The Condor 863641 0 The CooperOrnithological Society 1984

ARBOREAL DEAD-LEAF-SEARCHING OF THE NEOTROPICS

J. V. REMSEN, JR.

AND T. A. PARKER, III

ABSTRACT. -At least 11 speciesof birds in northern and southern are dead-leaf-searching“ specialists”: more than 7 5% of our foraging observations of these specieshave been of individuals searching for insects in dead, curled leaves suspendedabove ground in the vegetation. All known specialistsof this kind belong to the families Furnariidae and Formicariidae. An additional six species,here called “regular users,” exhibit dead-leaf-searchingbehavior in 25% to 75% of our foraging records. The number of specialists and regular users decreaseswith rising elevation in the Andes. Specialistsdisappear from the gra- dient between 2,000 m and 2,575 m, but regular usersoccur as high as 3,300 m, near timberline. As many as eight speciesof dead-leaf-searchingspecialists coexist in western Amazonia.

The degreeto which differencesin species searching dead leaves ranging in size (un- richness between temperate and tropical for- curled) from 6 x 4 cm to 45 x 30 cm. ests can be explained by increased variety of These dead leaves harbor more insectsthan resourceshas yet to be adequately quantified do green leaves (data from Panama; Gradwohl (Karr 1975, Terborgh 1980a). Certainly the and Greenberg 1982; Gradwohl, unpubl. data), year-round availability of fruit and nectar (Karr and thus potentially provide a high energyyield 1971, Morton 1973) and very large insects for those birds that search them. Such birds (Schoener 197 1) contributes markedly to the must expend more time than elsewhere in high richness of bird species in the tropics. searching for and handling prey because ex- Other resourcesand habitats absent from, or tracting insects from curled dead leaves often with restricted availability in, the Temperate requires probing motions and acrobatics not Zone that also contribute to tropical species involved in standard gleaning from green richnessare army ants (Willis and Oniki 1978), leaves. Dead leaves are also much less abun- oxbow lake edgeand permanently flooded for- dant, and more patchily distributed than, green est (Remsen and Parker, in press), bamboo leaves, increasing the average travelling time thickets (Parker 1982, Parker and Remsen, un- between foraging sites for a specialist. publ. data), and epiphytic plants (Terborgh 1977, Remsen and Parker, unpubl. data). Our STUDY AREAS AND METHODS purpose here is to point out the contribution In the austral winters of 1979, 1980, and to neotropical bird speciesrichness of the year- 198 1, we gathered foraging data on the insec- round availability of curled dead leaves sus- tivorous birds at six points (600 m, 1,350 m, pended in vegetation above the floor. 1,650 m, 2,575 m, 3,050 m, and 3,300 m) on These dead leaves are searchedby a variety of an altitudinal transect in the humid Andes of bird species, some of which are highly spe- Depto. La Paz, Bolivia. Each time a bird was cialized on this foraging tactic (Terborgh seen to touch a substrate with its bill while 1980a). foraging was considered as a single foraging Suspendeddead leaves may occur from the record; no more than three consecutive for- top of the canopy to near the ground, especially agingrecords were taken for an individual bird. where vegetation is dense enough to prevent These data allowed us to quantify the degree falling leaves from reaching the forest floor. of specialization on dead-leaf searching by a Dead leaves and the birds that searchthem are variety of speciesin a rich avifauna. We sup- most plentiful in vine tanglesand densethick- plemented these data with Parker’s foraging ets (Gradwohl and Greenberg 1982), where information from lowland Peru at the mouth high foliage density produces greater accu- of the Rio La Torre on the Rio Tambopata, mulations of dead leaves than elsewhere in Depto. Madre de Dios, 250 m, Manu National more open vegetation. Birds search leaves of Park, Depto. Madre de Dios, 250 m, and the all sizes as long as these are curled to form lower Rio Napo area, Depto. Loreto, 125 m., refugia for insects; we have recorded birds during the austral winter of 1983. Voucher

1361 DEAD-LEAF-SEARCHING BIRDS 37 specimensfor all speciesare deposited at the phological correlates among dead-leaf search- Museum of Zoology, Louisiana State Univer- ers in comparison to live-leaf searchers.The sity (LSUMZ). high frequency of hanging, clinging and prob- We grouped dead-leaf searchersinto three ing maneuvers required to reach dead leaves categories based on degree of specialization: and extract prey from them suggeststhat subtle (A) Specialists:species seen searching dead morphological differences relative to live-leaf leaves in more than 75% of our observations; foragersmay exist. The only stout-billed dead- we predict that Specialistsare obligate dead- leaf searchers,the Eubuccobarbets, often rip leaf searchers.(B) RegularUsers: species seen curled leaf clustersopen, rather than insert the searching dead leaves in 25-75% of our ob- bill into the leaves as do the more typical, servations;for speciessuch as the Eubuccobar- slender-billed species.Thus the barbets, who bets that feed on both fruit and insects, over are Regular Users rather than Specialists,often 25% of our non-frugivorous observationswere ruin the dead leaf as a suitable refugium for of individuals searchingdead leaves. (C) Oc- insects,whereas the obligate Specialistsoften casionalUsers: species for which lessthan 25% leave the leaf intact and suitable for new po- of our observationsare of dead-leaf searching. tential prey. Most dead-leaf searchersare dull- Although our samplesfor many of the species colored, primarily shadesof brown or buff, but are small, especially by standards for Tem- their coloration does not seem to differ in any perate Zone studies, we are reasonably confi- consistent way from that of non-dead-leaf- dent of our assignmentsto categories;before searchingconfamilials. The Eubuccobarbets, formal quantification beginning in 1979, we however, are among the gaudiest of tropical recorded the foraging behavior of most species birds. discussedbelow from 1974 to 1978 in con- Syntopic dead-leaf Specialistsdiffer in body junction with other studiesin Peru and Boliv- size, size of leaves examined, and foraging ia. height. The body weights of Myrmotherula antwrens (10 g) and small furnariids (15 g for RESULTS AND DISCUSSION Cranioleucaand )are approxi- Eleven speciesof Specialistsand six speciesof mately one-third to one-quarter that of the Regular Users were found at our study sitesin larger dead-leaf-searchingfurnariids (20-40 g northern Bolivia and southern Peru (Table 1). for Philydor,Automolus, Anabacerthia, Thri- Species from these sites that we found padectes),a difference that presumably pro- to be Occasional Users were: Piaya cayana ducesdifferences in prey sizestaken. The ant- (Cuculidae); Phaethornistuber (Trochilidae); wrens also search significantly smaller leaves aurovirens(Capitonidae); Sittasomus (Mann-Whitney U-test, P < .O1) than the large griseicapillus,Dendrocolaptes certhia (Den- furnariids; but we do not know whether small- drocolaptidae); Synallaxisazarae, Margaror- er leaves harbor smaller insects. At Serrania nis squamiger,Ancistrops strigilatus, Pseudo- Bellavista, where four furnariids of approxi- colaptesboissonneautii, Philydor pyrrhodes mately the samebody size coexisted,each Spe- (Furnariidae); Cymbilaimussanctaemariae, cialist fed at different but overlapping heights Thamnophilusschistaceus, T. aroyae,Pygip- above ground (Table 1). All differed signifi- tila stellaris, Dysithamnus mentalis, Myr- cantly from one another (P < .OOl, Mann- motherulalongicauda, M. iheringi,M. axil- Whitney U-test) except for Anabacerthiastria- laris, Terenura humeralis, Hypocnemis ticollisvs. Philydorerythrocercus (P < . 10). cantator, Myrmeciza hemimelaena(Formi- Examination of the elevational gradient in cariidae); Atlapetesrujinucha (Emberizinae); number of speciesof Specialistsand Regular Hemispinguscalophrys, Thlypopsis rujceps, Users at eight localities in northern Bolivia and Habia rubica, Tachyphonusluctuosus, Buth- extreme southern Peru (Fig. 1) showsthat the raupismontana, Tangara schrankii, T. chry- number of Specialistsranged from 2.5 to 4% sotis(Thraupinae); and Basileuterusluteovi- of the total resident avifauna up to 2,000 m; ridis(Parulidae). Specialistsdisappeared from the gradient be- All dead-leaf Specialistswere furnariids or tween 2,000 m and 2,575 m. Regular Users formicariids, the two largest families of pre- comprised 1 to 2% of the total resident avi- dominately foliage-gleaningbirds in the Neo- fauna at all elevations except 2,575 m, where tropics; sample sizes, however, were inade- no Regular Users were found. quate to test statistically whether or not these The number of speciesin both foraging cat- families contributed disproportionately, rela- egoriesdecreased in a more or lesslinear man- tive to other insectivorousfamilies, to the dead- ner with increasing elevation, paralleling the leaf Specialist category. Regular Users were general trend toward decreasingspecies rich- taxonomically more heterogeneous(Table 1). nesswith increasingelevation on the samegra- We could not find any conspicuous mor- dient (Remsen, unpubl. data). The slope for 38 J. V. REMSEN, JR., AND T. A. PARKER, III

TABLE 1. Dead-leaf Specialistsand Regular Users.

Percent total foraging Mean foraging height maneuvers searchmg dead in meters leaves (II. total observatmns) Localities’ kmae) SPECIALISTS Fumariidae SpeckledSpinetail Cranioleucagutturata 84.6 (26) 5, 6, 7, 9 11.3 (8-15) Plain Softtail ThripophagafuscicepY 100.0 (12) (6) 7 8.4 (6-l 1) Montane Foliage-gleaner Anabacerthiastriaticolhs 92.3 (39) 3,4 9.9 (1.5-26) Rufous-tailed Foliage-gleaner Philydor ruficaudatus 76.9 (39) 5, (6) 7 12.8 (9-21) Rufous-rumped Foliage-gleaner Philydor erythrocerc& 90.2 (82) 3,4, 6, 7, 9 12.0 (6-22) Striped Treehunter Thripadectesholostictus 100.0 (9) 3, 4 0.9 (0.2-l) Buff-throated Foliage-gleaner Automolusochrolaemus4 95.0 (20) 334, 5, 6,7 3.2 (1.5-8) Olive-backed Foliage-gleaner Automolusinfuscatus 86.1(63) 6, 739 4.7 (0.3-12) Chestnut-crownedFoliage-gleaner Automolusrufipileatus 100.0 (16) 5,6, 7, 9 4.3 (1.5-9) Formicariidae White-eyed Antwren Myrmotherula leucophthalma 95.3 (43) 6, 7 3.6 (l-10) Ornate Antwren Myrmotherula ornata 98.9 (9 1) 5, 6, 7, 8 5.3 (0.7-l 8) REGULAR USERS Capitonidae Eubuccoversicolor 100.0 (7)5 3,4 10.0 (l-20) Lemon-throated Barbet Eubuccorichardsoni 86.4 (22)5 5,6, 7,9 14.2 (12-24) Scarlet-hoodedBarbet Eubuccotucinkae 100.0 (12)5 (6) 7 6.6 (2-l 1) Dencrocolaptidae Buff-throated Woodcreeper Xiphorhynchusguttatus 34.4 (32) 5,6, 7,9 10.8 (1-21) Fumariidae Light-crowned Spinetail Cranioleucaalbiceps 25.0 (36) 1, 2 2.5 (1.5-3) Troglodytidae Moustached Wren Thryothorusgenibarbis 3. 4. 5. 6 5.3 (l-12)

S S&plesize does not include observations of frugivory. the Specialist’s regression (Fig. 1) was nearly the other Automolusspecies (“terra firme” vs. identical to that of all bird speciescombined river edge forest), as were the two antwrens at (-.0030 vs. -.003 1). The slope for Regular 250 m (M. leucophthalmain forest and M. Users was much lower, but with so few Regular ornatain bamboo thickets). At the 600-m site, User species,the difference means little. where leucophthalmawas absent, ornata oc- The number of speciesin both foraging cat- curred in all types of forest as well as in bam- egories was slightly inflated at the 250-m and boo thickets; this seemsto be an excellent ex- 600-m localities by greater habitat heteroge- ample of a “niche shift” and “ecological neity. At both localities, Automolusrujipilea- release.” At the 250 m site, Eubuccotucinkae tus was separated by habitat differences from was a scarceinhabitant of forest at oxbow lake DEAD-LEAF-SEARCHING BIRDS 39

10

9

8 -0 specialists

.----- 0 regular users

2000 ELEVATION (meters) FIGURE 1. Number of speciesof dead-leaf-searchingSpecialists and Regular Users at various elevations in humid forest in northern Bolivia and southernPeru. The solid line (y = -.00300x + 8.0) representsthat producedby a linear regressionof elevation against number of speciesof Specialists(r = -.96, P < .Ol) between 250 m and the lowest elevation, 2,575 m, at which Specialistsdisappear. The dotted line (y = -.OOl 1 lx + 4.0) representsthat produced by a linear regressionof elevation againstnumber of speciesof Regular Users for all eight localities (r = -.86, P < .Ol). The data for the 2,000-m site come from an LSUMZ collecting locality a few kilometers from the Bolivian border in extreme southeasternPeru (Abra de Maruncunca, 10 km SW San Juande1 Oro, Depto. Puno; courtesyT. S. Schulenberg); the two Specialiststhere are Anabacerthiastriaticollis and Thripadectesholostictus and the Regular User is Eubucco versicolor.The number of resident forest speciesat each site is: 48 at 3,300 m, 72 at 3,050 m, 79 at 2,575 m, 99 at 1,650 m, 116 at 1,350 m, 138 at 600 m, and approximately 300 at 250 m (the latter is the Rio Tambopata site; see footnotes of Table 1 for precise localities).

edges and E. richardsoni was found in other Furthermore, as canopy height decreaseswith types of forest. We found no such habitat rising elevation, opportunities for vertical division from 1,350 m to 3,330 m; at these stratification by the dead-leaf searchers de- montane localities the river-created habitats crease.In the low-stature at the 3,050- responsiblefor a high degree of habitat parti- m and 3,300-m sites,no potentially competing tioning in the lowlands (Remsen and Parker, bird speciessegregate by differencesin foraging in press) were absent. heights, other than by the broad categories, Why do dead-leaf-searchingspecies become undergrowthvs. trees(Remsen, unpubl.). This fewer at higher elevations? Although we lack contrastswith the vertical stratification of four direct data concerning productivity, our qual- speciesof dead-leaf searchersat 1,650 m (see itative impressions agree with those of Ter- above) and numerous examples among live- borgh (1977) for Andean elevational transects: leaf searchersat low elevations (Pearson 197 1, overall productivity declines markedly with 1977; Terborgh 1980b). increasingelevation, aswould be expectedwith Does insectdensity in dead-leaf clusters(and/ the decline in air temperatures. Janzen (1973) or density of dead leaves) above 2,000 m ele- has documented an overall decline in numbers vation decline below some thresholdvalue that and speciesof insects (above intermediate el- prevents specialization on dead-leaf search- evations) with increasing elevation in Costa ing? Gradwohl and Greenberg (1982, unpubl. Rica. Thus, available insect biomass, presum- data) have demonstrated how amenable the ably including that harbored by dead leaves, dead-leaf system is to quantification of the should decrease with increasing elevation. variables needed to answer such a question. 40 J. V. REMSEN, JR., AND T. A. PARKER, III

With most critical parameters so readily mea- 1. Pygiptila stellaris may also be a Regular surable (dead-leaf density, distance between User, but our sample size of observations is leaves, prey density) and with a spectrum of insufficient at present. degreesof specialization by the predators, this Skutch(1954,1960,1967,1969,1972,1982) systemwould be nearly ideal for testingmodels mentioned dead-leaf searchingfor the follow- of optimal foraging (e.g., MacArthur and Pian- ing Central American species,most of which, ka 1966, Pulliam 1974). when studied in detail, should prove to be only Our study areascontained very many species, Occasional Users: brachyura, Xen- yet several Neotropical speciesthat are Spe- ops minutus (Furnariidae); Taraba major, cialists were absent from our Bolivian-Peru- Myrmotherula schisticolor, Myrmeciza exsul vian study sites: Philydor atricapillus and (Formicariidae); Caryothraustes poliogaster Thripophaga macroura (Parker, pers. observ.), (Cardinalinae); Mitrospingus cassinnii, Eu- Myrmotherula fulviventris (Orians 1969, Wi- phonia imitans, Cyanerpes lucidus (Thraupi- ley 1971, Willis 1972, Gradwohl and Green- nae); Hylophilus decurtatus(Vireonidae); Ca- berg 1980, 1982), M. haematonota (Willis cicus uropygialis (Icteridae); and Psilorhinus 1972; 100% of 27 observations searchingdead mexicanus (Corvidae). To this list of Occa- leaves, Parker, pers. observ.), M. erythrura sional Users should be added Xiphorhynchus (Willis 1972; Pearson 1977; Parker, pers. ob- flavigaster (Dendrocolaptidae; Willis 1960); serv.), and M. gutturalis (Willis 1972). (We Margarornis rubiginosusand Syndactyla sub- suspectthat the remaining Myrmotherula with alaris (Furnariidae; Ridgely 1976); Thamno- a spotted throat, M. gularis, will also be found philus punctatus (Formicariidae; 9% of 165 to be a Specialist when its foraging behavior foraging records, Oniki 1975; Gradwohl and is studied.) Greenberg 1982); pallida, Phily- Other Neotropical speciesabsent from our dor rufis, Thamnophilus cryptoleucus,Dysi- sites, or for which we have few observations, thamnus stictothorax and Perenostola schis- that should be found to be Regular Users are: tacea (Parker, pers. observ.); and Piezorhina Eubucco bourcieri (Slud 1964, Orians 1969, cinerea (Emberizinae; 3.6% of 55 observa- Ridgely 1976, Parker, pers. observ.), Crani- tions, S. W. Cardiff, unpubl. data). We suspect oleuca erythrops (Skutch 1969) Thripophaga that virtually every Neotropical foliage-glean- berlepschi,Anabazenops fuscus, Philydor lich- ing bird investigatescurled dead leaves occa- tensteini,Thryothorus longirostris (Parker, pers. sionally. We encourage all investigators who observ.), T. coraya (100% of 10 observations study the foraging behavior of forest birds to searching dead leaves; S. W. Cardiff, T. S. present their results in such a way that the Schulenberg, and A. P. Capparella, pers. degree of specialization on dead leaves can be comms.), T. nigrogularis (78% observations quantified. searching dead leaves; n = 18; Parker, pers. Forests of temperate latitudes of course also observ.), and T. rutilus (Slud 1964, Morton have dead leaves suspendedabove ground, and 1980). Descriptions of the foraging behavior these are searched for insects by some bird of Cacicus [Amblycercus] holosericeus by speciesto varying degrees. In particular, the Skutch (1954) and Slud (1964) combined with Worm-eating Warbler (Helmitheros vermivo- our own limited observations on this species rus) seems to use dead leaves heavily; even in Bolivia, indicate that this speciesmay be a vagrantindividuals in California 3,000 km west Regular User. Willis (1960, 1966) indicated of their normal migration route search dead that the ant-tanagersHabia rubica, H. guttur- leaves (Remsen, pers. observ.). Perhaps it is alis, and H. cristata may be Regular Users; our not coincidental that in coloration and bill- limited data for rubica also support this. We size the Worm-eating Warbler resembles fur- suspectthat when studied thoroughly many or nariids more than most parulids. Other North most speciesof Cranioleuca, Philydor, Auto- American speciesthat we have observed for- molus, Hyloctistes,and Thripadectes(see Slud aging regularly in dead-leaf clusters are the 1964 for T. rufibrunneus) will be classifiedas Carolina Chickadee (Parus carolinensis), Regular Users or Specialists.Our limited ob- Northern Oriole (Zcterusgalbula), and Gold- servations for Automolus dorsalis and A. me- en-winged Warbler (Vermivora chrysoptera); lanopezusindicate that they are probably Spe- the latter also feeds this way frequently on the cialists. Parker’s foraging data from lowland wintering grounds in the Neotropics (Slud Peru in 1983 indicate that Capito niger (Cap- 1964, Willis 1972:129, Ridgely 1976, Morton itonidae) is a Regular User of dead leaves: of 1980) as do the Blue-winged Warbler (V. pi- 24 foraging observations (not including those nus), Prothonotary Warbler (Protonotaria ci- feeding on fruit), 20 (83.3%) were of individ- trea) (Morton 1980), and the Worm-eating uals searchingdead leaves. This speciesoccurs Warbler (Willis 1960, Lack and Lack 1972). at the 250-m point on the transect in Figure At temperate latitudes, the strong season- DEAD-LEAF-SEARCHING BIRDS 4 1

ality in leaf-fall limits the degreeto which birds MORTON,E. S. 1980. Adaptations to seasonalchanges may specialize on searchingdead leaves, and by migrant land birds in the Panama Canal Zone, p. perhaps the degree to which insectsspecialize 437-453. In A. Keast and E. S. Morton [eds.],Migrant birds in the Neotropics: ecology, behavior, distribu- on them for refugia. At tropical latitudes, leaf- tion, and conservation. Smithsonian Inst. Press, fall is presumablymuch lessseasonal, and dead Washington, DC. leaves suspendedin the vegetation appear to ONIKI, Y. 1975. The behavior and ecology of Slaty be available year-round in most areas. Thus, Antshrikes (Thamnophiluspunctatus) on Barro Col- orado Island, Panama Canal Zone. An. Acad. Bras. as our data from Bolivia and Peru show, as Cienc. 471477-5 15. many as eight dead-leaf-searching Specialists ORIANS, G. H. 1969. The number of bird speciesin some may be added to Neotropical forest commu- tropical forests. Ecology 50:783-80 1. nities by the continuous airailability of this re- PARKER,T. A., III. 1982. Observationsof some unusual source. rainforest and marsh birds in southeasternPeru. Wil- son Bull. 941477-493. ACKNOWLEDGMENTS PEARSON,D. L. 1971. Vertical stratification of birds in a tropical dry forest. Condor 73:46-53. Our fieldwork in Bolivia and Peru has been generously PEARSON.D. L. 1977. Ecoloaical relationshinsof small supportedby Babette M. Odom, John S. McIlhenny, and antbirds in Amazonian b;rd communities Auk 94: H. Irving and Laura M. Schweppe.For permissionto work 283-292. in Bolivia we are grateful to Ing. Carlos Aguirre and Dr. PULLIAM,H. R. 1974. On the theory of optimal diets. Arturo Castaiios,Direcci6n de Ciencia y Tecnologia, and Am. Nat. 108:58-74. to Dr. Ovidio SuarezMorales and Ing. Antonio Saavedra REMSEN,J. V., JR., AND T. A. PARKER,III. In press. Con- Muiioz, Academia National de Ciencias. For permission tribution of river-created habitats to bird speciesrich- to work in Peru, we are grateful to Eric Cardich Bricefio. nessin Amazonia. Biotropica. SusanMoller-Hergt, and others in the Ministerio de Ag- RIDGELY,R. S. 1976. A guide to the birds of Panama. ricultura. Prof. Gast6n Beiarano nrovided invaluable heln Princeton Univ. Press, Princeton, NJ. to our studiesin Bolivia.“We thank Tom and Jo Heindel SCHOENER,T. W. 1971. Large-billed insectivorousbirds: for their hospitality in Bolivia and to Max Gunther D. of a precipitousdiversity gradient. Condor 73: 154-16 1. Peruvian Safarisfor the opportunity to work at Explorer’s SKUTCH.A. F. 1952. Life historv of the Chestnut-tailed Inn, Rio Tambopata, in Peru. Judy Gradwohl, Russell Auiomolus. Condor 54:93-100. Greenberg, Eugene S. Morton, T. S: Schulenberg,Susan SKUTCH,A. F. 1954. Life histories of Central American Allen-Stotz, and DouglasF. Stotz made critical comments birds. Pacific Coast Avifauna No. 3 1. on the manuscript. we are grateful to the many people SKUTCH,A. F. 1960. Life histories of Central American who have participated in collectingforaging data and pre- birds, II. Pacific Coast Avifauna No. 34. paring specimensin Bolivia: A. P. Capparella, C. S. and SKUTCH.A. F. 1967. Life histories of Central American S. W. Cardiff, L. S. Hale, S. M. Lanyon, M. SgnchezS., highland birds. Publ. Nuttall Omithol. Club No. 7. T. S. Schulenberg,and D. Wiedenfeld. S. W. Cardiff, M. SKUTC~,A. F. 1969. Life histories of Central American B. Robbins, and T. S. Schulenberggraciously provided birds. III. Pacific Coast Avifauna No. 35. additional foraging data from Peru. SKUTCH,k. F. 1972. Studiesof tropical American birds. Publ. Nuttall Omithol. Club No. 10. LITERATURE CITED SKUTCH.A. F. 1982. New studies of trooical American birds. Publ. Nuttall Omithol. Club Nb. 19. GRADWOHL, J., AND R. GREENBERG.1980. The forma- SLUD, P. 1964. The birds of Costa Rica. Distribution tion of antwren flockson Barro Colorado Island, Pan- and ecology. Bull. Am. Mus. Nat. Hist. 128. ama. Auk 97:385-395. TERBORGH,J. 1977. Bird speciesdiversity on an Andean GRADWOHLJ., AND R. GREENBERG.1982. The effect of elevational gradient. Ecology 58: lOOi-1019. a single speciesof avian predator on the TERBORGH.J. 1980a. Causesof trooical soeciesdiversitv. of aerial leaf litter. Ecology 63:581-583. Proc. XVII Int. Omithol. Con&. (19?8):955-96 1. ’ JANZEN,D. H. 1973. Sweep samples of tropical foliage TERBORGH,J. 1980b. Vertical stratification of a Neo- insects:effects of seasons,vegetation types, elevation, tropical forest bird community. Proc. XVII Int. Or- time of day, and insularity. Ecology 54:687-708. nithol. Congr. (1978):1005-1012. KARR, J. R. 1971. The structure of avian communities WILEY, R. H. 1971. Cooperative roles in mixed flocks in selectedPanama and Illinois habitats. Ecol. Mono- of antwrens (Formicariidae). Auk 88:881-892. gr. 41:207-233. WILLIS,E. 0. 19&O. A study ok the foragingbehavior of KARR, J. R. 1975. Production, energy pathways, and two speciesof ant-tanagers.Auk 77:150-l 70. community diversity in forest birds, p. 161-176. In WILLIS,E. 0. 1966. Ecologyand behavior of the Crested F. B. Golley and E. Medina [eds.],Tropical ecological Ant-tanager. Condor 68:56-7 1. systems.Springer-Verlag, New York. WILLIS.E. 0. 1972. The behavior of Snotted Antbirds. LACK, D., AND P. LACK 1972. Wintering warblers in Omithol. Monogr. No. 10. _ Jamaica. Living Bird 11:129-l 53. WILLIS,E. O., ANDY. ONIKI. 1978. Birds and army ants. MACARTHUR,R. H., AND E. R. PIANKA. 1966. On op- Annu. Rev. Ecol. Syst. 91243-263. timal use of a patchy environment. Am. Nat. 100: 603-609. - - Museum of Zoology, Louisiana State University,Baton MORTON,E. S. 1973. On the evolutionary advantages Rouge, Louisiana 70803. Received 13 January 1983. Fi- and disadvantagesof fruit-eating in tropical birds. nal acceptance11 June 1983. Am. Nat. 107:8-22.