Condor 84~272-285 0 The Cooper Ornithological Society 1982

THE BEHAVIOR OF BLACK-BANDED ( PICUMNUS)

EDWIN 0. WILLIS

ABSTRACT.-In forests and nearby second growth at Manaus, Brazil, Black- banded Woodcreepers (Dendrocoluptespicumnus) regularly foraged with army as well as away from ants. Like -following Dendrocolaptescerthia and Hylexetastesperrotii in the same study area, they tended to wait or hitch slowly on vertical thick trunks, then sally to or peck at prey. They were intermediate in dominance statusand in use of perchesnear the ground, where ants flushed most prey. of all three speciesalso foraged high above the ground when the ants sent probes up trees. Sleek-headed (presumed female) Black-banded Woodcree- pers supplanted their ruff-headed mates and helped in their disputeswith neigh- boring pairs; one or both birds of a pair sang near roost sites in morning and evening; and pairs trespassedto a limited extent on their neighbors’ areas. Large young were out of the nest with parents as early as July and as late as January, then wandered away from them. Breeding birds molted at about the time young left. At other study sites, the related svecies D. hofmannsi and D. platyrostris behaved like D. pi&m&.

Woodcreepersof the genusDendrocolaptes are pers to my presence. I captured and color- among species of birds that regularly follow banded 13 individuals at the Reserva. I took army ants for flushed prey in neotropical for- weights with Pesola scales and cloaca1 tem- ests(Willis and Oniki 1978). The behaviorally peratures with a Schultheis thermometer. little-known Black-banded (D. Vocalizationswere recordedwith a Uher 4000- picumnus) proved to be an important species Report S tape recorder at 19 cm/s. in the ant-following guild that I studied for over a year near Manaus, Brazil. Here I report MEASUREMENTS AND SEXUAL on studiesof Black-bandedWoodcreepers, and DIFFERENCES compare their behavior with that of related I examined 638 museum specimensof Black- woodcreepers. banded Woodcreepers, measuringculmen and wing chord for most, and did not find strong STUDY AREAS AND METHODS sexual differences in color or measurements. I briefly studied Black-banded Woodcreepers Average wing length of males exceeded that of at severallowland forest localities:Nappi Creek females by 0.5-2.9 mm in seven of eight pop- and Bartica in Guyana (On&i and Willis 1972); ulations considered, but overlap was great Umbria, Leticia, and Mitu, Colombia; Limon- within each population. Moreover, the 0.5-mm cocha and Putuimi in Ecuador; Yarinacocha difference was in nominate D. p. picumnus, and Andoas in Peru; Diamantina, Palhao, Car- which is the subspeciesat Manaus. Females auari, Serra do Navio, Almeirim, and the weighed more than most males in small sam- Campina Reserve (near Manaus) in Brazil. At ples from Surinam (combined x = 88.7 g, Reserva Ducke, Brazil, I studied them almost II = 6) and eastern Peru (X = 86.3 g, y1= 5). daily from 3 July 1973 to 27 August 1974. From Nicaragua northward (X = 61.0, y1= 6) Reserva Ducke (entranceat 2”55’S, 59”59’W) and in the Colombian and Venezuelan Andes is a 10 X 10 km forest tract of the Instituto (X = 73.8 g, IZ = 4), the specimens of D. pi- National de Pesquisasda Amazonia (INPA), cumnus were small, but sexes overlapped just north of Manaus. Tall upland forests at widely in weight. 100-l 20 m elevation are broken here and there In Dendrocolaptes species at Reserva by medium-height forests in sandy valleys of Ducke, two types of birds formed pairs: a sleek- small streams at 50-60 m elevation. Experi- headed individual dominated one that ruffed mental forestry tracts line an entrance road to up its head and breastevery time its mate came a laboratory clearing, and neighboring areas near. In Barred Woodcreepers, one sleek- are irregularly cleared for maniac and other headed that I collected proved to be a crops (Willis 1977). female, while two collected ruff-headed birds By watching with binoculars from behind were males (Willis, unpubl.). I assumed,there- swarms of ants ( burchelli), I accus- fore, that “ruffed” birds were males and tomed individual Black-banded Woodcree- “sleeked” birds were females in Black-banded

[2721 BLACK-BANDED WOODCREEPERS 213

Woodcreepers also, even though both types remained frozen. (In many ,the thin weighed about the same at Manaus (K = 82.8 and high “keening” associatedwith freezing is g, yt = 14). difficult for humans to hear.) I noted freezing Collectors have not recorded the head ruff- in responseto alarm calls of other birds (Plain- ing of their D. picumnus specimens, but the brown Woodcreeper, Dendrocinclafuliginosa, precedingparagraph is consistentwith the fact 8 times; D. picumnus, 2; Black Nunbird, Mon- that more males (303) have been taken than asa atra, 2; keening of Rufous-throated Ant- females (240). Sleek-headed birds normally bird, Gymnopithysrujigula, 1). Once, one clung were retiring at Manaus, leaving the subordi- close to a tree as Little Chachalacas (Ortalis nate ruff-headedbirds to foragenear the ground motmot) flew past. Woodcreepers generally and near me. However, it is possiblethat males seem to be protectively colored, and the pale sing more than females at dawn and dusk (see streakson a Black-banded Woodcreeper may below), a factor that could also bias collecting have increased its resemblance to furrowed in favor of males. bark or to spaces between epiphyte leaves. Streaking appears to be correlated with rapid OBSERVATIONS OF BLACK-BANDED hitching along dark trunks in woodcreepers; WOODCREEPERS unstreakedand barred forms wait more or use VOICE well-lighted trunks. Song. Black-banded Woodcreepers at Manaus Fleeing. These woodcreepers often fled to sangdescending series of 15 to 20 simple chirps, avoid danger, sometimes by sidling quickly lasting almost 2 s (Fig. 1A): “kie-ie-ie-ie-ie-ie- around a tree trunk, then freezing or hitching ie-ie-ie-ie-ie-ie-ie-ee-eu-eu-er.” Barred Wood- upward on the far side of the trunk. During creepers (Dendrocolaptescerthia) at Manaus pauses,the bird often turned its head, flitted had a similar but slower song, of 8 notes in the tips of its wings or wrists, or peered around 1.4 s. A (Dendroco- the trunk (Fig. 2A). An isolated snarl might laptesplatyrostris)from Fazenda Barreiro Rico, break the silence. Another method was flight S&o Paulo, gave a slow evening song like that to other trees, without the usual downward of a Barred Woodcreeper, 12 notes in 2.3 s swing to a lower perch characteristic of for- (Fig. 1C), but one from Fazenda Klabin, Espir- aging moves. When the presumed danger was ito Santo, (courtesy of Michael Gochfeld) sang a human or low-flying raptor, the bird often a fast songlike that of a Black-banded Wood- climbed to large branches 20-35 m above the creeper, 18 notes in 2.3 s (Fig. 1C). ground. Once, at the sound of a distant gun- Rattles. Series of short chirps at 6-8 per s, shot, a bird dropped instead to a low trunk; fluctuating in volume and changingirregularly on another occasion, a woodcreeper snarled in quality “ee-ee-ee-ee-ie-ie-e-e-ie-ei-” and so and backed down a trunk when macaws (Ara on (Fig. lF, H), were frequent after disputes ararauna) flew over. The bird might sleek the or in responseto playback of recordings.They head or ruff it when near the observer; ruffing sometimes graded into songs. could refer to the following behavior pattern, Squeals. Loud series of several “squeeh” however, as it was most characteristic of a notes were given during fights. returning bird or one reluctant to move as I Snarls.A growlingnoise followed by a squeal, approached. “chauhhh-eesk,” (Fig. 1G) seemed to be an Snarling, usually with fleeing, was specifi- alarm call. At times the last note of a snarl was cally recorded at the approach of hawks and less piercing, perhaps a different call. falcons (Micrastur ruficollis, 4; Leucopternis Grunts. Short faint series,“ uk-uk-uk,” were albicollis, 2; unidentified hawk, 2; Buteogallus given at nearby competitors. Barred Wood- urubitinga, I), at the soundof avian alarm calls creepersin Panama grunted similarly. ( fuliginosa, 4) at a gunshot or Peepsongs.The brief songsof fledglingBlack- passingmacaws, to nets (4) and to me (5). The banded Woodcreepers (Figs. lD, E) consisted snarlsat me were usually a less sharp version, of 12 or so chirps in 1 s, the song at times apparently without a sharp rise in pitch at the rising and then falling rather than falling end. I noted fleeing or nervous wing-flitting in throughout: “wh-e-e-e-e-e-e-e-e-e-e-up.” response to me and to another person who came up to me (even though the bird involved ALARM BEHAVIOR had been ignoringme), to passingcars, to alarm Freezing. I commonly saw birds simply freeze, calls (Dendrocincla fuliginosa, Pithys albi- pressingtheir bodies againstthe limb or trunk frons, 2; Ortalis motmot) and to a hawk (Leu- and remaining immobile for several seconds copternisalbicollis). to minutes. Immature birds stoppedpeepsing- Other birds often performed alarm move- ing while they froze. Occasionallya bird snarled ments or calls following the sharp snarls of as it froze, but I detected no sound while it Black-banded Woodcreepers, even when the EDWIN 0. WILLIS

TINEID SECDIIDS FIGURE 1. Sonogramsof Black-banded (B-b) and Planalto (P) woodcreeper vocalizations. A. B-b, song, Manaus. B. P, song, Fazenda Klabin, Espirito Santo (courtesyof Michael Gochfeld). C. P, song at Fazenda Barreiro Rico, Slo Paul0 (note changeof scales).D and E. B-b, peepsongsof young, Manaus. F. B-b, rattles, Manaus. G. B-b, two snarls, the first from a distant bird and hencewith the growling first note faint, Manaus. H. B-b, two chirping rattles, Manaus. snarls seemed undirected or were related to The sharp claws clung tightly except during intraspecific disputesrather than to predators. flight attempts, and soon scarred one’s fingers. Some Black-banded Woodcreepers preened Taming. Naive woodcreepersoften hid high after freezing or fleeing, before resuming for- above the observer for an hour or two or aging. showed other alarm behavior, then began to Mobbing. I saw no evidence of mobbing, feed if other birds were foraging. After I had except that moderately tame males sometimes studied some individuals for several hours or ruffed their heads near me, while other birds made scattered observations on them over a snarled behind cover. few weeks, they became tame enough to Struggling. In the hand or net, captured descendand forage nearby without much wing- woodcreepersoccasionally startled the captor flitting. Young birds were lesswary than adults. with an abrupt snarl or attempt to fly. They Like Barred Woodcreepers in Panama but struggled little otherwise, and did not peck. unlike those in Amazonia, Black-banded BLACK-BANDED WOODCREEPERS 215

Woodcreepersusually continued following ants when the observer stayed nearby. None of the Dendrocolaptesobserved ever seemed curious or aggressiveunless I was near a good foraging site that the bird was approaching.

FORAGING BEHAVIOR Undisturbed Black-banded Woodcreepers sal- lied from trunks for much of their prey, and seldom investigated crevices to peck out prey in the fashion of true creepers (Certhiidae) or related woodcreepersin the genusXiphorhyn- thus. Black-banded Woodcreepers commonly followed army ants, but sometimes foraged alone or with bird flocks away from ants. Foraging over army ants. Black-banded Woodcreepers,foraging over army ants at Ma- naus, foraged below 1 m part of the time, espe- cially when competing Red-billed Woodcreep- ers ( perrotii, 120-g birds) were absent. Black-banded Woodcreepers foraged below one meter 24.7% of the time when Red- billed Woodcreepers were nearby (1,078 rec- ords) and 27.9% when they were not (2,480 records). Percentagesof records below 1 m (in 0.1 -m intervals) for individuals foraging near H. perrotii were: 0.4% on the ground, 0.7% up too.1 m,4.7%to0.2m, 5.5%to0.3m,4.1% to 0.4 m, 3.4% to 0.5 m, 1.5% to 0.6 m, 1.2% FIGURE 2. Black-banded Woodcreepers, from field to 0.7 m, 1.3% to 0.8 m, 0.5% to 0.9 m, and sketches. A. Peering after alarm call of a Plain-brown Woodcreeper.B. Resting, with belly on the trunk. C. Def- 1.3% to 1.0 m. For individuals foraging away ecation, requiring lifting of the tail from the trunk. from H. perrotii, the series of percentagedis- tributions was: 0.3, 0.3, 3.3, 6.3, 5.6, 5.4, 2.4, 1.5, 1.1, 0.7, and 1.4. Black-banded Woodcreepers foraged high from 2.1 to 3 cm, 3.9% to 4 cm, and 5.5% to above the ground somewhat more than did 5cm,and34.1,24.2, 16.4,7.3,0.8%from5.1 Red-billed Woodcreepers,but lessthan Barred to 15, 15 to 25, 25 to 50, 50 to 100, and 100 Woodcreepers. Percentages of records of D. to 200 cm diameter, respectively. These birds picurnnus l-10 m up were, respectively (in l- thus used 5-25 cm perches only 58.3% of the m intervals), 12.4, 8.6, 12.4, 9.2, 6.2, 4.5, 2.1, time and capturedonly 50.5% of 366 prey from 2.9, 3.0 when near H. perrotii and 14.2, 8.9, them. By contrast, Barred Woodcreepers over 9.8, 7.9, 5.4,4.9,4.2, 2.4,2.6 whenawayfrom army ants at Manaus used perches5-25 cm in it. From 10 to 15, 15 to 20, and 20-30 m up diameter for 73% of 937 records (Willis, the respective percentageswere 8.3, 4.4, 1.2 unpubl.). Perhaps the former species, being when near H. perrotii and 6.9, 4.1, 0.8 when dominant to the latter, was less likely to be away from it. chasedoff a perch under 5 cm in diameter; and Black-banded Woodcreepers used near-ver- Black-banded Woodcreepers often displaced tical perches most of the time. For 3,158 rec- their smaller (65-g) relatives at strong ant ords of perch angle, percentagesin 20-degree probes up very large trunks. intervals from 0” to 180” were (respectively) Black-banded Woodcreepers rarely used 0.7, 0.7, 0.7, 6.0, 88.9, 2.5, 0.3, 0.1, 0.03; one perchesunder 1 cm in diameter, which some- record (0.03%) was at 340”. (This notation is times bent under a bird’s weight. Perches up best understood by rotating parallel pencils, to 2 cm diameter often split the bird’s tail, so representingwoodcreeper and perch, from 0 that it rested on the fragile and quickly dam- or horizontal past 90” or vertical and then past aged inner webs rather than on the stiffened 180”to 360”; the “woodcreeper” startsand ends tips of the . Perches less than 3 cm above the “perch.“) diameter were good vantages for prey, repre- Diameters of 2,657 perchesof Black-banded senting 12.8% of 366 perchesbefore captures, Woodcreepersover army ants were distributed largely becausethe birds often had to use slen- 0.7% under 1 cm, 2.9% from 1.1 to 2 cm, 4.3% der saplingsin order to forage near the ground 216 EDWIN 0. WILLIS

TABLE 1. Placesand methods of foragingused by Black- TABLE 2. Vertical and horizontal distancesof sallying banded Woodcreepersover army ants. The valuesare per- Black-bandedWoodcreepers over army ants. The values centagesof 1,002 prey items. are numbers of records. Upward and downward move- ments are separatedby a slant line. Method Tossing Un- Placeof capture Sallying Gleaning Digging leaves known Vertical Honzontal distance(m) dxsance Cm) 1 2 3 4 5 Ground 29.8% 1.5 0.1 Trunk, liana, log 4.4 17.4 0.5 0.1 0.1 4 l/O Leaf, twig 23.4 4.6 0.3 3 O/l O/l Debris 0.4 0.7 2 2/l Termite, ant nest 0.1 0.2 1 18115 3/3 l/O Air 8.1 1.2 0 13 1 1 Not determined 6.0 1.2 aThe birdfliesapproximatelythe hypotenuseofthetwo indicateddistances. Total 72.1 26.6 0.7 0.2 0.4

mostly diving at or after flying or falling prey, (the best zone for ants). Trunks over 15 cm especially during emergencesof winged ants. diameter were even better vantages, repre- Diving after prey was about as common as senting57.4% of 366 perchestaken before prey. sallying upward (Table 2) whereas Barred When Black-banded Woodcreepers were Woodcreepers mostly sallied upward. hunting near Red-billed Woodcreepers, 28.5% Black-banded Woodcreepers seldom dug or of prey captures(n = 274) were on the ground, tossed leaves from cavities or termite nests when away from Red-billed Woodcreepers, above the ground, a rather frequent technique 36.3% (n = 633) were on the ground. Perches of large-billed Red-billed Woodcreepers. below 1 m were disproportionately repre- Gleaning wasless common than sallying,unlike sented among prey-capture records compared such related woodcreepers as the Chestnut- to records for all foraging perches, both near rumped Woodcreeper (Xiphorhynchuspardal- Red-billed Woodcreepers(28% of 124 records) otus) at Manaus but like Barred Woodcreepers and away from them (33% of 268 records). there. Barred Woodcreepersat Manaus capturedonly Prey (Table 3) of Black-bandedand of Barred about one-eighth of their prey on the ground woodcreeperswere similar at Manaus, except (Willis, unpubl.). Since Black-banded Wood- that no fast-flying cicadas nor moths fell prey creepers captured three times as many prey to the former, a slow-flying species,while more over ants as did Barred Woodcreepers,the for- winged ants did so. (Black-banded and other mer captured seven times as many prey on the large woodcreepers quickly chased Barred ground and over twice as many prey above the Woodcreepers away from ant emergences.) ground as did Barred Woodcreepers. Grasshoppers, katydids, roaches, and other Black-banded Woodcreepers captured prey orthopteranswere common prey, aswere other in diverse locations, using sallying and other large primitive arthropods on decaying trunks. techniques (Table 1). Sallying to the ground The modal prey length was less than the beak (the commonest type of capture in ant-follow- length, and D. picumnus captured smaller prey ing birds) represented nearly 30% of records, than did D. . (Prey size was estimated unlike 5-l 5% for Barred Woodcreepersin var- in the field asa fraction of exposedbeak length, ious localities (Willis, unpubl.). Sallying to 1B = 35 mm in this case;prey lengthsin Table foliage, the next most common type of prey 3 are head-to-vent and do not include pro- capture, was less than half as frequent as in jecting legs, antennae or cerci.) Black-banded Barred Woodcreepersat Manaus (where 5 5.6% Woodcreepers could, perhaps,take small prey of the 295 records were for that species).Sal- becausethey often hunted near army ants, at lying to palm fronds accounted for 42% (98/ ant emergencesor on the ground, while Barred 234) of my records of Black-banded Wood- Woodcreepers generally sallied to distant creepersthat were foliage-sallying at Manaus, foliage and ignored tiny prey. D. picumnus even more than for Barred Woodcreepers captured prey as far as 1.5 m away when sal- foliage-sallying there (24%). Pecking or glean- lying to the ground, but 10 of 35 records were ing prey from trunks or limbs was slightly more lessthan 0.2 m away, while 28 of 35 were less frequent by Black-banded than by Barred than 0.5 m off. Woodcreepers at Manaus (10.8%) but not as Black-bandedand Barred woodcreepersboth common as in the latter species in Panama hammered large prey against tree trunks or (36.7% of 294 records). shook and crushed them vigorously, but did Sallying to the air, as in the aerial sallies not dissectthem piece by piece on the ground. performed by Barred Woodcreepers, involved If the former had not had long bills, they could BLACK-BANDED WOODCREEPERS 217

TABLE 3. Types and sizes of prey taken by Black-banded Woodcreepersforaging over army ants.

Number of prey of given size (mm) PRV ? uo to: 10 20 30 40 50 60 70 105

Scorpion 4 I 4 Whip-scorpion 1 2 2 Spider 1 4 6 5 4 Centipede 1 3 1 5 4 4 Roach 11 5 17 7 5 Katydid 9 4 9 17 9 5 2 Beetle 5 Caterpillar 3 1 Ant, larva 1 winged 3: 3 2 Lizard 1 Unidentified 85 29 14 6 2 Total 65 88 52 57 48 25 17 10 4

not have handled any large prey, whereas ant- use large vertical perchesthan were crosswise- birds with much shorter beaks easily dissected clinging . suchprey on the ground. Black-banded Wood- Above the ground, both of thesewoodcreep- creepers gulped down the prey body and let ers hitched short distances and waited for a projectingwings or legsfall offto the antsbelow. few secondsto minutes looking for distant prey. Large katydids sometimes pushed vigorously They seldom examined the nearby trunk with their spiny legs,causing the woodcreepers minutely or pecked at it, woodpecker-style. to almost releasethem or to jugglethem upward These woodcreepers tended to follow ant for recapture.Tiny prey was sometimescrushed probes up trunks and to check certain types of in the tip of the beak rather than in the gape, sites where ants regularly flushed prey. They as if noxious; the bends of the wings went out visited palm fronds and crowns (229 records as in “anting” behavior. Scorpions were in my notes), rotten snagsor stubs(154) hol- crushed or hammered at the head end, often low “acariquara roxa” trees (128) slits or holes until the tail dropped off (see also Haver- in trunks (75) liana-crowded trunks (70) and Schmidt 1977). Whip-scorpions were pounded under or near epiphytes (5 5). Termite runways until the long legs dropped off. Large beetles (30) or termite and ant nests (11) attracted required much hammering and crushing;I once them, probably becausesuch runways are often saw a scarabrepeatedly pop from the beak and on dead portions of trunks. Fluted trunks (13) be retaught in mid-air. These woodcreepers and tree crotches (12) were visited. Occasion- rubbed hairy caterpillars on trunks or leafy ally one visited a treefall (6) but such sites trash. One green caterpillar was rubbed under generally had only horizontal and cluttered the tip of a wing several times. On another perchesin which antbirds were more at home. occasion,a bird capturing an ant with a larva, I saw Black-banded Woodcreepers flying back carefully dropped the ant and ate the larva. and forth around suchtangles, unable to enter. One roach was turned with the tongue so that They infrequently foraged on logs (4 records), it was head-first rather than tail-first before loose bark (4) and debris (2). One tore off a swallowing.Birds sometimeswiped their beaks philodendron leaf plastered on a trunk and after crushing large prey. peered underneath. Black-bandedand Barred woodcreepersthus Emergencesof winged ants above the ground foraged much alike at Manaus, except that the led to woodcreeper “circuses,” with as many former were more regular at ant swarms and asseven species fluttering about actively. Black- did more foraging near the ground. Both bandedWoodcreepers tended to rapidly replace worked similarly near the ground, looking back low-foraging White-chinned Woodcreepers over their shoulders from moderately large (Dendrocinclame&a) and subordinateBarred perchesand diving rather clumsily to snap up Woodcreepers, but Black-banded and Plain- prey fleeing the ants. They preferred fairly open brown woodcreepersstayed even if Red-billed understory, being larger and not so maneu- Woodcreepers scared them to the periphery. verable as antbirds, and perched only with dif- Chestnut-rumped Woodcreepers might hitch ficulty on slender growth. In dense under- up to get prey on trunks, and Long-billed growth, both woodcreepers tended to cease Woodcreepers(Nasica longirostris)might glean foraging near the ground and to hitch upward. on limbs. Black-banded Woodcreepers mostly However, they were heavier and better able to sallied downward for flying ants from near the 278 EDWIN 0. WILLIS

TABLE 4. Competitive interactions of Black-banded They rarely followed mixed-species flocks. Woodcreepersover army ants. Normally they hitched up tree trunks or major limbs, looked about at stops, and glided off in Number of interaction9 catenarycurves to other trees.Sallying to foliage SUP- Dis- plant- plkX- RC- lg- was recorded once. Svxies ing I%% Fight turn nore

Dendrocolaptespicum- COMPETITIVE BEHAVIOR nus 312 33 11 2 Interspecific. Black-banded Woodcreepers Dendrocinclafiliginosa 151 30 1 attacked a few birds over army ants, mostly D. merula 52 5 3 Dendrocolaptescerthia 32 27 2 2 woodcreepers, with a moderate number of Hylexetastesperrotii I44 I14 attacks on antbirds (Table 4). I seldom wit- Xiphorhynchuspardalo- nessedother attacks: trogons attacked a few of tus 17/3 7/l these woodcreepers near trogon nests, and a Nasica longirostris /3 /I 14 woodpecker attacked one that chasedit around Gymnopithysrufgula 43 8 Pithys albtfrons 21 1 5 a trunk away from ants. Shortly after this inci- Percnostolarujifrons 4 dent, the woodcreeperattacked several caciques Cacicushaemorrhous 5 (Cacicus sp.). Pharomachruspavoni- Attacks on antbirds were generally simple nus /4 Trogon violaceus /I “supplantings” or replacement of one bird by Monasa atra another near the ground. Some Rufous- Campephilusrubricollis 11: throated Antbirds chipped in alarm at sudden Turdus albicollis 1 attacks, but normally I heard no sound. These Ramphoceluscarbo 1 woodcreeperssometimes attacked antbirds in r Recordsaloneorbefore/denotedominancebyBlack-banded Woodcreep- en; recordsafter I denotedominance over these birds. food-catching attempts (six casesfor G. ruji- gula, four for Pithys albifrons, two for Perc- nostola rufifrons). Antbirds often moved off, “displaced,” at a slight lunge or stare by the exit holes. They then hitched rapidly up to just woodcreeper. They usually stayed within a below the exit holes again, or as close to the meter or two, even “returning” to the perch holes as Red-billed Woodcreeperswould allow just vacated by the woodcreeper in the caseof them to come. Meanwhile, Plain-brown P. albij-ons (a species that infiltrates persis- Woodcreepers sallied out from hiding places tently behind or near large competitors). Black- around peripheral vines and limbs. They and banded Woodcreepers near the ground occa- Black-bandedWoodcreepers were often the last sionally grunted at the small antbirds swirling to leave such a site, and reaped the benefits of nearby. Large antbirds suchas speciesof Phle- being the first at new emergences. gopsisin other regions seemed to be avoided, Between periods of watching or of activity for Black-banded or Hoffmanns’ (Dendroco- at special sites, Black-banded Woodcreepers laptes hoflmannsz] woodcreepers usually for- often wandered near swarms. They did not aged away from the ground whenever these spiral around trunks as much as did Barred antbirds were present. Woodcreepers,but otherwise similarly hitched Attacks by and on other woodcreepers up irregularly. Often they hitched up along the formed a considerable part of competitive upper sides of canopy limbs, only to glide off behavior. Gleaning trunk-foraging woodcree- with short bursts of flapping to other limbs or pers in the genera Xiphorhynchusand Nasica to lower trunks. Regularly, they came to any seemed pugnacious, and often initiated or activity of other woodcreepers or antbirds. received attacks.Chestnut-rumped Woodcree- Their open-wingeddownward glides,preceded pers, the smaller speciesat Manaus, occasion- at times by heavy flutters, sometimes scared ally hitched rapidly along toward Black-banded off small birds from such places. Meanwhile, Woodcreepers, pecking and dodging return Barred Woodcreepers wandered high and pecks. The latter normally chased the former, peripherally, and appeared mainly when a par- rather than vice versa. Long-billed Woodcree- ticular zone of activity became quiet. Both of pers, a large, limb-foraging species, initiated these woodcreepers seemed the same size in several attacks and fights with Black-banded the field, despite differencesin weight, and the Woodcreepers despite their rarity at Reserva two overlapped greatly in foraging over ants. Ducke (beingmainly a bird of floodplain woods Foraging awayfrom ants. At Manaus, Black- or varzeas). Away from Manaus, fairly large banded Woodcreepers infrequently foraged (60-65 g) Buff-throated Woodcreepers(X. gut- away from ants. They usually stayed high in tatus) sometimes stoppedtheir trunk- and epi- the canopy or in the forest mid-levels, where phyte-peering long enough to attack, engagein they wandered steadily as if searchingfor ants. a pecking duel, or drop in a fluttering aerial BLACK-BANDED WOODCREEPERS 219 duel with Black-banded or Hoffmanns’ wood- it. When attempting to feed, Barred Wood- creepers.The Buff-throated Woodcreeper usu- creepers were sometimes displaced (5 of 27) ally won, but at times a Hoffmanns’ or Black- or supplanted (1 of 32) by Black-banded banded woodcreeper snarled and jabbed back Woodcreepers. The latter, when feeding, were or supplanted the fast-climbing opponent. A sometimes supplanted (4 of 44) or displaced Strong-billed Woodcreeper (Xiphocoluptes (1 of 14) by Red-billed Woodcreepers. promeropirhynchus)was once seento displace Intraspecijc. Black-banded Woodcreepers a Black-banded Woodcreeper at Diamantina. lived in pairs, although male and female some- Medium-sized speciesof these trunk-climbing times fed at different coloniesof ants.The sleek- genera perhaps win at times over species of headed bird (female?) displaced or supplanted Dendrocolaptesbecause the latter cannot climb the ruff-headed bird (male?) when they were and dodge as rapidly, being adapted for sal- at the same ant raid, and the latter took what- lying rather than for trunk foraging. ever foraging sites were left. One evening, I Medium-sized (40-5 5 g) ant-following watched a pair threaten each other and fight, woodcreepersof the Dendrocincla were tumbling fluttering to the ground before sep- regularly attacked by Black-banded and Hoff- arating. Disputes between Black-banded manns’ woodcreepers. Uncommon low-for- Woodcreeperswere relatively frequent, for the aging White-chinned Woodcreepers at Ma- male and the female of a pair were antagonists, naus were attacked about one-third as often as while boundary disputes and territorial tres- common, high-foraging Plain-brown Wood- passersled to other attacks. Only 8 of 3 12 creepers, even though Black-banded Wood- intraspecific supplantings and 1 of 11 fights creepers spent only one-third as much time were connected with feeding chases. When near the ground as high above it. (Low-for- neighboring pairs met over ant swarms, noisy aging birds concentrate more than do high- chaseswith squealingusually ensued. Two pairs foraging ones, however.) Attacks were usually rarely stayed at a single ant raid unless they simple, rarely with a gape or snap of the beak, could forageover different branchesof the raid. and were usually without calls except for occa- Wandering individuals, including immatures, sional chattersor “stieking” notesfrom the sup- were chasedwhen encountered,but wandering planted dendrocincla, or a grunt from the birds often hid persistently at distant parts of attacker. Black-banded Woodcreeperswhirred a raid and hence followed jointly with terri- their wings loudly in some attacks. Occasion- torial pairs. ally one hitched rapidly up a trunk after a den- Disputes between mates, or between parents drocincla. Dendrocinclas alighting near Black- and dependent young, normally involved sim- banded Woodcreepers were likely to be driven ple chases, supplantings, or displacings with off, as were those that tried for prey nearby (17 little display or calling; after head ruffing or of 15 1 supplantingsand 2 of 30 displacingsfor sleeking and a grunt or squeal, the two sepa- D. fuliginosa, plus 3 of 52 supplantingsof D. rated. Young birds supplantedthe ruff-headed me&a, were in prey-catching attempts). parent in disputesbut avoided the sleek-headed Black-banded Woodcreepers occasionally parent’s lunges.One young grunted at the ruff- supplantedor displacedevasive Barred Wood- headed parent before supplantingit. At times, creepers, while Red-billed Woodcreepers, however, the sleek-headedparent fled from the which were larger, regularly supplanted both. young bird as it followed for food. Families Small birds regularly avoided larger ones in often flitted their wingswhen together, perhaps this group of ant-following woodcreepers, a sign of low-intensity aggression. reducingthe number of confrontations.Attacks Disputes between residents and nonresi- by Black-banded upon Barred woodcreepers dents, or between wandering birds, were often were much less frequent than their intraspe- vigorous and noisy. Squeals and songs, flut- cific attacks, and tended to be quieter: a grunt, tering aerial duels or chases,and fightserupted a growl, or rarely a squeal or two. Barred when a new bird or pair appeared at an ant Woodcreepers often hid by hitching quickly colony. The wrists and alulars were often around trunks. Red-billed and Black-banded spread,as were the body feathers.Sleek-headed woodcreeperssometimes did the reverse, sud- birds never spread the head feathers, but ruff- denly lunging around a trunk at a surprised headed individuals consistently did so. Ruff- smaller woodcreeper. Favored sites, such as headed birds often ruffed the throat and the hollow trees, rotten stubs, holes in trunks or breast feathers (i.e., all the streaked feathers) palm crowns, sometimes attracted first a in and out. Only the back, breast and lower Barred, then a Black-banded,and finally a Red- neck were ruffed out in sleek-headedbirds (Fig. billed woodcreeper. One Barred Woodcreeper 3A, B). Birds of both types spread their tails dropped 4 m at a look from a Black-banded and wing tips somewhat, and pushed the body Woodcreeper, then hitched slowly up toward out from the trunk. There was at times slow 280 EDWIN 0. WILLIS

they bowed their heads at each other. Their backsand scapularswere very ruffed out, their breast plumage was expanded but rather flat (except for a central groove), their wingtips were out and were flitted, but their wrists were concealedin the spreadbody feathers. Finally, the intruding young started singingloudly and fled down to near the ground. The resident young followed, the intruder snarled, and both began to forage over separate parts of the ant swarm. Once the intruder tried to supplantthe resident young, but it pecked back and stayed. The resident female was not present, presum- ably being at another ant colony. On a later occasion, the same resident ruff- headed bird was more aggressive, snarling loudly and ruffing his head at a quietly singing intruder, then chasingthe squealingbird about. Meanwhile, his sleek-headed mate flitted her wings and foraged; later she supplanted the intruder briefly. On another occasion,the same ruff-headed bird spread its head feathers at a trespassingruff-headed bird, then hitched up and chased it. In Barred Woodpeckersi males (ruff-headed birds) also feuded noisily on territories while females (sleek-headedbirds) disputed quietly; females dominated males despite seemingly FIGURE 3. Disputing Black-banded Woodcreepers, from field sketches.A. Sleek-headedaggressive bird jab- more vigorous display from males (Willis, bing the head toward an opponent above. B. Sleek-headed unpubl.). My observations of Black-banded bird threatening bird below. C. Sleek-headedsubmissive Woodcreepers suggestthat even trespassing bird. D. Ruff-headed submissivebird. sleek-headedimmatures seem capable of dis- placing r&I-headed resident adults at times. Black-banded Woodcreepers, like other forward and upward jabbing of the bill. Sleek- species of Dendrocolaptes, regularly sang at headed birds sang faintly, while r&-headed dawn and dusk. In 1972 to 1976, a pair of birds sang and squealed noisily. In one case, these birds roosted every night inside bro- two disputing ruff-headed birds gave only meliads 12 m up on a tree trunk outside the strangledsnarls as long as a Black-faced Hawk dining room at Reserva Ducke. Both birds sang (Leucopternis melanops) waited above, then as they approachedthe roost site at dark, occa- flew back and forth past each other with pierc- sionally even on the roost tree, before sepa- ing squealsfrom the losing bird. Submissive rately hitching up inside the bromeliads. At birds tended to retract the head and to cling dawn, a few songsmarked the departure of one close to the trunk (Fig. 3C, D). or both. Probably territories are marked by I watched a dispute among three banded these song periods near roost sites. Songs at birds in tall second growth, peripheral to the other hours marked disputes, either over or forest zone that was the center of the pair’s away from ants, and were infrequent. activities. One nearly-independent young (a sleek-headedindividual) disputed with a wan- REPRODUCTIVE BEHAVIOR dering sleek-headed young while the ruff- The behavior of mated birds and of singing headed parent foraged peripherally; when the birds is described above, see Competitive intruding young supplantedhim, he ruffed his Behavior, Intraspecific. head feathers and started singingfaintly, then Salmon (Sclater and Salvin 1879) found two hitched up and fled. The resident young and white eggs29 X 22 mm in a tree hole in north- the intruding one were silent except for a few ern Colombia. sharp snarls at first, and ascendedin displays I found several full-sized peepsingingyoung until out crosswiseon terminal twigs of the out of the nest between 6 July and early Jan- small tree. When on vertical perches, they uary at Manaus, and banded three of them. pointed their sleeked and extended heads and The singleyoung bird in each family normally necks skyward, but when on horizontal twigs followed the sleek-headedbird of the pair, but BLACK-BANDED WOODCREEPERS 281 in one case a young bird received food from both birds. The young resembled the adult except for fluffier plumage. For severalmonths, it peepsang repeatedly every 4 s or so as it wandered near the presumed female or on its own around ants. If the female captured prey and started to beat it on a trunk, the juvenile quickly flew over or hitched up to her. If she did not dodge away, it often fluttered its wings and gaped (Fig. 4A), then received the food as it hitched rapidly up past her. One feeding by a male was similar. The young then crushed and flailed the prey on its own, and swallowed it. Young birds sometimes pecked off and chewed bits of bark or leaves or stood on the ground (Fig. 4B), but gradually began to cap- ture prey items. They apparently stayed with their parents for several (3-4) months. Occa- sionally two young were together, with or with- out parents, in November or January. After- ward, they wandered away from their parents and disappeared from my study area. Other birds, thought to be immatures, wandered into the study area at this time (Fig. 5). Some young peepsangafter leaving their parents, as late as 21 January. FIGURE 4. Young Black-banded Woodcreepers, from field sketches.A. Fluttering while beggingfrom parent. B. Standingon the ground. MAINTENANCE BEHAVIOR AND MOLT Preening birds used vertical perches,although more perches were subvertical than in ordi- nary foraging (of 129 preening perches, 12% open in hot weather. In rain, an immature bird were 60-80” and 86% were SO-1007. Most hitched up under a dead Cecropia leaf caught preening perches were l-10 m up, and few in a tree fork and clung close under it with (10% of 129 records) were below 1 m. Perches head retracted, remaining for many minutes, 5-15 cm in diameter were used more often until the rain ended. (42% of 12 1 preeningrecords) than in foraging, One bird was seento cough up fragments of while perches under 5 cm diameter were sel- insect exoskeletons. Excreta were ejected dom used (8% of 12 1 records). Probably slen- forcefully outward and backward as the tail der, low, and overhanging perches were used was lifted well off the trunk for a fraction of a in foraging more than in preening becauseprey second (Fig. 2C). When army ants attacked could be captured from them, while larger, them, these woodcreepers jittered back and higher, and subvertical perches were used in forth from one leg to the other, pecked ants off preening because they were safer and con- their toes, or fled. served effort. Birds preened in the open forest Molt. Remiges were molted from the inner- understory, and did not go to densecover first. most primary (no. 1) outward and rectrices The head, whether ruffed or not, was were molted from the innermost pair outward. scratched over the wing (16 records). Once a Wing molt in Central America took about five bird wiped its head on a tree trunk. Resting months, with birds starting molt noted from birds clung close to the trunk, and body feath- June to mid-July and ones ending molt from ers were fluffed over the corners of the wing mid-October to mid-December. At Reserva (Fig. 2B). After preening or resting, stretching Ducke, one presumed male molted two pri- motions were much as in other birds. When maries in just over a month, and was in the bathing in a cavity, one bird was seento flutter, middle of wing molt in mid-November. His hitch upward to preen nearby, and then back mate wasnot so advanced in molt, for primary down into the cavity to bathe some more. no. 3 was new on each side in early November Another bird flattenedout and “sunned” briefly (she was still feeding their grown young at this when it hitched into sunlight on a tree limb. time). Another bird was ending molt in early Normally these woodcreepersremained in the February. Molt apparently followed breeding shade.Even so,they sometimeskept their beaks and was later than in Central America. 282 EDWIN 0. WILLIS

/hSecond Growth \,

JJ" ’ c-\ \ J \ Forest a- Edge

BLACK-BANDED WOODCREEPER

H - N = Feb.-Aug. i 974

0 meters 500 I

FIGURE 5. Wandering of an independent Black-banded Woodcreeper at Reserva Ducke. Kilometer posts along the highway from Manaus to Itacoatiara are indicated by triangles; a side road is indicated by a double dotted line; a stream is indicated by a line interrupted by dots.

Temperatures. Cloaca1 temperatures, not forests as well. In Central America, habitats counting one bird at 4 1.9”C, varied from range from bromeliad-crowded montane for- 43.0-44.o”c (K = 43.3, y1= 13). ests to semi-open pine-oak woodlands (Land 1970). SPATIAL BEHAVIOR At Manaus, pairs occupied different, some- In the Amazon region, Black-banded Wood- what overlapping areas that seemed to be ter- creepersprincipally occupiedupland forest and ritories (see Competitive Behavior, Intraspe- forest edge,where they rangedfrom large limbs cific). In the best-studiedpair (Fig. 6), the male, near the canopy down to the ground-the last female, or both birds occasionally wandered mainly when following army ants. At Manaus, well into the territories of other pairs for a few they regularly wandered from the forest into days when following an ant colony, but des- plantations, orchards, and tall or even semi- erted the colonies after the local pairs moved open secondgrowth. In northern Venezuela, I in. The young bird wandered with the pre- saw individuals (captured by J. Terborgh) in sumed female for several months, then wan- dense thorn woodland of the northern side of dered irregularly around the parental area. Ranch0 Grande National Park, and Paul Other presumed immatures (Fig. 5) moved Schwartz (pers. comm.) found them in tree- through the territories of settled pairs. One shaded coastal cocoa plantation and montane wandering presumed male gained a mate in BLACK-BANDED WOODCREEPERS 283

KM-25 A ~*YSRB

RESERVA DUCKE , BRAZIL A-F = July- Dec. 1973 G-L = Jan. - June 1974

FIGURE 6. Home area of a pair of Black-banded Woodcreepersnear Reserva Ducke, and early wandering of their young bird. Upper-case letters are records of a parent bird or pair; underlined letters indicate young and old bird(s); lower-caseletters indicate records of the young bird alone. Approximate areas of neighboring pairs are indicated by “home range edge” lines.

June 1974 and became settled; it was noted Nova Olinda do Norte and near Tangara da singingon several occasionsin June and July. Serra (Willis 1976) behaved like D. picumnus. Densities were low, being about five settled It was usually in floodplain forests or varzeas, males, five settledfemales, and four wandering but probably occurs in upland forests of the birds on 2.8 km2 of study area, or some 4 g/ region as well. Both forms of the snarling note ha, in May of 1974. Densities had been some- of D. hoffmannsi were much like those of D. what higher in July to November 1973, when picumnus;I did not note other calls.The sneez- several young birds were also present-over ing form of the call (“wh’kai”) was associated one per km2. Wandering birds were less with disputes,which the sleek-headedmember numerous at that time, however, being about of a pair always won; the sharp form of the one birdlkm2. I rarely found more than three call (“kaihh-jeep”) wasgiven in seemingalarm, Black-banded Woodcreepers at a singleswarm with wing-twitching and with hitching upward. of ants (Table 5). One sleek-headed bird ruffed the throat but not the crown. Of my 33 records of foraging, BEHAVIOR OF RELATED SPECIES one was a sally to the ground, 11 were sallies Hoffmanns’ Woodcreeper, a Brazilian species to palm fronds, 8 to other leaves, 4 each to studied briefly at Maloquinha, Borba, Coata, trunks and lianas, 2 to the air, and 3 to unre- 284 EDWIN 0. WILLIS

TABLE 5. Swarms of Eciton burchelli ants with given snarled; another gave a series of squirrel-like numbers of Black-bandedWoodcreepers. grunts, then started rattling. Planalto Wood- creepers foraged like Black-banded Wood- Number of swarms with given number of birds creepers,but were more often away from army Month 0 1 2 3 4 5 ants. Over ants, they often foraged low, sal- July 1973 3 3 4 7 lying to the ground for prey, for the ant-fol- August 18 18 16 lowing birds of most of their range were such September 61 22 7 small forms as fire-eyes (Pyriglena spp.) and October 19 7 17 15 tanagers(Trichothraupis spp.). I sawthem sup- November 36 12 18 11 4 December 28 11 9 13 5 planted by Buff-throated Woodcreepers (Mar- January 1974 25 25 15 11 1 aba, Bacabal), while they supplanted Plain- February 7 15 : 10 2 7 brown Woodcreepers (Bacabal), Thrushlike March 11 16 94 Woodcreepers (Dendrocincla turdina), fire- April 22 25 14 7 1 eyes, and Trichothraupis tanagers. Occasion- May 23 27 9 1 June 18 6 12 3 ally, especially when wary of an observer, July 40 7 18 5 Planalto Woodcreepersforaged high above the August 15 11 7 1 ground. Away from ants, they hitched slowly Total (Sept.-Aug.) 305 202 159 86 17 7 upward or looked about from medium-sized trunks and limbs from near the forest floor to the sub-canopy, sallied for prey or pecked it from nearby lianas or leaves, and glided to low corded sites. Of 12 records of gleaning, 4 were on other trunks. I watched one visit light poles from leaves, 3 from lianas, 2 from trunks, and at the forest reserve headquarters at Carlos 1 each from a palm frond, twig, and debris. Botelho Reserve (near S%oMiguel Arcanjo, S%o These woodcreepers competed with several Paulo State) for insects in the early morning. large antbirds near the ground, and foraged Others wandered low on the trunks of “cer- near the ground less than did Black-banded rado” (savanna)trees in the early morning both Woodcreepers at Manaus. One bird, foraging in Maranhao and Minas Gerais (Brejo Janu- away from antsat Nova Olinda do Norte stayed aria). In pairs, one bird ruffed the head while 20-30 m up on fairly large limbs. Another, at the other was sleek-headed;but birds were gen- Coata, while gliding down over ants and show- erally solitary. They sometimesfollowed mixed ing the yellowish wing linings, occasionally bird flocks, but not persistently. alarmed White-breasted Antbirds (Rhegma- It is possiblethat D. hofmannsi and D. pla- torhina hoflmannsz].Extending the neck, as I tyrostris are subspecies of D. picumnus saw in one bird that I alarmed (it sleeked its (including the yellow-billed and as yet unstud- crown after and before ruffling the crown) and ied D. p. pallescensand relatives of western another that had supplanted a White-chinned Brazil and Bolivia south to Argentina). None Woodcreeper, may have been aggressive of these ecologically and behaviorally similar behavior. Hoffmanns’ Woodcreeper sup- woodcreepers is likely to be more than mar- planted or displaced White-chinned, Plain- ginally sympatric, perhaps in adjacent terri- brown, Elegant (Xiphorhynchuselegans), and tories in differing habitats. The culmen and Barred woodcreepers but fought with or was wing of seven museum specimensof platyros- supplanted by Buff-throated Woodcreepers. tris from Maranhao are longer than normal for I studied Planalto Woodcreepersbriefly near that species,within the range of variation for Maraba (Pam), Bacabal and Colinas and Serra picumnus. The platyrostris that I observed at Negra (Maranhao), and at many localities in Maraba and in Maranhao had pale-tipped bills, southern Brazil west to central Mato Gross0 like picumnus rather than like dark-billed (180 km W Barra de Garcas). The average southern plutyrostris. It is possible that D. weight of these birds was 62.1 g in a sample picumnustransfasciatus, found from the Tapa- of 14 specimensfrom northern Argentina. Most joz River east to the Xingti River and close to songs and snarling notes were like those of Maraba, is more closely related to plutyrostris Black-banded Woodcreepers (Fig. 1). Regular than to D. p. picumnus north acrossthe Ama- singing by scattered birds at dawn and dusk zon River. revealed that more birds were present than I had suspected.Another song of pairs in Mar- DISCUSSION anhao and S%oPaulo, a seriesof “yourit” notes, While only one speciesof large woodcreeper reminded me of Barred Woodcreepers in Pan- follows army ants in most neotropical regions, ama. One Planalto Woodcreeper gave grunting three large speciesdo so over much of the cen- “i-i-i” notes as attacked by another, then tral Amazonian region. At Manaus, these BLACK-BANDED WOODCREEPERS 285 speciessubdivided their niche of “large ant- there. DouglasWechsler helped with field studies.A grant following woodcreeper” by dominance: Red- from the National Geographic Society allowed studies in Maranhao, and earlier studies in the Amazon were sup- billed chased Black-banded, which chased ported by the Frank M. Chapman Memorial Fund of the Barred woodcreepers. Black-banded Wood- American Museum of Natural History. Yoshika Oniki creepers foraged low over ants less than did helped in many ways. Red-billed Woodcreepers, but more than did Barred Woodcreepers.Causal relationshipsare LITERATURE CITED not certain, for the last speciesdiffered little HAVERSCHMIDT,F. 1977. Trois nouvelles esptces qui se in behavior between Manaus and Belem (Wil- nourissentde scorpions.Oiseaux 47:2 13-2 14. LAND, H. C. 1970. Birds of Guatemala. Livingston Pub- lis, unpubl.), even though the other two did lishing, Wynnewood, PA. not occur at Belem. ONIKI, Y., AND E. 0. WILLIS. 1972. Studiesof ant-follow- In behavior, D. picurnnusdiffers little from ing birds north of the eastern Amazon. Acta Ama- its closest relatives. It nevertheless conforms zonica 2: 127-l 5 1. to the general prediction (basedon ant-follow- SCLATER,P. L., AND 0. SALVIN. 1879. On the birds col- lected by the late Mr. T. K. Salmon in the state of ing antbirds) that, becausesocial groups attract Antioquia, United States of Colombia. Proc. Zool. attention, subordinate ant-following species Sot. Lond., p. 486-550. should be less social than dominant ones. WILLIS, E. 0. 1976. Effects of a cold wave on an Ama- Young D. picumnus stay with their parents zonian avifauna in the upper Paraguaydrainage, west- em Mato Grosso, with comments on oscine-subos- three to four months, longer than subordinate tine relationships.Acta Amazonica 6:379-394. D. certhia stay with parents but a shorter time WILLIS, E. 0. 1977. Lista preliminar das aves da parte than do domineering H. perrotii (Willis, noroeste e areas vizinhas da Reserva Ducke, Ama- unpubl.). zonas, Brasil. Rev. Bras. Biol. 37:585-601. WILLIS,E. O., AND Y. ONIKI. 1978. Birds and army ants. ACKNOWLEDGMENTS Annu. Rev. Ecol. Syst. 91243-263. I appreciatethe help of Grant GB-3292 1 from the National Rua 2 no. 2272, 13500 Rio Clara, SP. Brad. Received 5 ScienceFoundation for studiesat Manaus, and of the per- July 1980. Final acceptance29 April 1982. sonnelof the Instituto National de Pesquisasda Amazonia