July1983] ShortCommunications 739

Despite the above arguments, I believe the Red FFRENCH,R. 1973. A guide to the of Trinidad Siskin's establishment in Puerto Rico most probably and Tobago. Wynnewood, Pennsylvania, Liv- coincided with the period of its heaviest importation. ingston Publ. Co. This probably would have been in the 1930's when GUNDLACH,J. 1878. Apuntes para la fauna Puerto- the specieswas still relatively common in its native Riquena. Aves. Anales Sociedad Esp. Hist. Nat., lands and there was an extremely high demand for Madrid 2: 141-234, 343-422. it among Canary breeders.Discussions with long-term KING, W. B. 1981. Endangered birds of the world: residents in the region where the Red Siskin occurs The ICBP red data book. Washington D.C., could shed additional light on this matter. Smithsonian Institution Press. Regardlessof the length of time that the Red Sis- LEOPOLD, N. F. 1963. Checklist of birds of Puerto kin has inhabited Puerto Rico, its continued survival Rico and the Virgin Islands. Univ. Puerto Rico will depend on protective safeguards. While the Agric. Exp. Sta. Bull. 168. may have been a will-o'-the-wisp for orni- RAFFAELE,H.A. 1983. A guide to the birds of Puerto thologists, we cannot depend on it remaining so for Rico and the Virgin Islands. San Juan, Puerto pet traders. Rico, Fondo Educativo Interamericano. DE SCHAUENSEE,R.M. 1966. The species of birds of LITERATURE CITED and their distribution. Wynne- BIAGGI,V. 1970. Las Aves de Puerto Rico. San Juan, wood, Pennsylvania, Livingston Publ. Co. Puerto Rico, Editorial Universitaria. 1970. A guide to the birds of South Amer- CLAPP,R. B. 1975. Birds imported into the United ica. Wynnewood, Pennsylvania,Livingston Publ. Co. States in 1972. U.S. Fish Wildl. Serv., Spec. Sci. Rept. Wildl. 193. SUNDEVALL,C. J. 1869. Foglarne pa on Portorico, --, & R. C. BANKS. 1973a. Birds imported into efter Hr Hjalmarsons insamlingar framstallda. the United States in 1970. U.S. Fish Wildl. Serv., Ofversigt af Kongl. Vetenskaps-Akademiens Spec. Sci. Rept. Wildl. 164. Forhandlingar. TjugondespetteArgangen 593- 603. --, & ---. 1973b. Birds imported into the WETMORE,A. 1916. Birds of Porto Rico. U.S. Dept. United Statesin 1971. U.S. Fish Wildl. Serv., Spec. Agric. Bull. 326. Sci. Rept. Wildl. 170. DODWELL,G. T. 1976. Encyclopedia of canaries. Received19 August1982, accepted4 January1983. Neptune City, New Jersey,T. F. H. Publications.

A Comparison of White-bearded ( manacus)Populations and Lek Systems in Suriname and Trinidad

DEANNA H. OLSON • AND MICHAEL K. McDOWELL 2 •Departmentof Zoologyand 2Schoolof Oceanography,Oregon State University, Corvallis,Oregon 97331-2914 USA

The White-bearded Manakin (Manacus manacus)is of Suriname was compared to the Arima Valley pop- a neotropical, primarily frugivorous, lek-mating pas- ulation with respect to population density, morphol- serine. The ecology and behavior of M. m. trinitatis ogy, reproductive behavior, male mating success,and in the Arima Valley, Trinidad has been described lek characteristics. At least two resident males, with (Snow 1962; Lill 1974a, b). Snow (1962) found the display courts near each other, constitute a lek. A Trinidad population density to be unusually high display court is a circular area of forest floor, between with respectto mainland populationsand suggested saplings, cleared of leaf-litter by a male. A resident that this was due to either a high proportion of sec- male successfullydefends his court and a small area ondary forest, causedby limited clearing and lum- surrounding his court against intruding conspecific bering, or reduced interspecific competition on the males. [See Snow (1962) and Lill (1974a) for a more island. Secondaryforest may benefit survival and re- complete description of the display.] The production, becausemany of the trees abundant in lek characteristicsinvestigated in our study include secondaryforest are prime food sources(Snow 1962) interlek distance, lek area.. intercourt distance, and and there is an abundance of saplings on which the the number of resident males per lek. males depend for their display (Snow 1962, Lill Male dispersion in Suriname and Trinidad was 1974a). compared and related to the lek characteristicsin- There has been little research on M. manacus in vestigated. Male dispersion patterns have been de- mainland South America. In this study, a population scribed as a continuous gradient in the degree of male of M. m. manacusin virgin rain forest in the interior clustering. This gradient ranges from uniform fields 740 ShortCommunications [Auk, Vol. 100 of male territoriesthrough partial clusteringof males Four M. manacusleks were located adjacent to with smaller territories to highly clustered malesand granite plates and domes in dense vegetation char- large intercluster distances,the classicallek disper- acteristic of secondary growth. Males were not ob- sion (Bradbury 1981). We examined male clustering served to display solitarily or in areasdisturbed by and the possible influences of habitat resourcesand treefalls,but the displaysof severalmales were heard population density on male dispersion in M. mana- at two other possible lek locations in dense vegeta- tion near the edge of granite outcrops; these birds The study site was in the Raleigh Falls/Voltzberg were not observed.M. manacusdispersion may be Nature Reserve(approximately 4.5øN, 56.0øW),Suri- limited to areas with vegetation characteristicof sec- name, adjacentto the CoppenameRiver. The primary ondary forest. Secondary growth may supply ade- rain forest was interspersedwith unforested granite quate male display sites,food resources(Snow 1962), outcrops in the form of large, flat plates and domes and nesting sites for . rising to 100 m above the forest canopy Forestchar- The comparisonsof population density, interlek acterized by thick undergrowth and a low canopy distance, lek area, number of males per lek, and in- surrounded the outcrops. This apparent secondary tercourt distance between the two sites are shown in growth may be attributed to rocl•y soil at the edge of Table 1. Snow (1962) determined the population den- the granite outcrops. These areas comprised about sity in the Arima Valley to be 2.7 adults/ha. Lill one-third of our study site, which was located at the (1974b), 10 yr later, estimated that the density had baseof the Voltzberg granite dome within an area of declined by 20% and attributed the decline to the approximately 200 ha. reduction of secondary forest. The Voltzberg adult Data were collected at four leks from December population density was much less than either of these 1980 to May 1981. The averagenearest neighbor, in- Trinidad estimates,confirming Snow's (1962) obser- terlek distancewas determined with adjustmentsfor vation of a lower mainland population density. The topographic relief at the Voltzberg study site and average interlek distancebetween the four Voltzberg comparedwith the average two-dimensional interlek leks was much greater than the average of 11 Arima distance,approximated from Lill (1974b), in the Ari- Valley leks. The four Voltzberg leks consistedof 2, ma Valley. The number of adult resident males per 3, 4, and 5 resident adult males. Lill (1974a) described lek and the location of adult male display courtswere the number of resident males as ranging from 6 to determined by observation during the breeding sea- 50 in a typical Arima Valley lek. Snow (1962) found son. The extent of the breeding seasonin Suriname 205 display courts at seven leks, an average of 29 has not been described. In Trinidad, breeding gen- males per lek. The largest lek in Trinidad had 70 erally occursfrom March to August (Snow 1962, Lill display courts within an area of approximately 18.3 1974a). by 9.2 m (Snow 1962) and was smaller than the av- Snow (1962) estimated the population density of erage polygonal MLA at the Voltzberg site (which adult M. manacusby counting the number of active does not include the two-male lek). MLA was an ex- display courts at the leks to determine the number treme underestimateof the area surrounding the dis- of resident adult males and then calculating the pro- play courts in which resident males actively dis- portion of trapped nonresident males and females to played. The MLA of the three-male lek was 110 m2, trapped resident males. Population density was de- whereas the area in which all territorial and mating termined in the same manner at the Voltzberg study displays occurred was approximately 548 m2. The site from trapping data at leks. nearest-neighbor distance between resident male Males and femalesat two leks were uniquely color display courtsat the four Voltzberg leks ranged from banded, weighed, and measured for right-wing 6.2 to 21.5 m, with an average of 13.6 m, whereas the length. We visited a lek consistingof three color- Arima Valley courtswere, on average, 0.9-4.6 m apart banded resident males daily and observed it from a (Snow 1962). blind on days of high activity. The behaviorsof adult Morphological variation was found between Suri- males, females, and juvenile males at the lek were nam and Trinidad. Adult male wing lengths and recorded during 158 h of observation,between 0600 weights at the Voltzberg site were significantly dif- and 1800.Juvenile maleswere identified by their use ferent (P < 0.001) from those of the Arima Valley of male display patterns.Resident-male territory areas, population (Table 1). determined by locationsof agonisticencounters sur- We observed breeding activity from the onset of rounding individual courts,and lek area in which all our study in mid-December until March. Breeding display activity occurred were determined for this activity was most intense in December and January, three-male lek. The polygonal area enclosedby res- at the start of the short dry season,during which we ident-male display-court centers was used as the logged 114 h of observationtime between 0700 and minimum-lek-area (MLA) estimate for each lek at the 1700. Most of the female visits and copulations were Voltzberg site. A Zeiss MOP-3 System for Quantita- recorded between 1200 and 1600, during 75 h of ob- tive Digital Image Analysis was used to determine servation. area. The behavior patternsof birds at the Voltzberg site July 1983] ShortCommunications 741

TABLE1. Lek characteristics,population density estimates,and adult male wing lengths and weights from the Trinidad and Suriname study sites.

Arima Valley Voltzberg Average interlek distance (m) 371a 706 Number of resident males/lek 6-50 b 2-5 Lek area (m2) 167c 266 (MLA) Intercourt distance (m) 0.9-4.6 c 6.2-21.5 Population density (adults/ha) 2.19a 0.17 Adult male weight (g) Mean 18.20 c 17.05 a SD 1.24 0.46 n 185 9 Adult male wing lengths (mm) Mean 52.47 c 50.55 a SD 1.07 1.67 n 55 9

• Approximated from Lill (1974b). b Lill (1974a), 'Snow (1962). dp < 0.001;Student's t-test (Snedecorand Cochran 1980:96). during the breeding seasonwere similar to those de- and C. Male A maintained the largest territory, 202 scribed by Snow (1962) and Lill (1974a, b). Stereo- m 2, B's territory was 189 m 2, and C's was 36 m 2. Male typed male and female courtship displays, juvenile territory defense and territory size seemto be related displays,male-male aggression,and male-male dom- to male mating successat this Voltzberg lek. inant-subordinate relations were observed at the Male clustering at leks may be assessedby exam- Voltzberg leks. Lill (1974a) found a nonrandom mat- ining the number of resident males per lek, inter- ing distribution among males at the leks in Trinidad: court distance,and lek area. The largest number of a small proportion of resident males received the ma- resident males in a Voltzberg lek did not reach the jority of copulations. Nonrandom male mating suc- smallestnumber observedin the Arima Valley. In- cess has been recorded in other lek species[Sage tercourt distance,an indicator of male territory size, Grouse (Centrocercusurophasianus), Wiley 1973; ham- and lek area were greater in Suriname.The degree merheaded bat (Hypsignathusmonstrosus), Bradbury in male clustering at leks was less and interlek dis- 1977;Guianan Cock-of-the-Rock(RupicoIa rupicoIa), P. tances were larger in Suriname, but a classicallek W. Trail, pers. comm.]. Behavioral observations of dispersion was maintained. three males (A, B, C) made at one Voltzberg lek dur- Although a higher population density coincided ing the breeding seasonalso indicated skewed male with greater clustering of males at leks at the Arima mating success.Male A received 29 (88%) female vis- Valley, this is not an expected result. An increase in its and 11 (92%) copulations,male B received 4 (12%) populationdensity need not changemale clustering, female visits and 1 (8%) copulation, and male C was asdetermined by lek areas,numbers of malesper lek, unvisited. Lill (1974a) found a relationship between or male territory sizes. For example, an increasein male mating successand territory defense in the Ari- male density might result in an increasein the num- ma Valley. At the Voltzberg site, male A was most ber of nonresident, "floater" males or in the forma- successfulin defending his territory, which sur- tion of new leks in the area, neither of which need rounded his display court, contiguous with the ter- alter resident male clustering at existing leks. Be- ritories of males B and C. Male A always displaced causewe found lessmale clusteringand a lower pop- intruding males during encounters within his terri- ulation density in Suriname, however, a relation be- tory. A displacement was recorded when a male left tween these factors is suggested.Bradbury (1981) his perch upon the approachof another. Males A and proposed that male clustering arises from a pref- B equally displaced each other during encountersat erencefor clustersby females.If femalesare attracted the boundary between their territories. Male B did to larger clustersthen male clustering may increase not always displace one of the banded nonresident with population density. males from his territory. Male A always displaced Patchily distributedresources may limit lek size, male C during boundary encountersand often dis- the number of leks per area, or population density, placed male C within C's territory. Male C main- any of which could subsequentlyaffect male cluster- tained resident male statusbecause he displacednon- ing at leks. Secondaryforest may be a resourcelim- resident intruding males from his territory and iting M. manacuspopulations. Although the distri- remained at the lek throughout the breeding season. bution, abundance,and compositionof secondary There was no observed interaction between males B forest at the Arima Valley and Voltzberg study sites 742 ShortCommunications [Auk,Vol. 100 have not been determined and, consequently, a di- their support in all aspectsof our fieldwork. We are rect comparison between the sites cannot be made, grateful to STINASU, the Suriname Nature Conser- the Voltzberg site probably containsa much smaller vation Foundation, for their assistance and the use of proportion of secondarygrowth. The abundanceof their research facilities at the Raleigh Falls/Voltz- secondarygrowth may limit manakin display sites, berg Nature Reserve.We thank K. Fristrup, P. W. food resources,or nestingsites, which, in turn, might Trail, S. I. Rothstein,and especiallyA. R. Blaustein influence male dispersionpatterns. These and other for criticallyreading this manuscript.We are indebt- factors that may affect M. rnanacusdispersion and ed to R. P. Hard for the useof the ZeissMOP-3 Sys- population density, such as interspecificcompetition tem for Quantitativedigital ImageAnalysis and to C. and predation, should be investigated. Flackand S. Sargentfor typing the manuscript. Correlationsamong indicatorsof male clustering and populationdensity or habitatresources have been LITERATURE CITED reported in few other lek species.Few studiespro- vide sufficientinformation to examinethese specific BRADBURY,J. W. 1977. Lek mating behavior in the relationships.The number of malesper lek increased hammerheadedbat. Z. Tierpsychol.45: 225-255. with an increasein population density in the Lesser 1981. The evolution of leks. Pp. 138-169 in Prairie Chicken (Tympanuchuspallidicinctus, Hoffman Natural selection and social behavior (R. D. 1963),and the SageGrouse (Patterson 1952), but data Alexander and D. W. Tinkle, Eds.). New York, from thesestudies concerning male territory sizesor Chiron Press Inc. lek areasare not sufficientto assessmale clustering HOFFMAN, D.M. 1963. The Lesser Prairie Chicken fully. Bradbury(1981) found no relation between the in Colorado. J. Wildl. Mgmt. 27: 726-732. degree of male clusteringand density when inter- LEUTHOLD, W. 1966. Variations in territorial behav- specific comparisonswere made of several grouse ior of Ugartdakob Adenotakob thomasi (Neumann species.A relationship among population density, 1896). Behaviour 27: 214-257. male clustering, and habitat resources, has been L1LL,A. 1974a. Sexualbehavior of the lek-forming shownfor two lekking mammals.Clustering of male White-Bearded Manakin (Manacus manacustrini- territories increasedwith increasingpopulation den- tatis).Z. Tierpsychol.36: 1-36. sity and food-resourceavailability in the topi (Dam- 1974b. Social organization and space utili- aliscuskorrigum, Monfort-Braham 1975). In the Ugan- zation in the lek-forming White-Bearded Man- da kob (Adenotakob thomasi),single large territories akin, M. manacustrinitatis. Z. Tierpsychol.36: 513- were prevalent in marginal habitat, while small clus- 530. tered territories were more common in optimal hab- MONFORT-BRAHAM, N. 1975. Variations dans la itat, and seemedto develop in large populationswith structure sociale du Topi Damaliscuskorrigum high populationdensities (Leuthold 1966).In the topi Ogilby, au Parc National de l'Akagera, Rwanda. and Uganda kob, the degreeof male clusteringvar- Z. Tierpsychol.39: 332-364. ied along the whole gradient of dispersionpatterns, PATTERSON,R. L. 1952. The Sage Grouse in Wyo- from a uniform field of male territories to a classical ming. Denver, Colorado, Sage Books. lek. It is interestingto note that, although M. manacus SNEDECOR,G. W., & W. G. COCHRAN. 1980. Statistical male clustering was different in Suriname, the clas- methods, 7th Ed. Ames, Iowa, Iowa State Univ. sical lek dispersion was maintained. A more detailed Press. examinationof lek characteristicsin other lek species SNOW,D. W. 1962. A field study of the Black and may reveal a similar pattern of intraspecificvariation White Manakin, Manacus manacus, in Trinidad. in male clustering coincident with changesin pop- Zoologica 47: 65-104. ulation density and habitat resources. Such infor- W1LEY,R. H. 1973. Territoriality and non-random mation would further the understandingof the evo- mating in Sage Grouse, Centrocercusurophasianus. lution of male dispersionpatterns and the adaptive Anim. Behav. Monogr. 6: 85-169. significanceof leks. We wish to thank K. Fristrup and P. W. Trail for Received19 August1982, accepted28 January1983.