THE AUK A QUARTERLY JOURNAL OF ORNITHOLOGY
VoL. 100 J^•qv^I•¾1983 No. 1
FRUGIVORY AND POLYGAMY IN BIRDS OF PARADISE 1
BRUCE BEEHLER• Departmentof Biology,Princeton University, Princeton, New Jersey08544 USA
ABSTRACT.--Istudied fruit-feeding by nine speciesof birds of paradisein Papua New Guinea from July 1978through November 1980and gathered1,187 recordsof foragingat 31 speciesof treesand vines from 14 botanicalfamilies. Fruit consumedwas consistentlysmall- to moderate-sized (mean: I cm diameter), but fruit of different speciesof plants showed high morphologicaldiversity. I classifythe fruit of 31 plant speciesinto three morphological groups:capsule, fig, and drupe/berry.Each of the primarily frugivorousbirds of paradise was recordedtaking fruits from 10-21 plant species,including representativesfrom each class. The monogamousTrumpet Manucode and Crinkle-collaredManucode were fig specialists. More than 80% of their diet was figs. The polygamousspecies of paradisaeidswere more "generalized"fruit-feeders and took significantamounts of fruit from all threemorphological categories.The most important types of fruit among the polygamousbirds were capsular species(49% of diet). While fig specieswere visited by birds from many families, most nonfig trees hosted a narrower range of foragers,and two speciesof trees, Chisocheton weinlandii(Meliaceae) and Gastoniaspectabilis (Araliaceae), were visited only by birds of paradise. The frugivoroushabits of birds of paradiseare similar in severalrespects to thoseof the neotropicalcotingids and manakins.It is arguedthat while frugivory is an important com- ponent of the evolutionof polygamousarena display in thesebirds, it cannot,by itself, explain why some birds are polygamousand others monogamous.Frugivory in the tropics is a complexsyndrome that offers a number of ecologicalalternatives that, in turn, promote different behavioraladaptations. Received 22 March 1982, accepted30 June1982.
D. SNOW (1962a, b, 1971a, b) and B. Snow which malesare emancipatedfrom the nestand spend (1970, 1972, 1973, 1977) in their studies of the most of their time in epigamic display.... It is neotropical cotingas (Cotingidae) and mana- probablethat the evolution of the birds of paradise kins (Pipridae) discoveredan apparentrelation in the Australianregion was made possibleby a sim- ilar relationship between the birds and their food linking the habit of fruit-eating to non-pair- supply. bonding polygamy. D. Snow (1980: 1197) writes: There are fundamental differences between the demandsof frugivory and those of insec- Probably as a consequenceof the very nutritious tivory. Insectsare elusive, dispersedthrough fruitsthat they eat, and of theshort tim e perday that the habitat, and difficult to harvest, while fruit they need spendin foraging, somespecialized neo- tropical frugivores have evolved social systemsin is locally abundant, "patchy," and relatively simple to harvest in quantity (Lill 1976). Ani- mals that forage for insectsmust spend a large • Dedicated to the memory of J. Linsley Gressitt, portion of eachday satisfyingtheir dietary de- founder of the Wau EcologyInstitute. 2 Present address: MNH 114 (NHB Room 336), mands, while frugivorescan quickly fill their Smithsonian Institution, Washington, D.C. 20560 cropswith fruit and invest the remaining day- USA. light hoursin other activities,such as self-ad-
The Auk 100: 1-12. January 1983 2 BRUCEBEEHLER [Auk, Vol. 100
TABLE1. The birds of paradiseencountered at fruit treeson the Mount Missim study site.
Abun- Englishname • Latin binomen Weightb Wingc dance• Crinkle-collaredManucode Manucodiachalybatus 183 + 18 (3) 173 _+8 2-3 Trumpet Manucode M. keraudrenii 171 + 10 (8) 169 _+4 7-10 MagnificentRiflebird e'f Ptilorismagnificus 194 _+16 (2) 187 _+2 1-2 Black-billed Sicklebill Epimachusalbertisii 108 (1) 156 1-2 Superb Bird of Paradiseg Lophorinasuperba 89 _+ 5 (4) 136 _+2 2-4 Lawes' Six-wired Bird of Paradise Parotia lawesii 166 _+ 4 (4) 158 _+4 4-5 MagnificentBird of Paradise Diphyllodesmagnificus 90 + 8 (12) 112 + 2 6-7 RaggianaBird of Paradise Paradisaearaggiana 246 _+12 (3) 181 _+2 5-6 Blue Bird of Paradiseg P. rudolphi 178 (1) 156 1-2 • Nomenclaturebased on Diamond (1972)with someemendations to be usedin a checklistof New Guinea birds presentlyin preparation(B. Finch, H. Bell, and B. Beehler). •' Data includeweights of adult malesonly; listed are meanweight, SD, and samplesize. ½Males only; wing arc. dAbundance as measuredby the numberof individualsobserved using a fruiting tree duringa singleday. • Weightsfrom Gilliard and LeCroy(1967). f Wing measurementsfrom skins of threeadult males,preserved as studyskins, National Museum of NaturalHistory, Washington, D.C. • Data from Diamond (1972).
vertisement, mate competition, and defense of as well as interspecific differences, and com- a territory, display perch, or mate(s) (Snow pare theseobservations with thosefor frugivo~ 1976, Foster1977). The logic of this argument rous songbirdsfrom the neotropics.Finally, I is persuasive,although the generalizationas I examine the characteristicsof fruit eaten by have presentedit is oversimplified.The rela- polygamousarena-displaying birds and com- tionship between fruit-eating and polygamous ment on the relation of diet to behavior. arena-displayis supportedby the simple ob- servationthat the vast majority of polygamous STUDY AREA AND METHODS arena birds in the humid tropics are fruit-eat- I studiedbirds of paradiseon a 17-haplot of forest ers (Snow 1976,Bradbury 1981). on the southwestslopes of Mount Missim, at about Several authors have shown that fruit is a 1,430 m. The site is 9 km north-northeast of Wau, prominent part of the diet of some speciesof Morobe Province, Papua New Guinea (7ø16'S, birds of paradise (Rand 1940, Gilllard 1969, 146ø42'E).I worked on this site from July 1978 to Schodde 1976), and, in a few instances,the fruit November 1980, for 249 field-days.The study-siteis consumedby these birds has been identified mature Pre-montane Moist Forest (Holdrige et aL botanically(Cooper and Forshaw1977, LeCroy 1971) on highly dissectedand well-drained ridge- et al. 1980). To date, however, no systematic and-ravine topography.The area receivesan annual rainfall of about 2,000 mm (McAlpine et al. 1975) in accounthas been published on frugivory by a moderatelyseasonal pattern, with more rain falling birds of paradise. from November to April. The 17-ha site supports While studying the behavior and ecologyof about 80 speciesof trees and a community of about four speciesof birds of paradise in the moun- 125 speciesof resident birds. The plot is unusual in tains of centralPapua New Guinea, I gathered having nine residentbirds of paradise(Table 1). A data on fruit-eating by all bird speciessharing tenth species,Loria's Bird of Paradise(Cnemophilus fruit resourceswith these birds of paradise. The loriae), is an occasionalvisitor. Additional details on avian fruit-eating community included more the habitat and biota of the study-site are presented than 30 speciesof birds. In this paper I docu- elsewhere (Beehler 1983). ment the types of fruit eaten by nine species I documentedthe diet of the birds by: (1) recording of birds of paradise (Table 1) that share this birds visiting fruiting trees, and (2) collectingand mid-montane study-site,and I show that, while •dentifyingthefruit remains in fecal samples taken from captured birds. I made repeated observations the fruit-crop of some trees is consumedby a of fruit-feeding at plants producingfruit eaten by wide range of avian foragers, certain tree birds of paradise.In the early monthsof the study I speciesproduce fruit eatenmostly, and in some monitoredvisits by birds to many speciesof fruiting instanceswholly, by birds of paradise.In the plantsthat my researchassistants encountered in the discussionI point out somefamily-wide trends, forest.When I found a tree or vine that was regularly January1983] Frugivoryin Birdsof Paradise 3 visited by birds of paradise, I began making stan- speciesof birds of paradisethat thesewere not lost dardized observations. I recorded all birds that en- in the sampling bias. tered the tree to feed; for selectedspecies of para- The unit of measurein this projectwas the "feed- disaeidsI recordedlength of feeding-bouts,number ing-bout":a recordof a bird enteringa tree and tak- of fruit consumed, and interactions with other for- ing at leastone fruit. I discoveredearly in the study agers.I was assistedin this work by two field tech- that birds of paradiseforaged for fruit in a behav- niciansemployed to find new treesand to help with iorally stereotypedfashion (see Resultssection). the foragingcensuses. Their help madepossible the Consequently,I treat paradisaeidforaging bouts as simultaneous observations of avian foraging at sev- equivalent. Avian use of fruit resourcesare quanti- eral fruit-trees. fied and comparedin terms of theseforaging bouts. We recordedfruit-foraging by sittingnear the fruit- I collectedadditional data on fruit eatenby birds tree and watching birds enter, feed, and depart. We of paradiseby examiningfecal samples. I mist-netted identified birds using 7 x 35 and 8 x 44 binoculars, birds and placed them in a darkenedholding cage as well as a 157
TABLE2. Recordof feeding bouts by nine bird of paradise speciesat 31 food plants.
Birds of paradise•' Tree Fruit species Plant name typd' CCM TMA MRI BBS SBP LSW MBP RBP BBP totals Aglaia C ..... i -- 3 -- 4 Aporusa C .... I .... 1 Canthium D ...... i -- -- 1 Chisocheton • C 1 7 16 5 23 10 13 13 3 91 Cissushypoglauca • D -- 13 1 1 1 2 4 5 3 28 C. aristata D 1 ..... 1 -- -- 2 Dysoxylum•' C -- 1 ------1 12 17 -- 31 Elmerrillia •' C 1 8 -- 1 2 2 12 2 -- 28 Endospermurn•' D -- 2 -- -- 9 2 15 2 -- 30 Ficusdrupacea •' F -- 2 ...... 2 Ficusgul •' F 25 155 -- -- 1 12 8 45 1 247 F. odoardi •' F 19 55 2 -- 1 6 8 20 -- 111 Ficus #202 F 1 2 .... 1 -- -- 4 Ficus #181 F -- 7 ------2 -- -- 1 10 Ficus #217 F -- 9 ...... 9 Ficus #246 • F 1 4 .... 2 -- -- 7 Ficus #275 •' F 8 40 -- -- 2 1 5 13 4 73 Ficus #371 F 1 6 -- 1 4 4 -- 3 7 26 Gastonia •' D -- 17 7 -- -- 22 58 24 3 131 Glochidion C ...... 1 -- -- 1 Homalanthus •' C -- 4 8 -- 17 10 104 75 23 241 Myristica C -- 3 ------3 i 1 -- 8 Pandanus •' D .... 2 -- 5 2 1 10 Piper D -- 1 ...... 1 Scheffiera D -- -- 1 -- -- 47 2 -- 10 60 Sloanea aberrans C ...... 2 -- -- 2 S. sogerensis C -- 2 -- -- 2 2 4 1 -- 11 Sterculia C ..... 1 ------1 Syzygium D -- 8 ------1 -- 1 -- 10 Uvaria D -- 2 ...... 2 Zingiberaceae C 1 ..... 2 -- 1 4 Bird feeding totals: 59 348 34 7 65 129 261 227 57 1,187 "The three-letterabbreviations are formedfrom the first letterof the genericname plus the first lettersof the specificname, Decoded,they are: CCM - Crinkle-collaredManucode; TMA - Trumpet Mat•ucode;MRI - MagnificentRiflebird; BBS- Black-biIledSicklebill; SBP - Superb Bird of Paradise;LSW Lawes' Six-wired Bird of Paradise;MBP - MagnificentBird of Paradise;RBP - RaggianaBird of Paradise;BBP - Blue Bird of Paradise, •' The abbreviations for fruit-types are: C - capsule, F fig, and D = drupe/berry. •' Plants marked with this footnote are illustrated in Fig. 1. severaltimes in a morning, but eachvisit would the tree where they were feeding; to perform be relatively brief. such activities they invariably moved to adja- Initially, I measuredlength of foraging bouts cent vegetation.They preferred to feed in a tree for three species: Trumpet Manucode, mean that was already occupied by other foragers, length of foraging bout = 220 s + 89 SD, n = although this generalization does not hold for 18; Raggiana, mean = 178 _+ 122 SD, n = 26; smallfruiting plants that could accomodateone Magnificent, mean = 171 + 101 SD, n = 26. or two birds at a time. Once a foraging group These data show no significant difference (one- assembledin a fruit tree, foraging was kinetic. way ANOVA, F = 1.2, P = 0.3). In addition, I The bustle of activity was punctuated by rapid compared data for the number of fruit taken evacuations of the tree, in what I presume were per bout: Trumpet Manucode, mean number predator-escape maneuvers. Many times I of fruit eaten per bout = 7.6 + 5, n = 11; Rag- watched groups of birds flush from fruit trees, giana, mean = 5.4 + 7, n = 20; Magnificent, dropping nearly vertically toward the ground, mean = 5.3 + 6, n = 17. then veering into the nearestvegetation. I nev- The different species of birds of paradise er observedpredators in these instances,but tended to foragefor fruit in a stereotypedfash- predators were present on the study site, and ion. They rarely or never rested or preened in I did find carcasses of birds that had been con- January1983] Frugivoryin Birdsof Paradise 5
FIG
B
A
DRUPE/BERRY &c.
E
H
CAPSULE
I
M
Os L
Fig. 1. Selectedfruits from the threemorphological classes. Abbreviations are: S = seed;M = longitudinal/ medial section;X = transverse/medialcross-section. A. Ficusodoardi. B. Ficusgul. C. Ficus#275. D. Ficus #246. E. Gastoniaspectabilis. F. Pandanusconoideus. G. Cissushypoglauca. H. Endospermummedullosum. I. Homalanthusnovoguineensis. J. Dysoxylum cf. macrothyrsum.K. Elmerrilliapapuana. L. Chisochetonweinlandii.
sumed by foresthawks. I believe that the fear The fruit.--For the nine speciesof birds of of predation may be one of the forcesmodi- paradiseat the Mount Missim study site, I ob- fying fruit-foraging behavior in the speciesof tained 1,187 records(Table 2) of foraging at 31 birds of paradise. It is a conservativeforce, in- speciesof trees and vines of 14 botanicalfam- advertently suppressingthe acquisition of in- ilies (see Appendix for data on plants, their terspecifically different foraging methods taxonomy,and morphology).The fruit of the among the species(Howe 1979). 31 speciesvaried from 3 mm in diameter(edi- Birds of paradiseharvest fruit while perched ble portion) for Glochidionto 28 mm for Ficus and never take items in flight, as is character- odoardi.This difference in size did not appear istic of the prominent neotropicalfruit-eaters, to limit edibility to birds of paradise. The the cotingids and manakins (Snow 1976). smallestbird in the group, the MagnificentBird 6 BRUCEBEEnLER [Auk, Vol. 100
of Paradise, consumed both the smallest and erate (15,000for Cissushypoglauca), with fruit- largest fruits. The largest bird, the Raggiana ing seasonsfrom 2 to 12 monthslong and low Bird of Paradise,was recordedtaking the larg- synchronyof fruit ripening within the crop. est and second smallest fruits. The fruits of the capsulecategory ranged in Morphology of the fruits varied interspecif- size from 3 to 16 mm (mean 10 mm). Coloration ically (Fig. 1 a-l). I classifythe fruits into three is two-part in this class. The outer shell of the morphologicalgroups. (1) Fig (Fig. 1 a-d), fruits dehiscent capsule is one color, and the edible from plants of the genus Ficus,"an aggregate aril is invariably another color, often contrast- fruit formed by enlarged fleshy cups contain- ing with that of the capsule. In the speciesof ing numeroussmall achenes"(Johns 1976: 107). this class, the arils were: red (5); orange (3); The bird either swallowsthe entire fig or, when pink, brown, and white (1 each). The color of the fruit is large, pecks off fleshy pieces of it. the outer capsule ranged from brown (6) to (2) Capsule(Fig. i-l), fruit that has a seed with green (4) and orange-tan (1). Plants producing edible aril encased within an inedible, dehis- capsularfruit were predominatelycanopy trees cent capsule(included in this generalclass are: (6) or understorytrees (4). The fruit-cropswere legume, follicle, pod, and etario; Foster and very small (lessthan 300) to moderate(12,000), Gifford 1974, van der Pijl 1972). To feed, the with low synchronyof ripening and a season bird plucks the seed with attachedaril from the lastingfrom 11/2to 5 months. opened capsule and swallows it, later either Summarizing, the fruits eaten by the birds regurgitating or defecatingthe seed. (3) Drupe/ of paradise ranged in size from 3 to 28 mm berry (Fig. e-h), simple, compound, and ag- diameter (mean 10 mm). These measurements gregatefruits with seedssurrounded by fleshy of external diameter are not indicative of the edible pericarp. In all specieslisted in Table 2, volume of edible material because of the vari- the fruit (or fruitlet) in the drupe/berry classis ation in seed size; they are an important in- eaten whole (see Jensenand Salisbury 1972). dication, however, of the average size of the Of the 31 plant specieswhose fruit was taken nutritional "package" acceptableto foraging by the nine birds of paradise, 12 were capsular birds of paradise. The two most common fruit (412 observations), 10 were from the drupe/ colorswere red and orange. Van der Pijl (1972) berry class (275 observations), and 9 species notes that red and blue/purple are the most were figs (500 observations). popular colorsfor bird-dispersedfruits. Growth Figs ranged in diameter from 6 to 28 mm forms varied: canopytree (11), understorytree (mean 12 mm). Color of ripe figs varied from (7), vine (7), strangler (5), herbaceousshrub orange(4 species),green (2), red (1), and pink (1). (1) to whitish (1). The most commongrowth- The birdsand their feedingpreferences.--For form was hemiepiphytic strangler, which is five speciesof birds of paradise, I have fewer epiphytic in its early stages,then developsinto than 100 feeding observationsper species(Ta- a free-standingtree after it growslarge and kills bles 2 and 3). For these,I can make only guard- the host tree that originally supportedit. Five ed comments on feeding preferences.Of 59 specieswere stranglers, two were vines, and feeding observations of the Crinkle-collared two were canopy trees. Fig fruit-crops ranged Manucode,55 (93%) were at figs.This bird was from huge(100,000) to small(5,000); in the larger a fig specialist.Data on foragingby Black-billed species,with large crops,the fruit ripened syn- Sicklebill and Magnificent Riflebird (Beehler chronously over a season as short as 1983;unpubl. data) show that thesebirds were months. Many speciesof foragerswere attract- largely insectivorous.Both spent a preponder- ed to these large crops, and the figs were, in ant amount of time gleaningfor insectson bark most cases,simple to harvest, being arrayed and dead wood. This agrees with data from on the twigs and smaller limbs. Schodde(1976). The Superb and Blue birds of The 10 species in the drupe/berry category paradise were recorded at 12 and 11 speciesof varied from 6 to 20 mm (mean 13 mm). The fruiting plants, respectively,from all three fruit most common color of the ripe fruit was red categories. (4), with two each of yellow and white, and a The four speciesfor which I have more than single purple fruit. These fruits were displayed 100 observations each merit individual ac- on vines (5 species), canopy trees (3), and counts. Like the Crinkle-collared Manucode, the understory trees (2). Their crops ranged from Trumpet Manucode was a fig specialist. I re- small (2,000 for Scheffierapachystyla) to mod- corded the Trumpet Manucode at 21 speciesof January1983] Frugivoryin Birdsof Paradise
TABLE3. Paradisaeidpreference for fruits from the three morphologicalclasses?
Fruit classesb
Bird speciesc Fig Capsule Drupe/berry Crinkle-collared Manucode 55 (93%) 3 (5%) I (2%) Trumpet Manucode 280 (80%) 25 (7%) 43 (13%) MagnificentRiflebird 2 (6%) 24 (71%) 8 (23%) Black-billed Sicklebill 1 6 Superb Bird of Paradise 8 (12%) 45 (69%) 12 (19%) Lawes' Six-wired Bird of Paradise 22 (17%) 33 (26%) 74 (57%) MagnificentBird of Paradise 24 (9%) 151 (58%) 86 (33%) RaggianaBird of Paradise 81 (36%) 112 (49%) 34 (15%) Blue Bird of Paradise 27 (47%) 13 (23%) 17 (30%) 500 412 275
• See text for identification of fruit classes. b First numberis totalof field observationsof bird takingthis type of fruit; secondnumber, in parentheses,is percentageof bird's diet of fruit. c See Table 1 for full data on bird nomenclature.
fruiting plants;this list includesall nine species point of view. I include for analysisdata for all of figs. This fig componentmade up 80% of speciesof avian foragers that visited the food the bird's foraging records.The two speciesof plants. Fruit sourcesdiffered in the extent to Manucodia showed a foraging preference un- which they were dominated by birds of para- like any other birds of paradisein the study. dise. For analysis, I compare five food plants Lawes' Six-wired Bird of Paradise was re- that exhibit the range of variation. corded feeding at 18 speciesof fruiting plants Ficus#275 (Fig. lc) is a large stranglerwith from all three categories.The vine, Scheffiera a massivebranching canopy of limbs. Mature pachystyla,accounted for 36% of my observa- individuals produced a prodigious number of tions for this bird, and this bird made 78% of tiny figs, arrayedon the twigs and outerlimbs. the recordedavian visits to the Scheffiera. The figs ripened synchronouslyand were sim- I recorded261 foraging bouts of the Magnif- ple to harvest. Flocks of birds would visit and icent Bird of Paradise.This specieswas the nu- forageduring the peak period of fruiting. I re- merically dominant forager at two species of corded22 speciesof foragersfrom 10 bird fami- fruit trees: Homalanthus novoguineensisand lies. Birds of paradiseaccounted for 35% of the Gastoniaspectabilis. While the MagnificentBird recorded visits. of Paradise dominated at these two fruit Ficusgul (Fig. lb) is a free-standing tree, sources,the bird also visited 19 other species similar in most aspectsof its fruiting to the of fruiting plants. The bird's preferencefor figs previous speciesexcept for gul's smaller can- (9% of its diet) was low; capsularfruits formed opy and smaller crop. This fig attracted 15 the bulk of the diet (58%). speciesof foragers, and birds of paradise ac- The RaggianaBird of Paradisewas recorded countedfor 85% of the feeding records. taking fruit of 16 speciesof plants. It dominat- Homalanthusnovoguineensis (Fig. li), a small ed none of the larger fruit-tree speciesbut did subcanopytree, produceda modestcrop of fruit accountfor 55% of avian visits to Dysoxylum. that ripened over a 3-month period. The fruits Raggiana specialized on no single food item of this speciesrequired more effort to harvest but took small amountsfrom a broad array of than the two figs. For this fruit the bird of par- fruits from the three morphologicalcategories. adise plucked the entire capsule from its at- These nine birds of paradisecan be grouped tachment, held it with one foot, and pecked it into three categories:the primarily insectivo- open, taking out the two arillate seeds. I re- rous Black-billedSicklebill and MagnificentRi- corded 11 speciesof birds foraging at Homa- flebird; the fig-specialist manucodes (two lanthus;birds of paradisemade up 93% of these species); and the five species of fruit-eating records. generalists: Superb, Lawes' Six-wired, Mag- Gastoniaspectabilis (Fig. le) produceda large nificent, Raggiana,and Blue birds of paradise. crop of fruit (40,000 +) that was available over Bird foraging:the plant's perspective.--I now a short season (11/2 months). I recorded 131 examine fruit consumption from the plant's feeding visits to the tree and found that only 8 BRUCEBEEHLER [Auk, Vol. 100 six species of birds took the fruit, all birds of nership with forest trees as legitimate fruit-eaters paradise. there has been an extraordinaryproliferation of Chisochetonweinlandii (Fig. 11)was even more specieswith fantasticand beautiful plumagesand extreme in its attractiveness to birds of para- courtshipdisplays. dise. I recorded91 foraging bouts by nine bird species,all birds of paradise. Chisochetonpro- With respectto nesting ecology,several fac- duced a crop of about 4,500 fruits that ripened tors are involved: (1) feeding nestlings effi- over a 3-month season. The arillate seeds re- ciently, (2) nestpredation, (3) size of the clutch, quired some effort to harvest from their cap- and (4) the need to have a small and cryptic sules.The tree attractedno largecrowds of birds nest (Snow and Snow 1979, Willis 1979). It is but insteadwas visited by a smallsteady stream argued that, because nest predation in the of birds of paradise that rarely took more than tropics is high, individual investment per two or three arillate fruit per visit. nesting attempt should be low (hence a small clutch is best); the nest should be small and DISCUSSION cryptic to avoid attracting predators (again, promoting small clutchsize). Under thesecon- I believe aspectsof frugivory are of greatim- ditions, presumably one parent will be suffi- portanceto the evolutionof polygamyin birds cient to provision the nestlings, especiallyif of paradise.I review this issueagainst a back- the parent can efficiently feed the nestlings on ground of hypotheses that have been pre- regurgitatedfruit pulp (Snow and Snow 1979). sented in the literature to explain the evolution These factors promote male emancipation and of polygamous lek behavior in neotropical increasethe opportunity for polygamy. songbirds,and I considerthe birds of paradise If there are real links between nesting ecol- as examples that can either support or refute ogy and the evolutionof polygamy,one might these hypotheses. expect consistentdifferences between monog- Of the nine speciesof birds of paradise that amousand polygamousspecies. One might ex- I studied on the Mount Missim site, seven are pect that monogamous species build better- polygamous and two are monogamous (Gil- protectedor invulnerable nests, or have larger liard 1969, Beehler 1983). The two monoga- clutches,or both. Among the nine speciesad- mous speciesare congenericand apparently are dressedin this paper, there is no suchdichot- similar in many of their habits (unpubl. data). omy. Clutch size for the two monogamous In contrast,the polygamousspecies show both species is two eggs. For the 7 polygamous morphologicaland behavioral diversity (Gil- speciesit is two in 4 casesand one in 3 cases. liard 1969, Cooper and Forshaw 1977). The In fact, most forest songbirds in New Guinea Raggiana Bird of Paradise displays in a true produce a clutch of two, with occasionalde- lek. The Magnificentand Lawes'Six-wired birds viations to single- or triple-egg clutches(from of paradisedisplay at solitaryterrestrial courts. data in Rand and Gilliard 1967). The remaining four speciesdisplay solitarily in All nine speciesof birds of paradisebuild the middle levels of the forest vegetation. simple open nests attached to tree limbs, at Much has been written to explain the eco- variousheights above the ground. Again, there logical conditions that predispose tropical is no consistentdifference between monoga- songbirds to promiscuousarena behavior (D. mous and polygamous species. Becausethe Snow 1971a, 1976, 1980; Snow and Snow 1979, nine specieslive together in the sameforest on Willis et al. 1978, Willis 1979, Ricklefs 1980). Mount Missim and elsewhere, one can pre- The common theme in all of these relates to sume that nest predation is equivalent among two factors: (1) the advantagesof fruit-eating, the nine species,and any argumentabout the and (2) the exigenciesof nesting in the tropical relation between nesting ecologyand polyga- forest. my breaks down. Snow and Snow (1979) state Aproposof frugivory, Snow (1976:ix) writes: that the evolution of promiscuous arena dis- It is, apparently,the relief from the need to spend play has not developedin hole-nestingbirds most of the day in searchof food that has given the (another aspectof the nest-predationhypoth- fruit-eatersthe opportunity to develop other activi- esis).This, too, is contradictedby birds of par- ties to a degreenot seenin the insect-eaters.In a few adise. The King Bird of Paradise (Cicinnurus of the bird families that have entered into closepart- regius)nests in a hole and displaysin an arena January1983] Frugivoryin Birdsof Paradise 9 system (Gilliard 1969; Cooper and Forshaw How do paradisaeiddiets comparewith those 1977). of other speciesin other tropicalbird families? Addressingthe relationof frugivoryand po- The polygamousBearded Bellbird (?rocnias av- lygamy, Ricklefs(1980: 476) writes: erano) was recorded to take the fruit of 32 speciesof food plants,including five families ! would predictthat the fruits eatenmainly by pro- used by birds of paradise: Araliaceae, Myrta- miscuous speciesand those eaten by monogamous ceae, Rubiaceae,Euphorbiaceae, and Myristi- frugivores(mainly tanagersamong passerinesin the New World tropics)have distinctlydifferent patterns caceae (B. Snow 1970). The lek-displaying of dispersion, availability, and, perhaps, nutritional White-bearded Manakin (Manacus rnanacus) value. took fruit of 105 species,including members of five plant families also used by birds of para- Within limits, this suggestionappears to have dise: Zingiberaceae, Moraceae, Euphorbi- real merit, especially as it relates to birds of aceae, Araliaceae, and Rubiaceae (Snow 1962a). paradise.The two monogamousbirds of par- There are morphologicalsimilarires between adise in the group that I studied were both fig the specializedfruits from New Guinea and the specialists.The sevenpolygamous species took neotropics.Some of thesesimilarities are based a wide rangeof fruits, but many from the cap- on taxonomic affinities. The capsular species sule and drupe/berry categories.What can be of Meliaceaeand Myristicaceaein both regions made of thesedifferent dietary preferencesand are strikingly alike, with golden-brown cap- behaviors? Data on avian visitors to fruiting sular exocarpsand orange-redarils (cf. Viola plants show that figs are exploitedby a wide and Myristica in Central America and New rangeof bird families,while many of the fruits Guinea, respectively).In this samevein, these from the other'two categoriesare visited by a specializedfruits are arrangedon the plant in narrower segmentof the bird community, and patternsthat make them availableonly to a se- some are used only by birds of paradise. I sug- lect group of foragers,who often expendcon- gest that aspectsof this coevolutionaryrela- siderable energy to harvest these fruits [ex- tionship between birds of paradise and some emplified by Nectandra (Lauraceae) and its fruiting plants may contain an answer to why cotingidforagers (B. Snow1977) and Chisoche- the birds were able to evolve their remarkable tonand its paradisaeidforagers (Beehler 1983)]. polygamousmating systems.One aspectmay There are some nonconformities in the fruit- be nutritional. The capsular arillate fruits are polygamy relationship. Two of the birds of richer in proteinsand lipids and poorerin water paradise that I studied, the Black-billedSick- and carbohydratesthan the simpler and small- lebill and MagnificentRiflebird, take many ar- er fruits, typified here by many speciesof figs thropodsin their diets. In the caseof the Sick- discussed(Crome 1975, Foster1978, Frost 1980). lebill, most of its diet is arthropod matter. On Thus, much of a bird's nutritional needs can an intercontinental scale, the New Guinea be supplied by these specializedfruits. In ad- speciesthat are polygamoustend to take many dition, the specializedfood plants have, on the arthropodsin their diets, while some,or many, whole, longer fruiting seasonsthat offer reg- of the polygamousspecies in the neotropicsare ular visitors a small daily complement of ripe closeto being entirelyfrugivorous (Snow 1976, fruits (strategyI of Howe and Estabrook1977; Beehler1983). In New Guinea, speciesof cuck- seealso McKey 1975,Frost 1980,Beehler 1983). oo-shrikes (Campephagidae), honeyeaters The fruits are predictable, fixed in space, and (Meliphagidae), and berrypeckers(Dicaeidae) can offer nutritional rewards usually available form a large assemblageof the passerinefru- only from arthropods, a resourcethat is con- givorecommunity (Beehler 1981). No members siderably more costly to harvest (Snow 1976, of any of thesegroups, however, show any evi- Snow and Snow 1979). The specializedfruits, denceof the polygamousarena behavior ex- then, provide the nutrients needed for adult hibited by the birds of paradise.In fact, the and nestling nutrition, yet are simpler to har- majority of fruit-eating birds of the world are vest than the elusive insect resource.The gain monogamous,while mostof the speciesof Par- in time and nutrition can allow more time for adisaeidaeand Pipridae are polygamous. the female alone to care for the offspringand I believe the reason for this apparent anom- for the emancipatedmale to pursue a polyga- aly to be correctlystated by Bradbury (1981), mous mating strategy. whose studies of bats in tropical Africa offer 10 BRYCEBEEI-ILER [Auk,Vol. 100
other examplesof rainforestvertebrates that are tiffcationwas provided by ThaneK. Pratt,J. R. Croft, both frugivorousand polygamous.Bradbury E. E. Henty, and D. G. Frodin. For their intellectual argues that the peculiarities of frugivory are stimulationand guidancewith fieldworkI thank John important but do not constitutethe sole suffi- Terborgh,Henry Horn, and S. Dillon Ripley. M. S. cient factor promoting the evolution of arena Foster,R. F. Pasquier,and two anonymousreview- ersmade numerous improvements on earlydrafts of polygamy.It is the fruit diet that promoteslarge this paper.This work constitutesa part of my Ph.D. homeranges, with high overlap,which, in turn, dissertationresearch and was supportedby grants raises the potential for high interindividual from SmithsonianInstitution, Princeton University, contact rates. Frugivores share ranges and and Frank M. Chapman Fund. fruiting plants. Birds are highly vagile. Thus an individual malewho attemptsto attractand LITERATURE CITED courtnumbers of femaleswill have greatersuc- BEEHLER,B. 1981. Ecologicalstructuring of forest cessif the populationis frugivorous,rather than bird communitiesin New Guinea. Pp. 837-861 insectivorous, site-faithful, and territorial. As in Biogeographyand ecologyin New Guinea (J. discussedin the earlier arguments,frugivory L. Gressitt, Ed.). The Hague, Junk. also increases the likelihood of male emanci- ß 1983. The behavioral ecologyof four birds pation from nesting duties. These two phe- of paradise. Unpublished Ph.D. dissertation. nomena, occurring together, createan environ- Princeton, New Jersey, Princeton Univ. ment with high potentialfor polygamy. BRADBURY,J. W. 1981. Evolution of leks. Pp. 138- 169 in Natural selectionof socialbehavior (R. D. Given the diversityof fruit in the tropics(Fig. Alexander and D. Tinkle, Eds.). New York, Chi- 1), and the range of nutritional constituentsof ron. variousfruit types (Crome 1975,Frost 1980), in COOPER,W., & J. FORSHAW.1977. The birds of par- conjunctionwith great potential differencesin adise and bowerbirds. Sydney, Collins. production strategiesand spatial dispersions CROME,F. 1975. The ecologyof fruit pigeons in (Beehler 1983), it is evident that the term "fruit- tropical Northern Queensland. Australian Wildl. eating" subsumesa multitude of ecologicalsit- Res. 2: 155-185. uations that can, in turn, produce varied re- DIAMOND, ]. M. 1972. Avifauna of the eastern sponsesin an animal's behavior (Emlen and highlandsof New Guinea. Cambridge, Massa- chusetts, Publ. Nuttall Ornithol. Club. No. 12. Oring 1977). I agree with D. Snow (1980) that EMLEN,S. T., & L. W. ORING. 1977. Ecology,sexual frugivory is an obligatoryfield for study in the selection,and the evolution of mating systems. quest to understand the evolution of polyga- Science 197: 215-223. mous arena display in birds of paradise and FOSTER,A., & E. GIFFORD. 1974. Comparative other tropical forest songbirds. I stress, how- morphologyof vascular plants. San Francisco, ever, that the key to understandingthe frugiv- W. H. Freeman. ory-polygamyrelationship will require analy- FOSTER,M. S. 1977. Ecologicaland nutritional ef- sis of suchfactors as: (1) percentageof fruit in fectsof food scarcityon a tropicalfrugivorous a bird's diet, (2) spatial dispersionof the dif- bird and its fruit resource.Ecology 58: 73-85. ferent food plants, (3) size of fruit crops, (4) 1978. Total frugivoryin tropicalpasserines: intraspecificsynchrony of plants, (5) predict- a reappraisal.Tropical Ecol. 19: 131-154. FROST,P. 1980. Fruit-frugivore interactions in a ability of fruiting cyclesof individual plantsas South African coastaldune forest. Pp. 1179-1184 well as speciespopulations, (6) the ripening in Acta 17th Congr. Intern. Ornithol. (R. Nohr- synchronyof cropsand length of fruiting sea- ing, Ed.). Berlin, DeutscheGesellsch. son, and (7) nutritional constituents of the fruit GILLIARD,E. T. 1969. The birds of paradise and taken. These and additional factors define the bower birds. Garden City, New York, Natural spatio-temporalavailability and quality of a History Press. bird's food resources,which influenceranging --, & M. K. LECROY. 1967. Annotated list of and foraging patternsand ultimately affect so- birds of the Adelbert Mountains, New Guinea. cial and sexual habits. Bull. Amer. Mus. Nat. Hist. 138: 53-81. HOLDRIDGE, L., W. GRENKE, W. HATHEWAY, T.
ACKNOWLEDGMENTS LIA•G & J. Tosi. 1971. Forest environment in tropical life zones: a pilot study. Elmsford, New For help with fieldwork in Papua New Guinea I York, PergamonPress. thank Ninga Kawa, Michael Yamapao, the staff of HOWE, H. 1979. Fear and frugivory. Amer. Natur. the Wau EcologyInstitute, and the Wildlife Division, 114: 925-931. Office of Environment.Expertise in botanicaliden- --, & G. ESTABROOK.1977. On intraspecific January1983] Frugivoryin Birdsof Paradise 11
competition for avian dispersersin tropical trees. adise and bowerbirds, a resynthesis.Pp. 137- Amer. Natur. 111: 817-832. 149 in Proc. Intern. Ornithol. Congr. No. 16 (H. JENSEN,W., & F. SALISBURY.1972. Botany: an eco- Frith and J. Calaby, Eds.). Canberra, Australian logical approach. Belmont, California, Wads- Acad. Sci. worth. SNow, B. 1970. A field study of the BeardedBell- JoHNs,R. J. 1976. Common forest trees of Papua bird in Trinidad. Ibis 112: 299-329. New Guinea. Vol. 8, part 3. Angiospermae.Bu- 1972. A field study of the Calfbird Peris- lolo, Papua New Guinea, P. N. G. Forestry Col- socephalustricolor. Ibis 114: 139-162. lege. 1973. Notes on the behavior of the White LECROY, M. K., A. KULUPI, & W. S. PECKOVER. Bellbird. Auk 90: 743-751. 1980. Goldie's Bird of Paradise:display, natural 1977. Territorial behavior and courtship of history, and traditional relationshipsof the peo- the male Three-wattled Bellbird. Auk 94: 623- ple to the bird. Wilson Bull. 92: 289-301. 645. 1981. The genus Paradisaea:display and SNow, D. 1962a. A field study of the Black and evolution. Amer. Mus. Nat. Hist. Novitates 2714: White Manakin, Manacus manacus, in Trinidad. 1-52. Zoologica 47: 65-104. LILL, A. 1976. Lek behavior in the Golden-headed 1962b. A field study of the Golden-headed Manakin Pipra erythrocephalain Trinidad (West Manakin, Pipra erythrocephala,in Trinidadß W. Indies). Fortsch. Verhalt. 18: 1-83. I. Zoologica 47: 183-198. McALi'INE, D., G. KEIG, & K. SHORT. 1975. Cli- 1971a. Evolutionary aspectsof fruit-eating matic tables for Papua New Guinea. Division of in birds. Ibis 113: 194-202. Land Use Tech. Pap. No. 37. Melbourne, 1971b. Observationson the Purple-throat- C.S.I.R.O. ed Fruitcrow. Living Bird 10: 5-17. McKEY, D. 1975. The ecologyof co-evolved seed 1976. The web of adaptation. New York, dispersalsystems. Pp. 159-191 in Coevolutionof Quadrangle. animals and plants (L. Gilbert and P. Raven, Eds.) 1980. Regionaldifferences between tropical Austin, Texas, Univ. Texas. florasand the evolution of frugivory. Pp. 1192- PIlL , VAN DER, L. 1972. Principles of dispersal in 1198 in Acta 17th Congr. Intern. Ornithol. (R. higher plants, 2nd ed. New York, Springer-Ver- Nohring, ed.). Berlin, Deutsche Gesellsch. lag. --, & B. SNow. 1979. The Ochre-bellied Fly- RAND, A. L. 1940. Breeding habits of the birds of catcher and the evolution of lek behavior. Con- paradise Macgregoriaand Diphyllodes. Amer. dor 81: 286-292. Mus. Novitates. 1073: 1-14. WILLIS, E. O. 1979. Commentary. Condor 81: 324. , & E. T. GILLIARD. 1967. Handbook of New D. WESCHLER, & Y. ONIK•. 1978. On be- Guinea birds. Garden City, New York, Natural havior and nesting of McConnell's Flycatcher History Press. (Pipromorphamacconelli): does female rejection R•CKLEFS,R. 1980. Commentary. Auk 82: 476-477. lead to male promiscuity?Auk 95: 1-8. SCHODDE,R. 1976. Evolution in the birds-of-par- BRUCEBEEHLER [Auk, Vol. 100
X X XX X X X X XX
x