard in Nigeria; the calls of his bird were given on a single pitch and soundedabout as long as the contactcall notesof L. senegala. 3. Undirecteddisplay note: apparentlyhomologous with the squelchor "stip" note of other firefinchesis a barely audible hoarse blowing sound, accordingto Harrison (1962a: 263, 266, 268). zg. Distresscall: Adults that I handledin Nigeriagave a squawklike that of L. rhodopareia.

Lagonosticta ru[opicta 1. Alarm call: an abrupt call somewhatmore nasalin tone than that of L. senegala,the alarmcall is illustratedin Audiospectrograph21 a. A single call wasusually given by eachof four birds,male and female,that I caught in Nigeria and held in an aviary. 2. Contactcall: one call givenby a pair of captivesin Nigeria was a brief slurredharsh note 4-5 kHz in pitch. 3. Song: a jumbleof metallicnotes comprising the songrecorded in a malecaptured at Zaria, Nigeria (Audiospectrograph21 b, c) differsfrom the songsof mostother firefinches in the absenceof runsof the samerepeated motif. The songis variedalso by variationof pitchof the samenotes. Rapid changesin pitch at the beginningsand endingsof the whistledphrases appear to accountfor the metallicor nasalsound, and harmonicsare presentmore than in other firefinches. 94 ORNITHOLOGICAL MONOGRAPHS NO. 11

4. Wing whir: the male flies to the female with feathersor grassin the bill with a loud wing-whirring;the whir hasnot beennoted for speciesof firefinches other than L. rufopicta and L. senegala (Harrison, 1962a). 5. Distresscall: a captive that I seizedin the hand uttered a harsh squawking call.

VOCAL MIMICRY IN INDIGOBIRD SONG Singingindigobirds mimic the songsand calls of the firefinches;these mimetic songsinclude whistles and trills indistinguishablefrom those given by the firefinchesthemselves. The first record of the resemblanceof the songof a viduineto that of its host was apparentlythat of WoRers (1960: 24), who heard captive V. c. chalybeatasing a song similar to that of the risinginflections of L. senegala.Wolters also heard a male V. wilsonising a differentkind of whistledsong. This was apparentlythe first observationthat different kinds of indigobirdsmay have different songs,as earlier Benson (1948: 63) and Friedmann (1960: 70-71, 76, 82) had reported that the indigobirdsall sangalike. Vocal mimicryof the fosterspecies by the viduine fincheswas first clearly appreciatedby Nicolai (1961), who likewisenoted similarityof the songof a captiveV. chalybeatato that of the firefinchL. senegala. The calls so closelymimicked thoseof the firefinch he thought at first a firefinch itself was calling. He subsequentlydiscovered that mimicry of the host songoccurs in captivityin most speciesof viduine finches,although V. chalybeatawas the only indigobirdfor which this was reported and documentedwith audiospectrographs(Nicolai, 1964, 1967, 1968, 1969). In my field work the specificityof vocal mimicry of the indigobirdswas studiedin detail; most of the morphologicallydistinct forms were recorded in the field. Audiospectrographicanalysis of the recordingsshows a remarkable similarityin the resemblanceof the indigobirdsto the firefinchesin their vocalizationsand provide someindirect evidencefor the ontogenicdevelop- ment of mimicry.

Vidua chalybeata The most widespreadindigobird species,the Village Indigobird (Vidua chalybeata), mimics the SenegalFirefinch, L. senegala,almost exclusively. Of the 164 individual Village Indigobirds heard singing in the field, 163 mimickedthe vocalizationsof L. senegala.Examples of vocal mimicry are shownin Audiospectrographs2, 3, and 5 (comparewith Audiospectrographs 1 and 4 of L. senegala). Audiospectrographswere made of each different mimeticphrase of 50 indigobirdswhose songs were tape-recorded.The alarm calls, contactcalls, songs,and beggingcalls of L. senegalawere all sungin eachof the populationsof V. chalybeatastudied. The mimeticsongs and calls are similar to the firefinch vocalizationsand are indistinguishablefrom them 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 95 to the human ear. In playbackexperiments to captiveL. senegalathe fire- finchesrespond to the mimetic phrasesin the same manner as they do to their ownvocalizations. When I playedthe mimeticbegging call to a pair of firefinchesa week after they had successfullyfledged their own young,the two parentsflew to the taperecorder; the two youngdid not respond.Nicolai (1964: 153) has seencaptive firefinches respond to mimeticcalls. Young firefinchesdisturbed by a changeof locationresume their beggingwhen the mimetic firefinch song of an indigobirdis broadcast,and adult firefinches respondto the mimetic contactcall by giving the contactcall themselves,a responsewhich I have observedalso. The appropriatebehavioral responses of the firefinchesto thesemimetic phrases confirm the likenessof the indigobird mimicry to firefinchvocalizations; the responsesof captivebirds, however, do not necessarilyshow the function of vocal mimicry. Alarm call-mimeticnotes were usually given in a series(Audiospectrograph 2 a) althoughthey were given singlyalso. Mimetic alarm calls also intro- ducedthe mimetic songs,as the model calls do when given by the singing firefinchesthemselves (Audiospectrograph 2 b, c, d). The mimeticsongs of individualindigobirds were generallysimilar to each other thoughthey sometimesvaried in the numberof contactcall elements. Few large-scaleregional differencesin mimicry are evident in song,nor are they evidentin the recordingsof the songmodel firefinchesthemselves. The forms arnauropteryxin Transvaal,Rhodesia, and Malawi, centralisin Kenya, and neumanniin Nigeria as well as birds at Botswanaand imported captive nominate chalybeata (4), neumanni (2), ultramarina (2), and amauropteryx (8) in my aviaries all have mimetic songsof three or four contact notes, althoughin some of these, mimetic songsof five or more syllableswere heard or recorded as well. Both the model songsof L. senegalaand the mimetic songsof V. chalybeataare variable in pitch, duration, and changes in pitch in the individualsyllables. Individual male indigobirdsgave many variantsof mimetic song,not a singlestereotyped sequence of notes (Audio- spectrograph3). The beggingcall in all birds recordedresembled the call given by young L. senegalaof 12 daysor more of age (Audiospectrographs4, 5), neverthat of youngerbirds. Boththe alarm call element and the incompletely developed juvenilelocation call were given;the sequenceof thesecalls was variableas in the youngfirefinches. Also recorded were more prolonged, mature juvenile location calls similar to those given by independentjuvenile L. senegala (Audiospectrographs4 1, 5 b). Two other mimeticcalls are somewhatpuzzling. Audiospectrograph5 c showsa seriesof mimeticalarm calls irregularlyalternating with more prolonged, whistled calls resemblingthe immature juvenile location calls, recordedfrom V. c. arnauropteryxat Sabi Valley. The sequencehere is similar 96 ORNITHOLOGICAL MONOGRAPHS NO. 11 on the one hand to the fledglingL. senegalain a family party recordedin Botswana,but it also resemblesthe adult firefinch nest call (Audiospectro- graph4 g), and in fact thesetwo setsof callsare nearlyidentical (see p. 87). Also recorded in a male at Zaria was a series of notes each of which had two nonharmonicbands of energyas well as harmonicsof these;the notesrose slowly then rapidly and terminatedon a high slur or sometimesdropped (Audiospectrograph5 d). This recording appears identical to Nicolai's (1964: 144) recording of a V. chalybeata;Nicolai identified this as the mimeticnest call of L. senegalabut did not includean audiospectrographof the call as given by the firefinch. This version was noted in my recordings only in wild V. c. neumanniin Nigeria and in captiveV. c. ultramarinafrom Ethiopia. Very similar to this call is one given in severalother speciesof indigobirds (Audiospectrograph20). Each recordedsequence of indigobirdsong of three minutesduration or longerwas analyzedby ear and in part by audiospectrographfor the number of each of the distinctmimetic vocalizations. As determinedfrom songse- quencesin 50 birds the most frequentlyimitated firefinch vocalizations are the contactcalls and songs(40 birds) and alarm calls (38 birds), although in numberof syllablesand in total durationthe call most often mimickedwas the beggingcall. These long sequencesof beggingcalls are similar to the pattern of calling of the juvenile firefincheswhich sometimescall without interruption for several minutes. A singlemale at Mereriskywas the only V. chalybeatain the field heard to mimic a firefinchother than L. senegala.The bird, an adult in the greenish- blue breedingplumage and with the bright coral bill and feet characteristicof V. c. amauropteryx,was first seenon 31 December1966 singingat a regular call-site. For the next two dayshe was seenhere continually;the bird mimicked L. rhodopareia.The songwas recordedon 2 January. The bird was not disturbed;only one female was seento visit the male up to the time of his disappearance,uncollected, three weeks later. Recordingsshow the char- acteristicL. rhodopareiaalarm call and sevencontact call and songelements, five of whichwere recordedas well in L. rhodopareiaat Mererisky(Audio- spectrographs6, 11). No notesin the five minutesof recordingsnor in the hoursof songheard in the field resembledthose of L. senegala,nor were any mimeticnotes intermediate to the notesof the two firefinchspecies.

Vidua purpurascens All whitish-footedindigobirds in southernAfrica (south of Lake Malawi) and in easternAfrica (Kenya) that were heard to singmimicked the vocalizations of L. rhodopareia. The mimetic alarm call was similar in all of these birds in Transvaal, Rhodesia,Malawi, and Kenya. Audiospectrographs6 a, b, c, d and 7 a, b, c showsamples from Mererisky,Sabi Valley, and Sigor 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 97 recordedfrom song models and from the mimics which subsequentlywere collected. V. purpurascensgives a wide variety of whistlesand trills which mimic the call notesand songof L. rhodopareia.Calls of V. purpurascenswhich clearly mimicthe calls recorded and audiospectrographed of L. rhodopareiaat Meren- sky, Transvaal,are the calls describedfor that firefinch as "sisisisisisi,""ti-ti- ti-ti-ti," "titititi," "weee-eee,""wee-et, wee-et," "tu-tu-tu-tu,"the prolonged whistle,and the "L-shapedwhistle" (Audiospectrographs 6, 7). Also mimetic were the "whistle-buzz"(Audiospectrograph 7/) heardbut not recordedfrom L. rhodopareiaat MerenskyReserve and alsoa mimeticcall of severalshort whistlesfollowed by more prolongedones (Audiospectrograph7 ]) somewhat resemblingthe recordingshown of L. rhodopareiain Audiospectrograph 6 i. Other calls (slurred whistlesand slower whistled notes) recorded in V. purpurascensalso appearto be variantsof firefinchthemes. Regionalor local differencesin vocal mimicry of songwere evident in this species,though some other vocalizations were widespread. Trilled and slurred notes,audiospectrographically similar in generalform in at leastfour localities recorded,included the calls "titititi," "ti-ti-ti-ti," "weee-eee,"and a seriesof short whistles. Most localitieshad at least one firefinch-likecall unique to songsamples of that area. An exceptionwas Sigor,Kenya, where all calls recordedfrom male purpurascensmimicking L. rh. rhodopareiawere very similar to the calls of southern African L. rh. jamesonL Birds recorded at Merensky,Sabi Valley, Penhalonga,and Monkey Bay had no mimetic songs in common,and althoughvocalizations were recognizably of the samegeneral form (Audiospectrographs7, 8, 9, and 10), they all differedin detail. These local, characteristic,mimetic songswere usually shared among most local male V. purpurascens.Unfortunately only a few firefinchesthemselves were recorded,and I do not know directlyfrom recordingsof the firefinchesthem- selveswhether their songsare as local as thoseof their mimics. The mimetic purpurascenssongs recorded at Penhalongaand Monkey Bay differedfrom the mimeticsongs at Merensky,as theyoften had an abruptintroductory note or seriesof notesfollowed by a seriesof three or four prolonged,slurred whistles (Audiospectrographs9 and 10). Beggingcalls or nest calls were heard in most recordedsequences of V. purpurascens;a series recorded from a male collectedat SabiValley (Audio- spectrograph20 a) showsa typicalpattern in mimeticsong of severaldifferent seriesof mimetic phrasessung in succession.These mimetic calls are similar to thoserecorded in other speciesof indigobirdsin the time/frequencychar- acteristicsof the syllables(Audiospectrograph 20) and to the beggingse- quencesof L. senegalain the irregularalternation of shortnotes and prolonged,two-tone notes. This seriessuggests a developmentof the contact callsfrom the beggingcall in L. rhodopareiaas well as in L. senegala.The 98 ORNITHOLOGICAL MONOGRAPHS NO. 11 mimeticbegging calls closelymatch the beggingcalls that I have recorded and audiospectrographedfrom youngcaptive L. rh. jamesoni.

Vidua runetea The VariableIndigobird, V. funerea,mimics several motifs that I recorded from the firefinchL. rubricata. Audiospectrographsof the mimic phrasesare shownin Audiospectrograph14 as are other soundswhich probably mimic other vocalizations of this firefinch but which have not been recorded from the songmodel. Indigobirdsong phrases which clearlymimic the firefinch are the alarm call, the trilling call, the warbledwhistle, and the descending trill. This last vocalizationin V. •. [unereafrom Transvaalis similar in timing and sequenceto the trill of L. rubricatarecorded in Malawi (Audiospec- trograph 12 h) and heard from the firefinch itself in Transvaal. Bluish- plumagedV. [unereaheard at LusituRiver in easternRhodesia also sang a descendingtrill, but thesebirds were not tape-recorded.In V. [. codringtoni the trills differedin tempo and pitch and in a male recordedat Malawi the trillsare quitevariable (Audiospectrograph 16); someare similarto the firefinch modelrecorded there (Audiospectrograph12 h). Codringtonimimetic songswere curiouslyhigh-pitched compared with the mimeticsongs of other kinds of indigobirdsthat mimickedL. rubricata. No simpleslurred whistles or low trills were recordedfrom five codringtonimales at Penhalonga,Rho- desia,nor from a greenish-bluemale recordednear Zomba, Malawi. Prolonged(over 0.2 seconds)whistles with variabledegrees of modulation were recordedin vocalizationsof typical funerea,codringtoni, and nigerrima (Audiospectrographs14, 15, 16); theseresemble the prolongedwhistle and the warbled whistle (Audiospectrograph12k,/) of L. rubricata. Other whistlednotes recorded from codringtoniinclude a high, rapid, tinkling trill of nearly 8 kHz in pitch. Mimetic calls of the purplish-blue,white-footed indigobirds (V. [. nigerrima) at Lilongwe,Malawi, were similarto the callsof L. rubricatarecorded in the samelocality (Audiospectrographs15 a and 13 a; 15 b and 13 b; 15 d and 13 d; 15 ! and 13 c; and 15 e, g are downslurredwhistles as is 13 g). Thesesongs are somewhatsimilar to the alarm and contactcalls of L. rhodopareia,and beforethe vocalizationsof theseLilongwe birds were audiospec- trographedthey were misidentifiedas mimickingthis latter species(Payne, 1968a). The differentvocalizations recorded in the populationsof L. rubricata and V. [unereamay involvelocal or regionalgeographic differences in songand calls amongthe firefinches.Some of the vocalizations,on the other hand, are recognizablein differentareas, and theseinclude the alarmcalls. A trill that I tapedin Nigeria (Audiospectrograph14h), givenby a greenmale "nigeriae"of the speciesV. wilsoni) that mimickeda call I heardthere given 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 99 by a male L. rubricatapolionota itself was the call most dosely resembling Nicolai'srecording of a trill of L. rubricatahaematocephala from south-central Africa (Audiospectrograph12 i, j), and thus one complexfirefinch vocalization is similar in remote areas. Mimetic beggingcalls or nest calls of V. funereaare rather similar to those of otherkinds of indigobirds(Audiospectrograph 20). Probablythese mimic the youngof L. rubricata. In a seriesof callsof a V. f. codringtoniin Rho- desiathe higherfrequencies became progressively louder and the seriesgraded into a rapidly modulatedcall with a carrier frequencyof 7-8 kHz (Audio- spectrograph20 b); this seriessuggests that modulatedcall notesof the firefinch (and its mimic) may developfrom the beggingcall notes.

Vidua wilsoni For reasonsdiscussed later, all of the Pale-wingedIndigobirds of west Africa and the remainingregions between the Saharaand Red Sea and the Congoforests and northernKenya desertare regardedas a singlespecies, V. wilsoni. In this broad sense V. wilsoni includes the forms "wilsoni," "camerunensis,"and "nigeriae." Each of these forms has at times been regarded as a species(e.g., Mackworth-Praedand Grant, 1949; Bannerman, 1948; Payne, 1968b), but in the presentsection these names are intended insteadto indicatethe appearanceof the breedingplumage of the males ("wilsoni"is purple,"camerunensis" is blue, and "nigeriae"is green). All threepurplish ("wilsoni") Pale-wingedIndigobirds recorded in Nigeria and also two other purplishmales heard and collectedmimicked L. rara. Analysesof the recordingsshow mimicry of the characteristicalarm call of L. rara (Audiospectrograph19 a, b). It is possiblealso, on the basisof the songdescriptions of L. rara, to refer severalother vocalizationsof "wilsoni" to this firefinch. A recurringmotif is a seriesof low notesfollowed by variable high notesand thenprolonged repetitions of sharphigh notes;these last were the most distinctivefeature of "wilsoni" song mimicry in the field. Beggingcalls or. nest calls were heardbut were not recordedon tape during 20 hours of recording effort. Most bluish ("camerunensis")Pale-winged Indigobirds that I heard in Nigeria mimickedthe calls of the Black-facedFirefinch, L. larvata. Of the nine bluish males recorded seven had the dear slurred whistles of this firefinch. Motifs of L. larvata that were mimicked included "beri-beri," "hew- hew," and "whee-hew"(compare Audiospectrographs 17 and 18), and as in the firefinchitself many variationson thesethemes were given. Also clearly mimeticof Black-facedFirefinches were the alarmcalls (Audiospectrographs 17 a and 18 a, b). Two male "camerunensis"which mimicked L. larvata gavecalls like the undistinctivebegging calls or nest calls of other species (Audiospectrograph20). In onebird a sequenceof thesenotes graded into 100 ORNITHOLOGICAL MONOGRAPHS NO. 11 sharp,whistled notes resembling a foreshortenedversion of the elementsof song;this gradationsuggests that in L. larvataas well as in its mimic the contactcalls may develop from the beggingcall. Two otherbirds, one greenish- blueand one quitegreen, at Bauchiand Panshanumimicked L. larvataalso, as did several blue "camerunensis" in these localities. One bluishindigobird at Zaria and also one bird which was the greenest indigobirdtaken at Zaria mimickedthe Black-belliedFirefinch, L. rara. In- asmuch as each bird was collected minutes after it was recorded, there is no doubtthat bluishor greenish-blueindigobirds may mimic firefinchesother than the usualsong model of the blue ("camerunensis")population, which in northernNigeria is the Black-facedFirefinch, L. larvata. Both the alarm call of L. rara and also the varied notesand the rapid, high-pitchedharsh notescharacteristic of L. rara songwere clearlyevident in the mimeticsongs of thesetwo indigobirdsat Zaria (Audiospectrograph19 e). Bluish "camerunensis" in some other areas of Africa as well as in northern Nigeriaappear to mimicL. larvata,insofar as the transliterationsof the songs of thesebirds can be comparedwith tape-recordedsongs at all. Wolters (1960: 24) kept a captive,bluish, pale-footedmale indigobirdfrom an unknown geographicsource and listenedto its song,which he describedas a slow trill "zjiie, zjiie, zjiie." The sequence"zjiie, zjiie, zjiie" is perhapsthe sameas the whistledphrases "beri-beri" and "hew-hew-hew"of the songof L. larvata. Chapin'scomment (1954: 572) on "camerunensis"that the "song consistsof fine twittering'chwees'" likewisesuggests mimicry of L. larvata in the Uelle regionof the Congo,because the same songsmight be trans- literated in terms such as "chwees," "beds," or "hews." GreenishPale-winged Indigobirds of the form "nigeriae"at Panshanu mimickedL. rubricata. Two greenishmales were recorded and collected,one additional male was recorded, and one more male was heard to mimic this samefirefinch; all of thesebirds were at a call-siteby a turn in the road on the east slopeof the passone km abovethe villageof Panshanu.Mimetic callsrecorded from "nigeriae"here includedthe rapidly descendingwhistle, the slurred whistle,the descendingtrill, and the alarm call (Audiospectro- graph 14). A differentset of songmodels is apparentlyused by the Pale-wingedIndi- gobirdsin northernCameroon, as Nicolai (1968) heard severalbluish birds "camerunensis"mimic L. rara, and the only green bird ("nigeriae") noted was a mimic of L. larvata. The different color forms of Pale-wingedIndigo- birdsappear then to mimicdifferent foster species in northernCameroon than in northernNigeria, indicatinga geographicdifference in the behaviorof theseindigobirds. A male "nigeriae"of unknowngeographic origin heard in captivityin Nicolai'slaboratory also mimicked L. larvata (Nicolai, 1968). Althougheach form of V. wilsoni(greenish "nigeriae," bluish "camerunen- 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 101 sis,"and purplish"wilsoni") usuallymimicked a singlespecies of firefinchin northern Nigeria, at least one firefinch (L. rara) was mimicked there by males of each of the three forms, and one form of indigobird ("nigeriae") includedsome birds mimicking L. rara (at Zaria) or L. larvata (in Cameroon) rather than the songmodel of the birds at Panshanu,L. rubricam. The distribution of the mimeticsong behavior in thesethree forms suggeststhat in somerestricted areas they may behaveas distinctand separatespecies, while in other areas there is evidence that the boundaries between the forms are blurred. In any case,the varietyof songobserved in the smallsample available suggeststhat the forms of the V. wilsonicomplex show partial, incomplete correspondencebetween mimetic song and plumagecolor.

DISCUSSION Resemblanceof indigobird and fire/inch vocalizations.--The recordings and audiospectrographsof the indigobirdsshow a closeresemblance in their mimetic vocalizationsto the calls and songsof the firefinch hosts. Both in the field and in captivityit is sometimesimpossible to know whethera firefinch soundis givenby a firefinchor by an indigobirdunless the singingbird itself is seen. On occasion,even calls in the grasswhich soundedlike firefincheswere given by male indigobirdsfeeding with femalesbelow the call- site. In all featuresof the calls seen on audiospectrographsthe mimetic vocalizationsof the indigobirdsare indistinguishablefrom thoseof the firefinch song models. The sequencesof mimeticcalls also showsome similarity to sequencesof vocalizationsgiven by the firefinches,for examplein the sequenceof contact call elementsof L. senegala,in the long sequenceof trilled "ti" and "tu" notesof L. rhodopareia,and in the sequenceof notesin the descendingtrills of L. rubricam. The sameor similarsequences of syllablesare heardin the mimetic songsof the indigobirdsV. chalybeata,V. purpurascens,and V. /unerea, and in each form the sequencesthemselves are often repeatedin a long seriesboth by model and by mimic. In other respectsthe sequentialpatterning of mimetic indigobirdphrases is unlike the sequencegiven by an individualfirefinch. The sequenceof mimetic phrasessuch as alarm call-song-beggingcall is one which no firefinchitself is likely to give. Someseries of mimeticphrases resemble the combinedvocal- izations of two or more firefinches. In Audiospectrograph5, a mimetic indigobirdmay be imitating a group of beggingyoung rather than a single young,since the beggingnotes are irregularlyspaced and individualyoung firefinchesusually give regularlyspaced calls (Audiospectrograph4). Mi- meticphrases are often givenin jumbledsequence with no obviousregularity; thesesequences are like the nonmimeticvocalizations of indigobirdsin which the variantsof songare not givenin orderly sequence. 102 ORNITHOLOGICAL MONOGRAPHS NO. 11

TABLE 8

VOCAL MIMmRY OF INDmOBmDS • SOUTHERN AFmCA

Number Number o/ o/ mimics mimics Number Locality Indigobird Mimic song heard recorded collected

SOUTH AFRICA Downs /unerea rubricata 1 - - Hluhluwe /unerea rubricata 1 - - Kondowe purpurascens rhodopareia 3 1 3 amauropteryx senegala 3 - - Louw's Creek amauropteryx senegala 2 - 2 Marble Hall purpurascens rhodopareia 1 - - amauropteryx senegala 18 13 10 Merensky purpurascens rhodopareia 23 11 11 amauropteryx senegala 29 16 10 amauropteryx rhodopareia 1 1 - Ndumu amauropteryx senegala 1 - - Tzaneen /unerea rubricata 9 5 6 BOTSWANA Maun chalybeata senegala ! 6 6 14 Boro chalybeata senegala 2 1 1 Shorobe chalybeata senegala 2 - 2 RHODESIA Chipinga: purpurascens rhodopareia 1 - 1 14 mi. N.W. Lusitu /unerea rubricata 7 - 7 purpurascens rhodopareia 2 - 2 Sabi Valley amauropteryx senegala 12 7 11 purpurascens rhodopareia 10 7 7 Melsetter Road purpurascens rhodopareia 2 - 1 Penhalonga purpurascens rhodopareia !6 10 12 codringtoni rubricata 6 5 6 amauropteryx senegala 3 2 3 Salisbury: purpurascens rhodopareia 1 - - Atlantica

Mimetic songsin my experienceare not given in behavioral contexts appropriatefor the specificmimicry; firefinch alarm callsare not givenwhen the indigobirdsare alarmednor are nestcalls given at the nestof the host. Specificityo[ hostmimicry.--The mimic songof the indigobirdsis remark- ably constantamong males of similar appearance,and different taxa of indigobirdsliving in the samearea generally mimic the songsof differentkinds of firefinches. I heard the songsof 320 indigobirdsin the field in 57 different localities. Of thesesinging males the mimetic songwas heard well in 302 birds. No indigobirdmimicked any bird other than a firefinch, and no individualbird mimickedmore than one speciesof firefinch. All of the singing indigobirdswhich were observedfor more than a few minuteswere heard to mimic a firefinch. In the few caseswhere vocal mimicry was not heard, observationswere cut shortby 'disturbancesat the call-sites,and at all of 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 103

TABLE 9 VOCAL MIMICRY OF INDIGOBIRDS IN CENTRAL AND EAST AFRICA

Number Number o! o! mimics mimics Number Locality Indigobird Mimic song heard recorded collected MOZAMBIQUE Tete: 45 mi. N.E. purpurascens rhodopareia 3 - - amauropteryx senegala 1 - - Tete: 55 mi. N.E. purpurascens rhodopareia 2 - - MALAWI Mwanza amauropteryx senegala 1 - - Chileka amauropteryx senegala 2 Zomba codringtoni rubricata 1 • • purpurascens rhodopareia 3 - 3 Monkey Bay amauropteryx senegala 11 7 8 purpurascens rhodopareia 10 6 6 Lilongwe: city amauropteryx senegala 1 - - Lilongwe: airport (Mbabzi) nigerrima rubricata 4 2 4 Lilongwe: Likuni road nigerrima rubricata 2 2 2 Salima:20 mi. W. purpurascens rhodopareia 1 - 1 Salima: 5 mi. E. amauropteryx senegala 1 - 1 Grand Beach amauropteryx senegala 3 - - KENYA Malindi amauropteryx senegala 1 1 1 Nairobi centralis senegala 2 1 - Olorgesailie centralis senegala 3 3 3 Kisumu: 10-35 mi. E. centralis senegala 8 4 8 Kakamega: 22 mi. N. centralis senegala 1 I 1 Sigor centralis senegala 1 1 1 purpurascens rhodopareia 5 2 6 thesesites mimicry would likely have beenheard with additionalobservation time. The firefinchesmimicked by the singingindigobirds are listed in Tables 8, 9, and 10. Vocal mimicry was tape-recordedfor 138 singingmales; most were subsequentlycollected as were otherswhich were heard but not taped. The specimensin the hand in nearly all casesconfirmed the sight identifications in the field, althoughthe pale-wingedWest African forms intergraded morphologicallyand were not alwayscertainly recognizable even in favorable light until they were collected,and the forms nigerrimaand purpurascensin south-central Africa defied identification until color standards were available. Severalforms of V. chalybeatamimicked the songsof Lagonostictasenegala; these indigobirdsincluded V. chalybeataamauropteryx in southernAfrica, V. c. centralisin Kenya, and V. c. neumanniin Nigeria. Several forms oi indigobirdsmimicked L. rubricata: bluish-purpleV. [unerea [unerea (in- 104 ORNITHOLOGICAL MONOGRAPHS NO. 11

TABLE 10

VOCAL MIMICRY OF INDIGOBIRDS IN NIGERIA

Number Number of of mimics mimics Number Locality Indigobird Mimic song heard recorded collected Sokoto neumanni senegala 2 - 1 Gusau: 4 mi. W. neumanni senegaIa 1 - - Gusau neumanni senegaIa 1 1 1 Zaria: A.B.U. neumanni senegaIa 22 6 5 "camerunensis" Iarvata 9 6 8 "camerunensis" rara 1 1 1 "nigeriae" rara 1 1 1 "wilson?' rara 5 3 5 Zaria: city neumanni senegaIa 1 - - Dumbi Woods neumanni senegaIa 1 - - Kaduna: 20 mi. S.E. " camerunensis" Iarvata 1 - - Narode neumanni senegaIa 1 - - los: 26 mi. E. "camerunensis" Iarvata 1 - - PanshanuPass "camerunensis" Iarvata 1 - 1 "nigeriae" rubricata 4 3 3 Bauchi:25 mi. W. "camerunensis" Iarvata 3 2 3 RimanZayam neumanni senegala 1 - - Gombejunction: neumanni senegala 1 - - 13 mi. W. Gombejunction: neumanni senegaIa 1 - - 1 mi.S. Yola: 97 & 98 "camerunensis" Iarvata 2 - 1 mi. Kiri neumanni senegala 2 1 1 Numan: 6 mi. N.W. neumanni senegaIa 1 - Numan neumanni senegaIa 3 1 Ganye neumanni senegaIa 1 - - cluding only red-footed birds) in Natal and Transvaal, greenishV. f. codringtoni in Rhodesiaand Malawi, purplish-blueV. f. nigerrimain Malawi, and greenish"nigeriae" (here considereda form of V. wilsoni) in Nigeria. Justone form of indigobirdusually mimicked the songsof each of the other songmodel firefinches.V. purpurascensmimicked L. rhodopareia(both L. rh. rhodopareiaand L. rh. jamesoni), the bluish "camerunensis"form of V. wilsonimimicked L. larvata in Nigeria, and the purplishform of V. wilsoni, "wilsoni," mimicked L. rara in Nigeria. A few individual indigobirdsmimicked firefincheswhich normally were the song models of other kinds of indigobirds.There were only 4 of thesein the total of 302 mimickingmales heard in the field in this study, and the generalpattern of firefinch specificity in the mimetic songsof the indigobirdsis a consistentone. All of the speciesof firefinchesare mimicked(and presumablyparasitized) by the indigobirdswith the exceptionof L. ruiopicta.Its behaviormay make successfulparasitism unlikely. L. [rufopicta]nitidula may use the old nests of other birds (Mackworth-Praedand Grant, 1963: 649), whereasall other 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 105 firefinchspecies regularly build their own nests(Immelmann et al., 1965; Morel, 1969). A captivelone male L. rufopictarufopicta that I caughtin Nigeriaclosely accompanied a pair of L. senegalaas thelatter firefinches were buildinga nestin our house.The maleru]opicta spent many minutes at the nestentrance and repeatedlyhopped into the nest. No bachelorfirefinches of other species(L. senegala,L. rhodopareia)in our househave shownan interest in my nestbuildingL. senegalafirefinches. Morel's observations (1969: 82) of the poor timingof ovulationof indigobirdslaying in the re- usednests of firefinchessuggest a reasonfor the notableabsence of parasitism or mimicry of L. ru[opicta by the indigobirds-•in non-nestbuilding firefinchesthe behavioralstimulus of nestbuildingby the host male is absent, and the femaleindigobird does not undergoovarian development at the most appropriatetime before the firefinches lay. I alsonoted that my captivemale ru[opictawas more aggressivethan any other captivefirefinches, sometimes chasingother firefinches from the food, and possiblyL. ru]opictamay behave aggressivelytowards any viduinefinches near its nests.Perhaps another rea- sonwhy ru[opictais not parasitizedis that its songsare very differentfrom the songsof otherfirefinches. The indigobirdsmay nevermistake it for their own hostbecause its songsare so differentin form, or they may be actively excludedfrom consortingwith singingL. ru[opicta,as songin this speciesis associatedwith aggressivemotivation (Wolters, in Harrison, 1962b: 378; Nicolai, in Giittinger, 1970: 1017). In contrast,the mistakenidentification of other firefinchesmay lead occasionallyto the successfulexploitation by the indigobirdsof a new foster species. ./Ire the mimetic vocalizationslearned?-•Experimental studies of song developmentin the young indigobirdshave not been completed,but the learningof mimeticphrases is suggestedby severalfield observationsand by comparisonwith other birds. First, the specificvocalizations that are mimicked are those that a young indigobird would hear during its period of parentalcare. Secondly,imprinting on the songof other speciesis known for severalother kindsof finches,and the closephylogenetic relationship between theseand the viduinessuggests the possibilityof song-learningability in the indigobirds.Finally, severalindigobirds were heard and recordedin the field mimickingone speciesof host, whereasothers of the same morphological appearancemimicked other firefinch species,and morphologicallyintermediate birds had no intermediatemimetic songs. Analysisof indigobirdmimicry showsthat most but not all of the known firefinch vocalizations are mimicked. The mimicked calls include the alarm call, contactnotes, song, juvenile location call, and beggingcall of the young (especiallybegging calls of fledgedyoung firefinches). All of theseare given by firefinchesduring the period of parental care and are thus part of the immediateacoustic environment of youngindigobirds. Young indigobirdsare 106 ORNITHOLOGICAL MONOGRAPHS NO. 11 rearedtogether with the youngfirefinches and do not eject them or outcom- pete them for food; in more than 80 percentof all parasitizedL. senegala nestswhich produced young indigobirds the firefinchesfledged together with the young V. chalybeata(Morel, 1969; fig. 14). The beggingcalls of the firefinchyoung would be heard by most nestlingand fledglingindigobirds. Whether young indigobirdsthemselves give this beggingcall or a different one is not known. Perhapsthe newly hatchedyoung have an innate begging call distinctfrom any calls learnedlater. Songwas given throughoutmuch of the nestingcycle in L. senegalaobserved in captivity;one male firefinchin our housesang from the time its youngwere only three days old. Contact call notesare givenby both adultsnear the nestand are usedin the family group after fledglng. In one pair in our housethe male beganto give the nestcall and a strawdisplay and alsocopulated when his youngwere only two weeksout of the nestand werestill being fed and threeother pairs renested andthe female laid whenthe youngwere 13-14 daysout of the nestand being fed; so it is possiblethat a youngindigobird might hear a nestcall duringits time with its fosterparents. Calls of firefinchesnot heard from adult indigobirds were either vocal or mechanicalsounds produced by firefinch adults during brief periodsof display (e.g., wing whir, "stip" note) or were calls likely to be heard only once, the distresscall (if given when a snakeenters a firefinchnest the youngindigobird would have little opportunityto live long enoughto repeat this call of its former nestmates). The timingof the sensitiveperiod in whichsong is learnedin indigobirdsis suggestedby the rangeof callsthat they mimic. V. chalybeataadults repeat in their mimetic song almostthe entire sequenceof ontogenicdifferentiation of the nestlingbegging call of 12 days of age throughthe fledglingbegging call and the juvenilecontact call or roostingcall givenby independentyoung firefinches as well as the definitive adult contact calls. The inclusion of all of theseintermediate stages of the beggingcall-juvenile location call-contact call developmentalsequence suggests a semifivephase of a few weeks,from nestlinglife to independence,in the song-learningprocess of indigobirds.The occurrenceof developmentallyand structurallyintergrading syllables, rather than stereotypyof a few syllablesonly, alsois bestexplained by regardingthe mimeticvocalizations as learnedover a long period of time. The temporal aspectof the apparentsensitive period indicatesa durationof severalweeks, althoughperhaps not aslong as in the Chaffinch(Fringilla coelebs), Bullfinch (Pyrrhulapyrrhula), Cardinal (Cardinalis cardinalis), Song Sparrow (Melo- spiza melodia), and White-crownedSparrow (Zonotrichia leucophrys),all of which may continueto add to their repertoire songsheard during the severalmonths after fledgingand well into their independencefrom the adults (Thorpe, 1958; Nicolai, 1959; Dittus and Lemon, 1969; Mulligan, 1966; Mailer and Tamura, 1964). The fact that individualindigobirds mimic only 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 107 a singlespecies of firefinchand no otherbirds suggests some restriction of learningin timeand in responseonly to thefamily social group in whichthey arereared. The socialbond between the firefinches and the youngindigobirds is a strongone, and not onlydo the youngindigobirds and firefinchescall togetherand flock together,the firefinchesmay snuggleagainst the young indigobirdand preen it (Morel, 1969: 165). The abilityof finchesthat are closelyrelated to the viduinesto learnthe songsof otherspecies also supports the notionof learningof hostvocaliza- tionsby the indigobirds.A learningof songsof otherspecies has been noted for severalestrildids. A captiveyoung male StrawberryFinch (Amandava amandava)whose father seldomsang learned the songof a Blue Waxbill (Uraeginthusangolensis), a songquite unlike its own speciessong (Goodwin, 1960: 196-197). Young Zebra Finches (Poephila guttata = Taeniopygia castanotis)raised in nestsof domesticBengalese or SocietyFinches (Lonchura striata) learn the songof the BengaleseFinches and give this mimetic song as adults;the younglearn the songin a shortsensitive period of a few weeks and this learningis irreversible(Immelmann, 1965, 1967, 1968b, 1969a). Similarly the African Silverbill (Lonchura malabarica= Euodice cantans) reared by BengaleseFinches learn and sing the Bengalesesongs, and Ben- galesereared by Zebra Fincheslearn the Zebra Finch songfrom their foster parents (Immelmann, 1969a). In the sparrows(Passerinae) and weavers (Ploceinae)development of songmay be influencedby otherspecies as noted for the House Sparrow,Passer domesticus (Stoner, 1942:441 ), the Golden Sparrow,P. luteus(Nicolai, 1964: 192), and the Red-headedQuelea, Quelea erythrops (Kunze, 1961: 228). Since all three of these finch groups (Estrildidae,Ploceinae, Passerinae) include species which can learn the song of other speciesas youngbirds, it is likely that indigobirdsalso can learn the song of their hosts. Nicolai has raised Straw-tailedWhydahs (Vidua fischeri) in Africa with BengaleseFinches as foster parents. When I visited Nicolai's aviaries at Seewiesenin August,1970, he pointedout to me a singingStraw-tailed Why- dah. I heard it mimicking not its usual host (Uraeginthus ianthinogaster) that I had heard and tape recordedin Kenya but rather the BengaleseFinch, indicatingthat someof the mimickingviduine finches learn their mimeticvo- calizationsfrom their fosterspecies, and very likely the doselyrelated indigobirds do also. Finally, a few exceptionalindigobirds recorded in the field mimickedfirefinch speciesother than did their neighborsof similar appearance. One Transvaal V. chalybeata mimicked L. rhodopareia rather than L. senegala, and severalNigerian bluish and greenishV. wilsonimimicked L. rara rather than L. larvata or L. rubricata, the usual song modelsof thesecolor forms. Theseindigobirds probably learned these songs by beingreared with the fire- 108 ORNITHOLOGICAL MONOGRAPHS NO. 11 finch speciesthat sangthem. Since.some indigobirds that look alike mimic different speciesof firefinchesit appearsmore likely that different mimetic vocalizationsare learnedrather than controlleddirectly by differentgenomes in the morphologicallydistinct indigobirds. Furthermore, in west Africa whereconsiderable intergrading between the forms"wilsoni," "camerunensis," and "nigeriae"of V. wilsoni has taken place, the individual birds (even morphologicallyintermediate birds) mimic only a singlespecies of firefinchand do not have intermediatevocalizations. One bird at Zaria (RBP 4959) which was bluish-greenin color mimicked the songsof L. rara just as the local purpleindigobirds all did there. This bird was morphologicallyintermediate betweenthe forms "nigeriae"and "camerunensis,"yet it clearlymimicked the samehost as the local "wilsoni"form of V. wilsoni. Theseexceptional birds providethe strongestfield evidenceavailable that mimetic vocalizationsare learned in the indigobirds. The adaptive significanceof host specificityand of learning the songs of the foster firefinches.--The behaviorof indigobirdfemales in finding a host and in respondingand laying in the host nest appearsto be highly host spe- dtic. The behavioralreason that a femaleindigobird parasitizes only a single speciesof firefinchis likely to be that only one firefinchprovides the proper vocalstimuli matching the soundsof her fosterparents. The femalemay form a specifichost searchimage throughimprinting to her foster parentswhen she is young,thereby directing her later breedingefforts towardsthe male indigobirdand towardsthe firefinch with the same sounds.Following this, she mateswith the male indigobird,and his mimetic songsand the songsof the host firefinchesmay stimulateher sufficientlyto developlarge eggsand to ovulate. Host specificityis a form of highly conservativebehavior, as it involvesa traditionof responseto the samefoster speciesin generationafter generation.This conservatismensures that the host is appropriate---afire- finchresembling her fosterparents is indeedan appropriatehost for a female indigobirdbecause her own generationwas successfully raised by one like it. The sameset of geneticallydetermined adaptations for the successfulpara- sitismof the firefinchesmakes the indigobirdsunlikely to be successfulin the parasitismof other kinds of hosts. Indigobirdsare specializedto parasitize their estrildidhosts on two levels;first, the indigobirdsall resemblethe young of the genusLagonosticta (and related speciesin Estrilda) in the black mouthmarkings of the youngand in juvenalplumage. Second, the indigobirds eachimitate a particular,single species in their vocalmimicry and parasitize only this firefinchspecies, a finer level of host specialization.Probably also the mouth colorsof the youngare species-specificcoadaptations. Selective parasitismof the firefinch hostsin general,which all of the indigobirdsre- semblein the blackmouth spots, might result in a largernumber of offspring per femaleviduine than wouldindiscriminate parasitism of a wide variety of 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 109 otherkinds of estrildidspecies, most of whichhave dissimilar mouth markings andcolors, as described in Chapin(1954) andin Immelmannet al. (1965). Perhapsthe successof indigobirdparasitism of other firefinchesand of the Estrildaspecies with a patternof mouthspots might be considerable,but the fosterfirefinches may discriminateagainst nestlings with othermouth colors, and the responseof femaleindigobirds to a narrowset of vocalsignals of a singlespecies of firefinchhost ensures that her eggsare not laidin inappropriate nests. In the face of a diversityof possiblehosts, most with dissimilarmouth markingsand colorsof the youngor of behaviorin parentalcare, a female indigobirdrestricting her laying to a singlespecies of host may leave more offspringthan a lessdiscriminating female laying more eggsin the nestsof a variety of species. The adaptivesignificance of learningthe mimetic songsof the firefinches (rather than being programmedgenetically to mimic the firefinch songsin exactdetail without having heard the fosterfirefinches first) is probablyre- latedto the variabilityin time andspace of the hostsas part of the environment of an indigobirdpopulation. First, the host speciesof an indigobird populationmay becomelocally extinct with changesin climateor in the local habitat. If someyoung indigobirds are raisedin the nestsof an alternative host speciesthat is betteradapted to the new environmentthan the old host was,these indigobirds will be able to reproducesuccessfully themselves only if they are attunedto the appropriatehost. Learningallows the indigobirds to changetheir signalsmore easily than wouldselection of geneticallydeter- minedvocalizations. The abilityto adaptbehaviorally to changedconditions and a new hostmay allow the indigobirdsto persistwhere their formerhosts have disappeared.As discussedby Cooke (1964), Bakker (1964), and Morean (1966), much of Africa suffered drastic shifts in local climate and vegetationduring the late Pleistocene,and underthese rapidly changing con- ditionsan adaptablebehavioral response would likely have beenof greater adaptivevalue to theindigobirds than a relativelyinflexible, genetically canal- ized program of songdevelopment. More importantly,by learningthe host songs,successive generations of indigobirdsmay keep up with any local changesin firefinch dialectswithout havingto adusttheir genomes.The localhessof the geographicvariations in the songsand calls of the firefinchesis not well understood,but as discussed in the sectionon indigobirdsong dialects the vocalizationsof firefinchesmay differover only a few miles.The minorvariants in whichhost dialects may differfrom one anotherare the sortof detailsthat mightchange over a few generationsdue to learningmistakes in copyingthe songsof the earlier generation.If thehost changes the details of itssongs in timeand space mea- suredon so short a scale,then the indigobirdswould better be able to track the changesby learningand imitatingthe currentlocal host vocalizations.If 110 ORNITHOLOGICAL MONOGRAPHS NO. 11 the indigobirdswere completelydependent upon the processesof mutation and geneticrecombination followed by selectionfor genotypesthat coded mostclosely the songsof their hosts,they would much less be likely to keep pace with any rapid local changesin the host songs. Yet it would likely be of adaptivesignificance for each generationof indigobirdto copy any con- temporarychanges in host songbehavior, because a male with mimetic songs most like the current, local host vocalizations would be the one with the closestapproximation of songto the foster father of a female indigobirdof his own generationand local mimeticdialect and thus the most likely to at- tract the mostfemales and to leave a large numberof offspring. Males with the ability to adapt to changinghost vocalizationsas well as to retain the constantsong features of the more conservativehosts would likely have more offspringthan males restrictedby geneticcanalization to singingthe songs typicalof previousgenerations of firefinchesbut now out of use. It seemsto me that a male that learned its mimetic songwould be more likely to mimic songslike thoseof the fosterparents of any local female indigobirdsattuned to the local songdialects. Similarly,the ability of a female indigobirdto respondto the hostvocalizations in currentuse, againthrough imprinting in her early life, may make her more readily or more frequentlystimulated to ovulate.Hence she would leave more offspring when she hears firefinch songs like thoseof her fosterfather than would a femalewhose song responsiveness had been geneticallyselected for many generationsearlier. The latter would be maximallyresponsive to a songno longer resemblingvery closelythe current songsof the local hostpopulation. In both male and femaleindigobirds, natural selectionfor the specificimprintability upon the foster species'vo- calizationshas probablybeen important,and learningthus may be regarded as a behavioraladaptation allowing indigobirds to track any changingsongs of the firefinches,as well as to conserveappropriate traditions along with the unchangingfoster populations. Becauseeach indigobird may mimic the vocalizationsof severalindividual firefinches (V. purpurascensmimics the calls of both male and female L. rhodopareia,and V. chalybeatamimics both adult and youngL. senegala), the mimicryis not restrictedto imitationof a singleadult fosterparent but rather of the entire foster family group with which an indigobird is reared. Becausean individualindigobird has severalfirefinch signals,it may have a large repertoirewhich facilitatesspecies recognition of its foster firefinch or of its own mate. The significanceof the variation in the details of mimicked song(in the narrow sense)of L. senegalaby a male V. chalybeatais not clear from a developmentalpoint of view. Accordingto Morel (1969: 105) individualmale firefinches have individually characteristic songs, and youngindi- gobirdsmay respondselectively to thesesounds of their own foster family. Individualmale firefinchesin my aviarieshave stereotypedsongs which they 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 111 giverepeatedly when isolated from their mates,but at othertimes each male may vary its vocalizationsby alteringthe numberof contactnotes in the "songs,"the durationof eachnote, and the changein pitch in the notes. As differentmale firefinchesin the field often do have distinct songsnot shared by their neighbors,then perhapsthe indigobirdsmay learn firefinch songs from the firefinchesother than their fosterparents, because each indigobird male has severalmimetic songs (Audiospectrograph 3), not just one. The indigobirdsmay be sensitivespecifically to the songsof onlythe birdswhich sharemany of the featuresof the songsof their foster family and then may generalizetheir songsensitivity to other conspecificfirefinches of the same speciesas their fosterers.They alsomay learn somefirefinch sounds from hearingthe mimetic songsof the male indigobirdsmimicking their foster specieswhen the youngindigobirds visit the call-sitesof the singingmales. The adaptivesignificance of learningand singingthe songsof severalindividual firefinches(not just the foster father companion)probably is to be found in the shortlife expectancyof a firefinchand of its songtype. Indi- vidual L. senegalausually live for less than a year (Morel, 1969: 24), and from each successfullyparasitized nest only about one youngindigobird is reared (Morel, 1969: table 28), so indigobird song microdifferentiationon the level of individualfirefinch family groupswould quicklyend the repro- ductivefitness of overspecializedindigobirds. If an indigobirdwere attuned only to an individualfoster firefinch rather than to a population,it would likely leavefar fewer offspringthan would an individualthat had generalized its behavior to a larger number of suitableconspecific firefinch fosterers,as its foster parentsthemselves would probablybe dead the year after it was reared. If a femalewere sexuallyresponsive only to the songsmimicking her own individual foster father, not many indigobirdswould find mates. The variety of mimetic songsgiven by an individualindigobird suggests selection for a degreeof generalizationof learning song, and this in part counterbal- ancesthe selectionfor specificityin mimetic songbehavior resulting from the unsuitabilityof most other potential finches as foster species.

NONMIMETIC VOCALIZATIONS AND SONG DIALECTS Indigobirdsong includes chatters and complexnonmimetic songs as well as vocal mimicry. The harsh,buzzy nonmimeticsongs are generallysimilar in the differentspecies of indigobirdsand contrastsharply in soundwith the species-specificmimetic whistles; it was surprisingto me to hear a bird sing suchdissimilar songs. A simplechatter in V. chalybeatahas been figured (Nicolai, 1964:138), but the complexnonmimetic songs haye not previously been described.From tape recordingsmade in Africa in 1967 and 1968 the nonmimeticchatters and songswere audiospectrographedand about 2,600 nonmimetic songswere analyzed. 112 ORNITHOLOGICAL MONOGRAPHS NO. 11

I'.O Sec Audiospectrograph 22. Representative nonmimetic songs of an individual Vidua purpurascensat Merensky Reserve,Transvaal. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 113

kHz'

Audiospectrograph23. A nonmimetic song type shared by V. chalybeata males recorded at Marble Hall, Transvaal, during January 1967. Songs a and b were given by one individual; c, d, and e were recorded from three other males. All birds include mimetic alarm notes at times .05, .15, .2, and 1.4 secondsof the songs. The apparent differences in darkness and harmonics are due to different sensitivity settingsof the SOhO- graph. 114 ORNITHOLOGICAL MONOGRAPHS NO. 11

I(Hz'

ii I;0 Sec Audiospectrograph 24. Interpopulation differences in the nonmimetic songs of Vidua chalybeata amauropteryx: a, Monkey Bay, Malawi; b, Penhalonga,Rhodesia; c, Sabi Valley, Rhodesia; d, .Merensky.Reserve, Transvaal. Note the mimetic alarm notes in a at .1-.2 seconds. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 115

kHz'

!.I I'.0 Sec AudiospectrogrAph25. InterDoDu]ationdifferences in the nonmimeticsongs of Vidua chalybeata--acomparison of four subspecies:a, neumanni,Zaria, Nigeria; b, centralis,Sigor, Kenya; c, amauropteryx,Malindi, Kenya; d, okavangoensis,Maun, Botswana. 116 ORNITHOLOGICAL MONOGRAPHS NO. 11

kHz-

'.1 110Sec

Audiospectrograph 26. Nonmimetic songs of V. purpurascens (a, c, d, e) and V. /. nigerrima (b)--population differences: a, Sigor, Kenya; b, Lilongwe, Malawi; c, Monkey Bay, Malawi; d, Penhalonga,Rhodesia; e, Sabi Valley, Rhodesia. Note the mimetic notes in a ("tu-tu" series at .I-.8 seconds), in b ("ti" note at .6-.7 seconds), and in e ("titi" at .I-.3 seconds) by comparing these to Audiospectrographs 6 and 13. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 117

kHz'

'.1 I'.0 Sec Audiospectrograph2?, Nonmimetic songsof some indigobird mimics of Lagonosticta rubricata: a, b, "nigeriae" form of Vidua wilsoni, Panshanu, Nigeria; c, V. funerea codringtoni,Zomba, Malawi; d, V. f. codringtoni,Penhalonga, Rhodesia; e, V. 1. [unerea, Tzaneen, Transvaal. 118 ORNITHOLOGICAL MONOGRAPHS NO. 11

•.1 I•.0 Sec Audiospectrograph28. Nonmimetic son• types oœ"cam•r•sis" and forms'of V. wilsoniat Zaria, Nigeria: a, b, and c were recordedfrom three different "camerunensisy.and d ande wererecorded from a single"wilsoni." 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 119

kHz-

.I Ii0 Sec Audiospectrograph29. Nonmimetic song types of Pale-wingedIndigobirds at Zaria, Nigeria, which were blue or blue-green and mimicked Lagonosticta rara: a and c were taped from 4959, a blue-green"nigeriae," and b and d are from 4884, a blue bird of "camerunensis"plumage. Compare these songswith the "camerunensis"and "wilsoni" songsof Audiospectrograph28. 120 ORNITHOLOGICAL MONOGRAPHS NO. 11

•.I I•.0 Sec

Audiospectrograph30. Nonmimetic songs and wingbeat sounds of two speciesof indigobirdsin courtship display: a and b are from two individual Vidua chalybeata at Maun, Botswana, and c and d are from two successivehover displays of a single V. wilsoni at Zaria, Nigeria. The series of three low notes, given twice, at 1.2-1.8 secondsin c are calls of a different species. Note the variation in song and wingbeat betweenbirds and betweendisplays of a singleindigobird as well as variations in wing- beatsamong successive bounces in the samedisplay series. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 121

kHz

I I 0.1 1.0 Sec Audiospectrograph31. Variable introductorymimetic notes in nonmimeticdialect songsgiven by a single l,'idua chalybeataat Marble Hall. 122 ORNITHOLOGICAL MONOGRAPHS NO. 11

! O.I I.O Sec Audiospectrograph32. Above is song type K asdelivered by a maleVidua chalybeata at MarbleHall. An identicalsong type was given by 12 of the 13 malesrecorded, whilethe othermale gave the songshown below lacking a singlemimetic note. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 123

kHz

I I 0.1 1.0 Sec Audiospectrograph 33. Variation in the delivery of one mimetic contact note in three versions of a single song type given by one Vidua chalybeata at Marble Hall. 124 ORNITHOLOGICAL MONOGRAPHS NO. 11

kHz t!

I I 0.1 1.0 Sec Audiospectrograph34. Two similar, apparently homologous song types sung by two male V. chalybeataat Olorgesailie,Kenya. These similar songsdiffer in the introductory notes,in the pitch of the syllableat 0.6 seconds,and in the timing of the change in pitch of the long syllable at 0.7-0.9 seconds. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 125

kHz

0.1 1.0 Sec Audiospectrograph35. Two similar,apparently homologous song types both sharedby mostVidua purpurascens at Penhalonga. The lowersong may have been derivedfrom the upper one by a doublingof themimetic note at 0.6 seconds,an increasedmodulation of the descendingnote at 1.2 seconds,and an additionof a complexnote at the end,or similarlythe uppersong may be derivedfrom the lower. 126 ORNITHOLOGICAL MONOGRAPHS NO. 11

i I 0.1 1.0 Sec Audiospectrograph36. Repetitionof a songtype with a smalltime interval between songsin a Vidua purpurascensat Penhalonga.

I i 0.1 1.0 Sec Audiospectrograph37. Fusion of twounlike songs into a singlesong type. The phrasesfrom 0-0.6 and from 0.7-2.2 seconds were sting separately at other points in a singingbout by thisVidua purpurascens at Penhalonga. Occasionally they were sting as a single unit as shown here. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 127

c d

I I 0.1 1.0 Sec Audiospectrograph38. Possiblyhomologous songs of Vidua chalybeatain two nearbyareas in northernBotswana. Figures a andc wererecorded in Maun;b andd were recordedat Leomarin, six miles away. Each songtype was locally stereotyped, includingthe mimeticsong types c and d, which mimic the alarm call and juvenile location calls of Lagonosticta senegala. 128 ORNITHOLOGICAL MONOGRAPHS NO. 11

kHz

2_-

I I 0.1 1.0 Sec Audiospectrograph39. Complex songs of Vidua funerea nigerrirna recorded at Lilongwe,Malawi. The upper song(by 4541) and lower song (by 4534) have mimetic introductorynotes followed by nonmimeticnotes and may be homologous.The birds were about eight miles apart and sharedno identical song types; the song types shown are the most similar ones between the two birds.

kHz-

2 I•0' Sec Audiospectrograph 40. Chase call given by Vidua chalybeatain flight at Marble Hall, Transvaal. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 129

TABLE 11 SYLLABLE DURATION IN INDIGOBIRD CHATTER

Duration of syllablesin chatter (msec) Number of Species Locality birds Max. Min. Mean chalybeata Marble Hall 6 30 20 27.5 4.40 Merensky 9 50 20 28.9 5.46 Sabi Valley 7 55 20 31.4 11.2 Penhalonga 2 30 30 30.0 - Monkey Bay 7 30 20 27.1 3.64 Maun 6 30 20 25.0 5.76 Kenya (3 localities) 6 30 30 30.0 - Nigeria (3 locs.) 6 60 30 42.5 12.3 purpurascens Merensky 11 50 20 30.9 9.30 Sabi Valley 6 50 20 28.3 12.4 Penhalonga 6 40 20 31.7 7.20 Monkey Bay 6 60 20 33.3 15.6 Sigor 2 30 30 30.0 - funerea Tzaneen 5 60 20 31.0 - Penhalonga 6 40 25 33.3 5.40 Zomba 1 - - 50.0 - Lilongwe 4 55 25 40.0 - "camerunensis" Nigeria (2 locs.) 9 60 25 36.1 9.76 "wilsoni" Zaria 3 50 40 43.3 -

CHATTERS A "chatter" is a rapid sequenceof harsh notes. An indigobird chatter consistsof a sequenceof notes all on the same pitch and all of nearly the same duration. Each note coversa wide range of audible frequencies,as a result of rapid modulation, and this is responsiblefor the rough, harsh sound of the chatter elements(Marler, 1969). The chatteringindigobird repeatsthese syllables 8 to 16 times per second. Examplesof chattersare shown at the beginningsof nonmimeticsongs in Audiospectrographs22 a and 24 d. Chattersdiffer in number of notes,in the time and frequencypattern of individualnotes, and in the time betweensuccessive notes; within a single chatterseries all of thesefeatures were nearly constant. The individualnotes are composedof soundswith a wide frequencyenvelope and a seriesof harmonics;sound energy is concentratedat about 1.8 and 3.6 kHz in the chatternote. In notesof longerduration, the dominantfrequency sometimes decreasesthrough time. The zigzag appearanceof the upper and lower limits of the soundenvelope at eachharmonic in a note indicatesconsiderable frequencymodulation (FM). Amplitudemodulation (AM) is not quantita- tively determinableon the audiospectrographsbut is indicatedin the changes of thicknessand darknessof the frequencytrace through time, mainlyin the chatteredsyllables. Durations of the individualnotes range from 20 to 60 130 ORNITHOLOGICAL MONOGRAPHS NO. 11

TABLE 12 TIME INTERVAL BETWEEN SUCCESSIVE SYLLABLES IN INDIGOBIRD CHATTER

Mean periodicity o! calling in chatter (msec betweenonset of successivesyllables) Number of Species Locality individuals Max. Min. Mean t.95• chalybeata Marble Hall 6 87 45 75.7 16.7 Merensky 9 143 46 70.1 22.6 Sabi Valley 7 136 40 74.6 33.3 Penhalonga 2 98 80 89.0 MonkeyBay 7 83 37 67.3 1•.8 Maun 6 82 43 62.3 38.2 Kenya 6 80 60 76.5 8.8 Nigeria 6 165 80 115.8 40.8 purpurascens Merensky 11 150 45 78.7 21.2 Sabi Valley 6 115 49 70.8 25.1 Penhalonga 6 95 50 73.0 18.5 Monkey Bay 6 148 47 74.5 39.6 Sigor 2 72 70 71.0 - /unerea Tzaneen 5 176 46 80.0 - Penhalonga 6 97 56 77.0 18.7 Zomba 1 - - 85.0 - Lilongwe 4 140 48 89.8 - "camerunensis" Nigeria (2 locs.) 9 230 47 105.9 16.2 "wilsoni" Zaria 3 130 81 101.0 - msec. The number of the notes in each chatter series varies. Individual males sometimesgive 1 to 4 chatter noteswhen they return to the song post from a chaseof anothermale, and in their prolongedsinging bouts the samebirds give chatters ranging from a few harsh notes to more than 30 in a graded series of vocalizations. To comparethe characteristicsof chattersin different populationsand differentspecies of the indigobirdsI sampledone chatter from each bird that I had tape-recordedin the field. The chatterswere chosenat random by referenceto a table of randomnumbers. The chattersof all speciesof indigobirdsrecorded are quite similar. Audio- spectrographicanalysis showed no consistentspecies differences, although the Nigerian indigobirdshad somewhatslower chatters with longer syllables. As individual males show nearly the entire range of syllable duration in their various chatters,it is not surprisingthat no speciesdifferences were evidentin mean syllableduration (Table 11). Chatterscomposed of longer syllablesare usuallydelivered more slowly,with fewer syllablesper second, althoughwithin a single chatter seriesthe timing of the syllablesis very regular. All degreesof rates of chatteringare evident,and when a sample of chattersis comparedfor each indigobirdspecies (Table 12), no species differencesare apparent,even in sympatric,non-interbreeding forms. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 131

COMPLEX NONMIMETIC SONGS The othernonmimetic vocalizations given by singingmale indigobirds consist of seriesof dissimilarnotes given in rapid succession.Examination of hundredsof audiospectrographsshowed that soundswithin these series are usuallyseparated by at least 10 msec. Elementsof nonmimeticsongs that are separatedby at least10 msecfrom adjacenttracings are termedthe "notes" or "syllables."Some notes are simplewhile othersare quite complexand vary considerablyin pitch, duration,change of pitch, and modulation. Frequencymodulation is evidentin the chatteredand buzzy notesof the nonmimeticsongs. Rapid, regularfluctuations in pitch vary in a continuous manner and individualbirds each have a graded seriesof FM syllables rangingfrom 30 cps to 200 cps (the limit of temporalresolution of the audiospectrograph).Some syllablesshow little FM but more AM; some of theseare identicalto the chatter syllables. Syllableswith conspicuous FM (wherethe frequencyvaries more than 0.4 octaves)often vary in pitch of the cartierfrequency, as somenotes rise slowly,others drop rapidly,and othersdrop and thenrise or vary in othermore complex ways. Examplesof thesechanges are seenin Audiospectrograph22. Other classesof syllablesin whichFM is not evidentinclude slurred whistles. Some of theserise in pitch, somefall, andsome are complexwith severalchanges in pitch. Harmonicsare oftenevident in FM-dominatednotes but may also be presentin whistlednotes. Complexnotes occur in whichthe first half of the note is buzzyand shows much FM, whereasthe secondhalf of the note may show a gradualshift to a whistle. In additionto the characteristicnonmimetic notes, complex songs may includea few mimeticnotes identical in structureto the vocalizationsgiven in songbouts of pure mimicry. They seemto be incorporatedinto some complexsongs in the samemanner as are the nonmimeticsyllables. About a fifth of the complexsongs examined audiospectrographically had some mimeticmotifs. Thesemimetic notes were not readilyapparent in the field to the humanear, and I did not recognizethem until the songswere audiospectrographed.Examples of mimeticnotes in the complexsongs are evident in Audiospectrographs23 a-e, 24 a, 26 a, b, e, and others. Unlessotherwise indicatedI refer to both the complexsongs with mimeticnotes and those withoutas "nonmimeticsongs." Where the two kindsof nonmimeticsongs are differentiatedI refer to the songswith both kinds of notesas "mixed songs"to differentiatethem from the othernonmimetic songs. Songpattern is determinedby the sequenceof notesgiven as a unit. Usuallyno noteswere repeated within a songexcept for the introductory chattersand, when they occurred, the mimetic alarm notes and begging calls. The samesequence of syllablesin complexsong is repeatedtime after time 132 ORNITHOLOGICAL MONOGRAPHS NO. 11 in a stereotypedmanner by an individual male, and such a stereotyped patternis calleda "songtype." A songtype, by my definition,is a songin whichthe first threeor more distinctsyllables differ from thoseof all other songs,and songswhich differ in only one or two of the first three syllables are referredto as "variants"of thesesong typesrather than distinctsong types. Also referredto as variantsare songsthat are identicalin structure except for the presencein the middle or at the end of syllableslacking in other songsregarded as typical. A simplekind of songtype heard in many indigobirds,including both Vidua chalybeataand V. purpurascens,was a chatter followedby a complexbuzzy note (Audiospectrograph22 a). Another commonpattern in both of thesespecies is a rapid chatter,a pause,a series of longerchattered notes, and a final varied jumble of complexharsh notes and whistlednotes that changerapidly in pitch (Audiospectrograph22 b). The terminal flourishesof thesecomplex songs remind a North American bird-watcherof the song of Bell's Vireo (Vireo bellii) but are delivered more rapidly. Methods of analysis.---•l recordedsongs of every individual bird were analyzedaudiospectrographically, except where it was evident by ear that a bird was givingthe samesong type repeatedly,and then somerepetitions were countedbut not graphed. The songsof each bird were compared directly by arrangingthe audiospectrographson tables, holding each one besideevery other one, and matchingthe identical song types and their variants. This procedurewas repeatedfor the songsof all birds within a populationand for songsof birds in different populations.All variants of each songtype were tabulated. In addition, the songtypes for nine populations of three specieswere all measuredfor several features of their over-all structure,including total songduration, songduration less the introductory chatter, highestbasic frequency (excluding harmonics), lowest basic frequency,total number of syllablesper song type, and number of different kinds of syllablesin each song type. In caseswhere song type •variants involvedthe additionor omissionof syllables,I selectedas the "standard" song type the one which was sung most often by the local birds, and I regardedthe other versionsof this songtype as its variants. To compare the amountof repetitionor redundancyof syllableswithin a song type I divided the total number of syllablesby the number of different kinds of syllablesfor the "repetitionindex" (Thompson,1968: 270). Most song boutswere phrasedwith 140-800 msecbetween the songs. In caseswhere songswere sungin rapid successionor when syllableswere sungat long intervals, I have regarded the distinct song types to be the sum of the syllablesseparated in time by 140 msecor longer from other vocalizations. Variationin songsof an individuaL--Therepertoire of a male indigobird includesseveral distinct, complex nonmimetic song types. Each songtype 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 133 hasunique notes not foundin othersong types, although some notes occur in more than one songtype. The songtypes of a male are repeatedin preciselythe same manner, and two songsof an individual are shown (Audiospectrograph23 a, b) to illustratethis stereotypy. Tape recordingsof five minutesor longer of a songbout were made for 15 indigobirdsin the field, and these longer recordingswere examined aurallyand audiospectrographicallyto determine the numberand sequence of songtypes of an individual.The longestrecording made was of a V. I. codringtoni(RBP 4530) singingalong a roadsidenine miles south of the market at Zomba,Malawi. In eighthours of recordingattempts on 24 March 1967 I taped 18 minutesand 40 secondsof songin short sequencesoften brokenby silenceof the male,noises of passingpeople and cars,and heavy rainstormswhich continued much of the day. Nearly all of the 134 recorded nonmimeticsongs were audiospectrographed,and the otherswere identified by ear (with controlchecks) as identicalto a songjust given. Each song type was codedby letter. Two songswere merelylonger variants of other songtypes; C was the sameas R but had three additionalsyllables at the end, and S was the same as Q but had three more syllablesat the end. Variantsof the other songtypes were minor. A few were sometimesgiven withoutthe final syllable,and these are indicatedby a prime(') signfollow- ing the songtype codeletter. The sequenceof songtypes follows. Singing boutsinterrupted by flight are so indicatedwith a slant (/) followingthe recordedportions, chatters are shownas "ch" for singleseries or "chatters" for prolongedseries, while mimetic vocalizationsare indicated as "m": "ABC/DEEchFFFFFchFGGHIIIIIchJ JJ'JKchLKMMMM/ N'NNNchmmmchchchchchatters/mchOANNN/OchchFF/ chatters/chatters/IJ'JKKMM/mmmm/mmmm/PPPPE/ MMMMMMQA'AAQAQQ/BRCCCCDchEE'chFFFFFFFFF/N'MMMM/ mmmmmmmm/mmmm/mmmm/MAA'/ch/chJ'J'J'K/mmmm/ RABCAEEEEFFFFFG'G/chchEEE'EFFIIIIJ JJ'QSNNNP/." The vocal repertoireof this male included 17 distinct nonmimeficsong types. A few variantsoccurred, but no new distinctsong types appeared in the secondhalf of the recording,and it is thereforelikely that almostall of the songtypes of this bird were recorded. In additionto thesenonmimetic vocalizafions, the male had at least seven kinds of vocalizafions which mimicked the firefinch L. rubricata. No rigorousordering of songtypes was apparent.Some song types were repeatedin series(FF, MM, NN). The completesong type repertoirewas evidentlynot often exhaustedbefore the bird returnedto a songtype he hadalready sung, though the sequence of somesong types suggests a tendency to sing throughmost of the songtypes in a prolongedsession of .several 134 ORNITHOLOGICAL MONOGRAPHS NO. 11

ß ß

o (D

o o 5'0 ' NUMBER OF SONGS RECORDED Figure 17. Number of nonmimeticsong types of individualmale Vidua chalybeata as a functionof total numberof nonmimeticsongs recorded. minutesof singing.Mimetic phraseswere repeated,and differentkinds of mimeticnotes were often sung together in a seriesrather than being scatteredrandomly among the nonmimeticsongs. While this was a common patternin the singingbehavior of indigobirds,a few malesregularly did alternatemimetic phrases with nonmimetic songs. Some song bouts included 30 secondsor more of uninterruptedmimicry and alsoseveral continuous minutesof nonmimeticsong. Longeruninterrupted recordings of continuoussinging in otherkinds of indigobirdsshowed a patternof singingsimilar to that of the codringtoni. Representativeaudiospectrographs of vocalizations of a maleV. purpurascens (RBP 4412) at MerenskyReserve, Transvaal, in Audiospectrograph22 illustratethe great variation between some of thesong types of a singlemale. The songsequence in the recordingof 4 minutesand 18 secondswas the following(none of the songtypes were the sameas in the codringtoni;the letterscode an entirelydifferent set of songs): "ABBchmmCCCCCDCCmmCCCmmCCCCCCCmmCCCCmmCmmmmm- mmmmmmmmmmmmmEFEECC'chGchmHHFFEFGFBIJJ." 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 135

The sequenceof song types was irregular. Some songswere repeated (CCC), and some were alternated(E and F). The prolongedbout of mimicrylasted 52 seconds.As new songswere addedto the end of this recordingprobably the bird had additionalnonmimefic song types, but the songstructure, diversity, and repetitionof stereotypedsong types were typical of the longer recordings. Repertoiresize of an individualindigobird is indicatedin Figure 17, whichshows the numberof songtypes recorded in eachmale V. chalybeata taped in southernand easternAfrica. Where longerrecordings were made of a singingmale, more songswere availablefor analysis,and I found that more songtypes were given. As the plot of songsand songtypes showsa curve with little increasein number of song types given above 40 songs recorded,it appearsto be necessaryto recordat least40 songsof an indigobird to documentsuccessfully its songrepertoire. On the average,males for which40 or more songswere recordedhave somewherebetween 12 and 16 distinctnonmimetic song types, in additionto their mimeticvocalizations. Songsof different malesin a singlepopulation.--In the field, Karen and I thoughtwhile listening to malesin a singlearea that all birdssang the same nonmimeticsongs. One of the songtypes given by all of the V. chalybeata at Marble Hall was vaguelylike the interpretationwhich we gave it in the field--"dididdledeediddudididoo,didodwee." Audiospectrographanalysis of the songsconfirmed the impressionthat some,perhaps most, song types are sharedby most or all birds in a singlepopulation. The samesong type as sungby differentbirds is virtuallyidentical. These"dididdle . . ." songsof differentbirds were as similar as the successiverenderings of an individual bird (Audiospectrograph23). Only in minor detailsof variantswere the songs of different individuals at all different. One bird at Marble Hall omittedthe first introductorynote, and otherssometimes dropped the final syllables,as an individualbird may also do in a prolongedsequence of repeatinga songtype. Somebirds at Marble Hall had three versionsof one songtype, and a V. chalybeataat MerenskyReserve had five variantsof a songtype. More often,however, only one songvariant for eachsong type was found. To comparethe degreeof sharingof songtypes among individuals living in a single locality I have tabulated the number of times each male V. chalybeatarecorded at MarbleHall sangeach of thelocal song types (Table 13). This is the populationfor whichI recordedlocally the largestnumber of songsfrom a singlespecies. All 18 of theMarble Hall songtypes recorded were sharedby threeor morebirds, and at leastone songtype (K, the "dididdle"song type) was sungby all 13 birdsrecorded. Probably if a largernumber of songshad been recordedfrom each bird, most or all of 136 ORNITHOLOGICAL MONOGRAPHS NO. 11

+ + + +

•+ + +

+ • ++

+ 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 137 the songtypes would have been found to be sharedby all of the local indigobirds. Not only the songtypes but alsothe variantsof the songtypes were often sharedamong males in a population;for example,variant a of songtype H was recordedin 4 of the 10 birds known to sing H. Each of these4 also sangthe "standard"song type H as well as variant a (Table 13). A few variantswere not noted to be sharedlocally; variant )• of songtype H at MarbleHall wasrecorded in onlyone male. Other populationsof V. chalybeataalso had stereotypedsong types which were sharedlocally by mostor all birds recorded. In all nine localitieslisted in Tables 8, 9, and 10 for which two or more birds were recorded, the birds within a locality sharedsong types. The other indigobirdspecies likewise sharedsong types locally, althoughin generalthey did not share any song typeswith other sympatricspecies. V. purpurascensshared song types at all five localitieswith recordedsamples of songsfrom two or more birds. V. funerea shared song types at three of the four localitiessampled (not in the small sampleof songsof nigerrimaon the Likuni road in Malawi, but thesehad similarsong types). At the "nigeriae"site recorded near Panshanu, Nigeria, at the "wilsoni" site at Zaria, and at both "camerunensis"localities in Nigeria,each color form of V. wilsonilikewise shared among the local malesof similarappearance the samesong types. Differentmales in an area not only sharedeach song type, they also sang the song types in the same sequence;at least that is my impressionfrom listeningseveral times to the recordedsongs of eachbird and from examining the sequencein audiospectrographedsongs. Certain song types usually fol- lowedor alternatedwith other songtypes. Also, differentmales tended to repeat in seriesthe same song types. Table 13 suggeststhat the common song types of one bird are also the most frequently given song types of other birds. Individual male indigobirdswithin a population clearly do not emphasizeany individualitywith individualdifferences in their songs, inasmuchas all of the differencesbetween males in songtypes were also foundin the variantsof thesesong types given within the repertoireof an individual indigobird. Song variation amongpopulations: local dialects.--Both regional differencesin the detailsof nonmimeticsong types and local dialectaldifferences were found in the indigobirds.Each seriesof unique song types that are sharedby all birdslocally, and restrictedto a smallarea, comprises a song dialect. The two populationsof V. chalybeataamauropteryx recorded in Transvaalhad completelydifferent repertoiresof nonmimeticsongs. The two localitiesare separatedby about 120 miles,in which someseemingly suitablehabitat for chalybeataoccurs, such as aroundZebediela and along the OlifantsRiver at 24ø16'S,30ø00'E. I sawno indigobirdsthere however. 138 ORNITHOLOGICAL MONOGRAPHS NO. 11

Audiospectrographsof all 18 songtypes recorded from Marble Hall were comparedwith all 23 song types from Merensky Reserve,and there were no instancesof a common song type. Differencesin complex notes also were apparentwith restrictionof many of theseto a singlelocality. Local populationaldifferences in the complexnonmimetic song also were apparent in amauropteryxin Rhodesia;none of the songsof two males taped near Penhalongawere like sevenat the Sabi Valley ExperimentalStation. These two localitiesare separatedby 100 map miles and are both on the Sabi River drainagesystem; this form of indigobirdhas been collectedat severalintermediate localities along the river. At Monkey Bay, Malawi, amauropteryxsimilarly had distinctive,local versionsof the complexsongs, none of which were shared with other populationsstudied. Examples of the populationalvariation of song within this single subspeciesare shown in Audiospectrograph24. All populationsof other subspeciesof V. chalybeatathat were recorded likewisehad localunique song types. With the possibleexception of Nigerian V. c. neumanni,which had longersongs with shorter,simpler notes given in more rapid sequence,the differencesin nonmimetic song types among subspecieswere no greaterthan those amongdistinct populations of morphologicallyindistinguishable birds (Audiospectrograph 25 ). The other kinds of indigobirdsalso had highly localized song types. Samplesof recordedsong of V. purpurascensat MerenskyReserve, Sabi Valley, Penhalonga,Monkey Bay, and Sigor revealedsong types shared locally,but no songtypes were shared with otherpopulations studied (Audio- spectrograph26). V. funereawere recordedat Tzaneen,Penhalonga, Zomba, and Lilongwe; these birds comprisedthree distinct morphologicalforms recognizedas different subspecies.Examples of their songsare shownin Audiospectrograph27. Eachpopulation had distinctivesong types including birds in two different populationsof the morphologicalform codringtoni in Rhodesiaand Malawi. Similarly,the two forms of V. wilsoni recorded from more than one bird in more than one locality,"nigeriae" and "camerunensis,"had completelydifferent series of nonmimeticsong types at Panshanu and at Zaria in Nigeria. Althoughthe detailsof syllablestructure and sequenceof notes within the songtypes varied from placeto place,the generalfeatures of the nonmimeticsongs were similarin differentareas. Table 14 recordssome general songcharacteristics of the songtypes in nine differentpopulations, including three populationseach of V. chalybeata,V. purpurascens,and V. [unerea. On the average,song duration, frequency range, and songcomplexity were very similar amongdifferent populationsof the same species,even when different subspeciesare compared (the V. [unerea populationsfrom three different subspeciesareas). Variation among populationsof the same 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 139

Tt•BLa 14 SUMMARY OF CHARACTERISTICS OF NONMIMETIC SONG IN NINE POPULATIONS OF INDIGOBIRDS

Species Locality N Max. Min. Mean t.95•

I. SongDuration (sec.) cha•bea• Marble Hall 19 1.65 .60 1.25 .161 Merensky 23 1.50 .75 1.13 .155 Maun 18 1.70 .60 .90 .154 purpurascens Merensky 28 1.80 .80 1.08 .187 Sabi 13 1.65 .85 1.22 .177 Penhalonga 13 1.90 .75 1.28 .256 funerea Tzaneen 25 3.25 .60 1.48 .215 Penhalonga 16 1.95 .65 1.17 .228 Lilongwe 19 1.50 .60 1.02 .117 II. Song DurationMinus Introductory Chatter (sec.) cha•beam Marble HaH 18 1.65 .30 .92 .394 Merensky 23 1.30 .40 .83 .104 Maun 18 .85 .60 .73 .051 purpurascens Merensky 28 1.80 .80 1.08 .187 Sabi 13 1.65 .85 1.22 .177 Penh•onga 13 1.55 .40 .81 .218 funerea Tzaneen 25 2.30 .20 .93 .191 Penhalonga 16 1.95 .30 .81 .203 Lilongwe 19 1.05 .30 .65 .104 III. High Frequency (kHz) cha•bea• Marble Hall 19 7.8 5.2 6.74 .362 Merensky 23 7.5 5.0 6.55 .302 Maun 18 7.3 4.5 6.14 .358 purpurascens Merensky 28 8.5 5.8 5.96 .267 Sabi 13 7.0 3.8 5.60 .493 Penhalonga 13 7.5 6.0 6.83 .328 funerea Tzaneen 25 7.2 5.0 6.48 .276 Penhalonga 16 7.5 4.5 6.22 .352 Lilongwe 19 6.8 4.5 6.11 .303

IV. Low Frequency (kHz) cha•bea• Marble Hall 19 2.5 1.0 1.58 .201 Merensky 23 2.0 1.0 1.41 .106 Maun 18 2.0 1.0 1.44 .116 purpurascens Merensky 28 2.0 0.5 1.21 .206 Sabi 13 2.3 1.2 1.69 .191 Penhalonga 13 1.8 1.0 1.27 .163 [unerea Tzaneen 25 1.7 0.8 1.21 .318 Penhalonga 16 1.5 0.8 1.06 .627 Lilongwe 19 2.0 1.0 1.47 .133 Mean Repetition V. Number of Syllablesper Song Index chalybeam Marble Hall 18 11 3 6.32 1.21 1.04 Merensky 23 16 3 6.43 2.65 1.10 Maun 18 7 2 5.06 .650 1.05 purpurascens Merensky 28 9 3 4.82 .915 1.14 Sabi 12 16 4 7.08 1.98 1.20 Penh•onga 13 12 3 6.77 1.91 1.13 funerea Tzaneen 24 23 3 7.88 1.83 1.06 Penhalonga 16 14 2 6.69 1.73 1.07 Lilongwe 19 10 3 5.52 1.02 1.13 140 ORNITHOLOGICAL MONOGRAPHS NO. II speciesin the nomimetic songsclearly involves local songdifferences rather than broadregional differences that parallelthe morphologicaldifferentiation of the indigobirds. In a few areas indigobirdswere recorded in localities within 20 miles of each other. Although a small number of birds were involvedin some areas and the tapes of these are brief (less than five minutes) and hence incompletelydocument the songvariation of each bird, the recordingscon- sistentlysuggest a very local patternof songdialects. Along the road between 35 and 15 miles east of Kisumu,Kenya, two V. chalybeatawere recorded in villagesat mile 35, one at mile 20, and two at mile 15. Birds in the same village had songtypes in common,but birds in different villages did not sharea singlesong type. A male V. chalybeatataped in Numan, Nigeria, had no songsin commonwith another at Kid, 17 miles upstreamalong the GongolaRiver. The two localitiesare separatedin part by the Benue River which is nearly a mile wide at Numan. On the other hand, a bird of this speciesin the town of Penhalonga,Rhodesia, shared several nomimetic songswith anotherrecorded four milessouth. The very local nature of the song differencesis evident when the total numberof songtypes recorded in an area is comparedwith the range over whichthe birdswere recordedin any one locality. Table 15 showsthe song type data and the largestand mean distancesbetween any two birds recorded in a singlearea. The largestnumber of song types for V. chalybeatawas 39, at Monkey Bay. Even thoughrelatively few songswere recordedthere, the total number of song types was larger than it was in the sample taken at four localities where birds were recorded within a smaller area. Birds at the north and south extremesof the Monkey Bay sample area sharedno songtypes with each other or with the four birds recordedfrom three and one-half to six miles south of the northermost bird, althoughthese four all shared song types with each other. The Monkey Bay sample was taken along the roadsidedriving south of the town of Monkey Bay beginninga half mile south. Similarly,the numberof songtypes was high for the Kisumu area, althoughthe numberof songsrecorded was small. At Maun five birds were recorded within a mile of each other in the town, while a sixth was taped at "Leomarin" about 6 miles from the others. The birds within the village all shared song types, but the 13 song types recordedfrom the "Leomarin" bird were all distinct. The smallestnmber of song typeswas recorded at Marble Hall, where the largest nmbers of individual songs andbirds were recorded. The differencein totalnumber Of songtypes in these areas is not related to greater variation of the songsof individuals in the Transvaalsamples, as the mean nmber of songtypes per bird was the samein each (Table 15). No songtypes were sharedin any of theseareas amongbirds more than three miles apart. These data indicate that each 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 141

TABLE 1.5 DIALECT SONG TYPES TAPE-RECORDED IN VIDUd CHdLYBEdTd IN SOUTHERN AND EASTERN AFRICA

Distance between Number Total birds(miles) of Total song Mean number Locality birds songs typesx song types/bird2 Mean Maximum Marble Hall, Transvaal 13 670 18 11.5 (6) 0.64 1.2 Merensky, Transvaal 7 447 23 13.0 (6) 0.63 2.0 Maun, Botswana 6 256 32 14.0 (3) 2.31 6.0 Monkey Bay, Malawi 7 162 39 12.0 (1) 2.77 8.0 Sabi, Rhodesia 7 110 27 - 3.92 11.0 Kisumu, Kenya 5 92 31 - 12.00 20.0 z Number of song types in each locality excludes minor variants (1-4 syllables dropped at end, 1-2 syllables omitted at beginning). • Mean number of song types per bird includes only males for which more than 40 songs were recorded; numbers in parentheses indicate these birds. songtype is generallyrestricted to an area no larger than a few miles, and populations characterizedby song type homogeneityare restricted to a few squaremiles, even though the distributionof a speciesmay be continuousover many miles where the habitatis suitable. The songtype data for the threelocalities where V. chalybeatamales were recordedwithin the shortestdistances are comparedin more detail in Table 16. At Merensky,where birds were recordedover a longer distance, relativelymore of the songtypes were givenby only a singlebird, whereas relativelymore song types were sharedby the more proximatebirds at Marble Hall and Maun. Because the number of birds recorded was different at the threelocalities, I havestandardized the meannumber of birdssharing

TABLE 16 LOCAL SONG TYPES SHARED AMONG V. CH.4LYBE.4T.4IN SOUTHERN AFRICA

Mean number Maximum of birds NumberNumber of birdsgiving each song type distancesinging each of between songtype Locality birds 1 2 3 4 5 6 7 8 9 1011 12 13 birds(km)(transformed)• Maun, Botswana 5 - 5 4 8 1 1 8.5 Marble Hall, 13 3 5 1 - 1 - 3 4 1 2 6.7 Transvaal Merensky, 7 10 5 4 7 5 I 3 5.6 Transvaal

• An index of population homogeneity,is given by the mean number of birds sharing each local song type (• (row 1 q- row n(a.•)) q- total number of songtypes) standardizedto the Merensky sampleby multiplying this value by (13 .'-- number of birds recorded at locality) for the transformation. 142 ORNITHOLOGICAL MONOGRAPHS NO. 11 each local song type, and the adjusteddata suggesta close relationship betweenthe distancefrom bird to bird and the song heterogeneityin an area. The mean numberof birds sharingeach songtype was greateramong birdsrecorded in a morecompact area (Table 16). To find whether any song differencescould be detectedwithin an area of only a few miles and also whether any sharp breaks between dialects occurred,I determinedthe total number of song types and the number of sharedsong types for each combinationof two singingmale V. chalybeata at MerenskyReserve. All sevenmales recorded were includedin the analysis; the numberof nonmimeticsongs recorded for eachmale rangedfrom 46 to 81. Each male was an alpha-maleand was recordedon his original call-site; hence each male representsa different location within Merensky Reserve. With the aid of aerial photographsI plotted the distancebetween call-sites; the maps allow an accuracyto within 100 feet (see Figure 5). Comparing the number of songsshared between two males and the distance between them for each combination of males showed that males closer together sharedmore of their songtypes. The mean numberof songsshared between maleswithin 3,000 feet was significantlylarger than the mean number of songsshared between males at distancesfrom 3,100 to 7,200 feet, the longest distancebetween males in the sample(p ( .01; t-test). Calculation of a linear regressioncoefficient of percent songsshared betweenmales and the distancebetween them (feet x 100) gave • = 52.6 - 0.76 X. The scatterof points is rather wide, and the confidencelevels are large (F = 26.06; d.f. = 19). Different recordingsof the samebird at the samecall-site should give more than the 53 percentoverlap predictedby the equation;this low value indicatesthat more songsshould have been recordedfor each bird. However, the expressionmay be of interest in visualizingthe nature of songtype changeacross a dialect area. Using the regressionvalues to estimatethe distanceat which only 5 percent of the songtypes of maleswould be sharedgives a distanceof 5,800 feet, and at 6,200 feet only 1 percentof songtypes would be shared. In fact, two males that were 6,400 feet apart sharednone of their songs,but two that were 7,200 feet apart and at the far endsof the samplearea shared5 of their total of 19 distinctsong types (23 percent). Larger samplesof recordedsong will be requiredbefore suitableconfidence limits for the regressionmodel can be established.The generalpattern is clear, however;indigobirds on neighboring call-sitesshare most of the songtypes in their repertoires,and indigobirds further away shareproportionately fewer songtypes. No sharpboundary occursbetween song dialect areaswhere suitablehabitat is continuousover severalmiles. The constellationof songtypes shared by males shiftsgradu- ally amongthe neighborsof a bird at a call-siteto birds on more remote call-sites. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 143

Variation amongspecies in nonmimeticsong.-•Different species living in the samelocality were comparedto find whetherthey sharedcomplex nonmimeticsong types or otherfeatures in their songs.Localities where sympatricmales of two or more different specieswere recordedwere MerenskyReserve, Sabi Valley, Monkey Bay, Sigor,Penhalonga, and Zaria. Generallyeach specieshad its own songrepertoire, and no songtypes were sharedwith malesof other species.Three individuals,however, did sing the nonmimeticsongs characteristic of otherkinds of indigobirdswith which they lived. Two of thesewere males of the V. wilsonicomplex at Zaria. Although noneof the six "camerunensis"males recorded mimicking L. larvata at Zaria sharedany nonmimeticsongs with "wilsoni"mimics of L. rara, the one "camerunensis"and the one "nigeriae" which mimicked L. rara did. In RBP 4959 (a bluish-greenbird), two distinctnonmimetic song types were the only onesrecorded in five minutes. One of thesesongs was identicalto a songof a blue "camerunensis"mimic of L. larvata (Audiospectrographs 28 a and 29 a) while the other songwas identicalto one of two of the recorded "wilsoni" (Audiospectrographs28 d and 29 c). Five nonmimetic songtypes were recordedfor RBP 4884, a blue "camerunensis."Two of thesewere identical with the songsof RBP 4959, the other was sharedwith two "wilsoni" (Audiospectrographs28 a and 29 d), and two otherswere not sharedwith the songsamples of other Zaria birds, thoughone of these wassomewhat similar to songsof normal"camerunensis" (Audiospectrographs 28 b and c, and 29 d). Thus, both of the Zaria indigobirdswhich mimicked the "wrong"firefinch songs shared songs with both blue "camerunensis"and with purple "wilsoni." One of these(4959) was the greenestZaria bird and in color resemblestypical "nigeriae,"and althoughit is the Zaria bird least like the purple "wilsoni"in plumageand color, it sharedsome nonmimetic songswith "wilsoni." The other bird with nonmimeticsong types of anotherspecies was a V. chalybeatamale at Merensky that mimicked the songsof L. rhodopareia rather than L. senegala. From this male I recordednine song types; all of its songtypes were sharedlocally with V. purpurascens,but none with V. chalybeata. These observationssuggest that an indigobirdmay learn the nonmimetic songsof those indigobirdswhich mimic its own foster firefinch species, regardlessof the speciesof indigobirdinvolved. The problemof development of the nonmimeticsongs is discussedlater. To determine whether the sympatric, nondnterbreedingpopulations of indigobirdsare convergentwith or divergentfrom each other in their nonmimetic songs,I comparedthe song characteristicsof different speciesat Merenskyand Penhalonga (where V. purpurascensoccurs with V. chalybeata 144 ORNITHOLOGICAL MONOGRAPHS NO. 11 or V. ]. codringtoni,respectively) and the song characteristicsof different speciesliving in other areas where only one speciesis presentor at all common(Marble Hall, Maun, Tzaneen,and Lilongwe)in Table 14. Different speciesliving in the samearea do not showany tendencyto be more or less different from each other in song duration, frequency,or song complexity than the samespecies living in otherareas. No significantspecies differences are evidentin thesenonmimetic song characteristics (Table 14). Although differentspecies of indigobirdsliving in an area have distinctnonmimetic songs,the over-all featuresof songare very similar amongthe species,and the nonmimeticsong types differ only in the fine details of the individual syllablesand their sequence.As none of the song types are shared among differentpopulations of a singlespecies, the individualsong types are characteristicof local populationsrather than of species. In contrast to the mimetic songs,the nonmimeticindigobird songs provide no evident signal commonto all the differentpopulations of a species. Speciesvariation in the nonmimeticsongs and wingbeatsounds o! courtship display.--Fifteen hovering displaysgiven by male indigobirdswhile they were courtingother birds were tape recordedin the field. On eachoccasion the male sang a nonmimeticsong during the display. In all instances,the vocalizationand displayoverlapped in time as shownon the audiospectrographs;sometimes the songbegan after the hover did, and at other times the songpreceded it at the end of a prolongedchatter. In instancesof prolongedaerial display, the last part of hoveringwas unaccompaniedby song. One of the displaysrecorded was given to a visitingPasser griseus and anotherto an Amadina !asciata;the otherswere given to female indigobirds. In two instances,a mimeticvocalization was given immediatelyafter the displayas heardon tapes,but all recordedsongs of hoveringmales were strictly nonmimetic;no mimetic syllablesare evident in the audiospectrographsof any of thesesongs. Examples of the songsand the soundsof the wingbeatsof the bouncingmales are shownin Audiospectrograph30, and certaingeneral features of songare comparedin Table 17. Songsgiven during precopulatory display vary considerably.When birds of the samespecies and locality are compared(two "wilsoni" and also two "camerunensis"at Zaria; two V. chalybeataat Maun), it is clear that no singlespecies-characteristic song is given (Audiospectrograph30). The two V. chalybeatahad different songsin the single display recordedfor each bird, whereasthe two "wilsoni" shared a commor•song. Hovering and associatedsong were recorded in one male V. wit•onion threeseparate occasionson 1 September1968, at Zaria. Two of thesesongs were the same songtype, while the third was a differentsong type; this third songtype wasshared by anotherhovering "wilsoni" recorded on 29 July and collected. One "camerunensis"at Zaria alsohad two identicalsong types in two hovers 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 145 146 ORNITHOLOGICAL MONOGRAPHS NO. 11 and anothersong type on its third recordeddisplay. Neither was the same song type as in the other displaying"camerunensis" that I recorded. The songsgiven by malesin this displayare identicalto the nonmimetic songsgiven in prolongedsequences of advertizementsong and are not restrictedto courtshipcontexts (Audiospectrographs 30 c and 29 d; 30 d and 28 e). The temporalcomponents of the displayare evidentin the recordedwing- beat soundsof the male indigobird, and these are compared in Table 18. Probablyeach soundrepresents a distinctwingbeat. The numberof sounds in eachsingle series varies even between series in the samedisplay (Audio- spectrograph30 b) as well as between displaysby the same bird (Audio- spectrograph30 c, d), and thus variationsin this measurementof the display showit not to be consistentlydistinct in each speciesof indigobird. Similarly, the total numberof wingbeatseries and also the lengthof aerial displayare sot consistentcharacters of any one indigobirdspecies. No speciesdifferences in the numberof wingbeatsor in their frequencywere apparent. Perhaps the similar body size and wing size of the indigobird taxa preclude any but aerodynamicconsiderations from affectingthe aerial display. As is evident in the audiospectrographs,the sound of the wings in this display is not at all loud comparedwith the song, and the aerial bobbing display itself is mainly a visual signalto the female rather than an auditory one. The nonmimeticsong types given by male indigobirdsduring their courtship display are the same as those given during the song bouts on the call-sites, and as with singingmales on their sites the song types are sharedlocally amongdifferent birds. An individualsometimes may use one songtype and sometimesanother in its courtshipdisplay. Becauselocally the different males may sharetheir songtypes during courtshipdisplay, no individual recognitionamong indigobirdsis at all evident in their nonmimetic songs. Both the nonmimeticsongs and the frequencyof wingbeatof displayingmales vary from display to display in the same bird, and different birds as well as differentindigobird species are all similarin this behavior(Audiospectro- graph 30, Table 17).

REPERTOIRESIZE, MATING SYSTEMS,AND INDIVIDUAL RECOGNITION Each male indigobirdhas about 12 or more distinct nonmimeticsong typesand 6 or more mimeticsong types; thus he has one of the largestsong repertoiresknown among birds. The adaptive significanceof such a large number of songs in the indigobirds is unclear. In birds of the northern temperateregion which have many songs,large song repertoireshave been thoughtto provide the meansfor the recognitionof individualmales. Song repertoiresof large size have been describedin some monogamousbirds 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 147 with a strongpair bond,but it is unlikelythat the large repertoireof songs of an indigobirdis importantin individualrecognition. In the indigobirds a pair bondis absent,and also the completerepertoire of songtypes and their sequenceare to a largedegree shared among most local males. As indigobirds are polygynous,the femalesprobably select as matesthe individual malesmost likely to be of high fitness. Males may signaltheir fitnessnot by individualsong differences but by the behavioralcontext of song. The fact that a male singson a call-site is itself a sign of some high fitness, as alpha males at a site are the relatively older birds which have survived from year to year. The indigobirdsresemble some polygynoustropical hummingbirdsin their emphasisof local songdialects, their lack of individual songsignals, and theiruse of songsites at traditionalplaces where the males sing and which the femalesvisit for mating (Snow, 1968). Large song repertoiresare foundin somemonogamous New World finchesand sparrows, suchas the SongSparrow (Melospiza rnelodia) •and Cardinal (Cardinalis cardinalis), but in other equally monogamoussparrows such as White- crowned Sparrows (Zonotrichia leucophrys) each male generallyhas only a singlesong (Mulligan, 1966; Lemon, 1968; Marler, 1970), and no dear trend is apparentat the presenttime for monogamousspecies to have large songrepertoires. Indigobird socialstructure suggests no complexbehavioral interactionsthat might call for individual recognition. The information in an indigobirdsong which signals its speciesis containedin the mimeticsongs. Althoughthe nonmimeticsongs may distinguishthe mimic from the model for the females,having more than one nonmimeticsong type within a population seemsunnecessary and redundant. Perhapsthe large repertoire in songsresults in part from occasionalinput of dispersingbirds that arrive from other dialect areaswith new songsand it may be of interestto record songsof remote, isolated populationsto determinewhether they have a smaller number of song types.

ORIGIN OF NONMIMETIC SONGS AND SONG DIALECTS No experimentalstudies on song developmenthave been reported, and we have no direct evidencethat the indigobirdslearn their nonmimeticsong types from older indigobirdsliving in the same locality. Although some kindsof birdscan developstereotyped, complex songs through improvisation and crystallizationof songsheard when they are hand-rearedand can hear no other birds (Mulligan, 1966; Dittus and Lemon, 1969), it is unlikely that the indigobirdsdevelop their complex song types by improvising,because all birds in a locality shareeach of their songtypes with other local males, and it is most improbablethat each bird within a locality would by chance alone improviseprecisely the same very complex set of song types as its neighbors,yet in different localities improvise completelydifferent song 148 ORNITHOLOGICAL MONOGRAPHS NO. 11 types. Local learningof the traditionalsong types throughhearing the singingadult indigobirdsseems much more likely, for severalreasons. First, someof the complex,stereotyped songs include mimetic syllables as well as the nonmimeticsyllables. If indigobirdsacquire their mimetic vocalizationsby hearingtheir foster parentswhile they are young,then it seemslikely that complexnonmimetic song also is learned becausethe mixed complexsongs may includethese mimetic syllables. The arguments for the learningof mimeticsong apply here to the problemof the origin of the complexnonmimetic songs as well. Second,the indigobirds that sangthe wrongmimetic songs (birds mimicking fosterfirefinches other than the fosterspecies mimicked by otherindigo- birdswhich they morphologically resembled) also sang the nonmimeticsongs of the wrongkind of indigobird.It seemslikely that thesebirds had learned the nonmimeticsongs from other indigobirdsmimicking their own foster species. Third, the pattern of variationof songtypes from place to place suggests local learnedsong traditions rather than geneticspecializations of each local indigobirdpopulation, because if song type differencesamong populations were determinedby correspondinggenetic differences, the complexityof each songtype and the large repertoiresize of the indigobirdswould imply a huge proportionof unique, geneticallymonomorphic loci within each population. Such a pattern of highly monozygous,genetically specialized populationsseems improbable in these birds, and I know of no kind of sexuallyreproducing higher vertebrate with a high degreeof unique, mono- morphicgenetic loci amongeach local population. One observationof anotherspecies of viduinethat sangthe songsnot of its own speciesbut of the first viduineit had ever heard givessome direct evidencefor the learning of nonmimeticsongs in the viduines. A male Straw-tailedWhydah (Vidua fischeri), which I saw in one of Nicolai's aviariesat Seewiesenin 1970, I heard give not typical fischeri nonmimetic soundsbut rather the nonmimeticsongs of a male V. chalybeataliving in the adjacentaviary. This Iischeriwas raisedby a pair of BengaleseFinches in 1969 in Tanzania by Nicolai, in acousticisolation from other viduines, and wastransported back to Germanyand placedin the aviaries(J. Nicolai, pets. comm.). Apparently the songsof the l•. chalybeatamost closely approximatedthe critical missingfeatures in the soundexperience of the youngIischeri (the nonmimeticsongs of thesebirds are similar, though fischerihas longer, more complexsyllables and the notes are spacedmore widelythrough a song,in birdsI haverecorded in Kenya). Along with its BengaleseFinch vocalizations,the Iischerisang not only the nonmimetic indigobirdsongs but also the L. senegalafirefinch mimetic soundsof the indigobird,suggesting that it did not distinguishbetween the mimeticand 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 149 nonmimeticsongs of the viduinesin its specificsensitivity to learningthese songs.I am mostgrateful to J. Nicolai for makingthis observationavailable. Finally, in other songbirdswith local dialectsthe local song types are knownto be acquiredby the youngbirds by hearingthe songsof the adults in theirarea in all specieswhose song development has been studied (Thorpe, 1958; Marler, 1970; Marler and Tamura, 1964; Lemon and Scott, 1966; Dittus and Lemon, 1969; Thielcke, 1969a). The uniform importance of learningin the ontogeneticdevelopment of a local songdialect among song- birds suggeststhat learningis generallyimportant in most specieswhich havelocal dialects, including the indigobirds. The age span during which the indigobirdsare receptiveto indigobird nonmimeticsong and acquirethe sensoryinformation about the local song typesprobably begins while the birds are in the care of their foster parents and is completedafter severalweeks of independence.Probably the young indigobirdsare sensitiveto the nonmimeticsongs only of the adult indigobirds which mimic their own fosterspecies and in this way screenout the soundsof other species.Young indigobirdsare not raisedby their own parents and so cannot be sociallyguided or directedto their songs. However, the youngindigobirds may retain the selectivitynecessary for learningthe songsof only a singlekind of indigobirdthrough generalizing their song sensitivityto only thosebirds that mimic their foster firefinch species.In areas such as Merensky Reserve, two speciesof indigobirdsare equally abundant,and a youngbird might hear the songof the other speciesinstead of its own parental speciesnear its foster firefinch nest. The problem of the filtering of this diverseinformation to allow the learning of only the songtypes of a singleindigobird species may be a functionof earlyimprinting ontothe callsof the fosterparents. Thus, if an indigobirdhears the vocalizations of L. senegaladuring its nestlingor fledglingdays, it may then be receptivein its learningonly to the vocalizationsof neighboringmale indigobirdswhich mimic the samespecies of firefinch. Indigobirdsprobably learn their songtypes early in life beforeestablishing themselvesat a call-siterather than first establishingthemselves and then mimickingthe songtypes of the nearestneighbors. Evidence is found in the fact that the new, successivereplacements of male V. chalybeataat a call- site at Marble Hall where I had shot off the alpha male and his first few singingreplacements were progressivelyyounger birds, and thesereplacement birds had most likely never sung before, becausethey had never before acquireda call-site. The songsof thesefirst-year birds were fully formed andmatched precisely the songtypes of the oldermales at the samelocality. It is possible,on the otherhand, that theyhad heardand learnedthe songs merelya shorttime earlier,before the first maleswere shotoff. It doesseem likely, however,that the songof the indigobirdsis essentiallycompleted 150 ORNITHOLOGICAL MONOGRAPHS NO. 11 in large part within the first year of life. Two all-sparrowyplumaged singing birdswere tapedat the "nigeriae"site at Panshanu,Nigeria, after two males in breedingplumage were removed,and both of thesehad the same stereotypedsongs as the older males. One (RBP 4944) was successfullycollected; its skullwas only 18 percentpneumatized, and it wasprobably a bird hatched late in the previousbreeding season. Young, newly independentjuvenile indigobirds sometimes visit the call- sites of their own speciesshortly after they leave the foster firefinches. Severaltimes at Maun I saw a youngindigobird in juvenal plumagefly to the call-siteof a male V. chalybeata,perhaps attracted by the songof the the singingmale. By seekingout and visitingthe call-sitesof singingadults, the youngindigobirds may learn their local songtypes when they attendthe singingmales at the call-sites,where they may establisha socialbond with the singingmale companion. Often the adult males courted the young indigobirdsat the call-sites,hence thesevisits may involve somerisk to the young of being knockedoff the perch by the adults. The traditional use of specificcall-sites by the singingmales provides a settingfor a young indigobirdto return repeatedlyand learn the local dialectal nonmimetic songsas well as perhapslearning from the adult indigobirdsfurther details of its mimetic songs. How long the young indigobirdsremain sensitiveto the songsthey hear and can learn new songsis unknown. The sensitiveperiod may persistfor severalmonths, because a youngbird hatchedat the end of a breeding seasonmight otherwisenot have time to learn the local song types before all local adult malesceased their singing.The adult malesmolt and do not singduring the winterin southernAfrica. A youngbird that had not learned its songsthen might retain an ability to learn songsuntil early in the following breedingseason. On the other hand, probably most birds learn all of their songtypes before they are more than a few weeksindependent, because in each study area where severalmales were recordednone of the maleshad song types not sharedby their neighbors. If young indigobirdsdispersed from the localitywhere they were raisedand movedinto anotherarea while still in their sensitiveperiod, then they would learn the songsof the second dialectarea also. Thesemales when adult would then sing somesongs like their neighborsand othersunlike any local neighboringindigobirds. As no suchbirds were recordedin the field, probably most birds do not disperse from a dialect area after they have learned their nonmimeticsongs. If indigobirdscontinue to learn nonmimeticsong types after the first year, one might predict that birds of one year might have a smallerrepertoire size or a higher proportion of errors or song variants in their songs. To compare the songsof young and older males I tabulated the proportion of skull pneumatization(older birds are more completelypneumatized) and 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 151

TABLE 18 ANALYSIS OF VARIATION OF SONO TYPES IN A POPULATION OF VIDUA ½HAœYBEATA1

Number song type Percent variants Percent skull Total Number song types unpneu- number song Nonmimetic Mimetic with no Bird matized songs types (a-c) (d-g) variants

Farm a 40 68 17 8 1 47 Farm b 50 60 13 4 - 69 Farm c 65 14 5 2 - 60 Farm d - 27 7 2 - 72 Farm e 75 29 9 4 - 56 Farm ! - 17 7 2 1 57 Farm g 90 23 9 3 - 33 Farm h 90 93 6 2 2 33 Road a 50 38 9 2 - 78 Weir a - 131 12 8 2 13 Bridge a - 58 11 4 - 64 Fence a - 83 10 5 1 40 Hut a - 29 6 4 - 33

ß Symbols are the same as in Table 13. positionat a call-site(whether alpha or a replacement)for eachV. chalybeata tape recordedand collectedat Marble Hall (Table 18). Male h at the Farm call-site,the youngmale for whichthe mostsongs were recorded,had fewer songtypes than all five of the older a and b males for which more than 50 songswere recorded. At leastsome relatively young birds may have already acquireda large songrepertoire, however. A male V. chalybeatarecorded at "Leomarin,"Botswana, had 13 songtypes in 62 songsrecorded, and its skullwas 70 percentnonpneumatized, indicating an ageprobably not greater than a year. The proportionof variantsin the songsappears to be similar in all male indigobirdsregardless of skull pneumatizationor social rank (Table 18). Few birdswere tape-recorded for longperiods except the alpha males,and thusit is unclearwhether the youngermales have fewer song typesin a largesample of theirsongs and so may continue to addsome songs aftertheir first year, but the songtypes they do singare as well formedand haveno morevariants than do the songsof the oldermales. To findwhether adults may acquire additional song types I haveperiodically recordedthe songsof three male V. purpurascensand one V. chalybeata maintainedin my aviariesfor morethan two yearswith conspecificssinging othernonmimetic song types. The songtypes of eachbird have remained stablefrom year to year,and none has acquiredthe songtypes of another male,although they have prolonged auditory exposure to the othersongs. 152 ORNITHOLOGICAL MONOGRAPHS NO. 11

New songtypes in a populationmay arisefrom two main sources.First, alien birds may enter a population,establish themselves at a call-site,and sing the songswhich they learned in the area where they were reared. In this populationthe residentindigobirds or their youngmay then learn the new songsintroduced by the newcomersas well as the traditionalsong types. Second,new songtypes may be deriveddirectly from older songtypes locally by a processof culturalerrors. Mistakesin copyingthe songof the earlier generationsmay lead to the singingof new songtypes varying in detail, and thesemistakes may then becomeestablished by being re-copiedfaithfully by later generations.Because of the great homogeneityin songswithin a smallarea, mostnew songsin a dialectarea probablydevelop through mis- copying the songsof local residentsrather than by learning the songsof immigrantmales. If a large number (perhaps more than 10 percent) of birds in a population were immigrants,one would expect a comparable proportion of birds to have song repertoiresunlike their neighbors. In addition,as eachmale usuallyhas 12 or more nonmimeticsong types in his repertoire, one would expect an immigrant to have a large proportion of songtypes unshared by its neighbors.However, in the populationsstudied all malesshared songs with neighbors,and mostor all songtypes are shared by severalbirds. On the other hand, there is sufficientvariation within a populationin the songsof somebirds to suggestcopying errors in the learning and singingof the songtypes. Very local changesin songmay then become establishedthrough tradition as new songdialects. Most birds sing most song types in an unvarying, stereotypedmanner (Tables 13, 16, 18). However, someindividuals may vary theseby omission, addition,or alterationof individualsyllables or phrasesof a songtype. Several of the kinds of song aberrationsare illustrated and describedbelow. The songtype variantsinvolve mimetic syllables as well as nonmimeticsyllables. Alterations in the mimetic notes are remarkable as most songsdo not involve mimeticnotes and in thosethat do the nonmimeticnotes predominate. Somesong type variantsare restrictedto certainindividual birds, whereas othersong types may be givenin two or threevariations by an individual, and other variantsare sharedamong several local males. 1. Variable mimetic introduction. Many complexsongs are introduced by mimeticnotes, particularly by the mimeticalarm note in V. chalybeata. Audiospectrograph31 showsthree successiverenderings of the same song typeby a maleV. chalybeataat MarbleHall. Eachsong has the same number of notes, but the mimetic contact call notes are variable and are alternated betweensongs with the mimeticalarm notes. 2. Omissionof a note within a song. One of the song types at Marble Hall that wasshared by all 13 malesrecorded was givenin a secondversion 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 153 by one male, whichalone always omitted one mimeticalarm note (Audio- spectrograph32) in all eightof his tapedrenditions of this songtype. 3. Alterationof one syllable. Somesongs had notesgiven in different mannerby the samebird in successiverenderings of a song,and somesong typeswere given in stereotypedmanner by each bird but with a slight variationin thedelivery of oneof the syllables.Audiospectrograph 33 shows three variationsin the deliveryof a mimeticcontact call note by a male V. chalybeataat Marble Hall. Severalrenditions of this song were given with the contactcall note like that of the host,while occasionallythis syllable and no otherwas given with markedmodulation in pitchquite unlike the callsgiven by the firefinchitself. Thesevariations suggest a sourceof origin for someof the nonmimeticnotes in the indigobirddialect songs: variable alterationsof a mimetic syllable incorporatedinto a complex song by one bird, and a stereotypedimitation of one of thesevariants by youngindigo- birdshearing the variant songs in theiryoung life. 4. Alteration of several syllables,compounding the above variations. Two male V. chalybeatarecorded in the same tree during different weeks at Olorgesailie,Kenya, showedvariants of the structureof syllablesthat appearedto be in homologoussongs (Audiospectrograph 34). Another exampleinvolving the substitutionof mimetic notes as well as the alteration of nonmimeticsyllables is apparentin two songtypes, each given by most V. purpurascensrecorded at Penhalonga(Audiospectrograph 35). 5. Duplicationor repetitionof songsor of parts of songs. Sequences of songtypes given in a singingbout showthat somesong types are repeated. Occasionallythese songs were givenvery closelyone after anotheras a single song duplicatingthe terminal complex syllables(Audiospectrograph 36). Songduplication accompanied with an alterationof some syllableswithin the duplicatedportion may lead to novel combinationsof notesand to new song types. 6. Deletionof terminalsyllables. The omissionof the final one to four syllableswas the mostfrequent form of songvariation, and this was given bothby individualbirds in successiverenderings of the samesong type and by differentbirds, some of whichregularly dropped the syllablesand others regularlyincluded them in thesame longer song type (Audiospectrograph23). 7. Rapid alternationof differentsong types with a decreasein the time intervalbetween songs may leadto the fusionof two songtypes into a single one,as suggestedin Audiospectrograph37. When this occurs,the beginning or endingof the long mixedsong may then be dropped,resulting in a new songconforming to the typicalsong duration. All of thesevariations of songtypes were found among neighboring male indigobirdson the samecall-sites or on adjacentcall-sites. As each bird 154 ORNITHOLOGICAL MONOGRAPHS NO. 11

has many songtypes in its repertoireand as fewer than 100 songswere recordedfor most birds, a quantitativeanalysis of the frequencyof each form of song aberrationsis not now possible,but more than half of all variationsnoted were deletionsof terminal nonmimeticphrases and alterationsof introductoryor mid-songmimetic notes. The samekinds of variationsin songwere recordedin indigobirdsseveral miles from each other. Although birds did not share identical song types acrossthis distance,song type homologiescan still be seen between the dialectpopulations. Examples are seenin the songtypes recordedin V. chalybeata at Maun and six miles from Maun, Botswana (Audiospectro- graph 38) and in V. f. nigerrimarecorded one mile apart on the Lilongwe- Likuni road in Malawi (Audiospectrograph39). In the Maun birds local variations in the stereotypedmimetic vocalizationsare also apparent, as all birds recorded in the town of Maun shared the mimetic song type • Audiospectrograph38 c, whereasthe one bird recordedat "Leomarin" had a differentversion which was probablyderived from the Maun dialect (in this case several call-sitesoccur in the Maun population but only one at Leomarin, and the larger Maun populationis probablymore stable and may have providednew recruitsfor the Leomarin area). Whether or not these mimetic calls also reflect local differences in firefinch vocalizations is unknown,but as the syllablesare like thosegiven by beggingyoung fire finches out of the nest rather than by adult firefinchesit seemslikely here that the ritualized,dialectal mimetic songsare learned by indigobirdsfrom indigobirds. Birds recordedmore than a few milesfrom eachother had songtypes more different from each other. A few song types of V. purpurascensat Merenskywere rather similar to thoseof purpurascens17 miles down the Letaba River at Kondowe, but each differed in the structureof two or more of its syllables. As mostof the detailsof songdifferences between birds in differentneigh- borhoodsa few miles from each other are paralleled by song differences of birdswithin a singleneighborhood, it seemslikely that dialectaldifferences amongneighborhoods are derivedfrom simplecopying errors in a learning process.Some songtype variantswere recordedin only a singlebird, and thesevariants may have been first-generationcopying errors. If many of thesevariants persistin a population,the mistakesinherent in the song copyingabilities of the indigobirdsmay lead to the large song repertoire size, with some perhaps neuropsychologicallydetermined upper limit set on numberof songtypes that can be rememberedand sung. The samekind of songtype variationsoccur in indigobirdswithin a restrictedlocality and betweennearby localities, and the samebehavioral mechanisms responsible for mis-copyinglocal songtypes probably have beenthe sourceof the distinct 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 155 differencesin songtypes betweenlocalities, where a seriesof errors in the transcriptionof song types has led to the divergenceof an earlier common behavioral tradition.

SONG DIALECTS AND POPULATION STRUCTURE The numberof individualsthat are likely to interbreedwith each other is generally much less than the total number of individuals in a species (Ehrlich and Raven, 1969; Selander, 1970). To determine the structure of a population--thenumber of individualscomprising an effectivelyintra- breedinggroup and the distinctivehessof neighboringgroups--population biologistshave used mainly techniquesof capturingindividuals, marking them for recognition,releasing them, and finding where they may move during their lifetimes. Becausebirds are difficult to mark and recapture locally in large numbers,few studieson the populationstructure of birds have been completed.One studyof SongSparrows (Melospiza melodia) aroundSan FranciscoBay has shownthat populationsof small birds may indeed be very local, as the mean distanceover which a sparrow moves from its siteof hatchingto the placewhere it breedsmay be lessthan a mile (Johnston, 1956: 41). It would be of interest to know whether such local populationsare restrictedin their movementsbecause of the patchinessof their specialhabitats or becausebirds in a local populationactively behave to excludeany wandering birds from otherpopulations. Songdialects may providea new way to comparethe populationstructures of dialectalbirds. The localdialect differences in birdswhich have stereotyped songtypes given by all membersof a restrictedpopulation are generally learned by the young (from parents and neighbors)within the first few monthsof life. In birds that terminatethe sensitiveperiod for learning song before they disperse,the song types learned in early life may be behavioral markers that record the dialect area whence each bird comes, muchas numberedrings or bandsattached to a youngbird in the nest may be an individualmarker. Thesecrystallized behavior patterns may be used to trace the probableorigin of eachbird heard in an area, and sampling techniquesof song recording,estimation of population density, and de- terminationof matingsystems may permit goodestimates of the population structure. Song types may be interpretedas markers that show several features of population structure in dialectal birds. First, the number of birds in a "neighborhood"(Wright, 1969: 291) may be estimatedby countingthe numbersof birds sharingcommon song types that more remote birds do not sing,the ratio of singingand non-singingmales, and the femalesin an area. Second,the geneticallyeffective size of indigobirdpopulations may 156 ORNITHOLOGICAL MONOGRAPHS NO. 11 be estimatedfrom the data on neighborhoodsize and mating systems(see p. 157). Third, the way in which populationsreplace one anotherfrom areaswhere birds singone songtype to areasof other songsmay be shown by determiningwhether the repertoiresof two dialect areas intergradeor whetherthe birds in adjoiningareas all sing mutuallyexclusive song types along a boundary. Fourth, a rough notion of the amount of movementof individualsbetween areas may be graspedby determiningthe proportionof birds in an area with songtypes very different from their neighbors'song types, and the distancesinvolved in individual dispersalmay be estimated by comparing the distancesover which these odd songs are shared in different local populations.Finally, experimentalstudies may be carried out to find whetherthe local songdifferences are activelyused by the birds to sortthemselves out into distinctpopulations or are simplypassive markers of their home populations.All of these aspectsof populationstructure were consideredin the indigobirds.Experimental tests of the responsesof birdsto their songsare discussedon pp. 165-173, and the interpretationof indigobirdpopulation structure from their songsfollows.

ESTIMATES OF POPULATION STRUCTURE FROM SONG DIALECTS IN THE INDIGOBIRDS The songdata indicatea populationsystem of neighborhoodsthat intergrade with each other, rather than a seriesof discreet,noninterbreeding neighborhoods.As discussedearlier, indigobirdsmore than 10 miles apart sharedno songtypes, though some possibly homologous, similar songtypes may be tracedacross this distance.Within this distancesome differences are apparentin V. chalybeatasong types, as birdsonly 4 milesapart at Monkey Bay sharedno songtypes. Males within 3,000 feet of eachother at Merensky Reserveshared more songtypes, on the average,with each other than they did with singingmales separated by greaterdistances (3,100 to 7,200 feet), althoughalmost all birds sharedsome songtypes with every other male at Merensky. Dispersal of birds from one area to another appearsto be generally restricted,because no instanceswere found of birds more than a few miles apart sharingany songtypes. No songtypes were sharedin detail in any V. chalybeatarecorded more than three miles apart. (The samewas true for the other speciesof indigobirdsstudied, although I did find two V. purpurascensmales recorded at a dam on a smallcreek six milesfrom the Letaba River at Merensky Reserve sharing song types with males of their speciesalong the river.) A low rate of movementof birds is also indicatedby the fact that in all populationsadequately tape-recorded, every male sharedsong types with its neighbors,and at Marble Hall all 18 recorded 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 157 songtypes were shared,with at least one songtype being sungby all 13 malesrecorded. If a reasonablylarge proportionof birds movesmore than a few miles from the place where they hatch and learn their song types, somebirds in one localityought to have someor all of their songsunlike their singingneighbors. The data availablesuggest that somewhatless than 10 percentof all males move more than one or two miles from the place wherethey are hatchedto a call-sitewhere they sing. Althoughinterpopula- tion dispersalapparently is low, the isolationbetween populations character- ized by their songtypes is incompleteas there is a gradualreplacement of songtypes and a continuumof neighborhoodswhich singthem over a distance of a few miles. As male indigobirdschase all other males, regardlessof species,from the areasaroun•d their call-sites,it is unlikelythat malesdis- criminateselectively against males of songtypes of anotherneighborhood. Playback experimentsshowed no differencein the responsesof males or of females to the nonmimeticsong types of their own dialect or of other dialectsin the samespecies. Evidently a distanceof a few milesis sufficient for indigobirdsto generateand maintainlocal song differences between neighborhoods,even when neighborhoods are joinedby intermediateneighborhoods in suitable habitat. To find a meaningof the term populationwhich reflectsthe number of individualslikely to interbreedwith each other, I am usingthe term neighborhood,which Wright (1969: 291) definesas "the populationof a region in a continuumfrom which the parents of individualsborn near the center may be treatedas if drawn at random," that is, the area within which distance effectsare negligible.Birds which sharesong types are much more likely to interbreedwith each other (and to be geneticallysimilar) than birds which do not share songtypes, simply becausemost birds apparentlydo not move out of the home songtype localityinto other nearbylocalities betweenthe time theylearn their songsand the time they breed. Probably most indigobirdsfind their matesliving within a few miles of each other, in the samelocality where their parentsmated, as indigobirds(judging from their songs)are highlyfaithful to a locality,not only to a certaintree (the call-site) but alsoto a mile or a few miles of habitat. Indigobirdneighbor- hoodsthen consistof the populationsharing most of their songtypes with each other. The populationdensity of indigobirdsvaries among localities, but at MerenskyReserve with a continuumof overlapof songtypes over a few miles;there were about4 call-sitesof V. chalybeataper mile of river- bank habitat. At any one time only one male sang at the call-site,but malesmay replaceone another,and replacementmales generallysing the samesong types as alpha males (Table 13) sobelong to thesame behaviorally- recognizedneighborhood. Several females use each call-site. As at least 158 ORNITHOLOGICAL MONOGRAPHS NO. I1 four femalesat onecall-site may be layingon any oneday, andas histological examinationof the ovariesof breedingindigobirds that I have collected showsan intervalof a few daysbetween clutches, perhaps as many as 10 femalesmay use a call-siteover the courseof a few weeks. At one call-site at MarbleHall I collected9 malesin as manydays. Assumingequal numbers of malesand femalesin an area, the densityof indigobirdpopulations, at least at Marble Hall and Merensky,may be about 60 birds (one year or more of age) per mile of riparianbushland. Because song homogeneity is significantlygreater among birds less than a mile from each other than betweenbirds over twice this distance,the neighborhoodarea of indigobirds may be on a similarorder of magnitude.If we acceptthe sharingof song typesbetween birds two milesapart as evidenceof the extentof a neighborhood, then indigobirdneighborhoods may compriseabout 100 individuals. The effectivepopulation size is probablysomewhat less than neighborhood size, because not all individuals breed. Examination of over 80 female indigobirdscollected in the breeding seasonshows that almost all females breed, whereasonly the alpha males at each call-site mate. The effective sexratio is probablyaround 1: 6, and so the highlypolygynous mating system suggestsan effectivepopulation size of lessthan 100 breedingindividuals.

SIGNIFICANCE OF POPULATION STRUCTURE The population structure of V. chalybeata (and presumablythe other indigobirdsalso, as they all have local songdialects) appearsto consistof local pocketsof inbreedingindividuals along a linear range throughwhich behavioral differencesoccur over a few thousand feet. These local neigh- borhoodscharacterized by a set of characteristicsong types intergradewith one another,with birds at progressivelygreater distances sharing fewer of their song types with each other. The very local population structure, characteristicof the indigobirds,although apparently not recognizedpre- viously in avian populations,corresponds closely to the intuitively derived "isolationby distance"model of populationstructure developed by Wright (1946). Wright (1969: 300) suggeststhat in such a system,populations with a neighborhoodsize of lessthan 200 will be subjectto "not negligible" effectsof randomgenetic differentiation due to the samplingerrors of genetic segregationand recombinationin small populations,and he noted "if there is considerabledifferentiation of neighborhoods,from accidentsof sampling, this builds up differentiationof much larger areas," thoughthis processof geneticdifferentiation would be slow. The small neighborhoodsize of the indigobirdsmay havebeen important in their evolutionaryhistory, and some effectsare discussedon p. 263 and p. 294. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 159

The low dispersalrates may be responsiblefor the occurrenceof wide- spreadareas where firefinches are abundantbut their indigobirdmimics are absent.In KenyaL. rubricatais commonand widespreadin the highlands, but the onlykind of indigobirdthere is V. chalybeata,and it mimicsonly L. senegala.At somelocalities I foundfirefinches common but no indigobirds presentin the breedingseason (Brits, Transvaal; north side of Lake Ngami, Botswana),and Nicolai (1967) found the samein severalparts of Uganda and Tanzania. The presenceof unexploitedfirefinch populationssuggests an opportunityfor somedispersal of indigobirdsfrom their homearea. Over shortdistances, at least,I found both L. senegalaand V. funereain an area wherethey were absentonly a few monthsearlier. This was wherea forest at 1,200 feet elevation on the Lusitu River in eastern Rhodesia was still standingin October,1966, but wherepatches had been felled by April, 1967, and replacedby corn. Theseareas were thenoccupied by the firefinchesand indigobirds.Previously felled areas within a half mile were also occupied by thesebirds as theyhad beencollected there in 1965 (specimensin NMR; H. D. Jackson,pers. comm.). The very local nature of indigobirdpopulations, even when suitable habitat is widespread,may be related to the problemof finding a suitable mate. The host firefincheshave local song differencesmimicked by the indigobirds,and althoughI do not know from my limited field recordings of the firefincheswhether firefinch dialectsand populationsare as local as in the indigobirds,they may well be as indigobirdsonly six miles apart had different stereotypedmimetic songs (Audiospectrograph38). If a male indigobirddispersed and establishedhimself on a call-sitein a remote population,his mimeticsongs might be lesslike thoseof the localfirefinches than would be the mimetic songsof the resident males. This immigrant mightthen be lesslikely to gain residentfemales imprinted to the local firefinch dialectsthan wouldbe the residentmales. Similarly,a dispersingfemale mightbe stimulatedto a lesserdegree to undergoovarian development and ovulationif shewere out of hearingof the firefinch dialectlike that her foster parentssang, and she would leave fewer offspringthan a residentfemale. Both males and femaleswould have a better chanceof breedingsuccessfully in the area wherethey were rearedand wheretheir songbehavior was fixed on the local firefinches. Rather than favoring individualsthat range widely and colonizedistant areas, selection may favor individualswith a high degree of site tenacity. The questionof whetherfemale indigobirds seek out individualsof similar songdialects as matesand the possiblefunction of songdialects in maintaining differentpopulations as distinctgroups is discussedin the followingsection on the responsesof indigobirdsto playbacksof their recordedsongs. 160 ORNITHOLOGICAL MONOGRAPHS NO. 11

BEHAVIORAL CONTEXT OF INDIGOBIRD VOCALIZATIONS AND THE RESPONSES OF BIRDS TO SONGS

The behavior of the indigobirdsin relation to mimetic and nonmimetic song was studiedto determinethe significanceof these songsto the birds themselves.The numbersof each kind of song were determinedfrom tape recordingsof indigobirdsin different stagesof the breedingcycle. The calls associated with various behavioral contexts such as encounters between males or betweenmales and femaleswere noted. These two approachesI carried out mainly with Vidua chalybeataand V. purpurascensin the Transvaal. Also, I broadcasta seriesof tape recordingsof mimeticsongs and nonmimetic songsin the field to males and in captivity to females,and I noted their responses.These responsesprovided some evidenceon the importance of mimetic and nonmimetic vocalizationsin speciesrecognition.

ANALYSIS OF SONG RECORDINGS The behavior of male indigobirdschanges through the breedingseason. Priorto thefirst matings much time is spent•by the males in establishmentof dominanceat the call-sites.Males at this time usuallyignore females visiting the sites, whereas establishedmales in the breeding seasonusually court their female visitors. To determinewhether stud males have song quality different from that of conflictingmales before the breedingseason or after removal of the dominantmales from their sites, tape-recordedsamples of song of each of thesekinds of males were studiedaurally. The numbers of mimeticand nonmimeticvocalizations given by a bird duringa one-minute sample of the recording were scored. For each bird three one-minute sequencesof the songbouts were examined,except in pre-breedingseason malesin which one V. purpurascenstaped on different days providedtwo such sequences.Both chattersand mimetic phrasesseparated by a time intervalreadily apparent to the ear (about0.2 sec.) were scoredas individual vocalizations.Where mimetic elementsoccurred within the complex nonmimetic songsthese songs were scoredas singlenonmimetic songs. Table 19 recordsthe proportionsof mimetic and noumimeticsongs in V. chalybeataand V. purpurascensrecorded in Transvaal. They show no differencein frequencyof mimetic songthrough the breedingseason. The indigobirdsall had about 76 percent noumimeticand 24 percent mimetic songsin their songbouts. While the quality of songw. as apparentlyconstant, the alphamales in the breedingseason sang more regularlythan did disturbed males;recording sessions in whichthe birdsdid not singwere excludedfrom this analysis.Frequency of mimeticsongs was as high in non-breedingbirds prior to the breedingseason, even in molting males. A male V. chalybeata at Merensky and a male V. purpurascensat Monkey Bay, both halfway 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 161

TABLE 19 FREQUENCYOF MIMETIC ANDNONMIMETIC SONGS IN RELATIONTO THE BREEDINGCYCLE

Number Number nonmimetic mimetic Number Number vocalizations vocalizations Status Species birds minutes (mean___ •.95'•) (mean___

breeding 14.4 ñ1.4 4.5 ñ 1.8 Monthbefore V. purpurascenschalybeata 31 123 Breeding season alpha V. chalybeata 13 39 13.3 ñ 1.2 4.3 ñ 1.0 V. purpurascens 6 18 12.7 ñ 1.4 3.6 ñ 1.2

replacementsV. chalybeatapurpurascens 62 126 t 13.5ñ1.3 4.1ñ1.6 throughtheir prenuptialmolt, sang mimetic phrasesnearly as often as nonmlmetic ones. The absenceof seasonalchange in songmay be relatedto call-sitebehavior;even outside of the breedingseason the traditionaltrees are visited by malesand females,and particularlythe youngmay at this time frequent the tree to be usedas the matingsite. Inclusionof mimeticphrases in the songat all timesprobably increases the species-specificityof the birdsto a site,as it may resultin theirvisiting only a certainrestricted single call-site. The proportionof mimeticsong in the singingbouts of captivemales is considerablyhigher than in the wild birds. More thanhalf of all vocalizations in my captiveV. chalybeataand V. purpurascensrecorded in captivitywere mimetic, in contrastto a quarter of the songsof wild birds. The reasons for this difference are unknown.

BEHAVIORAL CONTEXT OF SONGS Changesin the proportionof mimeticand nonmimeticsongs were noted in the field from the beginningto the end of a periodof singing,though no quantitativedata were recorded.When the male flew to the call-siteand begansinging, most of his songswere nonmimetic,but near the end of a songbout more of the songswere mimetic. Often I had to wait for a few minutesof eachsong bout to passbefore I couldrecord a sequenceof mimetic vocalizations.Because the motivationof a singingmale is presumablymore agonlsticat the beginningof a songbout than at the end (whenhe stops singing,he flies from the call-site,indicating a waningof agonisticmotiva- tion), it seemslikely that mimetic and nonmimeticsongs may communicate informationabout the psychologicalstate of the male. At times when a male increasedhis agonisticbehavior (assumedthe call-site and chased 162 ORNITHOLOGICAL MONOGRAPHS NO. 11

TABLE 20 CORRELATION OF AGONISTIC AND COURTSHIP BEHAVIOR OF INDIGOBIRDS WITH THEIR VOCALLZATIONS

Number of times each song motif was given in several behavioral contexts

N onmime tic M imetic

Complex Context Chatter song P Supplanting attack 16 10 0 < .05 a chasesother male 24 53 0 < .001 a calls upon return 180 116 0 < .001 from a chase a calls and female 44 22 14 > .50 flies to call-site a hovers over female • 0 21 0 - a flies to ground when 0 0 68 < .00! female is on site

XTape recordings show 15 complex songs including 6 not noted in field observations; these are included. Field notes state 2 instances of mimicry, but tapes of both show that complex songs over- lapped with hover and the mimicry followed. other malesfrom it) he gave nonmimeticsongs, so thesenonmimetic vocal- izationsprobably signalan agonisticmeaning to the indigobirds. Malesoften gave mimetic songs when females visited them. Vocal mimicry wasnoted especially when a femaleflew from the call-siteafter an incomplete mating sequence.These observationssuggest that male indigobirdsdirect theirvocal mimicry mainly toward the mates. From my noteson behaviorand songduring continuous periods of observation of singingmales at their call-sitesand from behaviornotes and songs recordedon tape I tabulatedthe number of times mimetic and nonmimetic songscame with eachchange in the behaviorof the male. Both in male x male aggressivesituations and in courtshipthe frequencyof certainvocaliza- tions differedfrom a randomdistribution expected from the approximately 3:1 ratio of nonmimeticto mimeticphrases of the songbouts of wild males at the call-sites. The frequencieswith which vocalizationswere noted to occur in different behavioral contexts are recorded in Table 20, which includesonly the initial vocalizationgiven by a male in each context,e.g., the first vocalizationsung when an alpha male returnedto the call-siteafter chasinganother male (row 3). • Nonmimeticvocalizations were associated with aggressive behavior (supplanting attacks,males chasingmales, males returning to call-sitesafter a chase). One of the frequentlyrepeated song types given by one male chasing another is shown in Audiospectrograph40. In each of these agonistic contextsthe nonmimeticcall was given exclusively.The probabilitythat 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 163

thesenonmimetic vocalizations represent vocalizations drawn at randomby thebirds from the mixed repertoire of theirsong bouts is negligible(p < .05). Countersingingbetween two male V. chalybeataat neighboringcall-sites was heardat Marble Hall; the two malesregularly alternated songs and in this time were heard to give only nonmimeticsongs. Countersingingin birds generallyindicates territorial advertizement, that is, agonisticbehavior. Sup- plantingattacks were most often silent;in less than a fourth of all attacks seen did I hear nonmimeticvocalizations. If one regardsthe supplanting attacks as lower intensityagonistic situations than aerial chases,most of which were accompaniedby calling, the higher motivationalintensity of agonisticbehavior is accompaniedby an increasein nonmimeticvocalization. As the nonmimeticsongs were similar in their general pattern in the differentspecies of indigobirds,the uniformdispersion of breedingmales on call-sitesin areassuch as Merenskyand Zaria wheretwo or more kinds of indigobirdscoexist appears to restfitfrom the commonelements of nonmimeticsong shared between the species. On most occasionswhen femalesclearly appearedto fly to the call-site when the male gave a certainvocalization, the call givenwas a monotonic, prolongednonmimetic chatter. Chatteringmales often crouchedand leaned forwardwith the head feathers ruffled. Whenthey saw a femaleflying toward the call-site,they crouchedand chatteredin her direction. Most chatters did not immediatelybring in a female. As the chattersof differentspecies of indigobirdsare all similar,this call may attractfemales but it doesnot appearto be an isolatingmechanism among the differentindigobirds. Fe- males also sometimes chatter. Mimicrywas associated with courtshipthough not with the aerialhovering displayof the male. Mimicrywas givenboth regtfiady and exclusivelyby the malesas they flew to the grassand called,usually after courtingthe femaleon the call-site.Males mimicked continuously in the grassas they fed with the femalesby the call-site. Only when the male returnedto the call-sitetree did he againgive the harshnonmimetic songs. Males gave nonmimeticsongs in associationwith femaleswhen the femalesflew away from the siteand maleschased them. On no occasionwas mimicry given in an aggressivecontext with another male. The correlationof mimeticsong with courtshipbehavior is highlysignificant (p < .001, x2 contingencytest). Observationsof femalebehavior towards singing males in captivitywere not notablysuccessful. My captivemales sang less regularly than wild males, and my femalesgenerally did not breedeven when host nestswere available. On threeinstances I saw a captivefemale V. chalybeatafly to a male as he gave a mimetic song,and on each occasionthe male then hoveredbut the femaledid not solicitcopulation. One of the respondingfemales flew to a mimickingmale that wasin half non-breedingplumage, and this sug- 164 ORNITHOLOGICAL MONOGRAPHS NO. 11

TABLE 21 RESPONSESOF SINGING MALE INDIGOBIRDS TO RECORDED SONGSt

Responseof male indigobird Flewto or over playback Ceased No Playback song Chatter Silent singing effect Mimetic (firefinch species 0 1 3 3 mimicked by male) Mimetic (other speciesof 0 0 3 3 fire finch) Nonmimetic (conspecific) 6e 32 2 0 Nonmimetic (other species) 2 6 1 0 Pooled data of V. chalybeata, V. purpurascens, and V. Junerea. Dead conspecific male by recorder on one occasion. gestsan importantrole of mimeticsong (compared to the importanceof the maleplumage) in mateselection by thefemale indigobirds.

RESPONSES OF WILD MALES TO RECORDED SONGS Playbacksto maleindigobirds at their call-siteswere madeto determine whethermales responded differentially to the mimeticand nonmimeticsongs and to compareresponses to the songsof other species.Recordings were played at three-quartersof maximumplayback volume from concealment at distancesof 30-60 feet from the sites. No more than two playbackswere made to any one bird on a day, and when two were made the playback sessionswere at least 30 minutesapart. Table 21 summarizesthe response made by the male within the first five secondsof each playbacksession. The males usuallyflew directly toward the recorderwhen they heard the first nonmimeticsong. On fewer occasionsthe males becamesilent and peeredtowards the soundfor severalseconds before flying, and sometimes they chatteredand preened. Responseof a singingmale to a complexnonmimetic song of a species other than the subjectwas the same as that to a conspecific(usually the recordingof the subject). In both circumstancesthe male ceasedsinging and usuallyflew at the recorder,sometimes hopping down throughthe tree branchesand perchingquietly a foot or two from the recorder. When a deadmale was placed near the recorder,the call-sitemale attackedit whether or not the recordingwas broadcast. Playbacksof mimeticvocalizations did not elicit a consistentresponse. On one occasiona male V. purpurascensflew and hopped to the recorder, peeringat it. He then returnedto the perch and chattered. More often the malesceased singing and peeredat the sourceof the mimicry. This behavior 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 165 resemblesmales respondingto the calls given by firefinchessinging near the call-sites.Other malesseemed to ignorethe firefinches'calls. Mimetic playbacksof firefinchesother than the songmodel of the malesbeing tested evokedthe sameresponses as playbacksof mimicryof the songmodel. The inconsistencyof males in respondingto mimeticsongs supports the conclusion (p. 48) that vocal mimicry of the singingmales is not directedtowards the host firefinches.

RESPONSES OF CAPTIVE FEMALE INDIGOBIRDS TO SPECIES- AND POPULATION-SPECIFIC SONGS Althoughwild femaleindigobirds did not respondto recordedsongs that I broadcastin the field (becausethey habituallyused instead the established call-sites),captive females did respondin my aviarieswhen they had been isolatedfrom otherindigobirds and their firefinch hosts. To determinewhether the femalesdistinguish between the mimeticand nonmimeticsongs of their own speciesand otherindigobirds I testeda smallnumber of femaleswith recordedsongs. The behaviorof the aviary femalesprovides some experimental evidenceon the possiblefunction of mimetic song in the selection of a mate and a host by a female indigobird. In additionit was possibleto test the responsesof Vidua purpurascensto nonmimeticsongs of dialects of their own populationas well as of alien populationsof the same species. Indigobirdswere purchasedfrom a dealer who had recentlyreceived the birds from Rhodesia. In each shipmentof males and females, all males shared some nonmimeticsongs, and different shipmentshad males with differentsongs. The overlapof songsof the malesin each shipmentindicates that all birdsin a singleshipment were trappedor nettedfrom the samelocal neighborhoodin the field. Broadcastsof the recordedsongs of the males accompanyingthe femalestested thereforewere used to find the response of femalesto the dialect songsof their home population. Recordingsof other captivesand of wild birds taped in Rhodesiaand South Africa were used as samplesof songsof alien populationsof the same speciesin the playback experiments. Methods.--Individualfemales were taken from a groupof captivesthat had been kept on constantphotoperiod regimes mimicking the constant daylengthfound near the equator (LD 12:12). Indigobirdsmaintained in my laboratoriesfor more than two years on this regime have consistently remainedin breedingcondition (in breedingplumage and singing) for 10 monthsa year. All maleswere in breedingplumage at the time of the isolation of their females,and by so controllingthe reproductivecondition of the femalestheir responsivenessto the songswas presumablycomparable to that of wild femalesat the beginningof the breedingseason. None of the females was in molt at this time. Each female was isolated in an individual 166 ORNITHOLOGICAL MONOGRAPHS NO. 11 aviary or flight cage, usuallyout of sight and soundof any firefinch or indigobirdand alwaysisolated from the specieswhose songswere to be usedin playback. Three aviarieswere used. Aviary A was an indoorflight cage lighted by windowsand by supplementary,automatically timed fluo- rescentlighting which provided 12 hoursof lighteach day from 06: 00 to 18:00. AviaryB wasan outdooraviary measuring 15 x 20 feet and 10 feet high. Aviary C was anotheroutdoor aviary; it was out of sightof aviary B and at oneaviary I couldnot hear tape recordings played at theother. C measured 50 x 30 x 9 feet high. Each aviarywas provided with perchesarranged in a pattern symmetricalwith respect to the position of the tape recorders used in broadcasting.Birds were isolatedin each aviary for one to two weeks before any testswere made. The songsused in playbackwere copiedfrom indigobirdtape recordings madein the field or in captivity. Each tape had a five- or six-minuteintro- ductoryperiod of silencefollowed by alternating30-second bouts of song and 90-secondperiods of silence. Three pairs of tapeswere used. In pair 1 each tape had songsof one species(la = V. chalybeata,lb = V. purpurascens)copied from field recordings. The songson each tape were recorded with each 30-secondunit of sound for the playback comprising mimetic song, nonmimeticsong, or "mixed" songs (nonmimetic songsthat had one or more mimetic notes in them), in the following sequence: mimicry,mixed, mimicry, nonmimicry, mimicry, mixed. In pair 2 of the tapes the same songswere used and their sequencewas the same, except that the songson any one tape were of the mimetic songsof one speciesand the nonmimeticand mixed songs of the other. In the otherpair of tapes(3) the sources of songwere field- and captive-recordedV. purpurascens,V. chalybeata,and V. funerea. As in pair 1 of the tapes,the mimeticand nonmimeticsongs in pair 2 of tapeswere alternated,but mixed songswere omitted,and the sequence involvedfour boutsof mimicry and three boutsof nonmimeticsong on each tape. The mimeticsongs on each tape alternatedbetween V. purpurascens and V. funerea. The nonmimeticsongs alternated irregularly between own- dialectsongs and other-dialectsongs of V. purpurascensand the nonmimetic songsof V. chalybeata. Recordingsof the wild V. purpurascensand V. chalybeatawere copied from males taped at Penhalonga,Rhodesia, and the mimetic songsof V. funereawere from birds recordedat Tzaueen,Transvaal. The loudnessof all songswas the same on all tapes as indicatedby the vu-meter of the tape recorderwhen the playbacktapes were made, and in playbacksessions the volume and tone controlswere held in a constant setting. The positionsof tape recordersand tapes were alternatedto avoid training the birds to visit certain perchesand to avoid any possibleeffect of perch preferenceupon the resttits. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 167

In the teststwo tape recorders(Uher 4000-L) with paked tapeswere set at eitherend (in A and B) or side (C) of the aviary,playbacks were started at the sametime, and the responsesof the femaleswere observedfrom a blind. The songswere staggered in time with a 30-secondinterval of silence betweeneach of the 30-secondsong bouts which alternated back and forth betweenthe two tapes. Each female was testedno more frequentlythan once every two days. All testswere made between06:00 and 10:30 in the morning. The femalesoften respondedto the songplaybacks by looking at the recorderor by movingto it. The behaviorof femalesdirected towards the recordersduring each period of songor silencewas classified as no response, attentiveness,and approach,in the followingmanner. If a bird continued to dowhat it wasdoing at thebeginning of a test(feeding, perching, preening) and appearedto pay no attentionto the song,its behaviorwas calledno response.Approach was the mostcommon positive response to somesongs; herethe femaleflew to a perchor to the aviaryscreen within a few feet (in cageA within 12 inches)of the recorder,perched, and peeredquietly at the sourceof sound.This movement often involved a flightof 20 to 30 feet in the large aviary. Attentivenesswas the behaviorshown when a bird re- mainedon its perchbut lookeddirectly at the tape recorderin apparent responseto the broadcastsong. Often an attentivebird peeredfrom two or threeangles at the speaker.The responseswere pooledover all dayson whicha femalewas tested, regardless of whethershe responded at all posi- tively on each day. I carriedout two seriesof experimentsin the springand summerof 1970. Series1 involvedthe playbackof tapesof pairs1 or 2 to threebirds: a V. chalybeataamauropteryx in aviaryC from 27 April to 31 May (16 days of playbacks),a V. purpurascensin aviaryB from 7 to 30 May (9 days), and a V. purpurascensin flight cage A from 1 to 27 May (9 days). This lastbird did not respondat all on 24 or 27 May, and whenit wascaught on 31 May it wasabout 10 daysinto a heavybody molt. Playbackseries 2 involvedthe broadcastingof the pair 3 tapesto two birds,a V. purpurascens in aviaryA andanother V. purpurascensin aviary C, bothfrom 5 to 26 June (9 days). The samefemale was usedin both seriesin A; four femaleswere tested in all. Responseto mimetic songs.-•The responsesof captive femalesto the mimeticsongs of their own species(and host) are givenin Table 22. On nearlyhalf of all trials with mimeticsongs of their own speciesthe femalesimmediately approached a perch above the activespeaker. In only sixinstances of positiveresponse did the femalewait longer than 15 seconds to approach,and often sheapproached in lessthan 5 seconds.All females testedgave a dear responseboth in approachingand in becomingattentive 168 ORNITHOLOGICAL MONOGRAPHS NO. 11

TnBLE 22 RES?ONSES OF FEMALE INDIGOBraDS TO MIMETIC SONGS

Response Playback Source o/ series recording Approach Attentiveness No response own species 39 6 54 other species 3 3 91 silence 3 0 193 own species 29 10 33 other species 12 6 54 silence 2 0 142 x Number of song bouts of mimetic song played to each female: -/Y-B • 52, -/BG ----55, -/R ----89. Own and other species: V. chalybeata, V. purpurascens. a Number of song bouts of mimetic song played to each female: -/BG • 72, -/Y-R •--72. Own species= V. purpurascens, other species ---- V. ]unerea. to the soundof mimicryof their own species.Comparison of the number of positiveresponses and of no responsesof eachfemale to its own mimicry to that of the number of the control periodsof silenceshows that the responsivenessof females was significant (p < .05, x2 contingencytest). The femaleV. purpurascensignored the mimeticsongs of V. chalybeata and vice versa. Each female respondedsignificantly more often to the mimeticcalls of her own speciesthan to the mimicryof the other species (p < .05, x2 contingencytest). The selectivityof the femalesin responding to their own species'mimicry presumably is comparableto the behavioral selectionby wildfemale indigobirds of themales mimicking their host species, or of host firefinches. On the other hand, in series2 the two female V. purpurascenssometimes respondedto the mimeticcalls of V. /unereaas well as to the mimeticcalls of their own species.The femaleV. purpurascensresponded somewhat more often to the mimickedsounds of L. rhodopareia,their usual host, but their responsesto the soundsof L. rubricata(the songmodel of V. /unereain the playbackused) were as vigorousand rapid as to their own species. The responseto the L. rubricatamimetic motifs was considerablygreater thanthe responseof the femalesto motifsof L. senegala. Responseto nonrnirneticsongs.--Results of the responsesof femaleindigobirds to the noumimeticsongs of their own speciesand of other speciesare given in Table 23. Femalesflew somewhatless often to the nonmimefic songsthan to their conspecificmimetic songs. Although females responded on only 65 of the 197 nonmimeticplayback bouts (33 percent),the times whenthey did respondwere meaningful, especially compared with the lack of responseduring the controlperiods between song bouts and the playback of mimeticsongs of other species.No differenceis evidentin the responsivenessof femalesto the noumimeficsongs of their own speciesand of other 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 169

TABLE 23 RESPONSES OF FEMALE INDIGOBIRDS TO NONMIMETIC SONGS

Response Sourceof recordingx Approach Attentiveness No response own species,nonmimetic 23 (6 q- 17) 6 (1 q- 5) 70 (12 q- 58) own species,mixed 20 1 19 other species,nonmimetic 11 (6 + 5) 4 (1 q- 3) 21 (11 + 10) other species,mixed 13 2 22 Numbersin parenthesesgive the componentresults from series 1 and 2 respectively."Mixed" nonmimeticsongs, containing also a few mimetic syllables,were used only in series1. species(p > .50, x2 contingencytest). Theseresults suggest that nonmimetic songsof their own speciesand of otherspecies are aboutequally effective in attracting female indigobirds. The responseof femalesto mixed songswas similar to their responseto nonmimeticsongs, as femalessometimes approached speakers broadcasting the mixed songsbut lessregularly than they approachedthe mimeticsongs of their own species(Table 23). With the potentialspecies recognition signalsrepresented by the mimeticnotes a selectiveresponse of females to the mixedsongs of their own speciesmight have been expected,but no differencewas apparentin the frequencyor latencyof responsesto the mixed songs of own and other species. The series2 trials included a few mimetic songsof captive male V. pur- purascensreceived in the same shipmentsas the females tested. As the recordedsongs of thesemales differed from songsof malesin other shipments and from the songsof all indigobirdsrecorded in the field, the responses of the femalesto the songsof their shipment-matesare thoughtto showthe responsesof femalesto nonmimeticsongs of their own dialects. The two females tested in series 2 were from areas with different dialects. Table 24 comparesthe responsesof the femalesin series2 to their own song type dialects and to other dialects of V. purpurascens. The females approached

TABLE 24 RESPONSES OF FEMALE VIDUA PURPURASCENS TO NONMIMETIC SONGS OF HOME DIALECTS AND OTHER DIALECTS

Response Source of recording Approach Attentiveness No response

own dialect (captive males) 17 2 2O other dialect (captive males) 10 2 24 other dialect (wild males) 3 1 17 170 ORNITHOLOGICAL MONOGRAPHS NO. 11

TABLE 25 EFFECT OF PLAYBACKSEQUENCE ON THE RESPONSIVENESSOF FEMALE INDIGOBIRDS

Response Position o! trial Song in playback sequence Approach Attentiveness No response mimicry, own initial 15 0 19 subsequent 24 6 22 mixed, own initial 9 2 25 subsequent 11 1 19 both the songsrecorded from males from their own and from other song populations. No significantdifference in the frequency or latency of re- sponsewas evident. Although there was some tendencyfor the responseto other dialect songsto be lessfrequent in series2, the differencedisappears when the data from series 1 (Table 24, wild-recorded males from other dialectsystems) are also considered.The resultssuggest that femaleindigobirds respondno more stronglyto the nonmimeticsong types of males in their own song neighborhoodsthan to other dialect song types. Possibleeffect of song sequenceand location on the results.--To check for any possiblebias introducedby the playback designof using the same tapes repeatedlyin these experiments,I analyzed the results of female responsesfrom the series 1 tests to determinewhether femalesresponded to the nonmimeticand mixed songsmainly when these were played on the same side of the aviary as the mimetic songsof the female'sown species. It turned out that females approachedthe nonmimeticand mixed songsof their own specieswith the samefrequency regardless of whetherthe mimetic songsof their own specieswere played from the same or oppositeside of the aviary. The responseto the nonmimeticsongs of other indigobirdspecies similarlywas independentof the site of playbackof the mimeticsongs of the female subjectsin thesetrials. As songswere given in the samesequence in each playbackseries (the mimetic and mixed songsonly), I subsequentlycompared the number of times the female respondedto the first mimetic song bout with the number of times she respondedto the subsequentmimetic song bouts, to determine whetherthe sequenceof songsmay have affectedher responsiveness.The resultsof thiscomparison are givenin Table 25. On the average,the response to the first mimetic songwas nearly as often positive as was the responseto subsequentmimetic songs(44 percent versus58 percent responsiveness). Similarly, no noticeablechange in responsivenessto the mixed songsis evident (31 versus 39 percent). Probably longer sequencesof the same songtypes would result in a waningof the responsesof captivefemales in experimentalconditions, but in the playbacktrials used no changein 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 171 responsivenesswas evidentfrom one songbout to the secondor third song bout of the samekind of song,and for this reasonthe data for initial and subsequentresponses were pooled in Tables 22, 23, and 24. Discussion.--Femaleindigobirds respond strongly to the mimetic songs of their own species.In wild birds the responseto thesesongs may have two importantresults, depending on whetherthe singersare the maleindigo- birdsor the hosts. In the first instance,male indigobirdssing the mimetic songs,and the selectiveresponse of the aviary femalesmay show directly the role of mimetic song as a behavioralmechanism responsible for the assortafivemating of V. chalybeataand V. purpurascens.Second, the fire- finchesalso sing thesesongs, and the approachand attentivenessof female indigobirdsto the singingfirefinches in nature is probably involvedin the stimulationof ovarian developmentand in the finding of the host nests. Female indigobirdsmay be attractedto the vocalizafionsregardless of the identity of the singingbird, and wild females do approachboth singing male indigobirdsat the call-sitesand also their host firefinches. The responseof female V. purpurascensto the mimetic songsof V./unerea, as well as to its own species,suggests that femaleindigobirds do not use all of the potentialinformation about speciesdifferences in songin their selection of a mate (or host), but rather that they respondto certain more general featuresof song. The calls and songsof the firefinchesL. rhodopareiaand L. rubricata are similar in severalphrases, as describedearlier. The male indigobirdsmay imitate in perfect detail the minor differencesin song of these firefinches and may themselvesperceive the differencesfrom their songtutors in early life. But being able to parrot a difference(in males) and appreciatingthe differenceby respondingselectively to it (in females) are two distinct aspectsof behavior. Apparently only a portion of the species- specificinformation available in a song is actually used by the females in their own selectivebehavior. In my playback trials I recordedseveral different phrasesin each 30-secondsinging bout on the tapes, and hence it is not possibleto say to which phrasesthe femalesresponded, although I had the impressionthat the whistledslurs and the rapid warbles of L. rubricata were amongthe phrasesbroadcast at the momentthe female V. purpurascens flew to the speakers. These phrasesare especiallysimilar to some of the vocalizafionsof L. rhodopareia,the usual songmodel of this indigobird. The responsivenessof femaleindigobirds to playbackof songsis of interest in showingthat female birds may respond to songs,if the reproductive condition and the acoustic isolation of the birds are controlled and if each is kept in a large aviary in which it is free to perform many of its normal behaviorpatterns. Most otherstudies of the responsesof birdsto the recorded songsof their own specieshave been made with males. In their responsiveness to the songsof more than a singlespecies of firefinch the femaleindigobirds 172 ORNITHOLOGICAL MONOGRAPHS NO. 11 are similar to several other speciesin which the males may respond to electronicallyaltered songs of their own species,including alterations which removemany of the regularfeatures of a species'songs, resulting in broadcast versionsof song which no wild bird actually sings. European Robins ( Erithacus rubecula), Wrens ( Troglodytestroglodytes), Yellowhammers (Emberiza citrinella) and AmericanWhite-throated Sparrows (Zonotrichia albicollis) all respondto artificialsongs or to songsrecorded from real birds but alteredso they have only someof the acousticproperties of the songs of the wild birds (Falls, 1963; Bremond, 1968a, 1968b; Thielcke, 1970), and Ovenbirds (Seiurus aurocapillus),European Robins, and Indigo Buntings (Passerinacyanea) respondto broadcastsof the recordingsof wild birds in which the syllablesequence is alteredor the songsare playedbackwards (Falls, 1963; Bremond,1968a; Thompson, 1969). In the field a female V. purpurascensthat respondedto songsmimicking L. rubricata (as in series2 of the experimentabove) might possiblyend up mating with a male V. ]unerea and laying in the nest of L. rubricata. She might well also mate with a male of her own speciesthat mimicked the usual V. purpurascensfoster firefinch, L. rhodopareia. Perhaps, therefore, in areas where both of thesefirefinches and both hostsoccur nearby, and wherethe firefinchsongs are quite similar,behavioral and hencereproductive isolationbetween V. purpurascensand V. [unerea may break down or be nonexistent. Possiblefield examplesof this lack of isolation are discussed later (p. 282). The adaptivesignificance of the rather generalizedresponsiveness of the femalesto firefinchsong may be that firefinchesand their indigobirdmimics have differentsongs in differentlocalities, and hencea female that responded only to the local firefinch or indigobirdmimetic song dialectswould be at a disadvantagewhen it moved to another area. The more generalized responsemay permit dispersaland exploitationof firefincheswith slightly differentsongs in other areas,although a femalemight still be more strongly attractedto the firefinchsongs of her home area. The similar responseof female indigobirdsto the nonmimetic songsof their own speciesand to those of other indigobirdspecies suggests that localspecies differences that occurin the nonmimeticsongs have no important functionin the assortativemating of indigobirds.The resultsof the playback trials supportthe hypothesisthat mimeticcalls, but not nonmimeticcalls, are species-specificsignals involved in behavioralisolation between some kindsof indigobirdswhich live togetherand do not interbreed.The lack of a more selectiveresponse in femalesto the nonmimeticdialect songs of birds from their own home populationslikewise suggeststhe nonmimeticsong differencesmay be unimportantin maintaining any exclusivepopulation structurewithin a species.The analysis(p. 156) of the intergradingof the 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 173 nonmimeticsong types into adjacentneighborhoods indicated no suchbe- havioralisolation between the neighboringpopulations of a singlespecies of indigobirdanyway; rather the effectsof distanceand copyingmistakes accountfor the differencesin songbehavior of different populations. The similar but distinct nonmimeticsongs of males in different populationsof indigobirdsthus appear to have equal meaningfor a female: given similar mimetic songsof the males, a female away from her home neighborhood would be attractedto a male in anothersong neighborhood as stronglyas she would be in the site where she was raised, and a female in her home neighborhoodwould perhapsbe as stronglyattracted to a newcomer,once he was establishedon a call-site, as she would be to a male raised in her own neighborhood.Nottebohm (1970) has suggestedthat song dialects may be evolutionarypacemakers in promotingassortative mating in geographicallydistinct populations. Nottebohm's hypothesis may prove to be valid in birds which have in fact an exclusivepopulation structure. How- ever, in birds with intergradingneighborhoods with no active selectivebe- havioralexclusion of aliens,I wouldnot expectdiscrimination against males with differentsongs, such as the nonmimeticsongs of the indigobirds.

ASSORTATIVE MATING AND ISOLATING MECHANISMS

To find whether two or more sympatrickinds of indigobirdsbehave as distinct species,I watched their behavior at the call-sites to find whether onekind of femaleconsistently visited and matedwith a singlekind of male• that is, if they mated assortatively.In additionI have consideredhere the possibilitythat certain biological differences, such as habitat, breeding seasons, song,and visual appearance--amongthese indigobirds may be responsible for speciesisolation. These differences where they were foundmay be con- sideredpossible or potentialisolating mechanisms, and wherethey involve the differencesthat the indigobirdsseem to useas their own basesfor a choice of matesthey are regardedas effectiveisolating mechanisms.

ASSORTATIvE MATING The degreeof assortativemating in mixed populationsis a critical test of the presenceof behavioralisolation between two or more species. In earlierfield work and collectionsno distinctspecies differences between femaleindigobirds were known. In the absenceof known differences,White (1962, 1963a) suggestedthat indigobirdsare conspecificwith severalmale color morphsin an area. Traylor (1966: 60-61) pointedout that size as well as color differencesindicate that sympatricforms comprisedifferent species.The differencesin mimeticsong of each of theseforms provide additionalsuggestive evidence of differencesat the specieslevel. None of 174 ORNITHOLOGICAL MONOGRAPHS NO. 11

TABLE 26 ASSORTATIVE MATING IN THE INDIGOBIRDS

Number o//eraale visits to singingmale Locality Male (number [emales collected)

V. V. V. uniden- chalybeata purpurascens [unerea titled Merensky, Transvaal V. chalybeata 46 (6) 1 (1) - 9 V. purpurascens 3 (1) 37 (9) - 5 Sabi Valley, Rhodesia V. chalybeata 13 (7) 1 - 2 V. purpurascens0 17 •6) - 3 Penhalonga,Rhodesia V. chalybeata 3 - 0 0 0 V. purpurascens 0 5 (1) 0 0 V. funerea 0 0 12 (4) • 0 Monkey Bay, Malawi V. chalybeata 6 (4) 0 - 0 V. purpurascens 0 3 - 0 Sigor, Kenya V. chalybeata 3 0 - 0 V. purpurascens 0 11 (6) - 0 V. chalybeata (orange foot) V. wilsoni (gray foot) • Zaria, Nigeria V. chalybeata 7 (3) 0 0 V. wilsonF 0 17 (7) 0 x Three females were shot and lost; at least four females were involved. • Includes the forms "nigeriae," "camerunensis," and "wilsoni"; the females are indistinguishable. thesetraits alone conclusivelyestablishes the occurrenceof distinctspecies in the senseof Mayr (1963: 19-20), who regardedbiological species as interbreedingpopulations that are reproductivelyisolated from other populations. The mostconvincing evidence for separatesympatric breeding populations of indigobirdscomes from the differencesin morphologyof femaleswhich mate with the different forms of males. Observations of females at the call-sitesat the time of mating showedthat femalesmating with males of one form look similar to each other. These females in most areas were morphologicallydistinct from the femalesmating with other kinds of males; the femalesare describedin more detail in the sectionon geographicvariation and in the speciesaccounts. In areas where only a single kind of male indigobirdpredominated (Marble Hall, Tzaneen,Maun, Kisumu,Numan) the femaleswithin a population all resembledeach other in plumage, foot color, and bill color. For example,at Marble Hall all 67 femalesobserved to visit singingmale V. chalybeataamauropteryx had pink bills and pink feet, and at Maun where the only males were V. chalybeatasubsp. (described below) all 34 observationsof femaleswith foot and bill colornoted had whitishbills and pinkish feet, the same as the local males. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 175

Within localitieswhere two or more kinds of male indigobirdsoccurred, correspondinglydistinct females were recognizablein the field. Table 26 showsthe detailsof mate selectionby female indigobirds;each visit by a female to a call-site was recorded and several females were collected to verify the field identifications.At Merensky81 visitsby femaleswere made during conditionsfavorable for observationof foot and bill color. All but three visits (probablyinvolving one female) at call-sitesof male V. purpurascenswere made by white-billed,white-looted females, and all but two visitsof femalesto V. chalybeataamauropteryx sites were by pink-billed, pink-lootedfemales. Severalfemales, including the two identifiedon sight as mismatchesto their mates, were collected. The occurrenceof two distinct kindsof femaleindigobirds at Merenskyand the recordedassortative mating of thesefemales with malesof like appearanceprovides direct evidence that thesetwo indigobirdslocally behave as distinctbiological species. A similar high degree of assortativemating occursalso in other areas (Table 26) where two or more forms of indigobirdscoexist and where males (and their females) differ in bill color, foot color, or both. Red-billed V. chalybeataamauropteryx populations were most useful for observing assortativemating in the field as the femaleswere quite distinctivein their reddishbill color. Not all femaleindigobirds are morphologicallycharacter- ized. In Nigeria assortativemating separatingthe pale-footedV. wilsoni and the red-footed V. chalybeatawas evident. Females visiting the three forms of V. wilsoniwere morphologicallyindistinguishable from each other, and I regard theseforms as conspecific.The total observationsin Nigeria showed11 orange-footedfemales at call-sitesof V. chalybeata,whereas all females at the sites of male V. wilsoni of the forms "nigeriae" (2 visits), "camerunensis"(14), and "wilsoni" (2) had flesh-grayor whitish feet. The probability that data on mate selectionin wild indigobirdsindicate randommating between sympatric forms is lessthan 0.001 at Merenskyand SabiValley and is lessthan 0.05 at Zaria. In all of theseareas the indigobirdsbehave as goodspecies in their mate selectionas well as in the absence of morphologicalintergradation.

BREEDING BIOLOGY AND POTENTIAL ISOLATING MECHANISMS The nature of the isolatingmechanisms in the indigobirdshas not been studiedexperimentally except for a few of the songs,but informationfrom field observations,avicultural studies, and museumspecimens suggest which of the speciesdifferences may be responsiblefor the maintenanceof population structureand the limits of biologicalspecies in the indigobirds.Field observationsshowed that mate selectionis made by the femme in her choiceof a singingmale, and the role of behavior,especially the mimetic song,was studiedto determineits role in assortativemating. The occurrence 176 ORNITHOLOGICAL MONOGRAPHS NO. 11

,.• • ' . . ß ,;• -.• , •,,•

Figure 18. Grassy clearing in •r•chy•te•k• woodland; Palms ]i• the creek in the background. In the top of the lrec in foreground was a call-site of •Mu• fi•nere• codr#lglo•d; •go•lo•licla rllbriciilii safig ifi trig grass. T•e .site is near Zomba, Mala•L of assortativemating in mixed populations strongly suggestsa behavioral basis for speciesisolation. Other possibleisolating mechanismswere considered: do different indigobirds live together and have an opportunity to interbreed; do they breed at the same time of year; and does the host itself perhapsselect against hybrids or unusual parasitesin the nest? The answersto the questionshelp to characterizethe indigobirdsas biological speciesand also provide circumstantialevidence of the nature of the isolating mechanisms between them. Habitat differences.-•Each indigobird occurs with its host species of firefinch, and where the firefinchesare distributedallopatrically the indigobirds cannot interbreed. Within a locality some habitat segregationmay occur; throughout most of Africa Lagonosticta senegala and its parasite Vidtta chab'beatalive in towns and villagesand breed in the houses,whereas other species,such as L. rhodopareiaand V. purpttrascens,are bush birds generally breeding only away from human populations. At Penhalonga and south of Zomba I found V. Junerea codringtoni call-sites only along the banks of streams (Figure 18) while V. purpurascenscall-sites were on higher, drier groundparalleling the differencesin microhabitatof their hosts L. rubricata and L. rhodopareia. Some call-sitesof these two indigobirds 1972 PAYNE: PA•SITIC INDIGOBIRDS OF AFRICA 177

Figure 19. Grassy Acacia woodland one mile south of Monkey Bay, Malawi, the habitat of Lagonostictasene•,,ala, L. rhodopareia, V. chalybeata,and Vidua purpurascens. were within 500 feet of each other. In a few instances I saw a male of one speciestake over a call-site from another species. In the breeding season I have seenindividual female indigobirdsfly more than 1,500 feet from a call-site to a distant bush, so these minor differences in call-site location clearly do not effectivelyisolate the indigobirds.In other placeswhere two or more firefinchesoccur in the samehabitat, as at Merensky Reserveand Zaria, no differenceswere apparent in dispersionof the indigobirds. Both V. chalybeataand V. purpurascensperched in a singleacacia tree at Monkey Bay as did both local speciesof firefinches(Figure 19). Different species of indigobirdsthus do occurin the samelocal area, providingthe opportunity for interbreeding,although such interbreeding seldom occurs. Breedingseasons.•Possible temporal isolation of indigobirdswas studied by collectingbirds and examiningtheir reproductiveorgans and also by comparingthe time of year in which males of differentforms are in breeding plumage. In all areaswhere I saw males in breedingplumage, female indigobirds were laying eggs. Female V. chalybeata,V. purpt,'ascens,and V. [treetea all beganlaying in Januaryin the Transvaal;by February all three species were laying in easternRhodesia. V. chalybeataand V. purpurascensboth bred in Kenya in May and June. In Nigeria the femalesof V. chalyheata and of the "camerunensis"and "wilsoni"forms of V. wilsoniwere laying 178 ORNITHOLOGICAL MONOGRAPHS NO. 11 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 179

in July; the ovariesof somefemales of all forms of Zaria indigobirdshad somerecently ovulated follicles in the ovary, somefemales had an egg in the oviduct,and all kinds of indigobirdswere seento copulate. Severalmale indigobirdswere collectedin prenuptial molt. Testes are small at the beginningof molt and enlargewhile the glossyblack breeding plumagedelevops (Table 27). The testesare in breedingcondition (about 5 x 3 mm) when only a few sparrowyfeathers remain and the breeding plumageis almostcompletely grown. All males in breedingplumage had large testes.Large testesin a male V. chaIybeatain very worn breeding plumagethat I collectedon 21 June 1967 in Transvaal indicate that the testesremain large until the postnuptialmolt. Sincemales have large gonads throughoutthe time they are in breedingplumage, the breedingseasons of differentindigobirds may be comparedwith reasonfrom the presenceof males in breedingplumage in a population. The timingof the breedingseasons is evidentin museumspedmens from the seasonaldistribution of maleindigobirds in breedingplumage (Figure 20). Breedingis highlyseasonal in southernAfrica southof 24 ø lat.; all but one museumspedmen in breedingplumage taken there were collectedbetween November and May. The breedingseason becomes progressively longer towardscentral Africa mainly by the continuationof breedinginto later months. Within ten degreesof the equatorbreeding is mainly in the first eight monthsof the year, but a few males have been taken in breeding plumageat othertimes as well. Within two degreesof the equatorbreeding malesmay occurin all months. In northernAfrica breedingis mainly from July to December,six monthsout of phasewith southernbirds, thoughsome malesbreed into Januaryand February. The latitudinal trends parallel those of other African seed-eatingbirds whichbreed with or after the rains (Moreau, 1950: 252). Seed-eatingbirds raise their young both on insectsand on fresh, undried grassseeds which are most abundantafter the rains have sprouteda new growth of grassand the new grasshas seeded.However, it is unlikelythat rainsthemselves cause gonadalactivity and molt by controllingdirectly the physiologyof indigobirds. At Merenskyrains were late in the summerof 1966-67 and did not fall until the last ten days of 1966; by this time the indigobirdswere well into prenuptial molt and severalmale V. chalybeataand V. purpurascens werein full breedingplumage. On the otherhand, at Sigorgrass was high and lush (tsetse flies and local custom restrict cattle from the area) from recentApril and May rains which the local peopleregarded as unusually heavy,but half of the male indigobirdstaken had not completedthe prenuptialmolt and four of the six femaleshad not laid by the first of Junein 1967. The absenceof a closecorrelation in the field betweenthe beginning 180 ORNITHOLOGICAL MONOGRAPHS NO. 11

V.chal. ybeata L. $enegala

+ 12øN I I I SEN

6-12 I ! NIG 0-6 0-6S 4, 4,.I. - - v• KEN, UGA TAN ,• 6-12 ZAM, MAL 12-18 RHO + 18øS S. AFR V.purpura$cen$ L. rhodopareia 0-2 ø N

0-6øS , TAN 6-12 • ZAM, MAL 12-18 ] RHO +18øS ,m S. AFR •

V. runetea L. rubrico fa 2-6øS • , KEN 6-12 rm TAN , , S.E.CO•l 12-18 • • ZAM, MAL • RHO, MOZ + 18øS ' ' S. AFR nl.qerlae L. rubricata 2-10ON , , r• NIG

L. larvata 2-14øN I , , W. AFR •w//$on/" L. rata 2-14øN , , SUD, NIG

J FMAMJJASOND JFMAMJ JASOND Figure 20. Breeding seasonsof Vidua indigoblrds and Lagonosticta firefinches in Africa. The height of the figures for Vidua indicates 1, 2, or 3-]- male specimensin breeding plumage; - • one or more birds in postnuptial molt, q- -• one or more birds in prenuptial molt. Figures for Lagonosticta record one or more nesting records in the month. Nigerian L. senegalasources: Fry (1965), Smith (1966), Mackworth-Praed and Grant (1960). Sen • Senegal, Nig • Nigeria, Ken • Kenya, Uga -- Uganda, Tan • Tanzania, Zam = Zambia, Mal -• Malawi, Rho • Rhodesia, S. Afr • South Africa, S.E. Con ----southeast Congo, Moz --• Mozambique, W. Afr • westernAfrica. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 181 of rains and the molt and breedingof the indigobirdssuggests a response to conditions other than rainfall. Local observationsagree with the breedingseasons as determinedfrom specimens.At Nairobi, male V. chalybeatain breedingplumage have been takenin all monthsbut February,September, and December,and probably in every month somemales are breeding. In northernNigeria indigobirds (V. chalybeata)have been seen in breedingplumage in all monthsfrom July to March (Fry, 1965; Smith, 1966; N.J. Skinner,pers. comm.). At Mererisky my field observationsshowed some birds in breeding plumage by late Decemberand a few malesin worn breedingplumage until late June. A higherproportion of malesin sparrowyplumage were taken at equatorial latitudesthan at higher latitudesin my field work. In South Africa all of the sparrowyindigobirds taken in the breeding seasonwere females; all males taken from Januaryto July were in completebreeding plumage. In Rhodesia,Botswana, and Malawi of the 81 non-moltingmales only 4 (2.5 percent) were in partial or complete sparrowy plumage, while in Kenya and Nigeria 53 non-moltingmales included 7 (13 percent) in partial or completesparrowy plumage. In RhodesiaSmithers, Irwin, and Patterson (1960) have notedsome males beginning the postnuptialmolt while others remain in breedingcondition. Males are most out of phasewith each other duringthe more prolongedbreeding season near the equator. Each firefinch breeds at the same time at its indigobird song mimic (Figure 20), and within an area the different firefinch specieshave similar breedingseasons. Nesting records of firefinchesfor Figure 20 were taken from several sources: South African OrnithologicalSociety nest record cardsthrough June, 1966, and McLachlanand Liversidge(1957), for South Africa; RhodesianOrnithological Society nest record cards through February 1967 for Rhodesia;Brooke (1968) for Mozambique;Benson, Brooke, and Vernon (1964) and Benson and Irwin (1968) for Zambia and Malawi; Mackworth-Praedand Grant (1960) for Congo,Tanzania, Kenya, Uganda, Sudan,and Nigeria; van Someren(1916, 1956) for Kenya and Uganda; and Serle (1940, 1943, 1957), Bannerman (1953), Fry (1965), Smith (1966), andMorel andMorel (1962) for westAfrica. Speciesdifferences in the breedingseasons of firefinchesparallel their habitat requirements,inasmuch as L. senegalain southernAfrica breeds well into the dry season,whereas the moremesic-dwelling L. rubricatastops breedingwhen the dry seasonapproaches (Benson, 1963:631; Bensonet al., 1964: 102). In RhodesiaL. rhodopareiahas a long breedingseason with nests recorded in all months; L. senegala nests in all months but October and November,when it molts. However, the indigobirdsdo not have such a prolongedbreeding season. In equatorial east Africa firefinches breed in all months as do their viduine parasites. West African firefinches also 182 ORNITHOLOGICAL MONOGRAPHS NO. 11 show great speciesoverlap in breedingseasons within an area, thoughL. senegalanests later in the north than in the south. At Zaria, Nigeria, the breedingseasons of the local host firefinchesoverlap; in July and August, 1968, I saw active nestsand fledgedyoung of L. senegala,courtship behavior (straw display) of L. rata and L. larvata, and a female L. larvata flying with her tail cockedover her back--presumablyshe was incubating, as I have seen this posturein captive female L. senegalaflying from the nestin the incubationperiod. In Senegalthe time of breedingof L. senegalais notably constantfrom year to year in spiteof fluctuationsin rains and food supply (Morel, 1969: 27). As in the indigobirds,environmental conditions such as photoperiod or other predictablefeatures probably are more important in control of the timing of breeding. The geographicalvariations in the breedingseasons of the indigobirds and of their firefinch hostsgo hand in hand, and within any one area all hosts and parasiteshave greatly overlappingtimes of breeding. Thus no differencesin breedingseason of indigobirdsnor of the firefinch hostsavail- able to them function as temporalisolating mechanisms. Courtshipbehavior, colors, and song.--The assortativemating of indigobirds at the call-sitessuggests reproductive isolation as a result of differences in courtshipbehavior of the species.The role of mate selectionclearly lies with the female as male indigobirdsare undiscriminatingin their courtship displaystowards the femalesas they are also in agonisticbehavior towards othermales. Pinto and Lamm (1960:118) likewisehave reportedtwo male indigobirds(perhaps males of two different species;this was uncertain-- D. W. Lamm, pers. comm.) to court the sameindividual female in succession in Mozambique. Femalesevidently select their males on the basisof some differencesin the communicationsignals of the males. No speciesdifferences are evident in the visual signalsor their sequencein courtshipdisplay of the males. Nor are any species-characteristicdifferences apparent in the nonmimeticvocalizations, as the variationsof the nonmimeticsong types are local ones,not species-wideones. Only the species-specificvocal mimicry containsany consistent,characteristic species signals. Coexisting,noninterbreeding indigobirds also differ in other potential signals, however--the plumagecolors and bill and foot colorsof the males. Signals used by birds in mate selection within the indigobird species complex are more likely to be the speciesdifferences in vocal mimicry, for several reasons. First, the femalesin the field usually fly directly to a call-site from a distance of a few hundred feet. This behavior indicates mate selection based upona signalthat is receivedover a long distance.Plumage color of breeding malesis not readily apparentto the humaneye at severalhundred feet 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 183 evenwith binoculars,but mimeticsong can be heardby man at a distance of at least 300 feet. Secondly,plumage differences of malesare not consistentspecies-specific differences(see frontispiece).In some areasthe male breedingplumage in two sympatricspecies are identicalto the humaneye and to the spectro- photometer;no distinctspecies differences occur in plumagecolor of male V. purpurascensand V. f. funereain Transvaal,nor of V. purpurascensand V. chalybeatacentralis in northwesternKenya. Certain combinationsof bill and foot color distinguishthe three speciesof indigobirdsin SouthAfrica and Rhodesia,but outsideof the range of V. c. amauropteryxmale indigobirdsof all specieshave white bills and all but V. chalybeatamay havewhitish feet. In areassuch as northern Nigeria the threecoexisting forms of V. wilsoni (green, blue, and purple) all have white bills. Foot color in these forms is similar,although it is distinctlymore purplishin the blue-leatheredmales than in purplemales taken at Zaria. A further argumentagainst color as a speciesrecognition signal is the considerablegeographic variation in male breedingcolor in any one species.For example,V. chalybeatahas green, blue, and purplish-bluemales acrossNorth Africa, and since theseforms broadlyintergrade and all mimic the same host species,the plumagedif- ferencesclearly are not speciesisolating mechanisms. The combinationsof foot color and bill colorin all speciesknown to live togetherwithout interbreeding are locally distinct. For example,red-billed, red-looted V. chalybeata,white-billed, red-looted V. [unerea, and white- billed, white-footedV. purpurascensin Transvaal,in easternRhodesia, and in southern Malawi all differ in combinations of bill and foot color. Where the forms V. purpurascensand V. •. nigerrimado intergradein northern Malawi both have the samebill and foot colors. Possiblythe contrasting colorsof the bills and feet of somenon-interbreeding species may be per- ceivedas speciescharacters by the femalesand divergencehas been selected for in regionswhere selection against interbreeding is important,but I think the mimeticsongs are moreimportant than these colors in promotingassortative mating. Foot colordiffers in geographicallycomplementary populations of V. [unerea,though morphological evidence suggests interbreeding between theseforms (especially between nigerrirna and codringtoni). No experimental studieshave been carried out on the importanceof foot color (or bill color) as speciessignals in the indigobirds. A natural experiment was provided by a few mating visits by female indigobkdsto males of a different species,some of which sang a song unlike othersof their appearance(Tables 8 and 10). The one male indigobird in southernAfrica that I heard sing the "wrong" song (a male V. chalybeatathat mimicked L. rhodopareiainstead of L. senegala) had one female visit its call-site while it was there. The female was identified on 184 ORNITHOLOGICAL MONOGRAPHS NO. 11 sight as a white-billedV. purpurascens(a speciesthat normallymimics L. rhodopareia) and was at once collected (RBP 4323); in the hand she wasclearly white-billed and white-looted. Evidently the femalewas attracted to the male by his songrather than by his plumagecolor or red bill and red foot color. One wouldexpect an equalnumber of malesand femalesto be imprinted to the "wrong"song; such a femalewould visit a male of anotherspecies singingthe songstypical of other males of his appearance.Presumably imprintedon the wrong songwas one female V. chalybeataat Merensky which visited a V. purpurascenscall-site three times on 18 and 19 March 1966,and copulated with themale purpurascens. On morphologicalgrounds thisfemale (RBP 3995, seep. 27) is V. chalybeataamauropteryx. She had bright pink bill and feet, and she was laying. Mate selectionby the femalethus seemsto be dictatedmore by the song of a male than by his appearance.These observationsprovide the most direct field evidenceavailable of vocal mimicry as a behavioralmeans of maintainingthe assortativemating of sympatricindigobird species. Selectionby the host: its effect on speciesisolation of the indigobirds.- The fosterfirefinches may havean ultimateeffect upon reproductive isolation of differentkinds of indigobirdsby failingto rear the youngindigobirds with mouth colorsunlike thoseof their own young. Nicolai (1970: 929) found that youngindigobirds in the nestsof L. senegalaand of L. rhodopareiaeach had distinctlydifferent host-mimetic mouth colors. A youngV. chalybeata hatchedin the nestof L. rhodopareia,for example,might then be lesslikely to fledge than one in the nest of the usual host species,L. senegala. The observationthat there are very few indigobirdsof this specieswhich mimic firefinchesother than L. senegalasuggests that discriminationof young nestlingsby the fosterparents may in part accountfor the lack of interbreeding of V. chalybeatawith other indigobirdsin thesesame areas. The foster parentsmay alsoselect against hybrid young in the sameway, as the mouth colorsof hybridswould probablydiffer from those of the indigobirdsthat matchedtheir foster species.The beggingcalls of the young of different kinds of indigobirds,also, might be discriminatedby the adult fosterers, thoughthe beggingcalls of differentspecies of firefinchesand of their indigobird adult male mimics are only slightlydifferent from each other (Audio- spectrographs4, 5, 20). At any rate, an indigobirdfemale would only rarely lay in the nest of a speciesother than her normal host, so any behavioral discriminationby the foster parent would be too infrequent to accountfor the behavioralisolation of the indigobirdsthat was observedat their call-sites. For this reason, the suggesteddiscrimination by the firefinch hosts is not regardedas not likely to be importantin the maintenanceof indigobirdspecies limits at the presenttime, althoughthe behaviorof the hostsmay have been 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 185 importantin the evolutionof species-specificbrood parasitism and mimicry in the viduine finches. Resultsof interbreeding.--Thedifferent species of viduinesare not known to havegenetic barriers that mightprevent development of hybridizationand intergradation.Hybrid viduineshave been seenand collectedin the field (Priest,1936: 360, 364) and havebeen imported into aviculturalcollections from birdspresumably caught in the wild (Roberts,1926; Yamashina,1930; Abrahams, 1939; Yealland, 1959; Everitt, 1959; Harrison, 1963b; Alston, in Friedmann, 1960; and Strachan, in Winterbottom, 1967). These male hybrids,some of them namedas new speciesand genera,are blackishwith long central rectricesintermediate in length betweenthe short tails of the indigobirdsand the long tails of the other viduines. These have been con- sideredprobably hybrids between various species of indigobirdson the one handand Vidua paradisaeaor V. macrouraon the other;Abrahams supposed his bird to be a hybrid of V. paradisaeax V. regia. In SouthAfrica, W. D. Becker has bred hybrid viduineswhich are very similar to these all-black, long-tailedpresumptive hybrids from two different crosses,V. regia x V. chalybeataamauropteryx and V. regia x V. "funerea" (Winterbottom, 1965). Thesewere bred in an aviary with a pair of waxbills (Estrilda astrild) as the fosterparents. The songsof the hybridswere not noted. Winterbottom (1965) hasincluded a photographof the "purplewidow-bird" hybrid. As different viduine subgenera(Hypochera X Vidua) have been shown to produceviable offspring,the geneticsimilarities of all viduinessuggest that there is little interspecificgenetic incompatibility, though none of these hybridviduines have been bred to determinewhether they are reproductively fertile. The degreeof intergradationin morphologybetween two or more kinds of indigobirds(in the V. wilsoni complexor in the purpurascens-nigerrima complex, for example) suggeststhat some hybrid indigobirdsthemselves would be reproductivelyfertile. This matterslittle, as matingis assortative. In summary,female indigobirdsliving in areas where two or more kinds of malesoccur are highly assortativein their mating behavior in visiting the call-sitesof only a singlekind of male. Behaviorallythe forms V. chalybeata amauropteryx and V. purpurascensmated assortativelyand behaved as thoughspecifically distinct in Transvaal. Similarly some assortativemating was observedin V. chalybeataand V. wilsoniin Nigeria, in the forms V. funerea codringtoni,V. purpurascens,and V. chalybeataamauropteryx in Rhodesia,and in V. chalybeataand V. purpurascensin Kenya. The behavioral basisof mate selectionis most likely the femaleresponse to mimeticsong. Both field observationsof a female visit to the call-siteof a male singing the "wrong" mimetic songand aviary playbacksof tape recordedmimetic songindicate that female indigobirdsare attractedto songsthat mimic her fosterfirefinch species. Hearing thesesame songs is thoughtalso to stimulate 186 ORNITHOLOGICAL MONOGRAPHS NO. 11 ovarian developmentin the female viduines. Comparisonof the breeding seasons,habitats, plumage colors, and colorsof the bill and feet of different indigobirdssuggests that none of theseare as importantas mimetic songin maintainingthe reproductiveisolation within the indigobirdspecies complex.

DISTRIBUTION OF FIREFINCHES AND INDIGOBIRDS The speciesspecificity in indigobirdsong mimicry suggeststhat there should be close correspondencein the geographicalranges of these brood parasitesand the firefinch speciesthat foster them, both over large areas and within single localities. The geographicaldistributions of these birds were determinedprimarily from the localitiesof museumspecimens examined and secondarilyfrom the field observationsand publicationscited below. Localitiesof indigobirdsand firefinchesare listedfor each countryin Africa in AppendixB, and all but about 50 of theselocalities were preciselylocated or located to within 10 miles from various geographicsources [maps; gazetteers-•includingan unpublishedgazetteer of African bird localitiesby B. P. Hall; atlases, collectors'journals (Alexander, 1907; Bates, 1924; Erlanger mapsin BM[NH]; Grote, 1928; Neave, 1907, 1910; Adolph Fried- rich, Duke of Mecklenberg,1909); and correspondencewith severalcollectors and local residents]. The ranges based upon these definite localities are comparedin the maps of Figures 21-30. Distributionmaps of firefinches have alsobeen prepared by lmmelmannet al. (1965), and Hall and Moreau (1970) showmaps for both firefinchesand indigobirds;however, their maps are basedlargely on identificationsmade by othersand they do not specify the localitieswhere the birds were collected. Gazetteersfor all specimenand observationallocalities have beendeposited with the British Museum (Natural History), the Field Museumof Natural History, the National Museum of Rhodesia,and the Universityof MichiganMuseum of Zoology;these may be consultedfor the latitude and longitudeof each locality listed in Appendix B whichwas definitelylocated.

SOUTHERN AFRICA South A[rica.-•Although for many years the indigobirdsin South Africa were regardedas two species,field observationsin easternTransvaal showed that three distinct speciesoccur; two had been masqueradingunder the name of Vidua lunerea (Payne, 1968a). Roberts (1924: 188, 1940: 363) recognizedthe differencesin foot color in these birds, but his taxonomic treatmentsin generalhave been regardedas extremesplitting, and his species "Hypocheralunerea" and "Hypocherapurpurascens" were lumped by Mc- Lachlan and Liversidge (1957: 458) and Mackworth-Praed and Grant (1963: 665). Orange-footedindigobird mimics of Lagonostictarubricata are knownonly from the moistescarpment region where that firefinchoccurs, 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 187 whereaswhite-footed mimics of L. rhodopareiaoccur with this speciesin the drier lowveld and in the northern Transvaal bushveld. The distributional pattern show V. lunerea and L. rubricata in the Moist Subhumidand Humid moistureregions whereas the localitiesof V. purpurascensand L. rhodopareia lie mainlywithin the Dry Subhumidand more mesicparts of the Semi-arid moistureregions. In northern Zululand mean annual rainfall is similar to that of southernNatal (600-1,000 mm per year) but the low relief and greaterevapotranspiration in the north (Ady, 1965: 59) result in relative dryness,and both L. rhodopareiaand L. rubricata occur in Zululand. As no differencesin measurementsor plumagecolor in orange-footedV. [unerea and white-footedV. purpurascensare known, museumspecimens from north of the Tugela River in Zululand cannot be identified with certainty as V. purpurascens.At Hluhluwe I saw and photographedin color a male V. [unerea,and it had red-orangefeet and mimickedL. rubricata. In southern Natal and the easternCape Provincethe only commonfirefinch is L. rubricata and the only indigobirdmuseum specimens from these regions (excluding "Port Natal" birds, obviouslycage birds) are V. Iunerea; all indigobird specimensthat had the foot color recorded were red-footed. L. senegala and V. chalybeataare widespreadin the Transvaal; both are also known from northern Zululand at Ndumu. A sight record of an indigobird at AughrabiesFalls (Winterbottom, 1968: 254) probablywas of this species, perhapsof the white-billedform found in South-WestAfrica and Botswana. Both L. rubricata and V. lunerea have disappearedlocally in recent years in the easternCape Provincedue to lossof tall grasshabitat by overgrazing of domesticanimals. This firefinchwas knownformerly near SomersetEast, but V. junerea was uncommon there (James, 1925: 636); we saw neither there in 1965. Hewitt (1931: 73) regarded both as common except in overgrazedareas of the Transkei region. In severalweeks in the breeding seasonin 1965 and 1966 at Amanzi estateI found no L. rubricataalthough it had been reported "quite common" by Niven and Niven (1966: 83); herds of goats were destroyingthe grasshabitat and Mrs. Niven informed me that firefincheshad becomescarce in recentyears. V. Iunereais reported as "occasional,sporadic" and Mrs. Niven noted that indigobirdshad been absentfor the past few years. Skead (1964: 68, 1967: 89) noted that L. rubricatawas irregularin distributionin the easternCape, and G. Ranger has informed me that at Kei Road both finches were common in the 1920's but disappearedwith an increaseof sheepgrazing. South-WestAIrica.--Few firefinchesand indigobirdshave been taken in this dry country. L. senegalais most common,having been taken mainly in Ovambolandand in the Caprivi Strip, whereasL. rhodopareiais known only from the CuneneRiver. The only known indigobirdspecimen is a bluishV. chalybeatafrom Ongonga,Ovampo, a localityof L. senegala.J.M. 188 ORNITHOLOGICAL MONOGRAPHS NO. 11

L.$enegala

Figure 21. Distribution of Lagonosticta senegala.

Winterbottom (pers. commß) has birded extensivelyin South-WestAfrica for years without seeingan indigobird. Presumablythe aridity of the Kalahari and Namib deserts restricts the finches. Swaziland.--South of the area of overlap of V. purpurascensand V. runetea in Transvaal is a second area of overlap in Swaziland; these two indigobirdsoccur along the IngwavumaRiver. The firefinch L. rubricata is knownfrom the samelocality, while L. rhodopareiahas been taken in the next river systema few miles to the north. L. senegalaappears to be widespread,but no V. chalybeataare known. Botswana.-•In savanna and dry bushy country in northern and eastern Botswanaand near the Okavangoswamp firefinchesand indigobirdsoccur. L. rhodopareiaappears to be uncommonbut widespreadon the Chobe River, throughoutthe fringesof the Okavangoto Lake Ngami, and in the eastern corner of the country, whereasL. senegalais more widespread,occurring throughoutexcept in the Kalahari. V. chalybeatais common around the Okavango and down the Botletle River to Lake Dow as well as along the Chobe and Shashi rivers in the east. A singleV. purpurascensis known from Francistowndistrict. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 189

Figure 22. Distributionof Vidua chalybeata(excluding problematical Congo specimens from Kasai and Tanganika).

Neither firefinchesnor indigobirdshave been reportedfrom isolated villagesin the dry countryaway from the Ngamiwoodlands or the rivers. In his check-list of the birds of Bechuanaland, Smithers (1964: 156) recordedonly "Vidua funerea"from the country,but he informsme (letter, 19 May 1967) that this namewas used as the indigobirdsof southernAfrica were tentativelyall then regardedas conspecific,and within this area the form funereahad been describedearlier than the othersand was then thought to apply to all. Rhodesia. Acrossthe Rhodesianplateau and in the drier river valleys L. senegalaand L. rhodopareiaare widespreadas are V. chalybeataand V. purpurascens.In contrastL. rubricataoccurs only in humidmountainous areasof the easternedge of the country,and only in theseareas are there white-billed,red-looted indigobirds, V. funerea. The indigobirdsare relatively well known in Rhodesia both from field observationsand song recordingsdescribed here and from the activecollecting of the National Museum of Rhodesia. However, even in this area the host-parasiterelations are moreclearly seen from field observationsthan from the localdistributions of indigobirdsand firefinches.L. rubricatahas been collectedat all three 190 ORNITHOLOGICAL MONOGRAPHS NO. 11

L. rhodopare/a ' .. .•.L. • •iiii!i

Figure 23. Distribution of Lagonosticta rhodopareia. localitiesof V. funerea, L. rhodopareiahas been taken or observedby me at 11 of the 31 localitiesof V. purpurascens,and L. senegalaspecimens or observations are known for 17 of the 40 localities of V. chalybeata. Thus for fewer than half of the localitiesof the indigobirdsare there recordsalso for the known host firefinch. Each of the indigobirdswas taken more often with its known host firefinch than with any other firefinch, however, and the trend for specimensto document the concordancein distribution of parasite and host is evident. Mozambique.--Three host speciesof Lagonostictaoccur in Mozambique, but only two forms of indigobirdsare known. In all five localitieswhere V. chalybeata was taken with at least one firefinch, one of the firefinches was L. senegala;at four of these same localitiesL. rhodopareia also was taken. At one locality V. purpurascenswas also taken with L. rhodopareia. At Gorongozaat 1,000 feet elevation and at Zumbiti purplish indigobirds without data on foot color (presumablythese are V. purpurascens)were taken with or near L. rubricata; L. rhodopareia is not known from these mixed mesicareas (GorongozaMountain is "undifferentiatedmontane com- munity" and the Beira region is met by both Brachystegiawoodland and coastalforest-savanna mosaic; Keay, 1959). Malawi.--Three specieseach of host firefinchesand of indigobirdsare known for Malawi. Correspondenceof the distributionsis evident both 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 191

V.purpurascens

Figure 24. Distribution of Vidua purpurascens. withinlocalities where host and parasitehave been taken together and within the altitudinalranges of thefinches in thishilly country. V. chalybeatahas been taken at six localities with L. senegala,V. purpurascenswas collectedmost often with L. rhodopareia,and V. funerea was taken more with L. rubricatathan with other firefinches;all three indigobirds were thustaken mostoften with the firefinch known from my recordingsto be their songmodel in Malawi. In southernMalawi in the Nsanje (= Pt. Herald) District, R. C. Long collectedfor years in the roesichills by Chididi Mission and along the drier ShireValley. Table 28 indicatesthe frequencydistribution of specimens

TAa•.E 28 ALTITUDINAL DISTRIBUTION OF INDIGOBIRDS AND FIREFINCHES IN SOUTHERN MALAWI

Number of localitieswhere specimenswere taken Altitude V. ch. V. V. L. L. L. (feet) amauropteryx purpurascens funerea senegalarhodopareia rubricata

100-300 8 2 1 1 4 0 1000-1960 0 3 5 0 1 3 2000-2200 0 6 4 0 1 5 192 ORNITHOLOGICAL MONOGRAPHS NO. 11

ß•:'. :- ,• ß : . . :.•' ß .•: ß :...: . •;

Figure 25. Distribution of Lagonosticta rubricata. known from different altitudes in the Nsanje District, a narrow strip no wider than 20 miles lying betweenlatitudes 16ø30'S and 17ø10'S. V. chalybeatais most abundantin the flood plains, V. purpurascens(whitish feet, purplishplumage) occursboth in the low country and in the hills, and V. funerea (red feet, plumagegreen to blue) lives mainly in the hills. L. rhodopareia parallels the local distribution of V. purpurascens,and L. rubricataoccurs mainly in the hills with V. funerea. While averagetempera- turesthroughout this regionare similar,the mean annualrainfall (68 inches) at 1,960 feet altitude at Chididi is about twice the rainfall (35 inches) at 200 feet at Nsanje (Long, 1960: 88); and an altitudinalmoisture gradient is evidentwith more rainfall at higherelevations. As in SouthAfrica, Rho- desia,and Zambia L. rubricata and V. funerea are birds of wetter country whereasL. senegalaand V. chalybeataoccur in areasof densehuman populations. L. rhodopareiaand V. purpurascensin Malawi overlap L. rubricata and V. funereamore than in most areasof southernAfrica, perhapsdue to the mosaicof habitatsand diversetopography in southernMalawi. In northernMalawi occurtwo kinds of purplish-blue,pale-footed indigobirds, V. •. nigerrbnaand V. purpurascens.North of 15 ø latitude the birds above 4,000 feet elevation correspondin distributionto L. rubricata and at 3,800 feet near Lilongwethese birds, relatively more blue in male breeding 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 193

V. œunerea

Figure 26. Distribution of Vidua funerea (excluding the most highly problematical Congo birds from Kasai and Tanganika). plumage,mimic L. rubricata. At lower elevations,especially on the hot, dry plains south of Lake Malawi, the birds occur with L. rhodopareia,and these averagemore purplishand are V. purpurascens.These two forms appearto intergradewith eachother in Malawi. Zambia.--The four firefinchesin Zambia are ecologicallyseparated in habitat preference with L. rubricata occurring in wetter areas and L. rhodopareia in drier areas (Benson and Irwin, 1967: 118). L. senegala occursin hot dry areas,often with L. rhodopareiabut alsoin villages(Benson and White, 1957: 131), and the unparasitized firefinch L. (rufopicta) nitidula is found along major rivers. The indigobirdsdo not separateclearly with their hostsinto differentregions, largely because of the overlapbetween L. rhodopareiaand L. senegala. Excluding northeasternZambia, V. chalybeata was taken with L. senegala in six localities, but in five of these L. rhodopareiaalso was taken. Similarly V. purpurascenswas found with L. rhodopareiaat nine localities,but with L. senegalaat eight, includingfour where all three host firefinch specieswere collected. V. funerea was taken with L. rubricataat three localities. All of theseindigobirds were taken more frequentlywith their known songmodels in southernAfrica than with other fire finches. Distributionalpatterns in Zambia may be summarizedby recognizinga few major geographictrends. The speciesV. junereais greenin southeastern Zambia, Chilanga birds range from blue-greento blue, and Mwinilunga birds are purple-blue to purplish-blue. These last are clearly associated with L. rubricata, as north of 14ø S latitude in western Zambia this is the 194 ORNITHOLOGICAL MONOGRAPHS NO. 11 only host firefinch. The Mwinilungaarea has been well collectedby several workers(Benson and Irwin, 1967: xi); I met only L. rubricatathere in three weeksof field work in September,1966, and it is unlikely that other host firefinchesare in the region. The two specimensof V. f. codringtonifrom the LuangwaValley region (Chipako, Mulilo) are from a rather dry area, but alongthe adjacentescarpment the habitat may be sufficientlymesic for L. rubricamto be expectedalso (see Bensonand White, 1957; plate 9b). In contrastL. rhodopareiaand V. purpurascensare most often found in southernand southeasternZambia in the hot, dry valleys of the Zambezi and Luangwarivers. V. chalybeatain Barotselandand adjacentareas near Livingstoneand Balovaleusually have white bills; they resemble the Okavango birds rather than the red-billed form amauropteryx. Both of these forms of V. chalybeataare associateddistributionally with L. senegala. In the Northern Province and nearby Lake Lusiwasi three firefinches occur. Indigobirdsin that regionare difficultto identify. The speciesproblem is discussedin the sectionon the CentralAfrican speciescomplex; apparently both V. chalybeatacentralis and V. funereanigerrima occur there. Angola.--Three speciesof indigobirdsappear to be present in Angola, but as few specimenswith foot color recordedare known and no information is availableon song,future field observationsmay alter the presentunder- standing. Of the firefinches,L. senegalaand L. rhodopareiaoccur in the drier southernparts and L. [rubricam]landanae appears more commonly in the northern regionsin the same woodlandvegetational belt as does L. rubricata in northern Zambia. Red-footedindigobirds in Angola are all identified as V. chalybeataand in southwesternAngola occur togetherwith L. senegala. L. rhodopareia ansorgeihas been collectedsouth of 12 ø latitude in severallocalities including a high elevation(5,700 feet) in Huambo;one has alsobeen taken north at Dondo at 9ø38'S, but L. [rubricata] landanae is the more common form in the north of Angola. A purplish,pale-footed V. purpurascensis known from Gambos,Mossamedes, and anotheris from Huila, Huila; the second localityis alsoa siteof L. rhodopareia.Indigobirds north of 11ø30' latitude appearto be mostlyV. [unereanigerrima, and the mostcommon host firefinch is L. [rubricata] landanae.

EAST AFRICA Tanzania.•Three foster firefinch speciesoccur in Tanzania. Lagonosticta senegalais the only commonform in the dry Masai plains of the north, L. rubricataoccurs mainly in the southernhighlands and the northernmoun- tainsand by Lake Victoria, and L. rhodopareiais widelyscattered through central Tanzania. Bluish Vidua chalybeata in Tanzania includes both red- billed birds (V. chalybeataamauropteryxJ along the coast and white-billed i972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 195 birds (V. chalybeatacentralis) inland; both forms have reddishfeet. Both were taken with L. senegalaat ten localities,more than with any other firefinch. Greenish-glossed,red-footed V. lunereacodringtoni males have been taken at only two localities,at Mikindani on the southerncoast and at 5,150 feetat theriver near Iringa (Lynes, 1934: 128). The other indigobirdsof Tanzania are difficult to name; these are the birdswith bluish to purplishplumage and with whitish (not red) feet. I have not recordedsong and collectedthe singersthere as I have in otherparts of Africa, soI cannotdirectly match the songsand appearance of the Tanzanian birds with the birds sampledin other areas. Nicolai (1967) has studied indigobirdmimetic song in Tanzania, but he has not collectedspecimens of the singingbirds, and he has, I believe,misapplied the namesin his speciesidentifications, or he has at leastapplied them in a way differingin a consistentmanner from my own usage. In other areas of easternand southernAfrica where I have recordedand collectedthe indigobirdV. purpurascens,this form mimics L. rhodopareiaand occurs with it in relatively dry bushand acaciahabitats. In contrast,in my studyareas I found V. runeteato mimic and to occurwith L. rubricatain generallymore moist habitats.Hence I am applyingthe criteria of firefinchdistribution and habitat to aid in the identificationof the pale-lootedindigobirds in Tanzania. I have here regardedthe pale-lootedbirds in the drier parts, the birds found with L. rhodopareia,as V. purpurascens;the indigobirdsof moisterareas found withL. rubricataand described in the systematicsection I referto V. lunerea nigerrima,the samesubspecies that I foundto mimicL. rubricatain adjacent Malawi. The indigobirdswith pale feet and purplishto bluishplumage from the localitiesof Bagamoyo,Morogoro, Sunya, and Undis are known from relativelydry regions(Keay, 1959) andtheir plumage is morepurplish than blue,and in the first two of theselocalities the firefinchL. rhodopareiahas been taken. I regard these indigobirdsas V. purpurascens.The birds from Bukoba,Nyarunbogo, and Ukerewe,all by Lake Victoria, are the bluestof the indigobirds(excluding V. chalybeata)from Tanzania,and in thiswet regionthe firefinchL. rubricatais commonand L. rhodopareiais unknown. Theseindigobirds are V. [unereanigerrima; other localitieslisted for it in my Appendix B are in Brachystegia-Julbernardiamoist woodland biomesor at elevationsabove 4,000 feet, and for most of them L. rubricata is known.The remainingindigobirds cannot be assignedwith certaintyto speciesas their localities are in areasat theedges of majorvegetation types, or bothof the firefinchesare knownfrom the samelocality, or nothingat all is knownabout the habitator localfirefinches. The one exceptionis Usegua,a varied regionwhich is the type locality of V. purpurascens.To avoida nomenclatorialmorass the purplishbirds from Useguaare regarded 196 ORNITHOLOGICAL MONOGRAPHS NO. 11 as conspecificwith the purplishmimics of L. rhodopareiathat I recorded and collectedin other regionsof southernand easternAfrica. Nicolai (1967) reported that "purpurascens"mimics L. rubricata in Tanzania. His identificationswere basedentirely upon field observations andcomparisons with songsof captivebirds; he did not collectany specimens. Unfortunatelyit is difficult to distinguishbetween the forms V. f. nigerrima and V. purpurascenseven in museumspecimens, and specimensintermediate in color suggestssome introgressionbetween them. The habitat shownin Nicolai's photographof one locality in Tanzania appearedrather dry and suited to L. rhodopareiarather than L. rubricam (Payne, 1968a: 36); Nicolai writes me (in letter) that most of thesebirds were recordedin more humidhabitats, particularly along rivers, where L. rubricataoccurred. Prob- ably Nicolai's (1967) observationsof indigobirdsmimicking L. rubricata wereof the birdshere regarded as V. f. nigerrima.At any rate, the purplish- blue, pale-footedindigobirds of Tanzania appear to consistof two sibling forms,a bluer one that livesin moisthabitats with L. rubricataand a purpler form that lives in acacia steppehabitats with L. rhodopareia. Indigobirds occur throughoutthe collectiverange of these two firefinchesand are not restrictedto the rangeof only one of them. The problemis discussedfurther on pp. 239-240 and pp. 263-273. On distributionalgrounds it is unlikelythat Tanzaniaindigobirds identified on sight by Nicolai (1967: 311.) as "codringtoni"mimics of L. senegala were in fact this form. Codringtoniis much lesswidespread than V. chalybeata; only two museumspecimens are known from Tanzania. The form codringtonicoexists with V. chalybeataamauropteryx, the common mimic of L. senegala,in Rhodesia,Malawi, and Zambia without apparentinter- breeding. In Rhodesiaand Malawi these forms mimic different firefinches. V. f. codringtoniand V. chalybeatacentralis are similar in appearance (greenishor bluish-greenversus bluish or greenish-blue,with brown-black wingsversus dark brown wings, and both with white bill and reddishfeet). ProbablyNicolai's "codringtoni"was V. chalybeatacentralis. Field identification of the Tanzanianindigobirds certainly would be most difficult when not aidedby collectingspecimens. Kenya.--Although four host firefinchesoccur in Kenya, only two species of indigobirdsare known. L. senegalais widespreadboth in semiaridlow- lands and in the mesichighlands to 7,000 feet. L. rubricata and L. rhodopareia replace one another altitudinally,with L. rubricata commonin the mesichighlands and L. rhodopareiaalmost entirely in drier areasbelow 3,500 feet. L. rara occurslocally in westernKenya at Kakamegaand Mt. Elgon. Distributionof the dark-wingedindigobirds of the speciesV. chalybeata closelyagrees with that of L. senegalaas 25 of the 28 known localitiesof the indigobird are known localitiesof this firefinch as well; the other three 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 197 localities are all near known sites of L. senegala. A secondindigobird V. purpurascens,occurs in the thornbushlowlands in Kenya, a pale-winged, pale-footedmimic of L. rhodopareia.At three of the four Kenya localities of this indigobird,collectors have also taken the firefinchL. rhodopareia, and we visited the fourth (Kacheriba); it was similar in habitat (though now more heavily grazed) to Sigor about 30 miles east, with grassyopen woodlandof Acaciaand Euphorbia spp. and probably L. rhodopareiaoccurs there also as at Sigor, where we recordedthe indigobirdsmimicking this firefinch. No indigobirdin Kenya is known to mimic the other two firefinches occurringthere, even though one of these(L. rubricata) is widespread.All 20 maleindigobirds taken at 5,500 feet at Nairobi whereL. rubricatais about as commonas L. senegalalook alike and there is no suggestionof two morphologicallydistinct forms. Both of the Nairobi indigobirdsthat I heard mimickedL. senegalaas did all birds in the Kisumu region where both firefinches also occur. In one area at 7,800 feet elevation six miles east of Kericho, on the Kaisugu Tea Estates, L. rubricata occursin numbers but L. senegalais absent. Jenny Home informs me that indigobirdshave not beenseen there at all. There appearsto be in Kenya a widespread,abundant populationof L. rubricatathat is not exploitedby indigobirds. Uganda.--All five host firefinchesoccur in Uganda, but only one form of indigobirdis known. The most common firefinch is L. senegala,and at six localitiesit was taken with V. chalybeatacentralis. Two indigobirdsfrom westernUganda and one from Kampala are purpler than other local V. chalybeatabut they match them in size and they have darkish brown (not pale) wings. One of these birds (foot color "orange-horn") was taken togetherwith both L. senegalaand L. rubricata (Kabale) and none was taken with other firefinches. Near the northern border of Uganda in the vicinity of Nimule other kinds of indigobirdsare known, but in the absenceof definite records of "Nimule" birds collecteddearly south of the present border, they are discussedwith the Sudan birds. Some forms of the speciesV. wilsoni may occur within the northern bordersof Uganda, inasmuchas the Northern Guinea Woodlandsbiotype (the main biotype of theseindigobirds in westAfrica and the northernCongo region) extendsinto northernUganda and the hostsof these indigobirds--L. rara, L. larvata, and L. rubricata-- occur in Uganda. Burundi and Rwanda.--Indigobirdseast of Lake Kivu and the Congo forestsand westof Lake Victoria appearto be a singleform of V. chalybeata. Two firefinchesare widespreadhere, L. senegalaand L. rubricata. In six localitiesthe indigobirdsoccur with L. senegala.In two of theseL. rubricata also was collected but in no other locality were the indigobirds and L. rubricata taken together. 198 ORNITHOLOGICAL MONOGRAPHS NO. 11

Figure 27. Distribution of Lagonosticta larvata.

NORTH AND WEST AFRICA Ethiopia.--Four speciesof host firefinchesoccur in Ethiopia but only two kinds of indigobirdshave been collected. The most widespreadfirefinch is L. senegala;L. rhodopareiaoccupies the dry south,L. rubricata occurs in the mesichighlands, and L. larvata lives along the upper Blue Nile and Lake Tana. Nearly all indigobirdsare black-wingedV. chalybeata;these have been taken in ten L. senegalalocalities and one L. rubricatalocality and are thus closelyassociated with L. senegala.In northwesternEthiopia along the upper Blue Nile occursthe bluish form, "camerunensis,"of V. wilsoni; at Gallabatit wastaken together with L. larvata. Somalia.•The one specimen(Senckenberg Institut, Erlanger no. 2038) of indigobird is from Abrona and resemblesthe small V. chalybeata amauropteryxfrom coastaleast Africa. L. senegalais the only firefinch known from dry Somalia. Sudan.---In Sudanthe mostwidely distributedindigobird is V. chalybeata. In every locality where it was taken with a firefinch, the latter was L. senegala.The distributionsof the pale-wingedforms of V. wilsoni, on the other hand, do not clearly correspondwith thoseof the various speciesof firefinches;specimen localities are few. Bluish V. wilsoni have been taken with L. larvataonce, with L. rara twice, and with L. rubricataonce. Purplish .....•/_.rara ßD.": Figure 28. Distribution of Lagonosticta rara. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 199 on/" ß Figure 29. Distribution of Vidua wilsoni (form "wilsoni").

V. wilsoni have been taken at five localities in the south but not with any firefinch. GreenishV. wilsoniare mainly southernalso, as are the firefinches L. rata, L. rubricata, and L. larvata. A greenishindigobird ("nigeriae") has been taken also in the dry northwestat Kulme, Darfur, with L. larvata (Lynes, 1924: 670). On the Boma plateauin southeasternSudan perhaps severalspecies of indigobirdsmay live together;all of the knownhost fire- finchesincluding L. rhodopareiaoccur in this generalareaß Chad, Central Ajrican Republic.-•In West Africa the vegetationalbelts lying betweenthe Saharaon the north and the oceanon the southrun from east to west, and the distributionof firefinches and indigobirdsclosely parallelsthe rainfall and vegetationpattern. Indigobirdsfrom northern Chad are V. chalybeata, and at two localities these were taken with L. se'negala.

n/get/ae

.o o

Figure 30. Distribution of Vidua wilsoni. Above, green birds ("nigeriae"); below, blue birds ("camerunensis"). 200 ORNITHOLOGICAL MONOGRAPHS NO. 11

In Chad the "camerunensis" form of V. wilsoni occurs with L. larvata at Ir•na. In the Central African Republic only pale-wingedindigobirds are known, purplish "wilsoni" and bluish "camerunensis."L. larvata is apparentlythe most commonfirefinch and thus may correspondin distribution with the more common "camerunensis" form. Cameroom--Finch distributionis related to moistureregimes in Cameroon with L. senegalaand V. chalybeatafound in the semi-arid north while the southernmostspecimens near the border of the rain forest are L. rubricata and the green form of V. wilsoni. These last two occur togetherwith L. larvata,L. rara, and the blue and purpleforms of V. wilsoniin the intervening countryas well as in the highlandsof westernCameroon and Adamoua. As the three forms of pale-wingedindigobirds examined were taken with fire- finchesat only five localitieslittle more can be said about their distribution. Nicolai (1968) visitedNgaoundere for a few days and observedL. rara but no other firefinches. Severalbluish, pale-wingedindigobirds here mimicked L. rara and one greenishindigobird mimicked L. larvata. The singingindigobirds were not collected,but a blue "camerunensis"was netted in the brush. Nicolai has shown me the skin and I have comparedit with my seriesof Nigerianspecimens; the Ngaounderebird was about as blue as the Nigerian "camerunensis."The habitat and avifaunaaround Ngaoundere are generally like those in the Zaria region in northern Nigeria, and althoughthe relief is high the areais at the edgeof the relativelydry woodland(Bates, 1924: 23, plate 4; Keay, 1959), and L. larvata and L. rara are both to be expected around Ngaoundere. Nigeria.-•Indigobirds have been collected at more localities in Nigeria than in any otherWest African country,and in additionthe sightobservations of J. H. Elgood, C. H. Fry, W. Serle, D. Wells, and myself provide further useful information on local distribution. In the Sudan savanna of the Sokoto-Kano-LakeChad area the only firefinch presentis L. senegala,and all of the indigobirdstaken here are black-wingedV. chalybeata. These two occur also in the Guinea woodlandsespecially near villages;L. senegala extendssouth in the east as far as Numan and Ganye;in the west it has not been recordednear the coast. Two ANSP specimensof V. chalybeatafrom Lagos are from an area devoidof the usual host; the only firefinch known from the Lagos area is L. ruJopicta(J. H. Elgood, pers. comm.). Lagos was for many yearsa market centerfor live birds,and perhapsthe two indigobirds were trade birds. In northern Nigeria every town and village I visited had both L. senegalaand V. chalybeatahopping about the housesor feedingwith chickensor singingfrom smalltrees and powerlines. Occurringin the more humid regionsis L. rubricata. It is absentas far north as the Northern Guinea Woodlandsat Zaria (Fry, 1965, 1966; N.J. Skinner,pers. comm.; also my observations)where all other west African 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 201

Figure31. Habitatof PanshanuPass. Nigeria. from the call-siteof a male "nigeriae" that mimickedLagonosticta tubricata. Tall grasswas growing in rich blacksoil in well- watered crack• in the rocks. firefinches occur. More than half of the L. rubricata rccords are in a humid regionnear the Jos Plateauand two of the four recordsof the "nigeriae" indigobirdsare from thisarea (Figure31). L. rubricatawas also seen and tape-recordedat Panshanu,a localityof the '•nigeriae"form of V. wilsoni. Kiri, thetype locality of "Hypocheranigeriae," is drier andlies at the siteof emergenceof the GongolaRiver from the hills of the Kaltungoplateau into the hot plainsof the BenueRiver (Figure 32). In August,1968, whcnl visitedKiri the area was well grazedand devoidof tall grasson the hills and generallyhotter and drier than the L. rltbricatasite at PanshanuPass. In 1968,the only firefinchI sawat Kiri wasL. senegalaand the only indigobird was V. chalybeata.However, in August, 1904, when the type of nigeriae was collected,Kiri was a famine area, few peopleor domesticanimals were alive, and the grasswas tall (Alexander, 1907) and perhapswas then a suitablehabitat for other speciesof firefinches. Blue V. wilsoni (the form "ca•nerllttettsi3'")were taken or observcdwith the songmodel L. larvata at five localitiesin Nigeria, more than with any other firefinch. The generalpattern of distributionof L. larvata is similar to that of L. rara, and at Zaria the two firefincheshave nearly identicalhabitat 202 ORNITHOLOGICAL MONOGRAPHS NO. 11

:, •

Figure 32. Tall grass along the Benue River at Numan, Nigeria, looking upstream at •e confluence of the Gongola River below Kiri. Lagonosticta xenegala and Vidua chalybeatafed on fallen grassseeds on the path. requirements(Fry, 1966: 344). Both firefinchesoccur with L. senegalain the northern parts of their range and both occur with L. rubricata in the more humid southat Enugu, where again no habitat differencesare apparent (Serle, 1957: 680). Although the number of localitieswhere L. rara was seenor collectedwith an indigobirdis small, the agreementin distribution with its mimic at Zaria, the purplish "wilsoni" form of V. wilsoni, is clearer than with any other kind of indigobird, and Figure 33 shows a relatively large number of L. rata firefinches and "wilsoni" indigobirds in humid southernNigeria. A tendencyfor L. larvata firefinchesand "camerunensis" indigobirdsto be more widespreadin the Northern Guinea Woodland and for L. rata firefinchesand "wilsoni" indigobirdsto extend farther south into westernand midwesternNigeria is alsoapparent. Dahomey, Togo, Ghana, Ivory Coc•'t.---Across the Guinea Woodlands of west Africa occur the same forms of firefinches and indigobirds as in Nigeria. Distribution maps here show no clear associationof firefinches with any one pale-wingedindigobird; few specimenshave been taken. L. senegalaapproaches the coastin the DahomeyGap, a relativelyarid region extendingsouthward to the coast and separatingthe Upper Guinea and Lower Guinea forests (Moreau, 1966: 163). L. senegala is more common 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 203

wm z • • 204 ORNITHOLOGICAL MONOGRAPHS NO. 11 in the northernregions of thesecountries, and all recordsof V. chalybeata are north of 8 ø N latitude. Niger, Mali, Upper Volta.--In the semi-aridsavanna south of the Sahara, L. senegalaand V. chalybeataare foundnear many townsto the edgeof the Sahara as far north as Tombouktu and even into the desert at Air. One pale-winged,bluish "camerunensis" indigobird of the speciesV. wilsoniwas taken in the centralNiger delta; a specimenof L. larvata vinaceais known from a few hundredmiles upstreamat Bamako (Malzy, 1962: 58). A sexuallymonomorphic form of firefinch with an unnotchedouter primary is knownonly on rockyhills in centralMali. This bird, L. virata,is usually recognizedas a form of L. rubricata, which elsewhereis a bird of moist habitat;White, however (1963b: 204) regardsit as a form of L. rhodopareia. Probably virata is a relict form survivingin locally mesic areas and was isolatedby dryingconditions south of the Sahara(see Moreau, 1966: 53-60) from L. rubricata;Goodwin (1964: 104) regardsthe songof virata as similar to the songof L. rubricata. One greenishV. wilsoniindigobird ("nigeriae") apparentlywas taken in "SoudanFran•ais" by deCotte in the last century, andit may havebeen associated with L. ru. virata in mesicareas in Mall Sierra Leone.--Most of Sierra Leone lies in the roesic Guinea Woodland and in Moist LowlandForest and herethe mostwidespread firefinch (Figure 25) is L. rubricata (Serle, 1949: 124). L. larvata vinacea is known from one specimenwith a definitelocality in Sierra Leone, and L. rara is known at two others. Most indigobirdsare blue-glossedand pale-wingedbirds morphologicallyreferable to V. wilsoni (the form "camerunensis"). One black-wingedV. chalybeatawas taken from the mouth of the ScarciesRiver; its host, L. senegala,has not been noted there but is common upstreamin the drier northeast (Sefie, 1949: 124). Guinea, PortugueseGuinea.--In the hills of Fouta Djalon V. chalybeata is widespread;there the only firefinch specimensknown are L. senegala. Three host speciesof firefincheshave beentaken near the coastin Portuguese Guinea,L. senegala,L. larvata, and L. rubricata. Two forms of indigobirds are known from this area, V. chalybeataand purplish V. wilsoni. Senegal, Gambia.-•Between the Sahara and the Guinea Woodlands both L. senegalaand V. chalybeataextend along the SenegalRiver at the edge of the desertto the coastin Senegal. At most sites of V. chalybeatathis firefinchhas also been taken. V. chalybeataparasitizes L. senegalain Senegal (Morel, 1969). In Gambia L. larvata vinacea occurs; Hall and Moreau (1962: 354-355) list the firefinch for several unspecifiedlocalities. Old specimensof vinacea labelled merely "Senegal"or "Senegambia"and also a specimenfrom Niokolo-Koboin southeasternSenegal (Dekeyser, in Hall and Moreau, 1962: 378) suggestit is locally commonin the westernUpper 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 205

Guinea Woodlands. Bluish "camerunensis" males of V. wilsoni are known for Kuntair on the Gambia River.

CENXR^I• AFmC^ Gabon,Congo (Brazzaville).--On large-scale vegetation maps the countries of westequatorial Africa are largelyhumid forest regions and little habitat is suitablefor firefinchesand indigobirds (Chapin, 1932:104 et seq.;Keay, 1959; Rand et al., 1959: 233-234), but local clearingsin Gabon may have permittedpassage of someopen-country birds through the forestedregion (Rand et al., 1959: 236-238). In Gabonand Congo-Brazzaville(= Moyen Congo)the only known firefinch is L. rubricata.A singlespecimen of indigobird is known. A bluish-purplebird morphologicallyindistinguishable from westAfrican bluish-purple V. wilsoniwas taken at N'gabe,Congo (Brazza- ville), in the humidforest-savanna mosaic south of the equatorialforest; thisis the bird callednigerrima by Dekeyserand Derivot ( 1966-1968:418) and by its collector,Malbrant (in Malbrantand Maclatchy,1949: 421); I have followedthese authors in callingthe bird V. runeteanigerrima. Congo (Kinshasa).--Excluded from the huge central and western equatorialforest region of the Congo,firefinches and indigobirds live aroundthe northern,eastern, and southernparts of the country. North of the Congo River in the Ubangi,Moyen-Congo, and Uelle regionsoccur the threeroesic- countryfirefinches and the pale-winged indigobirds (blue, green, and purplish V. wilsoni) which live also in west Africa. Each of the color forms of V. wilsoniwas taken at least once with the firefinch speciesrecorded as the songmodel in Nigeria. Hall andMoreau (1970) recordpurplish V. wilsoni (the form "wilsoni") as the most widespreadindigobird in the Uelle; however, their localitiesare basedon the identificationsof Schouteden(1962: 137, 1963: 228) who namedwithout explanationor descriptionall indigobirds of the northernCongo as "Vidua funerea wilsoni" regardlessof their appearance.Schouteden's birds that I saw at MRAC includedsome green and blue males also. No L. senegalaor V. chalybeataare known from this region. Along the easternborder of the Congoin Ituri and Kivu provincesthe most common firefinches are L. rubricata and L. senegala, and L. rara is alsopresent in the north. All of the indigobirdscollected are V. chalybeata centralis. In Ituri, L. senegalais the most widespreadfirefinch; both indigobird specimenswere takenwith it. In Kivu somevariation in color occurs in indigobirdplumage, but two morphologicalforms are not clearly dis- cernible. The same number of indigobird localitieswere L. senegalasites as were L. rubricatasites, and this may indicatea meaningfuldistributional correlationbetween the indigobirdsand L. senegala,since L. rubricata is 206 ORNITHOLOGICAL MONOGRAPHS NO. 11 more widespreadhaving been taken at nearly twice as many localitiesas L. senegala. In Kinshasa(= Leopoldvilleprovince) indigobirds have been taken only at Boma (near the mouth of the CongoRiver) and up-river at Kwamouth. The Boma bird is small and purplish-blueand is morphologicallysimilar to the indigobirdspecimen from N'gabe, here called V. funerea nigerrima. The Kwamouth birds are blue and greenish-blue.As describedin the section on indigobirdvariation, they are small and in size resemblethe west African V. wilsonicomplex, but theyare darker-wingedlike the south-centralAfrican V. •. nigerrima,and they may be intermediatebetween these forms. For the present,however, I refer to them as V. •. nigerrima. Chapin (1932: 149) describedthe habitat at Kwamouth: "On my way down the CongoRiver, as we came out of the forest at Bolobo, and tarried a few hours there and at Kwamouth and Kunzulu, I was impressedwith the strong resemblance of this savanna to that of the Uelle." However, he noted no actual connectionof open woodlandto the Uelle districtand commentedrather upon the continuityof the equatorialforest betweenthese regions. The similarityof the habitat and of the birds of the open woodland (Chapin, 1932: 149) suggestssome possibleearlier connection through or around the forest. Firefinches are known from eight localitiesin Kinshasa,all of them along the CongoRiver. Two from Bas Congoare L. (rubricata) landanaeas in coastalPortuguese Cabinda and the othersare L. rubricata congica. Thus at the wet westernend of the Congo drainagesystem the indigobirdsare associatedwith L. rubricata. In Kasai provinceL. senegalaand L. rubricataare known from only a few localities;the latter form is perhapsmore widespread. A specimenin MRAC from Luluabourgwas reportedby Chapin (1954: 524) (also Hall and Moreau, 1970: 337) to be L. rhodopareia (= "L. ]. ]amesoni"), but all of the large Luluabourgfirefinches examined in my study were L. rubricata; Schouteden(1964) and Immelmann et al. (1965) mention no L. rhodopareiafor this area. At only four localitieswere indigobirdstaken with firefinches, and at two of these (Luebo, St. Joseph'sMission) both firefincheswere collectedby Fr. Callewaert. Indigobirdsin this region are variableranging in color from greento blue. At the presenttime we do not know whetherthey compriseone or two species;perhaps both V. chalybeata and V. funerea occur and interbreed. In Katanga south of 10ø S latitude are two localitiesof L. rubricam; at one site of two L. senegala,22 L. rubricata were taken also and evidently it is the more common species. The numerical dominance of L. rubricata here is similar to that in northwesternZambia. The two indigobird specimens in southernKatanga with foot color data had "pale flesh" and "flesh"feet. All of the southernKatanga indigobirdsare treated as V. funerea nigerrima. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 207

Severallocalities in UpembaNational Park in centralKatanga are known for L. senegalaand L. rhodopareiafirefinches, but no indigobirdsin breeding plumagehave been collectedthere. Chapin (1954: 524) recordsL. rh. ]amesonifrom two localitiessouth of 10ø S latitude in Katanga, but all L. rhodopareiaspecimens seen in the presentstudy from Katangawere taken north of 10ø S latitude (Kaluli, Kenia, Upembaregion, and Kabolo). In the Tanganikaregion of Katanga,as in Kasai, the indigobirdscannot be identifiedto specieson the basisof the morphologicalcharacters available.

DISCUSSION The distributionof each morphologicallywell characterizedform of indigobirdcorresponds reasonably well with the distributionof the firefinch known to be its songmodel and presumedto be its host. V. chalybeatais distributedwidely in Africa and occursin villagesin dry regionssuch as the edgesof the Kalahari and Sahara desertsalong with L. senegalafirefinches which obtain water and food from human activities. V. runetea and its songmodel L. rubricamlive in moisterregions near the edgesof evergreen forest in central Africa and along the continentalescarpment and on the coolerplateaus in southernAfrica. In southernAfrica, thesebirds overlap in intermediatehabitats with the typicallydrier-country forms V. purpurascens and its usualsong model L. rhodopareia.A similarhabitat difference is apparent in southernMalawi where V. funereacodringtoni and its model L. rubricamoccur mainly in the moist,higher altitudes whereas V. purpurascens and model L. rhodopareiaare more widespread. The variousforms of pale- wingedindigobirds (V. wilsoni) occurringin subsaharanAfrica north of the Congo do not seem to be correlatedwith single speciesof firefinches acrossthis entire range, but within certain regionsof the range each color form may be closelyassociated with a single firefinch. The blue form "camerunensis"occurs with L. larvatain Ethiopia,Nigeria, and Chad-Central African Republic, whereasthe blue birds in the Upper Guinea region (especiallySierra Leone) are more closely associatedin distributionwith L. rubricata. Purplish"wilsoni" and green"nigeriae" forms of V. wilsoniare lessdeafly associatedin their distributionwith a singlehost firefinch species than are the bluish"camerunensis" specimens. Trendsare evidentalso when individuallocalities at which both indigobirds and firefincheswere collectedand examined (or observedby me in the field) in southernAfrica, east Africa, and north and west Africa are compared(Table 29). In eachof theseregions V. chalybeatawas taken most often with its known host L. senegala. The data for southern Africa indicatea good correlationbetween the distributionsof V. funerea and L. rubricam and also between those of V. purpurascensand L. rhodopareia; a 2 x 2 contingencytest on the dataof rowsand columns2 and 3 of Table 29 2O8 ORNITHOLOGICAL MONOGRAPHS NO. 11

TnBLE 29 INSTANCES OF LOCAL SYMPATRY OF INDIGOBIRDS AND FIREFINCHES 1

d rea Vidua Lagonosticta

senegala rhodopareia rubricata Southern Africa ø chalybeata 52 37 7 purpurascens 33 33 10 /unerea 6 9 41 East Africa a chalybeata 50 4 17 purpurascens 3 3 0 [unerea 0 0 0 senegala rara larvata rubricata North and West Africa • chalybeata 60 3 7 7 "wilsoni "5 4 6 5 2 "camerunensis" 7 11 12 10 "nigeriae" 0 3 1 3 Excluding Angola, Katanga, Kasai, and Tanzania, as indigobirds there are hard to identify. South Africa, Botswana, Rhodesia, South West Africa, Mozambique, Malawi, Zambia. Kenya, Uganda, Ruanda, Burundi, Kivu, Ituri. Ethiopia, Sudan, Congo (Uelle, Ubangi, Moyen-Congo) to Senegal. "wilsoni," "camerunensis," and "nigeriae" are regarded as color forms of Vidua wilsoni. givesa p < .001, a highlysignificant correlation in local distributionbetween these indigobirdsand firefinches. The few localities which have been well collected in east Africa do not providedata sufficientto illustratea clear correspondenceof the V. Junerea- L. rubricata and the V. purpurascens-L.rhodopareia combinations in that region. In easternAfrica (excludingTanzania) examplesof V. chalybeata were taken at 72 localities,and specimensof L. senegalawere taken at 49 of the samelocalities. Most of thesebirds were collectedfor purposesother than documentingthe completelocal avifaunaor the concurrenceof indigobirds and firefinches,and viewedin this light the figure of 68 percentcoinci- denceis high. Particularlywhen comparedwith east African ploceidswhich do not have sucha closeinterspecific ecological relationship as do the mimetic viduines and their hosts, the local coexistenceof these specimensindicates more than a spuriouscoincidence. This is demonstratedby the case of the Chestnut Sparrow (Passer eminibey) in east Africa. This sparrow often appropriatesthe nests of the Grey-cappedSocial Weaver (Pseudonigrita arnaudi) in the same habitat (Payne, 1969). It also uses nestsof other ploceids,and it may sometimesbuild its own nestsat times, accordingto van Someren and van Someren (1945: 43-44), Fuggles-Couchmanand Elliott (1946: 345), and BeLts (1966: 528). In a sample of 52 localities where Chestnut Sparrows have been collected, the Grey-capped Social Weavershave been taken at only 13 of these,or only 25 percent (Payne, 1969: 304). 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 209

In west Africa one could not predict with confidencesolely on the basis of instances of local coexistence of the various color forms of V. wilsoni with their firefinch models in song that "camerunensis"is associatedwith L. larvata rather than with L. rara (or converselyfor "wilson?'). The Bar-breastedFirefinch L. rufopicta has a distributionrather similar to those of these other firefinches(White, 1963b; Immelmann et al., 1965; Mayr et al., 1968; Hall and Moreau, 1970) and only the absenceof any instances of mimicry of this firefinchin the dozensof indigobirdsheard and recorded within the rangeof rufopictaclearly showsthe associationof the indigobirds with the other speciesof firefinches. In central Africa where no field work has beencompleted on songit is unclearfrom the distributionsof indigobirds and firefincheswhich indigobirdforms are associatedwith L. senegalaand L. rubricata. The distributionalinformation thus showsgeneral trends which tend to confirm the restricted,specific host-parasite relations of the firefinchesand indigobirdsknown from song recordings,but the distributionaldata are of limited value in determiningrelations where more than a single species of host firefinch occurs.

VARIATION AND RELATIONSHIPS IN THE INDIGOBIRDS

The relationshipsamong the indigobirdshas been one of the most contro~ versialproblems in the systematicsof African birds. Extremesin taxonomic treatment have ranged from considerationof most plumage variants as distinct speciesto a conceptionof indigobirdsas a singlepolymorphic species. Mackworth-Praedand Grant (1949) recognizedeight species,Delacour and Edmond-Blanc (1934) recognizedsix, Sclater (1930) and Chapin (1954) admittedfive (different combinations),Friedmann (1960) and Wolters ( 1961 ) recognizedthree, and Traylor (1966) andWhite (1962, 1963a, 1963b), each for different reasons, had one.

BASIS OF IDENTIFICATION AND TAXONOMIC ALLOCATION OF SPECIMENS Earlier I regardedthe indigobirdsmimicking a commonspecies of host firefinchesas conspecific,and the birds mimicking different firefinchesas distinctspecies simply on the basisof their song(Payne, 1968a, 1968b). The reasoningbehind this notion was that mimeticsongs were the main isolating mechanismamong indigobirds, so thereforethe birds with different mimetic songswere behaviorallyisolated at the specieslevel (Nicolai, 1964, 1967). One must also consider,however, whether mimics of the same host species do in fact generallyinterbreed and whethermimics of differenthosts do not. For that reason,in addition to the sharingof mimetic songsI have used the followingcriteria to distinguishtaxonomic rank: (1) When allopatricforms 210 ORNITHOLOGICAL MONOGRAPHS NO. 11 sharethe samemimetic song and alsoshow morphological evidence of interbreeding,they are consideredconspecific. (2) Whenallopatric forms share the samesong model but showno morphologicalevidence of interbreeding, they are consideredto be differentspecies only if callingthem conspecific wouldbe confusingor wouldobscure important biological differences between them;otherwise they areregarded as conspecific. (3) When sympatricforms with differentsong models show no evidenceof interbreedingover muchof their commonrange, they are consideredto be differentspecies. (4) When someinterbreeding between populations with different songsis indicatedby the presenceof morphologicallyintermediate males, sympatric forms are consideredto be conspecificif the femalesare morphologicallyindistinguishable and to be different speciesif recognizablydistinct females show significant assortativemating behavior. I suggestthat formswith morphologicallyindis- tinguishablefemales are probablymore closelyrelated than forms in which the femalesare distinct. The criteria are similarto thoseof Short (1969: 90), exceptthat (for thesebirds) the distinctforms linked by specimensshowing somedegree of morphologicalintermediacy are regardedas conspecificif additionalevidence to the contrary, such as assortativemating, is lacking. Followingthe aboveguidelines, all of the forms here includedin V. chalybeataare conspecificby criterion1: The speciescategory V. [unereaincludes the forms codringtoniand nigerrima on the basis of criterion 1, but not the west African mimics of the same song model (the "nigeriae" individuals recordedat Panshanu,Nigeria, for example)becguse of criterion2 (see pp. 261-263). Undercriterion 3, V. chalybeata,V. pdrpurascens,and V. [unerea (and their races) are consideredto be specificallydistinct, even though some interbreedingmay occurin parts of their ranges. The forms called "nigeriae," "camerunensis,"and "wilsoni" are regardedas conspecific(V. wilsoni is the oldestname for this complex) under criterion4 (the femalesof theseforms are indistinguishablein a region where male plumagecharacters form more or lessdistinct clusters, suggesting insufficient isolation for the evolutionbf the distinctivefemales characteristic of someother sympatricindigobirds. The precedingparagraph summarizes part of the resultsof the following pageson variationand relationshipsand is intendedas a guideto this section. In contrastto Nicolai (1964), I think it necessaryto considerboth the morphologicalevidence of interbreedingbetween forms and their songs,not just the mimeticsongs alone, in describingspecies relationships among the viduines. If specieslevel taxonomicgroupings were made simplyon the basisof song, the resulting systematicarrangement of the indigobirdswould have several featuresundesirable in an arrangementintended to be of evolutionaryinterest. First, the rare individualmales that sangthe "wrong" host songwould be groupedwith other specimensof unlike appearance(and presumablyof no closephyletic relationship); for example,a male bird with the morphological 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 211 charactersof V. chalybeatarecorded singingthe song of L. rhodopareia wouldbecome a specimenof "V. purpurascens"for a taxonomistusing simply a songcriterion. Calling this bird "V. purpurascens"might be relevantfor notingits most probablefoster speciesand for predictingthe appearanceof its mate, but this would give no comfortto a systematistwho wantedto have his classificationreflect the phyleticrelationship of the birds classifiedor their similarityin morphologyto otherindividuals. Second, groupings based solely on songwould be of no use in describingvariation in the hundredsof museumspecimens for which no songdata were recorded. Third, and most importantly,a classificationbased solely on songwould be contradictedby any morphologicalevidence of interbreedingbetween forms that mimic different speciesof firefinches. Fourth, such a classificationwould be inherentlyun- stable,since recognition of speciesof the indigobirdswould be basedon a currenttaxonomy of the genusLagonosticta, a group with taxonomicprob- lemsitself and with localor regionaldifferences in songwithin a singlespecies. Strictly applied, a speciescriterion based solely on song would give a classificationin which the green,blue, and purple mimicsof L. rara at Zaria would be one species,but in which the green,blue, and purple mimics of another firefinchwould be consideredanother species of indigobird. And this conclusionwould follow even thoughno independentcriteria of speciesrelationships(assortative mating, morphologicaldistinctness) would be evident. From the field recordingsand collectionsof the V. wilsonicomplex (pp. 257- 261, Table 10) it is clear that different kinds of indigobirdsmay mimic differenthosts in differentregions, that locallyone kind of indigobirdmay mimic more than one host species,and that severalkinds of indigobirdsmay locally mimic a singlecommon host species.Here it would be inappropriateto apply a classificationbased simply on one song: one species.In addition,the playback experimentssuggested different degreesof behavioralattraction of female indigobirdsto the songsof different firefinches,and not an all-or-none responseto the firefinch species.In reality, while they do so in someareas, the indigobirdsof someother areasor over their entire range (in somecases) appear not to behaveas traditionalbiological species. The systematictreat- ment adoptedhere is a compromiseattempting to allow for the variation of the birds in nature, to honor the custom of taxonomists to find names that are mostuseful for referringspecimens to biologicallyreal groups,and to ful- fill the desireof evolutionarybiologists to have the namesof thesegroups reflect evolutionaryrelationships among living things.

ANALYSIS OF VARIATION Male indigobirdsare all similarin appearance.Their breedingplumage is black variouslyglossed with purple, blue, or green. All have white flank feathers which are occasionallyfluffed over the wingsbut which were not usedin 212 ORNITHOLOGICAL MONOGRAPHS NO. 11

TABLE 30 COLOR STANDARDS AND MUSEUM REFERENCE SPECIMENS

Color Museumreference specimens standards Description RBP MRAC AMNH BM(NH) NMR green 4437 RG2153 1928.7.20.248 1933.5.11.86 - blue-green 4444 68207 452275 1918.8.26.76 41771 bluish-green 4530 57692 453398 1932.8.6.759 17201 green-blue 4581 11200 161947 1929.2.18.479 62696 blue 4559 11199 1915.12.24.1720 1945.18.76 62693 purple-blue 4575 54519 - 1911.12.23.3303 49965 purplish-blue 4539 59177 264550 1937.12.19.578 21477 bluish-purple 4525 88307 452320 1911.5.30.442 35933 purple 4521 4614 - 60.12.31.131 46192 any conspicuousdisplay that I observed.Females and malesin non-breeding plumageare sparrowybrown birdswith streakedbacks, brownish or grayish breasts and flanks, and white bellies. Over-all variation in morphological charactersis slight amongthe indigobirds. Charactersused in the studyof morphologicalvariation were body weights, linear dimensions,plumage color, plumagepattern (in females), and color of the bill, feet, and iris. Body weightswere taken to the nearest0.1 g in the field for most of the birds collected.No birds showedsignificant visible de- positsof subcutaneousfat. The main differencesin body weight involved developmentof the reproductivetract in females, but since most females were breedingwhen collectedtheir weightswere comparedwithout adjust- ment of data. Linear dimensionsin mm were taken with calipersin the followingmanner: (1) wing length: chordof unflattenedwing; (2) tail length: distancefrom point of insertionof central pair of rectricesto tip of longest rectrix, exceptin birds where the central rectricesexceeded the othersby 3 mm or more, in which casethe next longestfeathers were measured;(3) bill length: distancefrom anterioredge of nostril to tip of bill; (4) bill width: distancebetween lateral surfacesof rami of lower mandibleimmediately posterior to horny surfaceof mandible;(5) tarsuslength: diagonalfrom posterior midpointof tarsometatarsusat the joint of tibiotarsusand tarsometa- tarsusto distal edge of distalmostundivided scute on anterior surfaceof the tarsometatarsusnear its junctionwith middle toe. Plumagecolor was comparedwith the color of certain "color standard specimens"in adult male plumage. The colorsof the color standardspeci- mensranged from greenthrough blue to purplish-blackand are representative of the rangeof colorsof indigobirdsespecially in the southernhalf of Africa. The color standardspecimens in UMMZ, alongwith matched,similarly colored specimensin other museumcollections designated "museum reference 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 213 specimens,"are listedin Table 30; thesespecimens may be consultedfor aid in identifyingindigobirds by the schemeused here. Brightnessof glossof adultmale plumage was judged in comparisonwith threecolor standard specimens (RBP 4437 = glossy,RBP 4575 = medium, and RBP 4521 = dull). The colorsof remiges(except the inner three secondaries,which are blackish in nearlyall populations)range from lightbrown through dark brownto black (RBP 4521 = light brown,RBP 4559 = dark brown, and RBP 4884 = black); comparisonof specimenswith thesebirds permittedthe recognitionof five categoriesof color (includingintermediates) of the flight feathers.The major plumagevariants are shownin the frontispiece. Color comparisonswere made under differentconditions including direct sunlight(whenever available--including examination of type specimensin the British Museum (Natural History)), north windowskylight, fluorescent lighting,and a Macbeth SuperColor MatchingSkylight (Model BX 848A). The degreeof differenceand the glossinessamong the color standardspeci- mensvaried with lightingconditions, but in all of the lightingconditions used the color standardspecimens retained the sameranking order for color and intensityof glossand thus comparisonis possibleof specimenscompared at different times and places. For colorimetricanalysis of bodyplumage I alsoused a Bauschand Lomb Spectronic505 recordingspectrophotometer, equipped with a visible reflectance attachment,at the Museumof Natural History, Universityof Kansas. Flatnessof the 100 percentline wasmaintained within limits of 1 percentand flatnessof the 0 percentline was maintainedwithin 0.25 percent. The mono- chromaticbeam spot was reduced by a diaphragmto a diameterof 13 mm at the sampleport. White referencestandards of 100 percentreflectance were preparedfrom bariumsulfate in a Bauschand Lomb powderpress. All speci- mensused in colorimetrywere unwashed;the mostobviously soiled or worn specimenswere rejected. Males were positionedwith the reflectancebeam recordingthe breast;care was taken to useonly the specimenswith no visible brownish,sparrowy feathers or pale gray featherbases. Femaleswere posi- tioned for both back and breast readings. The slight differencesin body positionin replicaterecordings showed no obviousdifferences with position- ing of the adult males,but replicatereadings of females(especially of the streakedbacks) varied, and the readingsI usedwere the lowestof the two or three taken for each bird. Previousspectrophotometric studies of bird plumage have made use of I. C. I. (= InternationalCommission on Illumination) trichromaticcoefficients (Selanderand Johnston,1967: 220). However,these coefficients, designed by psychophysicistsin 1931, are inappropriately applied to iridescent,glossy plumageor otherplumage which involves more than a singlepeak reflectance or indeedany deviationfrom a ratherspecific kind of spectraldistribution of 214 ORNITHOLOGICAL MONOGRAPHS NO. 11

I I I I I I I

4437 m

• • 45304444 • f43

x • 4581 / oLU 43 f43

• 3 4 zLU 45754559 • f

• 3 • 4539 /

• 452,• •

400' ' 560 ' 660 ' 700' WAVE LENGTH (rn•) Figure 34. Spectrophotometricdata showingthe color of the color standard specimens used in categorization of indigobird plumage color. Specimen numbers refer to birds listed in Table 30. Note the decrease in blue and green reflectance and the increase in red reflectance in the purplish specimens4539, 4525, and 4521 below compared to the greenish specimensabove; blue birds are in the middle of the figure. light (Judd, 1933; Hardy, 1936; Fox and Vevers, 1960). Different curves may give the sameI.C.I. values. As is evidentin the spectraldistribution of reflectancevalues in Figure 34, somemale indigobirds,especially the purplish specimens,have double peaks of maximumreflectance, one at the blue end 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 215 of the spectrumand one at the red end. Applicationof the trichromaticco- efficientsin thisinstance yields a computeddominant wave length (Xd) in the yellow-greenrange, an inappropriatevalue for a purplishbird havinga spec- tral reflectancewhich is lowest(not highest)within this range. Readingsof the spectralreflectance curve were therefore analyzed in the followingmanner for the male breedingplumages. (1) The dominantwave length d wasread directlyfrom the highestpoint of the left hand side of the reflectancecurve, and where this regionhad a broadplateau and no peakwas discerned, d wastaken from the midpointof the plateau. (2) Brightnessb wasdetermined by physicalintegration with a planimeter of the areaunder the reflectancecurve from 400•640 mt•. (3) An indexof colorpurity c wascalculated directly from the difference betweenthe reflectancevalues for d and for the lowestreading of the curve for eachbird; thusthe curveswith the moreclearly defined peaks of reflectance (greatestdeviation from flat gray) showthe higher indicesof color purity. For female indigobirdsthe reflectancevalues of the back and the breasteach provided two setsof values: (1) brightness,b, calculatedas for males,and (2) the slopeof the reflectancecurve, s, calculatedfrom the differencein reflectanceunits of the curveat 400 mt• and 640 mt•. The reflectance curve was generallyclose to a slopingstraight line for back and for breast,and the major differencesin colorbetween females are apparentfrom the400 mt•intercept and the slope of thecurve (see Figure 15). The physicalbasis of the differencesin color and glossof the breeding plumagesof maleindigobirds is duein part to the developmentof the barbules, whichare ratherbroad in the glossygreen feathers of "nigeriae"and are most narrowin the dull, blackishfeathers of "nigerrima,"as studiedmicroscopically by Auber (in White, 1963a). Traylor (1966: 70-72) reportsconsiderable variation in the color of wing and tail featherswithin individual birds. However, the variation in color of the remigesappears to me to be associatedmainly with a slightpost-mortem fadingof the exposedportions of the vanesand a fadingfrom repeatedex- posureto light in the museum.No museumspecimens collected in the period 1955 to 1968 showedany variation in this respect. Some birds showed occasionalblack, even glossyblack, rectrices,particularly the centralpair, whichwere sometimeselongated. One captivemale V. chalybeatathat I re- ceivedin normalbreeding plumage in 1968 moltedtwice in the followingyear, andalthough the new breeding plumage was normal and like thatof theprevi- ousyear in mostrespects the new setof rectriceswas markedly different. The centralpair of the new rectriceswere 8 mm longerthan the othersand were glossyblack, whereas all wereabout the samelength in the previousbreeding plumageof thebird (Figure35). In 1970 andin 1971the tail grewnormally 216 ORNITHOi OGICAL MONOGRAPHS NO 11

Figure 35. Captive male I/idua chalybeata ultramarina with elongated central rectrices. The arrow indicates the length of the normal tail feathers. Photographed by J. R. Purdue. again with the central rectricesdull blackish and no more than 1 mm longer than the lateral tail feathers. The hormonal condition of the living birds may have changedbetween the times of morphogenesisand pigment determination of these various feathers in successiveyears. The color of ingrowing body feathers in viduine finches is well known to be associatedwith luteotropic hormone (LH) and is sometimesused as a bioassayfor LH (Ortman, 1967; Ralph eta!., 1967a, 1967b; Hall, 1969). Presumablyin the normal molt cyclethe seasonalplumage change is causallyassociated with variation of LH or similar pituitary gonadotropinsin the blood. My captive indigobirdsregu- larly molt the inner two pairs of rectricesoverlapping in time with the in- growth of the black breedingplumage of the body and the presumedassoci- ated increasein LH. (These inner two pairs of rectricesare molted twice in 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 217 a year,as in my captivemale V. paradisaeaand V. hypocherina,whereas the outerrectrices are moltedonly once.) Variationsin thephysiological condition of the bird duringthe time of molt thusmay be responsiblefor the occasional black flight feathers. Colorsof the softparts were recorded at thetime of collectionand are also taken from specimenlabel data for other birds. Standardcolor references havenot been used by fieldcollectors, and some workers' terms such as "flesh" leavein doubtwhether others would regard the bird as pinkish,orangeish, or whitish-looted.Most male indigobirdshave feet which are unambiguously eitherwhitish or reddishwhen they are alive. Thesecolors change rapidly after death,and in somemuseum specimens for whichbill or foot colorwas recordedon the label at a considerabletime after death the specimendata may be of questionablevalidity. Post-mortemchanges may occurin a few hours. At Maun, Botswana,18 white-billedV. chalybeatathat I shot all had bright pink to reddish-pinkbills betweeneight hours and three days after death;after three days they were again white (I did not noticesuch a change in all white-billedindigobirds). Specimensmore than two monthspost-mor- tern seldomhad the samefoot color as they did when collected,though I coulddiscern a uniformlydarker hue after one year in the then-brownishfeet of V. chalybeataamauropteryx (red in life) than in the then-brownishfeet of V. purpurascens(whitish in life). After a few yearsthe distinctivereddish bill colorof specimensof V. c. amauropteryxfades to "horn" colorwhile the whitish bills of other forms often darken to the same horn color. Because colorsof bills and feet were recordedwithin minutesof death in my field work I placegreater faith in the validityof color data for thesespecimens than in the discordantdata of old museum specimens. All availablemuseum specimens of indigobirdswere examinedin the presentstudy. These totalled 1,865, including 302 collectedduring the field work and depositedin the Museumof Zoology,University of Michigan. Of the birds examined, most (1,476) were males in partial or complete breeding plumage. Most specimensof femalesand juvenilescollected in areaswhere more than a singletaxon occursare unrecognizableto species. Specimens collectedpreviously included only a few femalestaken with males and no meaningfulstudy of morphologicaldifferentiation in femalesof known form had been possible. Females of most kinds of indigobirdswere unknown (White, 1962, 1963a; Traylor, 1966). In the field work I collected116 adult females,all but 10 of them with a male at the call-site,and this sampleper- mits a comparisonof femalesof most of the morphologicalforms of indigobirds. Specialattention was givento the femalesin the studyof speciesrelationships,since to showwhether different forms that live togethercomprise morethan onebiological species it is necessaryto matchup eachbird with its mate. As in the paradisewhydah species complex (Vidua paradisaeaand its 218 ORNITHOLOGICAL MONOGRAPHS NO. 11 relatives;Payne, 1971 ), the indigobirdspecies often do have distinctfemales. Juvenilesof two indigobirdspecies were also collected;previously the only form for which juvenilescotfid be identifiedwas V. chalybeata;it lives in regionswhere no other speciesoccurs, and juvenilesfrom theseregions are naturally referableto this species.

Vidua chalybeata The Village Indigobird,Vidua chalybeata,is the most widespreadindigobird in Africa and extendsfrom Senegalacross the sub-Saharansavannah to eastAfrica and southbeyond the tropics. It is a geographicallyvariable specieswhose forms have only recentlybeen recognizedas conspecific(Traylor, 1966, 1968). Eight formshave beendescribed as subspeciesor species(aenea, amauropteryx,centralis, chalybeata, ignestii, neumanni, orientalis, and ultramarina). Of these,three are synonyms( aenea= chalybeata,orientalis = centralis, and ignestii= ultramarina). In additionto the five recognizablesub- speciesmentioned above, a sixth occursin southwesternAfrica and is here described.Additional trends in geographicvariation are better discussedthan given taxonomic names. Measurementsof male and femalespecimens of V. chalybeataare givenin Table 31. Males and femalesare generallysmallest in west and northern Africa; the largest birds are V. chalybeatacentralis in central east Africa. Along the east coast are small birds with reddishbills, and inland in south- easternAfrica the red-billed birds are larger. Birds in Ngamiland and in South-WestAfrica and southernAngola are againlarger than their red-billed counterpartsin southeasternAfrica. Two main groupsof V. chalybeataare evident,with respectto male breeding plumage,glossy birds in west and northernAfrica and dull birds in east and southernAfrica (Figure 36; see also frontispiece). In the westernand northernbirds the color changesgeographically from green and blue-green V. c. chalybeatanear the west coast (all variantsof green,blue-green, and blue occur in some single localities in Senegal Richard-Toll Station d'- Ornithologie,Diourbel, and Kirtaouna) to blue birds in most of west Africa. Mali birdsare intermediatein color and the divisionbetween bluish-green or green-blueV. c. chalybeataand the blueform neumanniis rather arbitrary. In northeasternAfrica no abrupt break occursin color from neumannito the more purplishV. c. ultramarinaof Ethiopia. Color in all of theseglossy birds is purer than in easternand southernAfrica, with mean b greaterthan 1.0 in mostpopulations. Male breedingplumage in the remainderof Africa is rather dull bluish. Color of flight feathersvaries in a similar geographicmanner. All northern malesfrom westAfrica throughEthiopia are black-winged.Even in birds of worn plumagemolting into the sparrowyoff-season plumage the flight feathers 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 219

SENEGAL

MALl

NIGERIA

SUDAN

ETHIOPIA

KENYA

RHODESIA

BOTSWANA

TRANSVAAL

3 4 5 6 400 450 500 0 I 2 • b d Figure 36. Color values of male breeding plumage in Vidua chalybeata: brightness (in2); c = index of color purity; d = dominant wavelength (mtz). Vertical lines give means, horizontal lines show ranges, and boxes are t.95% were blackishand were distinctlydarker than the correspondingfeathers of the pale-wingedforms of V. wilsoni. In fresh plumagea pale brown leading edge on thesefeathers appears, but even in the first month of the breeding seasonthe wings of live V. chalybeatain the field in Nigeria were much blackerthan the all-brownwings of other indigobirds;this edgewears off well before the breedingplumage is molted. Wing and tail color in central-east African V. chalybeatais medium or dark brown; in one bird from Nyeri these feathers were more blackish, though the bird otherwise resembledthe dark brown-wingedV. c. centralisof Kenya. V. c. amauropteryxare slighdypaler- winged, on the average. Foot color has been recordedon specimenlabels in more than 35 different terms by various collectors,including 25 color terms in Kenya alone. Most specimensof male V. chalybeatahad red, reddish,orangeish, or coral feet. Of the exceptionalcolors a number were indeterminate: "dirty livid horn," "hornschwinlich,"and "flesh"may indicatenearly any color. Othersinscribed on the labelssuch as light brown,yellow, or white are puzzlingas the plumage of these birds does not differ from others with reddish feet; post-mortem changesin color may have resultedbefore the colorswere recorded. A few were said to have black feet. Because of the observed seasonal variation of foot color in the Transvaalbirds taken before,during, and after breedingit is possiblethat the pale foot colorsnoted by somecollectors were due to collect- ing outsideof the main breedingseason. 220 ORNITHOLOGICAL MONOGRAPHS NO. 11

eq r--

eq Iq:• eq 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 221

•z<• 222 ORNITHOLOGICAL MONOGRAPHS NO. 11 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 223 224 ORNITHOLOGICAL MONOGRAPHS NO. 11

•0• 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 225

I I I I I 226 ORNITHOLOGICAL MONOGRAPHS NO. 11

e• Io• 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 227 228 ORNITHOLOGICAL MONOGRAPHS NO 11

Figure 37. Geographic variation in plumage pattern in female V. chalybeata. Specimens from left to right: RBP 4297, Merensky, Transvaal; RBP 4522, Monkey Bay, Malawi; RBP 4609, Maun, Botswana; RBP 4698, 34 mi. ESE Kisumu, Kenya; FMNH 83914, Debra Werk, Gojam, Ethiopia; RBP 4918, Zaria, Nigeria.

Bill color in museumspecimens of V. chalybeatawas recorded as whitish with few exceptions. The steely-glossedgreenish-blue males from southern Africa and along the east coasthave reddishbills; theseare paler in molting birds. The other exceptions,including a male collectedby van Somerenin Nairobi noted as red-billed, are probably birds that underwentpost-mortem changes.A Sudanbird noted as black-footedwas also saidto be black-billed (ANSP 93323); the bill and feet of the dried museum specimenlook no blacker now than the "horn-coloured" dried bills and feet of other Sudan V. chalybeata. Variation in plumagecolor of femalesparallels that of males in geographic pattern. Sub-Saharanbirds from semi-aridlocalities well north of the range of other indigobird species,and also known femalesfrom Nigeria which were shot with male V. chalybeata,are somewhatgrayer and less heavily streaked above than are female V. chalybeata from east or southernAfrica (Figure 37). In color the east African and southernAfrican femalesare not dearly separableexcept for birds from the Okavango regionwhich are notably grayer (less rufous) above and below than are neighboringV. c. amauropteryxfrom 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 229

back breast breast back

NIGERIA KENYA BOTSWANA RHODESIA TRANSVAAL

I • I o IO 20 30 0 I0 b s Figure 38. Color values of plumage of adult female V. chalybeata: b = brightness (in2), s -- index of color.

Rhodesiaor Transvaal(Figure 38). No strikingdifferences in femalecolor or pattern are evidentin plumagein the different racesof V. chalybeata. Bill and foot colorin the differentforms vary geographicallylike the colorsof the males. Descriptionso[ [orms.--The followingare brief descriptionsof the recog- nizablesubspecies of V. chalybeata.These names summarize only part of the geographicvariation, mainly variation which is evident over wide areas in relatively discontinuousform. Other details of variation are to be seen in Table 31.

Vidua chalybeata chalybeata (MQller) Fringilla chalybeataP. L. S. MQller, 1776, Des Ritters Carl yon Linne... vollstandiges Natursystem... , p. 166: "Brasilien" (Senegal substitutedby W. L. Sclater, 1930, Syst. /Iv. /iethiop., p. 807). Holotype: not located. syn. [Hypochera]aenea Hartlaub, 1854, J. f. Orn., 2, p. 115: Senegambia.Holotype: not located. Malesin breedingplumage are very glossyblue-green to greenin color, a few are green-blueor blue. Primaries,secondaries, and rectricesare black, borderedin freshplumage on the marginwith buff. Brownishfeathers evi- denfly retainedfrom an earlier plumagesometimes occur, especiallyon the major covertsand alula. Bill color is white, feet are reddish-orange,duller outsideof the breedingseason. Iris is brown. Femalesfrom centraland northernSenegal, where all male indigobirdsare the form V. c. chalybeata, are brown above and streakedwith dark brown. The wingsare brown. A femalefrom Senegalwas recorded with bill horn and feet reddish-brown. 230 ORNITHOLOGICAL MONOGRAPHS NO. 11

Malesin sparrowyplumage resemble females except for theirblack remiges andrectrices. The juvenal plumage is unstreakedbuffy brown in Senegalspec- imens;young birds in postjuvenalmolt retainthe brownjuvenal flight feathers for someweeks after acquisition of the streaked,sparrowy plumage. Foot and bill coloris gray until the time of postjuvenalmolt. Remarks.--Thecomplete intergradation between the bluestand the greenest black-wingedSenegal indigobirds indicates that the greenform "aenea"is not a form geographicallydistinct from the morebluish Senegal birds, and thusit is not recognizedtaxonomically as a subspecies;nevertheless, the presenceof greenindividuals of V. chalybeatais a real feature of indigobirdpopulations at the western end of Africa. From the descriptionof chalybeatagiven by Miiller it is impossibleto tell what kind of indigobirdwas beforehim. Foot and bill color was "blassroth" (reddish),and thismight refer to the red-billedindigobirds of southernAfrica or to any otherindigobirds whose bill color changedafter death. The plumage was "dunkel stahl/arbig-blau,"and the name chalybeatarefers to a blue iron mineral,chalybite; blue might describeany of severalforms of non-purplish indigobirds.The localityis unknownand no typeshave been designated. In spiteof the inadequacyof the initial description,it seemsclearly advisable to retainthe name in thesense of its useby nearlyall authorsof the pastcentury to refer to the bluishbirds from Senegal,a redesignationof a type locality "Brasilia"made withoutany explanationby Sclater. Vidua chalybeata neumanni (Alexander) Hypochera neumanni Alexander, 1908, Bull. Brit. Ornith. Club, 23, p. 33: Yo, near Lake Chad [northern Nigeria]. Holotype: BM (NH) 1911-12-23-3306, male in breeding plumage. Males are glossyblue, rangingfrom green-blueto purple-blue,and they haveblack flight feathers.Bill coloris white;live birdsin breedingcondition have orangefeet. Iris color is brown or dark brown. Females referred to this form (Table 31) are identified on the basis of the distributionof male neumanniand its host along the Nile in centralSudan and alsoon the basisof my collectionsfrom singingmale neumanniin north- em Nigeria at Zaria, Numan, and Kiri. These femalesare not clearly dis- tinguishablefrom femaleV. c. chalybeataexcept perhaps by a buffier color in the Nigerianbirds. Bill color variesin femalesin Nigeria. The upper mandibleof breedingseason females ranges from brown to horn-brownand grey-horn,and the lower mandibleis more often pale horn-creamor cream. Foot color in Nigerian femalesis orangeish.Color of the soft parts was not noted in breedingfemales from Sudan. A male (RBP 4897) in sparrowyplumage taken in a flock of eight adult malesin a harvestedmillet field at Sokotois similar to femalesin plumage exceptfor hisflight feathers, which are black.The bill wasorange-horn-brown 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 231 aboveand gray-creamybelow, the feet werelight orange,less bright than in males in breeding plumage. A juvenilefemale from Shendi,Sudan (AMNH 452469) is unstreaked brownabove and buffy below; bill andfoot coloris recordedas stoneand the iris as brown. Remarks.--V. c. neumanniintergrades with thewestern form chalybeataas birdsin Mali andFouta Djalon are intermediatein colorand in size(in wing lengthmore like thelarger Senegal birds). It alsointergrades with ultramarina, in the east. However,the area of northernAfrica inhabitedby glossyblue indigobirdsis sufficientlylarge (morethan 2,000 milesin length)to permit recognitionof neumannias a distinctiveform whichis relativelyuniform in appearancethroughout the area. Vidua chalybeata ultramarina (Gmelin) Fringilla ultramarina Gmelin, 1789, Syst.Nat., 1, pt. 2, p. 927: Abyssinia. Type: not located. syn.Hypochera ignestii Moltoni, 1925, Atti Soc. Italiana Sci., 64, p. 46: Dintorni di Gondor, Abissinia. Type: Museo Milano 24275, male in worn breedingplumage. The colorof manymales in Ethiopiais purple-blueor purplish-blue(see Table30) andis indistinguishablefrom color in somebirds taken 2,000 miles to thewest. OtherEthiopian birds are bluish-purple.The varietyof plumage colorsin Ethiopianbirds subtly parallels the situationof colorvariants in V. c. chalybeatain Senegal.Wing and tail coloris black. Birdsof the Ethiopian plateauaverage significantly larger in winglength than do birdsfrom Sudan to Nigeria, and it is on the basisof size as well as color variation that neumanniis regardedhere as taxonomicallydistinct from ultramarina.Bill color is white and feet of breedingmales are reddish. Femaleindigobirds from Ethiopia (excludingspecimens from northwestern and western areas where the bluish form of V. wilsoni also occurs) are included in Table 31 as ultramarina and are so identified on the basis of distribution. No females have been collected with known male ultramarina. Theseprobable females of ultramarinaare similarin plumageto the females of neumanni. No colors of bill, feet, or iris were noted on specimenlabels. Males with someglossy black feathersand alsoa partial sparrowyplumage have a color and patternof the brownishplumage similar to that of female ultramarina. Juvenal plumage is not known. Vidua chalybeata centralis (Neunzig) Hypochera chalybeata centralis Neunzig, 1928, Zool. Anz., 78, p. 113: Kissenji [Ruanda]. Holotype: Berlin Museum 278, male in breedingplumage. syn. Hypochera ultramarina (Gm.) var. orientalisReichenow, 1894, V6gel Deutsch- Ost-Afrikas, 3, p. 188: Paregebirge[_-- Pare Mountains, north-easternTanzania]. Holotype: not located. 232 ORNITHOLOGICAL MONOGRAPHS NO. 11

Male breedingplumage is dull to medium-glossedblue to purplish-blue. The wingsand tail are dark brown. Within the rangeof this form (Kenya and Tanzaniaexcept coastal areas, Uganda, Ituri, Kivu, Ruanda,Buruntil, and probablyeastern Katanga and Mweru Marsh of northeasternZambia) are the largestmales of the species,with meanwing lengths of morethan 67 mm in Ruanda,Burundi, Uganda, and Kenya. Bill coloris white, foot color is reddish, and the iris is dark brown. Femalesare grayishbrown, streaked with darkerbrown above; the underparts are grayishwith the belly white. Femalesrecorded in Table 31 are ascribedto this form on the basis of distribution (no other form of males is known to occur) for Ruanda, for Burundi, and for Nairobi, Kenya, and on the basisof birdscollected with singingmales of centralisin Kenya. Six birds fromKenya taken with malecentralis were all withinthe areafrom Kisumuto 34 miles east of Kisumu (RBP 4693, 4696, 4698, 4728, and LACM 56626) and one was taken at Olorgesailie(RBP 4684). In color and pattern these femalescannot be differentiatedfrom somepopulations of southernAfrica. Kenyafemales taken with malesof thisform havebill colorhorn to brown above,cream below, and foot color varyingfrom pinkishgray and darkish gray to orangeflesh and dark horn. The four that I collectednear Kisumu all had feet of pinkishor orangeishhues, as did a femaleseen at a centralis call-siteat Sigor,but the femaleshot at Olorgesailiewas noted as darkish- gray footedand was layingwith an eggin the oviduct. Iris coloris dark brown. Malesof thisform whichare moltingand hence show some definitive, iden- tifiable breedingplumage are markedvery like the femalesin the sparrowy nonbreedingplumage. No adequatelyidentified juveniles are known. As pointedout by Wolters(1960: 22) thename Vidua chalybeata orientalis (Reichenow,1894) is preoccupiedby Vidua paradisaeaorientalis Heuglin, 1871, if theseforms are unitedin the genusVidua, and hencethe morerecent name Vidua chalybeatacentralis (Neunzig, 1928) is the appropriateone.

Vidua chalybeata amauropteryx (Sharpe) Hypochaera amauropteryx Sharpe, 1890, Cat. Birds Brit. Mus., 13, p. 309: Rusten- berg, E. Transvaal [= Rustenburg,western Transvaal]. Holotype: British Museum (Natural History) 75-10-7-32, male in breeding plumage. Males of Transvaal, northern Zululand, Mozambique, Rhodesia, Zambia (from the Zambezi River below the falls northward to Chilanga and the LuangwaValley), and Malawi are green-blueto purplish-blue(the latter in more worn specimens)in colorwith a dull to mediumgloss. Wings are mediumbrown, often slightly paler than in eastAfrican centralis. Bill and feet are salmon-pinkto orange. Iris color is dark brown. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 233

Female amauropteryxtaken in Transvaalare slightlymore reddishthan those in Rhodesiaand Malawi. A few females are rather dark gray below, especiallyin a layingfemale (Figure 37) withblackish feathers from Monkey Bay (RBP 4522) thatI hadidentified in the fieldas a moltingmale. Bill and foot colorare pinkish. Iris coloris dark brown. Femalesidentified here as amauropteryx,with threeexceptions, are birdsthat were with singingmale amauropteryxwhen collected; the otherthree are tikewisedistinctive in having the same red bill as known mates of male amauropteryx. Malesin sparrowyplumage are very similarto the females,judging from birdsin molt whichshow both the dull bluishbreeding plumage and the nonbreedingplumage. One specimenis knownof amauropteryxin juvenalplumage (RBP 4763), a femalejust beginningthe postjuvenalmolt and collectedat Merensky, Transvaal,on 21 June 1967. The bird was shotwith a family groupof six L. senegala.The upperpartsare tawny,lightly streaked on the upperback with darkerbrown. The breastis buffy, the belly is white.The reflectivetubercules of themouth were completely resorbed. Bill andfoot colorwere pink, asdark as in a male V. c. amauropteryxin very worn breedingplumage collected on the same day. Remarks.--The small blue indigobirdsalong the coast of east Africa re- semblesouthern amauropteryx in their red bill color,although plumage gloss in somebirds is lessgreenish than in the birdsfrom the south. Bill color of birds seennear Dar-es-Salaamis red (Nicolai, 1967:311), and specimensof coastal red-billed birds have been collected south at Mikindani and north at Dar-es-Salaamand Bagamoyoin Tanzaniaand at Malindi in Kenya,and all of thesecoastal birds are small. They overlapthe amauropteryxfrom Zambia and Malawi in size and are barely differentin mean wing lengthfrom them. Plumageof the coastalmales in breedingdress suggests intergradation between amauropteryxand centralis. The single specimenavailable from Abrona, Somalia,a rather greenishbird, is similar to a Malindi bird in its brown wings and smallsize, and althoughits bill color was not recorded,the bird is in size and plumagemost like amauropteryxof coastaleast Africa. The red-billedpopulation at Malindi appearsto be local but stable. For many yearsthe only recordsof red-billedindigobirds at Malindi were sight observations(Pitman, in Chapin, 1954: 568). Two maleshave been taken recently(LSU 26921, RBP 4670) and the onefemale (RBP 4671) hasa bill reddishbelow and gray above. The female flew to the call-siteof the red- billed male at the Malindi police station,but shewas not in breedingcondition. Althoughthe coastalbirds are somewhatdistinctive in size, they also showrecombination of othercharacters (plumage color) of the two adjacent subspecies,and they need not be separatelynamed. 234 ORNITHOLOGICAL MONOGRAPHS NO. 11

Vidua chalybeataokavangoensis new subspecies Holotype: UMMZ 217,254 (RBP 4583), male in breeding plumage, collected 14 April 1967 at Maun, Botswana, at the edge of the Okavango Swamp. Description of type: Plumage dark, dull blue, wings dark brown, bill white, feet red. Wing 66 ram, tail 40, bill length 6.3, bill width 6.0, tarsus 15. Adult malesin breedingplumage are green-blueto blue with a medium- dull gloss. Wingsand tail are mediumbrown to dark brown. Feet are red, bill color is white, the iris is dark brown. Femalesare grayer(less bully) aboveand belowthan are femaleamauropteryx from Transvaal. Wing and tail are medium brown, the iris is dark brown, foot color is pinkish, and the bill is brown-horn. All of the females describedfor Botswanain Table 31 were shot within ten miles of Maun, most of them with males. One male from Maun has a few retained brown feathers in his breeding plumage,and these closelyresemble the sparrowyfemale plumage of this form. Four juvenilestaken near Maun from family groupsof L. senegalaare pale unstreakedbrown above,bully on the breast,and whitishon the belly. These areindistinguishable in plumagefrom juvenileindigobirds in otherareas. The bill color was whitish with the upper mandible gray. Feet were pale pink (RBP 4633, 4635) or dirty pink (RBP 4605) in birdswhich had not started the postjuvenalmolt and had retainedtraces of the reflectivetubercules at the base of the bill. Remarks.-•Traylor (1965: 381; 1966: 61) drew attentionto a few white- billed indigobirdsin southeasternAfrica, includingbirds from the Okavango region, and I visited Maun specificallyto study thesebirds. Traylot (1966: 61 ) thoughtthat the white-billedbirds were simply"occasional specimens... in which the bill is white insteadof red." The large sampleof birds taken at Maun and the scatteringof specimensfrom neighboringareas, including the specimenscollected in Botswanaby Traylot, showsthat nearly all of the birds of this region (V. c. okavangoensis)are morphologicallydistinct from the red-billedV. c. amauropteryxform in which Traylot includedthe birds from northwesternBotswana. Of the 17 males that I collectedat Maun and Shorobe, 16 were white-billed and one was red-billed. At least six other adult males were seen,all with white bills. Of eight additionalmuseum specimens from Nokanen,Sepopa, Maun (Boro), and Lake Dow (Kedia) with bill color noted on the label, all are white-billed, as is the dried bill in two additional birds from Khomo (Botletle River, Lake Dow) and one from Nata which was white-billedaccording to White (1962: 23). Thus the total sample of 34 males from northwestern Botswana indicates all but one to have white bills. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 235

The Okavangoarea birds differ further in being longer wingedthan male V. c. amauropteryx.The gray colorof the femaleplumage also is distinctive. The Maun birdsresemble the blue-glossed,red-legged birds from southern Angola at Huila and Gambos;southern Angola is geograplficallylinked by the CubangoRiver to the Okavangoof Botswana.Other Angolaspecimens of this form are from Fazendado Cuito, Quifandongo,and Serra do Mange. Andersson's(1872) specimenfrom Ondonga,South-West Africa, also be- longswith this form. This bird (BM[NH] 77-7-1-425) has usuallybeen calledamauropteryx but the originalbill colorwas not notedon the label and the wing sizeis large (67 mm). V. c. okavangoensisalso extends north of Ngamilandthrough Barotseland (Kalabo, Kazangula) as far as Balovale, as breedingmales from BalovaleDistrict (Lukulu) are white-billed. At Balovale Boma, P. L. Britton collecteda male (now in UMMZ) in nonbreedingplumage, quite gray, resembling the Maun femalesin color. Two femalestaken by M. A. Traylorat theLuashi-Zambezi confluence in Barotse- land alsoare quitegray and are includedas tlfis form in Table 31.

Vidua purpurascens Dusky Indigobirds(Vidua purpurascens)of southernand easternAfrica are characterizedby their pale, whitish feet and their large size; wing length averagesmore than 65 mm in all areas. All individuals that I heard and recordedin the field and also all ten males heard in captivity mimicked the callsand songsof L. rhodopareia.Within the speciesconsiderable if not con- spicuousgeographic variation occursin size and color, wlfich rangesfrom purplish-bluethrough glossy purple to dull purplish-black.The nature of the variationin this speciesis suchthat its biologyis better understoodif no subspeciesare proposed. A summaryof the geographicalvariation (Table 32) showsthat birdsfrom Transvaal, Rhodesia, and Zambia are slightly larger than birds from more northern and eastern populations. Transvaal birds tend to be dull bluish- purple,the birdsof the higherelevations (over 3,000 feet) in Rhodesiaand Zambia to be brighterand more bluish,and the birds of hot, low valleysof the Zambezi,Luangwa, and Sabi rivers to be lessglossy (more matt-black). In Rhodesiathe brightestand bluestmales known are from the easternhigh- lands around Umtali and Melsetter, and in Zambia the birds of the plateau are slightlybrighter than Zambezi and Luangwavalley birds. Birdsintermedi- ate in color between the bluish V. purpurascenstaken at Sigor, northern Kenya, and the more purplishTanzania birds occurin southeasternKenya and northeasternTanzania. All of these differences are most apparent in specimensin fresh, unworn plumage. 236 ORNITHOLOGICAL MONOGRAPHS NO. 11 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 237 238 ORNITHOLOGICAL MONOGRAPHS NO. 11

Female indigobirdscollected and identified by associationwith breeding malesof V. purpurascensshow no significantgeographic trends in size. The femalesfrom Sigor,Kenya, are slightlypaler than are Transvaalfemales but no charactersare known that would permit identificationto sourceof specimens from these areas. Bill color is whitish to brownish-white and foot color is pale. I mostoften recorded this as purplish-whitish-flesh(not orangeishor reddish). It appearspossible to differentiatebetween breeding female V. purpurascensand other specieswhich occur with it in Kenya, southernMa- lawi, andRhodesia by the foot colorof live birds. Femalesof V. purpurascens from theeastern Transvaal are not dearlyseparable from femaleV. [. [unerea in the same area on the basis of foot color. Description.--Onlya singleform hasbeen described for this species,which may be characterizedas follows. Vidua purpurascens(Reichenow) Hypochera purpurascensReichenow, 1883, J. f. Orn., 31, p. 221: "Usequa, Lindi" [Usegua,northeastern Tanzania]. Holotype: Berlin Museum 30709, male in breeding plumage. Males in breedingplumage are bluish,bluish-purple, purple, or purplish- black. Primaries, outer secondaries,and rectricesare brown. Bill is white. Feet are pale whitishto purplish-white.The iris is dark brown. Adult femalesare grayish-brownabove with dark brown streakson the bodyfeathers of the upperparts;underparts are brownish-gray,belly is white. Bill color is variableranging from light brown to whitish. Feet are slightly darker than in breedingmales and are flesh-whiteor pale purplish. Iris is dark brown. Molting maleshave a sparrowynon-breeding plumage similar to the female plumage. A juvenile (RBP 4767) taken at Merensky, Transvaal,on 22 June 1967 from a family group of Lagonostictarhodopareia is lightly streakedabove, buffy on the breastand whitishon the belly. Bill colorwas gray, white below andthe feet werecream-gray. Another juvenile probably of thisspecies judging by bill and foot color (RBP 4764) taken here on 21 June 1967 is in molt. The juvenalfeathers are similarto the otherbird, the feet are light gray, and the bill was gray-brownabove, white below. As the youngof V. c. amauropteryx are pinkish-billedduring postjuvenal molt theseyoung are thoughtto be V. purpurascens;in plumagecolor and size they are indistinguishablefrom juvenilesof V. chalybeata. Remarks.--Morphological differentiation of V. purpurascensfrom other speciesdepends largely upon the fidelity of the collectorin recordingfoot coloron the spedmenlabel at the time of collection.Several specimens from Transvaal,Zululand, and Mozambique are of questionableidentity inasmuch as foot color is not available;these birds are not includedin the present 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 239 analysis.In theseregions V. purpurascensdiffers morphologically in only a singleknown characterfrom V. [unerea,the feet are whitishinstead of red. Foot color is lesshelpful as an aid to identificationof females;among V. [unereaat Tzaneensome also had palefeet and are not distinguishablemor- phologicallyfrom V. purpurascensat Merensky. In Kenya male V. pur- purascensclosely resembles V. chalybeatacentralis and differs from it mor- phologicallyin foot color (whitishin purpurascens)and in wing color (paler brown primariesin purpurascens).Females in Kenya can usuallybe distinguishedfrom V. chalybeatacentralis by foot color. Two specimensrecorded as having"red" or "red-brown"feet are in all otherrespects representative of V. purpurascensand are tentativelyincluded here, NMR 26427 from Chalimbana, Zambia, and TM 25706 from Bindura, Rhodesia.The foot colormay havebeen recorded erroneously. In all of the 44 breedingmale indigobirdswith purplishplumage and pale wingsthat I collectedin Transvaal,Rhodesia and Malawi, foot color waspale (no closer to red than a pale mauveor pinkishcolor), and was distinctfrom the red feet of other indigobirdsin the same areas. The typelocality of purpurascensis a ratherlarge area, not a specificpoint. Reichenow(J. f. Orn., 1883, vol. 31, p. 221) givesthe localityas "Usequa, Lindi." "Usequa" has usually been interpreted(Friedmann, 1960: 60) as Usegua (= Useguha), a region north and inland from Bagamoionear the coastof northernTanzania (according to Chapin, 1954: 734, Useguais 5 ø 40' S, 38 ø 25'E to 6 ø 40'S, 37 ø 30' E). Reichenow'stype specimenwas ob- tained from Fischerwhen Fischerwas in Zanzibar, accordingto Cabanis (J. f. Orn., loc. cit., in an introductionto the descriptionof Hypocherapurpuras- cens). Cabanis stated that the bird came from the area of Somalis. On the otherhand the only "Lindi" on my mapsof Tanzaniaare far southat 10ø 00' on the coastand (Lindi Hills) by the southend of Lake Tanganyika. The label of the type (Berlin Museum30709) has the word "Lindi" written ap- parentlyafter the label was originallyinscribed, and two topotypicalspeci- mens,also collected by Fischer,lack mentionof "Lindi." Probablythe birds were indeedtaken from the Useguaregion in northeasternTanzania as this is lessthan 100 miles from Zanzibar. Also, Reichenowdescribed one other bird togetherwith purpurascensas new, and the other bird was from "Bagamojo, Nguru-Berge,"a mountainrange just west of the Useguaregion (Chapin, 1954: 712). It is unfortunatethat the exacttype locality cannotbe determined,because it is not entirelyclear whetherFischer's specimen represents an indigobird from a populationwhich mimics L. rhodopareiaor L. rubricata. Shouldthe holotypebe determinableas a memberof a populationmimicking L. rubricata,then a new namemay haveto be found to refer to the indigobirdspecies mimickingL. rhodopareiain other areas. No specimensof firefinchescol- 240 ORNITHOLOGICAL MONOGRAPHS NO. 11 lectedby Fischeror by anyoneelse from "Usegua"have beenfound. In the samegeneral region firefinchesare known from a few localities: L. senegala andL. rhodopareiaboth from Bagamoio and Morogoro, and L. rubricatafrom Turiani at 2,000 feet elevation and from localities more distant to the north and southof Usegua. The vegetationof the Useguaregion includesboth acaciasteppe and moisterhabitats (Keay, 1959), and it is impossibleto rule out eitherof thefire finches as possible hosts of topotypicalpurpurascens. The Useguabirds are amongthe purplestof all of the Tanzaniaindigobirds. There- fore on morphologicalgrounds they probablyare conspecificwith the more purplishbirds describedhere. Unlessadditional field work showsthat purplishindigobirds all mimicL. rubricataand not L. rhodopareiain the Usegua region,it seemspreferable to retain the name V. purpurascensfor the purplish to purplish-bluebirds, mostof whichmimic L. rhodopareia,rather than propose a new name for this species. Intergradation(and presumablygenetic introgression) of V. purpurascens and V. f. nigerrima is evident in Malawi and Tanzania.

Vidua funerea V. funerea is appropriatelynamed the Variable Indigobird, and the different populationsbelonging to this form have not often been groupedto- getherinto a singlespecies by earlier workers. Forms which are known or thoughtto sharethe samemimetic song include purplish-blue to bluish-purple nominate[unerea from SouthAfrica, greencodringtoni from the southernrift highlands,and geographicallyintermediate blue populationsbetween these two forms,as well as dull bluish-purplenigerrima from Angola and neighboring areas. Followingthese color changesnorthwards from Transvaaltowards the equator one finds purplish-blue blue-green blue-purplish--greenish- blue--a variablepattern indeed. In sizealso this speciesvaries geographically (Table 33) with the largestbirds beingthe greencodringtoni and purplish nominate[unerea and the smallestthe bluish and purplishbirds of the lower Congo from Kwamouth to Boma. The different forms are also unlike in color of the flight feathersand of the feet. With a knowledgeof their songs,it is possibleto reconstructthe relationshipsbetween these birds on morphological grounds. One new form is recognizedin the presentscheme. Femalestaken with malesalso show considerable geographic variation; subspeciesof V./unerea canin someinstances be morereadily distinguished from each other than from femalesof different species. Descriptionof forrns.--Theseveral subspecies that I regardas belongingto this speciesmay be characterizedas follows.

Vidua lunerea /unerea (de Tarragon) Fringilla Iunerea de Tarragon, 1874, Rev. Zool. Paris, p. 180: Natal. Holotype: not located. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 241

I I'• 242 ORNITHOLOGICAL MONOGRAPHS NO. 11

I I 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 243

I I I I I I I 244 ORNITHOLOGICAL MONOGRAPHS NO. II 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 245

I 246 ORNITHOLOGICAL MONOGRAPHS NO. 11

Adult malesin breedingplumage are medium-dullglossed with purplish- blue. Wingsbrown; rectrices brown. Bill white; feet orange-red;iris dark brown. Breedingfemales shot while with malesof thisform at Tzaneen,Transvaal, are gray-brownabove with dark brown streaks.Underparts are gray-brown with the bellywhite. Wing coloris brown. Bill colorsrecorded are white, whitish,pale pink-white,light gray, and grayish-white;the last two were in four femalestaken in Januaryearly in the seasonand the non-grayfemales were takenin Februaryand March. Foot colorsrecorded from freshlyshot birdsare orange-pink,pale pink, pink-orange,pale gray pink, pale rosygray, and pinkish-white. The iris is dark brown. Non-breedingplumage of malesis not known,nor is the juvenalplumage. Remarks.---In South Africa where this form occurs it differs from V. purpurascensin foot color of malesas well as in songand habitat, but no differencesin plumageare detectable.Some but not all female/unerea taken at Tzaneenhad feet of a distinctiveorange hue; the femaleswithout orange feet includedlaying birds. No plumagedifferences are evidentbetween females mated to red-footed V. [. [unerea and to white-footed V. purpurascens,the purplish-blueindigobirds of the eastern Transvaal. Most previous workers have not differentiatedbetween these two speciesin South Africa, evidently becauseof the paucityof data on foot color. V. /./unerea is known from spedmensfrom the easternCape Provinceas far westas SomersetEast. In Zululand,at Hluhluwe,I saw and photographed in color one of thesered-footed indigobirds. In Transvaalthe only specimens are from the humideastern edge of the escarpment.No purplishplumaged, red-footedindigobirds are known from Mozambique. The originaldescription of de Tarragon'sFringilla/unerea leavesconsider- ableroom for doubtabout the appearanceof the bird, and only the designa- tion of Natal as the type locality makesthe name referableto known kinds of living indigobirds.The bill and feet were saidto be whitish (blanchatres), but de Tarragon noted the specimenwas dried (dess•ch•), and hence the foot colorhad evidentlyfaded. De Tarragonregarded the bird as very similar to Fringillanitens Gmelin 1789 but worthy of speciesstatus because the latter was said to be from Brazil. The descriptionof Gmelin's Fringilla nitens (in his [unauthorized,post-Linnaean] edition of the SystemaNaturae) in turn appearsto be a description(caeruleo atra, chalybeo-nitens)of the samekind of bird describedas Fringilla chalybeataMiiller 1776, or Vidua chalybeata chalybeata(Miiller) of the presentwork. With the lack of holotypesof any of thesethree forms the only reasonablecourse to take, in the interestof a stablenomenclature, is to continueto regard/unerea (de Tarragon) as the name appropriateto the only commonform of indigobirdin Natal, the nomi- 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 247 natered-footed, white-billed, dull, purplish-blue glossed form of theVariable Indigobird. Vidua funerea lusituensis,new subspecies Holotype: UMMZ 217,255 (RBP 4571), male in breedingplumage, taken 8 April 1967 at 1,200 feet elevationin the Lusitu River valley southof Mt. Chimanimani, Melsetter District, eastern Rhodesia. Descriptionof type: Plumage glossyblue, wings blackishbrown, bill white, feet orange,wing 65 mm, tail 38, bill length 6.2, bill width 6.1, tarsus 15. Adult malesare brightblue; plumageis glossy.Wing and tail color is dark brown. Bill color is white, feet are orange,the iris is dark brown. One laying female (RBP 4579) taken with a male at the Lusitu River is brownishabove with darker brown shaft streaksand brownish-graybelow; the belly is white. Bill color was white and feet were orange,paler than in breedingmales. Two other femalesseen with males here also had orange feet. Non-breedingplumages of malesand the juvenalplumage are unknownfor this form. Remarks.--The bright blue birds of mountainoussoutheastern Rhodesia are known from eightmales and a female taken in the Lusitu valley and a male from Mt. Selinda. The blue color is intermediate between that of the purplish,dull nominatetunerea and bright greencodringtoni, and thus the geographicallyintermediate populations readily show the affinity of all of these mimics of L. rubricata. Likewise the one female that I collected is intermediatein graynessbetween females of nominatefunerea and codringtoni. The morphologicalhomogeneity in Lusitu birds indicatesa limited influx of genesfrom southernand northernforms; these males show no overlapwith eitherV. •. iunereaor V. •. codringtoniin plumagecolor, although wing color doesoverlap with malesfrom Penhalonga.Therefore the southeasternRho- desianform of V. iunerearepresents not a hybrid swarm,where extremesin color of greenand purplishwould be expected,but a stable,characteristic form. Recognitionof a name for this form may be helpful in emphasizing the presenceof populationswhich are intermediatebetween V. t. funereaand the form codringtoniand in pointing out the conspecificityof these forms. The slightlyshorter wing length of lusituensisis not significantlydifferent from other subspecieswith the presentlimited sample. Vidua )tunerea codringtoni (Neave) Hypochera codringtoni Neave, 1907, Mem. Lit. Phil. Soc. Manchester, 51, p. 94: Molilo's, near Petauke [----Mulilo at 14ø 02' S, 30 ø 58' E, Luangwa Valley, eastern Zambia]. Holotype: BM (NH) 1907-12-29-140, male in breeding plumage. Males in breedingplumage are glossygreen to blue-greenor greenish-blue. Wing and tail blackishto dark brown; bill white; feet red-orange;iris dark brown. 248 ORNITHOLOGICAL MONOGRAPHS NO. 11

Females(RBP 4456, 4528, 4529) taken with singingmale codringtoni whichalso were collectedare gray-brownabove with blackishshaft streaks, darkergray on the upperbreast and flanks, and white-bellied. Bill colorwas whitishin a femalewith an eggin the oviducttaken at Penhalongabut was grayabove and white below in two non-layingfemales shot at Zomba. The feet of all threewere orange; their iriseswere dark brown. Three otherfe- males(each shot but lost) wereseen at closerange at call-sitesof codringtoni in Rhodesia,and all of thesewere distinctlydarker gray below than other southernAfrican indigobirds; all had brightorange feet. A femalefrom Um- tali in the Umtali Museum (214) is probablycodringtoni as it matchesthe Penhalongafemale; its foot colorwas not recordedbut its plumagematches the other codringtoni. These are the only known female specimens. No non-breedingmales or juvenilesknown to be this form have been collected. Remarks.--Green codringtonihave been collectedin numbersin only two localities--thePenhalonga-Umtali region of Rhodesiaand the Nsanjeregion of Malawi. Traylor (1966:61 ) noted in the Malawi sample a correlation betweencolor and wing length,with the greenerbirds having the longerwings. A trend is alsoevident for the greenestbirds to have darker, blackishflight feathers.Traylor regardedbirds of smallersize and bluer color as evidence of intergradationbetween codringtoni and amauropteryx,mainly because amauropteryxis itselfyet lessgreen and is smaller. However,I think these bluishindividuals are betterregarded as lyingwithin the usualrange of variation of V. 1. codringtonior as showingintergradation with the bluer forms of V. runetea, suchas V. •. nigerrima. The size differencesof the birds grouped by Traylor into color classesshowed no statisticallysignificant differences betweenthe greenand the bluish-greenbirds, whereas both groupswere significantlylarger in wing lengththan amauropteryx.Second, all six green or bluish-greenbirds taken at Penhalonga,as well as the greenish-bluemale near Zomba, mimickedL. rubricata;the bluer birds sharedthe songof the greener birdsand all lackedthe mimeticsong of L. senegalawhich the V. c. amaurop- teryxwere heardto sing. The bluish-greenZomba bird is geographicallyand morphologicallyintermediate between the greenercodringtoni of the Nsanje region and the blue nigerrima at Lilongwe.

Vidua funerea nigerrirna (Sharpe) Hypochaera nigerrirna Sharpe, 1871, Proc. Zool. Soc. London 1871, p. 133: Angl. [= Angola]. Holotype: BM (NH) 78-12-31-810, male in worn breeding plumage and without data on foot color.

Malesin breedingplumage are purplish-blueto dull blue; mostare purple- blue.The primaries,outer secondaries, and rectrices are light brown to brown. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 249

Foot colorin my Lilongwebirds was whitish. Bill color is white, the iris is brown. Measurementsof type: wing,left 65 mm, right 66, tail 34, bill length 6.0, bill width 6.3, tarsus14. A male (NMR 26431) from Mwinilunga,Zambia, is in molt; its brown feathersindicate a non-breedingsparrowy plumage like that of the females. ThreeMRAC femalesfrom southwesternKatanga (Kasaji 2, Elizabeth- ville = Lumumbashi1) are indistinguishablemorphologically from females of V. purpurascensfrom Rhodesia;the Congolesebirds have no foot coloror bill colornoted. Anotherfemale (MRAC 11203) from Kwamouthis slightly more distinctlystreaked above. Becausemales in theseareas are all referable to V. f. nigerrima,these females may be identifiedon the basisof distribution. In addition,I shot a female (RBP 4540) from a V. •. nigerrimacall-site threemiles west of Mbabzi,near Lllongwe, Malawi, on 27 March 1967. The bill was white, and the feet were pinkish-white.Its measurementswere the same as in female V. purpurascens(Tables 32, 33). The bird was not dis- tinguishablein plumagefrom female V. purpurascens. No juvenilesare known that can be identifiedwith assuranceas this form. Remarks.--Thepoor conditionof the type specimendiminishes its useful- nessfor comparisonwith morerecently taken specimens in southernand central Africa. Examinationof the type in sunlightshows it to be indistinguishable in glossand wing color from someindividuals of V. purpurascens. Becausethe form nigerrima was describedearlier than was purpurascens, and becausetaxonomists generally have not distinguishedthese two forms, mostof the literatureof this centuryhas referredto all of the dark indigobirds of south-centralAfrica as nigerrima--eitheras a speciesor as a subspecies.As discussedhere, I regardnigerrima as conspecificwith V. /unerea and as specificallydistinct from V. purpurascens.The type localityof nigerrima aspublished by Sharpewas merely "Angola," but the originaltype label affixedby Sharpedesignated the localityas "GolungoAlto" and in his discussionof the collectionincluding this bird Sharpe(1871: 133) notedthat the collectionprobably came from northernAngola. Traylorhas pointed out that by 1870 ornithologicalexplorations in Angolawere unlikely to have takenbirds from regionsremote from Luandaor CuanzaNorte (Traylor, 1963: 13-14, and pers. comm.). The commonfirefinch in northernAn- gola is L. [rubricata]landanae, and the holotypeof nigerrimais similarin plumageto indigobirdsat Mwinilungaand southwesternKatanga, where againthe only common host species is L. rubricata.The onlyknown firefinch specimenfrom GolungoAlto is a femaleL. [rubricata]landanae in FMNH. For thesereasons I considerthe form nigerrimato be a subspeciesof V. funerea,the mimicof L. rubricatain otherparts of Africa. Birdsin northern Zambia,Katanga, and northern Malawi and morphologically indistinguishable from the nigerrimasample of northernAngola are associatedwith L. rubri- 250 ORNITHOLOGICAL MONOGRAPHS NO. 11

0 4

"wilsoni" - - - - '• amerunen$1$"• J

4-00 5O0 600 700 WAVE LENGTH (m•) Figure 39. Spectrophotometricdifferences in male breeding plumage color in the west African Pale-winged Indigobirds. Each bird is similar in color to the holotype of the form indicated. All specimensare from northern Nigeria ("nigeriae" : RBP specimen4946, "carnerunensis"= 4885, "wilsoni": 4960). cata (which forms a superspecieswith landanae)and theseindigobirds in turn intergradewith geographicallyadjacent forms of V. [unerea. Foot color of males of this form in northern Zambia is not well known. A male collectedby C. M. N. White at Mwinilungawas "flesh pink" in foot color. White (1946: 222) later reported additionalMwinilunga specimens but I have not found them; he notesone with "rose-coloured"feet and others with "whitish-pink"feet. Two nigerrimafrom southernKatanga (Dikulwe Valley) taken by S. A. Naeve have foot color data; one had "flesh" feet and the otherhad "pale flesh";none were orange. The foot color of Angola nigerrima is unknown; probably it is whitish. The form of greenish-blueV. [unereawhich replacespurplish-blue nigerrima in the central Congo is poorly known, and until field work has been carriedout in Kasai and Katangait seemsundesirable to describeany new forms. Vidua wilsoni The Pale-wingedIndigobirds of the Vidua wilsoni complex range from Senegalto Ethiopia and they vary greatlyin color. Blue, green,and purple birdswithin this grouphave each at timesbeen regarded as distinctspecies. Althoughthe firefinchhost species of theseindigobirds all vary from eastto 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 251 westand haverecognizable subspecies, the Pale-wingedIndigobirds show no comparableeast-to-west variation. Within a locality the green,blue, and purple birds may live together,and intermediateindividuals may occur. Spectrophotometricdifferences in malebreeding plumage color of "typically" (that is, bothlike the holotypesand alsolike manyother individuals) green, blue, andpurple birds are comparedin Figure39. In their mimeticsongs the three forms show considerablelocal differencesin the speciesof firefinch songmodel, but asdiscussed in the sectionon vocalmimicry some forms sing the songsof more than one firefinchmodel in one area, and the samemor- phologicalform may mimic differentspecies of firefinchesin differentparts of its range. The presenceof intermediateforms is significantin showingblue, green,and purple birdsnot to be simplemorphs in a polymorphicspecies, as the birds of intermediatecolors indicate a polygenicbasis of plumagecolor. Purplish birds ("wilsoni") are generally more distinct from blue ones ("camerunensis")and greenones ("nigeriae") than theselast two are from each other (Wolters, 1960, 1961; Traylor, 1966; the presentstudy). The other main considerationused in grouping all of these forms into a single monotypicbut variablespecies, V. wilsoni,is the indistinguishableappear- anceof the femalesmating with eachof the distinctivekinds of males. All of the femalesshot with identifiedmales in Nigeria were morphologicallysimilar; no differencesbetween females consorting with males of "wilsoni," "camerunensis,"and "nigeriae" at the call-siteswere found. The meaningsof the morphologicaluniformity of thesefemales may be two: first, the absenceof differencessuggests a lack of reproductiveisolation between forms sufficient to have led to any greatdegree of geneticdifferentiation between blue, green, and purple birds. Secondly,the similarityof femalesmakes it impossiblein practiceto demonstrateassortative matings of femalesin the field. Neverthe- less,males of one colorform tend to mimic a singlefirefinch locally, and to usethe samecall-sites, and future field studiesmay indeedshow a significant amountof assortativemating in this complex. Description.•The variable specieswilsoni may be describedas foliows.

Vidua wilsoni (Hartert) Hypochaera wilsoni Hartert, 1901, Nov. Zool., 8, p. 342: Yelwa, Nigeria. Holotype: AMNH 452337, male in breeding plumage. syn. Hypochera nigeriae Alexander, 1908, Bull. Brit. Ornith. Club, 23, p. 15: Kiri, R. Gongola [Nigeria]. Holotype: BM (NH) 1911-12-23-3302, male in fresh breeding plumage. syn. Hypochera chalybeatacamerunensis Grote, 1922, J. f. Orn., 70, 398: Weg Nola- Mbaiki, siid6stlichesNeukamerun [----Central African Republic]. Holotype: Berlin Museum 950, male in breeding plumage. syn. Hypochera chalybeata sharii Bannerman, 1922, Bull. Brit. Ornith. Club, 43, p. 29: Ram, Gribingui River, French Equatorial Africa [----Central African Republic]. Holotype: BM (NH) 1911-12-23-3308, male in breedingplumage. 252 ORNITHOLOGICAL MONOGRAPHS NO. 11

TABLE 34 MEASUREMENTS OF VIDUA WILSONI• FORM "WILSONI"

Male Geographicarea N Max. Min. Mean t.95•

Wing Sudan 8 65 62 63.13 .33 Nigeria 16 65 60 62.69 .30 other 8 65 63 63.63 .26

Tail

Sudan 8 37 35 35.75 .29 Nigeria 16 41 34 37.19 .49 other 8 40 36 37.25 .49

Bill Length Sudan 8 6.5 6.0 6.26 .05 Nigeria 16 6.5 5.7 6.17 .09 other 8 6.7 5.9 6.33 -

Bill Width

Sudan 6 6.5 6.0 6.26 .05 Nigeria 16 6.5 5.7 6. ! 8 .09 other 8 6.6 5.7 6.28 -

Tarsus

Sudan 8 15 14 14.25 .16 Nigeria 16 16 14 14.50 .15 other 8 15 14 14.38 -

Weight (g) Nigeria 5 14 13 13.4 -

Males vary greatly in color in these birds. To comparetheir mensural charactersI have separatedall males by eye into categoriesof green,blue, and purple by referring all birds as green or greener (less blue) than color standardspecimen RBP 4444 to "nigeriae,"all birds less green than 4444 but no more purplethan 4575 to "camerunensis,"and all birds more purplish (less blue) than 4575 to "wilsoni," althoughin fact birds of all intermediate hues are known. (The holotypesupon which the names "nigeriae," "camerunensis,"and "wilsoni"are basedfall into the color categoriesbearing their names.) Measurementsof thesebirds are comparedin Tables 34, 35, and 36. Little geographicvariation in size of any of the color forms is apparent. Within someareas the purplishmales, "wilsoni," are slightlysmaller in wing lengththan are the bluish or greenishbirds, and thesedifferences are statistically significant,as Traylor (1966: 63) has also pointed out. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 253

TnBLE 35 MEASUREMENTSOF VIDUA WILSONI, FORM "CAMERUNEN$1$"

Male Geographicarea N Max. Min. Mean t.95% Wing Ethiopia 2 65 64 64.5 - Sudan 10 67 60 64.80 .65 Congo 42 67 61 64.12 .22 Cameroon, C.A.R. 17 67 62 64.71 .30 Nigeria (north) 15 67 63 65.13 .24 Nigeria (south) 24 65 61 63.92 .22 Togo to Gambia 18 65 60 62.78 .34

Tail Ethiopia 2 39 37 38.0 - Sudan 10 40 34 37.5 .62 Congo 41 39 34 36.68 .20 Cameroon, C.A.R. 17 40 34 39.77 .38 Nigeria (north) 15 40 36 38.53 .39 Nigeria (south) 24 40 35 37.58 .27 Togo to Gambia 18 42 34 37.00 .45 Bill Length Ethiopia 2 6.2 6.1 6.15 - Sudan 10 6.5 5.8 6.23 .07 Congo 38 6.6 5.8 6.20 .03 Cameroon, C.A.R. 17 6.6 5.8 6.13 .08 Nigeria (north) 15 6.6 5.9 6.25 .05 Nigeria (south) 24 6.8 6.1 6.29 .01 Togo to Gambia 18 6.3 5.7 6.04 .05

Bill Width Ethiopia 2 6.1 5.9 6.0 - Sudan 10 6.5 5.7 6.14 .13 Congo 39 6.5 5.8 6.04 .16 Cameroon, C.A.R. 16 6.5 5.9 6.16 .05 Nigeria (north) 14 6.5 6.0 6.31 .03 Nigeria (south) 24 6.3 5.7 6.07 .04 Togo to Gambia 16 6.3 5.6 6.01 .06

Tarsus Ethiopia 2 15 14 14.5 - Sudan 10 16 14 14.90 .23 Congo 42 17 14 14.67 .12 Cameroon, C.A.R. 17 16 13 14.41 .17 Nigeria (north) 15 16 14 14.60 .16 Nigeria (south) 24 16 13 14.21 .15 Togo to Gambia 17 15 13 14.12 .12 Weight (g)

Nigeria 13 14 13 13.38 .20 254 ORNITHOLOGICAL MONOGRAPHS NO. 11

TABLE 36 MEASUREMENTSOF VIDUA WILSONI, FORM "NIGERIAE"

Male Geographicarea N Max. Min. Mean •.95ø•

Wing Sudan 2 63 62 62.5 - Cameroon 6 68 62 63.7 - Nigeria 6 68 63 65.2 - other 5 64 61 62.7 -

Tail

Sudan 2 38 33 35.5 - Cameroon 6 40 36 37.5 - Nigeria 6 39 35 37.0 - other 5 38 35 35.2 -

Bill Length Sudan 2 5.6 5.6 5.6 - Cameroon 6 6.2 5.7 6.03 - Nigeria 5 6.6 5.7 6.04 - other 5 6.1 5.6 5.85 -

Bill Width

Sudan 2 6.2 5.9 6.05 - Canaeroon 6 6.4 5.4 5.87 - Nigeria 6 6.5 5.8 6.18 - other 5 6.4 5.5 5.88 -

Tarsus

Sudan 2 14 13 13.5 - Cameroon 6 16 13 14.5 - Nigeria 6 15 14 14.7 - other 5 15 14 14.6 -

Weight (g) Nigeria 1 - - 13. -

The primaries,outer secondaries,and rectricesare pale brownin the purplish birds and slightlydarker brown on the averagein the blue birds taken in the same month in Nigeria. Green birds have pale brown flight feathers. Bill color in all males of all three color classesI collectedin Nigeria was white or whitish, and most collectors have used these terms to describe bill color, thoughone "camerunensis"was capucinebuff and another (the holotype of "camerunensis")was said to be black (schwarz). For the relatively uncommon"wilsoni" the termswhite, weiss, weisslich, colorless, upper washed out pink and lower bonewhite, pinky white,pinkish white, and pale cinnamon pink have all been used. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 255

Figure 40. Plumage of females of the Pale-winged Indigobird complex of Nigeria. RBP 4943, "nigeriae," PanshanuPass; RBP 4938, "catnerltttettsis,"Zaria; RBP 4881, "wilsoni." Zaria.

Foot color in male "wilsoni" that I collectedin Nigeria was uniformly more white than in male "ca, terunensis." The five Zaria "wilsoni" had foot color of pinkishwhite, pinky-purplishwhite, or purplishwhite. Bluish birds when collectedhad foot colors of light purplish, purplish, lavender-whitish,light purplish with slight brown cast, purplish gray, purplish white, and whitish light purplish, and green birds that I took had foot colors of whitish-slightly pinky gray and whitish-slightlypurplish gray. I did not collect purplish and greenishbirds on the same day so have no direct comparisonof their foot color, but the greenishbirds had foot colors similar to those of blue birds. Other collectors have used additional terms to describe foot colors in these birds, but with a few possiblyquestionable exceptions the colors have indicated whitish or pinkish-purple feet. In all forms the iris is dark brown, as in the other indigobirds. Adult female indigobirdsof the V. wilsoni complex were shot from the call- sites of green, bluish-green,blue, and purple males in northern Nigeria to determinewhether they were morphologicall)distinct. All femalesare grayish-brownabove with blacl•ish streaks along each feather shaft on theback. Underparts are grayish on the upper breast and flanks and whitish on the belly. Flanks are light gray. The primaries, secondaries,and rectricesare medium brown, the same color as in bluish males in all females. All of the nine females shot from known call-sites where the male also was collected are very similar in plumagecolor and pattern (Figure 40). Measurements 256 ORNITHOLOGICAL MONOGRAPHS NO. 11 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 257 and color of the bill and feet are listed in Table 37. No tendencyfor differencesin size or proportionsare evidentin the sample. All femaleshad horn- white bills, purplish-whitefeet, and dark brown irises. Even in freshly shot females taken with male "wilsoni" and "camerunensis,"the individuals held togetherin the hand within 30 minutes of collectionhad no differencesthat were apparent. Female V. wilsoni (all forms) can be distinguishedfrom female V. chalybeatain Nigeriaby the purplish-white(not orange)feet and by a slightlygrayer (lessrusty) brown color on the plumageof the back. Males in sparrowyplumage were shotfrom call-sitesof green,blue, and purplemales in Nigeria. Male plumageand colorof the softparts is the same as that of females,though flight feathersof male "wilsoni"are lighterin each of the two birds I shotthan in their females. Skull pneumatizationof the five sparrowymales, all taken in August,was lessthan 25 percentcomplete, suggestingthat all were first-year birds. All had small (2 mm or less) testes. One "nigeriae" (RBP 4943) had a skull about 10 percentpneumatized, and another (4944) was 18 percentpneumatized, but each sanga songidentical with bothof the two malesin breedingplumage preceding them as studmales at the call-siteat Panshanu.The five sparrowymales are indistinguishable from eachother exceptfor the slightlypaler flight feathersof "wilsoni"males (birdsshot from the samecall-sites as weremales in purplebreeding plumage). The juvenalplumage, like the mouthmarkings of the young,is unknown. Remarks.--One purplish male was excludedfrom the above descriptions andmeasurements; this bird (BM (NH) 74-2-11-29) is muchlonger-winged (68 mm) than any west African indigobird,the field label givesno locality or date, and the vaguelocality "Senegambia"written on the museumlabel attachedsometime later may be an error. Morphologicallythe specimenap- pearsto be a southernAfrican V. purpurascensin size, color, and the darker brown of the primaries.

RELATIONSAMONG THE BLUE, GREEN, AND PURPLE BIRDS The firefinch songmodels of the Pale-wingedIndigobirds that I heard or recordedin Nigeria were stronglyassociated with the appearanceof the singing males,as listedin Table 10 and summarizedhere in Figure 41 for the individualmales that were tape-recordedand collected.All purple males mimickedLagonosticta rara, mostblue and greenish-bluemales mimicked L. larvata,and greenbirds mimickedL. rubricata. However,a bluish-green male (RBP 4959) and a blue-greenmale (4884) each mimicked L. rara rather than L. larvataor L. rubricata,the songmodels of othermales resembling thesetwo, respectively. To determinethe extentof clumpingof charactercomplexes in malesof the pale-wingedgroup, I arrayedall availablemuseum specimens from locali- tiesnorth of the equatorialforests on the basisof color(comparison with color 258 ORNITHOLOGICAL MONOGRAPHS NO. 11

-l- I-- 66

Z 65 hi --I 64

Z 63

4437 4444 4530 4581 4559 4575 4539 4525 452 I

GREEN BLUE PURPLE

COLOR STANDARD SPECIMENS

Figure 41. Scatter diagram of color and size of Pale-wingedIndigobirds in Nigeria of known mimetic song. Open circles are birds that mimicked Lagonostictarubricata, closed circles mimicked L. larvata, and triangles mimicked L. rara. Color standard specimensare listed in Table 30.

n/ger/ae I camerune n$15 I ttlq[[$00/tt

68 '• 67 E • 66

• 65 z 64 ._1 63 z 62

I I I I I I I I I 4437 4444 4530 4581 4559 4575 4539 4525 4521 GREEN BLUE PURPLE

COLOR STANDARD SPECIMENS Figure 42. Scatter diagram of color and size of Pale-winged Indigobirds from Ethiopia to Gambia. Color standardspecimens are listed in Table 30. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 259

PURPLE BLUE GREEN

68 N

67 C N NC

66 N N NO

NO 0 GO NNN NNNNNNN NN N NN C E NCO CC

CO 0 NNN SO0 C NS •- 64 NO0 z NNC NN N CGO NGO 0 cs CG CSO 000

z 62 NNS c

61 GG G

400 410 50' DOMINANT WAVELENGTH,d (m/•) OF PLUMAGE

Figure 43. Scatter diagram showing the dominant color (d) of male breeding plumage and wing length in specimensof Pale-winged Indigobirds from Africa north of the Gulf of Guinea, the Congo forests, and the Kenya desert. Each letter indicates the locality of a specimen. C = Cameroon, Chad, and Central African Republic; G = Upper Guinea region; N = Nigeria; O = Congo; S = Sudan and Ethiopia. standardspecimens) and winglength in Figure42. The scatterdiagram shows one fairly distinctcluster of relativelysmall, purplishbirds ("wilsoni") and another rather diffuse cluster of birds ranging in color from blue to green. The pattern showsa considerabledegree of intergradation(parallel to the numberof morphologicallyintermediate specimens) between all of thesebirds, and it suggestsa greateramount of interbreeding(less isolation) betweenthe greenand blue birds than betweenthese and the purple birds. Scatterdia- gramsof Pale-wingedIndigobirds in Nigeria, Sudan,the upper Guinea region, and Cameroon-Chad,considered separately for each of theseregions, show similarregional patterns of intergradation.However, for the Congo,the purplishbirds and blue birdsdo not form distinctclusters but ratherform a continuous,intergrading series of specimens. Color comparisonsbased entirely on color standardspecimens give results similar to those based on spectrophotometricdata. Taking the dominant wavelengthd for all Pale-wingedIndigobirds which were analyzedon the spectrophotometeras the bestdirect singlemeasurement of colordifferences, I arrayedspecimens on the basisof colorand winglength again in Figure43. 260 ORNITHOLOGICAL MONOGRAPHS NO. 11

The diagramshows a color gradientranging from purple throughgreen with somesuggestion of a smallernumber of birds of colorsbetween purple (= violetd = 400 mv) andblue (d about450 mv) than betweenblue and green (d around50'0 mv). This clusteringis especiallynoteworthy because d is a weakermeasurement of differencebetween purple and blue than betweenblue and green (Figure 39). The specimensintermediate between purple and blue in the ranged = 410 to 420 mv are mainlyCongo (Uelle) birds,again suggestinga more continuousgradient of color rangingbetween purple and blue in the indigobirdsin the northernCongo than in other areasof the range of the Pale-wingedIndigobirds. The causalbiological significance of this geographic differencein the degreeof intermediacyof male plumageis completely unknown. The pattern of variation shown in Figure 42 agreeswith Traylor's(1966: 66) commentthat formsof the Pale-wingedIndigobirds may interbreedwith each other in someregions but not in others. Becausein much of their range the color forms of V. wilsoni appear not to intergradeuniformly but ratherto clusteraround purple or blue, theremay be a considerableamount of non-randommating amongthe Pale-wingedIn- digobirdsand a high proportionof recombinantsapproaching the phenotypes of extremesin color, suggestinga recombinationof "parental phenotypes" (Short, 1969: 92). The clusteringof phenotypeson a color gradientsuggests that morethan one speciesmight be recognizedin the Pale-wingedIndigobird complex,as theremay be somesort of reproductiveisolation within the complex. However, I prefer to treat the complextaxonomically as a singlespecies,for the followingreasons. First, the clusteringdoes not necessarilyindi- cate selectionfor the "parental phenotypes,"because we do not know the ancestralforms or the historicaldetails of differentiationwithin the complex. Second,the degreeof interbreedingwithin the complexis probablygreater than betweenthe southeasternAfrican V. funerea and V. purpurascens,be- causefemales of theselatter forms (V. f. codringtoniand V. f. lusituensisin contrast to-V. purpurascens)are morphologicallydifferentiated from each other but femalesmating with the color forms of V. wilsoniare not morphologicallydifferentiated from each other, at least not in the samplesavailable. The greatsimilarity of all femalesin the NigerianV. wilsonicomplex suggests a minimumof geneticdifferentiation between subpopulations marked by male plumagecolor or by their mimetic-songcultures. The minimumof differentiation in turn suggestslittle effective reproductiveisolation between these groups.They do not seemto be behavingquite like differentspecies, though they certainly are partially distinctfrom each other in vocal behavior. Until somemorphological correlates of femaleswith the appearanceor songof their maleshave been established,or until bandingstudies of femalesof known parentageand songtutelage have been completed,there will be no evidence 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 261 that the subgroupsof V. wilsoni are behavingassortatively in their mating systemsas distinct species.

RELATIONS OF V. WILSONI TO OTHER GROUPS In the presentsystematic arrangement V. wilsoniis regardedas specifically distinctfrom other indigobirds,with which, however,it most likely shares both songsand limited geneticinterchange. The calls and songsof one of the firefinch songmodels of this complex,L. larvata,are quite similarto the callsand songsof L. rhodopareia,and on the Boma plateau or adjacent Ethiopia it seemspossible that "camerunensis" mimics of L. larvata might interbreed with V. purpurascensmimics of L. rhodopareia,if theyoccurred together. The only non-colormorphological differencenoted betweenbluish "camerunensis"in Ethiopia and purple-blueV. purpurascensin northwesternKenya is the wing length, and this character overlapsin the two forms. The othergroup with whichthe Pale-wingedWest African Indigobirdswould be likely to interbreedis V. runetea, which usuallymimics L. rubricata in southernand south-centralAfrica. The green "nigeriae"form of V. wilsoni recordedat Panshanu,Nigeria, mimickedL. rubricata,and some of its mimetic vocalizationswere indistinguishablefrom mimetic vocalizationsof L. rubricatain southernAfrica. In addition, somesmall, brown-wingedindigobirds from near the southernedge of the great Congo forestsare morphologicallyindistinguishable from someV. wilsoninorth of the forests,and they may be intergradesbetween V. wilsoni and V. runetea. I have tentatively called all of these birds south of the forest V. runetea. Intergradationis not evident aroundthe easternedge of the equatorialforest, as east and south of the birds of the easternmostUelle all known male V. runeteahave dark brown (not pale brown) wingsand are not at all greenish. The lack of known geographicallyand morphologicallyintermediate populations in the large forestedcentral Congo region resultstrom the unsuitable habitat. Chapin(1932: 285) consideredthe equatorialforest to be "the most importantfaunal barrier in Africa." The lower UbanguiRiver is heavilyfor- ested along most of its banks for severalhundred miles (Adolf Friedrich, Duke of Mecklenberg,1909; Chapin, 1932) and it seemsunlikely to be an important avenueof gene flow in the indigobirds. However, the forest has not alwaysbeen a completelyeffective barrier to open-countrybirds. In a discussionof the avifaunal relationshipsof the savannabirds of Gabon and Moyen-Congo (= Congo: Brazzaville), Rand et al. (1959) comparedthe savannabirds southof the tropical rain forest with thoseof Angola (the savannasof thesetwo regionsare partly continuous) and of Cameroo.n(the savannaof Camcroon is separatedfrom open countryof Gabon-Congoby about240 milesor moreof forest[Keay, 1959]). 262 ORNITHOLOGICAL MONOGRAPHS NO. 11

Rand et al. describethe incompleteisolation of the savannaavifaunas of this area (p. 236):

Between the northernmostextensions of the southern Congo savannain Gabon and Moyen Congo and the southern edge of the Ubangi-Uelle savanna in the latitude of Yaounde, Cameroon, there is an almost unbroken stretch of forest that would appear to be a barrier for any of the savannabirds. There are, however,a few natural savannas in this region, and, probably more important, many man-made clearings that through continual burning have become grass-coveredand apparently act as stepping-stonesto maintain a continuous population flow between Cameroon and Gabon. Without this connection, it would be difficult to explain the fact that the birds of the Gabon savannas show a closer relationshipracially to the birds of Cameroon than to those of Angola. Moreau (1966:51 ) has describedthe probableshifts of vegetationbelts of Africa which might have affectedthe Congo forestsduring the Pleistocene. The Congo equatorialrain forestslie partly on Kalahari sands,which were redistributedthere by windsfrom southernAfrica during a dry period. By suggestingthat sandcould not be depositedover the modernforests (heavy rains would wash out the sandfrom the southerlywinds beforeit couldbe carriedfar) and by extrapolatingfrom a brief paperby de Heinzelin (1963) showingan arid period datingfrom 75,000 to 52,000 years ago in Africa, Moreau (1966: 51) has questionedthe existenceof any continuousCongo forestsin the Upper Pleistocene.The strikinglywidespread shifts of vegetation in Africa with the alternatingwet and dry periodsof the Pleistocenewere paralleled,evidently, in the Amazonian region of tropical South America. Now an unbrokenforest, the regionduring several dry climaticperiods was broken into smallerparts by more open vegetation(Haffer, 1969). One area that may have permittedmovements in drier timesis the Sangha river system.This connectssouthern Cameroon (including some open woodland areas;Keay, 1959) to the openwoodlands of southernCongo Republic (Brazzaville) and it, alongwith the "naturalsavannas" mentioned by Rand et al. (1959), may have beenassociated with the passageof open-country birdsacross the forest, which is narrowerin thisregion than it is at anypoint westof the rift area of the easternextremity of the Congo(Kinshasa). The resemblanceof the two indigobirdsfrom N'gabe (AMNH 345194) and Boma (MRAC 74937) to northernV. wilsonisuggests that theremay havebeen occasional movement, perhaps in drier geologicalperiods, of north- ernindigobirds across the forest. These two specimensare purplish-blue,they are smallerthan nearlyall southernAfrican V. junetea (Table 33), and their wingsare paler. The birdsfrom Kwamouth,in their smallsize, also suggest relationshipto the V. wilsonicomplex, but the largerange in variationin color of the Kasai birds (probably a form of V. [unerea) includessome birds greenerand somepurpler than the blue or green-blueKwamouth birds; the Kwamouthbirds are here consideredto be V. junetea. The possibleinter- 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 263 gradationbetween V. funereaand V. wilsonisuggests that thesemight be consideredconspecific. However, the equatorialforest now probablydoes serve as a barrier preventingthe interbreedingof the bulk of V. wilsoninorth of the forestwith the populationsof V. funereasouth of it, and the two forms are largelyallopatric, suggesting that they might better be termed a superspecies.A forestbelt of morethan 200 milesin breadthextending more than a thousandmiles from eastto westis probablyat the presenttime a highly effectivebarrier to movementof indigobirds,especially considering the local nature of their populationstructure indicated by their songdialects. Becausethe indigobirdsof the Kasai and adjacentregions have not been studiedin the field and no tape recordingsare availablefor the few specimens now in museums,I prefernot to introduceany new namesnor to lose infor- mationby lumpingthe smallerCongo birds south of the forestand thosenorth all in the samespecies. The variationamong the indigobirdsof Kinshasaand Kasaimay be not only a resultof interbreedingof V. wilsoniand V. funerea; the speciesV. chalybeatamay alsobe involvedin this complex,and short of lumpingall of the indigobirdsit seemsbest to await the resultsof field work in the Congo. Becausethe northern and westernforms of V. wiIsoni appear to be isolatedfrom southernindigobird populationsby the equatorial forest of the Congoregion and by the dry regionsof northernKenya and southern Ethiopia,and becauseit wouldbe confusingto lump wiIsoniwith V. iunerea but not with V. purpurascens(or vice versa), I have retained them as separate specieshere in the hope of providinga more convenientset of namesto use in describingthe biologicalrelationships among the indigobirds.Clearly all of the indigobirdsare closelyrelated.

THE V. FUNEREA--V. PURPURASCENS COMPLEX

The two speciesV. [unerea and V. purpurascensin southernAfrica are morphologicallyvery similarto eachother in size (Tables 32, 33 ) as well as in male plumage (Frontispiece;Figure 44). The only known morphological differencebetween these indigobird mimics of L. rubricamand L. rhodopareia, respectively,in SouthAfrica is the color of the feet. All SouthAfrican specimenswith foot color data were either clearlywhitish-footed or clearly reddish- looted, and all male indigobirdsthat mimicked one kind of firefinch were alike in their foot color. Some specimenslacking foot color data cannot be assignedto species(p. 238). V. •. funereaand V. purpurascensare not known to interbreed,and they seemto behavehere as discretesibling species, although the females may not all be distinguishable. Two areas of sympatryof these two forms are known, the Letaba River area in eastern Transvaal where V. [unerea mimicked L. rubricata and V. purpurascensmimicked L. rhodopareia,and 200 miles south along the 264 ORNITHOLOGICAL MONOGRAPHS NO. 11

back breast 'breast back

I0' ' 20' ' 30' ' ' ,'o' b s Figure 44. Color values of male breedingplumage in the V. purpurascens-V.funerea complex of southern Africa: b = brightness (in2); d = dominant wavelength c = index of color purity.

Ingwavuma River near Hlatikulu in southernSwaziland. At this secondlo- cality in 1969 M. O. E. Baddeleycollected two males which had "orange" feet and one with "pinkish white" feet; the term "pinkish white" was applied to bill color in all three. The two hosts,L. rubricata and L. rhodopareia, have both been collectedin southernSwaziland. These UMMZ specimensrepre- sent the southernmostlocality where the red-looted and the white-footed, purplish-blueindigobirds are known to live together. Except for foot color, V. [. funerea morphologicallyis more similar to V. purpurascensin SouthAfrica and Swazilandthan it is to the other subspecies of V. funerea (codringtoniand lusituensis)in Rhodesia. Figure 44 showsthe significantdifferences in brightnessb and dominant wave length d of the plumagecolor of male V. junetea from Transvaal and Rhodesia;the color purity c is also somewhatgreater in the bright Rhodesianbirds. Rhodesian V. [unerea, particularly V. •. codringtonifrom Umtali, are also blackish- winged,whereas both species in SouthAfrica are brown-winged.In Rhodesia

Figure 45. Color values of plumage of adult female indigobirds in northeastern Transvaal, South Africa: b = brightness (in2); s = index of color. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 265

Figure 46. Plumage of females of the white-billed species complex in southern Africa. RBP 4918, V. pt,'pttrascens, Mererisky, Transvaal: RBP 4380, V. 1. lunerea, Tzaneen. Transvaal: RBP 4579. 1'. t. lusituensis, Lusitu River. 1.200 feet elevation, Rhodesia; RBP 4456. V. I. cotb'ingtouœ2 miles S Penhalonga. Rhodesia. and southernMalawi the two speciesarc dissimilar in plumagecolor as well as foot color. The plumage of femalesis likewise more similar when the two speciesin SouthAfrica are comparedthan when racesof V. /unerea are comparedwith each other. Figure 45 showsthe absenceof any perceptiblecolor differences in femaleplumage of the three speciesin SouthAfrica. The somewhatdarker, grayer, more heavily streakedplumage of Rhodesian V. /unerea is evident in Figure 46. The feet of female Rhodesian V. 1. codringtoni and V. [. !usituensisare a vivid orange, whereas some of the breeding female V. I- [tmerea at Tzaneen had whitish feet. In south-centralAfrica from the Dedza highlandsof Malawi and the Lusaka area of Zambia northwards the morphologicaldifferentiation of indigobirds associatedwith L. rhodopareia and L. rubricata is much less marked, and perhapsconsiderable introgression occurs in Zambia, Malawi, and Ianzania south of the Congo forests and the Serengeti plains. If it were not for the assortativemating obsexxedin areas south of these regions, 1 would regard the purplish-blue,pale-footed indigobirds south of the equator as conspecific. When the plumage color and glossof white-looted males in bluish to purplish breedingplumage are compared for different altitudes in Malawi (Ta- ble 38), the birds at higher altitudesare more blue and glossyand birds at 266 ORNITHOLOGICAL MONOGRAPHS NO. 11

TABI•E 38 ALTITUDINAL VARIATION IN MALE PLUMAGE COLOR AND GLOSS OF V. I'URPURASCI•NS AND V. F. NIGERRIM.4 IN MALAWI

Number of specimensin each plumage class: Plumage class Altitude (X 1000 feet)• 0-I 1-2 2-3 3-4 4-5

Color purple-blue 0 0 0 0 0 intermediate 0 1 2 4 3 purplish-blue 0 1 0 2 0 intermediate 0 1 1 5 3 bluish-purple 0 6 1 1 0 intermediate 0 2 0 0 0 purple 3 4 7 7 0 Gloss low 3 10 9 4 1 low-medium 0 4 2 11 3 medium 0 0 0 4 1

x Localities included: 0-1,000'--Bwangu, Chikwawa, Port Herald (= Nsanje); 1,000-2,000'--Chididi Stream, Chinteche, Fort Johnston, Masona, Monkey Bay, Symon's (Central Shire); 2,000-3,000'-- Chididi Mission, Chididi Stream, Dai, Kazingizi, Mini Mini, Nyakamera, Salima 20 mi. W, Wangawanga Hill; 3,000-4,000'--Fort Lister, Lilongwe, Mbabzi (Lilongwe), Mkohoma, Mzimba, Njakwa, Zomba, Zomba 9 mi. S; 4,000-5,000'--Angomiland, Dedza, Mwangala. Other localities are excluded for lack of altitude data. lower altitudes are more purplish and dull. The bluest birds were taken mainly above 3,500 feet. As in the Nsanje area of southernMalawi, the elevationaldistribution throughout Malawi of the firefinchesL. rubricata and L. rhodopareia(listed in AppendixB) are similarto the distributionsof their mimics. Birds of intermediatealtitudes (2,000 to 3,000 feet) are intermediate, on the average,and overlap of the color and glossof the males at high and low altitudesin the males from intermediate altitudessuggests a lack of reproductiveisolation between glossier blue and dull purplebirds where intermediatehabitats and both hostfirefinch species are present. The only purple- blue or purplish-bluemales taken below 2,000 feet in Malawi were from Chinteche, on the west shore of Lake Malawi but well north of the acacia plain bordering the southernpart of the lake and in a considerablymore humidrainfall and vegetationregion (Keay, 1959; Ady, 1965); here L. rubricata occursand L. rhodopareiais apparentlyabsent. The color data of Table 38 suggestthat nigerrima(the bluer birds) and purpurascensare partially isolated but interbreed in local situations where both hosts are available. Evidence of intergradationbetween nigerrima and purpurascensis found also in a seriesof color reflectancecurves for representativespecimens taken along a northwest-southeastline acrosssouth-central Africa (Figure 47). Male nigerrimafrom Salujinga,northwestern Zambia, and from Kawambwa, northeasternZambia, are rather dull bluish. The short violet and blue re- 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 267

I, i I I I I i

• SALUJINGA •'"'•••WA _ -'•'•'•-..•LONGWE o o LILONGWEb f4

LILONGWEc •

LILONGWE d •

MONKEY BAY •-•

ZOMBA • _ .... •

ZOMBA

PENHALONGA 4

PEN HALO

400 500 600 700

WAVE LENGTH (m/•)

Figure 47. Intergradationbetween V. /unerea nigerrimaand V. purpurascensin Malawi. Spectrophotometricdata for indigobirdmales in breedingplumage. Solid lines indicate V. /unerea (mimics of L. rubricata recordedat Lilongwe and Zomba, Malawi, and Penhalonga,Rhodesia, and distributionallyassociated with L. rubricata at Salujinga and Kawambwa in Zambia). Dashed lines indicate V. purpurascens (recordedmimics of L. purpurascens).Note the variationamong the Lilongwebirds; some are most similar to northern Zambia nigerrima and others are closer to southern Malawi purpurascens. 268 ORNITHOLOGICAL MONOGRAPHS NO. 11 flectedwave lengthsgive the highestvalues, with a rise in the longerred wave lengths,especially in the Kawambwabird, giving a purplishcast to the plum~ age. Purplishbirds from Monkey Bay, Zomba, and Penhalongawere taken after they were heard to mimic L. rhodopareia;these V. purpurascensshow less reflectancein the blues and a relatively great reflectancein the longer red wave lengths. In theselast two localitiesrecorded mimics of L. rubricata were collected;the Penhalongabird figured is bright green and the Zomba bird bluish-green,and thesetwo orange-footed,black-winged birds represent V. [. nigerrima x codringtoni (closer to codringtoni) and V. [. codringtoni, respectively.The four recordedLilongwe mimics of L. rubricataare figured; theserange in color from birds indistinguishablein color from the northern Zambia nigerrima(a = RBP 4541, b = 4539) throughan intermediatebird (c = 4535) to a bird as purplish(d = 4534) as the Malawi V. purpurascens. The forms nigerrimaand purpurascensare more similarin color (Figure 47), wingcolor, and size than aredifferent forms (nigerrima, codringtoni) of indigobirdsmimicking L. rubricatawithin Malawi. Lilongwe,at 3,800 feet elevation,is only about40 milesfrom the hot, dry plainsat the southwestern edgeof Lake Malawi at an elevationof 1,600feet where V. purpurascensand L. rhodopareiaoccur. On a local scalemorphological intergradation (and geneticintrogression) between nigerrima and purpurascensmay take place alongthe LilongweRiver, which flows from Lilongweto Salimaand the lake. Female purpurascensand nigerrimain Malawi appearto be morphologically indistinguishable.One female purpurascensis known (NMR 94192), a bird shotby C. Long from a male purpurascensat 2,100 feet, Dai Village, NsanjeDistrict, southern Malawi, andin plumage,color of the softparts (bill horn, feet pale brown) and size (Table 32) it is very similar to the one nigerrimafemale available,a bird with white bill and pinkishfeet (RBP 4540) shotby me with a male nigerrimaat Mbabzi, Lilongwe,about 240 milesnorth and west of Dai. In Tanzania the situationseems to be like that in Malawi, with purplish- blue or bluish-purpleindigobirds, all with whitish feet, occurringwith L. rubricataand L. rhodopareia.Morphologically the museumspecimens available of all but the V. [unereacodringtoni, V. chalybeatacentralis, and V. c. amauropteryxare similar and do not fall into two distinctkinds, purplishand bluish,but rather all degreesof intermediacyof color occur. Nicolai (1967) found purplishbirds in Mikumi National Park, betweenMorogoro and Iringa and west of Moshi, to mimic L. rubricata,but at Iringa he later found purplish indigobirdsto mimic L. rhodopareia(Nicolai, pers. comm.). The distributionof indigobirdsand firefinches in Tanzaniasuggests that indigobirds in the'wetter areas are associated with L. rubricataand in thedryer areas are associatedwith L. rhodopareia. Around Lake Victoria, in the higher elevations of the north Pare Mountains,and in the moist woodlandsand highlands 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 269 of southernTanzania the purplish-blueindigobirds are distributionallyassociated with L. rubricata; theselocalities include Kindoroko, Mwagamira, Dar- es-Salaam,Karema, Bukoba, Ukerewe, and Nyarumbogo, and specimens from theselocalities are tentativelyidentified as V. f. nigerrima. Around the dry Masai steppethe firefinchis L. rhodopareiaand the pale-lootedindigobirds in the dryer areas (Morogoro,Undis, Usegua,Sunya, and [on the basisof Nicolai'ssong data] Iringa) are identifiedas V. purpurascens.Birds from the other Tanzania localities are not clearly referable in distribution or morphologyto eitherform. I can seeno consistentcolor differencebetween all the birds associated with L. rubricata and the ones associated with L. rhodopareia,although the birdsin moistregions near Lake Victoria are bluer and the Morogoro and Useguabirds are purpler than most Tanzania specimens. Nicolai (1967) has consideredthe pale-lootedindigobirds of east and southernAfrica to comprisetwo distinctspecies, one (which he calls "H. f. purpurascens")mimicking L. rubricata and the other (which he calls "H. nigerrima") mimickingL. rhodopareia. Although he has not collectedany "purpurascens"mimics of L. rubricata in the field he thought (1967: 314) these indigobirdswere more purplish in appearancethan his captive "nigerrima" mimics of L. rhodopareia. On a visit to Seewiesenin August, 1970, I heardother "nigerrima"mimicking L. rhodopareiaand observedthem at closerange in his aviaries. In colorthey were a closematch to my purpura- scensstudy skins from Rhodesia,with which I comparedthem on the spot, and it seemsmost likely that his "nigerrima"are the samekind of bird as my purpurascens.Because my specimensfrom Rhodesiaand Malawi that mimicked L. rhodopareiawere more purplishthan my specimensfrom the same areas that mimicked L. rubricata, and becausethe Malawi and Zambia specimenscollected from the range of L. rhodopareiaare nearly all more purplishthan specimensfrom the rangeof L. rubricata,it seemslikely that in Tanzania alsothe more purplishindigobirds are associatedwith L. rhodopareia and the bluer birds with L. rubricata. If there has been much introgressionbetween the forms, however,there may be no consistentcorrelation of morphologywith song. The holotypeof purpurascenstaken in Tanzania is clearlya purplishbird, and so is referableto the more purplishindigobirds from the morphologicalpoint of view, whereasthe worn holotypeof nigerrima was collectedin Angola at a knownlocality of L. [rubricata]landanae, and the holotypeis indistinguishablefrom indigobirdsin the rangeof L. rubricata in southernCongo and northernZambia, birds living where L. rubricatais the only hostfirefinch. Becausethe formsnigerrima and purpurascensintergrade with each other in color so that they may be indistinguishablein museum seriesexamined under controlledlighting and with a reflectancespectropho- tometer, I questionwhether any field worker could successfullycompare bluish-purpleand purplish-bluebirds seen at differenttimes and placeswith 270 ORNITHOLOGICAL MONOGRAPHS NO. 11 no standardcolor references; I surelycould not. Nicolai (1967) also stated that "nigerrima"has darker wing feathers. Museumspecimens examined in seriesshow no perceptibledifference to me, althoughone of the male L. rhodopareiamimics in Nicolai'saviaries did havevery dark, almostblackish, wings,a bird that had beenheld in captivityfor severalyears and had grown darkerwing feathersin successivemolts (Nicolai, pers. comm.). The other characterNicolai (1967: 314) mentionsas distinguishing"nigerrima" is a more slenderbody form. None of the taxa of indigobirdsthat I have handled have relativelylonger tails than any others,as documentedin the tablesof measurements,and I think there is no difference in the slenderness(schlankere K6rperform) of the variousindigobirds. In the absenceof specimensshowing the contrary,I suspectthat Nicolai'sL. rubricatamimics were the purplish- blue form calledhere nigerrima,and from observationsof his captivesI am certain that his L. rhodopareiamimics were the more purplishbirds here calledpurpurascens. Probably the pale-footedbirds of Tanzaniaare partially but incompletelyisolated by their distributionand mimeticsong, as in northern Malawi, and representboth nigerrimaand purpurascensand birds intermediatebetween these forms. Comparisonof specimenscollected in Tanzania whosesong has been recorded is necessarybefore any meaningfulanalysis of the extent of isolation can be carried out. The indigobirdsin Zambiain the areaswhere L. rhodopareiaand L. rubricata live togetherare variable in appearance,and some interbreedingmay occur betweenthe northern V. [. nigerrimaand the southern,dry-country V. purpurascens.However, examination of specimenssuggests essentially effective isolationbetween them. At Chilanga 13 male indigobirdspecimens (in the BulawayoMuseum) have been collected;these include 5 blue-glossed, red-billedV. c. amauropteryxas well as 8 whitish-billedbirds rangingin color from dull purple to blue-green(colors as defined in Table 30). As three speciesof firefinches( L. senegala,L. rhodopareia,L. rubricata) all live here (AppendixB) the indigobirdsmay be usingthree hosts. The 4 blue to blue- greenbirds with foot-colordata all had orangeor pink feet, while the 3 bluish- purple to purple birds all had pale ("white" or "flesh") feet. Bluish birds here are intermediatein plumagecolor between purplish-blue V. [. nigerrima and greenishV. [. codringtoni;the greatvariation in plumagecolor of the 5 blue to blue-greenspecimens suggests considerable variation of the recombi- nation of charactersof thesetwo forms of V. [unerea, and the Chilangabirds appearto representa populationwith intergradingcharacters of the forms nigerrimaand codringtoni.The reddishfoot color suggestsa stronginfluence of red-lootedforms of V. [unereasuch as codringtoni. The purpleand purple- blue birds (3) all had whitish or "flesh-coloured"feet, and these could con- ceivablybe allocatedto eithernigerrima or purpurascens,but giventhe close charactercorrelation of feet andplumage colors in the Chilangabirds I have 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 271 calledthese pale-footed birds V. purpurascens.The lack of recombinationof plumageand foot colorsin the Chilangabirds suggests to me a lack of hy- bridizationbetween local V. runeteaand V. purpurascens.Traylor (1966: 61) examinedmost of the Chilangaspecimens and considered them to represent a hybrid swarm between amauropteryxand codringtoni;he was not awareof the occurrenceof three kinds of indigobirdsin southernAfrica and did not mentionthe purplishspecimens. In mostareas of Zambiathe indigobirdsare morepurplish (purpurascens) in the dry southwhere L. rhodopareiaoccurs and more bluish (nigerrima) in wetter areaswith L. rubricata. Of the 48 specimensof purplishto bluish, pale-footedindigobirds in the rangeof L. rhodopareiain Zambiaonly one was in the color classas blue as color standardspecimen RBP 4539; this purplish-bluebird (NMSR 11893) was from Ft. Jameson,where both L. rhodopareiaand L. rubricataoccur. I have calledthis bird V. f. nigerrima; two otherspecimens (purple) from the samelocality (NMSR 4539, BMNH 1934-4-22-3)I havecalled V. purpurascens.Ft. Jamesonlies only about74 milesfrom Mbabzi, Lilongwe,Malawi, wherethe purplish-bluebirds mim- ickedL. rubricata,and the elevationwhere the Ft. Jamesonspecimen of nigerrimawas taken was 3,500 feet, aboutthe sameas at Lilongwe. One otherspecimen (from Mulanga,NMSR 30464) wasslightly more purple than colorstandard specimen 4539; this bird livesfar from any knownL. rubricata, and it is regardedas a V. purpurascens.On the otherhand, of the 17 specimensof birds(excluding V. chalybeata)in the rangeof L. rubricatain Zambia,only three were as purplish as 4539; thesethree were from Salujinga and Mwinilunga,where L. rubricatais the only host firefinch. Two from southeasternZambia (BMNH 1933-5-11-86from Chipakoin the Ft. Jameson districtand BMNH 1907-12-29-140from Mulilo, the holotypeof codringtoni) and two from Chilangawere greento bluish-green,and the other 10 were green-blueto purple-blue.The total sampleof Zambia birdsthus indicatesa smalldegree of interbreedingbetween V. purpurascensand anyform of V. funerea. The other two areaswhere V. purpurascensand V. f. nigerrimaoccur togetherare Angola and the southernCongo. In neitherof theseareas are there sufficientnumbers of specimensto determinewhether or not the two forms interbreed,though the two Congospecimens are considerablymore purplish thanany other indigobirds collected nearby, suggesting little or no intergradation. In Angolathe indigobirds[excluding four known red-footedbirds--V. chalybeataokavangoensis from Huila, Gambos,Fazanda do Cuito (Moco), andSerra do Mange(Moco)] rangein colorfrom purple-blue to bluish-purple. Thereis no cleardifference in colorbetween birds from L. rhodopareia localities(Huila, probablyalso the dry areaaround Gambos [Mossamedes]) and from L. [rubricata]landanae localities (Chitau, Dugue de Braganca, 272 ORNITHOLOGICAL MONOGRAPHS NO. 11

Luhanda), but someof thesebirds are worn and may have lost their original color. The museummaterial from Angola is insufficientto determine the degree of isolation there between purpurascensand nigerrima. The morphologicalsimilarity of nominatenigerrima from northernAngola to the indigobirdsin northwesternZambia and the adjacentlocalities in Katanga suggestno significantmorphological differentiation between birds mimicking L. [rubricata]landanae in Angolaand birds mimicking L. rubricatain Zambia and Katanga,and the lack of differentiationsuggests considerable gene flow and no important barriers of reproductiveisolation betweenthese birds. I regardV. funerea and V. purpurascensas distinctspecies because three of the four subspeciesof V. funerealive togetherwith purpurascenswithout extensiveinterbreeding. The black-winged,orange-footed males (V. •. codringtoni) and the purplish-blue,brown-winged, white-footed males (V. purpurascens)are sympatricin easternRhodesia in the Umtali area and do not interbreedthere to any significantextent. The maleshave different songs, and morphologicallydifferent kinds of femalesvisit thesetwo kinds of males. No specimens,male or female, displayany intermediacyin the characters examined.Also theseindigobirds appear to be specificallydistinct in south- easternRhodesia (Lusitu-Melsetter area), becausethere V. •. lusituensisand V. purpurascensdiffer in the sameways as do the Umtali birds, exceptthat lusituensisis blue insteadof green. No femaleV. purpurascenswere collected there with singingmales, althoughI did observetwo whitish-footedfemales visitingan activepurpurascens call-site in a cultivatedclearing at 1,200 feet elevationnear the LusituRiver. In addition,V. f. funerea(conspecific with codringtoniand lusituensis as shown by the samehost mimetic songs, the same foot color, and the morphologicalintermediacy of both male and female lusituensisfrom the intermediate geographic area) likewiseappears to be specificallydistinct from SouthAfrican V. purpurascens,although the morphological charactersare very similar in Transvaaland Swaziland. Some morphologicaldifferentiation is evidentin northernMalawi in the area where V. •. nigerrimaand V. purpurascensintergrade. It mustbe emphasizedthat the apparentintergradation of purplish-blueand bluish-purpleindigobirds, all withwhite bills and feet, in south-centralAfrica mayresult not fromintrogres- sionof two formsbut from a coincidentalresemblance of two isolatedspecies. In southernMalawi andeastern Rhodesia, and also probably in centralZambia (Chilanga),eastern Transvaal, and Swaziland,the existenceof areasof sympatry withoutinterbreeding, and the differentiationof the white-footedbirds in northernMalawi I think overrides,for taxonomicpurposes, any possible interbreedingbetween the forms V. •. nigerrimaand V. purpurascensin much of south-centraland eastAfrica. Short(1969: 98) hassuggested that the relative sizeof the areasof overlapwhere forms interbreed or do not may be a 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 273 good guidefor taxonomicdecisions of cases,such as this, where birds interbreed in some areas but not in others. As these areas south and north of the Zambezi River are not greatly dissimilarin extent, and as the femalesare stronglydifferentiated in the southernarea and behave(select mates) as distinct species,they are treatedin this study as two different species.

INTERGRADATION OF V. CHALYBEATA AND OTHER SPECIES Althoughin mostof its rangeV. chalybeatais morphologicallydistinct and doesnot interbreedwith other kinds of indigobirds,in the Congo V. c. cen- tralisappears to form a hybridswarm with V. [. nigerrimaand in eastAfrica it may interbreedwith one of the pale-footedforms of the V. [. nigerrima-V. purpurascenscomplex. Intergradationof centralisand nigerrima was suggestedby Traylor (1966) to occurin the Ruanda-Burundi-Kivuarea and intergradationof centralisand birdsthat he termed"[unerea" in northeastern Tanzania. No field work involvingboth collectingand tape-recordingindi- vidual males has been carried out in either of these areas, but examination of museumspecimens does suggest interbreeding. Intergradation and introgres- sionof V. chalybeatawith V. [unereais moststrongly indicated in the southern Congo. Indigobirds of northeasternTanzania have been thought to show interbreedingbetween V. chalybeataand V. "funerea"on the basisof the color of male breedingplumage (Traylor, 1966). However, when wing color, foot color, and song are all considered,the birds in this area appear to represent distinctspecies with little evidenceof hybridization. Approachingthe area of apparentintergradation from the north, where sometape recordingsand collectionswere made, may help resolvethe relationshipsof the Tanzaniabirds. The indigobirdmimics of L. rhodopareiaat Sigor,Kenya, all had whitishfeet and light brown wings,in contrastto other Kenyaindigobirds which mimicked L. senegalaand had reddishfeet and dark brown wings. In size, in male body plumage, and in female plumage these two forms (V. purpurascens,V. c. centralis) are indistinguishable.When the six male purpurascensfrom Sigor are comparedwith the 13 male centralis collectedin Kenya in May and June, 1967, no overlapin wing color is evident (Figure 48), though the darkest-wingedSigor purpurascensis a close matchto thepalest centralis. This centraliswas taken in ratherworn plumage at Olorgesailie;males taken within two degreesof latitude from Sigor south- ward were in lessworn plumageand showedno overlapin wing color. The bluish,pale-winged indigobirds of Sigorand Kacheribain scrubcoun- try betweenthe centralhighlands and the northernKenya desertappear to intergradewith the more southernpurplish purpurascens in southernKenya. Males without foot color data from Kibwezi and Bura, east and southeast of 274 ORNITHOLOGICAl MONOGRAPHS NO. 11

! ,[ /

Figure 48. Male indigobirds from Kenya, showing contrast between darkness of primaries and the body plumage. Specimens from left to right: RBP 4687, VMua chalybeata, 15 mi. ESE Kisumu. fresh plumage: RBP 4680, V. chalybcata, somewhat worn ph,mage. Olorgesailie: RBP 4711, I. chalybeata, Sigor, fresh plumage; RBP 4723, V. purpurascens,Sigor. fresh plumage; RBP 4719. V. purpltra.•ce,$.Sigor, fresh plumage. the Kenya highlandsin scrub country, were purpler by one color standard specimenthan the Sigorbirds; these had wingsas pale as the Sigorbirds and are identifiedhere as V. purpurascens.A bird (Bonn 61-632) from Lembeni, just southof the Kcnya-Tanzaniaborder, is slightlymore purplishthan two other Lembcnibirds (V. c. centralis) and bridgesthe gap betweenthe bluet Kenya V. pttrptlrascensand the more purplishpurpttrascens specimens of Tanzania, and anothermale from Kindoroko (Northern Pare Mountains) is closer(blucr) to the Sigorbirds; it had"stone colour" feet. Thesebirds may represent intergradesbetween two or more forms. In northernTanzania the purplishor bluish-purplespecimens have paler wings than do the bluish birds, and this trend suggestssome separationof species,but foot color data are few, and the situationis incompletelyunder- stood. Although most museumspecimens from the southernKenya-north- 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 275 easternTanzania regionlack data on foot color, the three centralisthat I collectedat Olorgesailie(southern Kenya) had light orangefeet, muchpaler than in the 10 centralisthat I collectedfarther north in Kenya. All three birds I heard to mimic L. senegala. Foot colors of these three birds were recorded as pinkish white-pink, light pink-orange,and pink, whereasthe other Kenya birdsI collectedhad red-orangeor orangefeet. Comparingthe plumagegloss of the Kenyabirds in my sampleunder a MacBethSuper Color MatchingSky- light I can seeno tendencyfor the Olorgesailiebirds to be more purplish(less greenish)than the other Kenya birds; they are all about equally blue. The pale foot color in thesesouthern birds suggestssome genetic influence of one of the pale-footedindigobirds (purpurascens or nigerrima) in east Africa between the Kenya highlandsand the northern Tanzania plains. Becauseall of the east African birds are more or less bluish it is rather difficult to demon- strate speciesintergradation using plumage characters. Perhaps, as Traylor (1966) argues,some intergradation between V. chalybeatacentralis and V. purpurascens (rather than runetea) has occurred in northeastern Tanzania, thoughat Sigor,north of the highlands,the two live togetherwith no evidence of interbreedingand their songsare distinct. Indigobirdsin the southernCongo region are morphologicallysimilar on the one hand to V. funerea nigerrima to the southwest and on the other to V. chalybeatacentralis to the northeastand east. Various interpretationshave beenmade of the relationshipsof the indigobirdsof Kivu, Kasai, and Katanga. Schouteden( 1964: 199, 1965: 91 ) regardedall as conspecificand named the Kasai birds V. funerea wilsoni and the southwesternKatanga birds V. •. funerea. Chapin (1954: 564-565) regardedthe Kasai birds as V. •. nigerrima and the Kivu and northeasternKatanga birds as V. c. centralis("Hypo- chera c. orientalis") and V. purpurascens("H. f. purpurascens"),while Traylot (1966) and Hall and Moreau (1970) have consideredthe birds to form a hybrid swarm. The puzzling complex, the behaviorally unknown Congo birds, may be compared with other populationswhere some song data are available. The color and size of some central African indigobirds are plotted in Figure 49. V. chalybeata centralis from Kivu and V. funerea nigerrima from Kasaji (southwesternKatanga) and the Mwinilunga district (northwesternZambia) overlapconsiderably in color and size. No morphologicalspecies distinctions betweenthese birds could be made here without locality data. It is possible that two speciesmay occur,with or without interbreeding,in Kivu, but the range of morphologicalvariation there is no greater than in central Uganda and the Kenya highlandswhere only centralisis known. Comparisonof the southernCongo indigobirdsshows that most from the Tanganika region of Katanga are morphologicallylike the Kivu centralis,and Lake Tanganyika 276 ORNITHOLOGICAL MONOGRAPHS NO. 11

70 Tanganika, Kivu

• 62 unga, 61 • Kasaji 4530 4559 4539 4521 4530 4559 4539 4521

BLUISH-GREEN BLUE PURPLE BLUISH-GREEN BLUE PURPLE COLOR STANDARD SPECIMENS Figure 49. Scatter diagram of color and size of indigobirdsof central Africa south of the Congo River. Kivu birds ---- V. chalybeatacentralis, Mwinilunga; Kasaji birds -- V. /unerea nigerrima; Kwamouth birds ---- V. funerea subsp.; Kasai, Tanganyika (Katanga), and northeasternZambia birds = V. chalybeata-V. /unerea complex. Color reference specimensare listed in Table 30. may be surroundedby thesenondistinctive blue indigobirds. The Tanganika birds are alsoinseparable from the Kasaji and Mwinilunganigerrima, though in wing lengththey are more like the majority of Kivu specimens.The most purplishspecimens are from Kabolo and "Tanganika"; the first of theseis a known locality of L. rhodopareia, and these and other purplish birds (Chiancey, Tembwe, Kasiki, Moba) may be purpurascens,as suggestedby Chapin (1954: 566). Kasai birds, taken mainly near Luluabourg, are generally smaller and greenerthan are Tanganika males. In Kasai these have been taken with L. senegalaand L. rubricata about equally often, and it is impossibleto assign the indigobirdsto either one of thesehosts on the basis of distributionalone. The larger and greenerKasai birds (especiallyone from Merode) are similar to V. funereacodringtoni, but codringtoniis geographicallyseparated from thesebirds by the purplish-bluepopulations of nigerrimain southernKatanga and northernZambia. The bluish-greenKasai birds are smallerthan codringtoni on the average.Wing coloris brownish,slightly darker than in greenish V. wilsonifrom the northernCongo and lighter than in southerncodringtoni. Comparisonof the southernCongo birds with neighboringforms suggests that most of the Tanganika birds are V. chalybeatacentralis but someof these may be V. funereanigerrima or even V. purpurascens.The greenishKasai birds are probablya form of V. [unerea,while the bluer individualsmay be 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 277 eitherV. funereaor V. centralis,or both of these. It is apparentin Figure49 from the wide rangeof plumagecolors and from the intermediatesbetween themthat intergradationbetween the speciesprobably occurs. If locallythe Kasaiindigobirds mimic two speciesof firefinches and if the mimicsoverlap oneanother in colorand size,the speciesintrogression that is apparentin museumspecimens might be convincing.In the absenceof behavioralinforma- tion it is not possibleto assignmeaningful species names to all specimensor to all populations. The bitfishindigobirds in northeasternZambia appearto be two species, V. chalybeatacentralis and V. [unereanigerrima. Two birds from Kaputa, Mweru Marsh area,are both slightlyless purplish (they are purple-blue)than birds from the rest of northeasternZambia. Foot colors were not noted by the collector. I think these two birds are V. chalybeata centralis. The local fosterfirefinch species at Kaputa is L. senegalaruberrima, and this form of firefinch is the host of the similar bitfish V. chalybeatacentralis in much of east Africa. The other indigobird specimensavailable from northeastern Zambiaappear to be V. [unereanigerrima. Theseindigobirds, from Ft. Rose- berry, Kasama,Kawambwa, Lake Lusiwasi,and Mporokoso,had breeding plumageof purple-blueto bluish-purple.Foot color in all but two was pale (white,pale flesh, pale horn); two bluish-purplespecimens from Kasamalack data on foot color. In the area where these birds were taken, the most common firefinchis L. rubricata,particularly in countryabove 3,500 feet in elevation (Benson and White, 1957: 131). In vegetationand moistureregime the area is similar to that of Mwinilunga District of northwesternZambia and of southernKatanga in the Congo(Keay, 1959; Ady, 1965: 53, 59), where the only kind of indigobirdknown is V. [unereanigerrima. All of theseindi- gobirdslook alike. The fact that the two Kaputa birds (V. chalybeata)were only slightlyless purplish than the other indigobirds(V. runetea nigerrima) suggestspossible limited intergradation between them in northeasternZambia. At the presenttime many of the centralAfrican specimensof indigobirds remainunidentifiable, and the patternof morphologicalvariation in the complex will remain enigmaticat least until tape recordingsare made and singing birdshave beencollected in the Congo.

DISCUSSION As seenin the precedingdescriptions, each kind of indigobirdlives togetherin some areas of Africa with one or more other kinds without interbreeding,but in somearea or anotherevery species of indigobirdappears to interbreedwith anotherkind of indigobird. This is true regardlessof the taxonomicdisposition of the indigobirds(unless they are all called conspecific). As the speciesare recognizedhere, eachof them intergradeswith V. [unereain part of its range. Althoughthe classificationadopted here differs 278 ORNITHOLOGICAL MONOGRAPHS NO. 11

codr•gtoni

amouropteryx ]r purpuroscens lus/tuensis fun!reopurpurascens OkOVOD• •OeDS/S

V.chalybeota V. funerea V.purpurascens V. wilsoni Figure 50. Summary of speciesrelationships among the indigobirds. Vertical lines show the connectionsbetween geographicreplacements of a single speciesand horizontal lines indicate the introgressing forms of indigobirds regarded as distinct species. The number of lines representsthe degree of interbreeding between the forms as evidenced by mimetic songsand specimensof intermediatecharacters. Forms of different species that are not connectedby horizontal lines are sympatric and apparently do not interbreed with each other. in detail from Traylor's,his generalconclusions (1966: 158-159) are ap- propriately cited here: The present study attempts to show that each form, directly, intergradeswith all other forms, and under this criterion they must all be consideredone species.It is not correct, however, to considerthe different phenotypesat one locality as merely morphs of a single species. In some areas two or more phenotypesbehave [morphological intermediatesare unknown] as distinct species,and they are linked only through mutual intergradation with a third form. Trinomial nomenclature is too rigid adequately to express this complex relationship, which must be described rather than listed in checklist form. I have summarizedthe complexpattern of relationshipsindicated in the presentstudy in Figure 50. In this figure the degreeof similarityin mor- 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 279 phology(and presumablyin evolutionaryrelationships) between geographically neighboringtaxa are roughlyproportional to the numberof lines con- nectingthem, and the verticaldistances between indigobirds reflect both their morphologicaldifferences and their geographicdistributions. The indigobirdsshow a pattern of variationwhich defiesan unambiguous statementof their relationshipsin terms of "species."Several different con- ceptionsof the natureof a specieseach add to an understandingof the indigobirds. First,a morphologicalconception of a species(as discussed by Mayr, 1963: 31) is useful in separatingmuseum specimens into groups;morphological forms often correspondto distinctsong groups and to assemblagesof birds that interbreedonly with each other. However, in some areastwo different population systems(V. funerea and V. purpurascensin northern Zambia; V. funereaand V. chalybeatain southernCongo) occurbut are not morphologically separableinto two distinctgroups on the basis of their morphology. Specimensfrom South Africa lacking foot color data are unidentifiableto speciesif they are oneof the two purplish,white-billed forms V. purpurascens or V. •. /unerea. Viewingthe indigobirds as host-specific races of a singlespecies rather than asdifferent species would emphasize the hostspecificity of indigobirdpopulations. However,in many areasdifferent kinds of indigobirdsdo live together withoutinterbreeding and behaveas distinctbiological species, and calling them differenthost-specific races would merely sidestep the problemof de- scribingtheir relationshipsin generalbiological terms. Second,the different indigobirdsare not geneticallyspecialized to any great degreeto parasitize certainfirefinches only; the term raceimplies genetically differentiated popu- lationswith each specializedto exploit a differentresource. In contrastto geneticallydetermined host-specific races of internalparasites, the host specificityof the indigobirdsis derivedfrom behavioralimprinting to the "host" firefinchesin each generation. A populationwhich acquires reproductive isolation through the processof learningmay be termeda culturalspecies. As in somehuman populations the effectof early experienceupon matingpreferences and behavioralisolation, throughits recurrencein generationafter generation,is to causeeach groupwhich learns the samesignals to mate assortatively.The culturalspe- ciesapproach is helpfulin visualizingthe matingrelationships among individual indigobirdsregardless of their morphology.Greenish-blue or bluish Pale-wingedIndigobirds in Nigeria whichmimic L. rara likely mate with female siblingsof purplishV. wilsoniwhich mimic L. rara rather than with the female siblingsof bluish birds which mimic L. larvata. In the absenceof knownmorphological correlates of femaleswith malesin this complex,the cultural specieshypothesis is not readily testable. 280 ORNITHOLOGICAL MONOGRAPHS NO. 11

The role of learnedmimicry in reproductiveisolation of the formsof indi- gobirdsclearly seems to be effective,even if it is not complete;the distribution of phenotypesof successivegenerations (museum specimens taken over many years) in most areasindicates that mixed indigobirdpopulations usually do not form panmictic mating groups. Finally, the biologicalspecies concept of Mayr (1963: 19-20) includes considerationof both geneticrelationships and matingsystems. In the indigobirds thesetwo elementsof the speciesconcept may be more distinctthan in mostbirds. On the onehand, the early experienceof eachbird determinesits matingrelationships and the populationswhich learn a commonmimetic song are regardedas a singlecultural species. In contrast,the similaritiesin morphologyof the birdsindicate their geneticrelationships and permitinterpreta- tion about the evolutionaryhistory of the group. Both setsof information whentaken together are moreinteresting and revealmore of the relationships, past and present,of the indigobirds,than doesthe data on songor on morphology alone, and I have tried to use this approachin describingthe indigobirds.

THE IMPORTANCE OF SONG BEHAVIOR AND GEOGRAPHIC ISOLATION IN DIFFERENTIATION AND SPECIATION OF THE INDIGOBIRDS The vocal mimicry in songof the indigobirdsis of evolutionaryinterest becausethe behavioraldifferences used as speciesisolating mechanisms are apparentlyacquired through imprinting to the fosterspecies. Nicolai (1964: 187-196, 1967: 310) has suggestedthat imprintingmay be sufficientto accountfor the originof new speciesof viduinefinches. It alsoseems likely to me that learningof the songsof the fosterspecies may explainthe interesting pattern of variation in the indigobirdswherein the kinds of birds that may behaveas distinctspecies in some regionsmay interbreedin others. Song imprintingmay havebeen as importantin its own way as geographicisolation of separatepopulations has been in the evolutionarydifferentiation of the indigobirds.

DETERMINATION OF MATING GROUPS BY IMPRINTING TO HOST SONGS Assortativemating among the indigobirdsappears to be largelya result of imprintingto the fosterparent companions and learningtheir songsand other vocalizations.The indigobirdspresumably learn their mimetic songsfrom the foster parents,and so the signalsused in speciesdiscrimination and mate selectionare newly acquiredin each generation. Only by keepingtrue to the behavioraltradition of matingwith individualsmost closely resembling in song the fosterparents and by parasitizingthe samespecies of hostis the discrete structureof separatebreeding groups within a mixed populationof indigo- 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 281 birds conserved.Because the vocal signalsused in mating are learnedby each generation,the matinggroups of indigobirdsmay be regardedas culturalspecies. The acquiredculture or complextradition involveslearning the song usedboth in courtshipand in selectionof the hostfirefinch that will provide in turn the parental care and song training for the next generationof indigobirds. The importanceof imprintingin subsequentmate selectionis not confined to the parasiticfinches. Sincethe earlierwork on imprintingby Whitman and Lorenz therehave been many experimentaldemonstrations of birds choosing as matesthe individualsresembling their fosterparents in preferenceto their actualparents (much as Freud suggested for humans;see Wickler, 1968: 200). Schutz(1965) has shownthat amongsexually dimorphic species of ducks rearedby fosterparents of other species,males often mate with other ducks that resembletheir foster mother. Harris (1970) found that gulls raised by other speciesof gullsoften chooseand mate successfullywith their fosterspecies rather than their own. In a field experimentwith Pied and Collared flycatchers (Ficedula hypoleuca and F. albicollis), when the eggs of the two specieswere switchedthe young raised by a foster parent of one species tended to mate with individuals of that species,as mixed speciespairs were found in the years after adoption (LiShrl, 1955). Whitman (1919) raised pigeonsin captivity and hybridizedsome of them by cross-fosteringwith different species. Domestic pigeons (Columba livia) tend to mate with other pigeonsresembling in plumagetheir fosterparents (Goodwin, 1958; Warriner et al., 1963). Nicolai (1964: 188-196) has artificially raised severalkinds of youngestrildid finches with fosterparents of otherspecies and has observed some effect on mate selection,but no effect upon the developmentof their song. Goodwin (1965: 314) found early experienceto affect mate selection in a Blue-cappedCordon-Bleu (Uraeginthus cyanocephalus) through imprinting to its Cordon-Bleu (U. angolensis)foster parents. Immelmann cross- fosteredyoung Zebra Finchesand BengaleseFinches; early experienceaffects both songdevelopment and mate selection.Zebra Finchesraised in the nests of BengaleseFinches sang the songsof Bengaleserather than their own, even if they were raised within sight and sound of their own species. The strong social bond between the young and the older birds feeding them appearsto determinethe social attachmentsand their species-specificselectivity of learning the songs. The sexual behavior of these cross-fosteredfinches as adults (especiallyin the males)was directedtowards the fosterspecies (Immelmann, 1965, 1967, 1968b, 1969a, 1969b, 1969c). Male mating behavioris similarly directedtowards the foster speciesin the Bengaleseand in the African Silver- bill, Lonchura malabarica (Immelmann, 1969b, 1969c). Finally, in a group perhaps more closely related to the viduines, sparrows of the genus Passer havebeen experimentally cross-fostered from the eggstage by Cheke (1969). 282 ORNITHOLOGICAL MONOGRAPHS NO. 11

One youngmale of P. domesticusfoster-reared by a pair of P. montanusbred successfullythe followingyear with a femaleP. montanus.All of this experimental evidenceindicates that the early socialenvironment in the young of many birds may determinethe mate selectionthat appearslater when the young bird becomesmature. The consequencesto the populationand specieslevels of species-specific imprintingin the indigobirdsmay involveboth the formation of novel mating groupsamong individuals imprinted to a newlyparasitized species of firefinch and also the breakdownof reproductiveisolation between indigobirds that had been usingdifferent hosts when one of the indigobirdgroups uses the nestsof the secondform. This ebb and flow of behavioral isolation among populationsis probably responsiblein large part for the observedpattern of differentiationand intergradationor introgressio,nof the indigobirds.

GEOGRAPHIC VARIATION IN BEHAVIOR AND THE BREAKDOWN OF REPRODUCTIVE ISOLATION BETWEEN INDIGOBIRDS The geographicvariations in the songsof somefirefinch speciessuggests a possiblebehavioral basis for the observationthat indigobirdswhich in some areas do not interbreedand behave as distinctspecies may in other areas interbreed. Geographicvariations in the songof the firefinchesthemselves is not well documentedfrom recordings,but the vocal mimicry of the indigobirds showedconsiderable differences between the contact calls and songs from locality to locality. These regional or local differencesin indigobird mimicry indicatethe presenceof correspondinggeographic differences in the contact calls and songs in the firefinches, because where both model and mimic were recordedthe local firefinch song model was mimicked precisely (for example,Audiospectrographs 6, 7, and 11 at Merensky,Transvaal). The data available suggesta regionalvariation in songparalleling the degree of isolation between V. purpurascensand V. funerea. In the eastern Transvaalthe songsof theseindigobirds (Audiospectrographs 7, 14) are so similarthat I did not distinguishthem until the secondyear of field work. Here both male and femaleplumage of the two speciesare indistinguishable, the only notable differencebeing male foot color: McLachlan and Liver- sidge (1957: 449) say that the songsof their hostsL. rhodopareiaand L. rubricataare the same. It was to the songsof TzaneenL. rubricata as mimickedby local V. f. funereathat the experimentalpurpurascens females from Rhodesiaresponded. In easternRhodesia and southernMalawi V. f. codringtoni singsprimarily the high, rapid trills in its mimetic songs,and these are unlike any commonlygiven by sympatricV. purpurascens(Audiospectro- graphs14, 16; 9). In theseareas V. purpurascensand V. funerea behave as distinctspecies. In contrast,the central Malawi V. f. nigerrima at Lilongwe had mimetic alarm notes given in rapid sequencemuch like L. rhodopareia 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 283 and its local V. purpurascensmimic at Monkey Bay, and the whistlesof the Lilongweindigobirds were prolonged into slurrednotes like thoseof the Mon- key Bay mimics (Audiospectrographs15, 10). The vocalizationsof these two formswere so similarthat I did not distinguishbetween them in the field, as I mistookthe Lilongwebirds for mimicsof L. rhodopareia,a speciesnot known to occur at Lilongwe. V. f. nigerrirnaand V. purpurascensshow evidenceof intergradationand introgressionin the Lilongwearea and elsewhere in Malawi, and somespecimens of mimicsof L. rubricataare morphologically indistinguishablefrom mimics of L. rhodopareialess than 80 miles distant (Figure 47). Regionalinterbreeding among the westAfrican pale-wingedbirds of the V. wilsonicomplex also may be causedby the similaritiesof the songsof their firefinch hosts. Although in some ways the songsof L. rubricata and L. larvata are distinctfrom each other, both have slurredwhistles, and additional featuresof songmay be similar. The songsof L. larvata are most similar to thoseof L. rhodopareia,which it replacesin southernSudan and Ethiopia. Thesetwo firefinchesboth have slurredwhistles that most commonlyrise at the end,whereas those of L. rubricataoften drop in pitch(Audiospectrographs 17, 6, 12). As southernAfrican femaleV. purpurascensrespond both to the slurredwhistles of songsmimicking their hosts,L. rhodopareia,and to whistles of songsmimicking L. rubricata,a similarsituation may occurin west Africa. Female"camerunensis" may respondto the songsof both L. larvata and L. rubricataand to male indigobirdsmimicking the songsof either of these. An indiscriminatingresponse by a female"camerunensis" in Nigeria would then lead to matingswith malesof either "camerunensis"or "nigeriae"and a lack of any rigorouslyassortative mating within the complex. On the other hand, the songsof L. rara and of its mimic, mainly purplish"wilsoni" in Nigeria, are distinctin beingcomposed of short,harsh syllables (Audiospectrographs 12, 14, 17, 18, 19). The relativelygreater degree of intergradationin plumage color in Nigeria betweenblue "camerunensis"and green "nigeriae"than be- tweeneither of theseand purple"wilsoni" (Figures 41, 42, 43) may be due to the greatersimilarity of songsof the host firefinchesof the first two, L. larvata and L. rubricata. In contrast,in the northern Congo, museumspecimens indicate considerable intergradationbetween the purple "wilsoni" and the other forms of V. wilsoni (Figure 43). Possiblyin this regionthe sametwo host species(L. larvata and L. rubricata) are mimickedby "camerunensis"and "nigeriae,"as theyare in Nigeria,and L. rara maybe mimicked,by "wilsoni,"but the songs of thelatter firefinch may be lessdistinct from the otherspecies than theyare in Nigeria. A captiveL. rara, of unknownsource, described by Nicolai (in Immelmann et al., 1965: 200), had slurred whistles, whereas none of the Nigeria birds had slurredwhistles in their songs. 284 ORNITHOLOGICAL MONOGRAPHS NO. 11

The breakdownof reproductiveisolating mechanisms between V. chalybeata and V. funereain the Congo,as shownby museumspecimens (Figure 49), may likewiseresult from someas yet undocumentedsimilarity there of the songsof their foster firefinchesL. senegalaand L. rubricata. Breakdown of isolation between indigobird speciesmight also result if a bird were imprintedto two speciesof firefinches. Nicolai purchaseda captiveindigobird in southern Kenya that mimicked the songsboth of L. senegalaand L. rhodopareia, and he has played his recordingof this bird for me. It mimicked severalmotifs of both firefinches. Hearing it evokedto me a scenario: at sometime, an indigobirdis fledgedby one species and then losesits foster parents and is adopted by a fledged family group of another speciesof firefinch. The young is then imprinted upon both firefinch foster species. When the adoptedorphan maturesand sings,he attracts and mates with females of two species,and other young indigobirdsin the neighborhoodlearn his enriched song repertoire and then mimic both firefinches themselves. This song tradition would lead to interbreedingbetween indigobirds reared by two different foster species. A third behavioral mechanismseems plausible also. If foster parents were tolerant of a young viduine with mismatchingmouth color patterns in certain areas, or if colors of the young varied geographically,then switchesof hostsor acceptanceof hybrid young may there allow interbreedingbetween the indigobirdspecies. The aboveparagraphs suggest that in three differentareas where different kinds of indigobirdsintergrade (kinds which do not often interbreedin other areas of sympatry), the songsof their host firefinchesare similar to each other in certainfeatures. Becausefemale indigobirds may be attractedto the songsof morethan one speciesif the songsare similar(Table 22, p. 171) and becausethey may then mate with more than one kind of male, the behavioral basis of the varying degreesof reproductiveisolation observedin different regionsof Africa may be due to the geographicvariation in the signalsof their hostsand of other speciesof firefinchesin the same regions. The adaptivesignificance of the regionalvariations in firefinchsong to the firefinchesthemselves is not known. The few recordingsavailable suggest that the songsof L. rubricata and L. rhodopareiamay be most distinctin regions where the two specieslive together (eastern Rhodesia, southern Malawi) and most similar in regionswhere one speciesreplaces the other (central Malawi). Becausefirefinches do not advertize their territories with song but rather use song mainly in sexual situations,the distinctivehessof firefinchsongs in regionsof sympatrymay be more importantin promoting speciesrecognition and decreasingthe probabilityof interbreedingamong firefinch species.If this is true, characterdivergence of firefinch songmay enhance the reproductiveisolation of their brood parasitesas well.

TWO MODELS OF SPECIATION The evidencefor geographicisolation as a necessarycondition for speciation in birds has been discussedat length by Mayr (1963). Observationof 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 285 the range in discretenessof populationsystems between a singlespecies homogeneousthrough space and a seriesof geographicallyseparate popula- tionsdifferentiated at the specieslevel are oftenregarded as evidenceof the importanceof geographicisolation in theprocess of speciation.Mayr (1963: 477) regardsgeographic separation as the only probableeffective isolating eventwhich would prevent genetic interchange between populations through timeto a degreesufficient to allowdifferentiation to the specieslevel. In most bird species,behavioral isolating mechanisms probably arose incidental to the differentiationof their populationsliving in differentareas. Differencesin behaviorpresumably have evolvedas a by-productof geneticresponses to differentenvironmental selective pressures. Even when behavioral differences betweenspecies may be learned[as in the meadowlarks,Sturnella magna and S. neglecta(Lanyon, 1957; Szijj, 1966)] thesedifferences in songtraditions may have originatedin populationsseparated from each other by long distances. The role of imprintingin the initial isolationof two reproductivelydistinct sisterpopulations, on the otherhand, need not involvegeographic isolation. Nicolai ( 1964: 188-196) has suggestedthat imprintingto differenthost speciesmay lead to speciationin the parasiticfinches. Not only doesimprinting equip each vidninewith its mimetic songs(= isolatingmechanisms), it also providesa meanswhereby a geneticallysimilar population of birds derived from a singleinterbreeding parental generationmay in the next generation be split into behavioralgroups each with different songs. These would form two distinct,non-interbreeding populations each mating with the birds imprinted to the samehost song. If the subsequentgenerations remain host- specificin theirparasitism and the youngare imprintedupon the samespecies of host,the absenceof interbreedingbetween the two populationswould eventually lead to the accumulationof geneticdifferences between them. At this stagein their evolution (after geneticdifferentiation) the populationswould be distinctspecies. This model accountsfor two functionsof mimeticsong, as a proximateisolating mechanism that preventsinterbreeding between different songtypes in eachgeneration, and as the ultimatesource of origin of populationdifferentiation. In vocalimprinting to the host,the proximatebe- havioral isolatingmechanisms of indigobirdpopulations may be identicalto the ultimate historical isolating mechanisms. As the indigobirdsare unusualamong birds in normallylearning their isolating mechanismsthrough imprinting to unrelatedspecies, the host firefinches,it appearsdesirable to comparethe biologicalcharacteristics of the indigobirdsin somedetail with eachof the two speciationmodels mentioned above to find whether either model, or both, agreeswith the naturalisticob- servations.Here I am following Alexander (1969: 499), who arguescon- 286 ORNITHOLOGICAL MONOGRAPHS NO. II vincinglythat modelsmay be discussedin terms of the evolutionarystages involved,and thesestages in turn may then be used as criteria to test the agreementof the model with the natural populations. Geographicspeciation.--Comparison of different populationsshows a seriesof someof the intermediateevolutionary stages of the kinds to be ex- pectedif speciationhas occurredin populationsisolated from each other. This comparisonsupports in part the notion of geographicisolation in the historical differentiationof the indigobirds. 1. Geographicvariation within a species.--Considerablevariation occurs betweenpopulations of the samespecies in size and color (V. chalybeata,V. funerea), indicating genetic differencesbetween some populationsthat are partiallyisolated from each other by distance(west and northeasternAfrican formsof V. chalybeata)or by unsuitableintervening habitat (separation of V. c. ultramarinaand V. c. centralisby desertsin northernKenya). In this last examplethe degreeof pheneticdifference between populations is proportional to the distanceseparating the two forms;only alongthe upperNile is evidence apparent(dark wings,purplish-blue plumage) of geneticcontinuity in the two main subspeciesgroups of V. chalybeata.Similarly in V. funereathe different forms are separatedby unsuitabledry habitat in the Limpopo River valley and the dry scrubof the Rhodesianplateau and alsoby the equatorial forestsof the Congo,if wilsoniis regardedas part of the superspeciescomplex. 2. Geographicvariation in the degreeof reproductiveisolation within a species,from scarcelymeasurable differences up to apparentlycomplete speciation.--As discussedin the last section,V. /unerea and V. purpurascensare isolatedreproductively in easternRhodesia, in southernMalawi, and in eastern Transvaal,but they intergrademorphologically and apparentlyhave interbred in northernMalawi and perhapsinterbreed in Zambia as well. Similarly, V. chalybeataand V. /unerea are isolatedin southernAfrica but apparently not in Kasai. The forms "wilsoni" on the one hand and "camerunensis" and "nigeriae" on the other show little evidenceof introgressionin Nigeria, but further east there are many birds in the V. wilsoni complex that are intermediate in plumage color. Although geographicvariation in the amount of interbreedingbetween different forms of indigobirdsdoes occur, the pattern doesnot suggestone of increasingamounts of reproductiveisolation between successivelymore remote areas, but rather one of adjacent or overlapping populationssometimes isolated and sometimesnot. When the nature of the geographicvariation is consideredin this light, it provideslimited supportfor criterion 2 of the importance of geographicisolation alone in causingthe reproductiveisolation. 3. The most closelyrelated forms are allopatricor narrowlyoverlapping in distribution.--Someindigobird forms in additionto speciesunits may be 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 287 mentionedhere, to answerthe questionof whether neighboringforms are morelike eachother than eachis like a non-neighboringform. On the sub- specieslevel V. chalybeataultramarina is most similar to V. c. neumanni, andgreen-glossed, black-winged "aenea" from the westend of Africa probably representa geographicisolate later invaded by the more bluish neumanni birds of the inland Sudanzone. In V. chalybeatain southernAfrica each subspecieshas a similarsubspecies replacing it geographically.In V. [unerea the varioussubspecies also replace one another over geographic gaps ([unerea and lusituensisin Transvaaland Rhodesia)or narrowlyoverlap with introgression(codringtoni and nigerrimain Malawi and Zambia). Certain other cognatespecies complexes in South Africa and Kenya may be interpreted as resulting from geographicisolation. V. purpurascensis more widely distributedwithin southernand easternSouth Africa than V. f. [unerea, a morphologicallysimilar form; thusV. •. [unereamay have been geographically isolated from V. purpurascensin the mesic area of coastal Natal and the easternCape. A similar historymay have unfoldedwith the V. chalybeata- V. [unerea complexin central Africa, and then have been complicatedby secondaryintrogression. 4. Speciesnot closelyrelated to each other share a commonpattern of geographicbarriers.--If two closelyrelated subspeciesor speciesgroups havedifferentiated from eachother on differentsides of a geographicbarrier, then one would expectto find other groupsof speciesor of subspecies(not close relativesof the first) with their own membersdivided by the same barriers,if the barrierscaused isolation and differentiation.Examples of geographicdistributions that are similarto thoseof someindigobird taxa are known for other African birds; for example,the cuckoo-shrikesCampephaga phoeniceaand C. [lava occurnorth and southof the Congoforests and meet in eastAfrica, muchas the northernblack-winged forms of Vidua chalybeata and the southernbrown-winged V. chalybeatado, and the boubou shrikes Laniarius[[.] [errugineusand L. [[.] aethiopicusreplace each other at the Limpopo River valley in southernAfrica, the same apparentbarrier that separatesVidua [. [unerea and V. [. lusituensis(distributions in Hall and Moreau, 1970: 52, 93). There are other distributionpatterns of closely related birds in Africa where speciesare not separatedat the geographic features lying between indigobird taxa, however. Because at least some taxa seemto be separatedat the sameboundaries between indigobird taxa, it seemslikely that isolation by geographicbarriers may accountfor the evolutionarydivergence of some indigobird taxa. 5. Characterdisplacement.--Only one instancethat mightbe interpreted as selectionfor divergenceof charactershas been noted in the indigobirds. Foot color may conceivablyfunction as a reproductiveisolating mechanism 288 ORNITHOLOGICAL MONOGRAPHS NO. 11 in reducingthe likelihoodof interspecificmating in certainindigobird species. WhereverVidua purpurascensis sympatricwith V. [. [unereaor codringtoni the foot colors of the two are distinct, and V. c. amauropteryxin the same areas of Rhodesia is distinct in bill color from either of these. In these areas no interbreedingis knownbetween the threespecies. In Zambiaand Malawi, on the other hand, V. purpurascensand V. [. nigerrima have similar foot and bill color,and theseforms appear to have no effectivebehavioral isolation becausespecimens from the areasof contactare morphologicallyintermediate. Perhapsexperimental dyeing of th'efoot color of the birds in theseregions would help to elucidatewhether the apparentlydivergent characters are of any importancein maintainingreproductive isolation. Cultural speciationthrough #nprinting to the fosterspecies.--Experimental verificationof the role of learnedmimetic songsin mate selectionthrough cross-fosteringof wild or captivebirds and their subsequentmating preferences would providea direct test of the cultural speciationmodel. In the absence of these experimentaldata, the comparativeapproach may be used, as it was for the geographicisolation idea. Insofar as the empirical featuresof indigobirdbiology and distributionmatch the stagesexpected if the birds speciateby switchinghosts, the comparativeapproach supports the model of cultural speciation. 1. Learned signalsas behavioralisolating mechanisms.--Mimefic song in the indigobirdsis associatedboth with courtshipbehavior and mate selection,so the mimeticvocalizafions are thoughtto be isolatingmechanisms between the indigobirds. The indigobirdsprobably learn their mimefie vocalizations,because: (a) the calls mimickedare the ones that a young indigobirdwould hear from its fosterparents and nestmates,(b) vocalizations typical of thoseat variousstages in the developmentof the calls of the young firefinchesas they mature into the adult calls are mimicked, (c) a few individual indigobirdsmimic unusualhosts, and this can be accountedfor only by learning,and (d) other ploceids,sparrows, estrildids, and some other viduines sometimeslearn the songsof other species,and someof thesebirds are effectivelyimprinted to their foster speciesand mate with them. 2. A high degree of assortafivemating of individualsimprinted upon the same mimetic song.--In several areas of Africa where mimics of two or morefirefinch species coexist, assortative mating is greaterthan 90 percent. In Nigeria a similarlyhigh degree of matingexclusiveness between red-footed V. chalybeataand the pale-wingedforms of V. wilsonialso occurs (Table 26). 3. A one-to-onecorrelation of host and parasitein racial and species differentiation.---Nicolai(1964, 1967) has proposedas his evidencefor imprintingin the speciafionprocess the supposedlyperfect match of the dif- 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 289 ferentation of each taxon of the viduines and that of its host. This notion to be bornout would require equal rates of evolutionin hostand parasite, and is similarto one advancedby parasitologistswho believethat host-specific parasitesevolve hand-in-hand with their hosts;the concepthas beentermed "Fahrenholz'Rule" and statedthus: "Parallelingthe evolutionand splitting up of the hoststhere is... a splittingup of the parasites"(Hermig, 1966: 107). Hennig (p. 111) commentsthat "eventhe most extremeadvocates do not assumethat the parallelismis so closethat everyprocess of speciation in the one correspondsto a processof speciationin the other." Implicit in Nicolai'sargument is a perfecthost-parasite correlation between all individual indigobirdsand their hostsin mimicry. The similaritieswhich are evident betweenhost and parasite in racial differentiationmay, however,be due to commongeographic barriers and selectivepressures, not to any obligate coupling of rates of divergence. Furthermore,it is evidentthat a few indigobirdsmimic the wrong songand imprintingon an unusualspecies of host may lead to geneflow from one mimeticpopulation to another,and as a result "Fahrenholz'Rule" is too rigid to describeindigobird population systems. Nevertheless, the data may be examinedto see whetherthey agreewith a 1:1 correlationof host and parasiteat the subspecificlevel. For presentpurposes I am following the firefinch subspeciestaxa as recognizedby Mayr et al. (1968) except for L. senegala;in this speciesI recognizeL. s. senegalain the senseof White (1963b: 202). The patternof geographicdifferentiation of V. chalybeatagenerally parallels that of its host L. senegala,although extensive clinal variation in southern Africa characterizesthe form L. s. "pallidicrissa"(not recognizedin Mayr et al., 1968) from L. s. rendalii, whereasthe correspondingsubspecies of indigobirds,V. c. okavangoensisand V. c. amauropteryx,intergrade only over a short distance.Another parallel is seenin the form L. s. somaliensis which correspondsin distributionto the small-wingedcoastal V. c. amauropteryx. L. s. senegalaextends eastward to Nigeria but the form V. c. chalybeataintergrades with V. c. neumannias far westas Guineaand Mali. In the otherspecies a clearcorrelation in morphologicaldivergence of races of host and parasiteis not at all evident. V. f. codringtoniand V. f. lusi- tuensisshare the hostL. rubricamhaematocephala, and a singlemorphological form, V. f. nigerrima,occurs through the rangesof L. rubricatacongica, L. ru. haematocephala,and L. [rubricata] landanae,even thoughthis last form of firefinch often is regardedas a distinct species(Mackworth-Praed and Grant, 1963: 646; White, 1963b: 206; Mayr et al., 1968: 330). Nor is a correlationevident in the other firefinch species.L. rh. rhodopareia intergradeswith L. rh. jamesoniin southeasternKenya where the bluish V. purpurascensintergrades with the purplish Tanzania birds, but no mor- 290 ORNITHOLOGICAL MONOGRAPHS NO. 11 phologicallyunique form of indigobirdis evidentin Angola in the range of L. rh. ansorgei,even though the songof ansorgeimay be distinct(Nicolai, 1967: 319). Finally the speciesL. rara, L. larvata, and L. rubricata are differentiatedfrom eastto westin the northernhalf of Africa, but the indigobirds of the V. wilsoni complexare not. A clearcorrelation of taxonof host and parasiteis seenin sometaxa but not in others. A 1:1 correlationbetween subspecies of indigobirdsand their firefinch hosts occurs in only 3 of the 13 indigobirdspecies and subspecies recognized(23 percent). Even those exampleswhich are similar do not necessarilyreflect a similarantiquity of host and indigobirdtaxa. In general the degreeof differentiationamong subspecies of the indigobirdsand their hostsdoes not suggesta commonevolutionary age for each,and the parasites haveevidently differentiated slower than their hosts. 4. Syrupatticspecies are more similarto each other than they are to allopatrictaxa.-• birdsin a singlepopulation lay in the nestsof morethan one firefinchspecies, two distinctbreeding populations, each mating only with birds imprinted to the same firefinch, may develop within an area. Eventuallysome of thesecultural groups may undergogenetic differentiation, but locallyone shouldfind evidenceof crypticspecies relationships. Probable examplesare the partiallyisolated forms of V. wilsoni ("wilsoni," "carnerunensis,"and "nigeriae") in Nigeria,V. f . funereaand V. purpurascensin South Africa, and V. chalybeataand V. purpurascensin Kenya; a similar situation may occurin Angolaand the southernCongo. In size and plumagecolor thoughnot alwaysin foot color the sympatriccognate forms are similar to eachother. This similarityis probablynot a parallel adaptationto common edaphicor climatic factors, for severalreasons. First, litfie correlationof plumagecolor and habitat is evident in the indigobirds. In South Africa V. c. amauropteryxis greenish-bluewhereas the other two speciesare dull purplish-blue.Indigobirds around the edgesof desertsmay be bright green (Senegal), bright blue (Nigeria), bright purplish (Ethiopian rift), or quite dark, dull blue (edge of Kalahari desert); the form "nigeriae"of V. wilsoni is bright greenbut lives mainly in mesichabitats. In size the sub-Saharan birds are small while the Kalahari birds are large. Female plumage is heavilystreaked near the Kalahari desertbut pale and lightly streakednear the Sahara, and juvenilesfollow the same pattern though they are more indistinctlystreaked than femalesin both regions. In addition,in regions where three specieslive together,more often two than three look alike, and this suggestsno stronginfluence of selectionby the environmentover- riding the differencesamong the species.The absenceof any evidentcor- relationbetween morphology and habitat suggeststhat the similarityof local siblingspecies results from their very closephyletic relationship. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 291

5. Local populationsof a singlemorphological form which mimic one host and other populationswhich mimic another.-•If one host becomes .ecologicallyreplaced by another,an indigobirdpopulation parasitic on the first speciesmay surviveby layingeggs in the nestsof a second.As indicated in Tables 8, 9, and 10, within a speciesall indigobirdpopulations that I found in the field were composedin any given area predominantlyof birds which shared the songs of the same foster speciesin their mimicry. An apparentcase of local, recentreplacement of one indigobirdby another kind is found at Kiri, Nigeria. A Pale-wingedIndigobird (form "nigeriae"of V. wilsoni) was taken there in 1907 (Alexander, 1907). Its probablehost L. rubricatawas not seenwhen I visitedKiri in 1968, and it had apparently beenreplaced by L. senegala,which was commonin and aroundthe village. The only indigobirdsseen in 1968 were the usualmimics of L. senegala,the black-wingedV. chalybeata. Probably as man altered the habitat in the intervening60 years,favoring L. senegalaover the other firefinches(see p. 201), the indigobirdV. chalybeatamoved into Kiri from other populations along the Benue and Gongola Rivers and replaced the indigobirdsof V. wilsoni that were imprinted to other firefinches. A few individualswithin a populationin some areas sang an unusual mimetic songunlike that of other individualsof the sameappearance (Tables 8, 10). A changein songmodel and in host in a local populationmay be responsiblefor Nicolai's (1968) report of song mimicry of L. rara by "camerunensis"in Cameroon;I found indigobirdsof similar appearancein Nigeria to mimicL. larvata. Nicolai alsofound "nigeriae"to mimic L. larvata at Ngaounderein Cameroon,whereas I found "nigeriae"to mimic L. rubricata at Panshanuin Nigeria. In the two models compared above, each of the criteria proposed as evidencefor the occurrenceof geographicdifferentiation and of differentiation throughimprinting is supportedby somepositive evidence from the biology and distributionof the indigobirds.Because support for the geographicisolation modelcan be foundin the patternof geographicvariation of indigobirds, the geographicisolation model cannot be disprovedas providing at least a partial explanationfor the morphologicaldifferentiation of this group. By the sametoken, the imprintingmodel explainsbetter someobserved features of indigobirdbiology. The two speciationmodels are not mutuallyexclusive. The geographicisolation model says nothing about the proximatemechanisms involved in change, but only about the distribution of populationsduring part of their history,and the imprintingmodel providesa meansof a switch of species-specificmating signalsregardless of whetherthe switchedpopula- tions may be geographicallyisolated during the time of their geneticdifferentiation. 292 ORNITHOLOGICAL MONOGRAPHS NO. 11

DISCUSSION The socialbehavior resulting from host-specificimprinting and the genetic differentiationof populationsliving in different areas are both causally involvedin the evolutionarychanges indicated by the patternof variationin the indigobirds.Because the notionsof the evolutionaryconsequences of imprintingto thefoster population and of geographicisolation are not mutually exclusive,we may comparethe interactionof thesetwo phenomenain ex- plaining some detailsof the history of evolutionin the indigobirds.The theoreticalproblems involved in the splittingof a singlepopulation into two geneticallydifferent groupsin most kinds of organisms(Maynard Smith, 1966) presentno barrierto the processof sympatricdivergence in the indigo- birdsbecause the discretesong characters to whichyoung indigobirds imprint are learnedfrom anotherspecies, not geneticdifferences inherited from their parents. A seriesof degreesof isolationbetween indigobirds seen in different areasmay provideevidence not only in supportof the role of geographic isolationin speciation(p. 286) but also evidencein agreementwith the importanceof host songdifferences in imprinting. Also, we may compare the ecological,geographical, and behavioral conditionspromoting genetic divergencein the indigobirds.In addition,we may considerthe probable interactionof behavioraland distributionalchanges in producingthe pattern of variation and the aspectsof indigobirddifferentiation that can best be accountedfor by geographicisolation and by imprinting. The effectof imprintingupon mating systems has beentaken into account in a series of mathematical models (O'Donald, 1960a, 1960b; Mainardi, 1964, 1967; Maynard Smith, 1966; Kalmus and Maynard Smith, 1966; Seiger,1967). The set of restrictivegenetic and ecologicalconditions sup- posedlyrequired for establishmentof geneticpolymorphism and subsequent divergenceof mating systemsthat was discussedby Maynard Smith (1966) has been consideredto precludeimprinting as a speciationmechanism (Se- lander, 1969). However, no bottleneckof Mendeliangenetics exists to negate the effects of imprinting on mating systemsand populationdivergence in the indigobirds,because imprinting is not based upon a genetic difference within indigobirds,but rather the indigobirdslearn their signalsfrom other species,the host firefinches. The mechanismproximally responsible for an ecologicaldivergence or exploitationof two independentniches (or foster species)is alsolearning rather than geneticpolymorphism. Maynard Smith (1966: 638) notes as one conditionfor the effectivenessof habitat selection in sympatricspeciation: "the choiceby a femaleof a place to lay her eggs [must depend]not directlyon her genotypebut on her own upbringing." Imprintingappears to determineboth the mating systemand the speciesof foster host for each generation. Complete reproductiveisolation of the 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 293 progenyof a singleinseminated female is plausiblewithin a singlegeneration if the young are imprinted to different speciesof hosts. The effectiveness of imprintingas a mechanismpromoting the divergenceof mating systems thuslies in learning;genetic differentiation of the matinggroups in the present model would follow rather than precedethe separationof a populationinto discretemating systems.In areas with two or more kinds of indigobirds whosemating has beenobserved in the field, mate selectionis highly assortative amongthe birdsimprinted to the samehost species. The ecologicalfea- turesof indigobirdmating systems and host-specificbrood exploitationappear to meet in large part the ecologicalconditions proposed by Maynard Smith for the occurrenceof speciationwithout geographicisolation, and the fact that the indigobirdsimprint to the songsof speciestotally different from themselvescircumvents the difficultiesof the conditionsof geneticpoly- morphismand dominanceof his model for sympatricspeciation. A seriesof stageslike the ones throughwhich two given, hypothetical indigobirdpopulations may have becomedifferentiated to the specieslevel throughimprinting to two speciesof foster firefinchescan be arrayed by comparingvarious mimetic song differencesand degreesof isolation found amongthe indigobirds in nature.(1) In the Malawi V. purpurascens-V.[unerea nigerrirnacomplex, the vocalizationsof the firefinch song models are very similar (sonagrams),and the two kinds of indigobirdsmimicking them are morphologicallysimilar and show little evidenceof being reproductively isolated(Figure 47). The presenceof morphologicallyintermediate specimens in the regionwhere one replaces another geographically suggests considerable introgressionbetween these forms. (2) In theNigerian pale-winged V. wilsoni complex, the purplish "wilsoni" and the bluish "camerunensis"occur together and are partially isolated, as morphologicallythe spedmensfall into distinctclusters with few intermediates(Figures 41, 42, 43). The mimetic songsof these two forms show that each is imprinted to a distinct firefinch species(Audiospectrographs 14, 18, 19). However,a few birdsare imprinted to otherspecies (in NigeriaI noteda bluishbird and a greenish-bluebird each mimickingsongs that the purplishbirds alsosang), and differentlycolored birdsmay mimicthe samespecies, suggesting no permanentisolation among color forms (Figure 41). Femalesof the two forms are morphologically indistinguishable,and my field observationsdid not show any assortative mating amongthe color forms of the pale-wingedcomplex. Birds of one morphologicalform also may mimic different hosts in different regions. (3) In Rhodesiathree kinds of indigobirdslive togetherwithout interbreeding, and eachmimics a differentspecies of firefinch,each with a distinctsong (V. [unerea codringtonimimics L. rubricata, V. purpurascensmimics L. rhodopareia,and V. chalybeataamauropteryx mimics L. senegala). The femalesare well differentiatedmorphologically, and matingis highlyassorta- 294 ORNITHOLOGICAL MONOGRAPHS NO. 11 tive. The parallelsequences between the degreeof songdifferences of the firefinchesand the stagesof differentiationand isolationof the indigobird mimicsprovide the samesort of patternaccepted by evolutionarybiologists-- varyingdegrees of reproductiveisolation in different geographicareas--as evidenceof theeffectiveness of geographic isolation in speciation(Mayr, 1963: 488-489; Alexander,1969: 499). It is evenpossible that differentiationof speciesmight ultimately result from youngindigobirds from a singlefemale becomingimprinted, some on one fosterspecies and someon the other. On mostoccasions when indigobirds switch hosts, the switchprobably is not followedby geneticdivergence of host-specificpopulations. A female indigobirdmight lay in an alternativefoster species'nest if the alternative specieshad a songmuch like that of her usualfoster species. Less often would a female lay in the nest of a firefinch havingdissimilar vocalizations, althoughthis might happenif a female had been stimulatedby watching her own fosterspecies build a nest and she then had beenprevented from layingthere because of the lossof the nestto a predatoror other disturbance. The small effectivepopulation size of indigobirds,as indicatedby the nonmimeticsong dialects (pp. 156-158) mightenhance the likelihoodof a behavioralswitch in mimeticsong and foster speciesbecoming established for severalgenerations, because a shift of a few individualscould often resultin scoresof offspringimprinted to an alternativehost in two or three years (individual female indigobirdsoften lay dozens of eggs each year, judgingfrom morethan 80 ovariesthat I haveexamined from birdsin breeding condition). The low ratesof interpopulationdispersal in eachgeneration indicatedby the local homogeneityof dialectalsong types suggest that local behavioralshifts would not be swampedby immigrantswith the old songs, nor would the shifted populationscatter rapidly into surroundingareas wherethere would be few if any matesimprinted to its new fosterspecies. A local populationof a few scoreto a few hundredindividuals, isolated by distanceas well as behavior,might undergosome geneticdifferentiation as a result of chancealone (Wright, 1969). Such an indigobirdpopulation might alreadyhave a distinctivegenetic makeup, without regardto its dif- ferentbehavior, as a resultof its smalleffective size and isolationby distance from otherconspecific indigobirds. While somegenetic differentiation in a local indigobirdpopulation might develop as a result of chance, as noted above, it seemsmore likely that importantgenetic changes sufficient to differentiatean indigobirdpopulation that has switchedits mimetic songand foster specieswill occur by selection whengeographic isolation accompanies the behavioralswitch. Geographical isolationfrom populationsmimicking the old foster specieswould enhance the behavioralseparation of the indigobirds,especially if the new foster firefinchwere to somedegree separated geographically from the old foster 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 295 species,thereby reducing the chancesfor behavioralswitches back to the old host. Indigobirdpopulations imprinted on differentsets of fostervocalizations mightundergo genetic divergence at a greaterrate if they were separated from eachother geographically, not only throughdifferential selection but also if in isolationthe vocalizationsof the two firefinch speciesdiverged. Whenthe indigobirdisolates then come together again through expansion of the rangeof one or more forms,they wouldbe isolatedto a greaterdegree if the songsof their hostshad becomemore different from other hosts. Conversely,however, character displacement (pp. 282-284) in the hostsongs in areasof sympatrymight more effectively provide the behavioraldifferences for cultural speciationof the viduinesthan would the lack of host song divergencein regionswhere only a singlehost occurred. Geographicisolation may have enhancedthe developmentof genetic differencesamong some indigobirdpopulations, but in the absenceof a behavioralswitch (imprinting) in the parasiticspecies it is unlikely that speciationwould result. Evidenceof the importanceof behavioralchanges in the evolutionarysplitting of indigobirdsis that all kinds of indigobirds thatmimic the samehost species appear to interbreedwith eachother wherever their rangesare in contact. For example,the well-differentiatedforms V. [unerea codringtoni(which probablydiverged from ancestralindigobirds duringa time whenit wasisolated in the southeastAfrican rift highlands) and the purplishsouthern V. [. [unereaand bluish-purplenigerrima all appear to have introgressedwherever they have comeinto secondarycontact. All of theseforms are mimics of Lagonostictarubricata, and hence geographic isolation,accompanied by strongmorphological differentiation but not by a switchin the fosterspecies has not by itself causedthe splittinginto different speciesof indigobirds.Similarly, the isolationof the green-glossed,form "aenea"and the purple-glossed,form ultramarinain northwestand in northeast Africa has not resultedin speciation,because each of thesetwo forms hasintrogressed with the morebluish-glossed forms of V. chalybeatawherever these occur. In all but one of the instances described in the sections on systematicsand variation,the neighboringtaxa of indigobirdsare known to intergradewith each other in every region of contact,provided that the neighboringtaxa mimic the samespecies of host (Figure 50). The possible exceptionis the V. wilsoni-V. [unerea complex,where evidencefor inter- breedingin the Congois equivocal(pp. 261-263). The relativecontributions of acquiredbehavioral changes and geographic isolationin the differentiationof indigobirdsmay also be seenby comparing speciesof indigobirds.In the V. wilsonicomplex, where two or three species possibly might be recognized,the ranges of the three forms "wilsoni," "camerunensis,"and "nigeriae"are largely coextensive from the Upper Guinea regionto Sudan,though only "camerunensis"is known from Ethiopia. The 296 ORNITHOLOGICAL MONOGRAPHS NO. 11 coextensivenessof the rangesof these forms today does not necessarily indicatethat they differentiatedin sympatry. Whether they developedin geographicisolation from other membersof the complex or whether they occurredtogether during their differentiation,the fact that they all extend acrossAfrica indicatesa considerableexpansion of the range in all forms after differentiation. Some, but not all, greenishspecimens are from the areasclosest to the Congoforests, and thesemay have differentiatedin the more humid geographicareas of northernTropical Africa. The large degree of genetic interchangebetween green, blue, and purple indigobirdsin the V. wilsonicomplex may be related to two featuresof African ecologythat forcedthe indigobirdsto switchhosts more than once in their history. The vegetationalbelts of west and north-easternAfrica between the Sahara to the north and the oceanor the equatorialforest to the south are relatively narrow and have undergoneconsiderab}e latitudinal displacementsin the past few thousandsor tens of thousandsof years. Moreau (1966: 42-60) hasdescribed the extensivePleistocene and post-Pleistoceneshifts of moisture regions in west Africa which left evidence of the advances of the desert (sand dunesnow coveredby savannavegetation) and of the wetter periods (relict distributions of roesic biota in the desert). Firefinch distribution indicatesa dependenceon thesemoisture-vegetation belts, and relict firefinch populationsare known in areassuch as the bend of the Niger River where the firefinch virata occurs as an isolated differentiate of Lagonosticta rubricata. When the vegetationzones shifted north or south somefirefinches may have been extirpatedlocally (especiallypopulations of L. rubricam, a speciesof humid habitats) and the indigobird parasitesof these lost firefinchescould have reproducedsuccessfully at the local level only by switchinghosts. The bluish-greenindigobird specimen ("nigeriae" of Lynes, 1924) from Kulme, Darfur, hundredsof miles nor.thof L. rubricata'spresent distribution,may representa relict form left strandedby a dry period. A greenishindigobird at Zaria mimickedL. rara; it alsowas north by about 50 miles of the nearest known L. rubricata. If greenish"nigeriae" originally differentiatedthrough imprintingto L. rubricata (the songmodel at Panshanu,Nigeria), then the greenbirds in the drier areasmay have been derivedfrom relict populations which usedL. rubricatain wetter climaticconditions of earlier years, but which switchedhosts as .thisfirefinch was replacedby the other firefinch species. Secondly,the number of coexistingfirefinches is greater in the Guinea Woodlandvegetation belt than elsewherein Africa; here as many as four speciessometimes live in the sameacre of grassy,brushy streamside habitat. With a high diversity of ecologicallyoverlapping firefinches the chancesof a female indigobird depositingan egg or two in the nest of a secondspecies of firefinchmay be correspondinglyhigh, and the resulting possiblecombinations of offspringof indigobirdparentage and firefinchsong 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 297 tutelagewould produce generations of birdswhich would mate with others of thesame mimetic tradition of song. Southof the equatorthe variousspecies of indigobirdshave ranges less coextensivethan do membersof the V. wilsonicomplex. Because V. chalybeataand V. runeteaeach occur in areaswhere the otherdoes not, they may havebeen geographically separated or parfly separatedduring their evolu.tionaryhistory. V. purpurascenslies largely within the rangeof V. chalybeata,but morphologicallyV. purpurascensis more similarto some formsof V. funereathan to theVillage Indigobirds, and these two probably sharea morerecent common ancestor than eitherdoes with V. chalybeata in southernAfrica. The formsV. funereanigerrima and V. purpurascens are very similarmorphologically and appearto interbreedwhere they occur together,but in Zambiaand Malawi, where they are bestknown, they are largelyseparated geographically with nigerrimaand its fosterspecies L. rubricatain thehigher, wetter plateau regions and purpurascens together with L. rhodopareiain the hotter,drier lowlands. These indigobirds may have beengeographically separated during their earlier history while they were undergoingthe minor morphologicaldifferentiation evident in south-central Africa. Probablynigerrima was more widespread in a wetterperiod in the past and at that time occupieda continuousregion of moisthabitat; now it is somewhatisolated at the higherelevations. Some of the easternRhodesian V. purpurascensare slightlymore bluishthan any V. purpurascensfrom centraland westernRhodesia, which is drier, and thesebluer birds may representthe geneticinflunce of nigerrimastranded in a later dry period whenL. rhodopareiareplaced L. rubricata.Later the lowland purpurascens may have invaded the dryish area and interbred with the switched-over populationof nigerrima; the hybrid swarm may have eventually been swamped by influxof themore purplish purpurascens. No goodmorphological or geneticdata are availableto permitan un- ambiguousretelling of thelocal history of all of theindigobirds in southern Africa,but thesefew examples indicate the importance of behavioralswitches andof geographicisolation and reinvasions acting together to producethe patternof variationseen in the present-dayindigobirds. No singlepattern of distributionand variationseems to me to be explainedbetter by a hypothesisof sympatric differentiation than by oneof somedegree of geo- graphicseparation during the developmentof morphologicalspecies differences. Comparingthe relativesignificance of the consequencesof host-specific imprintingand of geographicisolation in the indigobirdssuggests that im- printingis moreimportant as the initial event in thesplitting of a population andalso as the mechanism whereby some behavioral isolation among groups of geneticallyindistinguishable individuals is maintainedover many genera- 298 ORNITHOLOGICAL MONOGRAPHS NO. 11 tions. The occurrenceof both sympatricand allopatriccomplexes of cognate speciesmay be largelyaccounted for by the "capturing"of novelhost popula- tionsthrougl• imprinting. On the otherhand, geographic isolation is probably as a rule involvedin long-termpermanent genetic changes among the same populationswhich are ch,aracterized by behavioraltraditions of species- specificsong mimicry. It is conceivablethat permanentgenetic changes may occurin the absenceof geographicisolation in the mimetic viduines,but all of the observedinstances of morphologicaldifferentiation are readily ex- plainablein termsof geographicisolation of the songpopulations. Geographic isolationprobably has beenimportant in producingthe patternof speciation observedamong indigobirds with respectto the morphologicaldifferences of populationsnow sympatricand not interbreeding,but it has been less important in the ontogeneticand historicaldevelopment of the reproductive barriers among these species. The reproductivebarriers are, it seems, ontogeneticallythe resultof early experience,and the continuityof the same reproductivebarriers over many generationsis effectedthrough a conservative behavioral tradition. Geographicisolation appears to have been neither necessarynor sufficientfor geneticdifferentiation of the kind resultingin reproductiveisolation, becausethere is no completeisolation among the "species"of indigobirds(Figure 50). Imprinting appearsto be responsible for the assortativemating structureand lack of interbreedingfound in mixed indigobirdpopulations (where the indigobirdscomprise more than one song type and morphologicalform), both when isolationis completeas between V. chalybeataand V. purpurascensin Transvaaland when it is incomplete as in the V. wilsoni complexin Nigeria or in the V. iunerea nigerrima- V. purpurascenscomplex in south-centralAfrica. To whatever degree the indigobirdshave differentiated into groupsthat correspondwith the traditional biologicalspecies concept, imprinting seems very probable to have been a more important event in the origin of reproductiveisolation than has geographic isolation.

SUMMARY The indigobirds(Vidua chalybeataand its relatives)are small,parasitic ploceidfinches and form a group of morphologicallyvery similar kinds that interbreedin someregions of Africa but not in others. In the presentstudy their behavior,and especiallytheir vocalizations,was comparedin the field and their relationshipswere further investigatedin the museumto find the importanceof behavioraldifferences in this complexof siblingspecies. Indigobirdsare brood parasites,and they mimic the songsand callsof the firefinches (Lagonostictaspp.), which are their foster speciesor hosts. Each male generallymimics the songof a singlespecies of firefinch, and in a givenarea most malesof a speciesmimic the samekind of firefinch. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 299

Vocal mimicryis usedin matingbehavior, and field observationssuggest thatfemales select males by the males'mimetic songs. Playback experiments with captiveindigobirds resulted in femalesof two speciesresponding selectivelyto the mimeticsongs of their own species.As the mimeticsongs are probablylearned by the youngindigobirds from their fosterparents, mate selectionas well ashost selection is thoughtto be fixedbehaviorally through theexperience of beingreared by thefirefinches and hearing the vocalizations of the foster family. The indigobirdsare promiscuouslypolygynous. Each male singson a certainbush or tree,and the sameperch is usedthroughout a breedingseason andfrom year to year. At thesecall-sites females visit for matingand as many asfour females mate with a maleon a day. No pairbonds are formed, and the breeding females are not accompanied by themales to thefirefinch nests. Vocal mimicry by themale indigobirds is relatedonly indirectlyto the parasiticbehavior of the birdsas the adult indigobirdsdo not direct their vocalizationsto the firefinchesand the mimetic songis used mainly as an intraspecies signal by the indigobirds. In additionto vocalmimicry the indigobirdssing complex non-mimetic songs. The song repertoire of an individual male includes 12 or more stereotypednon-mimetic song types. All maleslocally may shareall of their songtypes, but theproportion of songtypes shared is lessamong birds 3,200 to 7,200 feet apart than amongcloser neighbors. Most birds more than a fewmiles apart shared no songtypes at all. Playbackexperiments with tape- recordednon-mimetic songs indicated that neither males nor females respond selectivelyto thenon-mimetic songs of theirown dialect area or of theirown species.From the featuresof the local songdialects several features of populationstructure are proposed:(1) indigobirdpopulations grade into one anotherrather than sharplyreplacing one anotherin space; (2) restrictionof songtypes to a smallarea alongwith estimatesof population densityindicates a neighborhoodsize of about100 adultsat the beginning of a breedingseason; (3) the effectivepopulation size, considering mating systems,is ratherless than 100 birds;and (4) dispersalfrom one song-type population to another is uncommon. Femaleindigobirds living in an areawhere only one kind of male occurred were morphologicallysimilar to each other, and in most areaswhere two or threekinds of maleswere found in the field, two or threemorphologically distinctkinds of femaleswere collected. By observingfemales as they visited malesat their call-sitesand by collectingthem, a highdegree of assortative matingwas establishedbetween the forms V. chalybeataamauropteryx and V. purpurascensin SouthAfrica and Rhodesia,V. chalybeatacentralis and V. purpurascensin Kenya,.and V. purpurascensand V. funereacodringtoni in Rhodesia.In theseareas these pairs of formsbehave as distinct biological 300 ORNITHOLOGICAL MONOGRAPHS NO. 11 species. In Nigeria the females of V. chalybeataneumanni likewise were morphologicallydistinct from females mating with males of the forms "wilsoni," "camerunensis,"and "nigeriae," but females of these last forms were indistinguishable.These three are regardedas largely sympatricmor- phologicalforms, without taxonomicstanding, of a singlespecies, V. wilsoni. Indigobirdsof the forms V. chalybeataamauropteryx, V. c. okavangoensis (a new form from northern Botswana), V. chalybeata centralis, and V. c. neumanniin the field mimickedthe firefinch L. senegala,as did captive V. c. ultramarinaand V. c. chalybeata.Morphologically each of theseforms more or lessgrades into the neighboringforms, and they are all regardedas subspeciesof a widespreadcommon species, V. chalybeata. The southernand easternAfrican purplish to bluish-purplebirds, all with whitish feet, V. purpurascens,mimicked L. rhodopareia in Transvaal, Rhodesia, Mozam- bique, Malawi, and Kenya. South African V. funereafunerea mimickedL. rubricataas did greenishV. •. codringtoniin Rhodesiaand Malawi; a geographicallyand morphologicallyintermediate form V. [. lusituensis,described as a new subspecies,also mimickedthis firefinch. These L. rubicatamimics are all forms of V. [unerea. In Malawi, the white-footed,purplish-blue indigobirds of moist habitats mimicked L. rubricata, and these indigobirds(nigerrima) also are regarded as a form of V. [unerea. The distributionsof indigobirdsand firefinchesare in agreementwith the behaviorallydetermined parasite-hostrelationships. In Nigeria, where the forms of the V. wilsoni complex were studied, most "camerunensis"mimicked L. larvata, but one mimickedL. rara. Green, pale-winged"nigeriae" mimicked L. rubricam at Panshanu, but another near Bauchi mimicked L. larvata and one at Zaria mimickedL. rara. All five purplish"wilsoni" heard at Zaria mimickedL. rara. In Nigeria eachcolor form of pale-wingedindigobirds, V. wilsoni,generally mimickeda distinctspecies of firefinch, but somefirefinches were mimicked by all threeforms of indigobirds.Regional differences in the firefinchspecies that are mimickedalso occur in thiscomplex. Although the population systems of indigobirdsdo not everywhereconform to a rigid conceptionof the nature of avian species,the recognitionof four speciesof indigobirdsis useful in comparingtheir behavior and their history. In the museum,1,865 specimensof indigobirdswere examined,including 302 that I collectedin the field, many of them femalestaken with their mates. Analysisof museumspecimens shows that mostof the forms here recognized as speciesdo not interbreedlocally with other species,but in one or more areas each species.apparently does interbreed with another species.All of thespecies intergrade with V. funereain part of the rangeof each. Comparison of the degreeof reproductiveisolation as shownby the paucityor absence of morphologicallyintermediate specimens and the mimetic songssuggest that indigobirdsare stronglyisolated in regionswhere their firefinch fosterers 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 301 havedistinct songs, but isolatedless or not at all in regionswhere the songs of the firefinchesare similar. In Rhodesiathe forms of V. funerea and V. purpurascensare morphologicallyrather different and the songsof their firefinchesL. rubricataand L. rhodopareiaare distinct,whereas in central Malawithese indigobirds intergrade morphologically and theirmimetic songs are very similar. Someevidence from experimentalplayback of songsto captivefemales in breedingcondition support the notion of a relative specificity(rather than an all-or-noneselectivity) in the responsivenessof females to the firefinch songs. Morphologicaland colorimetric characteristics of museumspecimens were describedand compared. Some members of eachindigobird species were less similarto eachother in thesecharacters than to other species.In a few areas, sympatricpopulations of two very similarspecies (such as V. chalybeata and V. purpurascensin Kenya,and V. funereaand V. purpurascensin Trans- vaal) suggestedthe importanceof hostselection and imprintingas an im- portanthistorical event in the initial divergenceof the two species.A female indigobirdlaying her eggsin the nestsof two speciesof firefinchesmight producetwo setsof offspringimprinted to differentsongs and hence reproductivelyisolated from eachother. As indigobirdsare generallyhost specitic in their vocal mimicry,such a rare event might ultimatelyhave led to. the accumulationof sufficientgenetic differences between their descendantsto permittheir recognitionas differentspecies. The smallpopulation size and low rates of dispersalin indigobirdpopulations would tend to minimize swampingof locally differentiatedpopulations and would promotelocal geneticdivergence in groupsof indigobirdsisolated behaviorally by their novel mimetic songs. This cultural speciationneed not have been accompaniedby markedgeographic isolation, although in all casesexamined somegeographic differences were noted among all four siblingspecies, sug- gestingpossible geographic isolation at sometime in their history.Learning of the mimeticsongs of the foster speciesmay have led to the original divergenceof indigobirdspecies as well as to their secondaryinterbreeding. 302 ORNITHOLOGICAL MONOGRAPHS NO. 11

LITERATURE CITED

ABRAHAMS,C.N. 1939. A supposedhybrid whydah. Avicult. Mag., Set. 5, 3: 203- 204. ADOLFFRIEDRICH, DUKE OF MECKLENBEKa. 1913. From the Congo to the Niger and the Nile. An account of the German Central African Expedition of 1910-1911. 2 Vols. Duckworth & Co., London. ADy, P.H. 1965. Oxford regional economic atlas: Africa. Clarendon Press, Oxford. ALEXANDEK,B. 1907. From the Niger to the Nile. New York and London, 2 vols. ALEXANDER,R.D. 1969. Comparative animal behavior and systematics.pp. 494- 517 in Systematicbiology (ed., C. G. Sibley). National Academy of Sciences, Publ. 1692. ANDERSSON,C.J. 1872. Notes on the birds of Damaraland and the adjacent countries of South-West Africa. John van Voorst, London. ANDREW,R.J. 1961. The displaysgiven by passerinesin courtshipand reproductive fighting: a review. Ibis, 103a: 315-348, 549-579. BAKKER,E. M. VAN ZINDEREN. 1964. Pollen analysis and its contribution to the palaeoecologyof the Pleistocenein southern Africa. Pp. 24-34 in Ecological studiesin southern Africa (ed., D. H. S. Davis). W. Junk, The Hague. BANNERMAN,D.A. 1948. The birds of tropical west Africa. Vol. 7. Crown Agents, London. BANNERMAN,D. A. 1953. The birds of west and equatorial Africa. Vol. 2. Oliver and Boyd, Edinburgh. BATES,G.L. 1924. On the birds collected in north-western and northern Cameroon and parts of northern Nigeria. Ibis, 1924: 1-45. BELCHER,C. F. 1930. The birds of Nyasaland. The Technical Press, London. BENSON,C.W. 1948. Geographicalvoice-variation in African birds. Ibis, 90: 48-71. BENSON,C. W. 1953. A checklist of the birds of Nyasaland. The Nyasaland Society, Blantyre. BENSON,C. W. 1963. The breedingseasons of birds in the Rhodesiasand Nyasaland. Proc. XIII Intern. Ornith. Congr., Ithaca: 623-639. BENSON,C. W., R. K. BROOKE,AND C. J. VERNON. 1964. Bird breedingdata for the Rhodesias and Nyasaland. Occas. Pap. Nat. Museums of Southern Rhodesia, 27B: 30-105. BENSON, C. W., AND M. P. S. IRWIN. 1967. A contribution to the ornithology of Zambia. Zambia Museum Papers, No. 1, Livingstone. BENSON,C. W., AND C. M. N. WHITE. 1957. Check list of the birds of Northern Rhodesia. Government Printer, Lusaka. BETmS,F.N. 1966. Notes on some resident breeding birds of southwestKenya. Ibis, 108: 513-530. BREMOND, J. C. 1968a. Recherches sur la semantique et les elements vecteurs d'information dans les signaux acoustiquesdu rouge-gorge (Erithacus rubecula L.). La Terre et la Vie, 1968: 109-220. BREMOND,J. C. 1968b. Valeur specifiquede la syntaxe dans le signal de defense territoriale du troglodyte (Troglodytes troglodytes). Behaviour, 30: 66-75. BROOKE,R. K. 1968. More on Sheppard'sbird work in the Portuguese Vumba. Ostrich, 39: 185-191. BRUNEL,J., ANDJ.-M. THIOLLAY. 1969. Liste pr61iminairedes oiseauxde C6te-d'Ivoire. Alauda, 37: 230-254, 315-337. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 303

CHAPIN,J.P. 1932. The Birds of the Belgian Congo. Part I. Bull. Amer. Mus. Nat. Hist., 65: x + 756 pp. C}•APIN,J.P. 1954. The Birds of the Belgian Congo. Part IV. Bull. Amer. Mus. Nat. Hist., 75B: ix + 846 pp. C}•EKE, A. S. 1969. Mechanism and consequencesof hybridization in sparrows Passer. Nature, 222: 179-180. COLLIAS,N. E., nHl) E. C. COLLinS. 1970. The behaviourof the West African Village Weaverbird. Ibis, 112: 457-480. COOKE,H. B.S. 1964. The Pleistocene environment in southern Africa. pp. 1-23 in EcologicalStudies in SouthernAfrica (ed., D. H. S. Davis). W. Junk, The Hague. CROOK,J.H. 1963. Comparative studieson the reproductivebehavior of two closely related weaver bird species(Ploceus cucullatusand Ploceus nigerrirnus) and their races. Behaviour, 21: 177-232. CaooK, J.H. 1964. The evolution of social organisation and visual communication in the weaver birds (Ploceinae). Behaviour, Suppl. 10:178 pp. CRooK,J.H. 1969. Functional and ecologicalaspects of vocalizationin weaver-birds. Pp. 265-289 in Bird Vocalizations (ed., R. A. Hinde). Oxford University Press. DEKE¾SER,P. L., ANI) J. H. DERtrOT. 1966--1968. Les oiseaux de l'ouest Africain. Vols. 1-3. Initiations et lgtudesAfricaines, Institut Fondamental d'Afrique Noire, Ifan-Dakar, No. xix. DELACOUR,J., nNl) F. EI)MoN•-BLANC. 1934. Monographie des veuves, Part 2. L'oiseau et la Rev. Fran•aise d'Ornith., 4: 52-110. DITTUS, W. P. J., AND R. E. LEMON. 1969. Effects of song tutoring and acoustic isolation on the song repertoires of cardinals. Anim. Behar., 17: 523-533. DIXON, K.L. 1961. Habitat distributionand niche relationshipsin North American speciesof Parus. Pp. 179-216 in Vertebrate Speciation (ed., W. F. Blair). Uni- versity of Texas Press, Austin. EHRLICH, P. R., AND P. H. RAVEN. 1969. Differentiation of populations. Science, 165: 1228-1232. EIBL-EIBESFELDT,I. 1970. Ethology, the biology of behavior. Holt, Rinehart and Winston, New York. EMLEN, J.T. 1957. Display and mate selection in the whydahs and bishop birds. Ostrich, 28: 202-213. EVERITT,C. 1959. The purple whydah. Avicult. Mag., 65: 96. FALLS,J.B. 1963. Propertiesof bird song eliciting responsesfrom territorial males. Proc. XIII Intern. Ornith. Congr., Ithaca: 259-271. FOX, H. M., ANDG. VEVERS. 1960. The nature of animal colours. Sidgwick & Jackson, London. FRIEDMANN,H. 1950. The breedinghabits of the weaverbirds.A studyin the biology of behavior patterns. Ann. Rep. SmithsonianInst. 1949:293-316. FRIEDMANN,H. 1960. The parasitic weaverbirds. U.S. Natl. Museum Bull. 223: viii + 196 pp. FRY, C.H. 1965. The birds of Zaria. IV. Residents,vagrants, and check-list(pas- serines). Nigerian Ornith. Soc. Bull., 8: 91-96. FRY, C.H. 1966. The ecological distribution of birds in northern Guinea savanna, Nigeria. Ostrich., Suppl. 6: 335-356. FUGGLES-CoUCHMAN,N. R., AND H. F. I. ELLIOTT. 1946. Some records and field- notes from north-eastern Tanganyika Territory. Ibis, 88: 327-347. GILLILAND, M. 1962. The flora of the Hans Merensky Nature Reserve. Fauna and Flora (Pretoria), 13: 41-49. 304 ORNITHOLOGICAL MONOGRAPHS NO. 11

GOODWIN,D. 1958. The existenceand causationof colour-preferencesin the pairing of feral and domestic pigeons. Bull. Brit. Ornith. Club, 78: 136-139. GOODWIN, D. 1960. Observations on avadavats and golden-breasted waxbills. Avicult. Mag., 66: 174-199. GOODWIN,D. 1964. Observations on the dark firefinch, with some comparisonswith Jameson'sfirefinch. Avicult. Mag., 70: 80-105. GOODWIN, D. 1965. A comparative study of captive blue waxbills (Estrildidae). Ibis, 107: 285-315. GOODWIN, D. 1969. Observations on two Jameson's firefinches. Avicult. Mag., 75: 87-94. GRANT,P.R. 1966. The coexistenceof two wren speciesof the genus Thryothorus. Wils. Bull., 78: 266-278. GROTE,H. 1928. Uebersicht fiber die Vogelfauna des Tschadgebiets.J. f. Orn., 76: 739-783. GOTTINGER,H.R. 1970. Zur Evolution yon Verhaltensweisenund Lautiiusserungen bei Prachtfinken (Estrilidae). Zeitschr. f. Tierpsychol., 27: 1011-1075. HALL, B. P., AND R. E. MOREAU. 1962. A study of the rare birds of Africa. Bull. British Museum (Nat. Hist.), Zoology, 8(7): 315-378. HALL, B. P., AND R. E. MOREAU. 1970. An atlas of speciation in African passerinc birds. Trustees of the British Museum (Nat. Hist.), London. HALL, M.F. 1962. Evolutionary aspectsof estrildid song. Symp. Zool. Soc. London, 8: 37-55. HALL, P.F. 1969. Hormonal control of melanin synthesisin birds. Gem Comp. Endocrinol., Suppl. 2, 451-458. HARDY,A.C. 1936. Handbookof Colorimetry. TechnologyPress, Mass. Inst. Tech., Cambridge. HARRIS,M.P. 1970. Abnormal migration and hybridization of Larus argentatus and L. ]uscus after interspeciesfostering experiments. Ibis, 112: 488-498. HARRISON,C. J. O. 1956. Some fire-finches and their behaviour. Avicult. Mag., 62: 128-141. HARRISON,C. J. O. 1962a. An ethologicalcomparison of some waxbills (Estrildini) and its relevanceto their taxonomy. Proc. Zool. Soc. London, 139: 261-282. HARRISON,C. J.O. 1962b. Solitary song and its inhibition in some Estrildidae. J. f. Orn., 103: 369-379. HARRISON,C. J.O. 1963a. Jameson'sfirefinch and the dark firefinch. Avicult. Mag., 69: 42. HARRISON,C. J.O. 1963b. An apparent natural hybrid between a combassouand a pin-tailedwhydah. Avicult. Mag., 69: 225-226. HARRISON,C. J.O. 1965a. The nestadvertisement display as a Passer/Ploceidaelink. Bull. Brit. Ornith. Club, 85: 26-30. HARRISON,C. J. O. 1965b. Allopreeningas agonisticbehaviour. Behaviour,24: 161-209. HEINZELIN,I. DE. 1963. Observationson the absolutechronology of the Upper Pleistocene.Pp. 285-303 in African ecology and human evolution (ed., F. C. Howell and F. Bourli•re). Aldine Publ. Co., Chicago. HENNING,W. 1966. Phylogeneticsystematics. University of Illinois Press,Urbana. HEWITT, J. 1931. A guide to the vertebrate fauna of the eastern Cape Province, South Africa. Part 1. Albany Museum, Grahamstown. HUTCHINSON,G. E., AND R. A. MACARTHUR.1959. On the theoreticalsignificance of aggressiveneglect in interspecificcompetition. Amer. Nat., 93: 133-134. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 305

IMMELMANN,K. 1959. ExperimentelleUntersuchungen tiber die biologischeBedeutung artspezifischerMerkmale beim Zebrafinken (Taeniopygia castanotisGould). Zool. Jahrb. Syst., 86: 437-592. IMMELMANN, K. 1962a. Biologische Bedeutung optischer und akustischer Merkmale bei Prachtfinken (Aves: Spermestidae). Verh. Deutsch. Zool. Ges. Saarbriicken, 1961: 369-374. IMMELMANN, K. 1962b. Beitr•ige zu einer vergleichenden Biologie australischer Prachtfinken(Spermestidae). Zool. Jahrb. Syst.,90: 1-196. IMMELMANN, K. 1965. Objektfixierung geschlechtlicherTriebhandlungen bei Pracht- finken. Naturwissenschaften, 52: 169. IMMELMANN, K. 1967. Zur ontogenetischenGesangsentwicklung bei Prachtfinken. Verh. Deutsch. Zool. Ges. GiSttingen, 1966: 320-332. IMMELMANN, K. 1968a. Zur biologischen Bedeutung des Estrildidengesanges. J. f. Orn., 109: 284-299. IMMELMANN, K. 1968b. Besonderheitenin der Gesangentwicklnngjunger Prachtfinken. Umschau in Wissenschaft und Technik, Frankfurt/M., 6: 179-180. IMMELMANN, K. 1969a. Song development in the zebra finch and other estrildid finches. Pp. 61-74 in Bird vocalizations.(ed., R. A. Hinde). Oxford University Press. IMMELMANN, K. 1969b. •ber den Einflnss fri•hkindlicher Erfahrungen auf dio geschlechtlicheObjektfixierung bei Estrildiden. Zeitschr. f. Tierpsychol., 26: 677- 691. IMMELMANN, K. 1969c. C)kologischennd stammesgeschichtlicheBetrachtungen zum Pr•igungsph•inomen.Zool. Anz., 183: 1-12. IMMELMAIqN, K., J. STEINBACHER,AND H. E. WOLTERS. 1965. VBgel in Kafig und Voliere: Prachtfinken. 2nd ed. Verlag Hans Liraberg, Aachen. IRWIN, M.P.S. 1952. Notes on some passerinebirds from Mashonaland, Southern Rhodesia. Ostrich, 23: 109-115. JAMES, H.W. 1925. Birds observed in the Somerset East District, Cape Province, Union of South Africa. Ibis, 67: 621-648. JAMES, H.W. 1970. Catalogue of the birds eggs in the collection of the National Museums of Rhodesia. Trustees of the National Museums of Rhodesia, Salisbury. JOHNSON,N.K. 1963. Biosystematicsof siblingspecies of flycatchersin the Empidonax hammondii-oberholseri-wrightiicomplex. Univ. California Publs. Zool., 66: 79-238. JOHNSTON,R. F. 1956. Population structure in salt marsh song sparrows. Condor, 58: 24-44, 254-272. JOt•V,DAIN, F. C. R., ANDR. SI-IUEt. 1935. Notes on a collectionof eggsand breedinghabits of birds near Lokoja, Nigeria. Ibis, 1935: 623-663. Jt•DD,D.B. 1933. The 1931 I.C.I. standard observer and coordinate system for colorimetry. J. Optical Soc. Amer., 23: 359-374. KALMUS, a., AND S. MAYNARDSMITH. 1966. Some evolutionary consequencesof pegmatypicmating systems(imprinting). Amer. Nat., 100: 619-635. KEAY, R. W.J. 1959. Vegetation map of Africa. Oxford University Press. KORNEat•?, A., AND J. H. WANSCHEa. 1967. Methuen handbook of colour. Ed. 2. Methuen, London. KUN•CEt, P. 1959. Zum Verhalten einiger Prachtfinken (Estrildinae). Zeitschr. f. Tierpsychol., 16: 302-350. KUNKEL, P. 1967. Displays facilitating sociability in waxbills of the genera Estrilda and Lagonosticta (faro. Estrildidae). Behaviour, 29: 237-261. Kt•NZE, H.D. 1961. K/Snnen Weber spotten? Gefiederte Welt, 60: 121-123, 133-136. 306 ORNITHOLOGICAL MONOGRAPHS NO. 11

LnCK, D. 1968. Ecological adaptations for breeding in birds. Meuthen, London. LnN¾ON,W. E. 1957. The comparativebiology of the meadowlarks (Sturnella) in Wisconsin. Publs. Nuttall Ornith. Club, 1:67 pp. LnN¾ON,W. E. 1966. Hybridization in meadowlarks. Bull. Amer. Mus. Nat. Hist., 134: 1-25. LnNYON,W.E. 1969. Vocal charactersand avian systematics.Pp. 291-310. in Bird vocalizations.(ed., R. A. Hinde). Oxford University Press. LelYiON,R.E. 1966. Geographicvariation in the songof cardinals. Canadian J. Zool., 44: 413-428. LEMON, R.E. 1968. The relation between organization and function of song in cardi- nMs. Behaviour, 32: 158-178. LEMON, R. E., riND D. M. SCOTT. 1966. On the development of song in young cardinals. Canadian J. Zool., 44: 191-197. LiSHRL,J. 1955. Beziehungenzwischen Halsband- und Trauerfiiegenschn•ipper(Mus- cicapa albicollis und M. hypoleuca) in demselben Brutgebiet. Acta Congressus Internationalis Ornithologici XI, 333-336. LON•, R.C. 1960. The birds of the Port Herald District. Part I. Ostrich, 31: 85-104. LORENZ,K. 1970. Studies in animal and human behaviour. Vol. 1. Harvard Uni- versity Press,Cambridge. LYNeS, H. 1924. On the birds of north and central Darfur, with notes on the west- central Kordofan and north Nubia Province of British Sudan. Ibis, 1924: 399-446, 648-719. LYNeS,H. 1934. Contributionto the ornithologyof southernTanganyika Territory. Birds of the Ubena-Uhehe highlands and Iringa uplands. J. f. Orn., 82, Sonderheft: 1-147. MnCKWORxH-PRAeD,C. W., riND C. H. B. GRANT. 1949. On the indigobirds of Africa. Ibis, 91: 98-102. MnCKWORTH-PRAeD,C. W., riND C. H. B. GRANT. 1960. Birds of eastern and northeastern Africa. Ed. 2. Vol. 2. Longmans, London. MACKWORTH-PRAED,C. W., ANDC. H. B. GRANT. 1963. Birds of the southernthird of Africa. Vol. 2. Longroans,London. MAINARDI,D. 1964. Effetto evolutivo della selezione sessualebasata su imprinting in Columba livia. Revista Italiana di Ornitologia, 34: 213-216. MAINARDI,D. 1967. Commento ad un recente studio sugli effetti evolutivi dell'ac- coppiamentopreferenziale basato sull'imprinting. Rivista Italiana di Ornitologia, 37: 51-60. MALBRANT,R., ANDA. MACLATCHY. 1949. Faune de l'equateur africain francais. Vol. 1. Oiseaux. Encyclop6dieBiologique 35, Paris. MALZY, P. 1962. La faune avienne du Mali (Bassin du Niger). L'oiseau et la Rev. Fran•aise d'Ornith., Vol. 32, Suppl. Special: 1-81. MARLER,P. 1967. Comparative study of song developmentin sparrows. Proc. XIV. Intern. Ornith. Cong., Oxford: 231-244. MARLER,P. 1969. Tonal quality of bird sounds.Pp. 5-18 in Bird vocalizations.(ed., R. A. Hinde). Oxford University Press. MARLER,P. 1970. A comparativeapproach to vocal learning: song developmentin white-crownedsparrows. J. Comp. Physiol. Psychol.Monogr. 71(2): 25 pp. MARLER,P., ANDM. TAMLIRA. 1964. Culturally transmittedpatterns of vocal behavior in sparrows. Science,146: 1483-1486. MAYNARDSMITH, J. 1966. Sympatric speciation. Amer. Nat., 100: 637-650. MAYR,E. 1963. Animal speciesand evolution. Harvard University Press,Cambridge. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 307

MAYR,E., R. A. PAYNTER,JR., ANDM. A. TRAYLOR.1968. Family Estrildidae. Pp. 306-390 in Check-listof birds of the world, Vol. XIV (ed., R. A. Paynter, Jr.). McLSCHLAN, G., SND R. LIVERSIDGE. 1957. Roberts birds of South Africa. Trustees of the SouthAfrican Bird BookFund, Cape Town. MoNaXD,A. 1951. Resultatsde la missionzoologique suisse au Cameroun: Oiseaux. Mem. Inst. Fran•ais d'Afrique Noire, Centre du Cameroun. SciencesNaturelies, 1: 59-122. MOREAU,R.E. 1948. Ecologicalisolation in a rich tropical avifauna. J. Anita. Ecol., 17: 113-126. MOREAty,R.E. 1950. The breedingseasons of African birds. 1: land birds. Ibis, 92: 223-267. MOREStY,R. E. 1966. The bird faunas of Africa and its islands. Academic Press, London. MOREL,G. 1959. Le parasitismede Lagonostictasenegala (L.) par ttypocherachaly- beata(M[iller). Ostrich,Suppl. 3: 157-159. MOREL,G., A•q•)M.-Y. MOREL. 1955. Au sujetdu parasitismede Lagonostictasenegala L. par Itypochera chalybeataMuller. Alauda, 23: 281-282. MOREL,G., AN•)M.-Y. MOREL. 1962. La reproductiondes oiseauxdans une r6gion semi-aride: la Vall•e du Senegal. Alauda, 30: 161-203,241-269. MOREL,M.-Y. 1964. Natalit• et mortalit6dans une populationnaturelie d'un passereau tropical,le Lagonostictasenegala. La Terre et la Vie, 1964: 436-451. MOREL,M.-Y. 1967. Les oiseauxtropicaux •l•vent-ils autant de jeunes qu'ils peuvent en nourrir? Le cas de Lagonostictasenegala. La Terre et la Vie, 1967: 77-82. MOREL,M.-Y. 1969. Contributionh l'6tude dynamiquede la populationde Lagono- sticta senegalaL. (Estrilidides) h Richard-Toll, (Senegal). Interrelationsavec le parasite Itypochera chalybeata (M[iller) (Viduines). Th•se, Facult• des Sciences de l'Universit6 de Rennes. MORRIS,D. 1958. The comparativeethology of grassfinches(Erythrurae) and manni- kins (Amadinae). Proc. Zool. Soc.London, 131: 389-439. MoYNlI-IAN,M., AND M. F. HALL. 1954. Hostile, sexual, and other social behaviour patternsof the spicefinch (Lonchura punctulata) in captivity. Behaviour,7: 33-76. MtrLLI•A•q,J. A. 1966. Singingbehavior and its developmentin the song sparrow Melospizamelodia. Univ. CaliforniaPubls. Zool., 81:76 pp. NEAVE, S. A. 1907. On a collection of birds from N. E. Rhodesia. Mem. Proc. ManchesterLit. Phil. Soc., 51(10): 1-104. NESVE,S.A. 1910. On the birds of Northern Rhodesiaand the Katanga district of Congoland. Ibis, 52: 78-155, 225-262. NEtmZI•, R. 1929a. Beitr•igesuz Kenntnisder Ploceiden. I. Rachenzeichnungenund Reflexionspapillender Nestjungender Spermestinen,ihr systematischeBedeutung und ihr biologischerWert bei dem Fiitterungsvorgang.Beitr. Fortpflanz. V6gel, 5: 7-17. NEtYNZm,R. 1929b. Zum Brutparasitismusder Viduinen. J. f. Orn., 77: 1-21. NiCOLAi,J. 1959. Familientraditionin der Gesangsentwicklungdes Gimpels (Pyrrhula pyrrhula L.). J. f. Orn., 100: 39-46. NICOLAI,J. 1961. Die Stimmen einiger Viduinen. J. f. Orn., 102: 213-214. N•COLAI, J. 1964. Der Brutparasitismusder Viduinae als ethologischesProblem. Pr•igungsph•inomeneals Faktoren der Rassen-und Artbildung. Zeitschr.f. Tierpsy- ½hol.,21: 129-204. NICOLM, J. 1965a. Der Brutparasitismusder Witwenv6gel. Naturwissenschaftund Medizin, 2: 3-15. 308 ORNITHOLOGICAL MONOGRAPHS NO. 11

NICOLAI,J. 1965b. Prachtfinken.[A 45-rpm recording]75-0932.5. KosmosLehrmit- tel, Stuttgart. NICOLAI,J. 1965c. Vogelhaltung--Vogelpflege.Second ed. Kosmos-Verlag,Stuttgart. NICOLAI,J. 1967. Rassen-und Artbildung in der ViduinengattungHypochera. 1. Hypochera [unerea purpurascens Reichenow--Brutparasit beim Dunkelroten Amaranten Lagonosticta rubricata (Lichtenstein). J. f. Orn., 108: 309-319. NICOLAI,J. 1968. Wirtsvogel-Beziehungender Hypochera-Formencamerunensis und nigeriae. Naturwissenschaften, 53: 654. NICOLAI, J. 1969. Beobachtungen an Paradieswitwen (Steganura paradisaea L., Steganuraobtusa Chapin) und der Strohwitwe(Tetraenura [ischeri Reichenow) in Ostafrika. J. f. Orn., 110: 421-447. NICOLAI, J. 1970. Unterfamilie Witwen. in Grzimeks Tierleben, Enzyklop•idie des Tierrieches,Vol. 9. V•Sgel3. Kindler Verlag, Z[irich, pp. 411-412, 417-419, 429. NIVEN, C. K., ANDP. N. F. NIVEN. 1966. The birds of Amanzi, District Uitenhage, C.P. South African Avifauna Series, 34: 1-55. NOTTEBOHM,F. 1970. Ontogenyof bird song. Science,167: 950-956. O'DONALD,P. 1960a. Inbreeding as a result of imprinting. Heredity, 15: 79-85. O'DONALD,P. 1960b. Assortative mating in a population in which two alleles are segregating. Heredity, 15: 389-396. OmANS,G. H., AN•)M. F. WILLSON. 1964. Interspecificterritories of birds. Ecology, 45: 736-745. ORTMAN,R. 1967. The performanceof the Napoleonweaver, the orangeweaver, and the paradisewhydah in the weaverfinch test for luteinizinghormones. Gem Comp. Endocrinol., 9: 368-373. PAYNE,R.B. 1967. Interspecificcommunication signals in parasiticbirds. Amer. Nat., 101: 363-375. PAYNE,R. B. 1968a. A preliminaryreport on the relationshipsof the indigobirds. Bull. Brit. Ornith. Club, 88: 32-36. PAYNE,R.B. 1968b. Mimicry and relationshipsin the indigobirdsor combassousof Nigeria. Nigerian Ornith. Soc. Bull., 5: 57-60. PAYNE,R. B. 1968c. The birds of toopane woodland and other habitats of Hans MerenskyNature Reserve,Transvaal, South Africa. SouthAfrican Avifauna Series, 56: 1-32. PAYNE, R.B. 1969. Nest parasitism and display of chestnut sparrows in a colony of grey-cappedsocial weavers. Ibis, 111: 300-307. PAYNE, R. B. 1970. The mouth markings of juvenal Vidua regia and Uraeginthus granatinus. Bull. Brit. Ornith. Club, 90: 16-18. PAYNE,R.B. 1971. Paradise whydahs Vidua paradisaeaand V. obtusa of southern and easternAfrica, with noteson differentiation of the females. Bull. Brit. Ornith. Club, 91: 66-76. PINTO,A. A. DA R., ANDD. W. LAMM. 1960. Contribution to the study of the ornithologyof Sul do Save (Mozambique),Part 4. Mem6rias do Museu Dr. Alvaro de Castro, 5: 69-126. POULSEN,H. 1956. Breedingof the Senegalcombassou (Hypochera chalybeataMiill.) in captivity. Avicult. Mag., 62: 177-181. PRmST,C.D. 1936. The birdsof southernRhodesia. Vol. 4. William Clowes& Sons, London. RALPH,C. D., D. L. GRINWICH,AND P. F. HALL. 1967a. Studies of the melanogenic responseof regeneratingfeathers in the weaver bird: comparisonof two species in responseto two gonadotrophins.J. Exp. Zool., 166: 283-288. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 309

RALPH,C. L., D. L. GRINWICH,AND P. F. HALL. 1967b. Hormonal regulation of feather pigmentation in African weaver birds: the exclusion of certain possible mechanisms.J. Exp. Zool., 166: 289-294. RAND,A. L., H. FRIEDMANN,AND M. A. TRAYLoR,JR. 1959. Birds from Gabon and Moyen Congo. Fieldiana: Zoology, 41: 221-410. RANGER,G.A. 1955. On threespecies of honey-guide;the greater(Indicator indicator) the lesser(Indicator minor) and the scaly-throated(Indicator variegatus). Ostrich, 26: 70-87. RIPLEy, S. D. 1962. Aggressiveneglect as a factor in interspecificcompetition in birds. Auk, 78: 366-371. ROBERTS,A. 1922. Review of the nomenclature of South African birds. Ann. Trans- vaal Museum, 8: 187-272. ROBERTS,A. 1924. Synopticcheck list of the birds of South Africa. Ann. Transvaal Museum, 10: 89-195. ROBERTS,A. 1926. Some changesin nomenclature,new recordsof migrants and new forms of S. African birds. Ann. Transvaal Museum, 11: 217-225. ROBERTS,h. 1939. Notes on the eggs of parasitic birds in South Africa. Part IL Ostrich, 10: 100-117. ROBERTS,A. 1940. The birds of South Africa. H. F. & G. Witherby, London. SCHOUTEDEN,H. 1962. La faune ornithologiquedes districts de la Mongala et de l'Ubangi. Contributions {t l'ornithologie de la R•publique du Congo, III. Docu- mentationZoologique, (Tervuren, Belgium), 3: 1-144. SCHOUTEDEN,H. 1963. La faune ornithologique des districts du Bas-Uele et du Haut-Uele. Contributions{t l'ornithologiede la R•publique du Congo, IV. Docu- mentation Zoologique,4: 1-241. SCHOUTEDEN,H. 1964. La faune ornithologique de la Province du Kasai. Contribu- tions h l'ornithologiede la R6publiquedu Congo, VI. DocumentationZoologique, 6: 1-212. SCHOUTEDEN,H. 1965. La faune ornithologique des territories de Dilolo et Kolwezi de la Province du Katanga. Contributions h l'ornithologie de la R6publique du Congo, VIII. Documentation Zoologique, 9: 1-96. SCHUTZ,F. 1965. SexuellePr•igung bei Anatiden. Zeitschr. f. Tierpsychol., 22: 50-103. SCLATER,W.L. 1930. SystemaAvium Aethiopicarum. A systematiclist of the birds of the Ethiopian Region. Part II. Brit. Ornith. Union, London. SEIGER,M.B. 1967. A computer simulation study of the influence of imprinting on population structure. Amer. Nat., 101: 47-57. SELANDER,R.K. 1969. The ecologicalaspects of the systematicsof animals. Pp. 213- 239 in Systematic biology (ed., C. G. Sibley). National Academy of Sciences Publ. 1692. SELANDER,R.K. 1970. Behavior and genetic variation in natural populations. Amer. Zool., 10: 53-66. SELANDER,R. K., AND D. R. GILLER. 1961. Analysis of sympatry of great-tailed and boat-tailedgrackles. Condor, 63: 29-86. SELANDER,R. K., AND R. F. JOHNSTON. 1967. Evolution in the house sparrow. I. Intrapopulation variation in North America. Condor, 69: 217-258. SERLE, W. 1940. Field observations on some Northern Nigerian birds. Part II. Ibis, 1940: 1-47. SERLE,W. 1943. Further field observationson Northern Nigerian birds. Ibis, 85: 413-436. SERLE,W. 1949. Birds of Sierra Leone (Part IV). Ostrich, 20: 114-126. 310 ORNITHOLOGICAL MONOGRAPHS NO. 11

SERLE,W. 1957. A contributionto the ornithologyof the easternregion of Nigeria. Part II. Ibis, 99: 628-685. SHARPE,R.B. 1871. On the birds of Angola. Proc. Zool. Soc.London, 1871: 130-135. SHORT,L.L. 1969. Taxonomicaspects of avian hybridization.Auk, 86: 84-105. SmLEY,C.G. 1970. A comparativestudy of the egg-whiteproteins of passerincbirds. PeabodyMuseum of NaturalHistory Bull. 32: vi q- 131pp., 36 plates. SICEAD,C.J. 1964. The birdsof King William'sTown district. SouthAfrican Avifauna Series, 15: 1-70. SKEAD,C.J. 1967. Ecologyof birdsin the easternCape Province. Ostrich, Suppl. 7: 1-103. SMITH,V.W. 1966. Breedingrecords for the PlateauProvince over 3000 feet, 1957- 1966. Nigerian Ornith. Soc. Bull., 3: 78-91. SMITHERS,R. H.N. 1961. Recordsof birds from the Lake DoT area of the Bechuana- land Protectorate.Occ. Pap.Nat. Museumsof SouthernRhodesia, 24B: 516-534. SMITHERS,R. H.N. 1964. A check list of the birds of the BechuanalandProtectorate andthe CapriviStrip. Trusteesof the NationalMuseums of S. Rhodesia,Cambridge. SMITHERs,R. H. N., M.P. S. IRWIN, ANDM. L. PATERSON. 1960. Additions and cor- rectionsto the checklist of the birdsof SouthernRhodesia. Oct. Pap.Nat. Museums of SouthernRhodesia, 23B: 232-256. SNOW,D.W. 1968. The singingassemblies of little hermits.Living Bird, 7: 47-55. SPARKS,J.H. 1963. Significanceof allopreeningin the red avadavatand its development in other birds. Nature, 200: 281. SPARKS,J. H. 1965. On the role of allopreeninginvitation behaviourin reducing aggressionamong red avadavats,with commentson its evolutionin the Spermestidae. Proc. Zool. Soc. London, 145: 387-403. STEIN, R.C. 1958. The behavioral, ecologicaland morphologicalcharacteristics of two populationsof the alder flycatcherEmpidonax traillii (Audubon). New York State Museum Bull., 371: 1-63. STEIN,R.C. 1963. Isolatingmechanisms between populations of Traill's flycatchers. Proc. Amer. Philos. Soc., 107: 21-50. STEINER,H. 1960. Klassifikationder Prachtfinken,Spermestidae, auf Grund der Rachenzeichnungenihrer Nestlinge. J. f. Orn., 101: 92-112. STONER,D. 1942. Longevityand other data on a captiveEnglish sparrow. Auk, 59: 440-442. SULLIVAN,G.A. 1970. Songof the finchLagonosticta senegala: interspecific mimicry by its broodparasite Vidua chalybeataand the role of songin host socialPontext. Thesis,University of Oklahoma.32 pp. SzlJJ,L. 1966. Hybridizationand the natureof the isolatingmechanism in sympatrie populationsof meadowlarks(Sturnella) in Ontario. Zeitschr.f. Tierpsychol.23: 677-690. THmLCKE,G. 1969a. Geographicvariation in bird vocalizations.Pp. 311-339 in Bird vocalizations(ed., R. A. Hinde). Oxford UniversityPress. THIELCKE,G. 1969b. Die Reaktion von Tannen- und Kohlmeise(Parus ater, P. major) auf den Gesang nahverwandter Formen. J. f. Orn. 110: 148-157. THOMPSON,W. L. 1968. The songsof five speciesof Passerina.Behaviour, 31: 261-287. THOMPSON,W.L. 1969. Songrecognition by territorialmale buntings(Passerina). Anita. Behaviour, 17: 658-663. THORPE,W.H. 1958. The learningof songpatterns by birds,with especialreference to the songof the Chaffinch Fringilla coelebs. Ibis, I00: 535-570. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 311

TINLEY, K.L. 1966. An ecologicalreconnaissance of the Moremi Wildlife Reserve, NorthernOkavango Swamps, Botswana. Okavango Willdlife Society,Johannesburg. 146 pp. TaaYLOR,M.A. 1963. Check-listof Angolanbirds. Publ. Cult. Co. Diam. Angola, 61: 250 pp. T•aYLOR, M. A. 1965. A collection of birds from Barotselandand Bechuanaland (concluded). Ibis, 107: 357-384. TaaYLO•, M. A. 1966. Relationshipsin the combassous(sub-gen Hypochera). Ostrich,Suppl. 6: 57-74. TaaYLO•, M.A. 1967. Cisticolaaberdare a good species.Bull. Brit. Ornith. Club, 87: 137-141. TmtYLo•,M.A. 1968. Family Ploceidae,subfamily Viduinae. Pp. 390-397 in Check- list of birds of the world, Vol. XIV (ed., R. A. Paynter, Jr.). VaN SOMEm•N,V. G.L. 1916. A list of birds collectedin Uganda and British East Africa, with noteson their nestingand other habits. Part II. Ibis, 58: 373-472. VaN SOMEm•N,V. G.L. 1956. Dayswith birds. Studiesof habitsof someEast African species.Fieldiana: Zoology, 38: 1-520. VaN SOMEREN,V. G. L., ANDG. R. C. VaN SOMEREN.1945. Evacuatedweaver coloniesand noteson the breedingecology of Euodicecantans Gmelin and Amadina fasciataGmelin. Ibis, 1945: 33-44. VERNER,J., ANDM. F. WILLSON. 1969. Mating systems,sexual dimorphism, and the role of male North Americanpasserinc birds in the nestingcycle. Orn. Monogr., 9: 1-76. VEILNON,C.J. 1967. JuvenileVidua runeteain family party of Lagonostictarubri- cata. Ostrich, 38: 290. VINCENT,A.W. 1949. On the breedinghabits of someAfrican birds. Ibis, 91: 660- 688. VINCENT,I. 1936. The birds of northern PortugueseEast Africa. Comprisinga list of, and observationson, the collections made during the British Museum Expeditionof 1931-32.--Part X (concluded).Ibis, 1936: 48-125. WALTER,H. 1964. Die Vegetationder Erde in/Sko-physiologischerBetrachtung. Band I. Die tropischenund subtropischenZonen. Gustav Fischer Verlag, 1ena. WARRINER,C. C., W. B. LEMMON,aND T. RAY. 1963. Early experienceas a variable in mate selection. Anita. Behav., 11: 221-224. WnrrE, C. M.N. 1946. The ornithologyof the Kaonde-LundaProvince, Northern Rhodesia.--Part IV. Ibis, 88: 206-224. WnrrE, C. M. N. 1962. Notes on the taxonomy of the indigo birds. Bull. Brit. Ornith. Club, 82: 22-26. WroTE, C. M.N. 1963a. The indigo birds. Bull. Brit. Ornith. Club, 83: 83-88. WHITE, C. M.N. 1963b. A revised check list of African flycatchers, tits, tree creepers, sunbirds,white-eyes, honey eaters, buntings, finches, weavers and waxbills. Gov- ernment Printer, Lusaka. WmTMaN, C. O. 1919. The behaviour of pigeons. Carnegie Institute, Washington. W•CKLER,W. 1968. Mimicry in plants and animals. World University Library, London. W•NTERBOTrOM,I.M. 1965. Note on the purple widow-bird. Ostrich, 36: 140-142. WXNTERBOTTOM,I.M. 1967. A further note on the purple widow-bird. Ostrich, 38: 54-55. Wn•TERaOTTOM,J. M. 1968. A check list of the land and fresh water birds of the westernCape Province. Ann. South African Mus., 53:276 pp. 312 ORNITHOLOGICAL MONOGRAPHS NO. 11

WOLTERS,H.E. 1960. Zur Systematikder Atlasfinken (Untergattung Hypochera der Gattung Vidua, Viduinae, Estrildidae,Aves). Bonn. Zool. Beitr., 11: 19-25. WOLTEaS,H. E. 1961. rrber die verwandschaftlichenBeziehungen der Arten der Viduinae. Vogelring, 30: 12-15. WOLTEaS,H. E. 1963. What is Lagonosticta rhodopareia Heuglin, 18687 Ostrich, 34: 177-178. WOLTERS,I-I. E., AND K. IMMELMANN. 1970. Familie Prachtfinken. in Grzimeks Tierleben, Enzyklop•idiedes Tierrieches,Vol. 9. V6gel 3. Kindler Verlag, Ziirich, pp. 427-445. WaXOaT,S. 1946. Isolation by distanceunder diverse systemsof mating. Genetics, 31: 39-59. Waxcx-xT,S. 1969. Evolution and the geneticsof populations. Vol. 2. The theory of gene frequencies. University of Chicago Press,Chicago. YAMASHINA,Y. 1930. On a new subspeciesof Prosteganurahaagneri. Tori, 6: 11. YeA[[•Uql),J.J. 1959. The nationalexhibition of cagebirds. Avicult. Mag., 65: 49-51. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 313

APPENDIX A

LOCALITY AND OTHER DATA FOR AUDIOSPECTROGRAPHS

Audio- Specimen spectro- catalog Tape graph Locality Date Kind of bird no. information•

la Zaria, Nigeria 26 July 1968 L. senega• not collected 25B: 019

Zaria, Nigeria 19 July 1968 .... not collected 25A: 038

c Norman, Okla. 11 Dec. 1969 R-R/ 34B: 133 (captive)

d Norman, Okla. 10 Oct. 1968 .... nestling - 31A: 397 (captive)

e Norman, Okla. 28 April 1968 .... R-R/ 31A: 057 (captive)

2a Olorgesailie, 22 May 1967 V.c. centralis 4680 22A: 266 Kenya

b 20 mi. E. 27 May 1967 ...... 4691 22II: 217 Kisumu, Kenya

c Merensky, 29 Jan. 1967 V. c. amauropteryx n.c. 9A: 218 Transvaal

d Zaria, Nigeria 1 Aug. 1968 V. c. neumanni n.c. 26A: 170

3a-e Penhalonga, 27 Feb. 1967 V. c. amauropteryx 4443 13: 584-634 Rhodesia

4a Norman, Okla. 13 July 1969 L. senegala - G2:118 (captive)

b Norman, Okla. 16 July 1969 .... - G2:176 (captive)

c Norman, Okla. 18 March 1969 .... G/- 32:004 (captive)

d Norman, Okla. 17 Sept. 1968 .... - 31A: 499 (captive)

e "Leomarin," 18 April 1967 .... - 21:200 Mann, Botswana

f Merensky, 23 June 1967 .... 24:245 Transvaal

Tape number, side, and counter number (from Uher tape recorder position meter). 314 ORNITHOLOGICAL MONOGRAPHS NO. 11

APPENDIX A (Continued)

Audio- Specimen spectro- catalog Tape graph Locality Date Kind of bird no. information•

4g Norman, Okla. 26 April 1968 L. senegala R-R/ 31A: 042 (captive) 5a Monkey Bay, 17 March 1967 V. c. amauropteryx n.c. 16:223 Malawi

b Mererisky, 29 Jan. 1967 ,, ,, ,, 4413 9:126 Transvaal

c Sabi Valley, 4 March 1967 ,, ,, ,, 4469 15:003 Rhodesia

d Zaria, Nigeria 1 Aug. 1967 V.c. neumanni n.c. 26A: 170 6a Merensky, 23 June 1967 L. rh. ]amesoni n.c. 24:237 Transvaal

b Monkey Bay, 17 March 1967 ,, ,, ,, n.c. 16:113 Malawi

c Sigor, Kenya 1 June 1967 L. rh. rhodopareia 4715 23: 042

d Marble Hall, 17 Jan. 1967 L. rh. ]amesoni n.c. 6:346 Transvaal

e Mererisky, 30 Jan. 1967 10A: 004 Transvaal

f Norman, Okla. 5 May 1969 ,, ,, ,, BrG/ 33A: 033 (captive)

g Mererisky, 23 June 1967 24:259 Transvaal

h Norman, Okla. 5 June 1969 ,, ,, ,, /BrG GI: 092 (captive)

i Norman, Okla. 6 Aug. 1969 ,, ,, ,, /BrG G2:414 (captive)

j Mererisky, 23 June 1967 24:261 Transvaal

k Merensky, 30 Jan. 1967 it l, •l n.c. 10A: 003 Transvaal

7a Mererisky, 1 Jan. 1967 V. purpurascens n.c. 5:575 Transvaal

b Sabi Valley, 4 March 1967 ,, ,, 4472 15:196 Rhodesia 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 315

APPENDIX A (Continued)

Audio- Specimen spectro- catalog Tape graph Locality Date Kind of bird no. information •

7c Sigor, Kenya 1 June 1967 V. purpurascens 47O8 23: 076

Merensky, 9 Feb. 1967 .... liB: 240 Transvaal

Merensky, 30 Jan. 1967 ,, ,, 4414 10A: 031 Transvaal

Merensky, 10A: 014 Transvaal

g Merensky, 10A: 031 Transvaal

h Merensky, 10A: 014 Transvaal

Sigor, Kenya 1 June 1967 ,, ,, 4708 23:079

Sabi Valley, 5 March 1967 .... 4489 16:011 Rhodesia

Merensky, 30 Jan. 1967 .... 4414 10A: 074 Transvaal

Merensky, 10A: 031 Transvaal

8a--e Sabi Valley, 4 March 1967 4441 15:327-348 Rhodesia

9a Penhalonga, 1 March 1967 4452 14:371 Rhodesia

Penhalonga, 27 Feb. 1967 444O 13:236 Rhodesia

Penhalonga, 4441 13:363 Rhodesia

Penhalonga, 13:343 Rhodesia

10a Monkey Bay, 18 March 1967 4512 16:583 Malawi

b Monkey Bay, ,, ,, ,, ,, ,, ,, 16:591 Malawi 316 ORNITHOLOGICAL MONOGRAPHS NO. 11

APPENDIX A (Continued)

,4udio- Specimen spectro- catalog Tape graph Locality Date Kind of bird no. information•

10c Monkey Bay, 17 March 1967 V. purpurascens n.c. 16:165 Malawi

d Monkey Bay, 20 March 1967 ,, ,, 4524 17:260 Malawi

e Monkey Bay, 17 March 1967 ,, ,, n.c. 16:147 Malawi

f Monkey Bay, 18 March 1967 .... 4521 17:191 Malawi

g Monkey Bay, 18 March 1967 ,, ,, 4512 16:601 Malawi

h Monkey Bay, 20 March 1967 ,, ,, 4524 17:262 Malawi

11a-h Merensky, 2 Jan. 1967 V.c. amauropteryx n.c. 6:000-044 Transvaal

12a Tzaneen, 25 June 1967 L. rubricam 4471 24:294 Transvaal

b, d, f, see Nicolai, ,, ,, 4531 18:038 g, i-1 1965b

c,h Zomba, 24March 1967 ,, ,, 4531 18:009,036 Malawi

13a-h Lilongwe, 26 March 1967 L. rubricam n.c. 19:082-095 Malawi

14a Panshanu, 30 Aug. 1968 V. wilsoni n.c. 27B: 152 Nigeria ("nigeriae")

b Tzaneen, 3 Feb. 1967 V.f. runetea 4424 10B: 172 Transvaal

c Panshanu, 30 Aug. 1968 V. wilsoni 4946 27B: 049 Nigeria (" nigeriae")

d Tzaneen, 2 Feb. 1967 V.f. runetea 4423 leA: 652 Transvaal

e Panshanu, 30 Aug. 1968 V. wilsoni n.c. 27B: 152 Nigeria ("nigeriae")

f Tzaneen, 2 Feb. 1967 V.f. funerea 4423 leA: 568 Transvaal 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 317

APPENDIX A (Continued)

,'1udio- Specimen spectro- catalog Tape graph Locality Date Kind of bird no. information•

14g Penhalonga, 1 March 1967 V. f. codringtoni 4454 14:410 Rhodesia

h Panshanu, 30 Aug. 1968 V. wilsoni n.c. 27B: 158 Nigeria ("nigeriae") i Tzaneen, 2 Feb. 1967 V.f. funerea 4423 10A: 568 Transvaal

Tzaneen, ,, 10A: 652 Transvaal

k Panshanu, 30 Aug. 1968 V. wilsoni 4946 27B: 060 Nigeria ("nigeriae") 15a Lilongwe, 27 March 1967 V. f. nigerrima 4541 19:243 Malawi

b Lilongwe, 26March 1967 ...... 4534 19:019 Malawi

½ Lilongwe, 27March 1967 ,, 4541 19:278 Malawi

d Lilongwe, 26March 1967 4534 19:004 Malawi

e Lilongwe, 4535 19:115 Malawi

f Lilongwe, ,, 19:118 Malawi

g Lilongwe, 27 March 1967 .... 4541 19:211 Malawi

h Lilongwe, 26 March 1967 .... 4535 19:118 Malawi

16a Zomba, 24 March 1967 V. f. codringtoni 4530 18:167 Malawi (9 mi. S)

Penhalonga, 1 March 1967 ...... 4454 14:409 Rhodesia

Penhalonga, 26 Feb. 1967 4437 12B: 075 Rhodesia

Penhalonga, ,, 12B: 072 Rhodesia 318 ORNITHOLOGICAL MONOGRAPHS NO. 11

APPENDIX A (Continued)

Audio- Specimen spectro- catalog Tape graph Locality Date Kind of bird no. information•

17a Zaria, Nigeria 31 Aug. 1968 L. larvata togoensis 4953 27B: 360 20 July 1968 ,, ,, ,, n.c. 25A: 179 9 Aug. 1968 ...... n.c. 26B: 184

d, f, g see Nicolai, L. I. vinacea 1965b

e Zaria, Nigeria 9 Aug. 1968 L. I. togoensis n.c. 26B: 194

18a Zaria, Nigeria 9 Aug. 1968 V. wilsom 4872 26B: 049 ("camerunensis")

b ,, ,, 29 July 1968 V. wilsoni 4855 25B: 115 ("camerunensis")

c .... 9 Aug. 1968 V. wilsoni 4872 26B: 087 ("camerunensis")

d .... 26 July 1968 V. wilsoni 4860 25B: 642 (" camerunensis")

e .... 9 Aug. 1968 V. wilsoni 4872 26B: 087 ("camerunensis")

f .... 29 July 1968 V. wilsoni 4855 25B: 073 (" camerunensis")

19a Zaria, Nigeria 20 July 1968 L. rara n.c. 25A: 160

1 Sept. 1968 V. wilsoni 4951 27B: 488 ("wilsoni")

1 Sept. 1968 V. wilsoni 4952 28: 144 ("wilsoni")

V. wilsoni ,, 28:134 ("wilsoni")

22 July 1968 V. wilsoni 4884 25A: 326 ("camerunensis")

20a Sabi Valley, 4 March 1967 V. purpurascens 4472 15: 224 Rhodesia

b Penhalonga, 1 March 1967 V. /. codringtoni 4454 14:409 Rhodesia

Tzaneen, 2 Feb. 1967 V.f. funerea 4423 10A: 652 Transvaal 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 319

APPENDIX A (Continued)

Audio- Specimen spectro- catalog Tape graph Locality Date Kind o! bird no. in[ormation•

20d Lilongwe, 27 March 1967 V. [. nigerrima 4541 19:280 Malawi

e Zaria, Nigeria 29 July 1968 V. wilsoni 4855 25B: 118 (" camerunensis")

21a Zaria, Nigeria 30 July 1968 L. ru[opicta 4853 26A: 026

b, c captive from 17 Sept. 1968 ,, ,, R/- 31A: 385, 391 Zaria, Nigeria

22a-e Merensky, I0 Feb. 1967 V. purpurascens 4412 lIB: 261-364 Transvaal

23a Marble Hall, 18 lan. 1967 V.c. amauropteryx n.c. 7:001 Transvaal

b Marble Hall ...... n.c. 7:021 Transvaal

c Marble Hall, 18 lan. 1967 ...... n.c. 7:153 Transvaal

d Marble Hall ...... 4335 7:501 Transvaal

e Marble Hall, ,, ,, ,, ...... n.c. 7:419 Transvaal

24a Monkey Bay, 17 March 1967 V. c. amauropteryx 4505 16:262 Mal•wi

b Penhalonga, 27 Feb. 1967 ,, ,, ,, 4443 13:618 Rhodesia

c Sabi Valley, 4 March 1967 ...... 4469 15:115 Rhodesia

d Merensky, 30 Jan. 1967 .... ,, n.c. 9:328 Transvaal

25a Zaria, Nigeria I Aug. 1968 V.c. neumanni n.c. 26A: 084

b Sigor, Kenya 2 June 1967 V.c. centralis 4711 23A: 215

c Malindi, 11 May 1967 V.c. amauropteryx 4670 22A: 026 Kenya

d Maun, 15Aprill967 V. c. okavangoens• 4591 20A: 344 Botswana 320 ORNITHOLOGICAL MONOGRAPHS NO. 11

APPENDIX A (Continued)

Audio- Specimen spectro- catalog Tape graph Locality Date Kind of bird no. information•

26a Sigor, Kenya I June 1967 V. purpurascens 4708 23:106 b Lilongwe, 27 March 1967 V. f. nigerrima 4539 19A: 152 Malawi c Monkey Bay, 20 March 1967 V. purpurascens 4524 17:266 Malawi

d Penhalonga, 1 March 1967 ,, •, 4452 14:312 Rhodesia

e Sabi Valley, 5 March 1967 .... 4489 16:026 Rhodesia 27a,b Panshanu, 23 Aug. 1968 V. wilsoni 4937 27A: 074, 104 Nigeria ("nigeriae") c Zomba, 24 March 1967 V. •. codringtoni 4530 18:210 Malawi (9 mi. S) d Penhalonga, 2 March 1967 , ,, . 4461 14:624 Rhodesia

e Tzaneen, 12 Feb. 1967 V. •. [unerea 4425 I IB: 395 Transvaal

28a Zaria, Nigeria 31 Aug. 1968 V. wilsoni 4959 27B: 223 (" ca•nerunensis") b ,, ,, 26 July 1968 V. wilsoni 4860 26A: 690 ("camerunensis") c .... 29 July 1968 V. wilsoni 4855 25B: 126 ("camerunensis") d, e .... 1 Sept. 1968 V. wilsoni 4951 28: 004, 27B: 458 ("wilsoni") 29a, c Zaria, Nigeria 31 Aug. 1968 V. wilsoni 4959 27B: 222, 266 ( "camerunensis") b, d .... 22 July 1968 V. wilsoni 4884 26A: 389, 356 (" camerunensis") 30a Maun, 15 April 1967 V.c. okavangoensis4590 20:207 Botswana

b .... 16 April 1967 ...... 4601 20:370 c,d Zaria, Nigeria I Sept. 1968 V. wilsoni 4951 28: 015, 093 ("wilsoni") 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 321

APPENDIX A (Continued)

.4 udio- Specimen spectro- catalog Tape graph Locality Date Kind of bird no. information•

31a-c Marble Hall, 18 Jan. 1967 V.c. amauropteryx n.c. 7: 103, 104, 106 Transvaal

32a Marble Hall, ,, ,, ,, .... ,, 4335 7:350 Transvaal

b Marble Hall, 23 Jan. 1967 ...... 4358 8:525 Transvaal

33a-c Marble Hall, 23 Jan. 1967 V.c. amauropteryx 4358 8: 581,474, 428 Transvaal

34a Olorgesailie, 22 May 1967 V.c. centralis 4680 22:268 Kenya

b Olorgesailie, 9 June 1967 ,, ,, ,, 4734 24:060 Kenya

35a, b Penhalonga, 26 Feb. 1967 V. purpurascens 4436 12: 305, 296 Rhodesia

36 Penhalonga, 27 Feb. 1967 V. purpurascens 4442 13:503 Rhodesia

37 Penhalonga,26 Feb. 1967 V. purpurascens 4436 12A: 280 Rhodesia

38a Maun, 14 April 1967 V.c. okavangoensis4583 19:363 Botswana

b "Leomarin," 18 April 1967 .... ,, 4616 21:094 Botswana

c Maun, 15 April 1967 ..... , 4585 20:092 Botswana

d "Leomarin," 18 April 1967 ...... 4616 21:044 Botswana

39a Lilongwe, 27 March 1967 V./. nigerrima 4541 19:229 Malawi (6 mi. NW)

b Lilongwe, 26 March 1967 ,, ,, ,, 4534 19:005 Malawi (2 mi. SE)

40 Marble Hall, 24 Jan. 1967 V.c. amauropteryx n.c. 8:610 Transvaal 322 ORNITHOLOGICAL MONOGRAPHS NO. 11

APPENDIX B

DISTRIBUTION OF FmEF•CHES AND INDIGOBIRDS Following are the lists of known localities of all indigobirds and host speciesof firefinchesin Africa. The localities are generally spelled as on the specimenlabels; where these differ considerably from other names for the same localities these also are included;for example,Kilima Nascharo(= Kilimaniaro). In parenthesesare regional names for localities which are particularly obscure or which may be confused with other places of the same name; for example, Kabinda (Sankuru Dist.). Vague localities are indicated in quotes; for example, "Senegambia." Gazetteers with latitude and longitude of all but about 50 of these localities have been depositedwith BM(NH) (London), FMNH (Chicago), NMR (Bulawayo), and UMMZ (Ann Arbor), and these may be consulted for the latitude and longitude of each locality as far as the localities were determined. I examined all specimensfrom these localities, with a few exceptions. Some or all Lagonosticta specimensin USNM, MCZ, Museum National d'Histoire Naturelie, MRAC (specimenscollected in 1968 and 1969), Cornell University, Ibadan University, Sencken- berg, and Bonn were kindly identified and listed by their respectivecurators. Lagono- sticta localities from the Albany Museum (Grahamstown) were unavailable. Juvenile L. rhodopareiaand L. rubricata are excludedfrom the list as they are morphologically indistinguishable;differences in primary emargination (see Roberts, 1922: 266; Wolters, 1963: 178; Immelmann et al., 1965: 173, 189) are restricted to the adults. Sight observationsare included and are indicatedby the initials of the observerwhich follow the locality: PB = Peter Britton, JHE • John H. Elgood, CHF ---- C. Hilary Fry, JN: lurgen Nicolai, RBP ---- Robert B. Payne, DW ---- David Wells. Additional localities taken from the literature are indicated by the citation following the locality. All Vidua records, unless otherwise cited, are from specimensthat I examined, from my sight records, or (Nigerian records) from birds caught in nets and examined in the hand by J. H. Elgood.

Lagonosticta senegala ALGERIA: Tamanrassat (Hoggar: captive?). ANGOLA: Arimba (Huila), Catumbella, Cavaco River, Ft. Quilengues, Giraul de Cima (Moca), Huila, Leba, Malambelo, Mulondo, Ponangkuma, Sfi da Bandelta, Tiambe (R. Giraul), Udje (= Uchi), Rio Capit•o (Huila). BOTSWANA: Bathoen Dam, Botletle River, Francistown, Kasane, Kabulabula, Maun, Maun 7 mi. NE, 11 mi. NE, Mochudi, Molepolole 4 mi. NW, Moremi (Tinley, 1966), Nata, Nata 8 mi. S, 5 mi. S, Ngoma, Sepopa 25 mi. SSE, Shashi River, Shorobe, Shorobe 20 mi. W, Tati River. BURUNDI: Bururi, Kitega, Muramuya, Nyanza L. Tanganyika, Usumbura (-- Bujumbura). CAMEROON: Mao Godi, Marua 35 mi. W, Rei Buba, Rei Buba 25 mi. N, Waza, Riggil (Cameroon ?). CENTRAL AFRICAN REPUBLIC: Mundjeffa, Wunnda. CHAD: Bol, Fort Lamy. CONGO: EQUATEUR (UBANGI): Libenge. CONGO: KASAI: Bakwanga, Kabambaie, Kabinda (Sankuru Dist.), Kabotebote (Luluabourg), Katombe, Kelonga (Kalonga), Luebo, Luluabourg, Lusambo, St. Joseph Mission. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 323

CONGO: KATANGA: Bukama, Kabuta, Kadia, Kasaji, Kalombwa, Katobwe, Katumba, Kiabo, Kiambi, Kikondja, Kilwa, Lwiro, Mabwe, Musosa. CONGO: KIVU: Baraka, Beni, Fataki, Idjwe, Kalika, Katana, Katava, L. Tan- ganyika N, NW, Lubango, Lulenga, Ngoma, Rumangabo, RussisiRiver, Tschibati. CONGO: ORIENTALE (ITURI): Boga, Bogoro, Bunia, Dele, Irumu, Kasenyi, Lendju, Mahagi Port, Nyangabo. COTE D'IVOIRE: Bouak6,Boundiali, Korhogo, Man, Nafoun, Ta16r6,Torgokaha, Wamelhoro; also Adiopodoum•, Bingerville, "Sub-Soudainen"(Brunel and Thiollay, 1969). DAHOMEY: Gaya. ETHIOPIA: Abou Beker, Addis Ababa, Adis el Kasun, Ahouna, Akaki River, Aliberet, Arba, Bejook, Belaner, Bergun, Caraina (?), Dagu Delali (Combochia), Dalasire, Dangila, Dawa River, Dawa-Aimola, Debra Wark (Gojam), Dida, Dire Daoua (• Diredawa), Dolo (Juba River), Duletoba,Errer, "EthiopiaS," "Ethiopia SE," Fejambiro,Gallabat, Gardulla, Giamo, Godessa,Gore, Hagar, Hanasch-Ufer, Haramagesee,Hardim L., Harrar (= Harar), HawashRiver, Holata, Irrku, Katyinwaka (= KichinWaka), Keren,Maraco, Maragaz, March River, Mojjio, Omo River, Owara- mulka, Sabata,Sadi Malka, Schebelli,Shoa, Sidu River, Yavello I00 mi. E, Zauday Grar. GAMBIA: Bathurst,"Gambia," "Gambia River," Kuntair. GUINEA: Fouta Djalon, Mamou. HAUTE VOLTA: FadauGurma 25 mi. W, Volta River. KENYA: Amala River, Anasa, Athi River, Bardamat, Baringo L., "British East Africa," Bungoma,Chesegon, Chogorio, Doinyo Narok (• Donje Erok), Elgon Mt., Embu, EscarpmentStation, Fort Hall, Isiolo, Juja, Kabete,Kacheriba (---- Kuchelebai), Kajiado, Kakamega,Kakamega 22 mi. N, Kampi yo Moto, Kapenguria,Karungu, Kavirondo,Kedong River, Khamosi,Kiambu, Kibwezi, Kikuyu, Kikuyusteppe,Kilifi, Kilima Mbogo,Kissaki, Kisumu, Kitui, Lamu,Limuru, Loita Plains,Machakos, Magadi L., Makindu, Malindi (RBP), Mathero's,Meru, Mombosasa,Msara, Muhoroni 4 mi. SE (RBP), Mumias-Yala,Muniuni, Nairobi, NaivashaL., Neng, Ngong, Nyambeni Range,Nyeri, Olorgesailie,Ruaraka, Rusinga, Saba Saba, Sigor, Simba, Sokoke Forest, Takungu,Talek R-JagertekR., Tana R.--Thika R., Tareta, Tharaka, Tumu Tumu, UasoNyiro R. North, UasoNyiro R. South,Wambugu, "West Kenya," "West Pokot," Whei Whei. MALAWI: BIantyre, Bwangu, Chileka, Chinteche, Dowa, Fort Hill, Kota Kota, Mandala-Blantyre, Matope Hill, Monkey Bay (RBP), Mphunzi, Mzimba, Nyambadwe, Pokera Stream,Salima 5 mi. E (RBP), Upper Shire R., Zomba. MALI: Ansongo,Bamako, Kara, Kulikoro, "SoudanFran•ais," Tombouctu. MOZAMBIQUE: Chibababe,Chicowa, Manica, Mapulanguene,Maputo, Maringua, Mossuril, "Mozambique,"Msusu, Mtogwe Mt., Panzilla, Tete, Tete 15 mi. N, Umbeluzi, Urema R.---6orongoza Reserve, Vallee du Ptmgoue, Vila Continha 8 mi. S, Vila Pereira, Zumbo, Zumbo~Kafue. NIGER: Agades, Aouderas, Timia, Zinder. NIGERIA: Agenebode, Agoulerie, Bauchi, Birnin-Kebbi (JHE), Bida (CHF), Bussa(JHE), Dan Gora (CHF), Denge (RBP), Dumbi Woods (RBP), Ganye (RBP), Gassagar,Giri, Gombe junction I mi. S (RBP), Gombe junction 13 mi. W (RBP), Gombejunction 27 mi. W (RBP), GongolaR. 25 mi. E Bauchi,Gusau, Gusau 4 mi. N (RBP), Hadejia (RBP), Ibi, Ilorin (JHE), Jagindi(RBP), Jos (JHE), Kaiama (JHE), Kaltungo15 mi. N (RBP), Kano,Katsina (JHE), Kiri (RBP), Kogum(RBP), Lokoja, 324 ORNITHOLOGICAL MONOGRAPHS NO. 11

Maiduguri (JHE), Maiduguri 45 mi. SE, Malamfatori (CHF), Narode (RBP), Natu L. (CHF), Niger R. at 11ø 30'N, Niger R. at Kaduna R.--Katcha (CHF), Numan, Numan 6 mi. NW (RBP), Pankshin (JHE), Panshanu(RBP), Panyam (JHE), Pategi (JHE), Potiskum (JHE), Riman Zayam (_-- Ziam) (RBP), Samara ExperimentalFarm (RBP), Shagunu (DW), Sokoto (JHE, RBP), Toro (RBP), Yankari (JHE), Yelwa (JHE), Yobe R.--L. Chad (JHE), Yola 100 mi. N, Zaria. PORTUGUESE GUINEA: Bissao. RHODESIA: Binga 23 mi. E, Birchenough Bridge (---- Sabi-Devuli), Bubi Hill, Buffalo Range,Bulawayo, Chipinda Pools, Chirundu,Chityas, Cyrene Mission,Dorowa, Enkeldoorn,Gache Gache, Hot Springs (Sabi-Odji), Khami, Kwenda Mission, Luma, Lundi River 1000', Malapati Drift 3 mi., Malinjinje Pan, Matopo Hills, Mola Camp, Mtoka, Mtekedsa,Nampini, Naunetsi,Nkai, Nkemanda,Pafuri 25 mi. upstream,Penal- verne, Penhalonga, Ruenya River, Rugiruhuru R.--Sijaria Mts., Sabi R. Lundi R., Sabi Valley Experimental Station (RBP), Salisbury, Salisbury 30 mi. E, Samalema Gorge, Sashi R.-•Shashani R., Sebungwe,Selukwe. Sengwa 11 mi. W, Sentinel Ranch, Tebekwe River, Turgwe River, Urntall, Umvuma, Victoria Falls 14-15 mi. W, Victoria Falls 30 mi. W, Victoria Falls 45 mi. W, Wankie, "Zambesi," Zambezi R. (15ø39'S, 30025'E), Zambezi R. (17ø45'S, 24ø12'E), Zambezi R. (15ø40'S, 29ø35'E). RWANDA: Akanyaru R., Astrida, Gabiro, Gisagara, Kagera, Kibingo, Kisenyi, "Kivu," "Kivusee,"Nsasa, Nuyundo, Rubona, Ruzizi. SENEGAL: Dagana, Dakar, Diourbel, Kirtaonda, Nianing, Richard-Toll, "Senegal," Thies, Tili-bu-Baker. SIERRA LEONE: Rotifunk, King Tom. SOMALIA: Bardera, Fanole, Mogadiscio,Salake. SOUTH AFRICA: CAPE PROVINCE: Barkley West, "Cape," Committees,de Rust (Oudtshoorn), Kimberley, Koegasbrug,Prieska, Sydney-on-Vaal,Vryburg. SOUTH AFRICA: NATAL: Bamangivade,Candover, Hibberdene, Mkuzi, "Natal," Otobotini, Pongola River, "Pt. Natal," Umfolosi. SOUTH AFRICA: ORANGE FREE STATE: Bloemfontein (3 localities). SOUTH AFRICA: TRANSVAAL: Bloemhof, Crocodile River, Elim (Zoutpans- berg), Hector Spruit, Irene, Klaserie R.--Olifants R., Kondowe (RBP), Louw's Creek (RBP), Malamala (Newington), Marble Hall Fisheries Station (RBP), Merensky Reserve, Mokeetsi, Morale R. (_-- Mutale), Newington, Pretoria, Rhenosterkop, Rhenosterpoort,Reitspruit (Marico), Rustenburg,Spruytskloof. SOUTH WEST AFRICA: Andara 15 mi. E, Andara 30 mi. W, Eupupa, Kabuta, Kapalm, Ondonga,Oshikango, Ovoquenyama, Rua Cana, Sambiu, SwaartboisDrift, Viool's Drift. SUDAN: Akoua, Atbara, "Bahr-el-Gebel,"Bahr el Ghazal, Bahr el Shagal, Bahr el Zeraf, Berber, Binue, Buram, Dongola, Duem (= E1 Duere), Dulgo, E1 Fasher, Fazaglo, Gaz-abu-Gumar, Gedaref, Gezirat al Fil, Gondokoru, Jebel Ahmeh Alga, Jebel Mara, Juba, Kallokitting, Kamisa, Khartoum, Lado, Melut, Merowe, Meshra el Rek, Musran Island, Naikhala, Rejaf, Roseires,Salimat el Alimat, Shendi, Shereik, Singa, Tauila, Toni, Wadi Naja, Wau. SWAZILAND: Hloye River N. Maloma, Ingwavuma R., Lubuli, Ranches Ltd., Stegi, Umbeluzi R.--Mlawala Station. TANZANIA: Bagamoio, Belun, Bismarkburg(= Kasanga), Bukoba, Dar-es-Salaam, Dongo, Engare Nairobi, Ihoho Forest, Iringa, Kakoma, Karema, Kibaya, Kiduna, Kigoma, Kilima Nascharo (: Kilimanjaro), Kilosa, Langenburg,Luweyia R. (---- Ruipa R.), Magononi am Rufu, Manyara (Lake), Mara River, Marangu, Matengo Plateau, 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 325

Mbuguni, Meru, Mikindani, Mikumi National Park (Nicolai, 1967), Mkomasi, Mnazi, Mortdull, Morogoro, Moshi, Ngomingi, Ngare-Mt., Njombe, Nyanza (L. Tanganyika), Pangani, Rombo, Rukwa, Rukwa~Mamba, Rukwa-Tumba, Tabora, Tindi, Tokuyu, Ujiji, Ukerewe, Uleia, Usambara, Weru Weru Ravine (Kilimanjaro). TOGO: Paio, Porto Seguro, Sebbe. UGANDA: Ankole, Arua, Bahr el Djebel, BudongoForest, Bududu, Buœundi, Butiaba, Buyala, Entebbe, Gondokoro, Kabale, Kabula Muliro, Kabunyala, Kampala, Katogo, Kibiro (= Kibero), Kisingo,Kyetume, Lubilia River, Luma, Masaka, Masindi, Mokia, Mondo, Mpanga Forest (= Kibale Forest), Mpumu, Mtesas (= Kwamtessa),Muanda L., Mubendi, Mushongero,Nyanza of Albert-Edward L., Ruimi Stream, Ruwenzori 3400', Sanja Mt., Semliki Valley, Sezibwa, Toro, "Uganda." ZAMBIA: Balovale (PB), Bulaya, Chama-Lundazi, Chiengi, Chilanga, Choma (Choma), Choma (Mweru), Fort Jameson, "Kaœue R.," Kama, Kapindi, Kaputa (Mweru), Katuta, Kazungula, Livingstone, Lochinvar (3 localities), Luangwa Valley (7 localities), Luano Valley, LuœupaRiver, Lumesi-Lundazi, Lundazi, Lusaka, Machili River, Mambova, Mankoya, Mazabuka, Mongu, Mumbwa, Munyamadzi River, Munyumbwe, Mupamadazi R. (12ø12'S, 31ø45'E; 12ø37'S, 32ø07'E), Mushelelwa, Nche- lenge, Ndola, Ng'ambwe Rapids, Ngwezi River, Ntengo (m Wamuna), Petauke, Senanga, Sesheke Boma, Shangombo, Sikongo, Simamba, Sinjembela, Sumbu, Zambezi R. near Lundi R.

Lagonosticta rhodopareia ANGOLA: Bongo River, Cabeta de Ladroes, Chingoroi, Dondo, Elandswater (Benguella), Fuima, Ft. Quilenges,Hanha, Huila, KabisomboRiver, Leba, Pungo Andongo, Sfi da Bandeira, Vila Flor. BOTSWANA: Francistown, Kasane, Kabulabula, Maun (RBP), Moremi (Tinley, 1966), Ngoma, Nokaneng,Sepopa, Shakawe, Shashi River, Tati River, Toten 14 mi. W (RBP). CONGO: KATANGA: Ganza, Kabengere, Kabalo, Kaluli River, Kasiba, Kinia (Marungu), Luœira(Kaswabilenga), Mabwe, Masombwe,Munoi. CONGO: LEOPOLDVILLE (KINSHASA): Matadi. ETHIOPIA: Bakora, Bodessa,Gardulla, Mega, Sagan R., Tertale, hr. Yavello 4000', 5000', 6000'. KENYA: Bura (Tana River), Bura (Teita), Doinyo Narok (= Donje Erok), Fort Ternan, GessimaRiver, Isiolo 8 mi. S, Kaimosi, Kajiado, Kapenguria,Kibwezi, Kilgoris, Kiliœi, Kitui, Mombasa, Mombasa Hills I0 mi. W, Mt. Garguez, Sagala-Teita, Sigor, Tareta, Tsavo, Urguess,Voi, Voi-Sagalla, Yala R. MALAWI: Bwangu, Chikwawa, Chiromo, Chisempere, Cidasonga, Fort Iohnston, Kanyimbe, Liwonde, Lodjwa, Monkey Bay (RBP), Mpata, Mzimba, Nakumba (: Nankumba), Njakwa, Ntakataka, Pokera, Sori, Symon's,Vintukutu, Zoa Falls, Zomba, Zomba 9 mi. S (RBP). MOZAMBIQUE: Ile, Manica, Maringua, Msusu, Panda, Rova, Santaca, Tambara Fort 16 mi. E, Tete 60 mi. N, Tica 25 mi. S (hr. Buzi R.), Umbelusi, Vila Vasco de Gama, Zinave, Zumbo. RHODESIA: Anglesea Farm, Balla Balla, Bedja Dip, Bembesi, Bikita, Bubi Hill, Bulawayo, Chewore R-Zambezi R., Chirinda, Chirundu, Chityas, Cyrene Mission, Essexvale, Essexvale-Bulawayo, Fort Rixon, Gairezi, Gurrugurru, Hot Springs (Sabi- Odji), Kana, Lamorna, Lonely Mine, Lundi River I000', Lusitu River 1200' (RBP), Magunje-Urungwe,Makwiro, Malapati Drift 3 mi., Malimasimbi, Matebe Hills, Matopos Mission (Matopo Hills), Mazoe Bridge, Mazohwe R. (Matopos), Mkien Farm, Mrowa, 326 ORNITHOLOGICAL MONOGRAPHS NO. 11

Mtoka, Mutema, Nampini, Naunetsi, Nata River, Nyahuvu, Nyamandholovu, Penhalonga 2 mi. S (RBP), Ramaguabane R.-•Shashi R., Rugiruhuru R.---Sijaria Mts., Rusape, Sabi Valley Experimental Station (RBP), Salisbury, Sanyati, Sebungwe, Selukwe, Shangani, Silozwane, Turgwe R., Turkmine, Umguza Forest, Urntall, Umvuli R., Umvuma, Victoria Falls, Victoria Falls fifth gorge, Victoria Falls 14-15 mi. W, Victoria Falls 30 mi. W, Wankie, Whitewaters, "Zambesi," Zambezi R. (15ø39'S, 30ø20'E). SOUTH AFRICA: NATAL: Ingwavuma, Mapuba Road, Mkuzi, Ndumu, Umfolosi, Waterberg. SOUTH AFRICA: TRANSVAAL: Blouberg, Brits, Brits 6 mi. SW, Hamanskraal, Hector Spruit, Houtbosrivier, Irene, Klaserie R.--Olifants R., Kondowe (RBP), Klein Letaba, Leydsdorp, Malamala (Newington), Magaliesburg, Marble Hall Fish. Station (RBP), Mariqua River, Merensky Reserve, Modderfontein, Mokeetsi, Motale R. (= Mutale), Newington, Northampton, Olifants River (K.N.P.), Pretoria, Rietspruit (Marico), Rustenburg, Settlers, Sand River. SOUTH WEST AFRICA: Eupupa, Swaartbois Drift. SUDAN: Boma Plateau, Towat. SWAZILAND: Assegai River, Hloye R.--N. Maloma, Komati R. nr. Bagelane, Lubuli, Nsoko, Stegi, Umbeluzi R.--Mlawula Stream. TANZANIA: Dodoma, Iringa, Kijango, Kilosa, Kisigau, Kunshinowi, Lolkisale, Mberera Hill, Mlali, Morogoro, Mwanasomano, Namalungo, Songea, Wemba, (-t-2 illegible localities). UGANDA: Moroto, Moroto Mt. ZAMBIA: Barotseland (17ø15'S, 24ø06'E), Chadizi, Chalimbana, Chilanga, Chi- pongwe, Chisomo, Chitungulu, Choma (Mweru), Fort Jameson,Kafue National Park, "Kafue-Zomba," Kapelembe, Kapindi, Kasaba, Kasama, Katombora, Kazembe, Ka- zungula, Kundabwika Falls, Livingstone, Lochinvar (3 localities), Luangwa Valley, Luapula River, Luanshya, Lundazi, Lusu Rapids, Machill River, Marble Hill Camp, Mazabuka, Mbala, Membe Stream, Mfubakazi, Miliyoti, Mumbwa Boma, Munyamadzi River (2 localities), Musaya Stream, Mutinondo River, Nyanje, Rutunsa, Sakargo, Sesheke Boma, Sumbu.

Lagonosticta rubricata (including landanae) ANGOLA: Ambaca, Canhoca, Cassai R., Chinchonxo, Chitau, Dala Ango, Dugue de Braganca, Fazenda Jerusalem,Ft. Don Carlos I (= Tembo Aluma), Gabela, Gabela 12 mi. SW, Golungo Alto, Landana, Luau River, Luhanda, N'Dala Tando (= Vila Salazar), Noqui, Pedreira, Pungo Andongo. BURUNDI: Musigabi, Usumbura (= Bujumbura). CAMEROON: Abong Mbang, Ankonolinga, Babadjou, Bafia (= Nnafi), Bamenda, Garou Boulai, Garua, Ibofi, Kumbo-Bamenda, Makondo Mafifigi, Ncongsamba, Ngikinda, Njassi, Ribao, Yaounde, Yoko (Monard, 1951), Yoko 50 mi. S, Yukuba. CENTRAL AFRICAN REPUBLIC: Mission Jean Dybowski (Kemo), Oberes Sannagebiet. CONGO (BRAZZAVILLE): Djambala, Haute Sangha. CONGO: EQUATEUR (UBANGI): Bobito, Bobutu, Boyasegase,Bozene, Bwa- manda. CONGO: KASAI: Bakwanga, Chikapa (Tshikapa), Dimbulembembwe, Gandajika, Kabambaie,Kabotebote (Luluabourg), Kasana,Kasende (Kasendi), Luebo, Luluabourg, Lusambo, Merode, Mkulwa, Ngombe, St. JosephMission. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 327

CONGO: KATANGA: Campia (MarunguMts.), Dikulwe Valley, Elizabethville (= Lumumbashi),Jadotville, Kaboko Mt., Kansenia,Kasaji, Kasangala,Kasapa, Kazembe,Lufira (upper valley), Munoi, Pelenge,Sakania, Shila-Tembo. CONGO: KIVU: Baraka, Beni, Bionga, Butembo,Ibachilo, Idjwe, Kahakaviro, Kamituga, Kananda, Kasongo,Katana, Kitutu, "L. Kivu," Lwiro, Mfumbira, Mufua, Mulembe,Namoya, Ngoma, Nyarukwangura, Tschibati, Uvira. CONGO: LEOPOLDVILLE(KINSHASA): Bokalakala,Boma, Kisantu (Bas Congo), Kunungu,Kwamouth, Leopoldville (= Kinshasa),Matadi. CONGO: ORIENTALE (ITURI): Boke, Djalasinda, Etembo, Irumu, Kasenyi, Mahagi Port, Mt. M6, Semliki. CONGO: ORIENTALE (UELLE): Bosodula,Faradje, Garamba, Kasale River, Kodja Hill (Gaima Range), Niangara,Nzoro (• Vankerchhovenville). COTE D'IVOIRE: Beoumi,Bouak6, Bou R. (S. of Kadioha); also Korhogo, Sipilou (Brunel and Thiollay, 1969). ETHIOPIA: Alghe, Arero, Bakora Chercher L., Chumwugar, Dobbana, Giamo, Gomma, Gummaro, Irrku Jawaha (= Yawaha), Maraco, Tana L. (= L. Tsana, L. Sanne). GABON: Monila, Tchibanga. GHANA: Accra, Ejura, Fanti, Kintampo, Mampong. GUINEA: Bossu (near Mt. Nimba). KENYA: Barsaloi,Chiyulu Hills, Chuka-Embu,Ithanga Hills, Jula Farm (Athi R.), Kakamega,Kakamega Forest Station, Kapenguria, Kericho 4 mi. E (RBP), Khamosi (Kaimosi), Kiambu, Meru, Nairobi, Ngong, Njoro, Rongai-Mau,Saba, Suna, Thika, Ukamba, Yala R. MALAWI: Chididi Hills, Chididi Mission, Chididi Stream, Chinteche, Chinteche 10 mi. W, Chitofu, Cholo, Dedza, Kabehe Hill, Kaluwa Hill, Kadukaduka,Kongwe, Kota Kota, Kota Kota 20 mi. W, Lilongwe(2 and 4 mi. E) (RBP), Lilongwe,Malosa Mt., MangocheMt., Mbabzi (Lilongwe) (RBP), Mlanje, Mphunzi, Mwanjati Hill, Mzimba, Ncuce, Nyika Plateau,Setala, Tengowapyoza, Vipya Plateau,Zomba, Zomba 9 mi. S. MALI: Bamako, Fiko, Kulikoro, Sanga. MOZAMBIQUE: Beira, Beira 6• mi., Chemezi, Estatuane,Fingoe, Furancungo, GorongozaMr., Jofane,Loren9o Marques, Luabo, Mapicuti, Mocuba l0 mi. W, Muanza S. Inwamingo,Mwira L., Namaacha, Netia, Rova, Santaca,Tambarara, Umbeluzi (• Umberusi), Vila Vasco de Gama, Zobue, Zumbo. NIGERIA: Aria Hills (Serle, 1940), Agwada (Serle, 1940), Ankpa, Aza, Enugu, Jos (JHE), Kaduna (JHE), Kafanchan (Serle, 1940), Kagoro Hills (Serle, 1940), Kogum (RBP), Panshanu (RBP). PORTUGUESE GUINEA: Bissao, Gunnal. RHODESIA: Haroni R., Himalaya, Honde R.•Mtarazi R., Inyanudazi River, Lusitu R. 1200', Melsetter, Muyuinga R. 2300', Penhalonga2 mi. S, PungweR. 2200', Rukonde Hill, Selinda, Umtali, Vumba. RWANDA: Akanyam R., Gakoma, Kisenyi, Mukada, Rand des Rugege Forest, Shaba. SIERRA LEONE: Benguema, Bo, Freetown, Gloucester, Knoll, Rokupr, Tungie. SOUTH AFRICA: CAPE PROVINCE: Adelaide l0 mi. N, Alexandria, Bolo-Kei Bridge, Committees,East London, Embotyi, Ft. Beaufort, Grahamstown,Gwanga (Peddie), Inthlyoyana, Kei Bridge, Kidd's Beach, King William's Town, Knysna, 328 ORNITHOLOGICAL MONOGRAPHS NO. 11

Kokstad 30 mi. E, Ngqueleni, Patensie, Pirie, Pt. St. Johns, Qora R. Bridge, Xora R. nr. mouth, Slippery Drift, Uitenhage, Umtamvuma Bridge. SOUTH AFRICA: NATAL: Coorobe Barton, Durban, Elandskop,Gollel, Gwaliweni, Hella Hella, Hibberdene, Howick, Ifafa R., Ingwavuma, Insusie Valley, Karkloof, Merrivale, "Natal," New Hanover, Kwabonambi, Melmoth, Ngoye Forest, Otobotini, Penicuik, Pietermaritzburg,Pinetown, Pongola R., "Pt. Natal," Richmond, Shongweni Dam, Spltzkop (Karkloof), Sweetwater, Table Mr., Ubombo, Umhlanga, Umhloti, "Umzilas Kingdom," Weenen. SOUTH AFRICA: TRANSVAAL: Barberton, Carolina, Legotgot (Barberton), Malamala (Newington), Modderfontein, Mokeetsi, Pienaars River, Pretoria, Rusten- burg, Tzaneen 6 mi. E, Wistern, Wakkerstroom 10 mi. E, Woodbush, Worcester Mine. SUDAN: Boma Plateau, Char 5 mi. N, Didinga Mrs., Imatong, Leone, Lolengi, Lotti Forest, Lotuke Mt., Nagichot, Nalagedi, Sakure. SWAZILAND: Ingwavuma R., Komati R. near Bagelane, Nsoko, Ranches Ltd., Stegi, Umbeluzi R.--Mwawula Stream. TANZANIA: Bukoba, Dar-es-Salaam, Kigoma, Kilima Nascharo (---- Kilimanjaro), Kilosa, Kitungulu, Kunbosa Forest, Lumbuti, Machame, Marangu, Marengo Plateau, Matogoro, Matombo, Matombo-Kizebbe, Mikindani, Mikumi National Park (Nicolai, 1967), Monduli, Moshi, Namalungo, Ngua, Njombe, Songea, Songea 75 mi. N, South Ulanga, Turiani, Ujamba, Uluguru Mts., Ussumi (Karagwe), Luweyia R. ---- Ruipa R., Hanang Mr. 10,000'. UGANDA: Ankole, Bugoma, Bugomba, Bundibuggio, Bwamba Valley, Entebbe, Entebbe 12 mi. W, Kabale, Kampala 30 mi. NW, Katwe, Kayonza, Kyetume, Masindi, Moyo, Mpanga Forest ---- Kibale Forest, Mpumu, Mubendi, Nkarara, Ntotoro, Pader, Sezibwa, "Uganda." ZAMBIA: Abercorn, Chilanga, Chimpili Plateau, Danger Hill, Fort Jameson, Fort Roseberry, Ft. Roseberry 18 mi. ENE, Fwaka 15 mi. NW, Kabompo Boma, Kasempa, Kawambwa 8 mi. ESE, Kawambwa 15 mi. N, Kitwe, Kondolilo Falls, Luamala R., Luanshya, Lusiwasi Lake (Serenje), Lusenga Plain, Luwenge, Mankoya, Marble Hill Camp, Mayau, Mkushi R., Mlembo, Mpika Airfield, Mporokoso 30 mi. ESE, Mterize R., Mttrnbwa Boma, Musolu R., Mutanda, Mwinilunga, Nalusanga,Ndola, Ntambu, Nyika Plateau, Salujinga, Solwezi, Luangwa Valley--Mpika.

Lagonosticta larvata

CAMEROON: Mayo Sala (Monard, 1951), Tibati. CENTRAL AFRICAN REPUBLIC (REPCENTRAFRICAINE): Bamingui R., Bosum, Kaja Djerri, Majim, "Ostkamerun," Ratu. CHAD: Ir6na. CONGO: ORIENTALE (UELLE): Faradje, Garamba. COTE D'IVOIRE: Bav6-Komo6; "Korhogo latitude" (Brunel and Thiollay, 1969). ETHIOPIA: Bunio, Dura R., Gallabat, Gelongol, Gumad R. 100 mi. SW L. Tana, Gumad R. 130 mi. SW L. Tana, Koko, Koscha, Sidisto, Takazai Valley. GAMBIA: "Gambia," Makka Niumi. GHANA: Binduri, Brumassi, Dokonkade, Ijura, Gambaga, Kintampo, Musarka, Wa. NIGERIA: Anara Forest Reserve (CHF), Ankpa, Bauchi (JHE), Bauchi 25 mi. W (RBP), Bima, Biu-Bauchi, Enugu, Hinna, Iseyin, Kaduna (JHE), Kogum (RBP), Kontagora (CHF), Lokoja, Maska Dam (RBP), Ruan Gizzo, Shagunu (DW), Tatara, Yankari (CHF), Yola 98 mi. NW (RBP), Zaria. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 329

PORTUGUESE GUINEA: Gunnal. SENEGAL: Niokolo-Kabo (Hall and Moreau, 1962: 378), "Senegal." SIERRA LEONE: "Sierra Leone," Tumbo. SUDAN: Boma Plains, Chak Chak, Famaka, Kulme, Roseires, Roseires25 mi. S. TOGO: Aledjo, Nuatja. UGANDA: Kamchuru, Moyo.

Lagonosticta rara CAMEROON: Banyo, Koubadje (Monard, 1951), Mayo Sala (Monard, 1951), Mboula, Ngaoundere(Nicolai, 1968), Ribao Plain, Sakdje (Monard, 1951), Tibati, Tibati-Ngambe, Tibati-Tingura, Tibati-Yoko, Yoko. CENTRAL AFRICAL REPUBLIC (REPCENTRAFRICAINE): Bangui, Bosum, Kaja Djerri. CONGO: ORIENTALE (ITURI): Ishwa (: Ischwa), Mahagi, Mahagi Port, Niarembe, "N. of Albert Edward." CONGO ORIENTALE (UELLE): Angodia, Api, Bosodula, Faradje, Gangala-na- Bodio, Garamba, Kibali R., Maude, Niangara, Tingasi. COTE D'IVOIRE: Mountains E of Mandinani, Ni•l•; also Bouak•, Korhogo (Brunel and Thiollay, 1969). KENYA: Kakamega, Kirui (Elgon). NIGERIA: Dororo, Dumbi Woods, Enugu, Indanre, Iseyin 15 mi. N, Kafanchan (Serle, 1940), Kishi 20 mi. N, Lokoja, "Niger River," Panshanu (RBP), Shagunu (DW), Zaria, Zaria-Bauchi mile 81 (RBP). SIERRA LEONE: Bumban, Saiama. SUDAN: Aloma Plateau, Bahr el Ghazal, Kajo Kaji, Mongalla, Nanga, Nimule, Tambura, Wau, Yei. TOGO: Aledjo, Nuatja. UGANDA: Bunyoro (• Unyoro), Fatiko, Foda, Kamchuru, Kibusi, Masindi, Nakwai Hills, Nyouri Jardin, Pader, Parosa, Seroti, Tiriri.

Vidua chalybeata ANGOLA: Fazenda do Cuito (Moco), Gambos, Huila, Quifandongo, Serra do Mange 1650 m. (Moco). BOTSWANA: Boro, Gaberones,Kasane, Kedia, Lake Dow, Maun, Maun 7 mi. NE, Nata, Nokaneng, Sepopa,Shashi R., Shorobe. BURUNDI: Katumba, Usumbura (•--Bujumbura). CAMEROON: Abarue, Mayo Sala (Monard, 1951), Rei Buba, Waza. CHAD: Abeche, Abilela, Besgongu, Bol, Fort Lamy. CONGO: KIVU: Bushangonya (= Usangora), Fizi, Kadjukju, Kalembelembe, Kasongo,Katana, Kibati, L. TanganyikaN, NW, Lulenga, Rutshuru,Tschibati, Uvira, Yamba Yamba. CONGO: ORIENTALE (ITURI): Bunia, Kasenyi. CONGO: KASAI and KATANGA: (see p. 333). COTE D'IVOIRE: BoundJail, Mankono, Touba (Bouet, in Brunel and Thiollay, 1969). ETHIOPIA: Abrer-der-Adieux, "Abyssinie," "Abyssinia," Addis Ababa, Adoshe- baital, Aliberet, Aramaio L., Aftale, Asmara, Baro Volya--Bonga Fork, Baroda, Baru, Chaadi Staati (Ghati Sati, Hadi-Saati), Chaucori, D[idaschamalka,Dambe, Dangila, 330 ORNITHOLOGICAL MONOGRAPHS NO. 11

Dejem (Gojam), Diedieem, Dire Daoua (= Diredawa), Gallabat, Galuda, Godessa, Gondar, Harrar (• Harar), Hawash R., Kassim R., Maraco, Mojjio, Om Hager, Omo River at 600', Sequela (= Zukwala), Setit, Shoa, Soddo, Tana L. (= L. Tsana, L. Sanne), Uba. GAMBIA: "Gambia," "Gambia R.," M'Boro. GHANA: Morago R. GUINEA: Conakry, Fouta Djalon, Kirita, Semini, Sokotou, Toubo. KENYA: Bungoma,Chuka, Chuka-Embu,Doinyo Narok (= Donje Erok), Elgon Mr., Embu, Enjemusi, Escarpment Station, Fort Hall, Isiolo, Kabete, Kakamega, Kakamega 22 mi. N, Kalini Thika, Kampi yo Moto, Karungu, Kavirondo, Kendu Bay, Kiambu, Kibwezi, Kikuyu, Kisumu, Kisumu 10 mi. NW, Kisumu 15 mi. ESE, Kisumu 20 mi. SE, Kitui, Kongelai, Lodwar 60 mi. W, Machakos, Malindi, Malindi 7 mi. N, Marich Pass, Meru, Muhoroni 4 mi. SE, Mumoni, Nairobi, Nyeri, Olorgesailie, Rusinga, Sigor, Taveta, Uaso Nyiro R. South, Voi. MALAWI: Chikwawa, Chileka, Chinteche, Chinteche 20 mi. S, Chiromo, Chitsa, Dedza, Fort Johnston,Kamangadazi, Kuziwaduka (= Nkuziwaduka), Lilongwe (RBP), Makanga, Makoko, Monkey Bay, Mphunzi, Mpimbi, Mzimba, Namadzi, Ndamera, Port Herald (= Nsanje), Ruo, Salima 5 mi. E, Symon's. MALI: Ansongo,Bamako, Fiko, Gao, Hambori, Kami, Kara, Kulikoro, Niafounke, Sanga, "Soudan Fran•ais." MOZAMBIQUE: Mocuba,Mossuril, Movene, Msusu, Mwira L., Namateche,Namapa, Tete, Xinavane, "Zambesi," Zumbo. NIGER: Agades, Dosso, Zinder. NIGERIA: Bauchi, Birnin-Kebbi, Denge (RBP), Dumbi Woods (RBP), Farniso, Fatika Shika (RBP), Ganye (RBP), Gombe junction 4 mi. W (RBP), Gombe junction 27 mi. W (RBP), Gombe junction 1 mi. S (RBP), Gombe junction 13 mi. W (RBP), Gusau, Gusau 4 mi. N (RBP), Gusau 50 mi. W (RBP), Gusau 62 mi. W (RBP), Jos, Jos 5 mi. NW (RBP), Jos 15 mi. NW (RBP), Jos 26 mi. E (RBP), Kano, Kiri, Lagos, Loko, Narode (RBP), Nasarawa (Serle, 1949), Numan, Numan 6 mi. NW (RBP), Pategi (JHE), Rabba, Riman Zayam (= Ziam), Shagunu, Sokoto, Talata Mafara 39 mi. W (RBP), Toro (RBP), Yo, Yobe R.-L. Chad, Zaria. RHODESIA: Bembesi, Binga 23 mi. E, Birchenough Bridge (= Sabi-Devuli), Bubi Hill, Bulawayo, Carrick Creache, Lupani, Malapati Drift 3 mi., Malapati Drift 10 mi. SW, Mazoe Bridge, Naunetsi, Penalverne, Penhalonga, Ruenya River, Rugiruhuru R.-- Sijaria Mrs., Sabi R., Sabi Valley Experimental Station, Sebungwe,Selukwe, Sentinel Ranch, Shabani Rd. Tebukwe R., Sinoia, Tuli. Tuli R., Turgwe R., Umguza Forest, Zambezi R. (17 ø45'S,24 ø12'E), Zambezi Valley (15ø39'S, 30ø20'E). RWANDA: Astrida, Gabiro, Kibingo, Kisenyi, Musha, Nsasa, Rubona. SENEGAL: Cap Vert, Dagana, Dakar, Diourbel, Fandene, Kirtaonda, Nianing, Richard-Toll, "Senegal," Thies. SIERRA LEONE: Mapotolon. SOMALIA: Abrona. SOUTH AFRICA: NATAL: Ndumu, "Pt. Natal." SOUTH AFRICA: TRANSVAAL: Blouberg, Bomgo Gorge (K. N. P.), Elim (Zoutpansberg), Hartebeesfontein, Irene, Kondowe (RBP), Leydsdorp, Limpopo R., Louw's Creek, Marble Hall Fisheries Station, Marikana, Merensky Reserve, Piena- arsrivierdam, Potchefstroom, Pretoria, Rustenburg. SOUTH WEST AFRICA: Ondonga. 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 331

SUDAN: "Bahr-el-Gebel,"Berber, "Blue Nile Province," Bunzuga,Debba, Dongola, el Abiad, E1 Erain, E1 Fasher,E1 Fifi, Gezirat al Fil, Insel Argo, JebelMarra, Jebelein, Kadugli, Kerma, Khartoum, "Kordofan,"Lado, Makwar, Medani, Merowe, Mongalla, Musran Island, Naikhala, Omdurman,Roseires, Sennar, Shendi, Singa, Talodi, Tauila, Wadi Naja, Wau, Zeidab. TANZANIA: Bagamoio, Bukoba, Dar-es-Salaam, Igawa, Ipande, Iringa, Kagehi, Kahama, Kakondo, Kibondo, Kigoma, Kisiwani, Kisii, Lembeni, Madibira, Malangali, Mara R., Marangu, Mikindani, Mkamala, Mkomasi,Moshi-Karanga R., Moshi S. Side Kilimanjaro,Mwanza, Mwaya, Nyanza (L. Tanganyika),Rungwa, Rukwa-Chinambo, Rukwa-Kafakola, Rukwa-Tumba, Tindi, Ukerewe, Utengale. Nicolai (1967) found orange-lootedbirds to mimic L. senegalaat Dar-es-Salaam,Kisangiro, Lembeni, and Mwanga. (see also p. 333). UGANDA: Butiaba, Entebbe, Jinja, Kabale, Kaiso, Kampala, Kayonza, Kibonwa (: Kibondwe), Kitwe, Masindi, Mpanga Forest = Kibale Forest, Mubendi, Mbuku, Tiriri, "Uganda." ZAMBIA: Balovale, Chaanga, Chikwa, Chilanga, Chipepo, Kakumbi, Kalabo, Kankomba, Kapalune, Kaputa (Mweru Marsh), Katangalika, Katombora, Kazungula (= Kazangulu, Kazungulu), Livingstone,Luangwa R. (2 localities), Luashi R., Lukulu, Lundazi, LupamadziR., Lusangazi,Magoye, Mazabuka, Mbuzi, Mulanga, Munyamadzi R. (3 localities), Munyumbwe, Mupamadzi R., Ntengo (= Wamuna), Rukuzi Dam, Tindi, Zambezi R. near Lindi R., Zawanga.

Vidua purpurascens ANGOLA: Gambos (Mossamedes), Huila (Huila) (see also p. 333). BOTSWANA: Francistown. KENYA: Bura (Teita), Kacheriba (= Kuchelebai), Kibwezi, Sigor. MALAWI: Bwangu, Chididi Mission, Chididi Stream, Chikwawa, Chileka Road, Dai, Fort Johnston,Fort Lister, Kazingizi, Magnua (species?), Masona, Matiya, Mini Mini, Mlanje, Monkey Bay, Nyakamera, Nyambadwe, "Nyasaland," Port Herald (= Nsanje), Ruo R.--Mlanje Mt., Salima 20 mi. W, Symon's, WangawangaHill, Zomba, Zomba 9 mi. S. MOZAMBIQUE: Luvio R.--Nampana, Mocuba, Movene, "Mozambique," Msusu, Mtogwe Mt., Namaucha, Nhauela, Tambarara, Tete, Vila Continha 8 mi. S, Zimbiti. RHODESIA: Anglesea Farm, Bindura, Birchenough Bridge (= Sabi-Devuli), Bula- wayo, Chipinga Road 2400', Chirinda, Darwin Mt., Hartley, Ingwesi Ranch, Kuramadzi R., Lusitu R. 1200', Matebe Hills, Melsetter, Melsetter Road 3600', Mphoenge Reserve, Penhalonga2 mi. S, Pombadzi R.--Lundi R., Que Que, Sabi Valley Experimental Station, Salisbury,Sashi R.--Shashani R., Sebungwe,Selukwe, Sentinel Ranch, Sinoia, Umguza Forest, Zambezi R. (17ø45'S, 24 ø12'E). SOUTH AFRICA: TRANSVAAL: Hartebeesfontein, Hector Spruit, Kondowe, Marble Hall Fisheries Station (RBP), Marikana, Merensky Reserve (see also p. 333). SWAZILAND: Ingwavuma R., Stegi. TANZANIA: Morogoro, Sunya, Undis, Usegua (see also p. 333). Nicolai (pers. comm.) heard white-footed indigobirds mimicking L. rhodopareia at Iringa. ZAMBIA: Chaanga, Chalimbana, Chama (Lundazi), Chiawa, Chilanga, Chilola Stream, Choma (Choma), Fort Jameson, Kanakazilui, Kankomba, Katombora, Living- stone, Lundazi, Lunzi R., Lusaka, Mambova, Meruz, Mukuni, Mulanga, Mumbwa Boma, Nyamudera, Sandwe, Senyati-Zambezi, Sinazongwe, Tembwe, Zambezi R. near Lundi R. 332 ORNITHOLOGICAL MONOGRAPHS NO. 11

Vidua funerea ANGOLA: Camabatela, Dugue de Braganca, Golungo Alto, Luhanda (5 km. N. Quela), Okasekenua (Chitau) (see also p. 333). CONGO: KASAI: (see p. 333). CONGO: KATANGA: Dikulwe Valley, Elizabethville (= Lumumbashi), Kando River, Kansenia,Kasaji, Kasapa, Plateau de Biano, Shila-Tembo (see also p. 333). CONGO: LEOPOLDVILLE (KIHSHASA): Boma. MALAWI: "Angomiland," Chididi Mission, Chididi Stream, Chikwawa, Chinteche, Cingoma, Dedza 5000', Dai, Kachere's (4000', Dedza), Kazingizi, Lilongwe-Likuni (2 and 3 mi. Lilongwe), Masona, Matope Hill, Mbabzi (Lilongwe), Mkohoma, Mwana- lunda Stream, Mwangala 5000', Ncuce, Njakwa, Mjazi, Nyamijale Stream, "Nyasa- land," Pinda Stream, Zomba 9 mi. S. RHODESIA: Lusitu R. 1200', Penhalonga 2 mi. S (Umtali), Selinda. SOUTH AFRICA: CAPE PROVINCE: Adelaide 10 mi. N, Bosch, Brandestron, "Cape of Good Hope," Committees,East Griqualand, East London, Kei Bridge, Komga, Lusikisiki (12 mi. to Holy Cross), Port Alfred, Somerset East. SOUTH AFRICA: NATAL: Amanzimtoti, Burg Mt., Drummond, Durban, Eschowe, Kloof, Mtunzini, "Natal," Park Rynie, Pietermaritzburg, "Pt. Natal," Richmond, Shong- weni Dam, Table Mt., Ubombo, Umhlanga, Umkomaas R., Umzumbe, Wartberg Road, Weenen. SOUTH AFRICA: TRANSVAAL: Downs, Tzaneen 6 mi. E (see also p. 333). SWAZILAND: Ingwavuma River. TANZANIA: Bukoba, Dar-es-Salaam, Iringa Highlands, Karema, Kindoroko (N. Pare Mrs. 5400'), Mikindani, Mwagamira, Nyarunbogo, Ukerewe; also Mikumi National Park, Mikumi-Iringa Road, and Moshi west, on the basisof Nicolai's (1967) observations of pale-looted birds mimicking L. rubricata. (See also p. 333.) ZAMBIA: Chambeshi, Chilanga, Chipako, Chisombwe, Fort Jameson, Fort Rose- berry, Kasama, Kasempa, Katangalika, Kawambwa, Lusiwasi Lake (Serenje), Mporokoso, Mulilo, Mwinilunga, Salujinga.

Vidua wilsoni, form "nigeriae" CAMEROON: Abong Mbang, Foumban 50 mi. SW, Garua, Mayo Sala (Monard, 1951), Sidiri Mt., Tibati. CONGO (BRAZZAVILLE) (= MOYEN CONGO): N'gabe. CONGO: ORIENTALE (UELLE): Faradje. GHANA: Yegi. MALI: "Soudan Fran•ais." NIGERIA: Enugu, Kiri, Kogum, Panshanu, Zaria. PORTUGUESE GUINEA: Gunnal. SENEGAL: "Senegal." SIERRA LEONE: Bintumane Peak. SUDAN: Kulme (Darfur), Nimule, Yei.

Vidua wilsoni, form " camerunensis" CAMEROON: Donenkeng, Kumbo-Bamenda, Mayo Sala (Monard, 1951), Nguru, Sidiri Mt., Tibati. CENTRAL AFRICAN REPUBLIC (REPCENTRAFRICAINE): Bengui, Mbru, Nola-Mbaike, Ratu, Yakota, "Rafiagebiet." 1972 PAYNE: PARASITIC INDIGOBIRDS OF AFRICA 333

CHAD: Ir•na, Madjingais. CONGO: EQUATEUR (UBANGI): Duma, Yakoma. CONGO: ORIENTALE (UELLE): Api, Auelldebietsches,Bafuka, Balingilingi Station (Nagero P. N. G.), Camp Aru, Congo-Nil kil. 999 (Faradje), Faradje, Gangala- na-Bodio, Niangara, Tukpuo, Zobia. COTE D'IVOIRE: Bav•-Komo•, Bouak•; Korhogo (Belllet, in Brunel and Thiollay, 1969). ETHIOPIA: Gallabat, Machigay. GAMBIA: Kuntair. GHANA: Kratschi, Yendi. GUINEA: Dabola, Fouta Djalon. MALI: Kara. NIGERIA: Abakaliki, Bauchi 25 mi. W, Enugu, Kaduna 20 mi. SSE (RBP), Kaduna •RR crossingS. (RBP), Panshanu,Samaru Experimental Farm (RBP), Yola 98 mi. NW, Zaria. SIERRA LEONE: Batkanu, Bendugu, Bumban, Makeni, Rokupr, Wurkufu. SUDAN: Abbi Obed, Lado, Malakal, Mongalla, Nimule, Rimo, Roseires. TOGO: Mangu, Paio.

Vidua wilsoni, form "wilsoni" CAMEROON: Dodo, Ndom. CENTRAL AFRICAN REPUBLIC (REPCENTRAFRICAINE): Nola-Mbaike. CONGO: ORIENTALE (UELLE): Gangala-na-Bodio. NIGERIA: Abeokuta, Agongu, Agoulerie, Amambara Creek, Enugu, Epe (JHE), Ibadan (JHE), Igbetti (JHE), Ilorin, Kishi 20 mi. N, Shagunu, Yelwa, Zaria. PORTUGUESE GUINEA: Gunnal. SENEGAL: "Senegal." SUDAN: Nimule, Rimo, Sheikh Tombe, Torit. TOGO: Kande 11 mi. S.

Problematical Specimens ANGOLA: Mombolo. CONGO: KASAI: Bena Dibele (Cie du Kasai: Luja), Bena Ndongo, Kabinda (Sankuru Dist.), Katombe, Lodja, Luebo, Merode, Mkulwa, Pania Mutombo, St. Joseph Mission. CONGO: KATANGA (east): Baudoinville, Chiancy (Marungu Mts.), Kabalo, Kasiki, Moba, Musosa, "Tanganika," Tembwe. TANZANIA: Birds of the nigerrima-purpurascens-centraliscomplex. Bagamoio, Dabaga, Kikoboga, Kilosa, Kododi, Lembeni, Luwiya R., Mamboio, Mfirisi, Miss. Uasse (? illegible), Mikindani, Mtaka River, Tingida (Kilosa). TRANSVAAL: Birds of the purpurascens-funereacomplex. Blouberg, Idalia, Rust- der-Winter, Rustenburg. ORNITHOLOGICAL MONOGRAPHS

No. 1. A DistributionalStudy of theBirds of BritishHonduras, by Stephen M. Russell.195 pp., 2 colorplates. 1964. Price $4.50 ($3.60 to AOU members). No. 2. A ComparativeStudy of SomeSocial Communication Patterns in the Pelecaniformes,by GerardFrederick van Tets. 88 pp., 4 text figures. 1965. Price $2.00 ($1.60 to AOU members). No. 3. The Birds of Kentucky,by RobertM. Mengel.Cloth bound, xiv + 581 pp., 4 color platesplus text figuresand vignettes.1965. Price $10.00 ($8.00 to AOU members). No.4. Evolutionof SomeArctic Gulls (Larus): an ExperimentalStudy of IsolatingMechanisms, by Neal GriffithSmith. 99 pp., 62 text figures. 1966. Price $2.50 ($2.00 to AOU members). No.5. A ComparativeLife-history Study of FourSpecies of Woodpeckers, by Louisede KirilineLawrence. 156 pp., 33 text figures.1967. Price $3.75 ($3.00 to AOU members). No. 6. Adaptationsfor Locomotionand Feeding in theAnhinga and the Double-crestedCormorant, by Oscar T. Owre.138 pp., 56 textfigures. 1967. Price $3.50 ($2.80 to AOU members). No. 7. A DistributionalSurvey of the Birds of Honduras,by Butt L. Monroe,Jr. 458 pp., 28 text figures,2 color plates. 1968. Price $9.00 ($7.20 to AOU members). No. 8. AnApproach to the Study of EcologicalRelationships among Grass- land Birds,by JohnA. Wiens.93 pp., 30 text figures.1969. Price $2.50 ($2.00 to AOU members). No. 9. MatingSystems, Sexual Dimorphism, and the Roleof Male North AmericanPassefine Birds in the NestingCycle, by Jared Verner and Mary F. Willson.76 pp. 1969. Price$2.50 ($2.00to AOU members). No.10. The Behaviorof SpottedAntbirds, by Edwin O. Willis,vi + 162pp., 3 colorplates, 27 textfigures. 1972. Price $6.00 ($4.75 to AOUmembers). No.11. Behavior, Mimetic Songs and Song Dialects, and Relationships of the ParasiticIndigobirds (Vidua) of Africa,by RobertB. Payne, vi + 333pp., 2 colorplates, 50 textfigures, 40 audiospectrographs.1973. Price$8.00 ($6.40 to A.O.U. members). Likeall otherAOU publications,Ornithological Monographs are shipped prepaid.Make checks payable to "TheAmerican Ornithologists' Union." For the..0nvenience of those who wish to maintaincomplete sets of Ornithological onographsand to receivenew numbers immediately upon issue, standing orders will be accepted.

Orderfrom: Burt L. Monroe,Jr., Treasurer,American Ornithologists' Union,Box 23447,Anchorage, Kentucky 40223