The Condor98:722-735 0 The Cooper Ornithological Society 1996

DISPLAY BEHAVIOR AND NATURAL HISTORY OF THE YELLOW-CROWNED (HETEROCERCUS FLA VIVERT’X PIPRIDAE)’

RICHARD0. PRUM~ Natural History Museum and Department of Systematics& Ecology, Universityof Kansas,Lawrence, KS 660452454, e-mail: [email protected]

JOSEPHD. KAPLAN BioDiversityInc., I6 Lafeyette St., Yarmouth, ME 04096

JAN E. PIERSON Box 248, South Harpswell, ME 04079

Abstract. The behaviorand natural historyof the Yellow-crownedManakin (Hetero- cercusflavivertex,Pipridae) were observed on theRio Ventuariin northernTenitorio Federal Amazonas,. Male flavivertex have a solitarylek breedingsystem, a displayrep- ertoireincluding five distinctvocalizations, several display elements, and a probableme- chanicaldisplay sound.The tail-shiver display,in which the cockedtail is prominently featured,has not beenobserved in otherHeterocercus species, but this displayelement may be an explanationfor the evolutionof the uniquetail shapein the genus.Otherflavivertex displayelements, such as the above-the-canopy display flight, are behavioral homologs shared with other Heterocercusspecies. Female H. flavivertex constructhanging cup nests that are similar in architecture,composition, and placementto othermanakin species.The diet of H. flavivertex is apparentlycomposed of a greaterquantity of insectsthan other manakin species.The evolutionof display,morphology, diet, and socialbehavior in Het- erocercusis discussed. Key words: lek; courtship;display: song; .

INTRODUCTION kins. The Yellow-crowned (H. flavivertex), Flame- The Neotropical manakins are well known for crowned (H. linteutus), and Orange-crowned (H. their striking sexually dimorphic plumage, elab- auruntiivertex) Manakins are distributed allo- orate courtship displays and lek breeding systems patrically around the Amazonian basin (Fig. 1; (Snow 1963b, Sick 1967, Prum 1990a, 1994). Snow 1979, Haffer 1986, Ridgely and Tudor Comparative studies of the manakins demon- 1994). All three specieslive in seasonallyflooded strate that lek breeding is primitive to the family, varzea forests, blackwater riparian or sandy soil and that many aspectsof manakin display and woodlands (Hilty and Brown 1986, Sick 1993, reproductive behavior vary phylogenetically Ridgely and Tudor 1994, R. 0. Prum, pers. ob- among closely related genera and speciesof man- serv.). The genus is characterized by dark green akins (Prum 1990a, 1990b, 1992, 1994). Al- upperparts, chestnut underparts, a sexually di- though the behavior of many manakin species morphic white throat patch with long side plumes, has been studied in detail, there are no published carotenoid-colored crown patch, and an unusu- accounts of the behavior and natural history of ally graduated tail with the outer rectrices short- almost half the speciesin the family (Prum 1990a, est (Fig. 2). The behavior of Heterocercuslintea- 1994). Additional work on these poorly known tus is known only from a single anecdotal ob- speciesis essential to advancing our understand- servation by Sick (1959, 1967, 1993). Sick de- ing of the evolution of behavior of the family. scribed the advertisement call of linteatus as pit- The genus Heterocercusincludes three species eew, and stated that linteatus displays by spread- that are among the most poorly known mana- ing the throat patch feathers to the sides like a “broad beard” (Sick 1993). There are no other published observations of the vocalizations, dis- I Received19 October 1995. Accepted 23 July 1996. plays, breeding system, nest, or diet of the genus. z Correspondingauthor. We conducted field work on Heterocercusflaviv- ~7221 HETEROCERCUS FLAVIVERTEX BEHAVIOR 123

FIGURE 1. Distribution of the three speciesof the genus Heterocercus:the Yellow-crowned Manalcin (H. fluvivertex) in diagonal hatching, Flame-crowned Manakin (H. linteutus)in horizontal hatching, and Orange- crowned Manakin (H. aurantiivertex)in cross hatching. The location of the study site, Campamento Camani on the Rio Ventuari, Venezuela, is shown by the dot. ertex on the Upper Rio Ventuari, Venezuela to territorial male jlavivertex was located on 11 address this lack of natural history data for the April. Structuredobservations of both males were genus. conducted daily until 20 April (by which time vocalization and territorial attendance had de- LOCALITY AND METHODS The observations were conducted at Campa- mento Camani (formerly called Junglaven Lodge) on the upper Rio Ventuari, Territorio Federal Amazonas, Venezuela (- 5” 08’N, 66” 1S ’W; 200 meters; Fig. 1). The locality is an area of mixed seasonally flooded varzea forest, white-sand woodland, open white-sand savanna, and terra firma forest. Over 300 speciesof have been observed at this site (Zimmer and Hilty, unpubl. data). Observations were made at Campamento Ca- mani from 10-l 1 March 1990 and 22 March- 20 April 1990. At the beginning of the obser- vations, nothing was known about the vocali- zations, behavior, or natural history of H. jla- vivertex. Becauseof this complete lack of infor- mation and the tremendous diversity of the avi- fauna at the site, Prum and Kaplan spent 10-l 1 March and 22 March-5 April 1990 searchingfor H. jlavivertex, frequently in inappropriate habi- FIGURE 2. Ventral view of the rectrices of a male Heterocercusjlavivertex (USNM 326942). The outer tats. On 6 April, Prum and Kaplan identified the rectrices (6-6) are less than half the length and width first territorial male flavivertex and began struc- of the inner rectrices (l- 1). The feather rachi are not tured observations of this individual. A second stiffened. Scale bar equals’10 mm. 724 R. 0. PRUM ET AL. clined significantly). Additional observations American Museum of Natural History (AMNH), were made at the same locality by Kaplan and New York, and Paris Museum of Natural His- Pierson from 16-22 February 1992. tory (PM). Additional observations were made Mist nets were deployed on 7-9 April 1990, of study skins in the Collecion Phelps, Caracas, and fourflavivertex were caught and banded with Venezuela. Museum specimens offlavivertex in distinctive color combinations. Three banded in- the Colleci6n Phelps were examined to investi- dividuals (Green-White, Yellow-Blue, and White- gate reproductive seasonality(this is the only sig- Red) were females or predefinitive males and nificant collection of jiavivertex’ with any infor- were never observed again. A territorial male in mation on reproductive condition). A drawing definitive plumage was banded Orange-Orange. of the gonadswas made on the data label of each The second territorial male was not banded to specimen at the time of preparation. Males with minimize disturbance at that display site. testes > 5 mm long, and females with ovaries > A total of 9 1.3 hours of structured observa- 5 mm or follicles > 2 mm were considered to tions of the two territorial males were made dur- be in breeding condition. The single nest col- ing the observation period. Focal individuals were lected was deposited at the United States Na- observed with binoculars and without blinds. tional Museum of Natural History (USNM), Observations were recorded using five minute Smithsonian Institution, Washington, DC. sampling periods. The following data were re- Illustrations of the displays were prepared by corded for each sampling period: presenceof the Carolyn Wooster from video tape recordingsand male, and frequency of exits and entrances, po- study skins borrowed from the National Muse- sition and height of the perch, the number of um of Natural History, Smithsonian Institution, perch changes, the number of the various vo- Washington, DC. calizations and displays, intraspecific encounters and aggressiveflights, preening, sunbathing, for- RESULTS aging and type of food item, and method of for- aging. Rare behaviors may have been underes- MORPHOLOGY timated during the initial observation days be- Maleflavivertex have deep green upperpartswith fore the behavioral repertoire was better known. a yellow central crown stripe. The underparts Daily observations started by 06:30 and contin- blend from warm cinnamon on the undertail co- ued until male territorial attendance and vocal verts to dark chestnut on the upper breast. The activity ceased (between 11:OOand 14:OO).Ad- throat is bright white with elongate side plumes. ditional afternoon observations (14:OOto 17:OO) Females lack the crown stripe and have a drabber were made on 6 observations days, but they in- but distinct gray throat patch. Both sexeshave dicate reduced territorial and display activity. a dark gray check patch that contrasts with the Foraging observations of male and femalefla- green upperparts and light throat. Some males vivertexwere made opportunistically throughout with reproductively active testeshave a female- the observation period. Observations of terri- like, predefinitive plumage (specimens in Colle- torial males were more thorough becauseof ad- cion Phelps). Time of acquisition of definitive ditional observations in their display territories. male plumage is unknown. Over 6 hours of acoustic tape recordings of The tail of both sexesis an unusual shape for vocalizations were made using a Sony TCM-5000 manakins (Fig. 2). The outer rectrices (6-6) av- and SennheiserME-SO and ME-88 microphones. erageless than half the length of the longest inner Over 4 hours of 8mm video tape recordingswere rectrices(1- 1) (Table l), and are much narrower. made using a Sony Sports Handycam CCD-SP9. The sexes show little size dimorphism, except Vocalizations will be deposited at the Cornell that males have significantly longer wings than University Library of Natural Sounds. Record- females (P < 0.002; Table 1). ings are referred to here by the original ROP tape Few body mass data are available for this spe- reference numbers which give the tape number cies. The three female/predefinitive male indi- and side. viduals banded weighed 21.5, 21.5, and 18.3 g, Vocalizations were analyzed using Canary 1.2 respectively. A single definitive plumage male from the Cornell University Library of Natural (00) weighed 2 1.3 g. Willard et al. (199 1) re- Sounds on a Macintosh Computer. Measure- ported no sexual dimorphism in mass in a sam- ments reported below are from specimensof the ple of five males (K f SD = 21.5 & 1.9 g) and

726 R. 0. PRUM ET AL.

TABLE 2. Frequency of vocalizations and displays of the Yellow-crowned Manakin (Heterocercusflavivertex) in Territorio Amazonas, Venezuela.

Obshwvrvion Obszy;tion % territorial Weeee--h-mm W&-CC Tail-shiver Aggressivefight attendanw b cellb displayb and chatterc&b

All observations 93.1 19 3.48 3.0 0.13 0.66 Peak activity” 30 4 5.76 3.72 0.13 0.60 Low activityd 9 3 78 0.34 0.01 0 0.15

mTerritorial attendanceis reportedas the % of observationperiods when a male was observedon territory. b Displaysand vocal frequenciesare reportedas averagenumber of observationsper how of observation. *Peak activity frequencieswere calculated from data collectedon four mornings(07:0&12:00) between1 l-14 April 1990. d Low activity frequencieswere calculated from data collectedon threemornings between 16-20 April 1990.

3 females (X + SD = 2 1.7 f 1.3 g) from southern resource-basedterritories in varzea forest for a Amazonas, Venezuela. period of two weeks. Territoriality was inferred by consistent attendance in an area over a num- DISTRIBUTION AND HABITAT ber of days. The two males were in attendance Heterocercusflavivertexis distributed in the Rio on their territories during 73% ofthe 1,096 5-min Negro drainage of eastern , southern observation periods on 19 observation days Venezuela and , the Upper Rio Orinoco (Table 2). Each territory included a core area 10 drainage of Venezuela as far north as the Rio m in diameter where males were observed > 75% Capanaparo, Estado Apure, Venezuela (Good- of the time in attendance.The entire area of perch win and Lentino 1992; specimen Collecion sites where territorial defense (i. e., vocalization Phelps, Caracas), and on the left bank of the or display) was observed was approximately 20 Amazon in Para, Brazil (Fig. 1; Snow 1979, Haf- m in diameter. fer 1986, Ridgely and Tudor 1994). All known Each male territory was solitary-i.e., com- localities are below 300 m in elevation. The pletely out of hearing distance from other male Colleci6n Phelps includes 104 study skins from flavivertex territories. The two territories ob- 12 localities along rivers throughout Territorio served were separated by -250 m of relatively Amazonas, Venezuela, representing essentially disturbed habitat without any intervening males. every locality sampled in Amazonas within ap- However, both territories were on the edge of a propriate habitat. large expanse of continuous habitat that lacked Heterocercusflavivertexwas observed at Cam- nearby territories. Twenty-two months later dur- pamento Camani in seasonally flooded varzea ing February 1992, 00 and an unbanded male forest of the Rio Ventuari and its tributaries, in occupied the same two territories and were ob- scrubby woodlands found on white sand soils served using the same calling perches. No ad- that border the varzea, and in seasonallyflooded ditional intervening male territories were found riparian forest along blackwater streamsand ox- at that time. bow lakes. Other sources report jlavivertex in Males perchedon tree branchesand horizontal similar habitats in other parts of its range (Hilty vines at heights of 3-6 m. Males defended ter- and Brown 1986, Ridgely and Tudor 1994). The ritories by vocalizing and performing aggressive sitesat Campamento Camani were characterized flight displays (see below). Male territories in- by a rather short canopy (- 8-l 5 m) and thick cluded a few fruit trees, but males frequently left undergrowth of vines, palms and small trees. H. the territory to forage. Female-plumaged birds flavivertex was never observed in terra firma for- were not observedin the territories for more than est in this area. A few individuals were observed a few seconds. foraging in small shrubs in open savanna that border the white-sand woodlands. These indi- VOCALIZATIONS viduals apparently were foraging opportunisti- Maleflavivertex have a repertoire of at least five cally at the edge of their preferred habitat. distinct vocalizations with several minor varia- tions. The full advertisement call is a loud, three- MALE TERRITORIALITY note, whistled weeee-pitch-ooo(Fig. 3) approxi- Two maleflavivertex-Orange-Orange (00) and mately 3 set long. The first note of the call rises unbanded (UB) - were observed defending non- slowly from - 4.5 to 5.25 kHz over 1.6 sec.After HETEROCERCUS FLA VIVERTEX BEHAVIOR 727

< 0.1 set note that descendsfrom 8.3 to 5.6 kHz (Fig. 5A). Two additional vocalizations-a querulous kweip and a loud keek-were heard on a few occasions immediately following performances of the tail-shiver display. The kweip note was recorded once. It consistsof three very short syl- lables in quick succession:5.9 kHz for 0.15 set, 6.33 kHz for 0.05 set, and 5.5 kHz for 0.05 set (Fig. 5B). The keek note was not recorded.

DISPLAY BEHAVIOR AND MECHANICAL SOUND PRODUCTION Three conspicuousdisplays were observed. Male FIGURE 3. Spectrogram(above) and waveform(be- low) of the full advertisementcall, weeee-pitch-ooo,of flavivertex perform a stylized, erect display pos- Heterocercusflavivertex (ROP- 1990-V1). Khz: kilo- ture while giving the full advertisement call- hertz(frequency); mPa: micropascal(pressure). weeee-pitch-ooo(Fig. 6A). The male perchesvery erect with its head pulled in and the beak up- raised, exposing the white throat patch. During a 0.6 set pause,the secondnote risesrapidly from the initial weeeenote, the male gradually opens - 4.5 to 6 kHz over 0.2 sec. Then, after a very its beak wider and wider. At the beginning of the brief pause of < 0.1 set, the third note descends explosive pitch-ooo notes, the male throws his slowly from 6 to 3.8 kHz over - 0.8 sec. The beak open very broadly, exposing the bright yel- amplitude, or volume, of the call increasesdra- low tongue and gape.At the same time, the wings matically in the first note of the call, peaks in the and body twitch conspicuouslywith physical ex- secondnote, and drops off gradually in the third ertion, as they do in the Screaming Piha (Lipau- (Fig. 3) giving the call a dramatic, punctuated, gus vociferans,Cotingidae; Snow 1982) when it or hiccuping cadence. The erect display posture calls. associatedwith this vocalization is describedbe- Territorial males give the weeee-pitch-oooand low (see Display Behavior). weeeecalls as advertisement and territorial de- Sometimes the full advertisement call includ- fense vocalizations. These vocalizations were ed a double introductory weeeenote. Males also given an average of 3.48 and 3.0 times per hour, gave a partial advertisement call. This call con- respectively, across all observation days (Table sistsof a single rising weeeenote that is identical 2). Calling frequency varied from 5.76 weeee- to the initial note of the full advertisement call pitch-000 and 3.72 weeeecalls per hour during but typically of lower volume. Males occasion- the morning (07:00-12:00) on four of the most ally give a double version with two weeeenotes active days, to 0.34 and 0.01 calls per hour, re- repeated. spectively during the mornings of three of the Males give an emphatic weer-weer . . . weer last observation days (Table 2). The highestcall- chatter call during aggressiveinteractions with ing frequency recorded was 7.6 weeee-pitch-ooo other individualflavivertex and occasionallyoth- calls per hour on one morning. er species.The aggressivechatter call is charac- terized by three to six sharp, squeaky notes that are successivelylower in volume and frequency. Each note is composed of two strong harmonics that rise rapidly in frequency over the first 0.06 set and then drop slowly in frequency over the last - 0.15 set (Fig. 4). Each subsequentnote in the series is slightly shorter and lower in ampli- 0.5 1 1.5 2 tude and frequency. Seconds Males also give a dry, emphatic chli, as they FIGURE 4. Spectrogramof theaggressive chatter call, fly from a perch immediately following the tail- weer-weer weer,of Heterocercusjlavivertex(ROP- shiver display (see below). The ch@ is a short, 1990-IA). 728 R. 0. PRUM ET AL.

A. frequency was 2.76 performances per hour be- tween 12:00-14:lO on a single observation day. The performance of the display was usually pre- 15- ceded and followed by an increase in advertise- ment vocalization. Usually, the male was ob- N 5 lo- served looking intently off into the forest im- mediately before performing the display. These 4 displays may have been directed toward an un- 5- seen female. On one occasion, J. Pierson saw an elaborate I I flight display from a trail through open savanna 0.5 1 .o bordering on white-sand woodland approxi- Seconds mately 2 km from the focal male territories. Fol- lowing a series of aggressivechatter calls by one B. or two unseen individuals, a malejlavivertex Ilew rapidly and rather vertically from one side of a trail up about 15-20 m above the woodland can- opy. It then flew horizontally in a semicircleabout 1O-l 2 m in diameter, and droppedabruptly down into the woodland on the other side of the trail. The entire flight took between 6 and 8 sec. No other flight displays were observed. However, an unusual whooshing sound was heard at least 3 or 4 times in the subsequent20-min period. The sound was a high-pitched, - 1 set, down-slurred I I 0.5 1.o whoosh. Seconds Prum or Kaplan did not observe this flight display. However, Prum tape recorded a sound FIGURE 5. Spectrogramsof: (A) the chtp call (ROP- in the territory of male UB that is strikingly sim- 1990-6B) and (B) the kweip note (ROP-1990-6B) of Heterocercusjlavivertex. ilar to the sound later described by Pierson. On that occasion, the male performed a tail-shiver display and flew off the perch giving the sharp In the conspicuoustail-shiver display (Fig. 6B), chfp note. One and a half secondsafter the chip, a male perches on a branch with the body hor- a descending whoosh sound was recorded (Fig. izontal and the tail cocked at a nearly right angle 7). The sound is 0.75 set long and contains two above its back. The white throat is erected into portions. The first part is a continuous whoosh a puff, and the wings are folded against the body. that steadily increases in amplitude. It is fol- The male moves the tail forward and back over lowed immediately by a short thud that punc- approximately 1S ’, and simultaneously rocks its tuates the end of the sound. The recorded whoosh body slightly in the oppositedirection, producing sound is reminiscent of an arrow flying by and a rapid, syncopated“ shivering” movement. The hitting a target; it may be produced by the bird male then flies conspicuously from the perch, flying very fast and stopping in flight suddenly. holding its entire body horizontal, flapping its The sound was not noticed at the time of re- wings very deeply, and giving an emphatic chfp cording, but is actually greater in amplitude call. An 8 mm video tape recording of a single (louder) on the recording than the chri, note that tail-shiver display documented four rapid tail was recorded under excellent conditions. and body twitching movements in less than one The association of this whoosh sound with a second. tail-shiver display supportsthe conjecture that it The tail-shiver display was observed twelve is a mechanical display sound. These observa- times on five different observation days between tions also indicate thatflavivertex may have two 07:15 and 15:55, for a total frequency of 0.13 flight displays. This mechanical whoosh flight displays per observation hour. The peak display display is apparently distinct from the above- HETEROCERCUS FLAVIVERTEX BEHAVIOR 729

A.

FIGURE 6. Display elements of Heterocercusfluvivertex.(A) the erect posture. (B) the tail-shiver display. See text for details. the-canopy display, since no whooshing noises NESTING were heard during the observation of the above- A female jlavivertex was observed constructing the-canopy display, and no above-the-canopy a nest on 3 April 1990 on Cafio Guayahe, a small flights were observed when the whoosh sounds tributary on the right bank of the Rio Ventuari were heard. Further, the possible recording of that is several kilometers upstream from Cam- the whooshwas made 1.5 set after male UB per- formed a tail-shiver display and flew from a perch in the territory. This is not enough time for the male to perform the 6-8 set long above-the-can- opy flight display.

AGGRESSIVE INTERACTIONS The chatter vocalization is apparently given when a male observes a conspecific in the vicinity of its territory. Typically, the resident male would suddenly peer intently at the surrounding vege- tation and then suddenly fly from the perch giv- ing the aggressivechatter call (n = 59). Aggressive flights were observed an average of 0.6 times per observation hour (Table 2). Males were so vig- I I I 1 ilant that on all but one occasion,the secondbird 2 3 1

was never seen or heard by the observers. In Seconds several cases, males directed aggressiveflights FIGURE 7. Spectrogram(above) and waveform (be- and chatter calls at individuals of other species low) of a ch@ display call (left) and a probable me- (e.g., Bucco tamatia). chanical whooshsound (right) (ROP- 1990-IA). 730 R. 0. PRUM ET AL.

TABLE 3. Observationsof foragingbehavior and estimates of diet compositionin maleHeterocercusflavivertex. The estimateof diet compositionincludes seed regurgitation events as indirect observationsof fruit foraging events.

Fruit Arthropods Total foraging events Seed regurgitation Observations 11 70 19 100 49 Frequencyhr- ’ 0.12 0.76 0.20 1.09 0.53 Estimatedpercent diet composition 40 46 14 - -

pamento Camani. The nest was a hanging cup members of the Melastomataceae,Lauraceae, and built in a fork of the lowest, descending branch Rubiaceae (identifications of herbarium speci- of a 15 m tall, broadleafed tree on the edge of mens collected is pending). Cafio Guayahe, which was approximately 8 m Male jlavivertex were often observed foraging wide at this site. The nest was placed 2 m above in and around their territories. In total, an av- the surface of the water on the lowest point of erage of 1.1 foraging events per hour were ob- the branch. The nest was not within a male ter- served (Table 3). In 19 instances, the food items ritory. The femalejuvivertex was observed add- could not be identified. On 11 occasions,males ing stiff, thin fibrous twigs to the nest, and shap- were observed foraging on fruits, which they ing the interior of the nest by pressing the nest seized in the bill during aerial sallies. In addition cup with her breast. The nest was observed again to these actual foraging events, males were ob- two weeks later. It had been abandoned soon served regurgitating seedswhile perched in the after the first observations and was then collected display territories (Table 3). Regurgitated seeds (USNM 48 163). varied in size by a couple orders of magnitude The collected nest may have been incomplete from < 1 mm to > 20 mm in length. when abandoned. The nest cup is transparent, On 70 occasions, we observed males eating and measures57.3 x 45.1 mm at the widest and invertebrates. Invertebrate foraging maneuvers narrowest exterior dimensions, and 22.6 mm included aerial sallies, gleaning from branches deep. The internal depth of the nest cup is 15 and leaves in flight, and, rarely, gleaning from mm. The branches supporting the nest are uni- branches while perched. Males pursued prey in formly 5 mm in diameter, and the center of the very active aerial salliesat all heightsin the forest nest cup is 46 mm from the fork of the branch. from l-10 m. Occasionally, individuals pro- The nest cup is composed of a variety of plant duced loud bill snaps during these aerial sallies. fibers including rootlets, small twigs, and a large Most invertebrate prey were flying insects too number of wiry, round, grass-like fibers. Some small to identify, but prey included a caterpillar plant fibers are very long and are bent and woven 4 cm long and a 2 cm long spider. A male was around into the nest cup without breaking. The once observed foraging for prey by hovering in edgesof the nest are attached to the branches by front of a web of an orb-weaving spider. The massesof spider web. maneuver was repeated five times in quick suc- cession without any apparent prey capture. DIET Heterocercusflavivertex were observed foraging PREENING AND SUNBATHING on a variety of fruits and invertebrates. Both The predominant activity of males on territory sexeswere observed foraging at fruiting trees in was preening. Males were observed preening in varzea forest with other frugivores (including 297 of 1,096 5-min observation periods, or 37% Black-headed Parrot Pionites melanocephala, of the 798 observation periods in which males Cuvier’s Toucan Ramphastos cuvieri, Many- were known to be present on territory. Males banded Aragari Pteroglossuspluricinctus, Span- preened by running the beak through their feath- gled Cotinga cayana, Pompador Cotinga ers. They also scratchedthemselves with the leg Xipholena punicea, raised over the open wing. Preening frequency atronitens, and Golden-headed Manakin Pipra was consistent throughout the day. On one oc- erythrocephalu).These fruit tree speciesincluded casion, male 00 was conspicuously wet when it HETEROCERCUS FL.AVIVERTEX BEHAVIOR 731

returned to its territory, having apparently bathed through December. Willard et al. (199 1) reported in standing water. enlarged testes and ovaries in specimens from Territorial males were observedsunbathing on southern Territorio Amazonas, Venezuela dur- 27 occasions(an average of 0.26 times per hour). ing February, March and April, and one female During sunbathing, males perchedbetween 5- 10 individual with small gonads in February. m high in a patch of sun, and spread and angled The Amazonas region of Venezuela is subject the wings and tail so that their upper surfaces to strongly seasonalrainfall, with prolonged dry were exposeddirectly to the sun. Frequently, the seasonsfrom December to April and rainy sea- males cocked their tails or lay prostrate against sons from April until November (Hilty and a perch to exposethemselves to the sun efficient- Brown 1986). These limited behavioral obser- ly. These postureswere so conspicuousand ste- vations and museum specimensindicate that the reotyped that they were initially interpreted as breeding season of H. jlavivertex in Amazonas, potential display elements. Repeated observa- Venezuela occurs during the dry season,proba- tions, however, confirmed these postures as hy- bly beginning by February and ending toward giene behavior. Bouts of sunbathing were ob- the end of April or May. served at various times from 08: 15 to 16:05 and lasted from 10 to 120 sec.Immediately after sun- DISCUSSION bathing, males returned to calling perchesin their BREEDING SYSTEM territories and began preening vigorously. Our observations indicate that Heterocercusjla- SEASONALITY vivertexhas a display polygyny breeding system, The observation of nest construction by a female like most other behaviorally known manakin on 3 April 1990 and male territoriality and dis- species.Males occupy small, non-resource-based play behavior between 6-23 April 1990 indicate defense territories and perform elaborate court- that H. jlavivertex were actively breeding during ship displays. Unlike other non-cooperative this month. During April 1990, male territorial manakins species, in which male display terri- activity and frequency of vocalization dropped tories are aggregated into groups of adjacent, dramatically (Table 2), indicating that at this site neighboring males, the territories of male jlavi- the end of the reproductive season may have vertex are apparently solitary. Both males ob- been approaching. For example, the frequency served defended territories that were out of of weeee-pitch-ooocalls averaged 5.76 calls per acoustic range of other males. hour during five mornings between 11-I 4 April, This apparent, complete dispersion of male whereas the frequency of this vocalization by territories could be hypothesized to be causedby these same individuals was 0.32 calls per hour low population density (Bradbury 198 1). Few during three mornings between 16-20 April (Ta- data are available onflavivertex population den- ble 2). Additional unstructured observations at sity, butflavivertex appearsto be relatively com- the site by Kaplan in 1992 indicate that male mon in appropriate habitat at this and other lo- vocal activity was even greater in mid-February calities (Hilty and Brown 1986). H. jlavivertex than in April. were collected regularly during expeditions to lo- Specimens from different years and localities calities throughout Territorio Amazonas by the throughout Amazonas, Venezuela indicate that Collecion Phelps. Evidence from populations of males are reproductively active in February, the White-bearded Manakin Manacus manacus March, and May (n = 36), and have smaller testes indicate that lek size can changewith population in January, July, and August (n = 6). A few male density, but other aspects of lek territoriality, specimens showed decreasedtestes size in May such as territory size and intermale distance, re- (n = 4). Female specimens had enlarged ovaries main unchanged (Olsen and McDowell 1983). in late February and March (n = 4), but other The behavior of male jiavivertex’ supports the specimens showed smaller ovaries in January, conclusion that male territories are solitary for February, March, May, and August (n = 12). All social rather than demographic reasons. Male reproductively active specimens of both sexes flavivertex responded extremely aggressivelyto fall between 11 February and 19 May, but no any conspecificindividuals near their territories. specimenswith gonad data exist for April or June, Aggregateddisplay territories with multiple ad- or during the rainy season from September jacent males require at least a moderate amount 732 R. 0. PRUM ET AL. of social tolerance among males. MaleJlavivertex in association with any nests, and apparently fe- are strikingly different from other manakins in males perform the parental care duties alone. that they show complete intolerance to conspe- cific males. This indicates that solitary display VOCAL AND DISPLAY BEHAVIOR territories are typical of the breeding system of The vocal and display repertoires of Heterocer- this species.Interestingly, a similar, solitary dis- cus jlavivertex are unlike those known in other play territoriality has been observed in the Or- genera and speciesof the family. The dramatic ange-crowned Manakin H. aurantiivertex on the advertisement call is unlike any vocalization Rio Napo, Ecuador (Prum, unpubl. data). Al- known from other manakin genera, but may be though more data are required to determine similar and homologous to the advertisement whetherflavivertex territories are uniformly dis- call of H. linteatus, describedby Sick (1993:501) tributed or clumped within the habitat, Hetero- as pit-eew. Interestingly, the advertisement call cercus territories are apparently distinct from of H. aurantiivertex is completely different from other behaviorally-known species of manakins the two other speciesin the genus(Prum, unpubl. in that territorial males are physically and so- data). The aurantiivertex advertisement call is a cially isolated from other territorial males. meandering, chattering trill of sharp chip notes. Prum (1994) hypothesized that solitary dis- This vocalization is a lengthened version of the play territoriality in Heterocercusis evolution- homologousaggressive chatter calls ofjluvivertex arily derived from the dispersedlek that is ple- but differs in function. Depending on which type siomorphic, or primitive, to the manakin family. of advertisement call is primitive to the genus, The solitary advertisement system has appar- the aggressivechattering call either evolved sec- ently evolved through the derived loss of male ondarily into a novel advertisement call in au- social aggregation. Prum (1994) proposed that runtiivertex, or a chattering advertisement call this breeding system be called a “solitary lek” was replaced by a derived, whistled advertise- since: (1) solitary leks, in this and possibly other ment call in a putative common ancestor offlu- instances, may be historically derived from tra- vivertex and linteatus. ditional, aggregatedclasses of lek territoriality The erect display posture of jlavivertex has and (2) solitary leks share the most important evolved in direct association with the advertise- feature with traditional, aggregatedleks-that is, ment call (Fig. 6A). The visual and acousticcom- males defend non-resource-based territories for ponents of this complex display are interrelated polygynous advertisement. and inseparable. In many regards,this display is similar to the loud vocal advertisements in the NESTING Screaming and Rufous Pihas (Lipaugus) or the The nest offlavivertex is typical in architecture, Capuchinbird (Perissocepkalustricolor; Cotin- composition, position, and size to other manakin gidae; Snow 1982). species(Snow 1962, Skutch 1969, Wetmore 1972, The tail-shiver display offlavivertex is unique Foster 1976, Dujardin 1987). Some manakin (Fig. 6B). No other known manakin display in- species build a flimsy cup that is transparent cludes a cocked-tailed, horizontal posture, a syn- enough for the contents to be visible from below copatedshivering movement, or a departurefrom (e.g., Pipra and A4unacus),similar to the possibly a perch with cklp note. The “wing-shiver” dis- incompletefluvivertex nest collected. In contrast, play of the Pipra aureola clade features a hori- the Long-tailed Manakin Ckiroxipkia linearis, zontal posture with the tail-cocked but the wings White-fronted Manakin Lepidotkrix serena, and are shivered rapidly rather than the body and tail Blue-crowned Manakin L. coronata build a more rocking rapidly back and forth (Snow 1963a, substantial cup that is heavily adorned with moss Schwartz and Snow 1978, Robbins 1983). It is or hanging dry leaves (Skutch 1969, Wetmore possible, however, that the horizontal compo- 1972, Foster 1976, Dujardin 1987). The use of nent of the flavivertex tail-shiver display is ho- spider webs and long plant fibers by H.Jlavivertex mologous with the horizontal posture shared by is very similar to Munacus manacus(Snow 1962). speciesof Pipra (Prum 1990a; character 4). If so, Nests of various manakins are also commonly this behavioral homolog has been elaborated in- placed over streamsbut thesestreams are usually dependently in various displays within the two smaller than the river where theflavivertex nest genera. TheJlavivertex tail-shiver display may be was found. Male jlavivertex were not observed similar to an undescribed display of Ckiroxipkia HETEROCERCUS FLAVIVERTEX BEHAVIOR 733 linearis (D. B. McDonald, pers. comm.), but any the wing is responsible for these mechanical such similarities are probably convergencesgiv- sounds (Table 1) . en the phylogenetic relationships of thesegenera Elsewherein manakins, only speciesin the ge- to other manakins which apparently lack such a nus Corapipo perform elaborate, above-the-can- display. opy flight displays with synchronized vocaliza- According to Sick (1993), linteatus also erects tions and mechanical wing sounds (Davis 1982, its white throat patch, asflavivertex does in the Prum 1986, 1990a). However, any similarities tail-shiver display. It is not known, however, between the displays of Heterocercusand Cor- whether linteatus or aurantiivertex perform the apipo are convergent and evolutionarily inde- tail-shiver display, or any behavioral homolog. pendent of one another. The only complete phy- In some other manakin species,unusual tail mor- logenetic hypothesis for the manakins indicates phologies are associatedwith specificdisplay el- that Heterocercusis closely related to the genus ements in which they are prominently featured. Pipra (sensustricto, excluding the serenagroup), The tail-shiver display ofjiuvivertex’ is a plausible whereasCorupipo is most closely related to Mus- explanation for the evolution of this unusual ius and Ilicura (Prum 1990a, 1992). These con- morphology in Heterocercus.The occurrence of spicuous displays probably serve a convergent identical tails in all three species of the genus advertisement function in the two genera. suggeststhat a homolog of the tail-shiver display DIET is also performed by linteatusand aurantiivertex. The jlavivertex above-the-canopy flight dis- A crude estimate of diet composition based on play is an impressive behavior that remains observations of foraging and seed regurgitation poorly known. A similar flight display, in which indicate that arthropods are a major component males Ay as high as 100 m above-the-canopy, of the diet of H. jiuvivertex’ (Table 2). These data has recently been observed in H. aurantiivertex are skewed by many factors. For example, fruit in Departemento Loreto, (J. Alvarez and resourcesare generally patchy and probably low- B. Whitney, pers. comm.). Apparently, this flight er in density within male territories where many display is a behavioral homolog shared by these of these observations were made. Also, males two species.The display probably evolved in the certainly foraged opportunistically within their common ancestor of the genusHeterocercus and territories on arthropod resourceswhich are typ- is probably performed by linteutus.Interestingly, ically more uniformly distributed than fruit. Ar- males of all Heterocercus species have signifi- thropods, however, are clearly an important food cantly longer wings than females, even though resource for jlavivertex. Although comparable they are not sexually dimorphic in other mea- data do not exist for other manakin species,ar- surements including tail shape (Table 1). The thropod foraging was observed more than an or- longer male wing length may have evolved as a der of magnitude less frequently than fruit for- consequence of selection for demanding flight aging in all previous observations of 15 other performance in aerial displays. A positive asso- speciesof manakins throughout the Neotropics ciation between wing length and flight displays (R. 0. Prum, pers. observ.). Interestingly, ob- is predicted by the general, positive correlation servation of foraging by male H. aurantiivertex between wing length and flight performance in indicate a similar, large component of arthro- birds (Greenewalt 1962). pods in the diet. Anotherjlavivertex flight display may be as- The striking differences in the diets of Heter- sociated with the production of a mechanical ocercusfrom those of other manakins may be sound. H. aurantiivertexdoesmake a mechanical related to the evolution of some of the other sound in association with a flight display (J. Al- differences in morphology, habitat, and social varez and B. Whitney, pers. comm.). These me- behavior. Heterocercus species have a longer, chanical sounds could be produced by either the narrower bill than do other manakins. This fea- wings or the tail. However, all manakin species ture is probably derived within the genus and that produce mechanical sounds have sexually may be an adaptation to insect foraging. Heter- dimorphic wings feathers (Prum, unpubl. data). ocercusspecies live in seasonally flooded forest The sexual dimorphism in wing length and the habitats, which are not preferred by other man- lack of sexual dimorphism in tail shape in Het- akin species.The shift in the diet composition erocercusare consistent with the hypothesis that may be related to the derived occupation of the 734 R. 0. PRUM ET AL. novel habitat within the family. Lastly, many seum of Natural History), G. Graves (Smithsonian In- proximate ecological models hypothesize a re- stitution), and C. Erard (Paris Museum ofNatural His- tory) kindly permitted R. 0. Prum to borrow or ex- lationship between lek size characteristics and amine specimensin their care. Behavioral illustrations resource distribution (Emlen and Oring 1977, were prepared by Carolyn Wooster. Illustration of the Bradbury 198 1, Bradbury and Gibson 1983, tail of H. flavivertex was prepared by Eric Wild. J. Foster 1983, Beehler and Foster 1988). The nov- Alvarez and B. Whitney kindly shared their observa- el diet in Heterocercusmay influence the distri- tions of H. aurantiivertex. M. S. Foster, D. B. Mc- Donald and M. B. Robbins contributed many sugges- bution ofreproductive resourcesin the genusand tions that greatly improved the manuscript. Data anal- may have resulted in the evolution or strategic ysis was supportedby NSF grant DEB-93 18273 to R. adoption of a solitary lek breeding system. Insect 0. Prum. foraging should lead to more uniform distribu- tion of reproductive resources and thereby to reduced home range size and home range over- LITERATURE CITED lap, which can result from clumped resources. BEEHLER,B. M., AND M. S. FOSTER. 1988. Hotshots, Both of these factors would favor reduced lek hotspots, and female preference in the organiza- tion of lek mating systems. Am. Nat. 131:203- size and increased male dispersion in either the 219. female choice (Bradbury 198 1) or hot-spot BRADBURY, J. W. 1981. The evolution of leks, p. 138- (Bradbury and Gibson 1983) models of lek de- 169. In R. D. Alexander and D. W. Tinkle [eds.], termination. Alternatively, the novel solitary lek Natural selectionand socialbehavior. Chiron Press, system may have evolved by sexual selection on New York. BRADBURY, J. W., AND R. M. GIBSON. 1983. Leks genetic variation for male social behavior itself and mate choice, p. 109-138. In P. Bateson [ed.], (Queller 1987, Prum 1994). Additional compar- Mate choice. Cambridge Univ. Press,Cambridge. ative researchon diet and social behavior ofHet- DAVIS, T. H. 1982. A flight-songdisplay of the White- erocercusand other manakins may help to fur- throated Manakin. Wilson Bull. 94594-595. DUJARDIN,J.-L. 1987. Decouverte du nid de la ponte ther test associationsbetween ecology, sexual se- du Manakin a front blanc (Pipra serena). LO-’ lection, and social behavior in the family. iseaux Rev. Fr. Omithol. 57:57-58. EMLEN,S. T., ANDL. W. OR~NG. 1977. Ecology, sex- ual selection,and the evolution ofmating- systems. From the late 19th century (Sclater 1888) until Science 197:.215-223. recently (Snow 1975, 1979), Heterocercus has FOSTER.M. S. 1976. Nestina bioloav of the Lona- tailed Manakin. Wilson Bill. 88:%0-420. - been viewed as an aberrant genus within the FOSTER,M. S. 1983. Disruption, dispersion, and manakin family, based on its larger size, atypical dominance in lek-breedina birds. Am. Nat. 122: plumage, and body and bill proportions. Most 53-72. recently, Prum (1990b, 1992) removed six ab- GcK~D~~N,M. L., AND M. LENTINO. 1992. Lista de las aves de1 Paraue National Cinaruco-Capana- errant genera from the manakin family, but paro, Estado Apure, Venezuela. 2nd ed. Soc.Con- strongly confirmed the placement of Heterocer- serv. Audubon Venez., Caracas. cus within the piprids as the closestrelatives of GREENEWALT,C. H. 1962. Dimensional relationships the type genus Pipra. These behavioral obser- for flying . Smithson. Misc. Collect. 144: vations further document that Heterocercus l-46: - HAFFER,J. 1986. Biogeographyof Neotropical birds, shares the derived lek social behavior, display p. 106-150. In T. C. Whitmore and G. T. Prance polygyny, and nest architecture that is typical of ieds.], Biogeography and quatemary history in other sexually dimorphic manakins, and con- tropical America. Clarendon Press, Oxford. firms its placement within the manakin clade. HILT, S. L., AND W. L. BROWN. 1986. A guide to the birds of Colombia. Princeton Univ. Press, ACKNOWLEDGMENTS Princeton, NJ. OLSEN, D. H., AND M. K. MCDOWELL. 1983. A com- The field observationswere supported by a Chapman parison of White-bearded Manakin (Manacus Fellowshin from the American Museum of Natural manacus) populations and lek systems in Suri- History, New York to R. 0. Prum. The field work was name and Trinidad. Auk 100:739-742. made possibleby the kind hospitality of Sr. C. Cidonio, Paw, R. 0. 1986. The displaysofthe White-throat- owner of Campamento Camani (formerly Junglaven ed Manakin Corapipogutturalis in . Ibis Lodge), the staff at Campamento Camani, and by the 128: 91-102. serendipitousfriendship of Sm. Ana Villanueva-Brandt, PRuM, R. 0. 1990a. Phylogeneticanalysis ofthe evo- Caracas,Venezuela. Field work was also aided by Ka- lution of display behavior in the Neotropical man- thy Phelps, Ramon Aveledo, and Miguel Lentino of akins (Aves: Pipridae). Ethology 84:202-23 1. the Collecibn Phelps, Caracas.R. Aveledo, M. Lentino PRY, R. 0. 1990b. A test of the monophyly of the (Colleci6n Phelps), G. Barrowclough(American Mu- manakins (Pipridae) and (Cotingidae) HETEROCERCUS FLA VIVERTEX BEHAVIOR 735

basedon morphology.Oct. Pap. Mus. Zool., Univ. SKUTCH,A. F. 1969. Life histories of Central Amer- Mich. 722: 144. ican birds. III. Pacific Coast Avifauna 35. PRUM,R. 0. 1992. Syringealmorphology, phylogeny, SNOW,D. W. 1962. A field study of the Black-and- and evolution of the Neotropical manakins (Aves: white Manakin, Manacus manacus,in Trinidad, Pipridae). Am. Mus. Nat. Hist. Novitates 3043: W. I. Zoologica 47:65-104. l-65. SNOW,D. W. 1963a. The display of the Orange-head- PRUM, R. 0. 1994. Phylogeneticanalysis of the evo- ed Manakin. Condor 65:44-48. lution of alternative social behavior in the man- SNOW,D. W. 1963b. The evolution of manakin dis- akins (Aves: Pipridae). Evolution 48: 1657-1675. plays. Acta Int. Omithol. Congr. 13:553-56 1. QUELLER,D. C. 1987. The evolution of leks through SNOW, D. W. 1975. The classificationof the mana- female choice. Anim. Behav. 35:1424-1432. - kins. Bull. Brit. Omithol. Club 95:20-27. RIDGELY.R. S.. AND G. TUDOR. 1994. The birds of SNOW,D. W. 1979. Tityrinae, Pipridae, Cotingidae, South America. Vol. 2. Univ. of Texas Press,Aus- p. 229-308. In M. A. Traylor [ed.], Check-list of tin. birds of the world. Vol. VIII. Mus. Comn. Zool.. ROBBINS,M. B. 1983. The display repertoire of the Cambridge, MA. Band-tailed Manakin (Pipra fasciicauda). Wilson SNOW,D. W. 1982. The cotingas.Cornell Univ. Press, Bull. 95:321-342. Ithaca, NY. SCHWARTZ,P., AND D. W. SNOW. 1978. Display and WETMORE, A. 1972. The birds of the Republic of related behavior of the Wire-tailed Manakin. Liv- Panama. Vol. 3. Smithson. Misc. Collections.Vol. ing Bird 17:5l-78. 150, Part 3. SCLATER,P. L. 1888. Catalogue of the birds of the WILLARD,D. F., M. S. FOSTER,G. F. BARROWCLOUGH, British Museum. British Museum, London. R. W. DICKERMAN.P. F. CANNELL.S. L. COATS. SICK, H. 1959. Die Balz der Schmuckvogel(Pipri- J. L. CRACRAFI,A&D J. P. O’NEILL. 1991. The dae). J. Omithol. 100:269-302. birds of Cerro de la Neblina, Territorio Federal SICK.H. 1967. Courtshin behavior in manakins (Pi- Amazonas, Venezuela. Fieldiana Zool. 65: l-80. pridae). A review. Li&ng Bird 615-22. SICK, H. 1993. Birds in Brazil. Princeton Univ. Press, Princeton, NJ.