The Structure and Dynamics of an Avifauna

LAND BIRD COMMUNITIES OF GRAND BAHAMA ISLAND: THE STRUCTURE AND DYNAMICS OF AN AVIFAUNA

JOHN T. EMLEN Department of Zoology University of Wisconsin

ORNITHOLOGICAL MONOGRAPHS NO. 24

PUBLISHED BY

THE AMERICAN ORNITHOLOGISTS' UNION

1977 LAND BIRD COMMUNITIES OF GRAND BAHAMA ISLAND: THE STRUCTURE AND DYNAMICS OF AN AVIFAUNA ORNITHOLOGICAL MONOGRAPHS

This series,published by the AmericanOrnithologists' Union, hasbeen established for major paperstoo long for inclusionin the Union's journal, The Auk. Publi- cationhas beenmade possible through the generosityof Mrs. Cafil Tucker and the Marcia Brady Tucker Foundation, Inc.

Correspondenceconcerning manuscripts for publicationin the seriesshould be addressedto the Editor-elect,Dr. MercedesS. Foster, Department of Biology, Universityof SouthFlorida, Tampa, Florida 33620.

Copiesof OrnithologicalMonographs may be orderedfrom the Assistantto the Treasurerof the AOU, Glen E. Woolfenden,Department of Biology,University of South Florida, Tampa, Florida 33620. (See price list on back and inside back cover.)

OrnithologicalMonographs No. 25, xi + 129 pp.

Editor of A.O.U. Monographs,John William Hardy

SpecialAssociate Editors of this issue,Frances C. James,Department of Biology, Florida State University, Tallahassee,and Ned K. Johnson,Museum of VertebrateZoology, University of California, Berkeley

Assistant Editor, June B. Gabaldon

Author, John T. Emlen, Departmentof Zoology,University of Wis- consin, Madison, 53706

First receivedDecember 1975; accepted,October 1976; final revision, December 1976

Issued November 22, 1977

Price $9.00 prepaid ($8.00 to AOU Members)

Library of CongressCatalogue Card Number 77-90792

Printedby the Allen Press,Inc., Lawrence,Kansas 66044

ii LAND BIRD COMMUNITIES OF GRAND BAHAMA ISLAND: THE STRUCTURE AND DYNAMICS OF AN AVIFAUNA

JOHN T. EMLEN Department of Zoology University of Wisconsin

ORNITHOLOGICAL MONOGRAPHS NO. 24

PUBLISHED BY

THE AMERICAN ORNITHOLOGISTS' UNION

1977 iv TABLE OF CONTENTS

PREFACE ...... xi

1--INTRODUCTION ...... 1

2--METHODS ...... 2

HABITAT MEASUREMENTS ...... 2

ANALYSIS OF HABITAT DISTRIBUTION ...... 3

BIRD POPULATION MEASUREMENTS ...... 4

WITHIN-HABITAT DISTRIBUTION AND ACTIVITY MEASUREMENTS ...... 4

3--GRAND BAHAMA ISLAND ...... 5

THE ENVIRONMENT ...... 5

Geology ...... 5 Physiographichistory ...... 5 Climate ...... 6

Recent history ...... 6

THE BIRDS ...... 7

Geographicdistribution of the breedingspecies ...... 7 Origins and routesof invasion...... 8 Colonizationpatterns ...... 10 Turnover rates ...... 11

4 THE HABITATS AND THEIR BIRD COMMUNITIES ...... 12

VEGETATION PATTERNS ...... 12

Generaldescription ...... 12 The structure of Grand Bahama habitats ...... 13 Groupingof standsinto habitattypes ...... 15 Alignment of standsalong gradients...... 19

THE BIRD COMMUNITIES ...... 21

Communitystructure ...... 21 Bird speciesdiversity ...... 25 Total bird density...... 28

DISCUSSION-•DIVERSITY AND DENSITY CORRELATES ...... 30

V 5-•BIRD DISTRIBUTION THROUGH THE HABITATS ...... 32

DYNAMICS OF DENSITY DISTRIBUTION ...... 32

Conceptsand models ...... 32 Habitat distributionpatterns ...... 33

HABITAT SELECTION ...... 36

Distributionby types ...... 36 Distribution along gradients...... 36

DISPERSION AMPLITUDES--SPECIALIZATION ...... 39

Dispersionof speciesthrough the 25 stands...... 39 Dispersionby types ...... 39 Dispersionalong gradients...... 41

DENSITY DISPERSION RELATIONS ...... 48

OVERLAP AND SIMILARITY ...... 51

Measuring overlaps ...... 51 Overlap and phylogeneticrelationships ...... 52 Overlap and geographicderivation ...... 52 Ecological spacing...... 54

6--THE PINE FOREST COMMUNITY-•SEASONAL CHANGES ...... 55

PROCEDURES AND DEFINITIONS ...... 55

SEASONAL CHANGES ...... 58

The breedingseason community ...... 59 The wintering community...... 60 The transientcommunity ...... 62

IMPACT OF THE WINTER MIGRANT INVASION ...... _-...... 63

7--SPATIAL DISTRIBUTION WITHIN THE PINE FOREST ...... 66

VERTICAL DISTRIBUTION-•LAYERS ...... 67

Vegetativestructure and avian utilizationat five levels...... 67 Populationstructure in the five layers...... 70

COMPARTMENT DISTRIBUTION ...... 70

Physicalcharacteristics of the compartments...... 72 Bird speciesdistribution through the compartments...... 73 Populationstructure in the compartments...... 75 Seasonalchanges in spatial distribution...... 76

8--GUILD DISTRIBUTION WITHIN THE PINE FOREST ...... 80

CONCEPTS AND DEFINITIONS ...... 80

PROCEDURES ...... 81

THE FORAGING GUILDS OF THE GRAND BAHAMA PINE FOREST ...... 83

Ground-gleaning herbivores ...... 83 Stemseed pluckers ...... 94 Fruit and bud harvesters ...... 94

Nectar sippers...... 95 Sap and cambium eaters ...... 96 Foliage browsers...... 96 Ground-gleaningcarnivores ...... 96 Ground pouncers...... 97 Flower probers...... 97 Shrubfoliage gleaners...... 97 Shrub stem drillers ...... 98

Pine bark and wood drillers ...... 98

Bark gleaners...... 98 Pine twig gleaners...... 99 Pine coneprobers ...... 100 Pine needle gleaners...... 101 Air sallyers...... 102 Air screeners ...... 103

GUILD BIOMASS AND FOOD ABUNDANCE ...... 103

GUILD STRUCTURE AND COMMUNITY DYNAMICS ...... 106

Guild structureand diversity...... 106 Dispersionof speciesthrough the guilds...... 106

SPECIES OVERLAP AND COMPETITION ...... 109

ACKNOWLEDGMENTS ...... 111

SUMMARY ...... 112

vii CONCLUSIONS ...... 115

LITERATURE CITED ...... 115

APPENDIX---SPECIES ACCOUNTS ...... 118

ooo V1H LIST OF FIGURES

Figure 1. Plant types and foliage compartments...... 4 2. Map of the westernBahama Islands ...... 6 3. Continental and Antillean ranges of the breeding land birds of Grand Bahama ...... 9 4. Faunal derivation of the breeding land birds of Grand Bahama ...... 10 5. Map of Grand Bahama Island showing extent of forests and location of survey stands ...... 12 6. Cross-island profile showing vegetation zones and positions of survey stands ...... 13 7. Submaturepine forests on Grand Bahama Island ...... 14 8. Young pine forest types and open stands...... 15 9. Other Grand Bahama habitat types ...... 16 10. Distribution of the 25 survey stands on canopy height-canopycover coordinates ...... 17 11. Distribution of the 25 standson habitat gradient diagrams ...... 20 12. Bird speciesdiversity plotted on habitat gradient diagrams ...... 26 13. Species number and diversity plotted against foliage height diversity at 25 sites ...... 27 14. Bird speciesnumber and diversityplotted againstvegetation density (volume) at 25 sites ...... 28 15. Total bird density plotted on habitat gradient diagrams ...... 29 16. Total bird density plotted againsttotal vegetationvolume ...... 31 17. Hypothetical density distribution through habitats of declining quality ...... 33 18. Decreasingdensity through 10 favored standsfor 34 common species...... 34 19. Habitat distribution of Grand Bahama specieswith respect to tree, shrub, and exposed ground cover ...... 40 20. Habitat distribution of Grand Bahama specieswith respect to vegetation height and arboreal foliage volume ...... 42 21. Habitat distribution of Grand Bahama specieswith respect to shrub type and shrub plus ground cover foliage volume ...... 44 22. Permanent resident densitiesplotted on habitat gradient diagrams ...... 46 23. Winter resident densitiesplotted on habitat gradient diagrams ...... 47 24. Density in preferred stand plotted against dispersionamplitude ...... 51 25. Ordination of the Grand Bahama bird speciesfor similarity of habitat selection ...... 58 26. Physioguomicstructure and compositionof the submaturepine forests...... 59 27. Temporal limits of the seasonalbird communitiesof Grand Bahama ...... 62 28. Monthly changes in size and composition of the pine forest bird communities ...... 63 29. Geographic sourcesof the migrant invaders•species ...... 64 30. Geographic sourcesof the migrant invaders--numerical representation.... 65 31. Relation of arian insect gleaner biomass to food resource levels in 7 compartments ...... 66 32. Vertical distributionof commonspecies in the pine forests...... 67 33. Speciesdiversity and densityin 5 equal layers of the pine forests...... 69 34. Relative abundanceof speciesin the 5 equal layers ...... 71 35. Winter invader and permanentresident representation in the 5 equal layers .... 71 36. Compartment distributiondiagrams for 30 pine forest species ...... 74 37. Relative abundanceof speciesin the 4 major compartments...... 75 38. Relative abundance of speciesin 16 foraging guilds ...... 85

ix TABLES

Table 1. Vegetation measurements in the 25 Grand Bahama habitats ...... 14 2. Bird densities at the 25 habitat sites--winter 1969 ...... 22 3. Relation of dispersiontypes and behavior typesin Grand Bahama land birds .... 35 4. Percent distribution of the 34 common bird speciesthrough the 25 survey stands ...... 37 5. Distribution of speciesthrough 7 habitat types on Grand Bahama Island ...... 48 6. Positionand dispersionof 5 pine foliage gleanersalong 9 gradientsof the pine forest habitat ...... 50 7. Habitat specialization along 7 gradients for various elements of the Grand Bahama land bird community ...... 50 8. Overlap in habitat selectionin winter ...... 53 9. Habitat overlap and geographic derivation ...... 54 10. Habitat overlap and residency status ...... 54 11. Physiognomic characteristicsof the submature pine forests ...... 56 12. Dominant tree and shrub speciesin submature pine forests ...... 57 13. Size, density and biomass of the 4 residency elements in the 3 seasonal pine forest commtmities ...... 60 14. Residencystatus, density, and biomass of member and visitor speciesin the 3 seasonalpine forest communities ...... 61 15. Vertical distribution of bird populations in the wintering commtmity of the pine forest ...... 68 16. Volume and structural characteristicsof the major habitat compartments of the pine forests ...... 72 17. Demographic characteristicsof the bird populations in the 4 major compartments of the pine forest ...... 77 18. Compartment preferencesof membersof the wintering and breedingcommu- nities ...... 78 19. Guilds and guild provincesrecognized in the submature pine forests ...... 82 20. Foraging guild distributionsof members of the pine forest bird commtmity .... 84 21. Speciesdensities in the foraging guilds ...... 86 22. Biomass of each speciesin the foraging guilds ...... 88 23. Percent speciesdensity compositionof foraging guilds ...... 90 24. Percent biomass composition of foraging guilds ...... 92 25. Foraging methods of members of the needle-gleaningguild ...... 102 26. Composition and structure of the 18 foraging guilds of the Grand Bahama forest communities ...... 107 27. Overlap in guild membershipin the wintering commtmity ...... 110 28. Overlap in guild membershipin the breedingseason commtmity ...... 111 PREFACE

This monographhad its beginning24 years ago when I undertooka study of the habitat distribution and structure of several communities of land birds in Central Africa. As I launchedinto this project it becameclear that available techniqueswere totally inadequatefor the quantitativeanalyses I neededand that it was up to me to developnew ones. Some progresswas made that year with a systemfor describingand measuring habitats (Emlen 1956), but I returned to the Stateswith no soliddata on eitherpopulation densities or habitat relations. Returningto other researchactivities, I postponedfurther fieldwork on these problemsuntil 1967 when I was able to devote a spring semesterto a studyof the structureand dynamicsof the winteringbird communitiesof a grass-bushland area in southernTexas (Emlen 1972). The vegetativestructure was relatively simplehere, and besidesimproving my techniquesof habitatdescription I worked intensivelyon developingcensus methods for smallland birds,methods that would provideabsolute density values by equatingvariables in transectcounts with the detectabilitycharacteristics of eachspecies (Emlen 1971). After testingthese habitat measurementand censustechniques in a variety of situations,I selectedthe extensivepineland forests of the southeasternstates and neighboringBahama islands for the type of studyI had tried to conductin Africa 15 yearsbefore. The work of William B. Robertsonin this area provideda valu- able backgroundon ecologicalconditions and faunal composition(Robertson 1955), and I beganthe studiesdescribed in this monographin Januaryof 1968. During the 20 yearssince the conceptionof this project,approaches to the study of communityecology have changedconsiderably. Important new conceptson the niche, competition,and communitydynamics have appeared,and emphasis has shiftedstrongly from descriptivestudies to model buildingand testing. Work in the old patternseems dimmed by the brillianceof thesenew studies. But I am repeatedlyimpressed that the imaginativecreations of theoreticalecologists rarely survivelong in this modernera of scientificferment unless they are built on solid empirical data. Most of the speculationsand interpretationsin this monograph will doubtlessbe ephemeralas our scienceprogresses; I only hope that the materials on which they are basedwill prove to be solid and clearly presented.

xi 1--INTRODUCTION

This monographrepresents an attemptto analyzethe populationstructure and dynamicsof a definitiveavifauna. Unlike a typicalfaunal studyit focuseson the densitiesand ecologicaldistributions of speciesthrough the habitatsand foraging substratesof the vegetationand undertakesto evaluateand interpretsome of the factorsunderlying community structure and regulation.My approachis descriptive with an attemptto presenta broadand balancedpicture of the entire system. I selectedan islandas the sitenot becauseof a particularinterest in the prob- lems of insularity,but becausean islandcommunity is shieldedby surrounding water barriersfrom the confusionof irregularingress and egressby speciesthat belongonly peripherally or reallydo not belongat all. As an islandsite for the study,Grand Bahamawas large enoughto provide an assortmentof habitattypes, yet smallenough to minimizethe confusingeffects of regionalnomadism and local geographicvariation. One type, lowland pine forest,was representedextensively while otherswere restrictedto small blocks or beltsadequate to hold onlylimited populations. Grand Bahama is particularly well situatedfor observingthe extent and impact of seasonalmigrant invasions on the structureand integrityof residentcommunities. Little ornithologicalwork has been done on Grand Bahama, but the avifauna of the region has been well coveredduring the past 50 years. The fieldwork was done during the months of January through May in 1968 and 1969. In 1968 I spentalternate 2-week periodson Grand Bahama and in southernFlorida censusingthe bird populationsand plotting their distribution throughthe vegetationof the pine forestsfor comparativematerial. In 1969 I spent the entire period on Grand Bahama, measuringcommunity structuresand populationdensities at 25 survey sites during January, February, and March, and conductingfurther censusand behaviorobservations in the pine forestsin April and May. I returnedfor brief visitsin Januaryand May of 1971 to check severaldetails and to collectmaterial on avian food resourcesin the pine foliage. The studycovered the small land birds only, thosethat interacteddirectly in exploitingthe resourcesof the terrestrialvegetation and its invertebrates.Other membersof this trophicassemblage, especially the lizards,should logically have beenincluded. They are a minor elementexcept in the shrubsand groundvegetation, but their omissionis due primarilyto my limitedversatility and lack of time. The rationale for ecologicalanalysis in this study is to recognizethree hierar- chicalsystems of decreasinginclusiveness through which dispersion can be charted: a habitat level, a within-habitat compartmentlevel, and a substratelevel. Indi- vidual birds are regardedas ranging through the various subdivisionsof these systemsbut concentratingtheir activities in those subdivisionswhere conditions are most favorablefor their morphologicaland physiologicalcharacteristics. The organizationof this monographis basedon thisapproach. After presentingback- groundmaterial in the first three chapters,! examineacross-habitat distribution in chapters4 and 5, and then compartmentand substratedistribution in the pine foresthabitat in chapters6, 7, and 8. Descriptivematerial on the bird specieswith scientificnames and quantitative 2 ORNITHOLOGICAL MONOGRAPHS NO. 24 data on ecologicaland distributionalcharacteristics is presentedin an appendix. Nomenclaturefor North American speciesfollows the A.O.U. Checklist (1957) and supplements(1973); for the Antilleanspecies I havefollowed Bond (1971).

2--METHODS

SELECTION OF SURVEY SITES

I first surveyedthe physiognomicand floristiccharacteristics of the vegetation of the islandto determinethe natureand rangeof habitatvariation and to provide a basisfor selectingappropriate tracts for intensivestudy. I then selected25 tracts representingthe full range of variation but each showingmaximum internal uni- formity of structureover at least20 ha. Sketchmaps of each tract were prepared, and transectroutes 0.9-1.8 km in lengthwere laid out bisectingthe most typical portion of the tract, avoiding edges, and taking advantageof existing trails and roadways.Three tracts in the submaturepine forests that dominatedthe island were used for intensivestudies of within-forestdistribution in 1968 (chapters 6, 7, and 8). The other 22 representedvarious habitat types and, togetherwith the 3 submature forest areas, were used for studies of across-habitatdistribution in 1969 (chapters4 and 5).

HABITAT MEASUREMENTS

The vegetationat each of the 25 sites was sampled along a series of habitat dimensionsof presumedimportance to the birds. Data were recorded according to a systematicplan in the 3 submaturepine forest tracts and less formally for the other 22 tracts. In the 3 submature forests the measurements were made at stationslocated by random number serieswithin each of 60 blocks (20 in each stand) 66 m long and 90 m wide, evenly straddlingthe transectroute. Each station point served as the center for measurements(trunk diameter, height, crown diameter,crown depth, and distance) of the nearesttree in each of three size classes(emergent, canopy, subcanopy)in each of four radial quadrants. Twenty points surroundingeach station point (at 3-m intervals) along 4 equi- spacedradii) were used for data on tree canopy, shrub vegetation,and ground vegetation. Tree canopy (presenceand type) was recordedwith a zenith sight- ing device (Emlen 1967); shrub type, height, crown diameter, and species,and ground cover type, depth, and density were recordedby plumb line contacts. A total of 240 data points was thus obtainedfor most arborealdimensions, and 1200 for the others. Becauseof time limitations,vegetation evaluations on the other 22 tracts were madeby lesssystematic sampling, pacing, and subjectivecomparison with diagram- matic representationsof canopycover. A shorthandsystem (Emlen 1956) was usedin whichvegetation type, foliage type, foliagedensity, canopy height, canopy cover, and patchinesswere recordedin formulae representingthe tree, shrub, and groundcover strata. For within-habitat'distribution studies the vegetationin all standswas apportioned into compartmentsbased on plant type, vertical zone, and horizontalzone in the tree, shrub, and-groundcover strata (Fig. 1 ). I recognized7 plant types 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 3 in the forestfrom deadstumps and emergentpines (toweringabove the dominant canopy) to ground cover, 5 vertical divisionsor levels in each plant from the lower stem to upper crown, and 5 horizontaldivisions arranged as concentric zonesaround each plant from the tree trunk or shrubcore to the peripheryand the emptyspace around and betweenplants. Any point in the entire spacebe- tweenthe canopytop and the groundis thus assignableto 1 of 175 compartments in this system.In practice,many compartmentswere nonexistent,undetectable, or unmeasurable,and for data analysisthe numberwas reducedby combinations and eliminations to 11. The volumeof standingvegetation (number of m3 of spacecontaining plant stemsof foliage) was calculatedfor the 3 vegetationstrata and the 1! recognized compartmentsof eachstand. Viewing the averagetree crown as a verticalcylinder with roundedtop and bottom, I multipliedthe mean crown depth in metersby 0.75 by the canopycover in m3 per ha for each tree classto give a volumetric densityvalue in m• per ha. A similar procedurewas used for the shrub and groundvegetation strata. Volumesfor all compartmentsper ha were derivedby calculatingthe percentrepresentation of each in its stratumand multiplyingby the stratumvolume. Volume of the air spacebetween plants was obtainedby subtractingall compartmentvolumes from the total spacebetween ground surface and tree tops. Values for bark and wood surfaceson tree trunks and stumps were calculatedin surfaceunits (m 2) by multiplying mean trunk circumference in the upperand lower half of treesby the trunklengths. The complexityof the habitat is important as a comprehensiveenvironmental parameterin communitystudies but is difficult to reduceto a simple,meaningful index. In this study I followed the procedureintroduced by MacArthur and MacArthur (1961), using the densityof vegetationin three horizontallayers correspondingto the trees,shrubs, and groundcover as the basisfor calculating an overall habitat diversityindex. Becauseof marked differencesin texture and screeningcharacteristics of the foliage in the three layers on Grand Bahama, however,and because of the presumedimportance of twigsand stemsas perching substrates,I used the presenceor absenceof standingvegetation in blocksof volumetric space as the basis of measurementsrather than the leaf surface or foliage screeningmeasures adopted by the MacArthurs. The information theory equation(H' = p• loge pi) was usedto calculatethe index values. Natural loga- rithms were usedin all diversitycalculations.

ANALYSIS OF HABITAT DISTRIBUTION

The relativepositions of the 25 standswith respectto three basic habitat features-canopyheight, canopy coverage, and foliagetype--are graphicallydis- playedin Figure 10. The groupingsof symbolsin this figureprovide a relatively objectivebasis for recognizing7 habitat typesuseful in general discussionsof speciesdistributions. This approachof plotting stand positionson bicoordinate gridswas also used as the basisfor tracingthe quantitativedistribution of species and of variouscommunity attributes along selected habitat gradients (Fig. 11). Densityor indexvalues for a speciesare simplyentered at the coordinateposition 4 ORNITHOLOGICAL MONOGRAPHS NO. 24

/. Emergent Pines •.•/ / ...... SubcanopyPines • / / ...... DeadPine Stumps I / / --- .... BroadleafTrees • ! • /'• //f'• /• • // •"•-----CanopyPines YJYYoYYYY¾/ /-----Shrubs(3size crosses) .,,o o.1,,,o,,,L,I o,l ooJQ 3o,H1,,,,9,.,1, 0o-T/o Qcl,/ /--Ground...... Cover SEVEN PLANT TYPES Crown M,I. Trunk Top• Upper OpenSpoce Lower-- - -- Air Shrubs Space Upper Lower GroundCover

FIVE FIVE ELEVEN COMPARTMENTS LEVELS CONCENTRIC ZONES

Fromm 1. Plant types and foliage compartmentsused in structural analyses of the forest habitat. M.O. = middle outside, M.I. = middle inside, L.O. = low outside, L.I. -----low inside. of eachstand in thesediagrams. The resultingpatterns provide graphic portrayals of densityand attribute distributions along the selectedgradients.

BIRD POPULATION MEASUREMENTS

Populationdensities for all bird specieswere obtainedfrom transectcounts convertedto absolutevalues by applyinglocally determineddetectability coeffi- cientsbased on the lateral distributionpattern of detectionpoints (Emlen 1971). All counts were started within a half hour after local sunrise and extended for from 1.2 to 2.1 hours. I alwayswalked the transectsalone, coveringthe route without deviationsat between1.0 and 1.6 km/hr. All detections,visual or audi- tory, were recorded.

WITHIN-HABITAT DISTRIBUTION AND ACTIVITY MEASUREMENTS

In addition to identifyingeach bird and estimatingits lateral distancefrom the trail, I was able to record basic informationon activity and positionwithin the habitat for most of the detections tallied on the census transects. These ob- servationsincluded activity when first detected(whether foraging, singing, resting, being agressive),plant type in which located, vertical position (level) in the plant, radial position(zone) in the plant, and heightabove the ground. Weather conditions(temperature, cloud cover, wind directionand speed,wetness of vegetation) were alsorecorded for all trips afield. Close attention to the task of detectingand recording all birds encountered on the morning transectcounts precluded sustained observations of behavior and 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 5 foragingmethods. Unscheduledobservation time later in the day was used for. suchwork as well as for measuringvegetation, finding nests, and mappingterri- tories. In these unscheduled observations I recorded the basic information de- scribedabove plus data on ageof the subjects,presence and activityof conspecific and allospecificassociates, the foragingperch, the food substrate,and the foraging method. After an observationor seriesof observationsin one place I evaluated the local habitat selectionby recordingthe type, size, and densityof foliage within approximately1 m of the bird, and the proportionof tree cover, shrub cover, and exposedground cover within an estimated36-m radius (1 acre). I used spacedobservation units rather than secondsof continuousobserving on the assumptionthat they would give a more balancedand representativepic- ture of foragingactivity by a population.In orderto furthersuppress biases related to atypicallocal conditionsor individualidiosyncrasies, I attemptedto distribute observationsas widely as possiblein theseactivity observations. Where one individual or an uncommonspecies remained for continuousobservation, I spaced the entries at a minimum of • min, and limited the record to 5 entries in a series. For compartmentdistribution analyses, the bias inherent in selectedobservation situationswas avoided by usingonly the transectcount tallies.

3--GRAND BAHAMA ISLAND

TI-IE ENVIRONMENT

Geology.-•Geologically,Grand Bahama Island is a long low ridge of oolitic limestoneof Pleistoceneage, risingat its highestpoint to only about 10 m above sea level. The groundsurface is heavily erodedwith innumerablepits and deep solutionholes. Soil is essentiallyabsent, the organicmaterials deposited from the vegetationleaching or settlingto thebottom of theholes. Physiographichistory.--Being a tectonicallystable border ridge of the Little BahamaBank (Fig. 2), the surfacearea of Grand Bahamaexpanded many fold when sea levelswere low during the Pleistoceneglacial epochsand alternately shrankto zero or nearlyzero duringthe interglacialperiod of high sealevel. Thus the terrestrialflora and fauna were probably annihilatedrepeatedly by marine inundation,the last suchepisode, according to data from southernFlorida, end- ing about80,000 yearsago (Alt and Brooks1965). No directland connections exceptwith neighboringAbaco Islandhave existedsince the tertiary. During the Pleistoceneice advances,however, the broad low surfaceof the Little Bahama land mass, of which Grand Bahama is a part, extendedto within 50 km of the Great Bahamaland massto the southwhich, in turn, approachedto within 35 km of Cuba. All of this complexhas been separated from the peninsulaof Florida sinceat leasttertiary times by the deepand relativelywide (105 km) rapidly flowingcurrent of the Florida straits. Sinceemerging from the deep seas of the Sangamon interglacial period (80,000 years ago) Grand BahamaIsland has fluctuatedin size from its present 1200 km 2 or smaller to 16,500 km 2, the area of the Little Bahama Bank of which it is a part. Estimatesof a sea-levelsubsidence to the -120 m mark duringthe last Wisconsinice advance(about 15,000 years ago) (Milliman and Emery 1968) 6 ORNITHOLOGICAL MONOGRAPHS NO. 24

FIGURE 2. Map of the western Bahama Islands showing their position with relation to neigh- boring land masses and to shallow "banks" that mark the extent of the island areas during Pleistoceneglacial periods. would indicate an expansionto about 15,000 km2 at that time. Since then the seahas been rising slowly, reaching a stageof -7 m about4,000 yearsago (Scholl et al. 1969). At this level the island of Grand Bahama must still have been six or eighttimes its presentsize and formeda continuousland masswith neighbor- ing Abaco Island. The major shrinkingin size has occurredover the past few thousandyears during which sea levels have been rising at about0.5 m per 1,000 years(Scholl et al., op. cit.). Clirnate.--The climateof Grand BahamaIsland is subtropicalwith tempera- turesranging from a mean of 20.3øC in Januaryto 28.2 in July (data from the Freeportairport). Annual rainfall averages1,216 mm and occursthroughout the year with about twice as much in summeras in winter. The trade winds producea dominantsoutheasterly air flow varyingin directionand intensityunder the influenceof cyclonicdisturbances to the north. Hurricanesoccur occasionally in late summer,the most recentsevere storms hitting the island in 1941 and 1949. Recenthistory.--Grand Bahama was almostundisturbed and essentiallyun- populateduntil 1929 when extensivetimber cuttingwas initiated. Most of the timberwas stripped from the islandfor mineprops or pulp duringthe late 1940's 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 7 and 1950's,but reproductionhas beenrapid sincethat time, and no extensive cutting other than local clearancefor real estatedevelopment has been done since 1959. Standscut between 1946 and 1949 were 10 and 12 m high in 1968 and possesseda lush shrubunderstory on the southhalf of the Island, lesslush on the north half. Standscut duringthe 1950's showedvarious stages of recovery re_flectingmean growth rates of roughly0.5-0.6 m per year. Height,density, and understoryvegetation are secondarilyinfluenced by irregularground fires and occasionalcrown fires. A few smalltracts of pine were apparentlydenuded by saltwater encroachments during storms. The principalmethod used in timbercutting in the 1940's and 1950's was to removeall treesother than saplingsand a few tall scatteredstems (about five treesper acre) usedfor anchoringthe haulinggear and supposedlyas seedtrees for forestreproduction. The remainingsmall treesand brushwere then usually levelledand burned. This procedurehas resultedin rather even-agedstands with a scatteringof slenderrelic "emergents"above the dominantyoung canopy. One patchof about4.4 km• on the northside was completely cleared and kept open for habitationand farmingduring the early lumberingperiod. Abandonedfor 10-12 years,it is now vegetatedwith bracken,coarse forbs, and tall grasses, and exceptfor the narrowcoastal strips and a few urbanareas and golf courses, constitutesthe only openland on the island. Furtherdisturbances caused largely by real estateoperations since 1959 include the constructionof the city of Freeport-Lucayawest of the Island'scenter, now coveringabout 30 km-ø, and a networkof unpaveddevelopment roads penetrat- ingin an irregularnetwork into nearly one-half of the Island'sinterior.

THE BIRDS

At the time of this study (1968-69) there were apparently33 speciesof land birdsexcluding birds of preybreeding on GrandBahama Island. Thirty of these were permanentresidents, and three were summerresidents migrating in from winter rangesto the south. An additional34 speciesmoved in from the north duringthe winterseason, and 8 morewere present briefly as transientsduring the springmigration period. Geographicderivation of the breedingspecies.--Grand Bahama, an oceanic island, containsno relict continentalspecies in its breedingfauna and has been populatedentirely by over-watercolonization from the neighboringcontinents and islands. Becauseof the relative recencyof its final emergenceas an island capableof supportinga terrestrialfauna, it presentsa goodopportunity to study the geographicorigins of its colonists.The two primary sourcesare obviously the North American continent to the north and west and the Antillean Islands to the south.The nearestland on the continentis 105 km awayacross the deepand geologicallyancient Florida straits (Fig. 2). Cuba,the nearestand alsothe larg- estof the Antilles,lies about 500 km away,but onelarge "stepping stone" (Andros Island) and a seriesof smallerones have been continuously available for dispers- ing birdssince Grand Bahama's emergence. The waterbarriers to the southand north of Androswere about 35 and 50 km wide respectivelyat lowestsea level 15,000years ago, and are 190 and 180 km wideat thepresent time. 8 ORNITHOLOGICAL MONOGRAPHS NO. 24

AbacoIsland, lying about30 km to the eastof Grand Bahama,is merelyan- other ridge on the Little BahamaBank, and, not surprisingly,has a very similar avian fauna. Originsand routesof invasion.--In the absenceof a direct record of invasions and colonizations,the presentgeographic ranges of the membersof the fauna providethe bestavailable indication of originsand routesof invasion.The direction andextent of the rangesof eachof the 33 membersof the breedingland bird fauna are graphicallyportrayed in Figure 3. Circlesto the left of the central circle in this diagramindicate continental affinities; circles to the right indicate Antilleanaffinities. The degreeof shadingin eachlateral circle reflects the closenessof phylogeneticrelationship of the relatedforms in eachof four increasingly remoteregions. Information incorporated in thesediagrams, together with analyses of avian dispersalin the area by Chapman (1891) and Bond (1963), provides a basisfor tentativedeductions on the geographicderivation of each memberof the Grand Bahama community. The deductionsare presentedin code form at the right of each speciesdiagram in Figure 3. A graphicrepresentation of the deducedorigins and routesof invasionis presentedin Figure 4. Of the 32 species(exluding the introducedHouse Sparrow), 7 are best representedby conspecificsor otherclosely related forms on the continentto the north- west and are accordinglyconsidered to have arriveddirectly from North America. Two of these,the wide-rangingMourning Dove and the Blue-grayGnatcatcher, have apparentlydeveloped no taxonomicallyrecognized morphological changes sincecolonizing Grand Bahama(Co). Four, the Hairy Woodpecker,the Brown- headedNuthatch, the Yellow-throatedWarbler, and the Pine Warbler have been recognizedby taxonomistsas subspecifically distinct on tho island(C•), and one, the BahamaYellowthroat, is consideredto have evolvedto full speciesstatus (C2). The ancestorsof 11 of the communitymembers appear to have come from North or Central Americarelatively recently but reachedGrand Bahamaby way of the Greater Antilles. All of thesehave changedmorphologically in the course of their Caribbeanperegrinations. The sequenceof eventsin the courseof their historyof dispersalrequires considerable speculation, but threespecies, the Cuban Nighthawk,Northern Mockingbird, and YellowWarbler, apparently subspeciated in the Greater Antilles before spreadingnorth into the Bahamas(C• A0); one, the Ground Dove, has severaldistinct races in the West Indies, all different from the continentalforms, and probablyhas subspeciatedat least twice in reaching Grand Bahama (C• A•). Five speciesare found in the Greater Antilles as well as in the Bahamasand belongto superspeciesor siblingspecies groups also occurring on the North Americancontinent. Their ancestraldispersal routes thus probablyincluded early invasions of the GreaterAntilles where they speciated beforespreading northward to the Bahamas.Two of these,the Black-whiskered Vireo and Olive-cappedWarbler, show no furthermorphological changes en route to the northern Bahamas (C2 A•). Two species,the Bahama Swallow and the Thick-billedVireo, I regard as belongingto superspeciesgroups with specifically distinctrepresentatives on severalAntillean islands as well as the Bahamas.An- cestorsof the Grand Bahama forms probably speciatedseveral times en route to theirpresent range (C2 A2). The interpretationof the BahamaSwallow (Cal- 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 9

White-cr,Pigeon A0 Brown-hd.Nuthotch ZenoidoDove A0 NorthernMockingbird CiAo MourningDove CO BahoreoMockingbird A• GroundDove CtAi Red-leggedThrush AI K.-W.Quoil Dove A0 Blue-groyGnotcotcher CO Smooth-billedAni AO Thick-billedVireo CgA2 CubonNighthowk CtAO BIk-whlskeredVireo C2A0 CubonEra. Hummer A1 BohomoBononoquit BohomoWoodstot A2 YellowWorbler C•Ao HoiryWoodpecker C• Boh.Yellowthr. Worbler C Red-bellßWoodp. C2AI Olive-coppedWorbler C2A0 GroyKingbird AO BohomoPine Worbler LoggerheodFlyc. AI BohomoYellowthroot C2 StolidFlycotcher C2Ai HouseSpotrow X0 Gr.Ar•tilleon Pewee C2At Striped-hd.Tonoger BohomoSwollow C2A2 BIk-focedGrossquit A Gr.AntilL Bullfinch

FIGURE3. Continental and Antillean ranges of the breeding land birds of Grand Bahama Island. The presenceof the sameor relatedforms (same subspecies,species, superspecies) is indicated at four increasinglyremote stationsin each of the two directions. (Except for the Bahama Swallow, taxonomy follows Bond (1971) and the AOU Checklist (1957) and supplementsthereof (1973). My deductionsconcerning the geographic derivation of each form on Grand Bahama are indicated by symbolsat the right of the diagrams. Symbols for degree of relationshipare: Solid circle = same subspecies; circle with cross= samespecies, different subspecies;circle with horizontalline = same superspecies,different species;diagonal line cuttingcircle = limited occurrence;vertical line cutting circle = recent colonizer. Symbolsfor deducedderivation are: C = North American continental origin, A ---- Antillean origin, CA = Continent via Antilles. X • introducedby modernman, /1 ----subspeciated with invasion,/2 • speciatedwith invasion, 70 = no change recorded. lichelidon)as congenericwith Lamprochelidonis basedon a compositeof taxonomists'opinions (L. Short in litt.). The 14 speciesclassed as of Antillean origin belongto generathat probably derivedat an earlyperiod from ancestralforms in eitherCentral or SouthAmerica (Bond 1963). Five of these, the White-crownedPigeon, the Zenaida Dove, the Key WestQuail Dove, the Smooth-billedAni, and the Gray Kingbird,now have dispersedranges in the WestIndies and showno taxonomicallyrecognized changes in their northward extensioninto the Bahamas (A0). Nine specieshave under- gonetaxonomically recognized changes in the Bahamas;eight of these,the Cuban EmeraldHummingbird, Loggerhead Flycatcher, Bahama Mockingbird, Red-legged Thrush,Bahama Bananaquit, Striped-headed Tanager, Black-faced Grassquit, and i0 ORNITHOLOGICAL MONOGRAPHS NO. 24

c- FROM AMERICAN CONTINENT Cc

CA- FROM CONTINENT C2 VIA ANTILLES C, ClA1 C C2A1 C2A2

A2 A• Ao A-FROM ANTILLES

FIGURE 4. Faunal derivation of the speciescomprising the breeding land bird fauna of the Grand Bahama Island. Each arrow represents a group of species. The width of the arrow indicates the number of species in the group. A cross line on the arrow shaft shows that subspeciationhas occurred; a double line, that full speciation has occurred. The symbolsat the base of each arrow match those used in Figure 3.

Greater Antillean Bullfinch, to the subspecieslevel (A•) and one, the Bahama Woodstar, to the specieslevel (A2). Colonizationpatterns.--Generalizations based on patternsof dispersionthrough island chains can provide suggestionson colonizationhistories and sequences. Of the seven speciescolonizing Grand Bahama directly from North America only two have pushedfarther into the Antilles than the first few northernBahama Islands (cf. Fig. 3) and these two, the Mourning Dove and the Pine Warbler, occurringas distinctsubspecies in the Greater Antilles, may have reachedthose sitesby an earlier invasionor a separateinvasion from Central America. Such patternsof limited and essentiallyunbroken penetration of the island chain sug- gestrecent invasion, i.e. an earlystage in taxondispersion as visualizedby Ricklefs and Cox (1972) in their taxon cycle model for archipelagos.From the other direction,the Southeast,Grand Bahamais at the end of a long islandchain of dispersion. The 25 speciesconsidered to have come by this route tend to show the extendedand broken distributionpatterns that characterizeadvanced stages in the taxon cycleof Ricklefsand Cox. Theseobservations suggest that membersof the Antillean faunal elementhave, in general,a longerhistory in the northernBahama Islands than thoseof the NorthAmerican element. Other factors must be considered,however, especially the vegetationpatterns through the Bahamasand the habitat characteristicsof 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 11 the bird species.Pine forestsdominate Grand Bahama and the other northern islands,while brushyvegetation cover the islandsto the southand east. With the exceptionof the MourningDove, a highly eurytopicspecies, all sevenof the colonizersfrom North America are pine forest inhabitantson Grand Bahama, eventhe Gnatchatcherand the Yellowthroat,which elsewhereare characteristically broad leaf forest and marshlandbirds respectively.This does not conform with Ricklef and Cox's generalizationthat recent (stage I) colonizerstend to occupy open, coastal, or disturbed habitats. On Grand Bahama these habitats are, in fact, occupiedby speciesbelonging to the Antillean element,species that according to the island penetrationand broken distributioncriteria, should be in ad- vancedstages of the taxoncycle. More direct evidencethat the bird speciesof the North American elementof pinelandbirds did not invade the northern Bahamasuntil quite recently comes from fossil pollen data indicating that pines were uncommon in the southern Florida vegetationuntil about 5,000 years ago (Watts 1971). T/he conditions of rising sea levels and permeablelimestone rock associatedwith this spreadof pine forestsin Florida must have also existedand could have producedsimilar effectson the Little BahamaBank. BahamianPines (Pinus caribaea)probably reachedGrand Bahamafrom the south (Howell 1972), but whetherthe present forestswere establishedbefore the postglacialperiod of sea level elevationis un- known. Avian fossil materialsfrom New ProvidenceIsland (Brodkorb 1959) indicatethat arborealbird speciesand hence trees were presenton the Great Bahama Bank to the southeastduring the last glacial stage of the Wisconsin glaciation. Turnoverrates.--The faunal historyof an island such as Grand Bahama is of coursemore than a progressivesuccession of colonizations.Many extinctionsand replacementshave undoubtedlyoccurred over the millenia, and some species may have had a historyof repeatedextinctions and recolonizations.MacArthur and Wilson (1967) proposedthat the number of speciespresent on an island at a giventime reflectsan equilibriumbetween the colonizationrate relatedparticu- larly to distancefrom colonizationsources, and the extinctionrate relatedpar- ticularlyto islandsize. Recentstudies on previouslysurveyed islands off southern California(Diamond 1969), in the SouthPacific (Diamond 1970), in the Virgin Islands (Robertson 1962), the Cayman Islands (Johnston1975), and on Mona Island near Puerto Rico (Terborgh and Faaborg 1973) suggestthat a considerable number of both extinctions and new colonizations have occurred on all of theseislands during the past 50-75 years. Suchdata, if accurate,suggest that turnoverrates of roughly 1.0-1.5% per year may be representativefor islandsof the size and positionof Grand Bahama,but other observers(Lynch and Johnson 1976, Lack 1976) have challenged these values as artifacts of modern human intervention. Historical data do not exist to permit direct estimatesof colonizationor ex- tinctionrates on Grand Bahama. One species,the West Indian Red-belliedWood- pecker,listed as an uncommonspecies in coastalscrub forest by Bond in 1936, was apparentlymissing in 1968-69, and two speciesnot noted by Bond, the MourningDove and HouseSparrow, were presentlocally and in small numbers 12 ORNITHOLOGICAL MONOGRAPHS NO. 24

Mangroveflats:[•

Hawksbill Creek LacayaBeach

FIGURE5. Map of Grand Bahama Island showingin A the distributionof pine forestsand mangroveflats and in B the locationof the 25 standssurveyed in this study. in the latter years. The BahamaMockingbird, not reportedon Grand Bahama by Bond, was collectedin 1960 (Schwartzand Klinikowski1963) and was seen by severalamateur observ. ers duringthe early 60's and by me in 1969.

4--THE HABITATS AND THEIR BIRD COMMMUNITIES

The term bird communityis usedin this chapterto designateall the birds that co-occupyan area of habitat at a particularseason and henceinteract directly with each other. The fauna of Grand Bahama Island is thus composedof an array of communitiesdiffering both spatiallyand temporally. Variation is con- tinuousalong both axes,but I have assigneda seriesof spatialcategories on the basis of localities or vegetation types, and of temporal categorieson the basisof the arrival and departureperiods of seasonalmigrants. In this chapter I examinethe spatialdistribution of the birds presenton the Island between10 January and 31 March 1969 while the winter migrantswere still present,i.e. the winteringcommunities.

VEGETATION PATTERNS

General description.--Mostof the interior of Grand Bahama Island (roughly 80% of the land surface) is coveredwith moderatelydense forests of Caribbean 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 13

1õ16 14 9 6 4 I 2 :5 13 1719212410222õ23 õ 7 8 1820 12 II SOUT. V V V V V VV VV VVV VV V V V VVV V CO•ST VNORTH COAST

$0 I00 75 75 200 200 500 I00 500

0 C

FmURE 6. Cross-islandprofile (schematic) transectingthe major vegetation zones and types; approximate average widths of zones are given in meters. The positions of the 25 survey stands with respect to the vegetation zones are indicated above the diagram; the numberscorrespond to those in Table 17 and the map in Figure 5.

Pine (Pinus caribaea). Much of this forest was clear cut during the late 1940's and early 1950's, but the very rapid growth has restoredthe appearanceof submature forest over most of the area. Broad tidal flats with open, low mangrove scrubseparate the pine forestsfrom the sea on the north or leeward side of the island. The higher windwardfront on the southernshore supports a seriesof parallel vegetationbelts generallyprogressing from a narrow beach backed by low sanddunes and occasionalpalms, through a narrowcoastal strip of halophytic grasses,herbs, and low denseshrubs and a strip of dense 3-m-high broad-leaf scrub with pocketsof cattail marsh to a wider band of 4-5 m scrub that blends back into the pines. The distributionof thesemajor vegetationtypes is mapped in.Figure5 and shownin a cross-islandprofile in Figure6. Photographsin Figure7 depictconditions in the submaturepine forestsas they appearedin 1968; thosein Figure 8 show the vegetationin recentlycut forests and in one older stand. Vegetationfeatures of other habitat types are shownin Figure 9. The structure of Grand Bahama habitats.--Measurements and evaluations of habitatfeatures considered to be significantto birdsare presentedin Table 1 for each of the 25 standssurveyed in this study. Includedare featuresof each of the threestrata in the vegetationplus an index of overallhabitat complexity or diversity. The figuresin each caseare averagesfor conditionswithin the stand. Mean tree height varied from 2 m in a recentlycut stand (stand #24-•est. 4 yrs.) to 16 m (max. 22 m) in the oldestof the submatureforests (#1-•est. 30+ yrs.). Trunk diametersin thesetwo standsaveraged 5.5 cm and 28 cm (max. 35 cm) respectively.Old cut stumpsrarely exceeded 35 cm in diameterat any site, suggestingthat this approximatesthe maximumsize attainedlocally by the Carib Pine in undisturbed forests. Exceptfor two badlydisturbed sites and one marginalpine-broadleaf thicket, tree canopycover varied from 19-45% with densitiesas greatas 60% in patches of up to 3fiha. In severalstands there were 2 distinctheight classes reflecting 2 major disturbancesspaced 8-10 years apart. In such casesthe lower stratum 14 ORNITHOLOGICAL MONOGRAPHS NO. 24 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 15 was generallyquite dense(35-40%) while the older treesproduced a 10-20% canopy. The shrub understoryof pinelandsvaried considerablybetween sites in com- positionheight and density,differences presumably being attributableto substrate water, fire history, and perhapswindborne moisture. North of a line running longitudinallydown the middleof the island,shrubs were generallylow, relatively sparse,and often dominatedby Palmetto (Thrinax microcarpa). South of this line and especiallynear the southcoast the shrub stratumwas relativelytall and lush with many Tamarinds (Lysiloma) century plants (Agave), and only a few palmettos. Broadleafshrubs reached their greatestheight and richnessbehind the dunes along the south coast where one study area (#9) was located. Here, often in standing fresh water, they formed nearly impenetrable closed canopy thickets or "coppers"3-5 m high, from which pines were largely or completelyexcluded. Where the water was brackishthe typical mixed copperswith Metopium, Bumella, Torrubia, Ilex, Rapanea, Annona, etc. gave way to monotypic thickets of red mangrove( Rhizophora). Grass dominatedin the ground stratum in roughly inverseproportion to tree 'and/orshrub canopy except in onecase (#25) wherebracken fern replacedthe grass,probably in responseto recentburning. An index of overallhabitat complexity,expressed as vegetationheight diversity (VHD), is presentedin the last column of Table 1. It was highestin the pine forest standswhere all three strata were well represented. It was lowest in the open or shrub habitats. Dense foliage in the closedcanopy thickets tended to shadeout the groundvegetation to producea nearlyone-stratum (shrub) situation. Grouping of standsinto habitat types.--The three vegetationalfeatures most commonlyincorporated into the descriptivenames applied to avian habitatsby ornithologists:canopy height, canopy cover, and foliage type of the dominant stratum,are presentedgraphically in Figure 10 for the 25 standsstudied on Grand Bahama Island. Canopy cover and height of the dominant stratum are plotted on the x and y axesrespectively, and the foliagetype, a complexmultidimensional feature,is indicatedqualitatively by symbols.The numberedsymbols representing

• Age in years (A) estimatedfrom sample ring countsand historical records. 2-4proportion of the area covered by trees (T), shrubs without trees (Sh), and ground vegetation or bare ground alone (G).---expressed as percents. • Tree density (D --• number per hectare). •4Proportion of age classes: M • mature (DBH • 20 cm), S ---- submature (DBH, 10-20 cm), and Y ----young (DBH ( I0 cm). g Canopy height (Fit.). (Top of canopy of the best represented age class.) •0 Volume of space occupied by pine foliage (mS/ha). nTotal shrub cover (C) as percent of ground surface covered by shrubs. (The value in column 3 is the percent of surface with shrubs not covered by tree crowns.) •2Proportion of shrub size classes in tenths. Shrubs • 2 m are classified as tall (T), less than I m as low (L). •a Proportion of foliage types in tenths (B = broadleaf), (P • palmetto), (F ---- ferns), (G = grasses). a•Volume of space occupied by shrubs (Vol.). Sum of area X depth of each height class (mS/ha). • Proportion of vegetation types in ground stratum in tenths. (G • grass, Fi • herbs, V •-- vines, L = litter, F = ferns). •* Volume of space occupied by ground cover (Vol.). Area X mean depth (mS/ha). •* Total volume of standing foliage in mS/ha. (Sum of columns 10 • 14 • 16). X*Habitat Diversity (vegetation height diversity--H' m Z P• log ø P•) where p• is the proportion of ma of habitat space containing standing vegetation that falls in the it• layer of three horizontal layers below the tree to• level--a ground layer 1 m deep (0.1 m), a shrub layer 4 m deep (I-5 m), and a tree layer 7 m deep (5-12 m). 16 ORNITHOLOGICAl MONOGRAPHS NO. 24

,C tD FIGURE 7. Submature pine forests on Grand Bahama Island. general view in stand 2 (20-25 years, medium density N38/35+ BPg/5, 5/10, 2/10 7 GVB); B, stand 3 (20-25 years, medium+ density N36,38+ BP10 2, 4 8, 2/107 BVG: C, shrub stratum in stand 2; D, ground cover in stand 2. (For explanation of descriptive code see Emlen 1956). 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 17

FIGURE 8. Variant pineland types. A, tin area with high tree density (near 3); B, pine- copper transition (4); C. a medium-dense stand with palmetto in understory (18): D. young pines in recently cut stand (24): E, an old stand (portion of !): F, old dead lreeq and scrub in cut and flooded area (7). ORNITHOLOGICAL MONOGRAPHS NO. 24

FIGURE 9. Other Grand Bahama habitats. A, coastal beach with tall grasses and dense shrubs (16); B, brushy marsh with tall grasses and cattails (14): C, dense broadleaf thicket (copper) behind the coastal dunes ½9): D, tidal flat,• with mangrove• (1 1 ). 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 19

DOMINANT 16- FOLIAGE TYPES

14-5 ? / Standard A- Pines \ / • Pine Forests 0- Broad Leaf 12- \ x I//•4 •20 •t x - Grasses, etc. I0- • • • Forest-Copper X •=/// -• Transition X • Coppers 6 • Young •9 ' 4 •---••'" /•.,• •, • • •Pines •0••-' Marsh • •4• Mongrove OldFields ..

• ..... • Ploins •,• • •

0 I0 •0 •0 40 50 60 TO 80 90 I00 COVER (PERCENT) OF DOMINANT STRATUM

F•o•E ]0. •cltncafion of the major habitat types on Grand •ahamas as determined by •c distribution of the 25 survey stands on canopy height and canopy cover coordinates and by foliage types Ond•catcdby symbols). the 25 standsin this diagramfall into 7 ratherdistinct clusters corresponding to subjectivelyconceived vegetation types considered in the first few paragraphs of this chapter.Two man-createdtypes, lawn (golf courses)and urbandevelop- ment, are not coveredin this study. Alignmentof standsalong gradients.•Diagramssimilar to thoseused for delineatinghabitat types in Figure 10 are presentedfor sevenadditional variable featuresof potentialsignificance to birds in Figure 11. The triangleat the top provides3 axesfor positioningeach of the 25 standswith respectto the relative amountsof tree, shrub,and groundcover exposedto the sky. The middle and lower rectanglesshow tree height matched with pine vegetationvolume, and understoryfoliage type matchedwith understoryvegetation volume. By plotting thesevalues on two or more axes,an appraisalof combinationdimensions on the diagonalscan also be made. Besidesrevealing the structuralrelationships of stands to each other in terms of specifieddimensions, these diagrams can be used as basesfor plottinghabitat distributionsof bird species(see Figs. 19, 20, 21) and of variouscommunity attributes (see Figs. 12 and 15). In the trianglediagram of Figure 11, three open field sites (10, 22, 25) and two coastalplain sites (15, 16) appearin the lowerleft corner,hugging the base line (no trees) but variouslydisplaced to the right (showingincreasing amounts of shrubcover). The two mangroveflats (11 and 12) fall near the centerof the base line, the marsh (14) with a mosaic of cattail and brush lies a little farther to the right,and the thicketsite or copper(9) fallsin the lowerright corner.The 16 pine forestsare scatteredover the heart of the trianglewhere trees, shrubs, and groundcover are all well represented.A site in the transitionbetween forest 20 ORNITHOLOGICAL MONOGRAPHS NO. 24

0 3 4 5 6 7 8 9 I 12 I I VOLUME OF PINE VEGETATION (1000m3/ho)

I00

u. 80- © Q © • SO- • 40-

• 2o-

o 3 4 5 6 7 8 9 I0 12 14 16 18 20 VOLUME OF VEGETATION IN SHRUB AND GROUND STRATA (lOOOrn3/ha)

FIGURE 11. Position of the 25 survey stands along seven habitat gradients. A, position with respect to the proportions of tree cover, shrub cover and ground cover open to the sky; B, position with respect to volume of standing pine vegetation (1000 mYha) and tree height (of dominant age class); C, position with respect to volume of shrub and ground cover vegetation (1000 m•/ha), and foliage type (percent of total ground cover that has broadleaf foliage vs. grass,palmetto, ferns• etc.). and coppet (6) appearsbelow and to the right, while a stand that failed to re- generateafter lurebering(7) liesbelow and near the baseline. In the central diagramof Figure 11 forest siteswith high pine foliage volume appearin the right half while fields,marshes, and mixed copperswith only scat- 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 21 tered pines are clusteredat the left. Seven sites where there were no pine trees are representedby a singlecircle with an x at the lower left corner. Tree height (vertical axis) in generalreflects the time intervalssince the last drasticlumbering at the site. Tree densitiesdecline as individualcrowns broaden with age so that canopycover generallyremains at between25 and 40%, and standingvegetation volume at around1,000 m3 per km2 throughthe growthcycle. Thus,while canopy height increases rather constantlywith time for 20-30 years, pine vegetationvolume is relativelyindependent of tree age and reflectsa variety of other environmental factors. In the lower rectangleof Figure 11 vegetationvolume is representedhori- zontallyon a log scale,and foliagetype, expressedas the ratio of broad-leavesto all other foliage (palmetto,grasses, ferns, etc.), on the ordinate. The highest vegetationvolumes occurred in the coppetsand coppet-pinetransition stands (9 and 6) appearingat the right end of the rectangleand highon the broadleafscale. Not far to the left of these,but low on the verticalscale, are the circlesrepresent- ing the marshhabitat (14), dominatedby cattails,and the two abandonedfields dominatedby brackenfern (25) and tall grass(22). The two beach tracts (15 and 16, but combinedand labelled15), and the old field (10), characterizedby relativelyshort and sparsegrass, are centrallylocated along the base line of the long axis. Pine forestshad relativelylight understoryvegetation and are clustered to the left of center. The two mangrovesites (11 and 12) with low vegetation volumesand highbroadleaf ratios, appear at the upperleft.

THE BIRD COMMUNITIES

The bird speciesfound in each of the 25 stands are listed in Table 2 with their population densities. The distributionof speciesacross the stands in this table is examinedin the next chapter. Only the communityattributes of structure, diversityand density,as revealedby readingdown the columnsare consideredhere. Community structure.--Community structure was analyzed with respect to speciescomposition and the proportions of migrant and permanent resident elements in the membership. 'Numericaldominance was held by thePalm Warbler in 12 of the25 communities includingmost of thosein pine forests,one on a mangroveflat (12), and oneon thecoastal strip (16). Otherwinter residents held the top numericalposi- tion in 8 more communities:the CommonYellowthroat in the marsh (14), one coastalstrip (15), andone old field (22); the Catbirdsin the coppetcommunity (9), and an old field with densebracken fern (25), the GrasshopperSparrow in a grassyfield (10), the Yellow-rumpedWarbler in a pine stand (4), and the Northern Waterthrushin one mangrovecommunity (11 ). Five communitieshad permanentresident species at the top: the Cuban EmeraldHummingbird dominatedin threeyoung pine stands(19, 21, 25), the Striped-headedTanager in the coppet-pinetransition community (6), and the Olive-cappedWarbler in one tall pine community(8). There are four seasonalresidency elements in the Grand Bahama avifauna: nonmigratorypermanent residents (hereafter designatedPR), winter residents or invadersmoving in from northern breedinggrounds for the winter months 22 ORNITHOLOGICAL MONOGRAPHS NO. 24

TABLE 2 DENSITY(INDIVIDUALS/KM 2) OF EACHBIRD SPECIESAT THE 25 HABITATSITES ON GRANDBAHAMA ISLAND--WINTER, 1969

Coef. of Sea- I 2 3 4 5 6 7 8 detect- sonal p-• p2 p2 p.• y2 Th • y• p• Species ability statusx 6.7a 8.12 6.7a 8.33 4.1a 5.8a 5.6a 5.5a White-crownedPigeon .30 P ..... 2.3 -- -- Mourning Dove .25 P Zenaida Dove .20 P 2.9 15 -- 12 -- 10 -- -- Ground Dove .20 P 6.0 -- 3.0 17 -- 3.5 -- -- Key West Quail Dove .20 P Smooth-billed Ani .60 P ..... 21 -- -- Cuban Emerald Hummingbird .14 P 21 46 200 59 173 123 52 293 Bahama Woodstar .12 P ...... 5.8 18 Hairy Woodpecker .35 P 16 11 10 11 17 24 4.0 12 Yellow-belliedSapsucker .22 W 5.4 6.8 5.5 .... 10 Gray Kingbird .60 S LoggerheadFlycatcher .30 P 9.8 1.7 4.0 4.8 9.7 2.3 -- 2.3 StolidFlycatcher .30 P -- -- 2.0 -- -- 7.0 -- -- Greater Antillean Pewee .30 P 6.0 18 4.0 9.7 23 18 2.3 2.3 Bahama Swallow .35 P Brown-headed Nuthatch .25 P 48 -- 17 .... 40 Gray Catbird .18 W -- 69 20 145 -- 188 -- -- N. Mockingbird .65 P -- 2.3 -- 9.7 -- 2.1 -- -- Bahama Mockingbird .60 P Red-legged Thrush .18 P 3.3 5.5 -- 1.5 -- 3.8 -- -- American Robin .50 W Blue-greenGnatcatcher .25 P 46 62 28 60 23 00 2.8 2.8 Thick-billed Vireo .36 P 20 36 3.3 42 33 65 -- 1.9 Black and white Warbler .17 W 14 15 7.0 14 -- 4.1 -- -- Worm-eating Warbler .10 W Orange-crownedWarbler .12 W -- -- 4.9 Parula Warbler .12 W -- 21 -- 24 -- 5.8 -- -- Yellow Warbler .20 W Magnolia Warbler .15 W ..... 4.6 -- -- CapeMay Warbler .12 W -- 25 -- 53 -- 12 -- -- Black-throated Blue Warbler .12 W ..... 5.8 -- -- Yellow-rumpedWarbler .20 W 3.0 99 12 313 20 55 15 15 Black-throated Green Warbler .12 W -- 4.2 Yellow-throated Warbler .20 P/W 54 40 24 56 34 11 -- 47 Olive-cappedWarbler .25 P 148 42 57 -- 120 7.0 -- 316 Pine Warbler .30 P 66 36 20 23 81 18 14 72 Prairie Warbler .15 W 20 26 16 55 -- 23 -- -- Palm Warbler .24 W 168 723 425 167 231 66 110 249 Ovenbird .12 W 4.9 -- -- 4.0 -- 18 -- -- Northern Waterthrush .15 W Common Yellowthroat .15 W 48 9.9 -- 29 13 14 -- -- Bahama Yellowthroat .20 P 27 20 15 37 63 18 16 65 Wilson's Warbler .12 W Redstart .15 W -- 30 8.0 19 19 ------Red-wingedBlackbird .60 W Bananaquit .20 P 39 111 128 109 39 52 7 44 Stripe-headedTanager .20 P 15 50 51 49 44 159 3.5 72 Greater Antillean Bullfinch .15 P ..... 25 -- -- Black-facedGrassquit .15 P 16 6.6 8.0 55 -- 97 -- 58 SavannahSparrow .12 W GrasshopperSparrow .12 W LincolnSparrow .12 W Total of species(S) 24 27 24 26 16 32 11 18 Bird speciesdiversity(H') 2.60 2.21 2.10 2.61 2.34 2.81 1.67 2.17 Total density(D) 807 1532 1073 1479 943 1164 232 1320 Permanentres. (species) 17 16 16 16 12 21 9 15 Permanentres. (density) 544 503 574 656 660 767 67 1046 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 23

TABLE 2 (CONTINUED)

9 10 11 12 13 14 15 16 17 18 p/Th2 F 2 Mn 2 Mn2 p2 Msh2 C2 Cs y2 p2 Species 2.53 3.1• 4.6a 4.23 1.43 4.73 1.43 1.23 4.6a 7.7a Winter res. (species) 7 11 8 l0 4 11 2 3 Winter res. (density) 263 1029 499 823 283 397 165 274 White-crowned Pigeon 5.4 Mourning Dove ------30 ZenaidaDove ...... 2.7 Ground Dove .... 3.8 -- 83 ------Key West Quail Dove ...... 51 -- -- Smooth-billedAni ..... 5.7 -- -- Cuban Emerald 58 9.4 -- -- 10 12 70 -- 31 70 BahamaWoodstar -- 11 ..... 108 87 4.3 Hairy Woodpecker 55 3.7 2.5 5.4 2.1 2.5 ------6.0 Yellow-bell. Sapsucker 44 ------6.6 Gray Kingbird -- -- 1.6 LoggerheadFlycatcher 5.4 ------4.9 ------8.7 5.3 Stolid Flycatcher Greater Ant. Pewee .... 7.3 .... 10 Bahama Swallow ..... 61 -- 4.2 -- Brn.-hd. Nuthatch .... 5.8 .... 25 Gray Catbird 314 7.2 ------54 28 ------N. Mockingbird -- 7.8 -- 1.5 -- 1.3 4.3 4.9 -- -- Bahama Mockingbird Red-leggedThrush 54 ...... 5.6 American Robin ..... 1.7 Bl.-g. Gnatcatcher 207 ------8.8 10 -- -- 3.5 2.1 Thick-billed Vireo 202 .... 35 -- -- 2.4 1.5 Black and white Warbler 28 Worm-eating Warbler .... 7.3 Orange-crowned Warb. -- -- 7.3 Parula Warbler 108 Yellow Warbler -- -- 13 Magnolia Warbler Cape May Warbler Black-throated Blue Warbler Yellow-rumped Warb. 186 13 -- 80 11 146 14 32 8.5 21 Blk.-thr. Gr. Warb. 14 Yellow-throated Warbler 8.1 ------3.7 ------4.3 3.4 Olive-capped Warb. ------30 20 -- -- 10 240 Pine Warbler -- 4.3 -- -- 15 -- -- 41 24 Prairie Warbler 151 8.7 ------17 19 -- 5.8 -- Palm Warbler 67 75 33 110 37 25 267 850 167 67 Ovenbird 108 Northern Waterthrush 97 -- 35 19 -- 125 -- -- Common Yellowthroat 202 -- 29 58 -- 228 445 237 29 Bahama Yellowthroat 143 6.5 -- 14 3.7 -- 14 -- 26 31 Wilson's Warbler 13 Redstart 97 .... 5.7 ------3.5 Red-winged Blackbird -- -- 1.5 96 -- 31 -- 5.3 -- -- Bananaquit 267 -- 8.7 33 7.3 30 139 -- 4.4 8.0 Stripe-headedTanager 154 ------15 .... 24 Greater Ant. Bullfinch 43 Black-facedGrassquit 3.2 -- -- 6.3 19 34 ------14 Savannah Sparrow ...... 69 ------GrasshopperSparrow -- 139 ...... 7.3 -- Lincoln Sparrow ...... 23 • -- -- Total of species(S) 27 11 8 12 18 19 13 7 15 20 Bird species diversity (H') 2.89 1.77 1.74 2.01 2.61 2.30 1.94 0.96 1.98 2.09 Total density (D) 2677 286 130 455 198 854 1226 1246 436 568 Permanent res. (species) 13 6 2 7 14 10 6 3 10 17 Permanent res. (density) 1205 43 11 92 136 220 361 122 218 476 Winter res. (species) 14 5 6 5 4 9 7 4 5 3 Winter res. (density) 1472 243 119 363 62 634 865 1124 218 92 24 ORNITHOLOGICAL MONOGRAPHS NO. 24

TABLE 2 (CONTINUED)

19 20 21 22 23 24 25 y2 pe y'-' F 2 ya y2 Total • Species 3.7a 1.3a 2.2a 1.93 3.1a 3.4a 2.0a Freq. 25.99 White-crownedPigeon ...... 2 0.27 MourningDove -- 38 ..... 2 1.68 Zenaida Dove .... 6.5 -- -- , 6 3.35 Ground Dove -- 31 7.0 .... 8 4.05 Key WestQuail Dove ...... 1 0.80 Smooth-billed Ani ...... 2 0.85 Cuban Emerald Hummingbird49 89 347 157 325 10'8 55 22 93.57 BahamaWoodstar 28 -- 286 117 76 25 16 12 21.17 Hairy Woodpecker -- 45 8.3 -- -- 17 -- 18 9.91 Yellow-belliedSapsucker ...... 6 3.18 Gray Kingbird ...... 1 0.07 LoggerheadFlycatcher -- 21 ------5.0 -- 13 3.60 StolidFlycatcher ...... 2 0.50 Greater Antillean Pewee -- 31 9.7 -- -- 30 -- 13 7.57 Bahama Swallow ------34 -- -- 165 4 6.83 Brown-headed Nuthatch ...... 5 8.48 Gray Catbird -- -- 87 100 -- -- 651 11 58.33 N. Mockingbird -- -- 6.6 62 2.0 -- 62 11 4.15 BahamaMockingbird ...... 3.2 1 0.07 Red-leggedThrush ..... 8.3 -- 7 3.44 American Robin ...... 1 0.08 Blue-greenGnatcatcher -- 38 2.3 -- -- 30 7.7 17 29.00 Thick-billed vireo -- 17 16 11 43 38 11 17 22.64 Black and white Warbler ...... 6 4.54 Worm-eatingWarbler ...... 1 0.40 Orange-crownedWarbler ...... 2 0.67 Parula Warbler .... 11 -- -- 5 6.75 Yellow Warbler ...... 1 0.60 MagnoliaWarbler ...... 1 0.27 CapeMay Warbler ...... 3 6.75 Black throated Blue Warbler ...... 2 1.58 Yellow-rumpedWarbler 11 157 222 30 20 7.5 -- 23 66.50 Black-throated Green Warbler ...... 2 0.67 Yellow-throated Warbler -- 16 29 -- 13 31 -- 15 21.75 Olive-cappedWarbler -- 75 17 -- 5.2 334 -- 14 72.00 Pine Warbler -- 62 24 -- 48 60 -- 16 27.87 Prairie Warbler ------27 -- -- 39 12 16.67 Palm Warbler 50 208 410 317 399 19 80 25 214.79 Ovenbird ...... 4 4.17 NorthernWaterthrush ...... 4 10.26 CommonYellowthroat -- 21 9.6 520 17 -- 128 17 50.07 BahamaYellowthroat 5.5 31 36 -- 117 61 154 21 32.15 Wilson'sWarbler ...... 1 0.33 Redstart -- 41 ..... 8 8.46 Red-wingedBlackbird ------20 ------5 5.84 Bananaquit -- 47 15 30 46 91 167 22 55.00 Stripe-headedTanager -- 79 -- 10 13 114 -- 15 36.75 GreaterAntillean Bullfinch ...... 2 2.07 Black-facedGrassquit -- 125 -- 27 8.7 51 -- 16 24.00 SavannahSparrow ------17 -- -- 16 3 1.58 GrasshopperSparrow ...... 16 3 4.83 LincolnSparrow ...... 1 0.33 Totalof species(S) 5 19 17 15 16 17 15 Birdspecies diversity(H") 1.36 2.67 1.95 2.02 1.91 2.31 1.98 Total density(D) 144 1172 1553 1479 1141 1030 1561 Permanentres. (species) 3 15 13 8 12 15 9 Permanentres. (density) 83 745 824 448 703 1003 641 Winterres. (species) 2 4 4 7 4 2 6 Winterres. (density) 61 427 729 1031 438 27 920 x Seasonalstatus: p = permanentresident, W: winter residentor visitant. 2 Vegetationtypes: p = submaturepine, Y • youngpine, Th : low tree thicket, F : old field, Mn : mangroveflat, Msh = Marsh, C : coastalshrub and grassas plottedin Figure 10. a Km of transect. 4 Totals: the density values in this column are for the total of areas covered by censustransects•the sums of the densities in individual stands. 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 25

(hereafterdesignated WR), summerresidents wintering elsewhere (SR), and in-transientmigrants (transients) encountered only in passagebetween their summer and winter homes(Tr). Only the first two of theseelements were present duringthe January-Marchperiod of this 1969 across-communitiesstudy. The ratio of PR to WR speciesvaried widely over the 25 communities(Table 2). In numbersof speciesthe PRs outnumbered the WRs in 20 communitiesin- cludingall of thosein pines,equalled them in one coppetand one mangrove community,and wereoutnumbered in the othermangrove and bothcoastal strip communities.In termsof total population(densities for all speciescombined), PRs dominatedin closed,wooded habitats but were dominatedby WRs in open situations.PRs outnumberedWRs in the youngpines 1.86:1, in the submature pines 1.35:1, and in the coppetcommunities 1.08:1. They were outnumbered by WRs in the old fields1.98:1, in the marsh2.88:1, in the coastalstrips 4.12:1, and in the mangrovecommunities 4.68:1. Bird speciesdiversity.--The diversityof a bird communitymay be expressed simplyas the numberof speciespresent (S), or it may be refinedto incorporate informationon the evennessor equitabilityof abundancesamong the member species(H') (MacArthurand MacArthur 1961). Measuredas number of species (S) diversityin the Grand Bahamacommunities averaged 17.3 speciesand rangedfrom 5 in a sparselyfoliaged, recently disturbed young pine stand(19) to 32 in the denselyfoliaged, tall shrubthicket or "coppet"(6) (Table 2). Using the informationtheory measure of diversity(H'), whichincorporates equitability of distributionalong with speciesnumber, diversity ranged from 0.96 bits in one of the coastalstrip communities (16) to 2.89 in a mixedcoppet (9). H • diversity valuesfor the 25 communitiesare presentedin Figure 12 in a form that shows the relationof diversityto eachof the 7 selectedparameters plotted in Figure 11. Trendscan be seentowards high diversityin shrub-dominatedstands and in standswith high vegetationvolume. Bird speciesdiversity is plottedagainst habitat diversity (both S and H') in Figure 13. The habitat diversityindex for this graph (VHD---see footnotefor column 18 in Table 1) showeda weak positivecorrelation with both the S and the H' indicesof bird speciesdiversity, weaker than might have been predicted by currentdiversity theory (MacArthur and MacArthur 1961). The major de- viantscontributing to this poor correlationwere, on one side, the coppetand the coppet-pinestands (9 and 6) whichwere essentiallyone-stratum habitats because of the suppressionof the groundvegetation by a very denseand tall shrubcanopy. On the other side, the two most deviantsites were the heavily disturbed(recently cut and slashed)pine stands(19 and 7) with vegetationthat, despiteits overall sparsity,had fairly equalshrub and groundstrata plus a few trees. The conditions at thesesites suggest that BSD might showa better correlationwith overall foliage volumethan with vegetationdiversity. When bird speciesdiversity is plotted againstvegetation volume (Fig. 14) a significantcorrelation is found. The correlationis particularlydear for species richness(Fig. 14A). Comparisonsof bird speciesdiversity across habitat types as delineatedin 26 ORNITHOLOGICAL MONOGRAPHS NO. 24

/•' A BIRDSPECIESDIVERSITY (H') / •o•\ e-•.•,-•.oo-• / ...... \ e-•.•,-•.•o-• / \ @-•.,,-•.•o-• / • •e • e-•.o,-•.,o-• / / e-,.,,-,.,o-,

= ß ß ¸ • e eee ß VOLUME OF PINE VEGETATION

,, ©

VOLUME OF VEGETATION IN SHRUB AND GROUND STRATA

FIGURE12. Distribution of bird speciesdiversity (H') along 7 habitat gradients. Symbols indicatethe diversity level in each of the 25 stands,as identified by positionin Figure 11.

Figure 10 providefurther insights into communitystructure and dynamics.Di- versity'(H') washighest in the coppetand pine-coppetcommunity type (2.85), followedby the standardpine (2.38), the marsh(2.30), youngpine (2.10), old field (1.92), mangrove (1.87), and coastal strip (1.45) communities. This sequencesuggests a progression toward low habitatstability, or at leastfrequent 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 27

3.00 - 9 e6 2 2O ßI '• eI ee4 25 2.50 24eee5 14 e2

18 • 20 2.00 25 ee 17 14: 20 el5 21e el2 ß 24ß8 eII ß e5 e2i e 7 t• e25 ß22 ,?. 1.50

,,m,•o 1.00 16 m. eII ß[6

19, 0.50

0 [ I I I I I I I I 0 [ I I I I I I I I 0 0.2 0.4 0.6 0.8 1.0 0 0.2 0.4 0.6 0.8 1.0 FOLIAGE HEIGHT DIVERSITY (H') FOLIAGE HEIGHT DIVERSITY

FIGURE 13. Bird speciesdiversity vs. vegetation diversity in the 25 survey sites. A, species number (S) plotted against foliage height diversity (H') (r: 0.085, P > 0.05); B, bird speciesdiversity (H') plotted against foliage height diversity (H') (r = 0.006, P > 0.05). disturbances,either natural or artificial. Had the series been extended to include urban developmentand lawnsit seemslikely that the trend would have continued. All of the precedingfigures on bird speciesdiversity are for the completewin- teringcommunities, i.e. the permanentresidents plus the winter invaders. With the departureof the WRs each spring, the diversity, of the Island's avifauna is reducedby about one third. This decreaseis most pronouncedin the mangrove and coastalstrip communities,where 29% and 56% of the winteringspecies are WRs, and least in the submatureand young pine communitieswhere only 28% and 25% are WRs. The magnitudeof theseseasonal changes is consideredagain in the next sectionon total populationdensity, and examinedin detail for the pine forest communityin later chapters. Total bird density.--Thesize of an aviancommunity may be expressedin terms of total populationdensity (the sum of all speciesdensities) or of community biomass.Population density values are usefulin studiesof communitystructure and dynamicsand are emphasizedin the presentacross-community comparisons. Biomassvalues including adjustments for variationsin food use rates with body size, are most appropriatefor analysesof consumer-resourcerelationships as consideredin later chaptersof this monograph. Calculatedtotal densityvalues for the 25 Grand Bahamacommunities ranged 28 ORNITHOLOGICAL MONOGRAPHS NO. 24

I I I I I I I 6•

3O 9 e6

• 25

14e 2 18 e8 I• 22• 17•5 ß •:• • •0 el2 14ee20 15• • •1 •11 •10 r• 21ee24 •7 23e e5

0 el9 elõ

hi • 10 ell el6

5-- el9

o • I I I I I I I I I I I o 4 8 12 16 20 24 28 0 4 8 12 16 20 24 28 TOTAL VEGETATION VOLUME (1000rn3/ha)

F[OURF,14. Bird speciesdiversity vs. vegetationdensity (volume) in the 25 survey sites. A, speciesnumber (S) plotted againstvegetation volume (r = 0.878, P < 0.01); B, bird speciesdiversity (H') plotted againstvegetation volume (r= 0.683, P <0.01).

from 130 birds per km• (53 per 100 acres) in a mangrovecommunity (11) to 2,677 (1,084 per 100 acres)in the coppetcommunity (9) (Table 2). Sixteenof the 25 standssupported densities between 800 and !,600 birdsper km•. Environ- mental factors associatedwith these wide differenceswere explored by both the habitat type and gradientapproaches. Densities were highestin the coppetha- bitattype with 1,921birds per km2 (778 per 100 acres),followed in orderby the coastalplains, old field, submaturepines, brushy marsh, young pines, and mangroves,the later with only 293 birds per km2 (Table 2). The effectof humandisturbance on total populationdensities was clearlyvisible in the pine forests.With timbercutting operations, submature forest communities averaging1,006 birds per km• were,in effect,replaced by youngpine communities with 767 birds per km•. This is essentiallywhat must have happenedon a vast scalein the Grand Bahama forestsduring the 1940's and 1950's, followedby a return to the higherdensities as the forestsrecovered in the 1960's. Figuringon this basis,17 of the 21 permanentresident forest species were reducedby the Grand BahamaIsland timber cuttingoperations, 3 of them completelyeliminated on the cut areas, while 4 speciesincreased. The effect on the winter invader 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 29 •" _• TOTALBIRDDENSITY(all sps.) / •,•\ ß-,•oo+ -, / \ O-,•o,-,•oo-• / k (• -,20I--,400- 3 / • •e • •-,oo,-,•oo-• / • k •- •o,-,ooo-• / •%%• _ k o-•o,-•Oo-• o u oo_,

IzJ IzJ

VOLUME OF PINE VEGETATION

,, ©

• ©© @ © o• ©0 0 VOLUME OF VEGETATION IN SHRUB AND GROUND STRATA

FIGURE15. Distribution of total bird density (all species)along 7 habitat gradients. Sym- bols indicate the densitylevel in each of the 25 survey stands. The position of the standsalong the 7 gradientsis taken directly from the calibrated diagrams in Figure I 1 where they are identified. specieswas even greater with 13 or 14 reduced,8 of themlocally eliminated, while 1 speciesbenefitted (see Table 3b). In the diagramsshowing overall populationdensity along a seriesof habitat gradients(Fig. 15) severaltrends can be detected.The centralbasal portion of 30 ORNITHOLOGICAL MONOGRAPHS NO. 24 the triangulardiagram again showsthat open (treeless)stands with low shrubs supportedfew birds, and that densitiesincreased with herbaceous(to left), shrub (to right), and arboreal(upward) cover. Thesetrends can alsobe seenon the horizontalaxis of the lower rectanglein the figure where total volume of shrub and groundvegetation is plottedon a logarithmicscale. As a moredirect test for the apparentpositive relationship between total population densityand the volume of the vegetativesubstrate, I plotted these two functionsagainst each other in Figure 16. The positivecorrelation in this graph (r = 0.354) is not significantat the 0.05 level.

DISCUSSION-•DIVERSITY AND DENSITY CORRELATES

Insofar as environmentalcomplexity implies a wide variety of nichescapable of supportinga varied assortmentof bird speciesthere is logic in the predicted correlationbetween bird speciesdiversity and habitat diversity. MacArthur and MacArthur's (1961) index of habitat diversity(FHD) has be•n usedby many investigators(MacArthur et al. 1966, Karr 1968, Recher 1969) and has served to demonstratethe predictedrelationship in a wide variety of situations.But other investigatorshave failed to find the expectedcorrespondence using FHD and have turned to other indicesbased on plant growth forms (Tomoff 1971) or someaspect of horizontalheterogeneity (Cody 1968, Willson 1974, Roth 1976). In the data presentedin this paper the variableshowing closest correlation with bird speciesdiversity in the Grand Bahamahabitats was total foliagevolume, a a factorapparently reflecting the amountrather than the complexityof vegetation in a stand. The amountof foliagein habitatshas alsobeen examined with positive resultsby Willson (1974). Her measureof vegetationquantity was the sum of the percentcover in eachof threevegetation strata (PCVS), rather than the total spaceoccupied by standingvegetation. In discussingher resultsWillson pointed out that an increasein PCVC generallyimplies an increasein habitat diversity as well as vegetative mass since additional strata or layers are inevitably includedwith high PCVC values. My total foliage volume is, in the same way, an indicator of the structural complexity of the vegetation as well as of volume per se. As I measuredit, foliage volume incorporatesthe full range of vegetativevariation found in all of the compartmentsand foragingsubstrates of the stand. A small volume stand such as a field or mangroveflat may for instancehave only one compartment,while a high volumestand may incorporate many compartmentsdistributed along both the vertical and horizontal planes. Thus habitat diversitymay still be and probably is the biologicallysignificant variablebehind BSD, my measuresof foliagevolume merely being a more sensitive indicatorof habitat complexitythan FHD in the Grand Bahamahabitats. If this interpretationis correct,the bestindex of habitat complexityfor avian community studiesmay indeedbe one basedon somesort of inventoryand quantitativeap- praisalof the compartmentsor guildprovinces present in a stand. In muchthe sameway that the structuralcomplexity of a habitat,measured as the diversityof structuralunits present,is thought to determinethe number of speciesin a community,the total quantityof structuralor substrateunits, regard- 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 31

2400

2000

z 1600 2 bJ ß 25 e2I

16 e8 1200 S15 e3 ß 6

e5 8ooL el

el8 el2 400 elO

ell ©19 I I I I I I I I I I I I I I o6 0.4' 0.8 1.2 1.6 2.0 2.4 2.8 TOTAL VEGETATION VOLUME (mSxlO6)

FIGURE 16. Total bird density (all species) plotted against total vegetation volume in the 25 survey stands (r = 0.354, P •- 0.082). lessof diversity,might be expectedto determinethe numberof individualbirds the habitat will support. The logic behind this propositionrests on two assumptions: (1) that there is at least a rough relationshipbetween the abundanceof foraging substrateunits and the abundanceof avian food organisms,and (2) that food abundanceis an important determinantof population density. My Grand Bahamadata showinga bettercorrelation of vegetationvolume with species numbersthan total density (Figs. 13 and 14) clearly do not supportthis proposition. They cannot,of course,be taken to disproveit, but they suggestthat more cautionmay be desirablein acceptingsome of the basic assumptionsof popular consumer-resourceand carryingcapacity models. They also suggestthat it may be more profitable to examine density-environmentrelations at finer levels Of communityorganization than the habitat level consideredin this chapter. An examinationof this problemat the guild level is undertakenin Chapter 8.

5--BIRD DISTRIBUTION THROUGH THE HABITATS

The quantitativedata collectedon the bird communitiesof Grand Bahama (Table 2) offer an exceptionalopportunity to examinethe dynamicsof density 32 ORNITHOLOGICAL MONOGRAPHS NO. 24 dispersionthrough the habitatsand to evaluatethe nature and mode of action of density-relatedsocial factors in populationregulation and distribution.They alsoprovide a basisfor analyzingthe distributionpatterns of all species,evaluating dispersionamplitudes and overlaps,and examiningspacing characteristics.

DYNAMICS OF DENSITY DISTRIBUTION

Conceptsand models.---Awidely accepted theory of habitatdistribution in birds and other mobile organisms(Lack 1933, 1940, Sviirdson 1949, Hilden 1965) visualizesindividuals as selectingamong the array of availablehabitat patches in an area on the basisof individualexperience and/or of innate responsesthat align preferencewith survival and reproductivesuccess. At the populationlevel such individualresponses result in movementstoward and densityincreases in the habitatswith qualitiesbest suited for the species'particular requirements. The basicfactor in habitat quality for birds is the physicalsubstrate (the vegetation) as it determinesthe abundanceand accessibilityof key resources,particularly food. But otherfactors including the presenceof otherbirds and particularly of conspecificsmay detractheavily from the overall quality of the habitat and effecta net value far belowthe potentialof the physicalenvironment. The detractingor suppressiveeffect on overallhabitat qualityof the population of birdsalready in residenceon an area varieswith the densityof that population (Fretwell and Lucas 1970). It may increaseprogressively with some function of densityas numbersadvance towards the resource-determinedcarrying capacity of the habitat (as in the logisticmodel of populationregulation), or it may be minimalat low densitiesand then appearrather abruptlyand completelyat some socialsaturation threshold level (as in the territory modelsof Kluyver and Tin- bergen 1953, Tinbergen1957, Brown 1969, and others). In the first instance habitatselection may be regardedas a singleresponse to the overall quality of the environment,reflecting a balance of positivelyvalent substratefeatures and negativelyvalent densitycharacteristics. In the secondinstance habitat selection would appearto be completedin two stages: first, a positiveresponse to physical substratefeatures and then a withdrawalimposed by the aggressivebehavior of residentbirds. When this occursdisplaced individuals must move on (spill over) to successivelyinferior sitesuntil they find one not alreadyfilled to the saturation level. The dynamicsof thesetwo systemsof populationdispersion and densityregulation are depicteddiagrammatically for a seriesof hypotheticalhabitat patches arrangedin order of decreasingsubstrate quality in Figure 17. DiagramA (free dispersion)shows the theoreticaldistribution of densities(solid bars) in the absenceof significantdensity factors, a purelyhypothetical situation in whichpopu- lationsshould be free to increaseby ingressor reproductionuntil they approximate the resource-determinedcarrying capacity of the habitats. Diagram B (modified free dispersion)shows the distributionof densitieswhere movement towards the betterhabitats is suppressedas somefunction of the densityof the birds already presentand the resource-determinedcarrying capacity. The bars indicating densitylevels in this seriesdecline less precipitously than in the free dispersal 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 33

A. B. C. FIGURE17. Three alternative models depictinghypothetical density distributionsand regu- lating mechanismsthrough a seriesof habitats arrangedin order of decreasingintrinsic quality. The heightof eachcolumn represents the carryingcapacity in termsof physical features. The height of the solid bars within each column representsthe presumed densitylevel attainedunder the assumptionsof each of the three models: A, established residentshave no direct effect on immigrants; B, establishedresidents lower the attractivenessof the habitat for prospectiveimmigrants; and C, establishedresidents actively repel prospectiveimmigrants as they arrive. model. Diagram C (socially restricteddispersion) shows the distributionof densitieswhere movementinto the best habitatsis arrestedrather abruptly at a thresholdlevel determinedby the saturationof availablespace with aggressive territoryholders. The barsin diagramC form a plateauof roughlyequal densities (through the high quality habitatsand down to the point where the resource- determinedcarrying capacity is lessthan the socialsaturation level of the species). Populationregulation and dispersionby the socialrestriction process (Fig. 17, modelC), while generallyassociated with and bestillustrated in territorialspecies and the breedingseason, need not be restrictedto thesesituations. Indeed, data on habitat distributionin membersof the Grand Bahama wintering community suggestthat it may have featuredin a numberof both permanentresident and winter residentspecies during the January-Marchperiod of studyin 1969. Habitat distributionpatterns.--The densitydata obtainedfor the 25 standsand presentedin Table 2 are organizedin the form of histogramsfor each of the commonspecies (> 3 birds/km•) in Figure 18. In each casethe standwith the highestdensity is shownat the left with the next nine standsaligned to the right in order of decreasingdensity. The variety of patternsin thesehistograms suggeststhat the variousmechanisms of densityregulation and dispersion(see Fig. 17) may differ considerablyamong the 34 speciesrepresented. Some show steep initial declinessuggestive of model #1 (FD) while othershave at least indications of a plateau of moderatelyhigh densitieslike that shown in the third (SRD) model. As a test for the applicabilityof my empiricaldata to those modelsI attemptedan independentand largely subjectiveevaluation of the aggressiveand localization characteristicsof each species. Four categorieswere established: localizedand aggressive(LA), localizedand passive(LP), nomadicand aggressive (NA), and nomadicand passive(NP). The histogramsin Figure 18 are groupedaccording to this classificationwith 14 speciesof type LA at the top, 34 ORNITHOLOGICAL MONOGRAPHS NO. 24

a. b. c. d. Bn:- hd. Mockb'd. Red-Leg. Thr.õ4 h. i I•1 Cam. • Bah. Ytht. •o• Ytht.

A. Localized, aggressive species (LA)

a. b. c. d. e. f. C.E.. ' :[• P•.rn

g. h. i j k. L Ovenb'd. 108

Localized, passive species (LP)

C. d Bah. Sw. BI. E•W. 165 W. SubmaturePines YoungPines Coppet k Mangraves Marshes Coastal / Fields

C. Nomadic, passive species (NP) FIGURE18. Histogramsshowing relative densitiesin descendingorder through the 10 favored stands for each of the 34 common (• 3 birds/kin2) species. Bar heights are relativeto the densityin the mostfavored stand, the first bar. The vegetationtype of the standsis indicatedby shading.Species histograms are arrangedin three groupson the basisof aggressivebehavior.

12of typeLP in thecenter, and 8 typeNP at thebottom. No birdswere assigned to type NA. Most (13 of 14) of the birdsplaced in the localizedand aggressivecategory (LA) arepermanent resident species showing traces of territorialbehavior, and, in somecases, the beginnings of breedingactivity. The onewinter resident speciesin thisgroup (Common Yellowthroat) showed evidence of localizationand occasionalaggressiveness suggestive of winter territorialism. Six of the 12 species in the localizedand passivegroup (LP) werepermanent residents, late nesters, 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 35

TABLE 3 RELATION OF DISPERSIONTYPES (3 MODELS) AND BEHAVIORTYPES (AGGRESSIVENESSAND LOCALIZATION)IN GRANDBAHAMA LAND BIRDS

Behavior type 2 Dispersion modelx LA LP NA NP Free dispersion(FD) 6 6 0 6 Modified free dispersion(MFD) 3 3 0 2 Sociallyrestricted dispersion (SRD) 5 3 0 0 x For description and graphic portrayal see Figure 17. -•LA = localized and aggressive, LP = localized and passive, NA = nomadic and aggressive, NP = nomadic and passive. and/or weakly territorial species. The other 6 were nonflockingwinter migrants of which individualswere seen or mist-netted(Fluck in litt.) repeatedlyin the samelocalities over periodsof daysor weeks. All but one of the 8 in the third group,the nomads,were winter residentspecies; most of them occurredin flocks that seemedto drift irregularly. The singlepermanent resident species assigned to thiscategory, the BahamaSwallow, was still rangingwidely at this season. The relation of localizationand aggressivecharacteristics to dispersionpatterns is presentedin Table 3. Specieswere assignedto the three dispersionmodels of Figure 17 accordingto the amountthey spilledover from their preferredhabitat type to a secondtype. If the drop in densityto the secondtype was greater than 50% the speciesis assignedto the FD model (Fig. 17A), if it was less than 50% but more than 20% it was assignedto the MFD model, and if it was less than 20% it was assignedto the SRD model. This is admittedlycrude sincethere is doubtlessconsiderable variation in the distinctivenessof thesehabitat typesfrom the birds' viewpoint. A positiverelation is indicated for the 26 localized species (LA and LP) with 8 (31%) falling into the SRD patterns;while none of the 8 nomadic speciesfell into this pattern. Similarly 5 (36%) of the 14 aggressive speciesfell in the SRD patternswhile only 3 (15%) of the 20 passivespecies did SO. The divisionpoints for these assignmentsare, of course,quite arbitrary, and a definite effect of either densityor socialfactors is still undemonstratedin any of the species.But I suggestthat the extent of overflow into secondaryhabitat types and beyondis indicativeof socialfactors in at least someof thesespecies unlessresponses to physicalhabitat featuresare far less specificthan commonly assumedin considerationsof habitat selection. Furthermore, the association of aggressivenessand localizationwith heavy overflow seem to support the con- tention that socialbehavior is a critical factor in at least someof the dispersion from preferredhabitat types. Another aspectof densitydistribution, conspicuous in many of the diagrams in Figure 18, is the irregularorder of habitat typesin the sequenceof declining densitiesfrom left to right. Thus a simplehabitat type such as submaturepines may appear at the head of a series,in the middle, and at the lower end, with standsbelonging to severalother typesintruding in the sequence(as in the Log- 36 ORNITHOLOGICAL MONOGRAPHS NO. 24 gerheadFlycatcher--diagram Ac). Such situationsare difficult to reconcilewith the assumptionunderlying all the modelsof Figure 17 that birds tend to move initially towardsstands of the physicaltype best suitedto their specialrequire- ments. The spatialproximity of inferior standsto standsof high quality and densitycould contribute to. this irregularsequence of types,or it is possiblethat my selectionof canopyheight, canopy cover, and foliagetype as key criteriafor delineatinghabitat types was poorlyconceived and quite unrepresentativeof the real situation. Finally one must rememberthat the habitat typeswere based on physicalfactors alone and do not reflect other modifyingfactors such as the presenceof other species.

HABITAT SELECTION

Distribution by types.--The habitat type(s) preferred by each speciesare indicatedin Table 5 by a line underthe highestdensity value(s) in each row. Top preferencewas unevenlydistributed among the 7 types; 22 of the 52 species preferredthe densely-follaged,and essentiallyone-layered copper habitat, a type foundon only a few hundredhectares on the entireisland. Only 8 speciesfavored the widespreadstandard pine habitat; smaller numbers reached top density in each of the other five types. Permanentresidents as a groupoutnumbered winter residentsin the three forest habitats but were outnumberedby them in the four open habitats. Among the 20 common (> 3.0 birds per km•) PR species,8 were most numerousin one or the other of the 2 pine habitats,and only 3 in one or anotherof the 4 open habitats. By contrast,7 WR speciesfavored open habitats,and only one favored the pinelands. Distribution along gradients.-•Density distributionalong the seven habitat di- mensionsrepresented in the diagramsof Figure 11 is plottedfor all the common speciestogether with otherdistributional data in Figures19, 20, and 21. I had thoughtthese plottings might revealclear patternsof distributionfrom which central tendenciesand variancescould be calculatedfor identifyingcritical dimen- sionsand revealingareas of competitionbetween related or convergentspecies. Most of the patternsare diffuse,however, the birds dispersingwidely and irregularly over large portionsor the full range of conditionsavailable to them along each dimension. Obvious specialistssuch as the shrub-orientedBullfinch or the grassinhabiting Savannah Sparrow present more restrictedclusters of points alongcertain dimensions, but thesecases are exceptional. Permanentresidents as a group increasedwith pine canopycover (upward in the trianglediagram) and pine-shrubcover (up and to the right), while the winter residentsdecreased along these gradients (Figs. 22 and 23). Differencesbetween PR and WR componentswere slight along the pine vegetationvolume axis, but WR speciestended to selectstands with taller trees (vertical axis in the central rectangle). The tendencynoted earlier (chapter4) for high overalldensities to accompanyhigh total vegetationvolumes (see Fig. 15) was more pronounced in the PRs than in the WRs. The partial habitatsegregation of permanentand winter residentpopulations 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 37

z <

I I I

I I I I

z <

z 38 ORNITHOLOGICAL MONOGRAPHS NO. 24

i • •] o F.•.•.o• o 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 39 revealedin thesedata could theoreticallyserve to dispersethe exploitationpres- sure on the resourcesof the island during the winter and early spring months. However, I saw no direct evidenceof WR speciesdisplacing PR speciesfrom preferredhabitats. In fact, the GroundDove, ZenaidaDove, and BahamaSwal- low, the threePR speciesthat showedpartial seasonalmovements between habitat types,all movedout to join the WRs in the openhabitat in winter.

DISPERSIONAMPLITUDES (SPECIALIZATION) The degreeof specializationa speciesin its selectionof habitatsas exhibited in its spreadthrough a seriesof standsor along a habitat gradientcould theoretically have an important beating on its interactionswith other species,its populationstability under changingconditions, and its overall successas a member of the avifauna. Dispersionof speciesthrough the 25 stands.---Speciesvaried greatly in the nature and extent of their dispersionthrough the 25 standsin this study (see Table 2). Nine specieswere recordedin only one stand while one, the Palm Warbler, was encounteredin all 25 of the standsin densitiesranging from 19 to 850 birds per km2. Relative populationstrength in each stand in percent total representationis shownin Table 4 for each of the 34 specieswith overall densitiesgreater than 3.0 individualsper km" From these values information theory indicesof distributionalamplitude (H') (Table 4, last 2 cols.) were calculatedon the assumptionthat both the number of habitatsoccupied and the relativedensity of birds in them shouldbe incorporatedin an index to be applied to considerationsof speciesoverlap and communitydynamics. Amplitudes ranged from 0.51 and 0.67 respectivelyin the ecologicallyrestricted Grasshopper Spar- row and Ovenbirdto 2.69 and 2.75 in the ubiquitousCuban EmeraldHumming- bird and Palm Warbler. Equitability (J' = H'/H' max) was consistentlyrather high (•> 0.75) for the specieswith 10 individualsper km• regardlessof the amplitude value. Dispersionby types.--Theuneven representation of habitattypes in the sample of standssurveyed creates strong biases in the measuresof distributionalamplitude presentedabove. For instance,a bird specializedfor pine foliagemight easily dispersethrough all of the 13 standscontaining 100 or more pine treesper hectare (see Table 1), while a speciesadapted to brushy marsh situationswould have only one standto selectfrom. An analysisof spreadthrough the sevenhabitat typesdelineated in Figure10 wastherefore made in whichmeans were calculated for the standardpine standsand other multi-representedtypes so as to equate themwith the lightlyrepresented types such as the marsh. Dispersionthrough the sevenhabitat types differed greatly from speciesto species(Table 5). Among the 28 permanentresident species 2 were encountered in all 7 typeswhile 6 were found in only one type each (column 8); distributionalamplitudes (H') variedfrom 0-1.75 and averaged0.77. Winter resident speciesshowed similar dispersion patterns with 3 of the 24 occurringin all 7 typesand 6 limitedto a singletype; distributional amplitudes ranged from 0 to 1.72 and averaged0.54. 40 ORNITHOLOGICAL MONOGRAPHS NO. 24

C)- SPECIES ABSENT (•)- 1- 8 BIRDS/km2 (•)- 9- 29 (•)-Q- 60-•0- 12459 • PROPORTIONSTREE,SHRUeOFe •- 125-249 / • GROUNDCOVER O- /o0sED\

White-crowned Pigeon Zenaida DOVe

Ground Dove Key West Quail Dove

Smooth-billed Ani Cuban'Era. Hummingbird Bahama Woodstar Hoiry Woodpecker

y- bellled Sapsucker Loggerhead Flycatcher Stolid Flycatcher Gr. Antilleon Pewee

Bahama Swallow Brown- hd. Nuthatch Gray Catbird Mocking bird

Bahama Mockingbird Red-legged Thrush Blue-gray Gnatcatcher Thick-billed Virea

FIGURE 19. Winter distribution of 44 specieswith respect to vegetation type. The 25 circles in each diagram representthe 25 survey stands,positioned with respect to the relative amounts of trees, shrubs, and ground cover open to the sky. The circles (stands) are identified by number in the large diagram at the head. The population density (in 8 density classes) is indicated for each stand by symbolsin the circles. A key to the symbolsis shownat the head. Further explanationis givenin the legendof Figure 11. 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 41

Black a White Warbler Parula Warbler Yellow Warbler Cape May Warbler

Black*thr. Blue Warbler Black-th• Green Warbler Yellow-rumped Warbler 'Yellow-throated Warbler

O{ive*capped Warbler Pine Warbler prairie Warbler Palm Warbler

Ovenbird No. Waterthrush Common Yellowthroat Bahama Yellow throat

Redstart Red Winged Blackbird Bancncquit Striped-headed Tanager

Gr. AntilleanBullfinch Grassquit Savannah Sparrow Grasshopper Sparrow

Dispersion along gradients.-•Dispersionalong the seven selectedhabitat di- mensionsand severalcombination dimensions are graphicallyportrayed for most speciesin the diagramsin Figures19, 20, and 21. The broad and irregulardensity distributionsseen in most of these diagramsindicate a general absenceof close environmentalrestriction with respectto the dimensionsadopted and, show the 42 ORNITHOLOGICAL MONOGRAPHS NO. 24

TREE FOLIAGE VOLUME ß HEIGHT {m)

, ,, ,'• ,'• ,: (iooo rn$/km2 )

oo &, o ©o .• o ooo'9 oo • o ©ooœ o• t White-crownedPigeon 2. Zenaido Dove

¸ © C• 0 0 0 0 0 O0© C• 0 0 O00 C•)0 $. Ground Oove 4. Key West Quail Oove

o ooo•oo o o • *•*½ • 6. Cuban Emeruld Hummingbird

o o o 0 00© •:) @ 7. Bahama Woodstot 8. Huiry Woodpecker o ooooo øe o o 0©0,9 © 9. Yellow-bellied Sapsucker I0. Loggerheud Flycatcher

0 0 (• 0 ¸ © • o ooo8 o o oo• II. Stolid Flycatcher 12, Gr. Antilleon Pewee

oo • o oo • © o oooœ o o ooo8' o I$. Bahama Swallow 14. Brown-headed Nuthatch

• 0 (• 0 © 0 0 o oo• øøo. o o ©o©œ zo 15.Gray Catbird 16. Mockingbird

oo &, o ©o o ooo,9 o o ooo,9 17.Bahama Mockingbird liB.Rod-legged Thrueh

o o©©,½© o ©©©eø 19.Blue-gray Gnatcatcher •0. Thick-billed Vireo

FIGURE 20. Winter distribution of 42 species with respect to vegetation height (vertical axis) and volume of pine foliage (horizontal axis). For explanation see Figures 11 and 19. futility of attemptingto apply traditionalstatistical methods for determiningvari- ancefrom the centersof distribution.Standard analytical methods were applied, however,to the distributionof five pine forest specialistswith limited dispersion along nine pine forest gradients(Table 6). Centers of distributionalong each 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 43

@ o c• o © o o i•,oo oO0 C•)o • 0 O00 C•)0 0 21. Block P- While Warbler 22. Porula Warbler

0 0 0•) 0 © o o 0 000 O00 • 0 O0 0 {•)'0 0 23. Yellow Warbler 24. Cope May Warbler

25. Yellow-romped Warbler 26. Yellow-lhraoled Warbler

0 @ @ 0 A• Olive-c•pped W•rbler 28, Pine Warbler

• 0 0 O• 0 • 29. Prairie Warbler 30. Palm Warbler

33. Common Yellowlhrool 54. Bahama Yellowthrool

©© .• o oo -0% o o ooo •ø o o ooo øøo 35. Redslarl 36. Red-winged Blackbird

o ©•©• • o ©oo• 37. Bananaquil :58, Striped -headed Tanager

oo 8• o ©• o• o ooo øøo o ©oo o©© 59. Gr. Anlillean Bullfinch 40. Grassquit

oo 0% o oo 0% o o oo oøø o o ooo øøo 41. Savannah Sparrow 42. GrasshopperSparrow dimensionwere calculatedfor thesespecies as meansby weightingthe value of eachstand's position on the gradient(multiplying it by the bird density)and dividingthe resultantmean by the highestvalue available to the birds,i.e. that of the standpositioned highest on the gradient(X = [f(x)/N]/ x Max.). The measureof dispersionis the standarddeviation from the actualmean (not the percentvalue), divided by andrepresented asa percentof themean (C.V. x 100) so as to allowfor comparisonsbetween species. Mean populationdensities for the pine standsand other (not pine) standsare givenat the bottomof the table for ready reference. 44 ORNITHOLOGICAL MONOGRAPHS NO. 24

SHRUB VOLUME x TYPE (BROADLEAF: OTHER)

o 9 i0 12 i• 20 25 ( tOO0m3/km 2)

oo oOo•O o© o%oO

oo©o d• © oooo oø oo o•ooO oo oooo© ;•. Groud Dove 4. Smooth-billed Ani • o• • •o •

5. Cuban Em. Hummingbird 6. Bahama Woodstar • o• • • o• • o •o• • • •o•O Z Nair• •o0•r B. L0•r•a• •l•at•r

S. Stolid Flycatcher I0. Gr. Antillean Pewee

II. Brown-headed Nuthatch 12. Gray catbird

ooo ooø oo oOo•O 14. Bahama Mockingbird

15, Red-legged Thrush 16. Blue-gray Gnatcatcher

FIGURE21. Winter distributionof 36 specieswith respectto shrub type (vertical axis) and volume of shrub and ground vegetation(horizontal axis). For explanationsee Figures 11 and 19.

In this table a low coefficientof variation (dispersion) relative to that shown by the other four specieson the same dimensionsuggests that the speciesis usingcues associated with that dimensionin selectingits habitat, while a high coefficientsuggests that the dimensionis relativelyinconsequential to the species in making its selection.However, complicatedecological relationships between variousparameters of the forestmay negatethe validity of suchinterpretations. 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 45

e el

17. Thick-billed Vireo IS. Black 8• White Wetbier •© 19. P,o•u{oWorble• 20. Yellow-rumpedWorbler 0© 21. Yellow-throoted Wetbier 22. Olive-copped Worbler

ooo

2•. Pi•e Worbler -24. Proifil

25. pQlm Werbler 26. Ovenbird

•0

For example,the low C.V. valuesfor the Brown-headedNuthatch along the pine canopy,tree density,and foliagevolume dimensions may indicatea key role for thesedimensions for habitatselection in the species,but the equallylow coefficients (relative to other species) along the shrub cover and shrub volume dimensions probablyshould not be interpretedin the sameway for this normallyaboreal 46 ORNITHOLOGICAL MONOGRAPHS NO. 24

.•. .• PERMANENTRESIDENT / F%.• • • \ e - .oo,-.2oo / ...... \ e- 8o.-,ooo / \ •- so,--8oo / • _e \ • so,-soo / • \ ©- 3o,-zoo / _ \ ©- ,o,-3oo

VOLUME OF PINE VEGETATION

VOLUME OF VEGETATION IN SHRUB AND GROUND STRATA FIGURE22. Distributionof total populationdensity along habitat gradientsfor the permanentresident species of the winteringcommunity of Grand Bahama. The points representingthe 25 standsare positionedas in Figure 11 where they are identified. species.They could indicate that when the birdsvisit the understorythey are very selective,but morelikely they simply reflect the effectof a densetree canopy on the growthform of the understory. For most of the speciesrepresented in Figures 19, 20, and 21 I evaluated dispersionor specializationin termsof the proportionof the availablerange of variationthat was occupiedalong each parameter gradient. A specieswas classed as a specialistwhen its distributionalong the indicatedgradient was restrictedto aboutone-third or lessof the range availableto it. Speciesdispersed over about two-thirdsof the rangewere classedas intermediates,and thosewell dispersedover the entire range, as generalists. Using thesecriteria there were more generalistsamong the winter residentsthan the permanentresidents on all of the dimensionsexcept groundcover exposed to the sky (Table 7, cols. 1 and 2). Further, when the 3 dimensionsgraphed in 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 47 WINTERRESIDENTDENSITIES e-,ooi-,•oo e- o,-,ooo @- o,-oo

VOLUME OF VEGETATION IN SHRUB AND GROUND STRATA FlOrroE 23. Distribution of total populationdensity for the winter residentspecies. the trianglediagrams are consideredtogether only 2 of the 12 specialists(17%) wereWRs while 15 of the 32 generalists(49%) were WRs. Similarly,when the 2 understorydimensions are combined,only 1 of the 6 specialists(17%) were WRswhile 11 of the 30 generalists(37%) wereWRs. Generalizationis a majority conditionin all cases;its higherincidence in winter residentsmay have evolved as an adaptationfor, or arisenas a consequenceof the migratoryway of life. Colonizersderiving from island sources might be expectedto be lessspecialized (broaderamplitudes) in theirhabitat responses than those deriving from conti- nentalareas where the numberof speciesand communitydiversity is character- isticallygreater, and habitatsare presumablymore finely apportioned among the membersof the fauna. The data from Grand Bahama,however, provide little or no supportfor suchan hypothesis.Using the H' measureof diversityacross habitat types (Table 5) as the measureof distributionamplitude, and an H' value of 0.90 as the arbitrarythreshold of specialization,7 of the 13 species 48 ORNITHOLOGICAL MONOGRAPHS NO. 24

TABLE 5a DISTRIBUTION OF SPECIES THROUGH THE SEVEN HABITAT TYPES ON GRAND BAHAMA ISLAND (PERMANENT RESIDENTS)

Standard Young Old Coastal Types (H t) Perre. res. species pines pines Coppers Mangrove Marsh fields sand (S) Div. White-cr. Pigeon -- -- 3.85 .... 1 0 Mourning Dove 4.75 -- -- 15.00 ------2 0.551 Zenaida Dove* 4.18 0.93 5.00 .... 3 0.932 Ground Dove* 7.60 1.00 1.75 ------41.50 4 0.650 K. W. Quail Dove ...... 25.50 1 0 Sm.-billed Ani -- -- 10.5 -- 5.70 -- -- 2 0.650 C. EmeraldHum'bd* 98.50 155.00 90.5 -- 12.00 73.80 35.00 6 1.569 Bah. Woodstar* 2.79 72.54 ------48.00 54.00 4 1.146 Hairy Woodpecker* 14.14 6.61 39.5 3.95 2.50 -- -- 5 1.160 Gray Kingbird ------0.80 ------1 0 Logg.Flycatcher* 6.73 3.34 3.85 .... 3 1.047 Stolid Flycatcher 0.25 -- 3.50 .... 2 0.245 Gr. Ant. Pewee* 11.04 9.29 9.00 .... 3 1.094 Bahama Swallow* .... 61.00 66.33 4.60 3 0.817 Brown-h. Nuthatch* 16.98 ...... 1 0 Mockingbird* 1.50 1.23 1.05 0.75 1.30 44.00 4.60 7 0.795 Bah. Mockingbird ..... 1.07 -- 1 0 Red-leg. Thrush* 1.99 1.19 28.90 .... 3 0.389 Bl-gr. Gnatcatcher* 30.96 11.76 153.00 -- 10.00 2.57 -- 5 0.873 Thk.-billed Vireo* 15.21 18.91 133.50 -- 35.00 7.33 -- 5 1.112 Yel.-throated Warb.* 30.51 15.90 9.55 .... 3 0.890 Olive-cap.Warb.* 113.50 69.46 3.50 -- 20.00 -- -- 4 0.991 Pine Warbler* 39.75 38.29 9.00 -- -- 1.43 -- 4 1.023 Bah. Yellowthroat* 28.71 46.36 80.05 7.00 -- 53.50 7.00 6 1.517 Bananaquit* 61.66 28.91 159.50 20.85 39.00 65.67 69.50 7 1.752 Str.-headedTanager* 44.38 24.93 156.50 -- -- 3.33 -- 4 0.880 Gr. Ant. Bullfinch -- -- 34.00 .... 1 0 Grassquit* 37.70 8.53 50.10 3.15 34.00 9.00 -- 6 1.488 28 species 21 18 21 7 10 12 8 Total P.R. 572.8 510.8 986.1 51.5 220.5 376.0 241.7

(54%) of Antillean derivationwere specialists,while 7 of 15 (47%) of the continentalspecies were specialists.In analysesof spreadalong the 7 habitat dimensionsof the habitat diagrams,species of Antillean origin had higher incidencesof specializationthan continentallyderived speciesin 3, lower in 3, and equalin 1 (Table 7). Lookingat recencyof colonization(Table 7, last 2 cols.) as indicatedby lack of recognizedspecific or subspecificdifferentiation, the very smallsample showed a higherincidence of specializationamong the recent arrivalson 5 of the dimensions,an equal representationon 1 (exposedground cover), and a lower incidenceon 1 (height of pine trees).

DENSITY-DISPERSION RELATIONS

The relation betweenhabitat tolerancelimits as revealedby dispersionindices (H • valuesin Table 5a) and densitytolerance limits as revealedby densitiesin the most preferredhabitat (maximumdensities in Table 5a) were explored 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 49

TABLE 5b DISTRIBUTION OF SPECIES THROUGH THE SEVEN HABITAT TYPES ON GRAND BAHAMAISLAND (WINTER RESIDENTS)

Standard Young Old Coastal Types Div. Winter res. species pines pines Coppets Mangrove Marsh fields sand (S) (H t) Yel.-belliedSapsucker 4.29 -- 22.00 .... 2 0.450 Catbird* 29.25 .-- 251.00 -- 54.00 249.40 14.00 5 1.181 American Robin .... 1.70 -- -- 1 0 B1.-and-Wh. Warb.* 6.25 -- 16.05 .... 2 0.593 Worm-eatingWarb. 0.91 ...... 1 0 Orange-crownedWarb. 0.61 -- -- 3.65 ------2 0.409 Parula Warbler* 5.63 1.57 56.90 .... 3 0.767 Yellow Warbler ------6.50 ------1 0 Magnolia Warbler -- -- 2.30 .... 1 0 CapeMay Warbler* 9.75 -- 6.00 .... 2 0.658 Blk.-thr. Blue Warbler -- -- 48.80 .... 1 0 Yel.-rumpedWarbler 78.88 43.43 120.50 40.00 146.00 16.67 23.00 7 1.715 Blk.-thr. Green Warb. 0.53 -- 7.00 .... 2 0.254 Prairie Warbler* 14.63 0.83 87.00 -- 17.00 24.90 9.50 6 1.279 Palm Warbler* 255.50 196.71 60.50 71.50 25.00 160.67 558.50 7 1.602 Ovenbird* 1.11 -- 63.00 .... 2 0.086 N. Waterthrush* -- -- 48.50 27.00 125.00 -- -- 3 0.900 Com. Yellowthroat* 13.49 9.80 108.00 43.50 228.00 216.00 341.00 7 1.536 Wilson's Warbler -- -- 6.50 .... 1 0 Redstart* 12.69 2.71 48.50 -- 5.70 -- -- 4 0.894 Red-w. Blackbird* ------49.00 31.00 6.67 2.65 4 1.007 SavannahSparrow ..... 11.00 34.50 2 0.552 GrasshopperSparrow* -- 1.04 ------51.67 -- 2 0.098 Lincoln Sparrow ...... 11.50 1 0 24 species 14 7 16 7 9 8 8 Total W.R. 433.5 256.1 912.7 241.2 633.4 751.3 994.7 TotalP.R. &W.R. 1006.4 766.9 1898.8 292.7 853.9 1127.3 1236.4

Values are individuals per km'-' in each type. *: species with total densities greater than 3.0 birds/kin 2. Values for the preferred habitat(s) of each species are underlined. by plottingthe recordedvalues for thesetwo parametersfor all speciesin Figure 24. The pattern that emerges shows an overall positive correlation between maximum density and habitat dispersion. But the regressionslope obviously changesnear midpoint, suggestingthat different regulatoryfactors may be operating under differentconditions of densityand/or dispersion. An examination of the overall pattern of points in terms of the models of dynamic density and dispersionregulation depicted in Figure 17 suggestsseveral relationshipswith heuristic implications. Of special interest is the absenceof points in the lower right quadrant. This seemsto indicate that no specieswith a high (> approx. 65 birds/km•) densityin its preferredhabitat type failed to spill over into secondarytypes in large numbers. Apparently the densityregulating factor(s), whether carrying capacity or social saturation, were functioning and causingoverflow in all high density species. In a similar vein the points in the lower left quadrantmay representspecies that for one reasonor anotherhad not reacted to either social saturationor carrying capacity levels in their preferred habitat. Points in the upper left quadrant may by the same token represent 50 ORNITHOLOGICAL MONOGRAPHS NO. 24

TABLE 6 POSITION AND DISPERSON OF FIVE PINE FOLIAGE GLEANERS ALONG NINE GRADIENTS OF THE PINE FOREST HABITATS 1

Brown-headed Blue-gray Yellow-throated Olive-capped Pine nuthatch gnatcatcher warbler warbler warbler

•a C.V. a • C.¾. • C.V. • C.V. • C.V. Pine tree parameters Canopy cover .87 10 .69 23 .72 26 .79 16 .66 28 Tree density .87 10 .71 24 .73 27 .80 15 .66 29 Tree height .77 23 .63 33 .60 46 .51 48 .49 57 Foliage volume .86 10 .69 23 .72 35 .79 15 .58 37 Understory parameters Shrub cover .49 11 .62 33 .60 44 .50 17 .59 26 Shrub height .24 104 .49 91 .38 86 .09 271 .18 183 Shrub type .96 23 .94 15 .92 17 .80 26 .69 38 Shrub volume .61 16 .74 28 .69 27 .54 24 .64 27 Ground cover .66 26 .68 31 .63 35 .68 26 .63 32 Population density X max X max • max • max • max Pine forests 10.5 48 25 62 27 56 107 334 44 72 Other habitats 0 0 27 207 3.2 11 2.3 20 3.0 18

Stands with greater than 20% pine cover. Position of the center (mean) of distribution on the gradient as a percent of the maximum available (x/x max). Standard deviation expressed as percent of the mean (tz/• X 100). specieswith low socialsaturation thresholds that had reachedthose thresholds in the preferred habitats and spilled over extensively,or specieswhose preferred habitatshad low carrying capacitiesthat had been filled and overflowed. And finally, pointsin the upper right quadrantmay representspecies that had high maximumdensity tolerances in habitatscapable of supportingdense populations but had neverthelessreached and exceededthose limits in their preferred types and spilledover into secondarytypes. The possiblerole of interspeciescompetition and other complexfactors operating

TABLE 7 HABITAT SPECIALIZATION ALONG SEVEN GRADIENTS IN VARIOUS ELEMENTS OF THE GRAND BAHAMA LAND BIRD COMMUNITY 1

Community elements 2

Seasonal Geographic Colonization status derivation recency Habitat parameters P.R. W.I. Cont. Ant. C.-A. Recent Old Vegetation layer open to sky Pine cover 37 29 67 17 42 50 40 Shrub cover 32 12 17 17 42 50 25 Ground cover 22 35 17 17 33 25 25 Pine vegetationvolume 30 24 50 0 42 50 30 Pine tree height 15 6 17 17 17 0 20 Understoryvolume 42 17 50 33 42 75 35 Understorycomposition 21 8 17 0 33 50 15 N = 27 17 6 6 12 4 20 x Values are incidence (%) of specialized species in the community where specialization is defined as distribution restricted to one-third or less of the range available. 2 Community elements analyzed are: seasonal status (permanent residents and winter residents), geographic derivation (continental, Antillean, and continent via Antilles), and recency of colonization (recent or old as suggestedby degree of taxonomic differentiation). For definitions of terms see text. 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 51

a.00 H' max. = 1.946

Y.-rump. W.ßBonq't. o

Palm W. o Boh. Ytht. ßC.E. Hum. Cam.Ytht. o GrassQt. ß ß

Pt. W. o e- 1.20 H.Wdp. ß B.Wdst. Th-bV e- •ewee ß ß C• •Log.ß F.Pine W.( o 1.00 ßo R- W. Bkb'd. ß0 - Cap.W. Cat b'd. e- _•en.D. I o No. Wthr. ß Y-th.W. I o ßTan. ß Red st. ßGnat. Mock.ø__ .• B_.S_.W. 0.80 -- ß oøPar. W. Saps.

C.M.W.oßAni ßGd. Dove '- 0.60 oB.I•W. ß o E M.D. Say.Sp. O-Cr.W. c 0.40 o R-L.Thr. •5-th. Gn.W.

ß•- --'---I[O•o•_Stolid See \E C3 IJLegendJGr. Sp.j ISBeJow/ o p! Vbmd;80m ! 120I I 160m m200 , / ,• I 300I I 400I I õ00I I PopulationDensity in the PreferredHabitat Type (Birds/kin2) Fmo-•E 24. Relation of dispersalinto secondaryhabitats (ecological amplitudes) to population density in the preferred habitat. Solid circles indicate permanent resident species; open circles indicate winter residents. The three permanent residents and four winter resident speciesgrouped near the lower left corner are, reading from the left: Magnolia Warbler, Yellow Warbler, Wilson's Warbler, Lincoln's Sparrow, Brown-headed Nut- hatch, Key West Quail Dove, and Greater Antillean Bullfinch. Symbols for species can be identified by referring to the list of speciesin the Appendix or in Table 2. withinhabitats is of coursebypassed in theseanalyses. The omissionis unfortunate, but until methodsfor evaluatingthese factors as separateentities and applying them as modifiersare developed,they must'be treatedsimply as aspectsof the habitatsaffecting both densityand dispersionin as yet unmeasuredways.

OVERLAP AND SIMILARITY Measuring overlaps.--Similar patterns of habitat responsein two or more membersof a fauna imply a certain amount of cohabitationor distributional 52 ORNITHOLOGICAL MONOGRAPHS NO. 24 overlap,especially when the distributionalamplitude of one or both of the species is broad. Furthermore, a high incidenceof overlap will tend to increasethe frequencyof encountersand behavioral interactionsbetween members of the two speciesand, at leastpotentially, raise the level of competitionfor resources in shortsupply and/or for favorablepositions in the vegetation.Habitat overlap or frequencyof cooccupationof habitats is thus a critical considerationin the dynamicsof a fauna such as that of Grand Bahama Island. In this analysiswe are concernedprimarily with the interactionsof individuals as representativesof species. The appropriatemeasure of similarity or overlap between any two speciesshould thus reflect the characteristicsof the species regardlessof the number of individualsactually involved in the interactions. I have thereforeused percent values for the distributionthrough the sevenhabitat types and determinedthe overlap for each speciespair by adding the overlaps for all caseswhere the two occurredtogether. I then divided the sum by the maximum possibleoverlap, always 100%, to obtain values suitable for direct comparisons.Such values have little value in themselvesbut hold considerable heuristicpotential for comparisonsof phylogenetic,geographic, seasonal, and perhapsother elementsin a community. A tabulation of overlaps between habitat types is presentedfor all pair combinationsamong the 34 specieswith densitiesgreater than 3 birds per km2 in Table 8. Overlap and phylogenetic relationships.--Closelyrelated species with their basicallysimilar morphologymight be expectedto be more similar in habitat distributionand showmore habitat overlap than remotelyrelated species,except as competitionin sympatrymight promoteecological displacement and accelerate evolutionarydivergence. Overlap values (Table 8) were very high (> 80%) between the 3 congenericwarblers of the permanentresident element, the Pine Warbler, Olive-capped Warbler, and Yellow-throated Warbler, but aside from this there is no evidencein the data of a positivecorrelation between phylogenetic relationshipand habitat overlap. The incidenceof high (•> 60%) overlap in Table 8 was 4/30 (13.3%) for congenericpairings, 9/97 (9.3%) for confamilial pairings,and 116/561 (20.7%) for all pairings. The slightlyhigher incidence for congenericsthan confamilialsin this seriescan be accountedfor entirelyby the high values for the 3 PR warblers mentionedabove. The Grand Bahama sampleis frustratinglysmall, but if this apparent negativecorrelation with phylogeneticrelationship is real, the three casesof high overlap in congeneric warblerscould reflect unresolvedcompetition following recent colonization,while the low overlap among confamilialsreflects relations between species that have been interactingon the Island over longerperiods of time. Overlapand geographicderivation.--Island colonists from the samegeographic source,e.g. eitherthe North Americancontinent or the Antilles,were presumably exposedto eachother in manycases before they immigrated. As group3they might thereforebe predictedto have evolvedgreater ecological independence and show less overlap than speciesconverging on an island from different sources. The data seemto supportthis prediction: I comparedthe incidenceof high overlap (> 80% in Table 8) in pairingswhere both specieswere of continentalorigin 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 53 54 ORNITHOLOGICAL MONOGRAPHS NO. 24

TABLE 9 HABITAT OVERLAP AND GEOGRAPHIC DERIVATION 1

Overlaps• 80% C-A C-C A-A A-A Instances 10 2 2 4 Incidence (%) 71.4 14.3 14.3 28.6 Pairs available 84 21 66 87 Expectedincidence 49.1 12.2 38.6 50.8 Actual/expected 1.45 1.17 0.37 0.56 x Comparison of the incidence of high (80%) overlap in habitat distribution between pairs of species'where the two came from different geographic sources (continent-Antilles) and where both came from the same source (continent-continent, or Antilles-Antilles).

(C-C), or Antillean origin (A-A) with thosein whichthe two camefrom different sources(C-A) (Table 9). Of the 14 PR speciespairs availablefor the analysis,4 had commonorigins and 10 had split origins. Four is only a little more than half that expectedfor the C-C and A-A categorieson the basisof availablepairings, while 10 is nearly 50% abovethat expectedfor the mixed (C-A) originscategory. The data thus appearto supportthis prediction. The possibilitythat some WRs might be dominatedby PRs and forced into partial segregationin less favorable habitats, as suggestedby Morel (1968) and Moreau (1972), or that WRs might simply avoid establishedresidents, territorial or not, in their preferredhabitats by settlingelsewhere, was testedby comparing the incidence of habitat overlap in the inter-element pairings, i.e. between PRs and WRs with the intra-element pairing of the PRs (Table 10). Matched against the ratios expected on the basis of available pairings,there is an indicationof the predictedlower incidenceof overlap in the pairingsbetween PRs and WRs. However,with no controlsfor suchfactors as differentfood habitsand distributionalamplitudes among WRs, this difference cannot be regardedas more than suggestiveof interaction-inducedsegregation. This questionis examinedagain at the within-habitatlevel in Chapter7. Ecologicalspacing.--The extent of habitat overlapbetween two speciesdepends on the similarityas well as the amplitudeof habitatresponses. Habitat similarity, or the closenessof the centersof ecologicaldistribution of speciesin multidimen- sionalhabitat spacemay be representedgraphically by an ordinationprocedure developedby Bray and Curtis (1957) and applied to bird communitiesby Beals (1960, 1973) and Emlen (1972) in which the coordinatesrepresent axes

TABLE 10 HABITAT OVERLAP AND RESIDENCY STATUS x

Overlaps• 80% P.R.-W.R. P.R.-P.R. W.R.-W.R. Instances 14 14 5 Incidence(%) 42.4 42.4 15.2 Pairs available 280 190 91 Expectedincidence 49.9 32.9 16.2 Actual/expected 0.85 1.29 0.94 aComparison of the incidenceof high(• 80%) overlapin habitatdistribution between (a) permanent residents and winter residents, (b) permanent residents and permanent residents, and (c) winter residents and winter residents. Expected incidence is the percent of high ovarlaps in the total sample of the indicated category. 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 55 of variationin overlaprather than specificvariables. The distancebetween any two pointsrepresenting species in suchan ordinationis thusa directreflection of similarityin habitat responsesas determinedby the amountof overlaprecorded for all speciespairs in the community.A two-dimensionalordination of this sort basedon the overlapsshown in Table 8 is presentedfor the sevenhabitat types of Grand BahamaIsland in Figure 25. I usedthe overlapsbetween habitat types rather than betweenstands because of the unequalrepresentation of the major typesin the sampleof standsavailable. The distancebetween any two pointsin this graph is an objectiveindication of the similarityin mean habitat responses of the two speciesthey represent,and any groupingof pointsindicates a cluster of ecologicallysimilar species.These distances are, of course,based entirely on the frequencyof habitat overlapor associationthrough the seriesand, as plotted, are independentof any presumptionsof causalfactors. Thus the axesin Figure 25 cannot be labelled. In Figure 25 most of the speciesare groupedclose to one of the terminal specieson the x axis (Parula Warbler), indicatingminimum overlap with the speciesat the other terminal(Grasshopper Sparrow). Similarlythey are grouped closeto the Brown-headedNuthatch and far from the Red-wingedBlackbirds, the two terminal specieson the y axis, calculatedafter and independentlyof the x ordination. Primarily pinelandspecies are, not surprisingly,grouped near the nuthatch,a pine forestspecialist, while open habitat speciestend to centernear the blackbird. Speciesrelated at the family level are well scatteredover the grid, but congenericstend to be clustered.Winter invaderspecies (open circles) seem to be moreperipheral than the permanentresidents (closed circles) in the overall scatterand showa generalconcentration near the Parula Warbler at the right end of the x axis. In a similar vein, many of the PRs tend to be clusterednear the nuthatchat the top of the y axis. Further analysisof the environmentalfactors implicatedfor the two axesin the figureand for additionalaxes not shownare not attemptedin this report.

6--THE PINE FOREST COMMUNITY---SEASONAL CHANGES

PROCEDURES AND DEFINITIONS In the rest of this monographI focus on the bird communitiesof the submature pine forests, the dominant and most complex of the habitat types on Grand Bahama. Most of the material for these analyseswas collectedin stands 1, 2, and 3 (Table 1) in 1968, supplementedwith data obtainedin the across habitat studies of 1969. Details of the within-habitat structure of these forests are presentedin Table 11 and summarizedgraphically in Figure26. The floristic compositionof the tree and shrubstrata is summarizedin Table 12. For tracing the seasonalchanges in communitycomposition, density and diversityin the forest I use the samebasic systememployed in the acrosshabitat studiesbut refine them for more preciseanalyses. Thus I recognizethree seasonal communitiesand four residencyelements: a winteringcommunity with permanent (PR) and winter resident(WR) elements,a breedingseason community with permanent(PR) and summerresidents (SR), and a transientcommunity in the spring (and fall) when PRs and persistingWRs and/or SRs are temporarily 56 ORNITHOLOGICAL MONOGRAPHS NO. 24

I I [ I

I I 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 57

TABLE 12 DOMINANT AND OTHER IMPORTANT TREE AND SHRUB SPECIES IN THE SUBMATURE PINE FORESTS

Tree stratum Lysilomalatisiliquum 3.3 Pinus caribaea 99.7 Ernodea littoralis 2.9 Lysilomalatisiliquum 0.3 Tetrazygiabicolor 2.7 Trema lamarkianum 0.5 Shrub stratum Ficus aurea 0.6 Metopium toxiferum 23.3 Swietianamahagoni 0.5 Thrinaxmorrisii 20.6 Durantarepens 0.9 Pteridium aquilinum 8.3 Solanum erianthum 0.5 Tabebuia bahamensis 4.0 Calliandra haematomma 2.1 Lantana ovatifolia 3.8 Smilax havanensis 1.8 Coccolobadiversifolia 3.7 Myrsinefloridana 1.3 Myrica cerifera 3.5 Eugeniaaxillaris 1.4 Agave bahamensis 0.6 Values are % representationin their respectivestrata (based on occurrenceat 1200 sample points). Identificationsare from Britton and Millspaugh (1920) with revisionsby W. Gillis. supplementedby migrantsin transitbetween northern and southernranges (Trs). The temporalrelations of thesethree communities and four residencyelements on Grand Bahamaare diagrammedin Figure25. March 31 was arbitrarilytaken as the terminationdate for the winteringcommunity, and all WRs after 1 April were groupedwith the transientelement to constitutethe transientcommunity presentduring April and early May. Density(individuals per km2) andbiomass (g per km2) valuesfor eachspecies in a seasonalcommunity were calculated on the basisof the numberof equivalent km• transectedwhile the particularspecies was present. For membersof the winteringcommunity, this period covered the entire90 daysin January,February, and March (both years). For the breedingseason community it includedall of April and May for the permanentresident species, but only the periodbetween arrivalon theisland and the end of May for the summerresident species. Members of the transientcommunity were presentfor varying periodsof time; for the residualwinter residentsit was from 1 April to their departure;and for in- transitmigrants it was from the arrival date to the departuredate (period betweenthe earliestand latest records for the specieson the island). Density and biomassvalues for each of the three seasonalcommunities are the sumsof the valuesfor each memberspecies. Total density,biomass, and diversityfor the entire assemblageduring any given time intervalwere obtained by addingthe valuesfor the overlappingcommunities as suchoverlaps occurred. In this studythere was a brief overlapof the winteringand transientcommunities in late March, and a longerperiod of overlapof the breedingand transient communitiesin April and early May (Fig. 27). Rare species,stragglers, and vagrantscreate problemsin bird community studies,particularly in establishingspecies lists and richnessindices. In this study the problem was handled by operationallydefining two classes: membersand visitors. Speciesencountered in the submatureforests over periodsof more than 15 daysand at frequenciesgreater than 1 in 10 transectcounts were designated members. Those that were presentfor shorterperiods or were recordedat lower 58 ORNITHOLOGICAL MONOGRAPHS NO. 24

100 Nuthatch

eO-Cap.W.

8O PineW.eY-th'W' Saps.o eeLog.F. 8.8W.W.o eGd.D. eTan. ß Pew.

60-- C.E.Hum• eB.Yethez•en' Thr. Gr'qt.eH.ßWdp. o u.eGnotC. o• PalmW.o Banqt. •yr.W. o o oGr. Sp. oC.•.W. Redst OPor. ß o , oPr.W. W. Catbd. ß eMockb'd. Th-bV 40--

ßBah. Wdst.

o oNo.Wth. Com.Yth't. 20-- ß Bah. Sw.

R- W. Bkb'd. o I I o I I 0 20 40 60 80 IOO

Fmu•E 25. Two-dimensional ordination of bird species wintering on Grand Bahama Island showing habitat spacingas revealed by distribution overlaps through the 25 survey stands. Closed circles are permanent resident species;open circles are winter resident species. frequencies,and those seen only outside the hours scheduledfor transectsor outsidethe surveystands were termedvisitors.

SEASONAL CHANGES

Progressivemonthly changesin the populationstructure, density, and biomass are shown in Figure 28. The number of speciesincreased with the arrival of in-transitmigrants and summerresidents to a peak in early May, then dropped abruptly with the departureof the winter residentsand in-transit migrantsto 62% of the midwinter level at the end of May. Densityvalues for all speciestogether declined almost linearly from 1,260 per km2 on 1 Januaryto 648 per km2 by 1 June (Fig. 28B). The major factor in this reduction was the large difference between the number of winter invaders (3 monthsmean of 470 per km2) and summerresidents (mean of 18 per km2). The numberof permanentresidents in the forestsactually increased slightly due to local movementsof someZenaida Doves, Ground Doves, and Bahama Swallows from other habitatson the island, an influx that more than compensatedfor a partial emigrationof Mockingbirdsand Cuban Emerald Hummingbirdsfrom the forests. The reductionin winter residentsbetween January and March is attributablelargely to a progressivedecline in Palm Warblers,particularly in 1968. 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 59

GroundCover :.'1--•

•-•: 15-1] F ...... Emer(jentPines //'.. '.'.....' '. '•.' . . ' •.',.,////://...//•'J//://'//'//'//.:•,•_Pines _Z '<--C(3nopy Pines • '• ...... SubcanopyPines •''½••.a•.•.•.• '•'.• •-• • 3.--•. .Sh'u•s \ • '. . . ,':"•.'_."<.7/•//.•/// • o i1•o•,,o•'-•..... ,..... , ..... ,--•-•,•-•-•?.•,"',•,Ground • :, Cover X'•_ •' •'7.•. o.o.•o//Y o .o cove.,oo shes VEGETATION PROFILE PLANT COVER

TREE STRATUM SHRUB STRATUM GROUND COVER

FIGURE 26. Physiognomic structure and composition of the submature pine forests on Grand Bahama Island. Values are means for the three survey stands.

Whetherthis declinerepresents an actual departurefrom the island, a diffusion into other local habitats, or mortality is not known. The remaining winter residentsdeparted during April and early May. In-transit migrantswere a minor element,even at the height of migration, and contributedonly four individuals per km• (< 1%) to the 1 May total. The maximumrepresentation of in-transit birds on any singletransect count was on 5 May 1968 when they constituted11 of the 119 birds recorded (9%). Biomass values also declined (Fig. 28C) but to a somewhatlesser extent, becauseof the relativelylarge body size of the immigrantdoves moving in from neighboringhabitats. The breedingseason cornmunity.--Permanent resident species dominated the breedingseason community of the submaturepine forestswith 19 of the 22 members and 5 of the 6 visitors (Tables 13 and 14). Of the three summer resident members,one, the Bahama Swallow, winters on Grand Bahama in other habitats,moving into the pine foreststo nest in early April. The other two, the Gray Kingbird and the Black-whiskeredVireo, are true migrants,moving north from winter homesin the southernCaribbean and South America in early May. A few kingbirds wintered in coastal habitats on Grand Bahama in 1969. The singlesummer visitor species,the Cuban Nighthawk, migratesin from the southin late April. 60 ORNITHOLOGICAL MONOGRAPHS NO. 24

TABLE 13 SIZE, DENSITY, AND BIOMASSx OF THE RESIDENCYELEMENTS COMPOSING THE THREE SEASONAL PINE FOREST COMMUNITIES

All Wintering comm. Breeding comm. Transient comm. comm. Specs. Dens. Biom. Specs. Dens. Biom. Specs. Dens. Biom. Specs.

Members Permanent res.: 19 599 7354 19 630 9000 - - - 19 Summer res. - - 3 18 476 - - - 3 Winterres. 13 570 6•06 - - - 13 107 1167 13 In-transitmigr. - ...... 3 37 1 Total 32 1169 13760 22 648 9476 14 110 1204 36

Visitors Permanent res. 5 5 - 5 Summer res. - 1 - 1 Winter res. 8 - 3 8 In-transitmigr. - - 6 6 Total 13 6 9 20 Grand Total 45 28 23 56 • Size = number of species,density = birds per km 2, biomass---- g per krn2. -•The Yellow-throatedWarbler is classifiedhere as a permanent resident speciesbut its population density and biomassare divided accordingto the relative abundanceof the permanentresident and winter resident races.

In termsof densityand biomass,the breedingseason community was heavily dominatedby permanentresidents. Summer residents, including the locally migrantswallow, comprised only 3% of the individualsand 5% of the biomass. The winteringcommunity.--The 24 permanentresident species of the Grand Bahamapineland community (19 membersand 5 visitors) were joined during the winter by 21 migrantinvaders (WRs) from the north (13 membersand 8 visitors)to producean overallwintering community of 45 species(Tables 13 and 14). One permanentresident, the Yellow-throatedWarbler, has a migrant population(race) aswell as the residentpopulation, so the total of winterinvader specieswas 22 (14 membersplus 8 visitors). In terms of individualsand arian biomassthe winter residentscomposed almost half (49%) of the winteringcommtmity. Without the dominantPalm Warbler, however,they would have constituted less than onefifth (17%) of the total. Thesewinter migrants came from breedingranges in North Americastretching from the mixed pinelandsof the Gulf Coast to the tundra edge in northern Canada(Fig. 29). Seventeenof the speciesbreed in the northernconifer-hard- woodlife area as mappedby Aldrich (1963), and of these,11 breedexclusively in that area. Eleven (4 exclusivelyand 7 partly) make their summer homes in the easterndeciduous forest area betweenthe Piedmont of central Georgia and the Great Lakes, and only 3 (all partly) summerin the mixed pinelandof the southern coastal states. MacArthur (1959) has noted that it is the eastern deciduousforests where the migratoryhabit is moststrongly developed; very few southernspecies are migratory.Although wintering here in a pinelandhabitat, abouta third of the species,including the dominantPalm Warbler,are primarily birds of brushlands or deciduous forests in their northern homes. 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 61

TABLE 14 RESIDENCY STATUS, DENSITY, AND BIOMASS OF MEMBER AND VISITOR SPECIES IN THE THREE SEASONAL PINE FOREST COMMUIqITIES

Wintering Breeding Transient a community community community Res. Dens. Biom. Dens. Biom. Dens. Biom. Days status birds/km: g/km 2 birds/km x g/km 2 birds/km •ø g/km 2 pres. Members' Zenaida Dove P 4.0 600 8.7 1300 -- -- Ground Dove P 3.5 122 9.4 328 -- -- C. Emerald Hummingbird P 106.00 318 84.0 252 -- -- Hairy Woodpecker P 10.6 530 8.7 435 Yellow-bellied Sapsucker W 3.5 158 -- -- 0.5 22 20 Gray Kingbird S -- -- 4.6 276 -- -- LoggerheadFlycatcher P 4.2 202 7.2 345 -- -- Stolid Flycatcher P 2.2 49 0.5 11 -- -- Greater Antillean Pewee P 17.0 170 21.9 219 -- -- Bahama Swallow S• -- -- 11.0 165 -- -- Brown-headed Nuthatch P 21.0 105 21.5 108 -- -- Northern Mockingbird P 3.2 153 0.4 19 -- -- Catbird W 7.7 292 -- -- 3.9 112 41 Red-leggedThrush P 3.0 210 7.2 503 -- -- Blue-gray Gnatcatcher P 49.0 245 48.6 243 -- -- Thick-billed Vireo P 12.6 202 25.2 403 -- -- Black-whiskered Vireo S -- -- 2.2 35 -- -- Bahama Bananaquit P 87.0 870 61.8 618 -- -- Black-and-White Warbler W 11.1 111 -- -- 10.7 107 47 Parula Warbler W 1.2 8 -- -- 1.7 12 41 Cape May Warbler W 1.0 8 -- -- 3.3 37 37 Blacksthroated Blue Warbler W 0.7 6 -- -- 8.2 66 41 Yellow-rumped Warbler W 13.5 135 -- -- 2.2 22 15 Black-thr. Green Warb. W 4.9 39 -- -- 2.2 18 19 Yellow-throated Warbler (res. race) P• 15.0 150 17.1 171 -- -- Yellow-throated Warbler (migr. race) W 45.0 405 -- -- 5.0 45 10 Olive-cappedWarbler P 95.0 665 121.0 847 -- -- Pine Warbler P 40.5 486 42.5 510 -- -- Black-poll Warbler Tr .... 3.1 37 Prairie Warbler W 11.4 91 -- -- 4.2 34 20 Palm Warbler W 452.0 4970 -- -- 31.0 341 36 Ovenbird W 1.5 29 -- -- 7.0 133 44 Common Yellowthroat W 9.7 107 -- -- 7.3 80 44 Bahama Yellowthroat P 30.3 485 31.2 500 -- -- Redstart W 6.7 47 -- -- 19.7 138 55 Striped-headedTanager P 62.5 1440 79.4 1820 -- -- Black-facedGrassquit P 32.0 352 33.4 368 -- -- Visitors: White-crowned Pigeon P N 6 N -- Mourning Dove P N N -- Yellow-billed Cuckoo Tr -- -- N Smooth-billed Ani P N N -- Chuck-wills-widow W N -- -- Cuban Nighthawk S -- N -- Bahama Woodstar P N N -- House Wren Tr -- -- T Ruby-crowned Kinglet Tr -- -- N Yellow-throated Vireo Tr -- -- N Blue-headed Vireo W N -- -- Worm-eating Warbler W T 7 -- T Orange-crowned W N -- -- Warbler Golden-wing Warbler W T -- -- Magnolia Warbler W N -- T Kirtland Warbler W N -- -- 62 ORNITHOLOGICAL MONOGRAPHS NO. 24

TABLE 14 (CoNTnqUED)

Wintering Breeding Transient a community community community Res. Dens. Biota. Dens. Biom. Dens. Biota. Days status birds/km 2 g/kin g birds/km 2 g/kin • birds/kin a g/km• pres.

Northern Waterthrush W N -- T Hooded Warbler Tr -- -- T Indigo Bunting Tr -- -- T Greater Anfillean Bullfinch P N N --

• A member is a species that is present over a period of more than 15 days and observed more than once per 10 transect counts. • A visitor is a species that does not meet the member criteria. a The transient community includes all migratory species occurring only in transit plus all winter residents remaining after 31 March. The number of days when each transient species was present is shown in the last column. The density and biomass values for transients are averaged for the entire 60-day transient period. 4 The Bahama Swallow is a permanent resident on Grand Bahama Island but was absent from the submature forests during the winter. 5 The Yellow-throated Warbler has a resident, Bahama race and a migrant race invading the same habitat in winter. The two are generally distinguishable in the field. e Indicates that the species was observed nearby, but not in the transect stands. * T (trace) indicates that the species was observed in small numbers (<1 per 10 transects).

The dominanceof the northern element among the winter invaderswas more pronouncedin the measurementsof densityand biomassthan of species(Fig. 28). Approximately 86% of the individualswere from the northern forest area, 13% from the easternor centralarea, and only 1% from the southernarea. Nine-tenths of the birds in the northern element were Palm Warblers. The transientcommunity. All 13 of the winter resident membersand 3 of the 8 winter residentvisitors remained for varyinglengths of time after 1 April (Table 13) and thusbecame members (or visitors) of the transientcommunity. These 16 specieswere joined by 7 in-transit speciesto produce a transientcommunity totaling 23 species,variously present in the standsfor from 10-55 days of the artificiallydelimited 60-day period (seeTable 14, last column). The residual winter invaders accounted for 97% of the individuals and 97% of the biomassin the transient community. Yellow-rumped Warblers and the migrant race of the Yellow-throatedWarbler droppedoff during the first half of April, and the dominantPalm Warblersbetween 6 and 16 April. But severalof the winter speciesincreased markedly for short periods during April or May as they receivedreinforcements from in-transit migrants. Waves of migrants,char- acteristicof migration on the continent,were only weakly developedexcept on

Mic•ratory Status CateGories In transit migrants ".:.":'!?Y:';'•':':. Winter residents -• ::'::.:.::'::' Permanentresidents --•///////////////////////AIIIll[111111[111111111 • Summer residents Illllllllllllll',', • I dan. I Feb.I Mar. I Apr.I I•lay I Communities: Wi nte r 117"/"//'•,breed incj r[TTT• ,transient :'•

F•OURE27. Assigned temporal limits and residence category structure of the wintering, breeding,and transientbird communitiesof the Grand Bahama submaturepine forests. 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 63

14000-• 48- 1200"• 12000i'l"• 32-

C C'"' C'"' residents 4o-J F M A M J 'øøø-l• J F tM IA IM •J 'øøøø1• J F M A IM 1J A-NUMBER OF SPECIES- B-DENSITY- C-BIOMASS (Individualsper kma) (gin per kma)

FIGURE 28. Monthly changes in the size and composition (by residence categories) of the bird communities of Grand Bahama submature pinelands, 1968-1969. The number of speciespresent is recorded, and the density and biomass are estimated for the first of each month from January to June. For the permanent residents I used the wintering community 2-year average for density and biomassas given in Table 2 for the 1 January, 1 February and 1 March values, and the breeding community average for the 1 May and 1 June values; the 1 April value was interpolated. For the winter residents I took the density and biomass values for the wintering community and scaled them according to my records of day-to-day changes on the transect routes to give the 1 January to 1 April figures. The 1 May figures reflect the decline from the 1 April figures as indicated on the transect record sheets. For the summer residents, I applied the breeding community values as given in Table 14 directly for the 1 June figures, and scaled these figures down according to my records of arrival dates for the four speciesto give the 1 May figures. Density and biomass values for the in-transit migrants are those for the Blackpoll Warbler, the only completely in-transit species, plus increments above winter levelsin persistingwinter residentspecies. one occasion,5 May 1968, when birdsbelonging to winter residentand in-transit speciescomprised 63% of the total count.

IMPACT OF THE WINTER MIGRANT INVASION

The impact of a heavy annual invasionof migrants into the domain of a supposedlybalanced bird communityraises a number of interestingtheoretical questions,especially in situationswhere habitat conditionsand food resourcesare quite stablethrough the annualcycle. If the numberof breedingspecies present on an island reflectsa delicateequilibrium between colonization and extinction rates as proposedby MacArthur and Wilson (1967), what will be the effect of a doubling of the speciesnumbers each fall? And if breeding densities are regulatedby availablefood resourcesas arguedby Lack (1954) and many others, what will be the effectof doublingthe overalldensities at a time whenfood supplies have not increased.9 My Grand Bahamadata provideno answersto the first question,but note that while speciesrichness (members and visitors) in the pine forestsincreased from 28 in summerto 45 in winter (transientsignored), bird speciesdiversity (H') was actuallyslightly lower (2.39 vs. 2.62) in winter. This reduceddiversity is clearlyattributable to the relativelylow equitabilityof the winterresident element-- over 77% of the total populationbelonging to a single species. I know of no 64 ORNITHOLOGICAL MONOGRAPHS NO. 24

,x ,x 17 o (11) o

FIGURE 29. Geographic sourcesof bird speciesmigrating to or through the Grand Bahama pine forests. The width of the arrows at their bases roughly represents the number of speciesin each regional element. The actual number is given, and, in parentheses,the number that belongs exclusively to the element. Open arrows represent the winter resident speciesmigrating in from (a) the northern coniferous-hardwoodforest region, (b) the central deciduousforest region, and (c) the southern mixed pineland region of eastern North America. The cross-hatched arrow represents the summer resident speciesfrom Antillean islands to the south. The stippled arrow representsthe species recordedonly as in-transit springmigrants. published data on seasonalchanges in H' values for closely delineated (single habitat type) communitieselsewhere. The effect of an annualinflux and subsequentwithdrawal of hordesof migrant individualson the stabilityof southernbird communitieshas been consideredby Moreau (1972), Morel (1968), Fretwell (1972), and others. Winter invaders often form a strongminority elementin winteringcommunities, and on occasion, as in severalof the Grand Bahamahabitats (see Table 4) may constitutea substan- tial majority. In the submaturepine habitat on Grand Bahamathe winter influx 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 65

a 86

o o o 0 0 0 0

0 0 0 O

0 0 0 0 0 0 0 0 0 X X X X x x 1 X X X X

6 FIGURE30. Numerical strength (relative population density) of the regional elements of birds migratingto or throughthe Grand Bahamapinelands. The three winter migrant elements(matching those in Figure 29) are calculatedas percentagesof the total of winter resident birds. The values for the summer resident and in-transit elements are figures equated against the winter residentpercentage values. As in Figure 29, the in-transit element in this calculationexcludes residual winter residents,but includes estimatesof the in-transitpopulations of winter residentspecies (where the migration seasonlevel exceededthe winter level). essentiallydoubled the overall density (599 to 1,169 birds) and biomass (7,354-13,760g). Thepossibility that this seasonal increase was accompanied by a concurrentand comparableincrease in food supplywas checked by sampling foliage,bark, and ground insect communities at the threestudy sites in January and May of 1971. The resultsas presentedand discussedin Chapter8 were negativein all cases(see Fig. 31). TentativelyI concludethat food was not a limitingfactor for communitypopulation density on GrandBahama, a proposition I havedeveloped further in a comparisonof Floridaand Bahamabird communities Emlen (1978, in press)and a comparisonof pine tree gleanerson two Bahama islands(Emlen in prep.). 66 ORNITHOLOGICAL MONOGRAPHS NO. 24

1.5

ß 1.0 Ground 0.8 ß Cover Shrub 0.6 Folioge ß Pine Crown Upper Upper-outside 0.4 -- ß Pine Crown Low-outside

0.2 -- Pine Crown Top

0.01 0.1 1.0 I0.0 FOOD RESOURCE (INSECT UNITS/m S OF STANDING FOLIAGE)

• 4.0-

o 3.0- Trunk - High ß ßTrunk-Low • 2.0-

1.0 I I I I 0.01 0.1 1.0 I0.0 FOOD RESOURCE(INSECT UNITS/cm 2 OF BARK SURFACE)

o3 0.5- C o3 ßJanuary o 0.3-- ß May

FOOD RESOURCE (INSECT UNITS/kg OF STANDING PINE FOLIAGE) FIGURE 31. Ratio of avian insect gleaner biomass to food resource abundance in selected compartmentsof the Bahama pineland habitat. A, winter data for five foliage compartments; B, winter data for two tree trunk compartments; C, winter and spring data for one tree foliage compartment (low-outside).

7--SPATIAL DISTRIBUTION WITHIN THE PINE FOREST

Bird speciesdiffer in their distributionthrough the spacewithin a stand of vegetation. As describedin Chapter 2, two methodsof partitioning this space were used for plotting speciesdistributions within the pine forests of Grand Bahama Island: (1) a seriesof horizontallayers were defined extendingfrom 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 67

A B C D œ F G H I J K L M N 0 P Q

9 ! 6

•o me, '- • T 3 • • ..=2-_•.• •- o• ß •

R S T U V W X Y Z AA BB 25 205 64 929 21 61 107 167 234 241 83

Fmo•œ 32. Vertical distribution in the vegetation of the 28 species with 20 or more records. Proportional representationin each of six 1.52-m (5-ft.) layers is graphically indicated for each species. The bottom layer is subdividedinto a ground level (0-0.3 m) and upper level (0.3-1.52 m) in all except the vegetation profile. Sample sizes are given at the top of each profile diagram. An overall community profile and a vegetation profile are shownat the lower right. the ground surfaceto the tops of the trees, and (2) blocks of habitat space or compartmentswere delineatedwith referenceto prominentrecurring features in the forest structure such as individual shrubs, individual trees, etc.

VERTICAL DISTRIBUTION (LAYERS) Data on height above the ground as estimated at the moment of detection were variouslygrouped for tabulation and analysis. These vertical distribution recordsfor each of the 28 specieswith 20 or more data pointsare graphedas a seriesof profile diagramsin Figure 32. The profilesat the lower right show the vertical distributionof all bird records (all speciespooled), and in the unshaded diagram, the volume of standingvegetation. The step interval used in these diagramsis 1.52 m (5 feet) except at the base where the first step has been subdivided into a 0.3-m (1-ft.) and a 1.2-m (4-ft.) unit; the effect of this subdivisionis to producea stemlikebase on severalof the diagrams. Speciesshow a wide variety of heightdistributions and varyingdegrees of specializationalong the gradient. Noteworthy is the strong representationof many speciesin the sparselyfoliaged layer betweenthe shrub tops and the bottoms of the tree crowns. This layer shows as a narrow "waistline" in the profile of standing vegetationvolumes. Vegetativestructure and avian utilizationat live levels.-•Changesin vegetation structurewith height are summarizedby groupingthe data into 5 layers of equal 68 ORNITHOLOGICAL MONOGRAPHS NO. 24

TABLE 15 VERTICAL DISTRIBUTION OF BIRD POPULATIONS IN THE WINTERING COMMUNITY OF THE PINE FOREST •

Five Layers (Height in meters) 0-3 3-6 6-9 9-12 12-15 Total

Structural elements Needles ++ ++ ++ Leaves Twigs ++ ++ ++ ++ Branches + ++ + + Vertical surfaces Horizontal surfaces + Total space(m 3) per ha 30,000 30,000 30,000 30,000 30,000 150,000 Volume of standing vegetation(m a) per ha 5,625 700 5,500 5,500 500 17,825 No. of bird species 36 35 31 24 22 38 Heavily representedspecies (20%) 11 4 1 0 3 19 Well-representedspecies (20%) 24 21 11 3 7 38 Total densityper km2 369.3 286.3 248.3 134.6 113.3 1151.8 (32%) (2555) (2255) (12%) (10%) Total bird densityper million ms of space 123.1 95.3 83.8 44.9 37.8 76.8 Total bird densityper million ms of standingvegetation 657 4090 452 245 2266 646 Bird speciesdiversity (H') 2.910 2.743 2.571 2.177 2.246 x Characteristics of the vegetation and bird populations as shown for five equally thick (3-m) layers of space between the ground surface and the tree tops. depth(3 m) betweenthe groundsurface and the tree topsin Table 15 and Figure 33. The lowest layer containsessentially all of the ground cover and shrubs. The secondencompasses the relatively open spacebetween the shrubs•and the tree crowns;the third, the lower half of the pine canopy;the fourth, the upper half of the pine canopy;and the fifth, the topsof the taller crownsof the canopy plus the scatteredemergent crowns. Emergent crowns above the 15-m level are omittedin this analysis. While the amountof spacein each of the five layersis equal, the amount and type of substrateavailable for bird use differs considerably.In the basal layer, 5,625 ms of the 30,000 m3 of space(19%) was occupiedby relativelydense and diversevegetation, dominated by leavesand twigs over a rather denselymatted, heavily pitted ground surface (Table 15). Perchesand escapecover were well distributedfor smallbirds, and food was providedby seed-producinggrasses and forbs as well as nectar-producingblossoms and many intricately divided insect- supportingforaging surfaces. Seed-eatingdoves and grassquits,low-foraging insectivorouscatbirds, vireos, ovenbirds and yellowthroats,and blossom-frequenting hummingbirdsand bananaquitswere bestrepresented in this layer, and with others comprisedthe largestand richestelement in the forest. Eleven species,designated "heavilyrepresented" in Table 15, had more than half of their numbersrecorded in this layer. The second layer (3-6 m) contained only about % as much standing vegetationper unit of spaceas the basal layer. Large horizontalbranches and verticaltrunks replaced the leavesand fine twigswith firm lookoutperches and 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 69

0 I • 0 40 80 I•0 Bird Sp.Diversity (H') Density (Birds/million m$)

A. B. C. D. FIaORE 33. Number of species (S), species diversity (H •) and density per million m3 of space (Dr) of the winter bird community of submature pine forests compared with the volume of standing vegetation in each of five equal horizontal divisions of space (layers) between the ground surface and the tree tops. (Emergent crowns above 15 m are not included.) broad foragingsurfaces. Visibility was clear over relativelylong distanceswhile escapecover was spatiallyremoved. Despitethe sparsityand low diversityof the foliage,birds were numerous and variedin this subcanopylayer, whichwas second only to the basallayer in numberof species,bird speciesdiversity, and density (Fig. 33). Prominentamong the well-representedspecies were the insect-chasing GreaterAntillean Pewee (51% of all recordsfall in this layer) and StolidFly- catcher(34%), the branch-perchingGround Dove (30%), the ground-pouncing Mockingbird(50%) and Red-leggedThrush (31%), the open-trunk-gleaning Black-and-whiteWarbler (45%), and the subcanopybranch and low pine foliage gleaners:gnatcatcher (32%), ParulaWarbler (73%), Cape May Warbler (40%), Yellow-rumpedWarbler (39%), Black-throatedGreen Warbler (72%), Black- throated Blue Warbler (32%), Black-poll Warbler (32%), Prairie Warbler (37%), Palm Warbler (28%), redstart(39%), bananaquit(29%), and Striped- headed Tanager (34%). The third and fourth layers encompassedthe pine crownsof the canopy. Vegetation,occupying about 18% of the total spacein eachof theselayers (Table 15), consistedalmost entirely of pine needlesand the subhorizontaltwigs and branchessupporting them. Foliagedistribution was patchy, each crown constituting a roughly circular patch with needlesdominating peripherally and branches dominatingcentrally. Of the two layers,the lower had thinnerfoliage (sparser and more scatteredneedle clusters) and heavier branches. It thus formed a transition betweenthe open subcanopyof the secondlayer and the relativelydense foliage of the fourth. The numberof species,diversity, and densityof birds wasless in thesethan in the lowerlayers, and appreciablyless in the fourthlayer (upperhalf of the crowns)than the third layer (lower half). Two species,the Hairy Woodpecker(50%) and the Yellow-throatedWarbler (36%), reached theirmaximum densities in the thirdlayer, and representation was above average for 7 more: StolidFlycatcher (33%), Black-and-whiteWarbler (21%), Olive- cappedWarbler (25%), Black-pollWarbler (36%), Pine Warbler (29%), Palm Warbler(27%), and redstart(38%). Only 3 speciesreached above average densities in the fourth layer; the Brown-headedNuthatch with 25%, the Olive- cappedWarbler with 35%, and the Pine Warbler with 28%. The fifth and highest layer closely resembled the fourth in structure and 70 ORNITHOLOGICAL MONOGRAPHS NO. 24 texture, except that the patcheswere more widely scatteredand collectively provided less than one-tenthas much standingfoliage. Also present were oc- casionaldead branchesand stubsproviding high lookoutperches. Two species, bothof thembirds that fly out for aerialinsects from openperches, reached their highestnumbers in the tops of the tall canopyand emergenttrees: the Bahama Swallow(100%) and the Gray Kingbird (56%). Also well representedwere the Loggerhead Flycatcher (29%) and three high foliage gleaners: the Brown- headedNuthatch (53%), the Olive-cappedWarbler (29%), and the Pine Warbler (22%). The overall picture of bird distributionwas one of gradual and fairly even declinein speciesnumber, species diversity, and total populationdensity upward throughthe five layers (Fig. 33). Vegetationdensity per cubic unit of space, representingthe quantityof substratein a layer for perching,foraging, etc., varied considerablyand irregularlyhowever, with highs in the first and fourth layers 8-10 times thosein the secondand fifth. Reflectingthese irregularities bird- substrateratios varied greatly through the layers(Table 15) leavingthe impression that, in this context at least, bird numbers and densitiesare much more closely relatedto spaceper se than to the amountof substrateavailable for perchingand foraging. Populationstructure in the five layers.--Becauseof great differencesbetween speciesin overallpopulation density, species with a low percentagerepresentation in a given layer often outnumberedwell-represented species in that layer. For example, the dominant (most numerous) member of the wintering community in the basal layer was the Palm Warbler with 241 individuals recorded per hectare, yet this constitutedonly 25% of its numbers, a lower basal layer representationthan that shownby 20 lessabundant species. The numericaldominance structure of the wintering community of the pinelands in each of the five 3-m layers is presentedin Figure 34. The height of each segmentin the columnsshows the relativefrequency of occurrenceof eachspecies in the layer. The distributionsof abundancefollow expectedlogarithmic declines. Of the total bird populationin each layer, 80% belongsto between25 and 29% of the speciespresent. The Palm Warbler is the dominantspecies in the first three layers and is in secondposition in the other two where it falls behind the Olive- cappedWarbler. The divisionof habitatspace into 5 layersalso provides a meansfor analyzing the distributionof the impact of seasonalimmigration within the community (Figure 35). The influx of migrantswas heaviestin the lower tree canopyand subcanopylayers (3-9 m) where winter residentsoutnumbered permanent residentsabout 4:3 during the winter months;it was lightestin the high tree layer whereWRs were only half as numerousas PRs. Summerresidents consti- tuteda very minorelement in all layersexcept the uppercanopy where they were 5.4% as numerousas the permanentresidents.

COMPARTMENT DISTRIBUTION

The vegetationcompartments used in this study (see Fig. 1) providea series of divisionsand subdivisionsof the habitatspace, unequal in volumebut directly 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 71

N ß 1395 N = 1056 N= 862 N=486 N=417 36 SPECIES :55 SPECIES 51 SPECIES 24 SPECIES 22 SPECIES I00 I 26 Sps.I ! 24,3%,I 80

0 O-$rn $-6m 6-9rn 9-i2m 1•-15m (UPPER• {EMERGENTS) 6 SHRUBS/ CROWNS/

FIGURE 34. Relative abundance (numerical dominance) of the bird speciescomprising the wintering community in each of five equal (3-m) layers between the ground surface and the tree tops. comparableas units of living spacefor bird populations. As units they are more natural than the equal horizontal layers examined in the precedingpages and combineaspects of substrateuse and preferencenot possiblewith the heterogen- ously structured,horizontal layers. The compartments,on the other hand, are not amenableto the serial arrangementthat made it possibleto trace within-habitat trendsin speciesdistribution and communitystructure. Someof the analysesthat

PR Wl PR Wl PR Wl PR WI PR Wl PR Wl O-:Sm :5-6m 6-9m 9-12m 12-15m TOTAL

SHRUBS,/ •CANOPY/ •CANOPY/ • CROWNSJ

FIGURE 35. Numbers of winter residents compared with permanent residents in the five horizontal layers of the standardpineland habitat in winter. Lower portions of shaded columns show numbers of Palm Warblers. 72 ORNITHOLOGICAL MONOGRAPHS NO. 24

TABLE 16 VOLUME AND STRUCTURAL CHARACTERISTICS OF THE MAJOR HABITAT COMPARTMENTS IN THE PINE FORESTSx

Tree crowns Shrubs Trunk Ground Air space2 Shell Core Shell Core cover space Total Total space (1000 mYha) 0.39 8.20 4.00 2.83 1.42 1.38 152 170 (%) (0.2) (4.8) (2.4) (1.7) (0.8) (0.8) (89.4) Structural elements Needles 0 +++ + 0 0 0 0 Leaves 0 0 0 +++ + ++ 0 Twigs 0 +++ + +++ ++ ++ 0 Branches + + +++ 0 ++ + 0 Vertical surfaces +++ 0 0 0 ñ 0 0 Horizontal surfaces 0 0 0 0 0 4- 0 x Plus signs give a rough appraisal of the representation of each structural element from light(+) to heavy(+++). • 2 Volume of cylinaers 0.30 m in diameter incorporating each trunk. follow resemblethose used in the discussionof the layers, others will explore aspectsof the dynamicsand regulationof habitat distribution. Physicalcharacteristics o! the compartments.--Theamount of spaceincluded within eachmajor habitatcompartment is presentedin Table 16 togetherwith an appraisalof the structuralelements or perchingsubstrates in each. The trunk compartment,divided at the base of the crown into upper and lower subcompartmentsfor some analyses,comprises the compositeof spaces around each of the tree trunks and stumpsin the forest. Arbitrarily defined as the spacewithin imaginarycylinders 0.30 m in diametersurrounding each trunk, the total volume is only 390 ms per ha2, or 0.2% of the 170,000 m• of space presentbetween the groundsurface and the canopytop. Vertical wood and bark surfacesare the dominant structural elementsin this compartment,but the basal few centimetersof branchesalso contributean importantperching and foraging substratefor certain species. Pine tree crowns,potentially divisible from the field data into 12 subcompart- ments,comprise the secondmajor compartment.The upper and outer portionsor shellsof the crownscontain a ratherdense matrix of needlesand twigs;the inner portionsor coresare dominatedby horizontaland subhorizontalbranches with smallerquantities of needlesand twigs(Table 16). The volumeof spaceoccupied by crownshells and coresis 8,200 m2 and 4,000 m• per ha2 respectively,or 4.8 and 2.3% of the total habitat space. Broadleafedtrees, as alreadynoted, were smalland comprisedonly 0.1% of the tree stratum in the standardforests. Becauseof their low representationand smallsize, they are incorporated into the shrub compartment in thepresent analysis. Shrubs,potentially divisible into 15 subcompartmentson the basisof field data, constitutethe third major compartment.The upper and outer portionsof shrubs are characterizedby manyleaves and twigsand collectivelyoccupy 2,830 m• per ha. The loweror basalportions are dominatedby branchesand heavytwigs and occupy1,420 m• per ha. Shrubsare, of course,close to the groundsurface and 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 73

thereforemore available than trees as escapeand perching cover for ground- foragingbirds. The thin layer of spaceover the groundsurface occupied by trailing vines and herbaceousfoliage comprisesthe groundcover compartment.Varying in depth, it is consideredto incorporatea total of 1,380 m3 per ha, thus being the smallest compartmentexcept for the tree trunks. The structure is characterizedby a reticulumof subhorizontaltwigs and slenderstems interspersed with small leaves and grassblades over a highlyirregular, deeply pitted limestone base. The airspacebetween and surroundingthese 6 vegetation-filledcompartments occupied152,000 m3 per ha or 89.4% of the total spacebelow the tree tops. Being without restingsurfaces, the utility of this spaceis limited to passagebetween perchesand brief salliesfor capturingflying insects. Bahama Swallowsand Nighthawks,the only aerial screenersin the pinelandscommunity, do nearlyall of their foragingabove the tree tops and henceoutside any of the compartments in this system. Becauseof its generallytransitory significance as substrate,and particularlybecause of the biasesintroduced by the high detectabilityof birds when in it, the airspaceis ignored in most of the compartmentdistribution analysesin this study. Bird speciesdistribution through the compartments.--Distributionthrough the compartmentsof the submaturepinelands is shownfor the 30 specieswith 20 or more recordsin Figure 36. From the plethora of positionaldata on the field record sheets,10 compartmentswere selectedfor these representations,the 7 structurallycharacterized in Table 16 with the first 3 further subdividedon the basisof position.Diagrams in the top row of the figureidentify the compartments and give the pooled data for speciesgroups, including one for all species combinedand one for foliagevolumes in eachcompartment. The figuresin each box are frequencyvalues, expressed as percentsof the total sampleavailable for the speciesor group. Samplesizes are given to the right of each diagram, followedby an indexof distributionalamplitude (H •) and an indexof compartment preference(observed frequency in the favored compartmentdivided by expectedfrequency, assuming equal distribution through the four major compartment categories,tree crowns,shrubs, trunks, and groundcover). With 4 major compartmentsthis latter index will rangefrom 1.0 for a speciesnot deviating from an evendistribution to 4.0 for a speciesfound entirely in a singlecompartment. Amongthese 30 speciesthe Mockingbirdshowed the mostdiverse compartment distribution(H' = 0.91) and the Olive-cappedWarbler the least(H' = 0.07). Of the 30 species4 favoredthe trunk compartment,12 the tree crowns,11 the shrubs, and 3 the groundcover. The 4 trunk specialistswith their index of preference were the Yellow-belliedSapsucker (3.4), Hairy Woodpecker(3.0), Black-and- whiteWarbler (2.7), andBrown-headed Nuthatch (2.1). Tree crownspecialists includedthe Olive-cappedWarbler (3.8), LoggerheadFlycatcher (3.7), Pine Warbler(3.4), andpewee (3.2). The principalshrub species in orderof specialization were the Common Yellowthroat (3.4), catbird (3.1), Bahama Yellow- throat (3.0), and GreaterAntillean Bullfinch (3.0), and the principalground compartmentspecialists were the grassquit(3.4) and Ovenbird(3.2). 74 ORNITHOLOGICAL MONOGRAPHS NO. 24

INDEX TO COMPARTMENT TOTAL BIRDS TOTAL elRDS TOTAL BIRDS COMPARTMENTS VOLUMES (ALL SPECIES) (PERM. RES.) (WINTER RES.)

.85 .67 .42 .55 16

Grid. Dove C.E. Hum. Y b. Saps. Hairy Wdp. Logg. Flyc. GA. Pewee

79 24 76 21 .55 .91 .51 .71

Brn.- hd. Nut. Mocker Catbird Red- L. Thrush Gnotcatcher Th. b. Vireo

Bk.8•W.Warb. Parula Warb. Cape May W. Y-rumped W. YelI.Tht.Warb, Dlive Cap.W.

Bk. Poll Warb. Pine Warbler Prairie WOrb. Palm Warbler Ovenbird Redstort

.29 .5•

Corn. Yelltht. Bah. Yelltht. Bananaquit St[ hd. Tan. Grassquit G.A. Bullfinch F[ouaE 36. Distribution of detection positionsthrough ten vegetationcompartments for the 30 specieswith 20 or more position records. All routes and seasonsare pooled. Figures in the diagramsare percent values. Sample size, an index of distributionalamplitude (H'/H' max.), and an index of compartmentpreference (4d/D) amongthe four major compartment categories are shown at the right of each diagram (see text for further explanation). Compartmentabbreviations: TrU •_ upper trunk, TrL = lower trunk, Top • crown top, UI = upper inside, UO = upper outside, LI = lower inside, LO • lower outside, ShU •- upper shrub, ShL •_ lower shrub, Gd = ground surface, Cn = crown. 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 75

WINTERING COMMUNITY BREEDING SEASON COMMUNITY

% TRUNKS CROWNS SHRUBS GD.COV TRUNKS CROWNS SHRUBS GD. COV

i[_•.•.P c• 50 B[•.•.•.•. 25 c[•_•.•St I O.Cap. D IWarbler N: 233 1174 1359 330 74 632 246 46

FmtJRE 37. Relative abundance (dominance) of the bird speciesof the wintering and breeding seasoncommunities in the four major habitat compartmentsof the standardforests.

Winter residentspecies as a group were lesswell representedthan permanent resident birds in the trunk (3.3% vs. 8.8%) and shrub (23.3% vs. 29.0%) compartments,and better representedin the ground cover (18.9% vs. 4.8%). Thesedifferences were largelydue to the tree and ground-foragingproclivities of the stronglydominant Palm Warblers amongthe winter invaders. Distributionwithin the tree crowns varied greatly from speciesto species. Among the 18 specieswith 10 or more recordsin the crowns, 16 were more frequentlyencountered in the core compartmentsthan in the outer shell, even thoughthe corecontained less than half as muchspace. Of the 16 core species,4, the nuthatch, thrush, vireo, and Yellow-throated Warbler were most numerous in the upper portion and 12, the Hairy Woodpecker,Loggerhead Flycatcher, pewee, Black-and-whiteWarbler, Yellow-rumpedWarbler, Cape May Warbler, Palm Warbler, Prairie Warbler, Black-poll Warbler, redstart,bananaquit, and tanager in the lower portion. The two speciesbest represented in the shellspace were the Olive-cappedand Pine Warblers, both concentratingin the upper-outersub- compartment. Populationstructure in the compartments.--Thenumerically dominant species of the breedingseason and winter communitiesand their percentrepresentation in each of the four major compartmentsare shownin Figure 37. These dominance rankingsdiffer considerablyfrom the specializationrankings in the same compartments(see Figure 36) becauseof great differencesin the overall densities of the species. In the wintering community, Palm Warblers dominated in the crown, shrub, and groundcover compartments,and the Brown-headedNuthatch in the trunk compartment.Among the four dominantspecies in eachcompartment three were permanentresidents. In the breedingseason community, there was less overlap betweencompartments, each of the four having different prime dominants. Demographicdata for the four major compartmentsare presentedin Table 17. In the breedingseason community, the tree crownshad the largest number of species,the highestdiversity value (H'), and the highestbird densityper areal unit. Winter residentsincluded 11 speciesusing the shrubcompartment against nine in the crowns,and these,superimposed on the permanentresidents, shifted the balanceof speciesnumber and diversityfrom the crownto the shrubcompartment duringthe winter months. The numberof individuals(density) amongthe 76 ORNITHOLOGICAL MONOGRAPHS NO. 24 winter residents,however, was strongestin the tree-crowncompartment where they roughlyequalled the residentpopulation. In the shrubs,the winter residents nearlyequalled the numberof permanentresident shrub birds persistingfrom the summer,but a seasonalmovement by certain permanentresident species down from the tree crowns raised the total winter densityin the shrubsto more than double that of the breedingseason. Densitiesexpressed as birds per m• of compartmentspace were highestin the trunk compartmentand lowestin the groundcover (last rows in Table 17). The reversal of relative rank between trunk and crown compartmentsfrom that recordedfor the areal densitymeasurement is, of course,due to the much smaller spaceoccupied by the trunk compartment.These volumetric density values will be used in considerationsof substrateselection and crowding dynamicsin the next chapter. Seasonalchanges in spatial distributions.--Quantitativedistributional analyses suchas thosepresented above have limited utility beyondpure descriptionexcept as they can be usedto measuredifferences between localities, habitats, years, or seasons.In this sectionI attempta comparisonof the compartmentdistributions of the permanentresident species of the Grand Bahamapine forestsbefore and after the winter migrantsdeparted for their northern homes. The pertinent data for eachspecies in the four majorcompartments are presentedin Table 18. In the trunk compartmentthe departureof three migrant species,the Yellow- bellied Sapsucker,Black-and-white Warbler, and Yellow-throatedWarbler (mi- grantrace) with a combineddensity equal to about41% of the winteringcom- partmenttotal, triggered no majorshifts among the remainingpermanent residents. There was, however, a slight increasein trunk foraging by severalprimarily foliage-relatedspecies to effecta smallrise in the total numberof speciesand speciesdiversity (Table 17). Two summerimmigrants, the Gray Kingbirdand the BahamaSwallow spent much time perchedhigh on stubsnear the top of the tree canopy;although assigned to the trunkcompartment, these birds had little or no interactionwith other trunk compartmentspecies. Amongthe permanentresidents frequenting the tree trunk compartment,the two dominantspecies, the Hairy Woodpeckerand the Brown-headedNuthatch, showedno appreciablechange in theirpreference for trunksbetween winter and summer(Pref. 2.92-2.94 and 2.11-2.20 respectively).I haveno gooddata on changesin the compartmentdistribution of the thirdpermanent resident species of thetrunk compartment, the Yellow-throated Warbler, since, as alreadynoted, the permanentresident population was supplemented in winter by an invasion of migrantsnot alwaysdistinguishable in the field from birds of theresident race. My noteson positively identified birds indicate that the resident birds were more stronglyoriented to the trunkcompartment than the migrants.This impression issupported by thehigher preference for trunks (1.01) shownby the breeding com- munitybirds (all PRs) thanthe mixed wintering community (0.61). In the treecrown compartment the departureof 10 species(1, an interhabitat shifton the island),and the arrivalof 1 summerresident produced a net decrease in speciesnumber from 29 to 20, a decreasein individualsper km2 from 644 to 395, and, becauseof a more equitabledistribution of numbers,an increasein 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 77

TABLE 17 DEMOGRAPHIC CHARACTERISTICS OF THE BIRD POPULATIONS IN THE 4 MAJOR COMPARTMENTS OF THE PINE FOREST

Wintering community Trunks Crowns Shrubs Ground cover Total Number of specied 10 26 27 17 31 Permanent residents 8 18 17 12 20 Winter residents 3 9 11 5 12 Summer residents -- -- Strongly represented species (•> 50%) 4 9 14 2 29 Permanent residents 2 6 8 2 -- Winter residents 2 3 6 0 -- Summer residents -- -- Well-representedspecies (•> 20%) 4 18 22 7 31 Permanent residents 2 13 12 6 -- Winter residents 2 5 10 1 -- Summer residents -- -- Bird Diversity (H •) 1.767 1.931 2.285 2.093 2.389 Permanent residents 1.458 2.193 2.034 1.972 2.404 Winter residents .473 .278 .952 .645 .998 Summer residents -- -- Density-1 (#/km 2) 45.1 644.3 402.1 51.6 1143.1 Permanent residents 31.5 324.1 245.3 29.5 630.4 Winter residents 13.6 320.2 156.8 22.1 512.7 Summer residents -- -- Density-2 ( #/million ms) 117.1 52.8 94.6 37.5 302.0 Permanent residents 81.9 26.6 57.7 22.3 188.5 Winter residents 35.2 26.2 36.9 15.2 113.5 Summer residents -- --

Breeding season community

Trunks Crowns Shrubs Ground cover Total Number of species• 13 20 17 12 23 Permanent residents 12 19 15 12 21 Winter residents -- -- Summer residents 1 1 2 0 2 Strongly representedspecies (•> 50%) 4 9 3 0 16 Permanent residents 4 8 2 0 -- Winter residents -- -- Summer residents 0 1 1 0 -- Well-represented species (•> 20%) 6 19 10 6 23 Permanent residents 5 18 9 6 -- Winter residents -- -- Summer residents 1 1 1 0 -- Bird diversity (H') 1.900 2.479 2.299 2.110 2.617 Permanent residents 1.827 2.312 2.141 2.110 2.596 Winter residents -- -- Summer residents ------.507 Density 1 (#/km:) 43.7 395.2 169.5 39.5 647.9 Permanent residents 42.7 392.6 166.8 39.5 641.6 Winter residents -- -- Summer residents 1.0 2.6 2.7 0 6.3 Density-2 ( #/million m:•) 114.5 32.4 39.9 30.5 217.3 Permanent residents 111.9 32.3 39.2 30.5 194.9 Winter residents -- -- Summer residents 2.6 0.2 0.7 0.0 22.4

Species recorded on more than 5 occasions. 78 ORNITHOLOGICAL MONOGRAPHS NO. 24

TABLE 18 COMPARTMENTDISTRIBUTIONS AND PREFERENCES (UNDERLINED) OF MEMBERS OF THE WINTERING AND BREEDING COMMUNITIES 1

Winter community (before migrant exodus) Trunks Crowns Shrubs Gd. Cover N PreferenceAmplitude Permanent residents Zenaida Dove - 1.3 - 2.7 6 2.70 .46 Ground Dove - 0.4 1.4 1.8 10 2.00 .69 Cuban Emerald Hummingbird 2.1 32.9 62.5 5.3 165 2.43 .64 Hairy Woodpecker 7.8 1.4 1.5 - 73 2.92 .56 Stolid Flycatcher - 1.3 0.9 - 5 2.36 .49 Loggerhead Flycatcher 0.4 3.8 - - 24 3.62 .22 Greater Antillean Pewee - 15.0 1.2 0.5 40 3.59 .28 Brown-headed Nuthatch 11.1 9.9 - - 40 2.11 .50 Bahama Mockingbird 0.3 0.9 1.3 0.6 26 1.68 .91 Red-legged Thrush - 0.4 1.3 1.3 7 1.73 .72 Blue-gray Gnatcatcher 0.4 17.6 30.4 0.5 130 2.49 .58 Thick-billed Vireo - 1.4 11.2 - 37 3.56 .25 Yellow-throated Warbler 2 4.3 12.5 0.3 - 52 3.07 .50 Olive-capped Warbler - 94.1 1.0 - 293 3.96 .04 Pine Warbler - 37.3 2.0 0.4 132 3.76 .18 Bahama Yellowthroat 0.3 2.1 20.0 7.6 71 2.67 .61 Bananaquit - 30.5 53.9 1.7 146 2.50 .53 Stripe-headedTanager - 25.6 36.9 - 103 2.36 .49 Grassquit - 2.5 15.0 7.1 41 2.44 .64 Greater Antillean Bullfinch - 2.1 8.0 5.9 16 2.00 .70 Total 26.7 293.0 249.3 35.4 1510 2.06 .73 Winter Residents Yellow-bellied Sapsucker 2.5 0.4 0.7 - 20 2.78 .59 Gray Catbird - 0.7 5.9 1.2 75 3.03 .52 Black-and-White Warbler 11.1 - - - 11 4.00 .00 Yellow-throated Warbler 4.9 33.3 4.5 - 93 3.12 .49 Prairie Warbler - 4.5 6.8 0.2 52 2.37 .54 Palm Warbler - 302.8 122.0 18.1 815 2.73 .54 Yellow-rumped Warbler - 7.7 5.7 0.1 185 2.28 .52 Other foliage warblers - 1.9 4.9 0 29 2.88 .50 Common Yellowthroat - 0.3 6.9 2.5 31 2.85 .50 Redstart - 3.8 2.8 - 36 2.30 .49 Total 18.5 355.4 153.6 22.1 1347 2.50 .64 Grand Total 45.1 648.4 402.9 57.5 2587 2.25 .69

Breeding season community (after migrant exodus) Trunks Crowns Shrubs Gd. Cover N Preference Amplitude

Permanent residents Zenaida Dove - 3.8 .9 3.9 9 1.81 .69 Ground Dove 0.5 5.5 .5 3.0 20 2.32 .70 Cuban Emerald Hummingbird 0.8 39.5 36.1 5.0 66 1.94 .66 Hairy Woodpecker 6.4 2.3 - - 26 2.94 .42 Stolid Flycatcher 0.1 0.4 - - 8 3.20 .36 Loggerhead Flycatcher - 7.2 - - 7 4.00 .00 Greater Antillean Pewee - 11.0 5.7 1.5 46 2.42 .63 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 79

TABLE 18 (CONTINUED)

Breeding season community (after migrant exodus) Trunks Crowns Shrubs Gd. Cover N Preference Amplitude

Brown-headed Nuthatch 11.8 9.7 - - 38 2.20 .50 Bahama Mockingbird .... 0 - .00 Red-leggedThrush - 4.0 0.9 2.2 16 2.25 .68 Blue-gray Gnatcatcher 1.0 24.8 22.4 0.5 122 2.04 .60 Thick-billed Vireo - 9.1 16.1 - 45 2.56 .47 Yellow-throated Warbler 2 4.3 12.5 0.3 - 52 2.92 .47 Olive-cappedWarbler 1.2 115.0 1.2 2.4 196 3.84 .15 Pine Warbler 1.3 37.8 1.3 1.3 75 3.63 .22 Bahama Yellowthroat - 6.6 24.2 8.2 24 2.48 .66 Bananaquit - 41.4 19.8 - 74 2.71 .45 Stripe-headedTanager 2.4 52.4 23.8 0.8 112 2.64 .57 Grassquit - 9.0 12.0 12.0 22 1.45 .79 Greater Antillean Bullfinch - 0.5 0.9 1.1 19 1.76 .76 Total 29.8 392.5 166.1 41.9 977 2.48 .70

Summer residents Gray Kingbird 1.0 2.6 0.5 - 2.54 .64 Bahama Swallow • 12.9 - - - 4.00 - Black-whiskered Swallow - - 2.2 - 4.00 - Total 13.9 2.6 2.7 - 2.90 .56 Grand Total 43.7 395.1 168.8 41.9 2.43 .73

x Values are birds per kma. Sample sizes, compartment preference (4d/D) and indices of distributional amplitude (Ht/H• max.) are shown in the last three columns. 2 The winter community values for the permanent resident Yellow-throated Warbler are arbitrarily treated here as direct repetitions of the breeding season community values and added to the winter resident values. a The Bahama Swallow is considered a summer invader for this tabulation. bird speciesdiversity from 1.93 to 2.48 (Table 17). With thesechanges in communitystructure the amountof permanentresident activity in the crownsapparently increasedappreciably in 9 species,decreased appreciably in 3, dropped out in 1, and held fairly constantin 6. The basis for these changescould have been intercompartmentshifts, interhabitat shifts, movementsto or from the island, mortality, or errors in the censusoperations. Unfortunately my data are inade- quate to providemore than crude guesseson what was involvedin the individual cases. In specieswhere increasesor decreasesin activity in the crown department are clearly matched by reciprocal changesin other compartmentsin Table 18 I interpret the record as indicatingan intercompartmentshift. Where informal springobservations in other habitatssuggest reciprocal seasonal changes with those recordedin the pine forestsI suspectan interhabitatshift. On this basisI suspect that there was a partial shift up from the shrub and ground cover compartment of the pine forest in the Cuban Emerald Hummingbird, Red-legged Thrush, Blue-gray Gnatcatcher,Bahama Yellowthroat, bananaquit,Striped-headed Tan- ager, and grassquit,and a partial shift down into the shrubcompartment by the Greater Antillean Pewee. I also suspecta movementinto the pine habitat from more open situationsby the Ground Dove and Bahama Swallow and a movement out to the densecoppets by the Greater Antillean Bullfinch. In the shrub compartmentthe springexodus involved 157 birds of 11 species 80 ORNITHOLOGICAL MONOGRAPHS NO. 24 from a winter communityof 402 birds and 27 species;2.7 birds of 2 species enteredas summerimmigrants. Concurrent with this reductionin overall density, the activity of permanentresidents in the compartmentdeclined from 245 tc 167. Largestdecreases were by the Cuban Emerald Hummingbird,bananaquit, Striped-headedTanager, and Greater Antillean Bullfinch, attributedprimarily to shiftsof activity up into the tree crown compartment. In the groundcover compartment the departureof 5 specieswith 22 individuals per km: left a communityof 12 permanentresident species with 30 individuals. Thesepermanent residents increased to 40 birds largely as a result of increased groundactivity by severalarboreal warblers and shrub-inhabitinggrassquits. These data are basedon a singleseason's observations and in many caseson smallsamples. Insofar as they are representative,however, they indicatetrends of considrableinterest to studentsof communitydynamics, trends that may be summarizedby reducingthe confusingarray of valuesto a commondenominator. Startingwith 100 birds in the winteringcommunity of the pine forests,spring migrationremoved about 45 winter residentsand addedone summerresident for a net reductionto 56 birds. This breedingcommunity consisted of 29 crown- frequentingbirds reduced 50% from 57, 22 shrubbirds reduced 39% from 35, 2.8 trunk birds reduced41% from 4.7, and 2.7 groundbirds reduced43% from 4.5. This precipitousdecline in overall densitywas accompaniedby a shift in the distributionof activity by the permanentresident species among the four habitat compartments.The principal net movementwas away from the shrubs (a decreaseof 32%) up into the crowns(an increaseof 21%) and down to the groundcompartment (an increaseof 42%). This responseis oppositeto what wouldbe predictedfrom a modelbased on compensatoryadjustments to migrant departures.Again, the expectationsof traditionalcompetition theory are not supported.

8--GUILD DISTRIBUTION WITHIN THE PINE FOREST

CONCEPTS AND DEFINITIONS The segregationof specieswithin a habitatmay be basedon functionalas well as physicalcriteria. Thus, speciesthat feed, sing,nest, or seekshelter on dif- ferentsubstrates of the vegetation,or that use the substratesand their resources in differentways, are bufferedagainst (functionally segregated from) potential competitorseven when they are occupyingthe samehabitats and the samecom- partmentswithin those habitats. In thisstudy I recognizedfour functional systems: foraging,nesting, singing, and resting,corresponding to the four major activities recordedin my field notes. For each of thesesystems I delineateda seriesof functional"provinces" or departments,and the birds that were found foraging, singing,etc., within a provincewere collectivelytermed a "guild" (Root 1967) with a namecorresponding to the province.These provinces and the guildsof speciesthat occupyor exploitthem combineboth functionalposition (the re- sourceor substrateselected and the way it is used) and spatialposition (compartment). They may thusbe usedas comprehensiveunits for analyzingecological distributionand the dynamicsof interactionsbetween species within a habitat suchas the submaturepine forests under discussion. 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 81

The numberof guildsto be recognizedin a systemdepends on the analysisto be undertakenand the degreeof refinementdesired for that analysis.Something betweena lumpingof all variationsinto two or threebroad categoriesand a split- ting to the level whereeach species fits a separateguild can providea usefulbasis for studyingfunctional relationships between species. At theseintermediate levels, a specieswill, as a result of its behavioralvariability and flexibility, character- isticallyhold membershipin severalguilds concurrently. Empirically based assignmentsof fractionalmemberships for each speciesin the communitycan then be made and usedto providea comprehensiveempirical record of the species'foraging behavior,a record that is roughlyequivalent to the realizedniche (trophic aspects)of the speciesas conceivedby Hutchinson(1957). Such analyseswith fractionalguild assignmentsare possible,of course,only when data are collected concurrentlyon all the speciesin the community and through all compartments of the habitat. A list of the guildsrecognized in this studyis presentedin Table 19. Semi-independentconditions of competitionand behavioral exclusionpresum- ably exist in each of the four functionalsystems. However, my data are adequate to quantitativelyanalyze segregationand overlap in only one, the foraging guild system.

PROCEDURES

Delineation of the foragingguilds of the Grand Bahama submatureforests and assignmentsof speciesin the systemwere based on the position in which birds were observedin the field with respectto three ecologicalcriteria: compartment position, food type, and foraging method. Recognizing 10 compartments (as defined in the last chapter), 10 food types (seeds,fruits, buds, leaves, nectar, sap, and flying, crawling,hiding, and burrowinganimals) and 7 major foraging methods(gleaning, plucking, probing, drilling, pouncing, sallying, and screening), there cotfid theoreticallybe 700 foraging guilds among which to distributethe species.In actualitythe threecriteria are not independent,and mostof the potential combinationsdo not exist. Edible fruits and buds, for instance,are found in the Bahamapinelands only in shruband groundcover compartmentsand taken only by pluckingmethods. The 18 foragingprovinces and their guildslisted in Table 18 thusrepresent essentially all of the naturalcombinations of compartments,food types,and foragingmethods to be found in the pine forest habitat of the Island. Sincemembers of most of the bird speciesin the forest foraged in several vegetationcompartments and used several types of food and feeding methods, each was regardedas holdingmembership in severalof the 18 foragingguilds. The fractionof a species'foraging activity assignable to eachguild was determined by referenceto all availabledata on food habits,foraging behavior, and compartment distribution. Food habits data were obtainedfrom a few freshly killed specimensfound in the courseof the study, but primarily from the published recordsof stomachcontents summarized in Martin et al. (1951), Bent (1939, 1940, 1942, 1948, 1953, 1968), and Wetmore (1916). The principal datum takenfrom theseworks was the ratio of plant and animalmaterial in specimens collectedas closeas possibleto the locality (i.e. Puerto Rico and/or Southeastern U.S.). Compromiseswere necessarywhere data were few, and in several cases 82 ORNITHOLOGICAL MONOGRAPHS NO. 24

TABLE 19 GUILDS AND GUILD PROVINCES RECOGNIZED IN THE SUBMATURE PINE FORESTS OF GRAND BAI•.MA ISLAND

Foraging guilds Nesting guilds 1. Herbivores 1. In cavities a. Ground seedgleaners a. In prepared excavations b. Stem seedpluckers b. In natural crevices c. Fruit and bud harvesters 2. On horizontal substrates d. Nectar sippers a. On ground in open e. Sap and cambium eaters b. On ground in dense cover f. Foliage eaters c. On branch 2. Carnivores d. On twig spray--tree a. Ground insectgleaners e. On twig spray--shrub b. Ground insectpouncers 3. In subvertical crotches c. Flower insectprobers a. In tree branches d. Shrub foliage gleaners b. In tree twigs e. Shrub stem drillers c. In shrub twigs f. Tree bark and wood drillers g. Tree bark gleaners Resting guilds h. Tree twig gleaners 1. On exposedperches i. Cone probers a. Above the dominant canopy j. Needle gleaners b. Within the tree canopy k. Air sallyers c. Below the canopy 1. Air screeners d. On shrubs e. On the ground Singing guilds 2. In densefoliage 1. From exposed (open) perches a. In needles a. Above tree canopy b. In leaves b. Below tree canopy c. In ground cover 2. From foliage 3. On vertical trunks a. In trees b. In shrubs 3. From the air (Flight songs) it was necessaryto make best guesseson the basisof availableinformation for other speciesknown to have similarfeeding habits. More detailedinformation on the type of plant or animalmaterial taken by the birdscould generally be determinedwith considerableconfidence from observations on feedinglocation and foragingmethod. A total of 1,912 specialobservations of feeding activity (independentof the censusoperations) provided most of this information,but a few subjectiveadjustments of these records were necessary becauseof differentialobservability in certain situations.Most of the foraging observationson Palm Warblers, for instance,were made of birds in tree crowns, yet data on first detectionsindicated that thesebirds foragedmore in the shrub and groundvegetation than in trees. For most situations,the tallies of compartment distributionin the censustransect records could not be used in making foraging guild assignmentsbecause time could not be taken then to observe specialactivities and examine foraging substrates. In the absenceof a singledirect method for assigningguild memberships, I used all data availablefor each species,relying most heavily on my specialforaging observations.I then subjectivelydivided the total foragingactivity of eachspecies into tenthsand indicatedfurther instancesof unusualbehavior as "traces" (T). For the analysesthat follow, I have multipliedthe pine forest densityand biomassvalues for eachspecies (see Table 2) by the bird's relativerepresentation 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 83

in eachguild. Thisprovides absolute values of guildroles and a basisfor deter- miningthe overalldensity and biomassof eachguild as well as the contribution of eachspecies to it. Table 20 showsthe foragingguild distributionin tenthsof total foraging activityfor each of the 33 speciesoccurring as regularmembers of the avian communitiesof the submaturepine foresthabitat. A "T" in the table refersto observationsof occasionalactivity in a guildby speciesconcentrating their foraging in other guilds. In the descriptionsand discussionsthat follow I will use the followingterms to reflect the distributionof specializationof speciesin the foragingguild system.A "primarymember" of a guild is one that doeshalf or moreof its feedingin that guild;a "secondarymember" is one that feedsfrom two- to four-tenthsof the time in the guild, and a tertiarymember is one that does only one-tenthof its feedingin the guild. The CubanEmerald Hummingbird is thusa primarymember of the nectar-sippingguild, a secondarymember of the flower-probinginsectivore guild, a tertiarymember of the needleinsect-gleaning guild and showsa traceof activityin the pine twig insect-gleaningguild.

THE FORAGING GUILDS OF THE GRAND BAHAMA PINE FOREST The membershipsand relative strengths of dominantspecies in eachof the 16 occupiedforaging guilds of the winteringcommunity are graphicallyportrayed in Figure38. The calculatednumber of individualsper km2 is givenfor eachspecies in Table21, and the biomassin gramsper km• in Table22. Tables23 and 24 presentthese density and biomassvalues as the percentagecontribution of each speciesto the guild composition. In the descriptionsof guild structurethat follow, relative abundanceor numericalranking of the speciesin a guildis expressedin termsof dominance. The dominantand subdominantmembers are the specieswith 40% or more,and between15 and 39% representationin the guildrespectively; subordinate species are those furnishingbetween 5 and 15% of the guild membership.Species contributingfrom 1 to 4% are classedas incidentalmembers, and thosewith only tracesof activityin the guild are regardedas visitorsrather than members.In thissystem a speciesof smallbody size may, of course,be dominantin the density ranking,yet subordinantor evenincidental in the biomassranking. Ground-gleaningherbivores.-•It is generallyimpossible to determinejust what a ground-gleaningbird is ingesting,and thus to separateseed-eating from insect- eating.Published accounts of foodhabits for eachspecies and considerationsof feedingequipment, particularly the bill, wereused in combinationwith personal observationsto delineatethe ground-gleaningherbivores of the Grand Bahama pinelands. Two permanentresident species, the ZenaidaDove and the GroundDove, apparentlylimit theirforaging almost entirely to thisguild (Table20, column1); two others,the Red-leggedThrush and the grassquitare secondarymembers. In the winter thesefour are joined to a limited extentby the primarily insectivorous catbirds and Mockingbirds. The dovescomprised 45% of the individualsin the guild in winter (Table 23a, column 1) but 79% of the biomass(Table 24a, column 1). After the winter 84 ORNITHOLOGICAL MONOGRAPHS NO. 24

TABLE 20 FORAGING GUILD DISTRIBUTION OF MEMBERS OF THE PINE FOREST BIRD COMMUNITY ON GRAND BAHAMA ISLAND •

Herbivores Carnivores (mainly insects)

Foraging guilds: a b c d e f a b c d e f g h i j k 1 N a Ampl. 4 ZenaidaDove 9 - 1 7 .11 Ground Dove 10 21 .00 C. Emerald Hum'bird 6 3 T - 1 23 .35 Hairy Woodpecker 2 1 2 4 1 85 .51 Yel.-bellied Sapsucker 1 3 6 .41 Gray Kingbird - 2 8 - 6 .17 Loggerhead Flycatcher - 4 T 6 - 14 .27 Stolid Flycatcher - 1 1 8 - 15 .22 Gr. Antillean Pewee - 2 - 1 T T 7 - 33 .31 Bahama Swallow 10 17 .00 Brown-hd. Nuthatch 1 4 3 1 1 97 .49 Bahama Mockingbird 1 - 6 1 2 - T T - 14 .45 Gray Catbird 1 - 5 3 T - 1 13 .44 Red-legged Thrush 3 - 3 4 17 .38 Bl.-gray Gnatcatcher T 3 2 1 - 4 T - 169 .51 Thick-billed Vireo ! - 1 - 7 1 5O .33 Blk.-whiskered Vireo 1 8 1 3 .22 Blk.-and-Wh. Warbler 10 T T 50 .07 Parula Warbler 1 9 T 19 .15 Cape May Warbler 1 2 T 2 T - 4 1 - 38 .58 Blk.-thr. Blue Warbler 1 8 1 2 .22 Yellow-rumped Warbler 4 T T 3 - 1 2 - 36 .51 Blk.-thr. Green Warbler 6 1 - 3 7 .31 Yel.-throated W. (migr.) T 1 2 1 6 T - 104 .45 Yel.-throated W. (res.) 7 2 T 1 T - 48 .35 Olive-capped Warbler T 1 T 9 T - 351 .22 Pine Warbler 1 T T 9 T - 89 .22 Prairie Warbler 5 T - 5 51 .28 Palm Warbler 1 1 3 - T 1 T T - 3 1 - 226 .80 Ovenbird 10 T 14 .04 Common Yellowthroat 8 2 5 .17 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 85

TABLE 20 (CoNTiNUED)

Herbivores Carnivores (mainly insects) Foraging guilds: a b c d e f a b c d e f g h i j k 1 N a Ampl. 4 Bahama Yellowthroat 2 6 - 1 1 57 .38 Redstart 4 T 1 5 - 69 .36 Bananaquit T 4 2 2 2 81 .50 Str.-headed Tanager T - 9 T 1 32 .18 Black-f. Grassquit 2 $ 1 1 1 43 .47 Number of specieswith 1/10 or more 6 1 14 5 1 0 8 6 4 16 1 3 7 9 2 17 8 1 Total number of species 7 1 16 5 1 1 10 7 6 18 1 3 13 17 6 18 14 1 (1912) •Estimated to nearest tenth from observations of foraging activity in 1968, 1969, and 1971 and from published information on food habitats. : For identity of foraging guilds see Table 19. a N = sample size of foraging activity observations. 4 Amplitude: guild dispersion amplitude (Ht/H t max.). residentshad gone,i.e. the breedingseason, these 2 speciescomprised 66% of the individuals and 87% of the biomass (Tables 23b and 24b). I have very lime informationon the food items taken, but doves, particularly the ZenaidaDove, clearly concentratedon relativelylarge seeds. They fed while

16 16 121 150 2 160 7 57

A B C D E F G H Ground Stem Fruit Nectar Sap Ground Ground Flower Seeds Seeds Eaters Sippers Eaters Insects-G[. Insects-Pn. Insects

114 I 5 56 41 7 .351 68

I J k L M N 0 P Shrub Stem Bar k Bur k Twig Cone Needle Air Insects Insects Drillers Gleoners Gleaners Probers Gleaners Sallyers FmURE38. Membershipand dominancestructure in the 16 foragingguilds of the wintering communityof the standardforests. The figure aboveeach columnis the total density (birds per km'•) in the guild. The dominant and subdominant members are named in order from the bottom up; the figures inside each column at the top give the number of additional specieswith 1% or more representation and, in parentheses,the number of visitors, i.e. specieswith less than 1% representation. 86 ORNITHOLOGICAL MONOGRAPHS NO. 24

I I I I I i I I I I f I I I I• I I I I I I I I I I I I I0 I0

I l

I I

I ,"l•. I I I cq

I I c-] [,•

I I I •1•o III I I I I I ooo

IlOO •11 I I I I

i,q. i l

I [ I I I I I 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 87

I I I

•11 II II

Jllll•

Ii iIiI I

IIIIIII I I I IOO IlOO 88 ORNITHOLOGICAL MONOGRAPHS NO. 24

I I I I I I I I I I I I I I0 I I I I I I I I I I I I I•

ler• I I I I I II/-i

I I I r-.

I I

I I I

III III I •1• Ill

I I I I I I

I I I

I I I I I 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 89

I I I I I I I I I I I I I•

Z O

I• I I I I I I I I

I I I I IO

III I II IO

I lb-.

I IIIII III I I I I•

I I I 90 ORNITHOLOGICAL MONOG•PHS NO. 24

I I I I I I I I [ I I I I I I I I I I I I I I I I I I I IO

z I•1 I f I•

Io I I I I I I

I I I I I

IIII o

IIII III 111•10 11 I I I

IIII III IIII10 •11 I I l

•1111 III

• I•111 •I•

¸ I IIIII III llll• Ill I I I I l ¸

m•lllll •11 I I I I I 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 91

I I I I I I I I I I I I I I•

I I I

I I I •-

I I I

r i

I I I I I I

I I I I I I I I I I

Ill

cq I•

I I I

• I I 92 ORNITHOLOGICAL MONOGRAPHS NO. 24

IIIIII IIIll I II I I I I I I I I I I I I I I•

I I I IIIIIIii

IO I I I I I I I I I I IIIIII11

I I I I I

I I

I I I I

I• I I •'

I I I I I 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 93

I III I i

IIIII I I I I 94 ORNITHOLOGICAL MONOGRAPHS NO. 24 walkingrapidly on fairly open ground,primarily roadbeds and road shoulders, turningand jabbing at surfaceobjects at the rate of 20-30 pecksper min. Grass- quits, althoughprimarily stem seedpluckers, were often observedbiting and hullingsmall grass seeds on the ground.The other three speciesare assigned membershipin this guild only becausethey are known to eat fallen fruit and smallamounts of seed. Their foragingmethods suggest that seedsare generally ignored. This summarysuggests that eachmember of the guild occupiesa relatively separatesubprovince and that interspeciesoverlap of food resourcesis slight. Ground Doves, aside from their assumedselection of smaller seedsthan Zenaida Doves, tend to foragein more open sectionsof the forest. Stem seed-pluckers.--Thisguild is distinctiveand readily delineated. It has only one memberin the Grand Bahamapinelands, the grassquit,and this member doesabout half of its foragingin the guild. The birds generallywork from the ground,reaching up to nibbleat the tiny seedsin the grassheads, or leapingup to biteoff a seedor to pull the wholehead down within reach. Seeds are apparently bittenoff the headsand hulledon the spotwith rapid mandibularmovements. A grassquitwas once seen perchedon and feeding on the seedsof a thistle, an uncommonplant on the study areas. Fruit and bud harvesters.--Berries of many varieties were common and widespreadin the shrubstratum of the Bahamapinelands through both the winter and springmonths, and new budswere numerous,particularly in the early months of the year. Sevenspecies of permanentresident birds and 8 of winter residents were observedfeeding on fruits and/or buds,but only 3 of these,one permanent residentand 2 winterinvaders, were primary membersof the guild. The Striped-headedTanager, the fourth mostcommon bird of the breeding seasoncommunity, specializedon fruits and buds and constituted87% of the individuals(Table 23b) and 79% of the biomass(Table 24b) of this guild at that time. Secondaryand tertiary fruit-bud eatersof the breedingseason com- munitywere the Hairy Woodpecker(a pine trunk driller), the Red-leggedThrush (a ground-gleaninginsectivore), the Thick-billedVireo (a shrubfoliage gleaner), and the grassquit(an herb and grassplucker). Bananaquitswere seenon fruit once. The Striped-headedTanager also dominatedthe fruit-bud eatingguild in the winteringcommunity, with 47% of the individualsand 54% of the biomass.It was closelyfollowed by the abundantPalm Warbler with 37% and 21%. Because of their low densities,the two primary fruit eatersamong the winter residents,the catbirdand Mockingbird,were incidentalmembers of the guild with only 3% and 2% of the membershipand 6% and 4% of the biomass,respectively. The Yellow- rumpedWarbler contributedan additional5% of the membershipand 2% of the biomass.Lesser members were the Yellow-belliedSapsucker, the Parula Warbler, Cape May Warbler, and the migratoryrace of the Yellow-throatedWarbler. Striped-headedTanagers perched quietly in small groupsnear the berrieson which they fed, reachingout to bite then mandibulatethe pulp and allowingthe peelsand husks,when present,to fall to the ground. Feeding apparentlyoccupied only a small fraction of the daylight hours, and much time was spent by the 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 95 tanagersquietly perchinghigh in the pines. Other fruit eatersof the breeding seasonwere specialistsin other guilds and apparentlyencountered little or no competitionfor fruits or buds. The two dominantsof the wintering community guild fed on different items. The Palm Warbler, a highly eurytrophicbird at this season,took nine-tenthsof its food in other guilds and assumedimportance in the fruit-eatingguild only becauseof its very largenumbers. Like the lesscommon Yellow-rumped Warbler, it focusedlargely on the Wax Myrtle berry, a fruit of secondaryimportance to the tanager. Catbirdsand Mockingbirdsfed mainly on fleshy berries such as those of Smilax and Lantana, fruits also harvestedby the tanagers. In view of the leisurelyfeeding habits of all thesespecies and the continuousabundance of fruits, significantcompetition probably did not occur during the periodsof this study. Nectar sippers.-•Blossomsof many types were abundanton Grand Bahama Islandthrough all the winterand springmonths. No measurementswere attempted, but the successiveblossoming of variousspecies of Ernodea,Calliandra, Agava, Lysiloma,Tetrazygia, Tabebouia, and Acacia appearedto offer an abundantsupply of nectarfor the membersof this guild. Of specialinterest were the spectacular bloomsof Agave braceanain the pine forestsin Februaryand March. The huge headsof showy yellow flowers,literally drippingwith nectar, attractedmany winterresidents as well as permanentresidents and createdmany "food territory" situationswith intensivefighting and chasing(Kale 1967, Emlen 1973). It wasgenerally impossible to determinewhen a flower-probingbird wassipping nectarand whenit was taking smallinsects. A crudedecision on this matterwas madeon the basisof publisheddata on stomachcontents of relatedspecies. The error in this decisionmay be considerable. Three speciesconstituted the nectar-sippingguild in the breedingseason. The only primarymember was the CubanEmerald Hummingbird which, with an overalldensity of 106 birdsper km2, contributed 62% of the guildmembership and 30% of the biomass.A strongsecondary member was the bananaquit,with 30% of the membershipand 50% of the biomass.The third member,the ground- skulkingBahama Yellowthroat, commonly probed the smalltubular flowers of Ernodea and contributedthe remaining 8% of the individualsand 20% of the biomass. Two specieswere added to thisguild during the wintermonths, the abundant PalmWarbler and the relativescarce Cape May Warbler. PalmWarblers, although onlytertiary nectar sippers, added 45 membersper km2 to constitute30% of the winterguild membership and 44% of the biomass,while the CubanEmerald Hummingbirdheld its dominantnumerical position with 43% of the membership, but had only 16% of the biomass.These wintering warblers concentrated their nectarsipping on theAgave blossoms where they came into closecontact and frequentaggressive encounters with the hummingbirdsand bananaquits. The irregular changingof floral distributionin the forest had an obviouseffect on the distributionof nectarsippers. This, and the frequentaggressive encounters particularlyon the Agave heads,indicated interference within and betweenthe speciesbelonging to the guild. The BahamaYellowthroat, concentrating on the 96 ORNITHOLOGICAL MONOGRAPHS NO. 24 low-lyingand widely scatteredErnodea blossoms, was essentiallyindependent, but the otherfour speciesof the winteringcommunity clearly competed for this particularresource at times. Sap and cambiumeaters.--Apparently only one species,the winter resident Yellow-belliedSapsucker, exploits this food resource.In my limited observations I saw sapsuckersfeeding on berriesonce, and assume,on the basisof published accountsof food habits in North America, that the birds take considerableinsect foodfrom the trunk and bark of pine trees. In my bestestimate this spedesis a primary memberof the sap and cambium-eatingguild and comprisesa density in the guild of about 1.8 birds and 79 g per km-ø (Tables 21 and 22). Foliage browsers.-•Exceptthat this guild has importance in certain avian communitieselsewhere, I wouldmerge it with the fruit andbud harvesters,for my only evidenceof foliagebrowsing is one casein the Striped-headedTanager. A bird was watched as it nibbled for several secondsat the succulenttip of a growingSmilax stem. Ground-gleaningcarnivores.---As akeady noted it is often impossibleto tell the nature of a ground gleaner'sprey. On the basis of publishedrecords on stomach contents in related species,the Bahamian doves are assumedto be primarily herbivorous,and the mimids, thrushes,and warblersprimarily insectivorous when ground gleaning. Birds that pounce on ground-inhabitinginsects are treatedin anotherguild. The ground-gleaningcarnivore guild apparentlyhas three memberson Grand Bahamaduring the breedingseason and sevenduring the winter. In the breeding seasonthe Bahama Yellowthroatis the only primary ground gleaner. The Red- leggedThrush takes slightly less than half of itsfood in thisguild, and the grassquit a smalleramount. Severalothers, including the Pine Warbler and Blue-grayGnat- catcher,obtained some insectfood by ground gleaning. The dominant member of the guildis the BahamaYellowthroat, with 76% of the membershipand 56% of the biomass. The rather scarceRed-legged Thrush contributes11% of the guild'spopulation and 37% of its biomass,while the grassquitcontributes 13% and 7%. In the winterthe guild,with 7 members,is dominatedby winter residentsto the extent of 86%. The principalmember of this invadingforce is the Palm Warbler which, with 25- 30% of its feedingin this guild constitutes79% of the individualsand 69% of the biomassof the guild. Much lessimportant as members are the Mockingbird(< 1% and 1%), the CommonYellowthroat (5% and 4%), and the Ovenbird,a primarymember (1% and 1%). Two distinct size classescharacterize the ground-gleaningcarnivores, and the largermimids and thrushespresumably take different(larger) prey than the warblers.Among the largermembers, the Red-leggedThrush and the Mocking- bird foragein opensituations, the thrushalong roadways and the mockingbirds primarilyin clearingat the forestedge. Among the warblersthe samevariety of feedingsituations may be recognized,leading to segregationof the potentially competingguild members. Bahama Yellowthroats frequent the deeplitter under shrubs,actually tunnelling under vines and leaves.Ovenbirds feed primarily on the litter surface as do Palm Warblers, but the latter tend to select more open 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 97

situationsin the forest. Finally,the grassquits,Pine Warblers and otherincidental membersof thisguild are found mostly on nearlybarren ground as along roadways or in recent burns. Becauseof this diversityof feedingsituations and the relatedvariations in foragingbehavior, there appears to be very little directoverlap in the foraging nichesof the various guild members. Groundpouncers.---A second method of foragingon terrestrialarthropods is to pounceon them from an elevatedperch. This method,referred to as "drop to ground"by Orians(1969), is effectiveon a somewhatdifferent prey faunathan groundgleaning and is generallypracticed by a differentgroup of birds. For these reasonsI haverecognized its practitionersas a distinctguild in this analysis. Four permanentresidents, one summerresident, and one winter resident practicedground-pouncing in the Grand Bahamapinelands. None of these, however,used pouncing as theirprincipal method of foraging,and the total guild membershipwas very small. In the breedingseason community all four members of the flycatcherfamily, the Gray Kingbird,Loggerhead Flycatcher, Stolid Flycatcher,and GreaterAntillean Pewee, occasionally dropped to the groundto captureinsects, and the Thick-billedVireo was observedto do it severaltimes. Mockingbirdscommonly fed in this way in the open terrain of their summer habitatand wereobserved doing it in the pine forestin winteron one occasion. Flower probers.--It is well knownthat many,perhaps all, flower-probingbirds take a certain number of insectsas well as nectar; in some casestheir principal food sourceis apparentlyinsects. It is thereforeappropriate to recognizea flower-probingcarnivore guild, althoughI have insufficientdata to quantifyits membershipand its overallsignificance in the Bahamapineland community. In the absenceof informationon food habitsI will guessthat three prominent flower probersof the breedingseason community are secondaryor tertiary membersof thisguild. Theseare the CubanEmerald Hummingbird, the Bahama Yellowthroat,and the bananaquit.The abundanceof thesespecies suggests that the total membershipof the guild is quite large,perhaps about 50 individualsper km2. This groupis apparentlysupplemented in winter by 3 migrantspecies, the Palm Warbler,Cape May Warbler,and Black-throatedBlue Warbler, and I will guessthat this supplementamounts to 10% or lessof the total. Shrub foliage gleaners.-•After an unsuccessfulattempt to differentiatestem and leaf insectgleaners in the shrubstratum I lumpedall suchbirds and added those that foraged in the occasionalLysiloma or Gumbo Limbo (Bursera sirnarouba)tree (rarely more than 5 m tall) into a singleguild of shrub foliage gleaners.Insect and spiderforaging usually could be separatedfrom berry-eating without difficulty. No lessthan 7 speciesin the breedingseason community and 15 in the wintering communitydid some shrub-foliagegleaning. Two membersof the breeding season community, the Thick-billed Vireo and Black-whiskeredVireo, were primary membersof the guild. The Thick-billedVireo contributed37% of the membershipand biomass respectively. Bananaquitswith two-tenths of their activityin the guildand Striped-headed Tanagers with one-tenthwere subdominants in terms of both numbers and biomass. The gnatcatcher,a subdominanthumeri- 98 ORNITHOLOGICAL MONOGRAPHS NO. 24 cally, was a subordinatein biomass.Greater Antillean Peweesand Bahama Yellowthroatsdid a small amountof foragingin this guild. Catbirdsand 8 of the 11 speciesof migrantwarblers joined the guildof shrub foliagegleaners in winter. With them the total membershipand biomassmore than doubled. As in 6 other guilds,the principalmember of this invading contingent,constituting nearly three-quarters of its total, wasthe Palm Warbler, whichdominated the winteringguild with 40% of its total membershipand 40% of its biomass. Two major foragingmethods were notedamong the shrubfoliage gleaners. The twovireos advanced slowly through the foliage, perching on a twigor branch and peeringabout for 5 or more secbefore leaping or flutteringa half meter or moreto snatchan insectfrom a twig or leaf surface.Warblers and Gnatcatchers movedmore rapidly, generally making their final jumpfrom a distanceof only a few centimeters.The differencesin thesetwo foragingprocedures and in the manyvariations of themprobably function variously for differenttypes of insect prey. Combinedwith differencesin compartmentdistribution and in bodysize, theypresumably spread the pressuresof bird predationon shrubfoliage insects over many behavioraland morphologicaltypes. Shrubstem drillers.--Shrub stems constitute a minor segmentof the vegetation of a pinestand, and the guildwhich exploits the insectpopulations on Grand BahamaIsland is apparentlylimited to onespecies, the permanentresident Hairy Woodpecker.Of the 85 observedfeedings of this bird 17% were on shrub stems. No consistentdifference between the sexeswas noted in this sample. Pine bark and wood drillers.--The excavatingand extractingof insectsfrom within or beneathbark appearsto warrantthe recognitionof a specialguild, distinctfrom that encompassingthe gleaners,which coverthe bark surfaceor probe under the edgesof bark flakes (next section). Membershipin the bark- and wood-drillingguild on GrandBahama includes only the Hairy Woodpeckerand the Brown-headedNuthatch in the breeding season.The Woodpeckerdoes about two-tenths of its foragingin the guildbut constitutes45% of the membershipand 88% of the biomassof the guild. The Brown-headedNuthatch, a much smallerbird with limited capacitiesin wood drilling,allotted about one-tenth of its foragingto thisguild and contributed 55% of the guild'smembership and 12%of its biomass. In the winterthe guildis joinedby the Yellow-belliedSapsucker, a sap and cambiumeater, which contributesless than one-tenthof the guild's winter membership. Barkgleaners.--The bark-gleaning guild is quiteclearly delineated in its spatial distribution,the bark surfacesof pine trunksand branches,and in its food composition,which is almostentirely anthropods. The feeding behavior dimension is lessclearly defined in that the distinctionbetween pecking and drillingis sometimes difficult to make. The compositionof this guild is surprisinglycomplex on Grand Bahama Island. In the breedingseason community there is one primary member,the residentBahama race of the Yellow-throatedWarbler, three secondarymembersin theHairy Woodpecker, Brown-headed Nuthatch, and Blue-gray Gnatcatcher, and 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 99 one tertiary--the Pine Warbler. LoggerheadFlycatchers, Greater Antillean Pewees,and Olive-cappedWarblers were alsoseen foraging on bark surfaceson a few occasions.The mostnumerous bark gleanerin the submatureforests was the Yellow-throatedWarbler, with 31% of the membershipand 28% of the guild biomass;the gnatcatcherand Brown-headedNuthatch were dose behind with 26% and 23% of the membership,and 12% and 10% of the biomassrespectively. Hairy Woodpeckerscontributed only 9% of the membershipbut a significant 39% of the biomass,while the Pine Warbler was representedwith 11% of the membershipand 12% of the biomass. In the winteringcommunity, migrants made up 34% of the individualsand 33% of the biomassof the 9-memberguild. The mostimportant member of this contingentwas the Black-and-whiteWarbler (20% and 16%). The migrant race of the Yellow-throatedWarbler, identifiedunder favorablecircumstances by the contrastingwhite of its underparts,was three times as numerousas the resident race in the winter communitybut, in marked contrastto the latter, fed only occasionallyin the bark-gleaningguild. The redstart,an uncommonmember of the winteringcommunity, fed only occasionallyon bark in 1968 and 1969, but as an abundanttransient briefly dominated that guild in May of 1971. Three major methodsof foragingwere used by bark-gleanersin the Grand Bahamapinelands: pecking or probinginto cracksor under the edgesof bark flakes,flushing the prey and chasingit, and pouncingon the bark surface.The first was the commonestand was usedmost often by the dominantand subdominant Yellow-throatedWarbler, Nuthatch, Hairy Woodpecker,and Black-and- white Warbler. The flushing method was commonlyused by gnatcatchersand was the principalmethod of the redstart. Both of thesespecies spread and flash strikinglymarked tail patterns toward the surfaceas they hitch along. Actual captureswere not often seen,but a tumblingchase was obviouslya responseto the flushingof an insect. Yellow-throatedWarblers also used this method on occasion. The third method,pouncing on the trunk, was usedby kingbirdsand peweesand occasionallyby gnatcatchers,redstarts, and others. While the identityof the prey wasnot determinedin my observations,I suspect that the three methodstend to functionfor different typesof prey and that these differences,together with prey size selectionby large and small bird species, resultin a dispersionof predatorpressures and a reductionin potentialcompetition for any specificinsect type. Pine twig gleaners.---Forthe purposesof this studypine twigs are definedas branchesless than 2 cm in diameterwith few or no greenneedles. Such twigs are generally 2-3 years old and up to 1 m long. The surface is gray and scrofulouswith old needlescars, in contrastto the large brown plates and flakes (bark) on the surfacesof older and larger branches.Many birds, particularly needle-gleaners,use both live and dead twigsfor perching,but guild membership as here definedis limited to the relativelyfew birds that actually forage on the twig surfacesor drill into them. Examinationof severalhundred twig samples revealeda very sparsepopulation of spidersand small insectsamong the old needle scars. The twig-gleauingguild in the Grand Bahamapinelands contained 6 primary 100 ORNITHOLOGICAL MONOGRAPHS NO. 24 and secondaryplus 3 incidentalmembers in the breedingseason, and 9 members plus 7 incidentalsin the winter. No speciesspent more than 30% of its foraging activityin the guild, and only 2 permanentresidents, the Brown-headedNuthatch and the Yellow-throatedWarbler, and 1 winter resident the Yellow-rumped Warbler,took as muchas 20% of their food from twigs. The 3 dominant and subdominantmembers of the guild during the breeding seasonwere the Olive-cappedWarbler, the Brown-headedNuthatch, and the Blue- gray Gnatcatcherwith about 43%, 23%, and 18% of the membershiprespectively, and 38%, 14%, and 11% of the biomass. The Hairy Woodpecker (3% and 20%), the residentrace of the Yellow-throatedWarbler (12% and 14%), and the summeringBlack-whiskered Vireo (1% and 2%) had minor representationin the guild, while hummingbirds,pewees, and Pine Warblers were each recorded foragingon twigs on a few occasions. Winter residentsraised the total membershipfrom 28 to 41 birds per km2 and the biomassfrom 220 to 367. Prominentamong this winter contingentwere the Yellow-rumped and migrant Yellow-throated Warblers (19% and 22% of the membersand 11% and 22% of the biomassrespectively) and the ubiquitousPalm Warbler, which thoughfeeding only incidentallyon twigs contributed8% of the winter membershipand 9% of the biomass. Pine twigs were apparentlypoor sourcesof food and served as no more than secondaryor incidentalforaging sites for any species.They were usedextensively as lookout perchesfor flycatchingsallyers, however, and as accessroutes to the relativelyrich foliageand bark substratesby mostif not all the arborealbirds of the forest. Their occasionalor incidentaluse as food substratesby many speciesis, therefore,not surprising.Dead twigs appearedto offer more food than live twigs, but no consistentrecords were kept to permit quantitativecomparisons. Foraging methodsincluded direct searchingand peckingby most species,drilling by the nuthatchesand woodpeckers,flushing and chasingby redstartsand gnatcatchers, and hoveringby hummingbirds,pewees, and severalwarblers. Pine cone probers.•The small (5-7 cm) cones of mature Caribbean Pines are retainedon the twigsand branchesin large numbersfor severalyears, opening their scalesafter the first year to presentlarge irregular elustersof deeply creviced feedingsubstrates for insectivorousand pineuseed-eatingbirds. Conesare reported to be importantsources of insectsand seedsfor titmice in England (Gibb 1960) and for nuthatches,warblers and other speciesin North America (Morse 1966). The resource was, however, very weakly exploited on Grand Bahama Island during the period of this study. Even the Brown-headedNuthatch, a habitual exploiterof coneswhere it has been studiedon the continent,was rarely seen foragingon cones. In the breedingseason the only consistentmember of the cone-gleaningguild was the Brown-headedNuthatch, and it did lessthan 10% of its foragingon cones. All three speciesof residentarboreal warblers often hoppedover conesor paused on them to singbut were observedprobing the cone surfaceor creviceson only a few occasions. The migrant Yellow-throatedWarbler frequentedcones during its winter stay in the pine forestand apparentlytook abouta quarterof its food in this guild. It 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 101 was thus the dominantcone-gleaner during the winter, with the nuthatcha strong subdominantand five warbler speciesrecorded as occasionalvisitors. Pine needlegleaners.--I have includedall types of foragingin needleclusters in this guild. The dividingline betweentwig and needlecluster is not alwaysclear to an observeron the ground,but all birdsthat peckedat the baseof or out into the needlesof the previousseasoh's growth are included. About half of the pinelandbird speciesforaged at least occasionallyin pine needles.In the breedingseason community two species,the Olive-cappedand Pine Warblers,were primaryneedle gleaners; and two, the gnatcatcherand the bananaquit, weresecondary members. Cuban Emerald Hummingbirds, Stolid Flycatchers, Brown-headedNuthatches, Thick-billed Vireos, Yellow-throatedWarblers, and grassquitsalso foraged in needlesfor minor portionsof their food. The dominantneedle gleaner in the breedingseason, contributing 55% of the membershipand 49% of the biomass,was the Olive-cappedWarbler. It was followedby the Pine Warbler (19% and 29%), the Blue-grayGnatcatcher (10% and 6%), and the bananaquit(6% and 8%). The invasionof winter migrantsdoubled the membershipof the pine-gleaning guild and more than doubled its biomass. The main contributer to this increase was the Palm Warbler which, with about 30% of its feeding in the needles, introduced 136 individuals and 1,492 g per km• to the guild. This constituted 39% of the membershipand 46% of the biomassof the wintering guild. Other migrantmembers of the winteringguild were Cape May Warbler, Yellow-rumped Warbler, Black-throatedGreen Warbler, Yellow-throatedWarbler (migrant race), Prairie Warbler, and redstart. Of these the Yellow-throated Warbler and the Prairie Warbler were primaryneedle gleaners, but becauseof their relativescarcity constitutedonly 8% and 2% of the membershipand 8% and 2% of the biomass respectively. Three foragingmethods were employedby needlegleaners: (1) reachingout from the twig baseof a cluster,(2) crawlinginto the needlesand peckingeither inward toward the shootor outward,and (3) hoveringoutside the cluster (Table 25). The first of theseis usedby nearlyall membersof the guildand may thusbe consideredthe leastspecialized. The winteringPalm, Yellow-rumped,and Yellow- throatedWarblers, which feed largelyin rather differenthabitats on their breeding groundin the north,fed almostentirely in thisway whenin the needlecompartment on Grand Bahamaas did mostof the residentspecies. Crawling or tunnellinginto the needleclusters was the dominantforaging method of the Olive-cappedWarbler and an importantsecondary method in the Pine Warbler (Table 25b). Thesewere the two primaryneedle-gleaners of Grand BahamaIsland. The methodis apparently rarely usedby other species. Hovering to snatch prey from the distal portions of needle clusterswas the principal needle-gleaningmethod of the Cuban-EmeraldHummingbird and was observedon several occasionsin the Thick-billed Vireo and the Cape May Warbler. It wasalso used on occasionby nearlyall of the needle-gleaningspecies. Peckingfrom the basaltwig limits a bird to animalsin the basalportions of a needlecluster, while hoveringmakes the outer portionsof needlesavailable. The heart of the cluster, particularly the basesof the needle bundles and the central 102 ORNITHOLOGICAL MONOGRAPHS NO. 24

TABLE 25 FORAGING METHODS OF PROMINENT MEMBERS OF THE NEEDLE-GLEANING GUILD •

Winter a Spring • Peck from Crawl a Hover Peck Hover Peck base in

Permanent residents Olive-cappedWarbler 0 95 2 I01 14 53 Pine Warbler 0 37 0 32 25 7 Yellow-throated Warbler (PR) 0 2 0 3 3 0 Other permanentresidents 4 18 0 I I 11 0 Total permanentresidents 4 152 2 147 53 60 Winter residents Palm Warbler 8 48 - - 48 0 Yellow-throatedWarbler (WR) 0 25 - - 25 0 Other winter residents 12 34 - - 34 0 Total winter residents 20 107 - - 107 0

• 1968 and 1971 data pooled. • Incidence of hovering and pecking from a perch before vs after the winter migrants departed. 3 Incidence of reaching into a cluster from the base vs crawling into the cluster. (Unfortunately this division was not incorporated onto my field tally sheets until May of 1971, but incidental entries in my notebooks in both 1968 and 1971 indicate that I never saw the "crawling in" foraging method in the migrant species.) part of the shoot, are perhaps most readily reached by the needle crawlers. The food supplyis thusparcelled out to a variety of speciesaccording to the foraging methodsemployed. The small size of the Olive-cappedWarbler may be an asset in its exploitationof the heart of needleclusters, but the Pine Warbler, a large species,has adoptedthe methodto at least a limited degreeon Grand Bahama. No birds on Grand Bahamawere seento foragewhile hangingbeneath the twigs like titmice. Air sallyers.-•The term air sallyeris appliedto birds that prey on flying insects by sallyingout from fixed perchesto snatchtheir prey in midair. The guild is distinguishedby the vegetationcompartment occupied (between trees) and the foragingmethod. The food itemstend to be adult stagesof Diptera Hymenoptera, Lepidoptera,and other free-flyinginsects. The flushersand chasersof the foliage- and bark-gleaningguilds might be classifiedhere, but the food resourceexploited in thesecases seems to justify the assignmentgiven them in this report. Most tyrant Flycatchersare primary air sallyers,and the guild is often closely associatedwith them. Many other speciessally for flying insectsat times,however, and no fewer than 15 specieswere recordeddoing so in this study. The three permanentresident and one summerresident Tyrannid of the Grand Bahamabreeding community were all primary air sallyers. They also dominated the guild during the winter, with three residentwarblers and the Gnatcatcher contributingminor supplements. In winter the guildwas trebledin both membershipand biomassby invading species.The Palm Warblersdid much sallying,especially toward evening, and becauseof their large numbers came to dominate the guild with 67% of the membershipand 60% of the biomassduring the winter months. Redstartsand 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 103

Yellow-rumpedWarblers contributedan additional 5% and 4% to the winter membershipand 3% and 3% to its biomass. Mockingbirdsand several other warblers made lesser contributions. During the breedingseason the Gray Kingbird capturedlarge insectsalong roadwaysin the moreopen areas. LoggerheadFlycatchers apparently took similar prey but tendedto foragein the denserareas of the forest. StolidFlycatchers and peweestook smallerprey, capturingthem at lower levels and on shortersallying flights than the two large species. Insufficientobservations (15 records) were obtainedon the StolidFlycatcher to permita meaningfulcomparison of its foraging behavior with that of the Pewee. Air sallyingby warblers in the wintering communitywas, as noted above, largely restrictedto specialtimes and specialsituations. The swarmingof small dipteransat duskwas responsiblefor a large part of their air sallying. On several occasionsPalm Warblersseemed to drop all otherforaging, gliding out in innumer- able salliesto snapup small insectsabove the shrubsor betweenthe tree crowns. With this concentrationof feedingin brieflyfavorable situations, it is doubtfulthat they seriouslyinterfered with the foragingof the primarymembers of the guild. Air screeners.--Theterm air screenersis applied to birds that beat back and forth capturingflying insectson the wing. Swifts,swallows, and nightjarsare the usualmembers of this guild, and threeof these,the Chuck-willsWidow, the Cuban Nighthawk, and the Bahama Swallow,were observedat one time or another in the Grand Bahamapineland. Bahama Swallows, the only regular air screenerof the standard pineland community,foraged in more opensituations and especiallyalong the southcoastal beachesduring the winter,but they enteredthe pine foreststo nest in cavitiesin high dead pine stumpsand to forage above the tree tops or at lower elevations along roadwaysand forest edgesin April and May. These birds with over 11 individualsand 165 g per km2 probablydid not competesignificantly with any of the other regularcommunity members.

GUILD BIOMASS AND FOOD ABUNDANCE

A guildanalysis based on the sumsof fractionalguild memberships of all species in a communityhas a numberof usefulapplications for communityand population ecology. First, it providesa new and promisingquantitative approach to the elusiveproblem of consumer-resourcerelationships, relationships that underlie manybasic and popular hypotheses of competitionand populationregulation but remainessentially untested in naturalsituations. Measurements of the food supply availableto a selectedspecies are rarely attemptedin nature becauseof the near impossibilityof identifyingand quantifying all the preyorganisms under the highly complexand seasonally changing conditions of naturalenvironments. The foraging guild/guildprovince approach bypasses these difficultiesby transferringthe analysisto a levelwhere the resource component is a singleintegrated unit, a prey community,delineated by naturaland readilytraced spatial boundaries, and the consumercomponent is identified by definition in terms of those resource boundaries. 104 ORNITHOLOGICAL MONOGRAPHS NO. 24

The invertebratepopulations, i.e. prey communitiesfor insectivorousguilds, werecensused in theGrand Bahama pine forests by samplingthe foraging substrate in sevenspatially defined crown, trunk, shrub, and ground compartmentsin Januaryof 1969 and in one crowncompartment in Januaryand May of 1971. The 1969 datawere used for between-compartmentcomparisons in a singleseason, and the 1971 datafor between-seasoncomparisons in a singlecompartment. For the 1969 data Dr. NormanSloan and I cut and gentlylowered three typical pine trees of the canopystratum. All needlesand twigs were then clipped and transferredwith a minimumof disturbanceto a seriesof plasticbags, one for each of the recognizedcrown compartments. Representative segments of the trunksand brancheswere similarlycut and bagged.Shrub foliage and groundcover materials were clippeddirectly from the standingvegetation. All materialswere fumigated and shippedto Dr. Sloan'slaboratory in Michiganwhere all edible organisms were removed,identified to genusor family, and counted. Since our objectivewas to obtain comparableestimates of the standingfood suppliesin thesecompartments, and sincetime was a practicalconsideration, we did not attemptto weighthe collectionsbut simplycategorized the specimensin each compartmentsample into four size classes(>8 ram, 4-8 mm, 2-4 ram, <2 ram) and multipliedthe numberof specimensin eachclass by 4, 3, 2, and 1 respectively.The sum of the four products,while in no way representingprey biomass,gives a singlevalue in "insectunits" suitable for cross-guildcomparisons. I also assignedspecimens to one of six generalmorphological types designated "moths," "flies" (both Hymenopteransand Dipterans), "bugs" (including Cole- opterans,Orthopterans, etc.), "spiders" (including centipedesand pseudoscor- pions), "ants,"and "larvae" (includingpupae and eggs). Although extremelycrude from the point of view of an insect taxonomist,I consideredthese categories, based on generalform, activity,and approachabilityby birds, more appropriatefor the objectivesof the analysisthan the usuallycited phyletic taxa. One of the morphologicaltypes, "ants," was omitted from the tabulationson the premisethat antsare largelyunacceptable as food by the foliage• and bark-gleaningbirds under consideration. For the between-seasonspine foliage comparisonof 1971 Virginia Emlen and I collectedsamples of foliage directly from the lower outer crown compartment of 20 canopystratum pines. A metal hoop 50 cm in diameterbacked by a large plasticgarbage bag and mountedat the end of a 4-m pole wasquietly slipped over the end of each selectedbranch. The branch was then clippedby cord-operated pruningshears and the foliagesample in the bag loweredand tied off for temporary storagewithout fumigation. From 5 to 10 brancheswere taken from each of 20 trees in 2 stands. We examined all foliage sampleswithin three hours after collecting,vigorously shaking the branchesand needleclusters over a large white tub into whichmost of the organismsof edible sizefell and from which they were removedand storedin specimenvials for subsequentclassification and identifica- tion. Twigs and needleswere alsoexamined closely after the shakingprocedure for persistingspecimens. The same systemof classificationinto four size classes andsix morphological categories was used as that appliedin the between-compartment study. 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 105

Avian biomassin g per km2 is plotted againstprey densityin insectunits per m3 (between-compartmentfoliage comparisons), cm 2 (between-compartmentbark comparisons),or kg of foliage (between-seasonsfoliage comparisons)in Figure 31 (p. 66). The trend of increasingavian biomasswith increasingfood supply predictedby the traditional consumer-resourcemodel was not found in these comparisons.Among the three pine canopy compartments,avian biomasswas lowestin the crown tops where insectpopulations reached their highestdensities (Fig. 31A). The upper portionsof tree trunks had fewer insectsbut supported more birds than the lower portions (Fig. 3lB). In the interseasoncomparison wherethe collectionswere most extensiveand the samplingwas best distributed, the winteringcommunity had nearly 50% more avian insectivorebiomass at the time when the resourcebase was lowest (Fig. 31C). Food supply is generallyregarded as a key factor in prevailing theoriesof populationregulation and relatedconcepts of intraspeciescompetition and carrying capacity (Lack 1954, Wynne-Edwards1962). It also plays a central role in modern theoriesof interspeciescompetition, ecological segregation, ecological release,character displacement, and evolutionarydivergence. The failure of the observationsdescribed above to supportthe predictionsof competitiontheory therefore demands critical attention. There are, of course, factors other than food to be considered, and the foragingguilds of the Bahamianpine forestsmay be atypicalor otherwiseinap- propriate for testingconsumer-resource models. Also, at a more specificlevel, factorsother than food supplymay have affectedthe attractivenessof someof the compartmentsor substratesused in the between-compartmentcomparisons. For instance,the failure to conformto the predictionsof prevailingtheory in Figure 31A andB couldbe attributedto structuralor positionalfactors such as foliage density,foliage type, or proximityof escapecover on the trunksor ground. The comparisonin Figure 31C betweenwinter and summerratios in a singletree crownsubcompartment appears to avoid thesecomplicating factors. With sub- strateaccessibility and exposureto predationequated by identicalconditions of foliageand position, a seasonalinflux of invaderstheoretically could not occuron an equalor reducedresource base unless the initial (breedingseason) populations were below capacity. Althoughmy data are admittedlysketchy, and all variablesare obviouslynot accountedfor, I considerthat the failureof the data to supportthe predictionsof popular competitiontheories suggeststhat factors other than food levels are importantlyinvolved in populationregulation in at least someof the guildsof the Grand Bahamapine forests.This view is alsosupported by observationsmade in 1971 comparingconsumer-food supply ratios in the Grand Bahamapine forests with thosein closelysimilar pine standsin Florida where populationdensities weremuch lower (Emlenin press)and by observationsmade in 1976 comparing compartmentdistributions of the pine tree insectivoreson Grand Bahama with thoseon neighboringAndros Island (Emlen MS).

GUILD STRUCTURE AND COMMUNITY DYNAMICS A secondapplication of the summedfractions approach to foragingguilds is to tracethe ecological distributions of eachspecies within a communityand quantify 106 ORNITHOLOGICAL MONOGRAPHS NO. 24 the areasand extentof ecologicaloverlap between species. Extensive overlap in the exploitationof food resources,foraging substrates,and foraging behavior implies more frequent interactionsbetween species and, at least potentially,in- creasedcompetition for resources,a factor that is traditionallyof central concern in considerationsof communitydynamics. Guild data from the Grand Bahama pine forest communitieswere analyzedfor informationon guild structureand diversity,species distributions, and speciesoverlap. Guild structureand diversity.--Theguilds of a bird communitymay differ in the numberof speciesthey contain, the equitabilityor evennessof representationof these species,and bird speciesdiversity, a combinationof speciesrichness and equitability. These indices,especially equitability, are potentiallyuseful as indi- catorsof the amountof resourcesharing and of speciesinteraction rates in the variousguilds. Values are givenin Table 26 for each of the 18 foragingguilds of the Grand Bahamapine forest. Thesevalues have little biologicalusefulness for direct, between-guildcomparisons since their resourcebases or provincesvary in size and richnessand cannot be equated. They may be useful, however, in between-elementcomparisons. They reveal, for instance,that PRs as a group weremore evenlydistributed (higher equitability) than WRs in 9 of the 10 guilds with multiplemembership. The equitabilityindex could also have somepractical utility in revealingwhere my systemof local guild categorizationcould profitably be refined. For instance,the highequitability value in the bark-gleaninginsectivore guildof thewintering community in Table26 (J' = 0.91) suggeststhat two rather distinctforaging categorieswere presentthat could usefullybe recognizedas separateguilds (also seeFig. 38). No suchsplitting is indicatedfor the ground insectgleaners where J' -- 0.38. Dispersionof speciesthrough the guilds.--As alreadynoted, most bird species in the Grand Bahama submatureforests distributed their foraging activitiesover severalguild provinces.The extentof dispersionthrough the systemis shownfor eachspecies in an indexof amplitudein the last columnof Table 20. Theseindices show the proportionof the maximumpossible diversity of distributionthat a speciesachieves (J' = H'/H' max.), i.e. it is a measureof the evennessor equitabilityof distributionthrough the 18 guildswhere the maximumnumber of guilds for a speciesis limited to 10 by the systemof allottingrepresentation in tenths. Trace representationswere not consideredin thesecalculations. Low values,such as thosefor the two doves (0.00 and 0.11), the Bahama Swallow (0.00), or the Black-and-whiteWarbler (0.07) indicate high specializationin one or a few of the guilds;high values such as those for the gnatcatcher(0.51), the PalmWarbler (0.80), or the bananaquit(0.50) indicatewide dispersalof activitythrough many guild provinces. The dispersionof a speciesthrough the systemof guilds (high amplitude) is attributedprimarily to ecologicaland trophicplasticity. A populationsaturation- overflowmodel suchas that appliedto the betweenhabitat analysesin chapter5 doesnot seemappropriate at the guildlevel, but guilddispersion may be facilitated by populationpressures in somesituations. It mightwell havebeen operating, for instance,in the Palm Warblersthat assumedthe dominantposition in no lessthan 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 107 108 ORNITHOLOGICAL MONOGRAPHS NO. 24 four of the guildsof the winteringcommunity (Fig. 38 and Table 23A). Spatial and/or functionalproximity of the foragingsubstrates were probablykey factorsin this caseand in four instancesof doubledominance in the breedingseason community (Table 23B), the Cuban Emerald Hummingbird(nectar-sipping and flower insectprobing), the Greater Antillean ?ewee (air sallyingand ground pouncing),the Brown-headedNuthatch (bark gleaningand coneprobing), and the Olive-cappedWarbler (needlegleaning and pine twig gleaning). Specializationand generalization(low and high dispersionamplitudes) were widely and rather evenly distributedthrough the guilds,with specialistsand generalistsoften occurring together in the sameguild. Defininga specialistas a specieswith a guilddistribution amplitude of lessthan 0.23 and a generalistas one with an amplitudevalue of more than 0.43, winter residentshad about the same numberof specialists(6) as permanentresidents (5) and similarmean amplitudes (0.35 and 0.32). Summerresidents were apparentlymore specialized,with a mean amplitudeof 0.13, but in view of the smallnumber of speciesin the element(3), no significanceshould be attachedto the differenceat thistime. Another comparisonof interest is that between speciesof continentaland Antilleanorigin. Speciesthat colonizeislands, i.e. the Antillean elementin the Grand Bahamaavifauna, are theoreticallymore likely to be generalistsbecause of greateradaptability to novel conditions(MacArthur and Wilson 1967) and a tendencyto expandin ecologicalamplitude when released from the pressuresof high continentalbird speciesdiversity (Grant 1966, Lack 1970). On Grand Bahamaone might thus expectrelatively more generalistsand fewer specialists amongthe speciesof Antilleanorigin vis-•t-vis continental origin. The data do not supportthis prediction. Among the 7 speciesclassed as recently derived from continentalstock (Fig. 3), 3 were generalistsand 2 were specialists;among the 8 from Antillean stock2 were generalistsand 2 were specialists.The groupof 7 speciesclassed as comingto Grand Bahamafrom the continentby way of the Antilles contained1 generalistand 4 specialists. The top choiceof a speciesamong the foragingprovinces of a habitat (the guild of primarymembership for the species)should be a goodplace to look for evidence of segregationarising from interspeciescompetition for food resources.I would predictthat the speciesmost successfulin establishingthemselves in a community would be thosethat were preadaptedor that quicklybecame adapted to available foragingprovinces not alreadypreempted or heavilyexploited. This selectionof uncrowdedprovinces should, all elsebeing equal, lead to a rather even dispersion of primary membershipsthrough the local guild system. In contrast to this predi.ction,.the 35 primaryguild memberships in the Grand Bahamapine forests were concentratedin 6 of the 18 guilds,and 7 guildshad no primarymembers at all. Of the primarymemberships 25 fell in 5 of the guilds,5 in each;1 guildhad 4, 2 had 2, and 2 had 1 (Table 19). Specialistspecies as defined above were disperseda little more widelythan generalists(through 8 vs. 6 guilds),but 3 specialistspecies had their primary membershipssuperimposed in one guild (shrubfoliage gleaners), and therewere 2 doubleoccupancies in eachof 4 other guilds.Admittedly, the largeamount of cohabitationrevealed by thesedata may 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 109 reflect grossinequalities in my partitioningof the functionalspace, but even allowinggenerously for this, the predictionsof the hypothesisof segregationand dispersionof primary membershipsin functionalspace do not seemto be met. Predictionsof general competitiontheory also failed to find support from observationsof distributionadjustments by permanentresident species as winter migrants withdrew from foraging provinces they had dominated. Permanent residentcompensatory responses to the exodusof winter residentsexceeded 10% of the lossin only one guild;permanent residents actually declined by 22% as the migrantsleft in one guild (Table 26).

SPECIES OVERLAP AND COMPETITION

Insofar as the guild provincesadopted for this analysisapproximate natural subdivisionsof the forest'savailable food resources, the extentof overlapin guild membershipbetween two speciesreflects the extent of joint use of a common foodsupply by thosespecies. Similarly, the sumof the overlapsbetween a species and all the other membersof the communityindicates the extent to which that speciesshares its food resourceswith the rest of the community.Such joint useof commonresources may be equated with potential competition,theoretically becomingreal competitionwhen and where suppliesare limited with respectto total demand. Indices of guild overlap for all of the 526 two-speciescombinations of the winteringcommunity and the 211 combinationsof the breedingseason community are presentedin Tables27 and 28. The valueswere obtainedby summingthe tenthsof overlapin eachcolumn of Table 20 anddividing the total by the maximum scorepossible for all overlaps--alwaysten. For example,the Cuban Emerald Hummingbirdand the Bananaquitin the winteringcommunity are recordedas overlappingto the extent of 0.4 in the nectar-sippingguild, 0.2 in the flower- probingguild, and 0.1 in theneedle-gleaning guild. With 10 tenthspossible in cases of completeoverlap (perfect coincidence),the index of overlap for these two speciesis 7 tenths,as shownat the propercoordinate point in Table 27. Theoreticallycompetition will tend to suppressoverlap and preventthe occurrenceof highoverlap. The data do not supportthis prediction. Mean overlapfor all speciespairs in the winteringcommunity was 1.58, appreciablyhigher than the 1.06 overlapexpected by chance.The comparablevalues for the breeding season were 1.20 and 0.69. The concentrationof relativelyhigh overlapvalues along the diagonalmargin of the triangleof valuesin Tables27 and 28 reflectsa similarityof foraging behaviorin relatedspecies which, because of theroughly phylogenetic arrangement of the list of speciestend to have their coordinatepoints grouped in this zone. Withinthis zone, particularly high values are clusteredwhere the sequentiallylisted membersof naturaltaxonomic units (e.g. the flycatchersor the Dendroicawar- biers)form associatedblocks of coordinatepoints. This distributionsupports the traditionalview that specieswith similarforaging structures, particularly the bill, tend to have similarfeeding habits. Data are insufficientto test the prediction of competitiontheory that closely related congenerics tend to segregateecologically. 110 ORNITHOLOGICAL MONOGRAPHS NO. 24

z• o •: •ll• 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 111

TABLE 28 OVERLAP IN GUILD MEMBERSHIP DISTRIBUTION IN THE 21 MEMBERS OF THE BREEDING SEASON COMMUNITY OF THE PINE FORESTS•

Z G CH G L S G B B R B T BY O P B B SG P ZenaidaDove Instancesof majoroverlaps between species P Ground Dove 9 P CubanEmerald OverlapPRXPR PRxSI SIXSI Tot. Hummingbird 0 0 9 (90%) 2 2 0 4 P HairyWoodpecker 0 0 0 8 (80%) 2 1 0 3 S GrayKingbird 0 0 0 0 7 (70%) 3 1 0 4 P Loggerhead 3-6 17 1 0 18 Flycatcher 0 0 0 0 8 1-2 59 11 0 70 P Stolid Flycatcher 0 0 1 0 9 7 None 87 23 2 112 P Greater Antillean Pewee 0 0 0 0 9 8 8 Mean- S Bahama Swallow 0 0 0 0 0 0 0 0 Recorded 1.14 1.49 0.00 1.20 P Brown-headed Expected 0.63 1.07 0.00 0.69 Nuthatch 0 0 1 6 0 0 1 0 0 PRed-leggedThrush 4 3 0 2 0 0 0 0 0 0 P Blue-gray Gnatcatcher 0 0 1 3 0 0 1 1 0 4 0 P Thick-billed Vireo 1 0 1 1 1 1 1 2 0 1 1 4 P Black-wiskered Vireo 1 0 0 1 0 0 0 1 0 1 1 4 7 P Yellow-throated Warbler 0 0 1 5 0 0 1 0 0 7 0 4 1 1 P Olive-capped Warbler 0 0 1 1 0 0 1 0 0 2 0 5 1 1 2 P Pine Warbler 0 0 1 1 0 0 1 0 0 1 0 5 1 0 2 9 P Bahama Yellowthroat 0 0 3 0 0 0 0 1 0 0 4 1 1 1 0 0 0 PBananaquit 0 0 7 0 0 0 1 1 0 1 0 4 3 2 1 2 2 4 P Striped-headed Tanager 1 0 0 2 0 0 0 0 0 0 3 1 2 1 0 0 0 1 1 PGrassquit 0 2 1 1 0 0 1 0 0 1 4 1 2 0 1 1 1 0 1 1 Letters at column heads match the initial letters in the list of specieson left margin.

Only oneof the 11 casesof high overlap(7-9 tenths)in the breedingcommunity involvestwo congenerics,however, and here the two (Olive-cappedWarbler and Pine Warbler) are not very closelyrelated. Becauseof differentamounts of exposureto interactionone mightpredict that pair combinationsinvolving 2 permanentresident species (interacting through the year includingthe breedingseason) would show more divergence(have a lower overlap)than combinationsof 2 winterresident species or combinationswith one permanentresident and one winterresident species. Data in Table 27, summarized in the box at the upperright, generallysupport this prediction.The mean overlap valueof PR x PR pairingswas 1.22, whilethat for WR x WR pairingswas 2.02, and for PR x WR pairings,1.61.

ACKNOWLEDGMENTS

Most of the work reportedin thispaper was conducted with supportfrom grants from the Universityof WisconsinGraduate School and the National Science Foundation(GB-15304). Transportationand operatingexpenses were covered primarilyby grantsfrom the AmericanPhilosophical Society and the Frank M. ChapmanMemorial Fund. Facilitiesand logisticassistance on the Islandwere 112 ORNITHOLOGICAL MONOGRAPHS NO. 24 providedby the ColonialResearch Institute of Freeport,Grand Bahama. I partic- nlady thank Virginia Emlen, William Gillis, Paul Fluck, Dorothy Rand, and Norman Sloanfor assistanceon variousaspects of the field work. Many helpful commentsand criticismson the manuscriptwere offered by Edward Beals, J. Merritt Emlen, Timothy Moermond, Ronald Pulliam, and William Robertson. Usefulsuggestions on specialtopics were offeredby JamesBond. The patienceand artisticskills of CherylHughes and Donald Chandlerare gratefullyacknowledged.

SUMMARY AND CONCLUSIONS

Introduction, Methods, and Background.--An intensive study of ecological distributionwas conductedon the land bird populationsof Grand Bahama Island, a 1200 km2, low-lying,pine-covered island located 105 km off the east coast of southernFlorida. Shrunkenby risingsea levels from about 15,000 km: duringthe last pleistoceneice age,the islandnow contains33 breedingland birds (excluding birdsof prey). About one-halfof thesederive from the Antilles, one-fourthfrom Central America via the Antilles, and one-fourth, mostly pine forest species, directlyfrom North America. Absolutedensity and biomassdata were obtainedfor across-habitatdistribution analysesfor all species,residents and migrants,at 25 sitesrepresenting the range of habitat conditions on the island. More intensive studies of within-habitat distributionwere conductedat three sitesin the dominanthabitat type, submature pine forest. The Habitats and Their Bird Communities.--Proceduresfor delineating,measuring, andanalyzing the structureand diversity of habitatsare outlinedand applied to the 25 stands. Techniquesfor plottingbird distributionsalong single and compoundhabitat gradients are describedand applied.Prominent aspects of the composition,structure, diversity, and density of thewintering bird communitiesin the standsare examined and compared between stands and types and along habitat gradients. Bird speciesdiversity is shownto be correlatedwith severalgraded features of habitatstructure and with vegetationdiversity as measuredby the relativequantity of foliageat 3 levels(foliage height diversity). The bestcorrelation found for bird speciesdiversity was, however, with totalvegetation volume, measured as cubic metersof standingfoliage per km2. The logicalbasis for bird diversity-habitat complexityrelationships is discussedand the demonstratedbest correlation with vegetationvolume attributed to complexityfactors associated with vegetation volume as it was measured. Total communitydensity, the sum of all speciesdensities, showed definite trendsalong several habitat gradients, particularly those of increasingtree and shrubcover. Community density was, however, only weakly correlated with total vegetationvolume. The logicalbasis for a totalbird density-vegetationvolume correlationis discussedand the demonstratedpoor correlationon Grand Bahama citedas a basisfor cautionagainst uncritical acceptance of widelyheld assumptions concerningconsumer-resource relations in nature. HabitatDistribution.---Concepts of the dynamicsof densitydistribution through a mosaicof habitatpatches are reviewed.Models are presentedin whichas yet 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 113 unestablishedbirds are viewed as moving into the physicallybest patchesuntil (1) factors associatedwith increasingdensity reduce the net quality of those patchesor (2) aggressiveresistance by residentsterminates further immigration. Density distributionthrough a seriesof habitat patchesof decreasingphysical quality is visualizedas progressivelydeclining in the first instanceand holdinga constantlevel until carryingcapacity falls below social saturationin the second. The patternsof densitydistribution through the top ten standsfor each Grand Bahama specieswere comparedwith these models. Relatively aggressiveand localizedspecies showed more evidenceof a socialsaturation plateau in these comparisonsthan nomadic and passive species, suggesting that the socialsaturation phenomenon,well knownin breedingseason territorial behavior, was also operating in somemembers of this winteringcommunity. In a plottingof maximumdensities against dispersion amplitudes, species with high absolutedensities in their preferredhabitat types were always dispersed broadlyinto secondarytypes. Thesedispersals are tentativelyattributed to overflow from capacitypopulations. Species with low absolutedensities in their preferredhabitats and narrow dispersalamplitude are tentativelyregarded as speciesthat had not filled their preferred habitat to either social saturationor resource-determinedcarrying capacity levels. Those with low absolutedensity andbroad dispersal amplitude are regardedas speciesthat had dispersedbecause of (for them) low densitytolerances or low capacityhabitats. Winter densitiesvaried greatly amonghabitat types. Permanentresidents as a group outnumberedwinter residentsin the forest habitatsbut were outnumbered by them in the open habitats. Densitydistributions for most specieswere unexpectedlyirregular and diffuse along the seven selectedparameter gradients. Permanentresidents increased relative to winter residentswith increasingpine cover, decreasingtree height and increasingvegetation volume, but no evidenceof seasonaldisplacement of one groupby the other was detectedin any habitat. Overlap in habitat distributionwas no greaterin closelyrelated than remotely related speciesexcept within one group of congenericpine habitat warblers. It was no greaterin pairingsof continentalor of Antillean speciesthan in mixed geographicpairings, but was greater in pairings of permanentresidents and of winter residentsthan in mixed PR-WR pairings,supporting a predictionthat species in the two seasonalelements would tend to segregate. A two-dimensionalordination of habitat spacingfor all the speciesof the Bahamasw. intering communityprovides a base for evaluatingthe similarity of habitat selectionresponses between the species. The Pine Forest Community--SeasonalChanges.--The bird fauna occupying the pine forest, the most extensiveand most complex of the Grand Bahama habitats, changed markedly in composition,diversity, and biomass between January and June, as winter residentsdeparted, summer residents arrived, and transientspassed through. The wintering community,divided almost equally betweenpermanent residents and winter migrants,was nearly twice as large in numbersof species,individuals, and biomassas the breedingseason community. Summerresidents and transients were minor elements in theseseasonal changes. 114 ORNITHOLOGICAL MONOGRAPHS NO. 24

The greatannual increase in numberof specieseach fall occursin the absence of any obviouschange in habitat complexity,and the near doublingof total populationdensity and biomass,in the apparentabsence of any increasein basic foodresources. The impactof increasedspecies numbers on thepermanent resident populationsmay be dulledby a low equitabilityin thewinter migrant element. The impactof increaseddensity and biomass could be moreimagined than real if, as I propose,food supplyis lesscritical as a limitingfactor than commonlysupposed. SpatialDistribution Within the Pine Forest.--The spacewithin the pine forest habitatwas partitioned into 5 equallayers from the groundsurface to the tree tops, and 10 (or 4 for someanalyses) unequal compartments delineated with reference to prominentrecurring features of the forest vegetation. Speciesdiversity and total densitydeclined gradually and evenlyupward through the five equallayers, while foliagedensity was high in the first, third, and fourth layersand low in the secondand fifth. Bird distributionthus appearedto reflect responsesto spaceper se ratherthan to the quantityof perchingand foragingsurfaces.Winter resident species were relatively best represented in the subcanopyand lower canopy,the third and fourth layers. Graphic representationsof compartmentdistributions reveal preference,dispersion,and overlapcharacteristics of 30 commonspecies and variousspecies groups. A comparisonof the compartmentdistribution of permanent resident species before and after the annual departure of winter migrants (i.e. the winter and breedingseason distributions) reveal considerableshuffle between compartments andhabitats including a net shiftfrom the shrubcompartment into the crowns and downto the groundvegetation. The directionof this movementis oppositeto that predictedby modelsof compensationadjustment and traditionalcompetition theory. Guild Distribution Within the Pine Forest.--In this chapter the ecological distributionof each pine forestspecies is delineatedin termsof the substratesand resourcesits membersexploit and the way they exploit them. For this purpose the forest habitat of Grand Bahama was divided into 18 functionallydefined foragingprovinces and the avain communityinto 18 matchingforaging guilds. Local observationsof foraging behavior were used to assigneach bird species fractionallyto as many of theseguilds as it was seento enter. Most of the guilds thus containedone or more primary or specialistmembers and a number of secondarymembers and visitors. Areal densitiesfrom the transectcensus were multipliedby membershipvalues in tenthsto give directly comparableabsolute densitiesfor each guild and each guild member. The structure,diversity, and densitycharacteristics of each of the 18 Grand Bahamapine forest guildsare described. Invertebratefood resourceswere sampledquantitatively in five pine foliage compartments,two understorycompartments, and two trunk compartments.The prey organismdensities calculated from thesesamples were then matchedwith the measured biomass of avian consumersin those compartmentsfor an across- provincecomparison of consumer-resourceratios. Resultsfailed to show the positivecorrelation predicted by traditionalcarrying capacity theory. In another test the invertebratefood resourcesof a single pine foliage compartmentwere 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 115 sampledbefore and after the winter migrantsdeparted from the island, and again there was no positive correlationof avian biomasswith food supply. An H' index of the dispersionamplitude of speciesthrough the 18 guild provincesof the forest provided a basis for defining levels of specializationin foraging behavior. The winter residentand permanentresident elements had similarmean specialization values, but the winter residentelement was more varied. A prediction of relatively high amplitude in speciesof Antillean origin based on attributescommonly associated with colonizationand low diversityisland communities was not supported.A predictionthat provincepreference should be widely distributedamong species(to avoid crowdingin a few provinces)was not supported. Contrary to the predictionsof competitiontheory no generalcompensatory shift in provincedistribution by permanentresident species was detectedwith the springdeparture of winter migrants. Measuresof total guild overlap (reflectingjoint resourceuse) for the members of each speciespair combinationshowed an overall incidenceappreciably higher than that expectedby chance. This, again, appearsto be in oppositionto the predictionsof traditionalcompetition theory. High overlapwas correlatedwith phylogeneticrelationship at the family level. A few instancesof very high overlap (70-90%) occurredbetween congenerics. In line with competitiontheory pre- dictions,overlap was lessin pairingswithin the permanentresident element than in pairingsof permanentresidents with winter residentsor pairingsbetween two winter resident species.

LITERATURE CITED

A•DmCH, J. 1963. Geographic orientation of American Tetraonidae. J. Wildl. Mgmt. 27: 529-545. ALT, D., ANDH. K. BROOKS. 1965. Age of the Florida marine terraces. J. Geol. 73: 406- 411. AMEmCAN Om•XTHOLOGISTS'UNION. 1957. Check-list of North American birds, fifth ed. Baltimore, Amer. Ornithol. Union. --.. 1973. Thirty-second supplement to the American Ornithologists' Union checklist of North American birds. Auk 90: 411-419. BEALS,E.W. 1960. Forest bird communities in the Apostle Islands of Wisconsin. Wilson Bull. 72: 156-181. --.. 1973. Ordination: mathematical elegance and ecological naivet•. I. Ecol. 61: 23-35. BENT, A.C. 1939. Life histories of North American woodpeckers. U.S. Natl. Mus. Bull. No. 174. 1940. Life histories of North American cuckoos, goatsuckers,hummingbirds and their allies. U.S. Natl. Mus. Bull. No. 176. ---. 1942. Life histories of North American flycatchers, larks, swallows and their allies. U.S. Natl. Mus. Bull. No. 179. 1948. Life histories of North American nuthatches, wrens, thrashers and their allies. U.S. Natl. Mus. Bull. No. 195. 1953. Life histories of North American wood warblers. U.S. Natl. Mus. Bull. No. 203. 1968. Life histories of North American cardinals, grosbeaks, buntings, towbees, finches, sparrowsand their allies. U.S. Natl. Mus. Bull. No. 237. BoN•, I. 1963. Derivation of the Antillean avifauna. Proc. Acad. Nat. Sci. Philadelphia 115: 79-98. 1971. Birds of the West Indies. Boston, Houghton-Mifflin Co. 116 ORNITHOLOGICAL MONOGRAPHS NO. 24

BRAY,J. R., AND J. T. CURTIS. 1957. An ordination of the upland forest communitiesof southernWisconsin. Ecol. Monogr. 27: 325-349. BRITTON, N. L., AND C. F. MILLSPAU•3H. 1920. The Bahama flora. New York, Hafner Publ. Co. BRODY:ORB,P. 1959. Pleistocene birds from New Providence Island, Bahamas. Bull. Florida State Mus. 4: 349-371. BROWN,J.L. 1969. Territorial behavior and population regulation in birds. Wilson Bull. 81: 293-329. CHAPMAN,F.M. 1891. The origin of the avifaunas of the Bahamas. Amer. Naturalist 25: 528-539. COPY, M. L. 1968. On the methods of resource division in grasslandbird communities. Amer. Naturalist 102: 107-147. CRITCHFIELD,W.B. 1966. Geographic distribution of the pines of the world. U.S. Dept. Agr. Misc. Publ. 991. DIAMOND, J. M. 1969. Avifaunal equilibria and species turnover rates on the channel islands of California. Proc. Natl. Acad. Sci. 64: 57-63. 1970. Ecological consequencesof island colonization by southwestPacific birds. Part 2, The effect of speciesdiversity on total population density. Proc. Natl. Acad. Sci. 67: 1715-1721. EMLEN, J. T. 1956. A method for describing and comparing avian habitats. Ibis 98: 565-576. 1967. A rapid method for measuringarboreal canopy cover. Ecol. 48: 158-160. 1971. Population densities of birds derived from transect counts. Auk 88: 323- 342. 1972. Size and structure of a wintering avian community in southern Texas. Ecology 53: 317-329. 1973. Territorial aggression in wintering warblers at Bahama agave blossoms. Wilson Bull. 85: 71-74. 1978. Density anomalies and regulation mechanisms in land bird populations on the Florida peninsula. Amer. Nat. 112: in press. in preparation. Densities and foraging distributions of birds on two Bahama islands. FRETWELL,S. D. 1972. Populationsin a seasonalenvironment. Princeton, New Jersey, Princeton Univ. Press. --, AND H. L. LUCAS,JR. 1970. On territorial behavior and other factors influencing habitat distribution in birds. Part 1, Theoretical development. Acta Biotheoretica 19: 16-36. GIBB, J. A. 1960. Populations of tits and goldcrests and their food supplies in pine plantations. Ibis 102: 163-208. GILLIS, W. T., R. BYRNE, AND W. HaRmSON. 1975. Bibliography of the natural history of the Bahama Islands. Atoll Res. Bull. 191. Ge,ANT, P.R. 1966. The density of land birds on Tres Marias Islands in Mexico. Part 1, Numbers and biomass. Canadian J. Zool. 44: 805-815. HILDEN, O. 1965. Habitat selectionin birds. Ann. Zoologici Fennici 2: 53-75. HOWELL,T.R. 1972. Birds of the lowland pine savannaof northeasternNicaragua. Con- dor 74: 316-340. HUTCHINSON,G. E. 1957. Concluding remarks. Cold Spring Harbor Symp. Quant. Biol. 22: 415-427. JOHNSTON,D.W. 1975. Ecological analysisof the Cayman Island avifauna. Bull. Florida State Mus. 19(5). KALE, H. W. 1967. Aggressivebehavior by a migrating Cape May Warbler. Auk 84: 120-121. KAR•, J. R. 1968. Habitat and avian diversity on strip-mined land in east-central Illinois. Condor 70: 348-357. , AND R. R. ROTH. 1971. Vegetation structure and avian diversity in several new world areas. Amer. Naturalist 105: 423-435. 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 117

KLUYVER,H. N., ANI) L. TINBERGEN.1953. Territory and the regulationof densityin titmice. Archiv. Need. Zool. 10: 265-287. LAc•c,D. 1933. Habitat selectionin birds. J. Anim. Ecol. 2: 239-262. 1940. Habitat selectionand speciationin birds. Brit. Birds 34: 80-84. 1954. The natural regulationof animal numbers. London, Oxford Univ. Press. 1970. The numbersof bird specieson islands. Bird Study 16: 193-209. •-. 1976. Island biology illustratedby the land birds of Jamaica. Berkeley, Univ. of Calif. Press. LYNCH,J. F., ANI) N. K. JOHNSON.1976, Turnover and equilibria in insular avifaunas, with specialreference to the California Channel Islands. Condor 76: 370-384. MACARTHUR,R.H. 1959. On the breeding distributionof North American migrants. Auk 76: 218-228. , ANDJ. W. MACARTHUR. 1961. On bird speciesdiversity. Ecology42: 594-598. , •q• E. O. WILSON. 1967. The theory of island biogeography. Princeton, New Jersey,Princeton Univ. Press. , H. RECHER,ANI) M. COl)y. 1966. On the relation between habitat selectionand speciesdiversity. Amer. Naturalist 100: 319-332. M•mTIN, A. C., H. S. ZIM, ANI) A. L. NELSON. 1951. American wildlife and plants. New York, McGraw-Hill. MILLIMAN, J. D., ANDK. O. EMERY. 1968. Sea levels during the past 35,000 years. Science 162: 1121-1123. MORF_AU,R.E. 1972. The Palaeartic-African bird migration system. New York, Academic Press. MOREL, G. 1968. Contribution h la synecologiedes oiseaux du sahel senegalais.Paris, Mem. Office de la R6chercheScientifique et Technique outre-Mer, 29. MORSE, D. H. 1966. Foraging relationships of Brown-headed Nuthatches and Pine War- blers. Ecology 48: 94-103. ORGANS,G. 1969. The number of bird speciesin some tropical forests. Ecology 50: 783- 801. RECHER,H. 1969. Bird species diversity and habitat diversity in Australia and North America. Amer. Naturalist 103: 75-80. RICKLEFS,R. E., ANDG. W. Cox. 1972. Taxon cyclesin the West Indian avifauna. Amer. Naturalist 106: 195-219. ROBERTSON,W. B., JR. 1955. An analysis of the breeding bird populations of tropical Florida in relation to the vegetation. Unpublished Ph.D. Thesis, Champaign, Univ. Illinois. •.. 1962. Observations on the birds of St. John, Virgin Islands. Auk 79: 44-76. ROOT, R. B. 1967. The niche exploitation pattern of the Blue-gray Gnatcatcher. Ecol. Monogr. 37: 317-349. ROTH, R.R. 1976. Spatial heterogeneityand bird speciesdiversity. Ecology 57: 773-782. SCHOLL,D. W., F. C. Ce•mUEnl>, SR., ANI) M. STraYER. 1969. Florida submergencecurve revised: its relation to coastal sedimentation rates. Science 163: 562-564. SCHW•TZ, A., ANI) R. F. KLINnCOWSKI.1963. Observationson West Indian birds. Proc. Acad. Nat. Sci. Philadelphia 115: 53-77. SVXm)SON,G. 1949. Competitionand habitat selectionin birds. Oikos 1: 157-174. TERBORGH,J., AND J. FAABORG. 1973. Turnover and ecological release in the avifauna of Mona Island, Puerto Rico. Auk 90: 759-779. TINBEROEN,N. 1957. The functions of territory. Bird Study 4: 14-27. TOMOFF,C.S. 1971. Arian speciesdiversity in desert scrub. Ecology 55: 396-403. WATTS, W. A. 1971. Postglacial and interglacial vegetation history of southern Georgia and central Florida. Ecology 52: 676-690. WETMORE,A. 1916. Birds of Puerto Rico. U.S. Dept. Agr. Bull. 32. WILLSON, M. F. 1974. Arian community organization and habitat structure. Ecology 55:1017-1029. WYNNE-El)walruS,V.C. 1962. Animal dispersionin relation to social behavior. Edinburgh, Oliver and Boyd. 118 ORNITHOLOGICAL MONOGRAPHS NO. 24

APPENDIX

SPECIES ACCOUNTS

The material presentedbelow provides a complete listing with annotations of the land bird speciesthat I observedalong the transectroutes and elsewhereon Grand Bahamaduring the monthsof Januarythrough May of 1968 and 1969. Most of the ecologicaland population data are taken directly from tables and figures in the main body of the text and assembledhere for ready reference under the species'names. Additional data on habitat distributionare presentedin graphicform in Figures 19, 20, and 21. Birds of prey, not includedin the formal list, consistedof a few winteringAmerican Kestr½ls (Falco sœarvcriu•), a small residentpopulation of Red-tailedHawks (Butco ]amaiccnsis),one Barn Owl (Tyro alba), and a few dozen of Turkey Vultures (½at•art½•aura). The symbolsdenoting geographic derivation are taken directly from Figure •, the letters indicatingthe source (C -- Continental,A = Antilles), and the subscriptnumbers, the amount of evolutionrecognized (0 = no changerecorded by taxonomists,1 m subspecific change, 2 m full specificchange). Geographicderivation is given for permanent resident and summerresident speciesonly. The coefficientof detectability(C.D. 412) for a speciesis the frequency(percent) with which birds were recorded in a band 824 feet (250 m) wide along the transect trails. Values were calculatedby plotting the lateral distributionof all detectionpoints along the trails, proi½ctingthe densityof detectionsin a narrow strip near the trail, where full coverage is assumed,to 125 m (the wide band), and dividing the projectednumber by the number actually recorded in the band (Emlen 1971). C.D. values were used to translate the trail countsper km of trans½ct(in the wide band) directly to density per km•. All calculated densitieswere convertedto metric units in this report. C.D. valuesfor the speciesin this study ranged from 10% to 65%; data were inadequateto calculatemeaningful C.D. values for the rarer species. Under the heading"status" I recordedthe seasonaloccurrence of each specieson Grand Bahama, whether a permanentresident, a winter resident, a summer resident, or an in-transit visitor (occurringonly as a migratingtransient) (Table 2). I then listed the mean density of the speciesfor all censustrans½cts during the winter period, 1 January to •1 March (Table 2). The rank positionof the common½stspecies is given in parenthesesfor the first 15. For rare speciesI indicated only the number of records obtained. For summer resident speciesI gave May densitieswhere these were available. Under habitat distributionI listed in order, the one or more (of seven) habitat types most heavily occupiedby the speciesand gave the densityin birds per km• for each of these types. In parenthesesI added a J• value (H•/H • max) indicatingthe amplitudeof dispersion of the speciesthrough the seventypes. Values range from 0 (found in only one habitat type) to 1.00 (evenly dispersedthrough all 7). Data are taken from Table $; they do not necessarilymatch thosein Table 18 which coversonly the • standsstudied intensively for within-habitat distribution. Distributionalong sevenselected habitat gradientsis shownfor most of the speciesin the diagramsin Figures19, 20, and 21. The small circlesin thesediagrams represent the positionof each of the 25 standsalong the specifiedgradients, and the symbol within the circlesgives the numericalstatus of the speciesin that stand. A key identifyingthe stands by position and explainingthe symbolsis presentedat the head of each figure. Further explanationsare given in the legend for Figure 11. Under "habitatoverlap" I listeda few speciesthat sharetheir habitatsmost heavily with the speciesunder consideration. The valueafter a species'name is the percentof overlap with the indicatedspecies through the 7 habitat types as taken from Figure 10. The number of specieswith overlapvalues greater than an indicatedhigh level is givenin parentheses. Under "verticaldistribution in pines"I simplyreferred the readerback to the species diagrams in Figure •2. Under "compartmentdistribution in pines"I named the maior compartments(among the 4 r½coguized:crowns, shrubs, trunks, ground) in which the speciesis bestrepresented duringthe winter seasonand gavethe calculatedindex of preferencefor that compartment 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 119

(4d/D) where the highest possiblescore is 4.0. The amplitude of dispersionthrough the 4 major compartmentsis presentedas a J' value (H'/H' max) in parentheses. Under "foraging guild distribution" I listed the one or several foraging guilds (among the 18 recognizedin Table 20) in which the speciesis best represented,and then gave a J' value reflecting the amplitude of dispersion of the species in question through the 16 guilds. J' values run from 0 (all foraging is in one guild) to 1.00 (even distribution). Under "foraging guild overlap" I listed the 2 or more speciesshowing greatest similarity to the one under consideration in their choice of food and foraging methods. The values are the sumsof tenths of overlap through the 16 recognizedforaging provinces (Table 27). The total of overlap with all other species,given in parentheses,is an indicator of the overall level of foraging interaction with other community members; values range from 0 in the highly specializedBahama Swallow to 88 in the eurythrophicPalm Warbler. White-CrownedPigeon, (Columbia leucocephala) Geographic derivation--A0. CD412--30%. Status--permanent resident, 0.27 birds per km -•. Irregular. Habitat distribution--coppets. 3.9 birds per km: (J'= 0). Habitat gradients---Figs: 19.1, 20.1, 21.1. Zenaida Dove, ( Zenaida aurita zenaida) Geographic derivation--A0. CD412--20%. Status---permanentresident, 3.4 birds per km•. Habitat distribution---coppets--5.0,submature piness4.2 (J• = 0.48). Habitat gradients---Figs: 19.2, 20.2, 21.2. Habitat overlap---Hairy Woodpecker--80, Cape May Warbler--79 (8 over 70%). Vertical distribution in pines---Fig. 32A. Compartmentdistribution in pines--ground--2.7, crowns----1.3(J• _-- 0.46). Foraging guild distribution--groundseeds----9 (J': 0.11). Foragingguild overlapsGround Dove--9, Red-leggedThrush---4 (total 25). Mourning Dove, ( Zenaida macroura carolinensis) Geographic derivation---CO. CD412--25%. Status---permanentresident, 1.68 birds per km-•. Habitat distribution--mangroves---15.0,submature pines--4.8 (J' = 0.28). Ground Dove, (Columbina passerina bahamensis) Geographic derivation--C1 A1. CD412--20%. Status---permanentresident, 4.1 birds per km 'ø. Habitat distribution--Coastalbrush---41.5 (winter), submaturepines--7.6 (J' = 0.33). Habitat gradient--Figs: 19.3, 20.3, 21.3. Habitat overlapsPalm Warbler--62, CommonYellowthroat--42 (3 over 30). Vertical distribution--Fig. 32B. Compartmentdistribution--ground--l.8, shrubs---1.6(J' _-- 0.69). Foraging guild distribution--ground seeds----10(J': 10). Foragingguild overlap•Zenaida Dove--9, Red-leggedThrush--3 (total 16). Key West Quail Dove, (Geotrygon chrysea) Geographic derivation--A0. CD412---10%. Status•permanent resident, 0.80 birds per km •. Habitat distribution---coastalbrush 25.5 (J' = 0). Habitat gradient--Figs: 19.4, 20.4. Yellow-Billed Cuckoo, ( Coccyzusa. americanus) Status---in-transitvisitor, 5 records,6 May (1968)-25 April (1969). Smooth-BilledAni, ( Crotophagaani) Geographic derivation A0. CD412--60%. Status--permanent resident, 0.85 birds per km :. 120 ORNITHOLOGICAL MONOGRAPHS NO. 24

Habitat distribution---coppets•10.5, marsh--5.7 (J' = 0.33). Habitat gradient--Figs: 19.5, 20.5, 21.4.

Chuck-wills-widow, ( Caprimulgus carolinensis) Status•winter invader, 6 recordsin March and April. Habitat distribution--submature pines.

Cuban Nighthawk, ( Chordeilesminor gundlachi) Geographic derivation--C2 A0. Status•summer resident (not recorded on transect counts), first spring record---29 April (1969).

Chimney Swift, ( Chaetura pelagica) Status•in-transit visitor, 4 records: 6, 7, 9 May, 1968, 27 April, 1969.

Cuban Emerald Hummingbird, ( Chlorostilbonricordii bracei) Geographic derivation--A1. CD412--14%. Status--permanentresident, 93.6 birdsper km• (2nd). Habitat distribution--youngpines•155, submaturepines--99, coppers--90 (I' = 0.81). Habitat gradient--Figs: 19.6, 20.6, 21.5. Habitat overlap--Bahama Yellowthroat--72, Greater Antillean Pewee--72 (6 over 60%). Vertical distribution--Fig. 27C. Compartmentdistribution--shrubs--2.4, crowns--l.3 (I' = 0.64). Foraging guild distribution--nectar--6, flower insects•3 (I' = 0.35). Foragingguild overlap--bananaquit--7, Cape May Warbler--3 (total 29).

Bahama Woodstar, (Calliphlox evelynae evelynae) Geographicderivation--A2. CD412---12%. Status•permanentresident, 21.2 birdsper km'• (15th). Habitat distribution--youngpiness72.5, coastalbrush--54.0 (I' = 0.59). Habitat gradient---Figs: 19.7, 20.7, 21.6. Habitat overlap---PalmWarbler--59, CommonYellowthroat--55 (5 over 40%). Belted Kingfisher, (Megaceryle alcyon) Status winter invader along waterways, 8 records. None seen on transcot routes.

Red-Bellied Woodpecker, ( Centurussuperciliaris) Geographicderivation--C1 A1. Status•permanent resident(Bond 1971). None recordedin 1968 or 1969.

Yellow-Bellied Sapsucker,(Sphyrapicus v. varius) CD412----22% Status•winter resident,3.2 birds per km2. Habitat distribution---coppers--22.0,submature pines---4.3 (J' = 0.23). Habitat gradient---Figs: 19.9, 20.9. Habitat overlap---Northern Waterthrush--89, Yellow-rumped Warbler--63 (7 over 40%). Compartmentdistribution---trunks 2.8, shrubs--0.8 (J' = 0.59). Foragingguild distribution--sapand cambium--5 (J' = 0.41). Foraging guild overlapsHairy Woodpecker--5, Brown-headedNuthatch--4 (total 38).

Hairy Woodpecker, ( Dendrocopusvillosus piger) Geographicderivation--C1. CD412---35%. Status•permanentresident 9.9 birdsper km-•. Habitat distribution---coppets•39.5,submature pines•14.1 (J' = 0.60). 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 121

Habitat gradient Figs: 19.8, 20.8, 21.7. Habitat overlapsStriped-headed Tanager--90, Black-and-white Warbler--87 (6 over 70%). Compartment distribution--trunks-•7.8, shrubs---0.6 (J' = 0.56). Foraging guild distribution--Bark gleaners4 (J' = 0.51). Foraging guild overlapsBrown-headed Nuthatch•6, Yellow-throated Warbler--5, Yel- low-bellied Sapsucker--5 (total 45). Gray Kingbird, ( Tyrannus d. dominicensis) Geographicderivation--A0. CD412--60%. Status•common summer resident (two winter records 10 January 1968, 14 March 1969). Habitat distribution--mangrove--0.80.Open pine woodlandsaround Freeport. Compartmentdistribution--(breeding season)-•crowns•2.5 (J' = 0.64). Foraging guild distribution--air insect sallyer--8, ground insectpouncer--2 (J'= 0.17).

LoggerheadFlycatcher, ( Tolmarchus caudi/asciatusbehamensis) Geographicderivation A1. CD412--30%. Status--permanentresident, 3.6 birdsper km2. Habitat distribution--submaturepines--6.7, coppets--3.9 (J• = 0.54). Habitat gradient--Figs: 19.10, 20.10, 21.8. Habitat overlapsGreater Antillean Pewee--89, Yellow-throated Warbler--89 (7 over 60%). Vertical distribution--Fig. 32D. Compartmentdistribution--crowns--3.6, trunks--0.4 (J' • 0.22). Foragingguild distribution--airinsect sailyet--6, groundinsect pouncer--4 (J' ----0.27). Foragingguild overlapsGreater Antillean Pewee--8, StolidFlycatcher--7 (total 28).

Stolid Flycatcher, (Myiarchus stolidus lucayiensis) Geographicderivation C2 A1. CD412---30%. Status--permanent resident,0.50 birds per km2. Habitat distribution---coppets-•3.5, submaturepines--0.3, (J' • 0.13). Habitat gradient Figs: 19.11, 20.11, 21.9. Compartment distribution-•crowns•2.4, shrubs--l.6 (J' ----0.49). Foraging guild distribution--air insect sailyet--8, ground insect pouncer--1 (J' ---- 0.22). Foraging guild overlapsGreater Antillean Pewee--8, Loggerhead Flycatcher--7 (total 41).

Greater Antillean Pewee, (Contopus caribaeusbahamensis) Geographic derivation--C2 A1. CD412--30%. Status•permanent resident, 7.6 birds per km2. Habitat distribufion--submature pines--11.0, young pines--9.3, coppets--9.0 (J' • 0.56). Habitat gradient Figs: 19.12, 20.12, 21.10. Habitat overlap--Yellow-throated Warbler--83, Pine Warbler--80 (7 over 60%). Vertical distribufion-•crowns--3.6, shrubs•0.3 (J' • 0.28). Foraging guild distribution--air insect sallyer--7, ground insect pouncer--2 (J' • 0.31). Foragingguild overlapsLoggerheadFlycatcher--8, Stolid Flycatcher--8 (total 42). Bahama Swallow, ( Callichelidon cyaneoviridis) Geographicderivation--C2. CD412--35%. Status--permanent resident, 6.8 birds per km:. Habitat distribution-•old fields--66.3, marsh--61.0, coastal•4.6 (J'= 0.42). Moved into submatureforests for breedingin late April. Habitat gradient--Figs: 19.13, 20.13. Habitat overlap---Mockingbird---54,Common Yellowthroat 50 (4 over 50%). Foraging guild distribution--air insectscreener--19 (J' • 0). 122 ORNITHOLOGICAL MONOGRAPHS NO. 24

Barn Swallow, ( Hirundo rustica erythrogaster) Status--in-transit visitor. One record, 13 May 1968.

Brown-Headed Nuthatch, ( Sitta pusilla insularis) Geographic derivation---C1. CD412---25%. Status--permanent resident, 8.48 birds per kmL Habitat distribution--submature pines--17.0 (J' • 0). Habitat gradient Figs: 19.14, 20.14, 21.11. Habitat overlapsCape May Warbler--62, Olive-cappedWarbler--55 (4 over 40%). Vertical distribution--trunks•2.1, crowns--l.9 (J• = 0.50). Foraging guild distribution--bark insectgleaner---4, pine twig gleaner--3 (J' = 0.49). Foraging guild overlap---Hairy Woodpecker--6, Yellow-bellied Sapsucker-•4 (total 49).

House Wren, (Troglodytesaedon ssp.) Status•winter resident,1 bird present,6-21 April 1969. Habitat distribution--submature pines.

Northern Mockingbird, ( Mimus polyglottosorphaeus) Geographic derivation--C1 A0. CD412--60%. Status--permanent resident, 4.15 birds per km 2. Habitat distribution---old fields--44.0, coastal-•4.6 (J'= 0.41). Common in Freeport suburbs. Habitat gradient Figs: 19.16, 20.16, 21.13. Habitat overlap--GrasshopperSparrow--83, Common Yellowthroat-•41 (5 over 30%). Vertical distribution--Fig. 32G. Compartment distribution--shrubs--l.7, crowns•l.2 (J' = 0.91). Foraging guild distribution--fruit and bud eater--6, ground insect pouncer--2 (J' = 0.45). Foraging guild overlap--Gray Catbird--7, Red-leggedThrush--5 (total 45).

Bahama Mockingbird, ( M imus g. gundlachii) Geographic derivation--A0. CD412--60%. Status•permanent resident. One record, 18 March 1969. Habitat distribution---old fields--l.07 (J' = 0). Habitat gradient--Figs: 19.17, 20.17, 21.14.

Gray Catbird, (Dumetella carolinentis) CD412•18% Status--winter resident, 58.3 birds per km: (Sth). Habitat distribution---coppets•251, old fields•249, marsh--54 (J' = 0.61). Habitat gradient--Figs. 19.15, 20.15, 21.12. Habitat overlapsPrairie Warbler--70, Bahama Yellowthroat--68 (4 over 55%). Vertical distribution--Fig. 32H. Compartment distribution--shrubs•3.0, crowns--0.4 (J' = 0.52). Foraging guild distribution--fruit and bud eater--5, ground insect gleaner--3 (J' = 0.44). Foraging guild overlapsMockingbird--7, Red-legged Thrush--7 (total 62). American Robin, (Turdus migratoriusspp.) CD412--50% Status--winter resident, irregular. Five records (15-16 March, 1969). Habitat distribution--marsh

Red-Legged Thrush, (Mimocichla p. plumbea) Geographicalderivation A1. CD412---18%. Status--permanentresident, 3.44 birds per km:. 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 123

Habitat distribution-•coppets•38.9, submaturepiness2.0 (J' = 0.20). Habitat gradient--Figs: 19.18, 20.18, 21.15. Habitat overlap-•Parula Warbler--95, Ovenbird 92 (7 over 70%). Vertical distribution--Fig. 32I. Compartment distribution--shrubs--l.7, ground--l.7 (J' = 0.72). Foraging guild distribution--ground insect gleaner 4, fruit and bud eater--3 (J' = 0.38). Foraging guild overlapsGray Catbirds7, Northern Mockingbirds5 (total 51).

Blue-Gray Gnatcatcher, ( Polioptila c. caerulea) Geographicalderivation--CO. CD412•25%. Status•permancnt resident,29.0 birds per km2 (10th). Habitat distribution-•coppets•153, submaturepiness31.0. (J' = 0.45). Habitat gradient--Fig. 19.19, 20.19, 21.16. Habitat overlap•redstart--93, Striped-headedTanager--90, Cape May Warbler--88 (8 over 80%). Vertical distribution Fig. 32J. Compartment distribution--shrubs•2.5, crowns•l.4 (J' = 0.58). Foraging guild distribution--needle insect gleanerst4, shrub foliage gleaner--3, bark gleanerst2 (jr = 0.51). Foraging guild overlapsPrairie Warbler--7, Black-throated Green Warbler--7, Cape May Warbler--6 (total 85).

Water Pipit, ( •lnthus spinoletta rubescens) Status•winter resident on golf coursesnear Freeport--none seen on transect routes.

Cedar Waxwing, ( Bombycilla cedrorum) Status•in-transit visitor flock of 50 seen in Freeport suburbs on 3-5 May 1968.

Starling, (Sturnus v. vulgaris) Status•irregular winter visitor, flock of 100 seen 18 March 1969.

Thick-Billed Vireo, (Vireo c. crassirostris) Geographical derivation--C2 A2. CD412-•36%. Status•permanent resident, 22.6 birds per km2 (13th). Habitat distribution-•coppets•133, marsh-•35, Young pines--18.9 (J• = 0.57). Habitat gradient--Figs: 19.20, 20.20, 21.17. Habitat overlap•redstart--83, Blue-gray Gnatcatcher--82, Striped-headed Tanager--81 (8 over 70%). Vertical distribution Fig. 32K. Compartmentdistribution--shrubs--3.6, crowns-•0.4 (J' • 0.25). Foragingguild distribution--shrubfoliage gleaners--7, (J'= 0.33). Foraging guild overlap--Parula Warbler--8, Black-throated Blue Warbler--7, Black- throated Green Warbler--7 (total 72).

Yellow-Throated Vireo, (Vireo flavi[rons) Statu•winter resident, two records 15 March 1969, 25 April 1969.

Black-WhiskeredVireo, (Vireo altiloquusbarbatulus) Geographicalderivation--C2 A1. Status---summerresident, common. Arrived 3 May 1968, 5 May 1969. Habitat distribution--restrictedto coppetsand around high shrubsand broad-leafedtrees in disturbedopen pineland. Compartmentdistribution--shrubs •l.0 (breedingseason). Foraging guild distribution--shrubfoliage gleaners---8,(J' = 0.22). 124 ORNITHOLOGICAL MONOGRAPHS NO. 24

Bananaquit,( Coerebaflaveola bahamends) Geographicalderivation--A2. CD412--20%. Status--permanentresident, 55.0 birdsper km• (6th). Habitat distribution-•coppets---160,coastal--70, submature pines--62 (J' = 0.90). Habitat gradient--F!gs: 19.39, 20.37, 21.31. Habitat overlapsPrairieWarbler--76, BahamaYellowthroat--76 (10 over 60%). Vertical distribution Fig. 27Z. Compartmentdistribution--shrubs--2.5, crowns--l.4 (J' = 0.53). Foragingguild distribution--nectarsippers•4, Flower insectprobers•2, Shrub foliage gleanerst2 (J' = 0.50). Foraging guild overlapsCuban Emerald Hummingbird--7, Cape May Warbler--6 (total 60).

Black-and-WhiteWarbler, ( Mniotilta varia) CD412--7% Status---winterresident, 4.5 birds per km2. Habitat distribution---coppers--16.1,submature pines--6.3 (J' = 0.30). Habitat gradient Figs: 19.21, 20.21, 21.18. Habitat overlapsStriped-headedTanager--88, redstarts88, Hairy Woodpecker--87(5 over 80%). Vertical distribution Fig. 32L. Compartmentdistribution--trunks•4.0 (J' = 0). Foragingguild distribution--barkgleaners--10 (J' = 0.07). Foraging guild overlap--Yellow-throatedWarbler (PR)--7, Brown-headedNuthatch•4 (total 22).

Worm-Eating Warbler, (Helmitheros vermivorus) CD412--10% Status•Winter resident,3 records,8 May, 1968; 13 March, 1969;25 April, 1969. Habitat distribution--submaturepiness0.91 (J• = 0).

Golden-WingedWarbler, (Vermivora chrysoptera) Status--winter resident, one record, 10 March 1968.

Orange-CrownedWarbler, (Vermivora celata) CD412--12% Status--winter resident,0.67 birds per km 2. Habitat distribution--mangrove--3.7,submature pines--0.6 (J' _• 0.21).

Northern Parula Warbler, (Parula americana) CD412---12% Status---winter resident, 6.8 birds per km•. Habitat distribution---coppets--56.9,submature pines--5.6 (J' = 0.39). Habitat gradient Figs: 19.22, 20.22, 21.19. Habitat overlap--Red-legged Thrush--95, Ovenbird--90, Blue-gray Gnatcatcher--83 (6 over 70%). Vertical distribution Fig. 32M. Foragingguild distribution--shrubfoliage gleaner--9 (J•: 0.15). Foraging guild overlap•Thick-billed Vireo--8, Black-throatedBlue Warbler--8 (total 55).

Yellow Warbler, ( Dendroica petechia gundlachi) Geographicalderivation•C1 A1. CD412---20%. Status--permanentresident, 0.6 birdsper km2. Habitat distribution--mangroves--6.5 (J• _• 0). Habitat gradient Figs: 19.23, 20.23. 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 125

Magnolia Warbler, ( Dendroica magnolia) CD412--15% Status--winterresident, 3 records: 15 January1969, 21 April 1969, 25 April 1969. Habitat distribution---coppets--2.3 (J': 0).

Cape May Warbler, (Dendroica tigrina) CD412--12% Status--Winter resident, 6.75 birds per km2. Abundant transient in early May 1968, late April 1969. Habitat distribution--submaturepines---9.3, coppets--6.0 (J": 0.33). Habitat gradient Figs: 19.24, 20.24. Habitat overlapsBlue-gray Gnatcatcher--88, Zenaida Dove--79 (4 over 70%). Vertical distribution Fig. 32N. Foraging guild distribution--needle gleanerst4, shrub foliage gleaners--2, nectar sippers---2 (J' = 0.58). Foraging guild overlapsPalm Warbler--7, Bananaquit--6, Prairie Warbler--6 (total 84).

Black-Throated Blue Warbler, (Dendroica caerulescensssp.) CD412--12% Status---winterresident, 1.58 birds per km2. Influx of transientsin late April-early May. Habitat distribution---coppets•48.8 (J' = 0). Habitat gradient--Fig: 19.25. Foraging guild distribution--shrubfoliage gleaners--8 (J•: 0.22). Foraging guild overlap--Parula Warbler--8, Thick-billed Vireo--7, Black-throated Green Warbler--7 (total 55).

Yellow-Rumped (Myrtle) Warbler, ( Dendroica c. coronata) CD412--20% Status---winterresident, 66.5 birds per km • (4th). Habitat distribution--marsh--146,coppets--121, submature pines--78.9 (J'= 0.88). Habitat gradient--Figs: 19.27, 20.25, 21.20. Habitat overlap•grassquit---80, Bananaquit--69 (5 over 60). Vertical distribution--Fig. 320. Compartment distribution--crowns--2.3, shrubs---1.6 (J'= 0.52). Foragingguild distribution--fruit and bud eaters--4, pine twig gleanerst3, (J• = 0.51). Foraging guild overlap--Striped-headedTanager---4, Gray Catbirds4, Northern Mock- ingbird---4 (total 62).

Black-ThroatedGreen Warbler, (Dendroicavirens spp.) CD412--12%. Status--winter resident,0.67 birds per km•. Habitat distribution--coppets---7.0,submature pines--0.5 (J• ----_0.13). Foraging guild distribution--shrubfoliage gleaners---6,pine needle gleaners--3 (J• = 0.31). Foraging guild overlap--Prairie Warbler--8, Blue-gray Gnatcatcher--7, Thick-billed Vireo--7 (total 80).

Yellow-ThroatedWarbler, (Dendroica dominica flavescens (a) anddominica (b)) Geographicalderivation (a) C1, CD412 (a and b)--20%. Status---(a)permanent resident (flavescens)--ca. 14 birds per km•; (b) winter resident (dominica)--ca. 7 birds per km2. Total 21 birds per km2 (14th). Habitatdistribution--submature pines--30.5, young pines--15.9 (J• = 0.45). Habitat gradient Figs: 19.28, 20.26, 21.21. Habitat overlapsLoggerheadFlycatcher--89, Olive-cappedWarbler--85, Pine War- bler--84 (5 over 70%). 126 ORNITHOLOGICAL MONOGRAPHS NO. 24

Vertical distribution--Fig. 32P. Compartmentdistribution-•crowns•3.1, trunks--0.5 (J' = 0.49). Foraging guild distribution--(a) bark gleaners--7, twig gleaners---2(J' = 0.35); (b) needle gleanerst6, pine twig gleanerst2 (J' = 0.45). Foraging guild overlap•(a) Brown-headedNuthatch--7, Black-and-whiteWarbler--7 (total 50); (b) Olive-cappedWarbler--7, Pine Warbler--7, Blue-grayGnatcatcher- 6 (total 61). (Assignmentto foragingguilds is very approximatesince identifica- tion to subspecieswas often impossiblein the field.)

Olive-CappedWarbler, ( Dendroica pityophila) Geographicderivation--C2 A0. CD412--25%. Status--permanent resident--72.0 birds per km 2 (3rd). Habitat distribution--submaturepines•114, youngpines--69 (J' = 0.51). Habitat gradient--Figs: 19.29, 20.27, 21.22. Habitat overlapsYellow-throatedWarbler--85, Pine Warbler--80 (4 over 70%). Vertical distribution Fig. 32Q. Compartmentdistribution-•crowns•3.96, shrubs•0.04 (J' = 0.04). Foragingguild distribution--needleinsect gleanerst9, twig gleaners•l (J' = 0.22). Foragingguild overlap--Pine Warbler--9, Yellow-throatedWarbler (WR)--7 (total 52).

Blackpoll Warbler, (Dendroica striata) CD412--20%. Status•in-transitvisitor; 15 April-13 May 1968;21 April-18 May 1969. Vertical distribution--Fig. 32R.

Pine Warbler, (Dendroica pinus achrustera) Geographicalderivation--C 1. CD412-•30%. Status•permanent resident,27.9 birds per km2 (1 lth). Habitat distribufion--submaturepines-•40, youngpines--38 (J' = 0.52). Habitat gradient--Figs: 19.30, 20.28, 21.23. Habitat overlapsYellow-throatedWarbler--84, Pine Warbler--80, Greater Antillean Pewee--80 (4 over 70%). Vertical distribution--Fig. 32S. Compartmentdistribution-•crowns•3.76 (J' = 0.18). Foraging guild distribution--needleinsect gleaners--9, bark insect gleaners•l (J' = 0.22). Foraging guild ovedap•Olive-cappedWarbler--9, Yellow-throatedWarbler (WR)--7 (total 50).

Kirtland's Warbler, (Dendroica kirtlandi) Status--Rare winter resident, never seen in 500 hours of field observation but one captured by Dr. Paul Fluck in April 1969. Habitat distribution--Capturedbird wasin submaturepine habitat.

Prairie Warbler, (Dendroica d. discolor) CD412--15%. Status•winter resident, 16.7 birds per km•. Habitat distribution-•coppets•87,old fields•25, submaturepines--15 (J' _--0.66). Habitat gradient-•Figs: 19.31, 20.29, 21.24. Habitat overlapsGray Catbird--90, Thick-billed Vireo--79, Bananaquit--76 (7 over 70%). Vertical distribution--Fig. 32T. Compartmentdistribution--shrubs•2.4, crowns•l.6 (J' = 0.54). Foragingguild distribution--shrub foliage gleaners--5, needle gleaners--5 (J' = 0.28). Foraging guild overlap--Black-throatedGreen Warbler--8, Blue-gray Gnatcatcher--7 (total 77). 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 127

Palm Warbler, (Dendroica palmarum ssp.) CD412---24%. Status--Winter resident214.8 birds per km• (1 st). Habitat distribution--Coastal brush---559, Submature piness256, Young pines-•197 (J' = 0.82). Habitat gradient--Figs: 19.32, 20.30, 21.25. Habitat overlap--GroundDove--62, Common Yellowthroat--61,Bananaquit--60 (6 over 50%). Vertical distribution---Fig. 32U. Compartmentdistribution---crowns--2.7, shrubs--l.1 (J' = 0.54). Foraging guild distribution--groundinsect gleaners--3, needle insect gleanerst3 (J' = 0.80). Foragingguild overlap--CapeMay Warbler--7, Gray Catbird--6, BahamaYellowthroat --5 (total 88).

Ovenbird, ( Seiurusaurocapillus) CD412--12%. Status--Winterresident, 4.2 birdsper km•. Transientinflux in late April and early May. Habitat distribution-•coppets--63,submature pines•l.1 (J' = 0.04). Habitat gradient--Figs: 19.33, 20.31, 21.26. Habitat overlap--Red-legged Thrush--92, Parula Warbler--90, Blue-gray Gnatcatcher- 75 (8 over 60%). Vertical distribution Fig. 32V. Compartment distribution--Ground---3.6. Foraging guild distribution--ground insect gleaners--10 (J'= 0.04). Foraging guild overlapsCommon Yellowthroat--8, BahamaYellowthroat--6 (total 26).

Northern Waterthrush, (Seiurus novaboracends spp.) CD412--15%. Status---winterresident, 10.3 birdsper km•. Transientsin early May. Habitat distribution--marsh--125, coppets•9, mangroves--27 (J' = 0.46). Habitat gradient--Figs: 19.34, 20.32, 21.27. Habitat overlap--Yellow-bellied Sapsucker--87, Yellow-rumped Warbler--64 (2 over 50%).

Common Yellowthroat, (Geothlypis tricha, ssp.) CD412--15%. Status---winterresident, 50.1 birds per km• (7th). Habitat distribution--coastal brush--341, marsh---228, old fields•216 (l' = 0.79). Habitat gradient--Figs: 19.35, 20.33, 21.28. Habitat overlap--Palm Warbler--61, Bananaquit--57 (5 over 50%). Vertical distribution Fig. 32W. Compartment distribution--shrubs--2.9, ground--l.0 (I' = 0.50). Foraging guild distribution--ground insect gleaners--8, shrub foliage gleanerst2 (J' = 0.17). Foraging guild overlap---Ovenbird--8, Bahama Yellowthroat•7 (total 47).

Bahama Yellowthroat, (Geothlypia roatrata tanned) Geographicalderivation--C2. CD412--20%. Status•permanent resident,32.2 birds per km• (9th). Habitat distribution-•coppets--80.1, old fields•53, young pines--46 (J' = 0.78). Habitat gradient--Figs: 19.36, 20.34, 21.29. Habitat overlap--Bananaquit--76, Cuban Emerald Hummingbird-•72, Prairie Warbler-- 62 (9 over 60%). Vertical distribution--Fig. 32X. Compartment distribution--shrubs--2.7, ground cover--l.0 (J'= 0.61). 128 ORNITHOLOGICAL MONOGRAPHS NO. 24

Foraging guild distribution--ground insectgleaners--6, nectar sippers--2 (J' = 0.38). Foraging guild overlapsCommon Yellowthroat--7, Ovenbird--6, Palm Warbler--5 (total 47).

Hooded Warbler, (Wilsonia citrina) Status--winter resident, three records: 2 April 1968, 8 April 1968, 11 April 1968.

Wilson's Warbler, (Wilsonia p. pusilia) CD412--12%. Status--winter resident, 0.3 birds per km:. Habitat distribution---coppers--6.5 (J': 0).

Redstart, ( Setophaga r. ruticilla ) CD412--15%. Status--winter resident, 8.5 birds per km 2. Transient influx in late April and early May. Habitat distribution--coppets•49, submaturepines--12.7 (J' = 0.46). Habitat gradient--Figs: 19.37, 20.35, 21.30. Habitat overlap--Blue-gray Gnatcatcher--93, Striped-headedTanager--90, Black-and- white Warbler--88 (7 over 80%). Vertical distribution--Fig. 32Y. Compartment distribution--crowns--2.3, shrubs--l.7 (J' = 0.49). Foraging guild distribution--shrub foliage gleanerst4, air sallyerst4 (J' = 0.36). Foraging guild overlapsGreater Antillean Pewee--6, Stolid Flycatcher--6, Thick- billed Vireo--5 (total 69).

House Sparrow, (Passer d. domesticus) Geographical derivation--XO. Status--permanent resident. Habitat distributionsCommon in urban Freeport and around hotels. None seen on the transect routes.

Bobolink, ( Dolichonyx orizivorus) Status--in-transit visitor 27 April (1969)-11 May (1968). None seen on transect routes. Habitat distribution--mainly in coastal brush and grassyareas behind dunes.

Red-Winged Blackbird, ( Agelaius phoeniceusbryantO CD412--60%. Status--winter resident,5.8 birds per kmL Habitat distribution--mangroves•49, marsh--31 (J' = 0.52). Habitat gradient--Figs: 19.38, 20.36. Habitat overlap--Yellow-rumped Warbler--46, Bahama Swallow--45 (2 over 40%).

Baltimore Oriole, (Icterus g. galbula) Status--winter resident. Uncommon in Freeport suburbs. None seen on transect routes. Striped-Headed Tanager, (Spindalis zena townsendi) Geographical derivation--A1. CD 412--20%. Status--permanentresident, 36.8 birds per km2 (Sth). Habitat distribution-•coppets•156, submature piness44.4, young pines--24.9 (J' = 0.45). Habitat gradient Figs: 19.40, 20.38, 21.32. Habitat overlap--Blue-gray Gnatcatcher--90, Redstart--90, Hairy Woodpecker--70 (6 over 80%). Vertical distribution Fig. 32AA. Compartment distribution--shrubs--24, crowns•l.6 (J' = 0.49). Foraging guild distribution--fruit and bud eaters--9 (J' = 0.18). Foraging guild overlapsNorthern Mockingbird--6, Gray Catbird--6 (total 41). 1977 EMLEN: GRAND BAHAMA BIRD COMMUNITIES 129

Rose-BreastedGrosbeak, ( Pheucticusludovicianus) Status--in-transit visitor, 4 records; 25 April to 9 May 1969.

Indigo Bunting, ( Passerina cyanea) Status--in-transit visitor, 12 April (1968)--6 May (1969) Habitat distribution Irregular at brushy sites near Freeport. None seen on the transect routes.

Black-Faced Grassquit, ( Tiaris b. bicolor) Geographicalderivation---A1. CD412--15%. Status--permanentresident 24.0 birdsper km'"(12th). Habitat distribution--coppets•50.1, submature pines--37.7, marsh---34 (J': 0.76). Habitat gradient Figs: 19.42, 20.40, 21.34. Habitat overlap•Yellow-rumped Warbler--80, Bananaquit--72 (2 over 70%). Vertical distribution--Fig. 32BB. Compartment distribution--shrubs•2.4, ground cover--l.2 (J' = 0.64). Foraging guild distribution--stem seed pluckerst5, ground seed gleaners--2 (J' : 0.47). Foraging guild overlap--Red-legged thrush•4, Palm Warblers--3 (total 38). Greater Antillean Bullfinch, ( Loxigilla v. violacea) Geographical derivation--A1. CD412---15%. Status--permanentresident, 2.1 birds per km'ø. Habitat distribution--coppets--34 (J' = 0). Habitat gradientsFigs: 19.41, 20.39, 21.33. Compartment distribution--shrubs--2.0, ground cover--l.5 (J' = 0.70). SavannahSparrow, ( Passerculussandwichensis savanna) CD412--12%. Status•winter resident,1.6 birds per km'2. Habitat distribution-•coastal brush--34.5, old fields--11 (J' = 0.28). Also common in grassylots in suburbanFreeport. Habitat gradient: Figs: 19.43, 20.41, 21.35.

GrasshopperSparrow, (,'lmmodramus savannarumpratensis) CD412--12%. Status•winter resident,4.8 birds per km'•. Habitat distribution--old fields•51.7, youngpines•l.0 (J' = 0.05). Habitat gradient--Figs: 19.44, 20.42, 21.36. Lincoln Sparrow, (Melospiza I. lincolni) Status--winter resident. Two records: 6 May 1968; 8 February 1969. Habitat distribution---brush behind coastal dunes. No. 15. FunctionalAnatomy and AdaptiveEvolution of the FeedingApparatus in the Hawaiian HoneycreeperGenus Loxops (Drepanididae), by Law- rence P. Richardsand Walter J. Bock. vii + 173 pp., 14 text figures+ 26 plates.1973. Price$9.00 ($7.50to AOU members). No. 16. The Red-tailedTropicbird on Kure Atoll, by RobertR. Fleet.vii q- 64 pp.,34 textfigures, 5 tables.1974. Price $5.50 ($4.50 to AOU members). No. 17. ComparativeBehavior of theAmerican Avocet and the Black-necked Stilt (Recurvirostridae),by RobertBruce Hamilton, vi + 98 pp., 18 text figures.1975. Price$7.50 ($6.00 to AOU members). No. 18. BreedingBiology and Behavior of theOldsquaw (C!angula byemalls L.), by RobertM. Alison,vi + 52 pp., 13 textfigures. 1975. Price$3.50 ($2.50 to AOU members). No. 19. Bird Populationsof AspenForests in WesternNorth America,by J. A. DouglasFlack, viii + 97 pp.,frontis., 56 textfigures, appendix. 1976. Price $7.50 ($6.00 to AOU members).

No. 20. SexualSize Dimorphism in Hawksand Owls of North America,by Noel F. R. Snyderand James W. Wiley,vi + 95 pp., frontis.,14 text figures, appendix.1976. Price$7.50 ($6.00 to AOU members). No. 21. SocialOrganization and Behavior of theAcorn Woodpecker in Central CoastalCalifornia, by MichaelH. MacRobertsand Barbara R. MacRoberts, viii q- 115 pp., 23 text figures,2 appendices.1976. Price$7.50 ($6.00 to AOU members).

No. 22. Maintenance Behavior and Communication in the Brown Pelican, by RalphW. Schreiber,viii + 78 pp.,38 textfigures. 1977. Price $6.50 ($5.00 to AOU members). No. 23. SpeciesRelationships in the AvianGenus Aimophila, by LarryL. Wolf, viii q- 220 pp., 17 textfigures q- 10 plates,long-play phono disc album. 1977. Price $12.00 ($10.50 to AOU members). No. 24. Land Bird Communities of Grand Bahama Island: The Structure and Dynamicsof an Avifauna,by JohnT. Emlen,xi + 129pp., 38 textfigures, appendix.1977. Price$9.00 ($8.00 to AOU members).

Like all otherAOU publications,Ornithological Monographs are shippedprepaid. Make checkspayable to "TheAmerican Ornithologists' Union." For the convenience of thosewho wish to maintaincomplete sets of OrnithologicalMonographs and to receive new numbersimmediately upon issue,standing orders will be accepted. Order from: Glen E. Woolfenden, Assistant to the Treasurer AOU, Depart- mentof Biology,University of SouthFlorida, Tampa, Florida •620. ORNITHOLOGICAL MONOGRAPHS No. 1. A Distributional Study of the Birds of British Honduras, by StephenM. Russell 195 pp., 2 color plates. 1964. Price $7.00 ($5.50 to AOU members). No. 2. A Comparative Study of Some Social Communication Patterns in the Pelicaniformes, by Gerald Frederick van Tets. 88 pp., 4 text figures. 1965. Price $3.50 ($2.50 to AOU members). No. 3. The Birds of Kentucky, by RobertM. Mengel. Cloth bound,xiv + 581 pp., 4 color plates plus text figures and vignettes. 1965. Price $15.00 ($12.50 to AOU members). No. 4. Evolution of Some Arctic Gulls (Larus): an Experimental Study of Isolating Mechanisms,by Neal Griffith Smith. 99 pp., 62 text figures. 1966. Price $4.00 ($3.00 to AOU members). No. 5. A Comparative Life-history Study of Four Speciesof Woodpeckers,by Louise de Kiriline Lawrence. 156 pp., 33 text figures. 1967. Price $6.00 ($4.50 to AOU members). No. 6. Adaptations for Locomotion and Feeding in the Anhinga and the Double-crested Cormorant, by Oscar T. Owre. 138 pp., 56 text figures. 1967. Price $6.00 ($4.50 to AOU members). No. 7. A Distributional Survey of the Birds of Honduras, by Burt L. Monroe, Jr. 458 pp., 28 text figures, 2 color plates. 1968. Price $14.00 ($11.00 to AOU members). An Approachto the Study of EcologicalRelationships among Grassland Birds, by JohnA. Wiens. 93 pp., 30 text figures. 1969. Price $4.00 ($3.00 to AOU members). No. 9. Mating Systems,Sexual Dimorphism, and the Role of Male North American PasserineBirds in the Nesting Cycle, by JaredVerner and Mary F. Willson. 76 pp. 1969. Price $4.00 ($3.00 to AOU members). No. 10. The Behaviorof SpottedAntbirds, by EdwinO. Willis,vi + 162pp., 3 color plates, 27 text figures. 1972. Price $9.00 ($7.50 to AOU members). No. 11. Behavior, Mimetic Songsand Song Dialects, and Relationshipsof the Parasitic Indigobirds (Vidua) of Africa, by RobertB. Payne,vi % 333 pp., 2 color plates, 50 text figures, 40 andiospectrographs.1973. Price $12.50 ($10.00 to AOU members). No. 12. Intra-island Variation in the MascareneWhite-eye Zosterops borbonica, by Frank B. Gill, vi + 66 pp., 1 color plate, 31 text figures. 1973. Price $3.50 ($2.50 to AOU members). No. 13. Evolutionary Trends in the Neotropical Ovenbirdsand Woodhewers,by Alan Feduccia, iv + 69 pp., 20 text figures. 1973. Price $3.50 ($2.50 to AOU members). No. 14. A Symposiumon the House Sparrow(Passer domesticus) and European Tree Sparrow (P. montanus) in North America, S. Charles Kendeigh. chairman. vi q- 121 pp., 25 text figures. 1973. Price $6.00 ($4.50 to AOU members). (Continued on inside back cover)