The Condor98:169-719 0 The CooperOrnithological Society 1996

ECOLOGICAL RELATIONSHIPS OF TWO IN HISPANIOLA: EFFECTS OF HABITAT AND FLOCKING’

STEVEN C. LATTA~ AND JOSEPH M. WUNDEIUE, JR. International Institute of Tropical Forestry, U.S.D.A. Forest Service, P.O. Box 490, Palmer, Puerto Rico 00721

Abstract. We studied microhabitat use, foraging and social behavior of Broad-billed (Todus subulatus)and Narrow-billed (T. angustirostris)Todies in two areasof sympatry in the Cordillera Central of the Dominican Republic. Solitary Broad-billed and Narrow-billed Todies occupied distinct microhabitats in both shade coffee plantations and native pine forest while generally sharing similar foraging strategies.In both habitats, Broad-billed Todies foraged higher in the vegetation and occurred in more outer horizontal positions with lower foliage density than did their congener.Movement ratesand feedingrates differed significantlybetween the two ,with the Narrow-bill being the more active species. Changesin foragingbehavior by both speciesof todies were observedwhen they associated with mixed-speciesflocks in pine forest. We noted a decreasein some measuresof spatial overlap of todies in interspecificflocks, but other feeding behaviors tended to converge. Key words: foraging behavior;mixed-species flocks;Hispaniola; Todus; insectivores;hab- itat use.

Sinopsis. Estudiamos el uso de1microhabitat, comportamiento alimenticio y social de Todussubulatus y T. angustirostrisen dos areas simpatricasde la Cordillera Central de la Republica Dominicana. Aves solitarias de ambas especiesocupan distintos microhabitates en plantacionesde cafe de sombra y en 10sbosques nativos, mientras comparten estrategias de forrajeo generalmentesimilares. En ambos habitates T. subulatusse alimenta m&s alto en la vegetation mientras se encuentraen posicioneshorizontales m&s hacia afuera, con una densidad de follaje menor que su congenere.Las razones de movimiento y alimentaci6n difieren grandementeentre las dos especies,siendo el T. angustirostrisla especiem&s activa. Se observaron10s cambios en el comportamiento de alimentaci6n cuando estasespecies se asociabanen bandadas de especiesmixtas en 10sbosques de pinos. Notamos una merma en el solapamiento espacial en las bandadas interespecificas,pero otros comportamientos de alimentacmn mostraron una tendencia a converger.

INTRODUCTION solitary individuals with individuals in mixed- Patterns of resourceuse among avian sympatric species flocks (Morse 1970, Austin and Smith congenersmay changein the presenceof mixed- 1972, Alatalo 1981), and color-banded species foraging flocks. Plocking behavior, hy- which forage solitarily and as group members pothesized to increase fitness through enhanced (Valburg 1992). Plocking may result in the con- foraging or decreasedlikelihood of predation (re- vergence of foraging behavior of flock partici- viewed by Powell 1985, Hutto 1994) may fa- pants (Pearson 197 1, Buskirk 1972 cited in Pow- cilitate the exploitation of foraging locations or ell 1985, Valburg 1992) and a reduction in re- tacticsthat a would not otherwiseuse (Krebs source partitioning (Morse 1970, Powell 1985), 1973,Buskirk 1976,Valburg 1992),ormayforce indicating possible social learning and copying a bird to modify foraging behavior because of (Krebs 1973, Morse 1978). Decreased foraging potential competition with so many birds in close overlap in flocks is generally interpreted as in- proximity (Hutto 1988). Changesin foraging be- dicating interference competition (Alatalo 198 1). havior have been suggestedin studiescomparing Studies of how resource partitioning is affected by the interaction of species,social organization, and habitat have not been published for the trop- ics (but see With and Morrison 1990). I Received11 March 1996. Accepted17 July 1996. * Present Address: Division of Biological Sciences, Hispaniola supportstwo closely related species 110 Tucker Hall, University of Missouri, Columbia, of todies and provides an ideal system for study- MO 65211, e-mail: [email protected] ing resourcepartitioning and the effectsof habitat

[7691 170 STEVEN C. LATTA AND JOSEPH M. WUNDERLE, JR. and social organization on foraging behavior of ious citrus species(Citrus spp.), and banana or congeners. Broad-billed (Todussubulatus) and plantains (Muss spp.) were also scattered Narrow-billed (T. angustirostris)Todies, thought throughout some plantations where they provid- to have evolved in isolation when Hispaniola ed an intermediate layer above the coffee. In a was separatedinto two islands during Pleistocene few plantations, an occasionalpine (Pinussp.) or glacier melts (Kepler 1977), continue to be geo- palm (Roystoneasp.) extended into the oversto- graphically isolated over much of the island (Ke- ry. The predominant variety of coffee (Coffea pler 1977, Dod 1981). The Broad-billed arabica)in the plantations was the traditional occursfrom sealevel to 1700 m and prefersxeric “tipica” variety, although “caturra” predomi- scrub habitats and broadleaf and pine forests. nated in some of the larger plantations. Most of The Narrow-billed Tody is found at higher ele- the shadecoffee plantations were relatively small vations (800-3,200 m) and most often occursin (X = 1.9 ha, range = 0.1-9.7 ha). moist forests and ravines. The two speciesare Coffee is cultivated in areas that were origi- sympatric, however, in mangroves and wet forest nally pine (P. occidentalis)forest. Pine forests, on the Samana Peninsula, over a broad range many of which have been selectively logged, re- (465-1,730 m) in the Sierra de Baoruco, and in main on the steeper slopes and ridge-tops, with portions of the Cordillera Central (Kepler 1977; variable amounts of broadleaf understory that pers. observ.). Both speciesmost frequently oc- may be affected by fire, cutting, or grazing. cur as single birds or in pairs throughout their Broadleaf specieswere not sampled but included ranges, but join mixed-species flocks in the pine Ribessp., I. Vera,mango (A4. indica),citrus (Cit- forestsof the Cordillera Central (Latta and Wun- russp.), Cecropiasp., Syzygiumjumbos, Chia derle 1996). sp., Miconia sp., as well as numerous unidenti- In this study we compared the foraging be- fied shrubs. Pine forest was sampled at six sites havior of Broad-billed and Narrow-billed Todies on slopesnear Manabao (90%1,050 m) and six in two areasof sympatry in the Cordillera Central sites near Jarabacoa(643-779 m). Although the of the Dominican Republic to determine if re- area of the pine forest was not measured, the sourceuse and foraging behavior varied by spe- remnant patches were all estimated to be > 100 cies, habitat, or social organization of the birds. ha. We compared foragingbehavior between the two Foliage height profileswere determined for each species in shade coffee plantations and native habitat where feeding observations were collect- pine forest, and within specieswhen foraging sol- ed (Wunderle and Latta, in press).Foliage height itarily and in mixed-speciesflocks in pine forests. profiles for the two habitat types showedgeneral We hypothesized that flocking allows similar structural similarity with a fairly open canopy, a speciesto copy foraging locations and behaviors, densercoffee or mixed broadleaf understory, and and that todies gain foraging benefits from such very little in the way of an intermediate layer. In behavior. We predicted that (1) when foraging the shadecoffee plantations, Zngaprovided most solitarily in similar habitat each specieswould of the shadeoverstory, with greatestcover in the occupy a distinct microhabitat and interspecific lo-12 m height category. Pine foliage predom- behavioral differences would be consistent be- inated in the canopy of the pine forest, although tween habitats, (2) foraging behavior and micro- scatteredbroadleaf trees also extended into the habitat use by the two specieswould converge canopy, and pine sites had a slightly higher can- in mixed-species flocks, and (3) foraging rates opy with greatest cover in the 15-20 m height would increase in the presenceof mixed-species category. The broadleaf understory of the pine flocks. forest provided a sparserfoliage cover within the different height categoriesthan that provided by STUDY AREA AND METHODS coffee in the shade coffee plantations, except Todies were studied in fourteen shade coffee within the first 1.0 m where foliage cover was plantations in the vicinity of Manabao and Jar- greater in the pine sites. abacoa, La Vega Province, Dominican Republic at elevations of 540-850 m. Shade coffee plan- FEEDING BEHAVIOR tations were characterized by a predominant Foraging observations were made in the non- overstory of ZngaVera, although mango (Man- breedingseason from October-March 1993-l 995 giferaindica), avocado (Perseaamericana), var- while walking slowly through the sites. Obser- ECOLOGICAL RELATIONSHIPS OF TODIES 171 vations were made from 07:00-12:00 in coffee Becausecanopy heights varied somewhat be- plantations, but extended to 18:OOin pine forests tween sites,bird height relative to canopy height where feeding activity occurredall day. Foraging was calculated and used in further analyses. In observations were made by walking slowly all statistical analysesfoliage density scoresof O- through the habitat until a foraging bird was lo- 2 and 4-5 were combined to reduce the number cated. The first foraging event five secondsafter of sparsecells. Similarly, for analyses of feeding an individual was initially detectedwas recorded behavior in pine forestsand analysescomparing to avoid a bias toward the more conspicuous behavior in pine to behavior in coffee, sample feeding techniques. Only a single foraging event sizes within a cell were increased with the fol- was recorded per individual per day to reduce lowing condensations:(1) foraging methods were the problem of autocorrelation inherent in se- combined into three categories:sally-strike, sal- quence data (Wagner 198 1). In all cases,the “so- ly-stall and sally-hover, and all others, (2) pri- cial organization” of the bird (i.e., whether the mary substrateswere reclassifiedas either broad- bird was solitary or a member of a mixed-species leaf or pine, (3) secondary substrates were re- flock) was noted. A mixed-species flock was de- classifiedas broadleaf-related (leaves or twigs) or fined as comprising at least two heterospecifics pine related (needles or twigs) and then elimi- within 25 m of one another and moving together nated as a habitat variable in pine forests as the for at least five minutes. test became identical to that of the primary sub- Terminology for foraging maneuvers and for- strate, and (4) horizontal positions 1 and 2 were aging site classificationswas derived from Rem- combined. sen and Robinson (1990). Aerial foraging ma- Movement rates (i.e., changes in perch posi- neuvers encountered included sally-strike, sally- tion) and feeding rates were determined by fol- stall, sally-hover, sally-pounce, and jump-up lowing feeding todies and recording the time of (leap). All perch maneuvers were combined into each feeding attempt into a cassette recorder. a single category (glean) and included reach-up, Birds were followed for up to 10 min or as long reach-out, probe, and glean. For each foraging as consecutive feeding attempts could be ob- maneuver in the coffeeplantations, the direction served. Rates were calculated as the averagetime of the feeding maneuver (upwards, downwards, between moves or feeding attempts, but no rates horizontal) was noted, and the distance moved were determined unless there were at least four by the bird (from the perch site to the food item) consecutive moves or feeding attempts. Rates was estimated by eye to the nearest 0.15 m. were log-transformed to achieve normality prior At each foraging site (i.e., the location of the to statistical testing. food item when attacked by the bird) the follow- Both Broad-billed and Narrow-billed Todies ing were noted: (1) estimated height of the bird are sexually monomorphic, so data from both above the ground, (2) estimated height of the sexeswere lumped. Lumping of the sexesmay tallest tree within 15 m (canopy height), (3) hor- be justified becausein the closely related Puerto izontal position of the bird (1 = inner l/3 of tree, Rican Tody (T. mexicanus), which are reliably 2 = middle l/3 of tree, 3 = outer l/3 of tree), sexed by eye color, Kepler (1977) did not find (4) the estimated amount of light passingthrough any difference in body size or morphology. an imaginary 2 m diameter sphere surrounding STATISTICAL ANALYSES the foragingsite (0 = 100% of light passesthrough, 1 = 95%99% of light passesthrough, 2 = 75% The software package SYSTAT Version 5.2 94% of light passesthrough, 3 = 25%74% of (Wilkinson 1992) was used to perform various light passesthrough, 4 = 5%24% of light passes statistical tests described in Sokal and Rohlf through, and 5 = O-4% of light passesthrough, (1981). A probability of Type I error of 0.05 or (5) the location of the prey item, or the “primary lesswas acceptedas significant but greatervalues substratetype” (for example a tree speciesor air), are shown for descriptive purposes.Where data and (6) the “secondary substrate” or the location were not normally distributed nonparametric of the food item within or on the primary sub- statisticswere used. A Pearson’s correlation ma- strate (for example leaf, twig, trunk, fruit, flower). trix was used to verify the independence of all In coffee plantations we also noted (7) substrate variables associated with foraging behavior in side (top or bottom) and (8) substrate condition coffee and pine habitats. (live or dead). Independent samples t-test (with pooled var- 772 STEVEN C. LATTA ANDJOSEPH M. WUNDERLE, JR.

0.80 n BROAD-BILLEDTODY iaNARROW-BILLED TODY 0.70

b 0.60 g # 0.50

g 2 0.40

E 5 0.30 z s 2 0.20

COFFEE-SOLITARY PINE-FLOCK FIGURE 1. Mean relative foragingheight (-t SE) of Broad-billedand Narrow-billedTodies in shadecoffee habitat,and as solitarybirds and flockparticipants in native pine forest.Relative foraging height is calculated by dividing absoluteforaging height by maximum canopyheight within 15 m of the foragingbird.

iances) was used to test for interspecific differ- native pine forestin the northern Cordillera Cen- ences in the means of attack distances, and to tral. Todies of both speciesappeared to be ter- test for interspecific and intraspecific differences ritorial in thesehabitats, as evidenced by singing in the means of foragingheights, movement rates, and territorial displays, and by recapturesof col- and feeding rates. A 2 x 2 Test of Independence or-banded birds within the non-breeding season with a x2 statistic or a Row x Column Test of and up to two or more years post-banding. In Independence with a G-statistic was used to test pine forests, where insectivorous mixed-species for the independence of various feeding behav- foraging flocks of migrant and resident birds are iors and feeding site characteristicsamong spe- common (Latta and Wunderle 1996) todiesjoin ciesand among social organizations within a spe- these flocks as the flocks move through tody ter- cies.In testswhere cellswere sparse(< 5) a G-test ritories (pers. observ.). Similar foraging flocks with Yates’ correction for continuity was used. were never observed in coffee plantations. Log-linear models were used to examine the de- greeof associationbetween species(Broad-billed FEEDING BEHAVIOR IN COFFEE or Narrow-billed), social organization (flocking Broad-billed and Narrow-billed Todies occupied or solitary), and each of four variables describing distinct microhabitats in shadecoffee plantations feeding behavior in pine forest habitat: foraging while sharing generally similar foraging strate- method, horizontal position, foliage density, and gies. Broad-billed Todies foraged higher in the substratetype. Where a three-factor interaction vegetation (t = 11.O, df = 242, P < 0.001; Fig. was not present, tests for conditional indepen- 1) and occurred in more outer horizontal posi- dence were also run. tions (G = 14.3, df = 2, P = 0.001; Fig. 2), with lower foliage density (G = 23.5, df = 2, P < RESULTS 0.001; Fig. 3) than did Narrow-billed Todies. Broad-billed and Narrow-billed Todies were Foraging substrates of the two speciesdiffered found to inhabit shade coffee plantations and (G = 74.3, df = 3, P < 0.001; Fig. 4), with the ECOLOGICAL RELATIONSHIPS OF TODIES 713

INNER THIRD MIDDLE THIRD OUTER THIRD •i !%I

NBTO

PINE-FLOCK

BBTO

NBTO

PINE-SOLITARY

NBTO

COFFEE-SOLmARY

BBTO

0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0 90.0 100 PERCENTAGEOF MANEUVERS FIGURE 2. Horizontalposition of foraging Broad-billed (BBTO) and Narrow-billed (NBTO) Todies in shade coffee plantations, and as solitary birds and flock participants in native pine forest.

Broad-bill most often (69.3% of feeding move- = 103, P < 0.001; Fig. 6) between the species, ments) found in the Zngu overstory, while the as did feeding rates (t = 4.9, df = 66, P -c 0.001; Narrow-bill fed primarily (61.2%) in the coffee. Fig. 6), with the Narrow-billed Tody being the There was no difference (x2 = 0.9, df = 1, P = more active species. 0.35), however, in the use of secondary sub- strates, with both speciesusing leaves far more FEEDING BEHAVIOR IN PINE frequently (94.0% and 89.4%, respectively) than Interspecific differences in feeding behaviors of branches or trunks as attack sites. The two spe- solitary birds observedin coffeeplantations were cies did not differ in their foraging maneuvers (G similar to those of solitary birds in pine forest = 7.1, df = 4, P = 0.13; Fig. 5), and there was habitat. Broad-billed Todies foraged higher in no interspecific difference in use of live as op- the vegetation (t = 5.7, df = 92, P < 0.001; Fig. posed to dead substrates(x2 = 0.0, df = 1, P = 1) and occurred in positions with lower foliage l.OO), or lower as opposed to upper substrate density (x2 = 7.3, df = 2, P = 0.025; Fig. 3) than surfaces(x*= lS,df= 1, P=0.22).Attackflight did Narrow-billed Todies, but there was no sig- directions, however, differed significantly (G = nificant difference in horizontal positions (x2= 6.9, df = 2, P = 0.03), with Broad-bills directing 0.9, df = 1, P = 0.35; Fig. 2) or use of foraging attacks nearly equally both horizontally (48.9%) substrates(x2 = 0.3, df = 1, P = 0.56; Fig. 4) by and upwards (42.2%), whereas most (73.9%) of the two species.Interspecific differences in for- the attacks by Narrow-bills were directed hori- aging maneuvers was found (G = 20.8, df = 3, zontally. Attack flight distances also differed (t P < 0.001; Fig. 5), with the Broad-bill using more = 3.4, df = 111, P = O.OOl), with mean attack sally-strikes and sally-hovers than the Narrow- distances of 0.86 m and 0.49 m, respectively. bill which used a higher proportion of near-perch Movement rates differed significantly (t = 7.7, df jumps when feeding. There was no interspecific 774 STEVEN C. LATTA AND JOSEPH M. WUNDERLE, JR.

HIGH DENSITY MODERATE DENSITY LOW DENSITY a cl ’ ’ (O%-24%) (25% - 74%) (75% - 100%)

NBTO

PINE-FLOCK

BBTO

NBTO

PINE-SOLITARY

BBTO

NBTO

COFFEE-SOLITARY

BBTO

i. I. I . 8. I I. I., I ., ., 0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0 90.0 100 PERCENTAGE OF MANEUVERS FIGURE 3. Foliage density surroundingforaging Broad-billed (BBTO) and Narrow-billed (NBTO) Todies in shade coffee plantations, and as solitary birds and flock participants in native pine forest. Foliage density was measured as the percentageof light passingthrough an imaginary sphere 2.0 m around the foraging site and ranged from O-l 00%.

difference in use of live as opposed to dead sub- respondingincreases in the movement and feed- strates (x2 = 0.4, df = 1, P = 0.52), or lower as ing rates of flocking Narrow-bills were recorded, opposedto upper substratesurfaces (x2 = 0.3, df but theseincreases were non-significant (t = 1S, = 1, P = 0.58). Movement (t = 7.8, df = 41, P df = 33, P = 0.14; t = 0.8, df = 29, P = 0.45), < 0.001; Fig. 6) and feeding rates (t = 4.3, df = respectively. 27, P < 0.001; Fig. 6) differed significantly be- Log-linear models used to analyze data on mi- tween the species,as they did in coffee habitat, crohabitat characteristicsand foraging behavior with the Narrow-bill being the more active tody. in pine forests showed no significant three-way interaction (G = 2.1, df = 1, P = 0.15) between EFFECTS OF SOCIAL ORGANIZATION ON species,social organization, and horizontal po- FORAGING BEHAVIOR sition of feeding todies, and testsfor conditional Flocking birds foraged significantly higher than independence also were non-significant. A sig- solitary individuals in the same pine habitat nificant interaction was found between species (Broad-billed Todies: t = -5.2, df = 66, P < and horizontal position (G = 16.8, df = 4, P = 0.001, Narrow-billed Todies: t = -5.6, df = 92, .002), indicating that speciesoccupy distinct hor- P < 0.001; Fig. 1). Intraspecific comparisonsalso izontal positions regardless of social organiza- showeda significant increase in the mean move- tion. Log-linear analysis found the presenceof a ment rate (t = -3.2, df = 3 1, P = 0.003), but three-way interaction (G = 5.4, df = 1, P = 0.02) not the mean feeding rate (t = - 1.6, df = 17, P between species, social organization, and sub- = 0.12) of flocking vs. solitary Broad-bills. Cor- strate use. This suggeststhat the degree of as- ECOLOGICAL RELATIONSHIPS OF TODIES 715

100 n COFFEE q INGA q PINE q OTHERBROADLEAF 0 AIR

2g 70.0

5 60.0

E 0b 50.0

z &t 40.0

z 2 30.0

z 20.0

BBTO (n=179) NBTO (n=67) BBTO (n=49) NBTO (1145) BBTO (n=19) NBTO (n=49)

COFFEE-SOLITARY PINE-SOLITARY PINE-FLOCK

FIGURE 4. Substrateuse by foragingBroad-billed (BBTO) and Narrow-billed(NBTO) Todiesin shadecoffee plantations,and as solitarybirds and flockparticipants in native pine forest. sociation between species and social organiza- terized by short sallies to the undersides of live, tion differed for different substrates.A test for a broadleaf leaves. The two speciesare ecologically three-way interaction between species,social or- segregated,however, because of the use of dif- ganization and foliage density was not significant ferent foraging strata. Narrow-billed Todies con- (G = 3.7, df = 2, P = 0.16), but tests for con- sistently foragedlower than their congener,using ditional independence found significant inter- denser vegetation both in shade coffee planta- actions between social organization and foliage tions and in pine forests.Different foraging strata density for each species(G = 13.3, df = 4, P = dictated the use of different substratesin coffee 0.0 l), and speciesand foliage density for a given plantations where Narrow-billed Todies were social organization (G = 17.9, df = 4, P = 0.001). found virtually always in the coffee understory, A final log-linear analysis showeda strong three- whereasBroad-billed Todies, foraginghigher, did way interaction between species,social organi- not use coffee (which generally reachesno more zation and foraging method (G = 10.2, df = 2, than 3-4 m in height), and were more often seen P = 0.006), again indicating the degree of asso- in the Zngu overstory. This was not the case in ciation between species and foraging methods pine forest where the deciduous understory is differed with social organization. taller and where some deciduoustrees even reach into the canopy. DISCUSSION Choice of microhabitat may have an effect on RESOURCE PARTITIONING AMONG some foraging methods. Robinson and Holmes SOLITARY BIRDS (1982) suggestedthat vegetation structure may Broad-billed and Narrow-billed Todies appear limit foraging maneuvers by affecting how a bird to forage in a generally similar manner charac- moves through its environment, how the vege- 776 STEVEN C. LATTA AND JOSEPH M. WUNDERLE, JR.

70.0-r

n SALLY-STRIKE &j SALLY-STALL qSALLY-HOVER qJIJMP-UP 0 GLEiAN

60.0

z 50.0

g

$ 40.0 E 8

g 30.0I- s z Y cr: 20.0,- p

10.0I-

0.0I-l- BBTO (n=179) NBTO (n=67) BBTO (w49) NBTO (1145) BBTO (1149) NBTO (n-19)

COFFEE-SOLITARY PINE-SOLITARY PINE-FLOCK FIGURE 5. Foragingmaneuvers utilized by foragingBroad-billed (BBTO) and Narrow-billed(NATO) Todies in shadecoffee plantations, and as solitarybirds and flockparticipants in native pine forest. tation can be searchedand at what rate, and how prey that did not appear to be taken by the small- easily prey can be captured. In this study, coffee- er Narrow-billed Tody. Different food types also dwelling Narrow-billed Todies exhibited shorter, may be reflectedin the variation between species and more often horizontal, attack flights than in bill size (Greenberg 1985, Moermond and Broad-billed Todies-foraging behavior that may Denslow 1985, Sherry 1990). be expectedin microhabitats with denser foliage. Similarly, in pine forest, Narrow-bills used a EFFECTSOF SOCIAL ORGANIZATION ON higher proportion of near-perch jumps than the FORAGING BEHAVIOR longer sallies of the Broad-bill. Both Broad-billed and Narrow-billed Todies have Differences in foraging behavior and differ- a high propensity to join mixed-species foraging ential use of substratesmay be related to the use flocks in pine forest in the Cordillera Central. In of different food types. Hespenheide (1973) a related study (Latta and Wunderle 1996) 7 1% showedthat larger birds can be expectedto select of encounterswith Narrow-billed Todies and 6 1% larger prey, and this may be assumed to be the of encounterswith Broad-billed Todies were with case for the todies. Food items were rarely seen individuals foraging in flocks. Both speciesav- in this study, but Kepler (1977) reported that the eraged 1.3 birds /flock when the specieswas pres- feeds primarily on dipterans ent. Mixed-speciesflocks in this habitat averaged and coleopterans,and may take nearly any small 7.1 + 0.2 different speciesand 11.3 f 0.5 in- insect. Insect surveys in these same habitats dividuals. Seventy percent of birds encountered (Wunderle and Latta, in press) show a similar in mixed-species flocks were permanent resi- range of prey available. Broad-billed Todies were dents and 64% of all flocking individuals were seen on occasionto prey on larger orthopterans, insectivores.Although most individuals were not lepidopterans, and lepidopteran caterpillars- color-banded, our impression was that Broad- ECOLOGICAL RELATIONSHIPS OF TODIES 717

SO.0

n MOVEMENT q FEEDING 70.0

E .P fg 60.0 2 P -g 50.0 0 B ii 40.0 f s a 4 30.0 8 P 2; 20.0

ii

10.0

0.0 BBTO NBTO BBTO NBTO BBTO NBTO

COFFEE-SOLITARY PINE-SOLITARY PINE-FLOCK FIGURE 6. Mean movement and feeding rates (k SE) by foraging Broad-billed (BBTO) and Narrow-billed (NBTO) Todies in shade coffee plantations, and as solitary birds and flock participants in native pine forest. Rates are calculatedas the mean time between consecutiveactivities.

billed and Narrow-billed Todies often occurred trast, the use of horizontal positions by the two in pairs and were among the few flock-joining species in flocks tended to converge. Narrow- speciesto maintain smaller individual territo- billed Todies continued to occupy less exposed ries-joining the flock only when the flock passed horizontal positions (inner) relative to the Broad- through their territory. bill, but showed a trend towards the use of the Most measures of the impact of interspecific more exposed(outer) positions. Broad-billed To- flocking behavior on resourcepartitioning indi- dies showed an opposite trend, though this may cate that flocking results in a reduction or con- reflect movement from the outer canopy of the vergence in foraging niche diversity (Powell broadleaf understory to feeding locations in the 1985). Changesin spatial parameters-especially lower reachesof pine trees. decreasesin foragingheight diversity- have been Log-linear analysis suggestedthe presence of most frequently reported when birds join mixed- changesin substrate use and density of foliage speciesflocks (Pearson 197 1, Buskirk 1972 cited used by flocking todies, which are parameters in Powell 1985). In this study, however, we did that may have been related to thesespatial shifts. not observe a reduction in all measuresof spatial A strong interaction between species,social or- overlap among tody species in mixed-species ganization, and substrateuse was probably a re- flocks. Both Broad-billed Todies and Narrow- sult of the upward vertical shift of the flocking billed Todies foraged significantly higher in the birds in the vegetation, and the corresponding vegetation when participating in mixed-species increaseduse of substrate-typesby Narrow-billed flocks, but the height at which one tody species Todies that were previously used primarily by foraged relative to the other increased. In con- solitary Broad-billed Todies. A similar reasoning 778 STEVEN C. LATTA AND JOSEPH M. WUNDERLE, JR. likely explains the result of the log-linear analysis todies to forage in more exposedsites and utilize of foliage density. In the presenceof mixed-spe- riskier behavior-foraging faster with less vigi- cies flocks both speciesmoved higher in the veg- lance. Caution must be taken in interpreting these etation. Narrow-billed Todies generally re- resultsin terms of flocking advantages,however, mained in the broadleaf understory, but in more as environmental conditions were uncontrolled exposed (outer) horizontal positions where fo- and may have differed independently of flocking liage density scoreswere lower. The Broad-billed behavior. Tody, in moving upwards, increased its use of In conclusion, we found that the coexistence pine (particularly the inner positions), where fo- of two speciesof todies in shade coffeeand pine liage density scoresdecreased only marginally. habitats was facilitated by the use of distinct for- Several studieshave reported changesin avian aging strata, the use of different foraging sub- feeding behavior resulting from association with strates (in coffee plantations), and different for- mixed-speciesflocks (Morse 1970, Valburg 1992), aging maneuvers (in pine forest). Behavioral dif- including the convergenceof the foraging behav- ferencesmay have reflected constraints imposed ior of individuals within the flock (Turner 1965, by microhabitat differences within these strata. Buskirk 1976, Morse 1978, Herrerra 1979, Val- The occurrence of todies in mixed-species for- burg 1992), but only Buskirk (1972, cited in Pow- aging flocks resulted in the convergencein mea- ell 1985) has recorded convergence of foraging sures of some spatial parameters and in feeding maneuvers by insectivores. The data presented behavior, indicating potential shifts in resource here suggestthat todies in mixed-species flocks utilization, but the use of distinct strata by the may not only adjust some spatial parameters,but specieswas maintained. Further studiesthat elu- some foraging maneuvers may converge as well. cidatebehavioral differencesbetween flocking and The observedthree-way interaction between spe- non-flocking individuals, and the costsand ben- cies, level of social organization, and foraging efits of flock association, are warranted. method reflectsthe presenceof interspecific dif- ferencesin foraging maneuvers in pine habitat, ACKNOWLEDGMENTS with Narrow-billed Todies using more near-perch We gratefullyacknowledge the fine assistanceprovided maneuvers than Broad-billed Todies, but these by Teodoro Lam, Esteban Terranova, and Eduardo differencesdecrease among flocking birds when Vazquez. We thank the numerous coffee plantation ownersfor permissionto work in their plantations.The both speciesmore commonly utilize a sally-hov- manuscriptbenefited from the constructivecomments er or sally-stall technique. of J. Faaborg, R. Hutto, J. M. Meyers, and two anon- This trend towards convergence of some spa- ymous reviewers. Funding was provided by the Na- tial parameters and foraging behaviors by two tional Fish and Wildlife Foundation and the John T. and Catherine C. MacArthur Foundation. speciesof todies in the presenceof mixed-species flocks suggeststhat todies find enhanced foraging LITERATURE CITED opportunities amidst flocks. Movement rates and feeding rates did increase for both Broad-billed ALAT~, R. V. 1981. Interspecific competition in tits Pam spp.and the GoldcrestRegulus regulus: and Narrow-billed Todies in flocks in this study, foraging shifts in multispecific flocks. Oikos 37: but the increase was significant only for move- 335-344. ment rates of Broad-billed Todies. Convergence AUSTIN, G. T., AND E. L. SMITH. 1972. Winter for- of foraging behavior as well as rate increasesmay aging ecology of mixed insectivorous bird flocks in oak woodland in southernArizona. Condor 74: indicate that the birds benefit from the flushing 17-24. of insectsby the activity of other flock members, BUSKIRK,W. H. 1972. Ecology of bird flocks in a supporting the food facilitation hypothesisof in- tropical forest. Ph.D. diss., Univ. of California, terspecificflock formation (Powell 1985). But to- Davis. dies may also benefit from predator avoidance. BUSKIRK,W. H. 1976. Social systemsin a tropical forest avifauna. Am. Nat. 110:293-310. Sharp-shinned Hawks (Accipiter striatus), which DOD, A. S. 1981. Guia de campo para las aves de la occur in the pine and shade coffee habitats, reg- Republica Dominicana. Editora Horizontes, San- ularly prey on small birds. In the more sparsely to Domingo, Dominican Republic. foliated pine forest tody foraging behavior may GREENBERG,R. 1985. A comparison of foliage dis- crimination learningin a specialistand a generalist be constrained by this predator. Protection af- speciesof migrant warbler (Aves: Parulidae). Can. forded todies by way of mutual vigilance or the J. Zool. 63~773-776. confusion of large flocks(Powell 1985) may allow HERRERA, C. M. 1979. Ecologicalaspects of hetero- ECOLOGICAL RELATIONSHIPS OF TODIES 179

specific flock formation in a Mediterranean pas- POWELL,G. V. N. 1985. Sociobiologyand the adap- serine bird community. Oikos 3385-96. tive significanceof interspecificforaging flocks in HESPENHE~E,H. A. 1973. Ecologicalinferences from the Neotropics. Omithol: Monogr. 36:713-732. morphologicaldata. Annu. Rev. Ecol. Syst.4:2 13- REMSEN. J. V.. AND S. K. ROBINSON.1990. A clas- 229. sification schemefor foragingbehavior of birds in Hurro, R. L. 1988. Foraging behavior patterns sug- terrestrialhabitats. Stud. Avian Biol. 13:144-l 60. gest a possible cost associatedwith participation ROBINSON.S. K.. ANDR. T. HOLMES. 1982. Foraaine in mixed-speciesbird flocks. Oikos 5 1:79-83. behavior of forest birds: the relationshipsamong Hurro, R. L. 1994. The composition and social or- searchtactics, diet, and habitat structure.Ecology ganization of mixed-species flocks in a tropical 63:1918-1931. deciduousforest in western Mexico. Condor 96: SHERRY,T. W. 1990. When are birds dietarily spe- 105-l 18. cialized? Distinguishing ecological from evolu- KEPLER,A. K. 1977. Comparative study of todies tionary approaches.Stud. Avian Biol. 13:337-352. (Todidae)with emphasison the PuertoRican Tody, SOKAL,R. R., AND F. J. ROHLF. 1981. Biometry. W. Todusmexicanus. Publ. Nuttall Omithol. Club 16. H. Freeman, San Francisco. KREBS,J. R. 1973. Sociallearning and the significance TURNER.E. R. A. 1965. Social feedine in birds. Be- of mixed-speciesflocks of chickadees(Parus spp.). haviour 24: l-46. Can. J. Zool. 51:1275-1288. VALBURG,L. K. 1992. Flocking and fiugivory: the LATTA,S. C., AND J. M. WUNDERLE,JR. 1996. The effect of social groupingson-resource use in the composition and foraging ecology of mixed-spe- Common Bush-Tanager.Condor 94:358-363. cies flocks in pine forests of Hispaniola. Condor WAGNER,J. L. 1981. Visibility and bias in avian 98~595-607. foraging data. Condor 83~263-264. MOERMOND,T. C., ANDJ. DENSLOW.1985. Neotrop- Wnxr~so~, L. 1992. SYSTAT: the system for sta- ical avian fiugivores: patterns of behavior, mor- tistics. SYSTAT Inc., Evanston, IL. phologyand nutrition,with consequencesfor fruit WITH, K. A., AND M. L. MORRISON. 1990. Flock selection.Omithol. Monoar. 36:865-897. formation of two Parids in relation to cyclical seed MORSE,D. H. 1970. Ecologicaiaspectsof somemixed- production in a pinyon-juniper woodland. Auk speciesforaging flocks of birds. Ecol. Monogr. 40: 107:522-532. 119-168. WUNDERLE,J. M., JR., AND S. C. LATTA. In press. MORSE,D. H. 1978. Structureand foraging patterns Avian abundance in sun and shade coffee plan- of flocksof tits and associatedspecies in an English tations and remnant pine forest in the Cordillera woodland during winter. Ibis 120:298-3 12. Central, Dominican Republic. Omitologia Neo- PEARSON,D. L. 1971. Vertical stratification of birds tropical. in a tropical dry forest. Condor 73:46-55.