The Condor95:99&999 0 The Cooper Ornithological Society 1993

FRUGIVORY AND COEXISTENCE IN A RESIDENT AND A MIGRATORY ON THE YUCATAN PENINSULA ’

RUSSELL GREENBERG SmithsonianMigratory Center, National ZoologicalPark, Washington,DC 20008

DANIEL K. NIVEN Department of Ecology,Ethology and Evolution, Universityof Illinois, Champaigne,IL 61820

STEVEN HOPP AND CAROL BOONE Department of Psychology,Emory and Henry College,Emory, VA 24327

Abstract. The White-eyedVireo (Vireo griseus)and the MangroveVireo (V. pallens) are two of the most commonspecies of insectivorousbird on the YucatanPeninsula. Mangrove Vireo pairsmaintain year-longterritories primarily in scrub,whereas individual White-eyed defendterritories in a broad rangeof terrestrialhabitats. The two speciesshow a strongreciprocal distribution along a successionalgradient from regeneratingpasture and old field to maturedry forest.Within secondgrowth scrub,White-eyed Vireos are primarily associatedwith patchesof trees. Despite minor morphologicaldifferences typical of migrant- residentcomparisons, we found no significantdifferences in the ratesof different locomotory movement, in the relative proportion of attack type used, nor in foraging substrate.The major foragingdifferences were a lower reducedattack rate and greateruse of fruit (primarily Burserasimaruba) in the White-eyed Vireo. Playback experiments failed to draw significant interspecific responsesto song playback and elicited weak responsesto conspecific and heterospecificchatter calls. Simultaneouscall-in experimentsand opportunisticobservations failed to provide evidence of interspecificaggression. White-eyed Vireos, however, consis- tently chased Mangrove Vireos out of fruiting Bursera. We hypothesizethat the extensive useof Burserafruit allows White-eyed Vireos to over-winter in mature forest on the Yucatan Peninsula, a habitat that apparently cannot support breeding by vireos in the White-eyed Vireo complex. Key words: Vireo;frugivory; Neotropicalmigrant; foraging; aggression.

INTRODUCTION portant role in the distribution of overwintering Competition forms the backbone of most theo- migrants. ries regarding the winter distribution of migra- The Mangrove Vireo (Vireopallens) and White- tory . Competition hypotheseshave focused eyed Vireo (Vireo griseus) are closely-related on interactions between closely-related migra- (Hellmayr 1935, Paynter 1955, Hamilton 1958, tory species(Salamonson 1957, MacArthur 1958, Blake 1968, Parkes 1990). A few authors have Cox 1968, Chipley 1980) or between migratory considered them conspecificbecause of their al- and resident components of a local avifauna lopatric breeding distribution and morphological (Willis 1966, Leek 1972, Waide 1980, Ricklefs similarity (Hellmayr 1935, Paynter 1955, Ham- 1992) with little attention given to ecological in- ilton 1958, Smithe and Paynter 1963). However, teractions between closely-related sympatric res- most workers familiar with both speciesin the ident and migratory species(or populations with- field considerthem to be very closely-related spe- in a species,but seeBarlow 1980). However, the cies, part of a larger complex including several period during or immediately following specia- West Indian species(Chapman 1896, Miller et tion is when selection for separation by macro- al. 1957, Hamilton 1962, Russell 1964, Monroe habitat or geographymay be most intense (Lack 1968). 1944, Cox 1968). Therefore, potential interac- Both speciesare among the most common in tions with congenersmay have a particularly im- terrestrial habitats of the Yucatan Peninsula (Paynter 1955, Lynch 1989). Mangrove Vireos occur throughout the year in early stagesof sec- I Received4 September1992. Accepted 19 July 1993. ond growth and White-eyed Vireos can be found ECOLOGICAL RELATIONSHIPS IN VIREOS 991 throughout the winter (October-mid April) in a primary no. 1. Body mass was obtained from variety of habitats. specimensnetted in northern and central Quin- This paper summarizes the habitat distribu- tana Roo. We found no significant difference in tion, ecologicalmorphology and foraging ecology fat scoresfor the two species,so the massesare of the two speciesof vireos. We also report ex- directly comparable. periments testing for responseto playback song and chatter (both White-eyed Vireo and Man- HABITAT DISTRIBUTION grove Vireo are territorial and sing in the non- We established six one-km census transects as breeding season) and to the close approach of part of a broader study of the distribution ofbirds heterospecificterritory holders. in the Sian Ka’an BiosphereReserve and Vicinity (Greenberg 1992). The transects were estab- STUDY SITE lished in a partially active and partially aban- The researchwas conducted as part of a project doned farm and included: a one year-old cattle on the ecology of migrants in and around the pasture (cleared, but with no planted grass);an Sian Ka’an Biosphere Reserve, on the road to area of 5-7 year-old grazed second-growthhab- Vigia Chico, northeast of Felipe Carrillo Puerto, itat (grazed acahual); an area of 5-7 year-old un- . The vegetation is a mosaic of grazed second-growth habitat (ungrazed aca- mature and regenerating tropical dry forest (ap- hual); a semi-deciduous (short) forest (canopy ht. proximately 1.2 m rainfall/year), which has two = 10.2 * 0.2 m SE); and a medium stature sub- distinct phases:a low canopy subdeciduousfor- perennial (tall) forest (canopy ht. = 13.2 + 0.3 est and a medium canopy subperennial forest m). Transects were censusedweekly between 1 (Olmsted et al. 1983). The forest types differ in October 1987 and 30 March 1988, and 25 Oc- overall stature (Greenberg 1992) as well as de- tober 1988 and 30 March 1990. We conducted gree of leaf loss during the long dry season(Jan- censusesby walking the transect and counting uary-June). Although forest covers most of the all birds heard or seen within 20 m of the tran- area, there are small areas of recently cleared sect. We recorded the location of each vireo to cattle pasture and milpa, and larger areas of re- the nearest 50 x 40 m transect unit. Repeated generating cattle pasture and burns, mostly 5-7 censuseswere used so that a profile of the dis- years of age at the time of the study. tribution of vireos could be compared to the pat- tern of vegetation structure. The distribution along the ungrazed acahual transect, where both METHODS specieswere common, was compared with val- ECOLOGICAL MORPHOLOGY ues for several vegetation variables measured for We compared the external morphology of the each 5 x 40 m unit. Measurements were taken two speciesby measuring study skins in the U.S. for 10 evenly spacedcircular plots (2 m radius) National Museum of Natural History collection. per unit and include the number of stemsof vary- To reduce possible confounding effectsof sexual ing dbh (O-l.0 cm, 1.1-2.0 cm, 2.1-5.0 cm, 5.1- dimorphism, we compared measurements of 10.0 cm, 10.1-t cm dbh), maximum plant height, males. We measured 21 specimens each from estimated ground cover (to nearest 10%) and to- Mangrove Vireos from the Yucatan Peninsula tal number of trees greater than 5 m ht./transect and Peten and White-eyed Vireos from scattered unit. localities in the eastern United States. Because we found no statistically significant difference in FORAGING DATA the measurements of Vireo griseus griseus and We gathered foraging data by observing in ap- V. g. noveboracensis,we pooled White-eyed Vir- propriate habitat and recording the location, lo- eo measurements to compare to Mangrove Vir- comotory movements, and foraging attack meth- eos. One of us (DN) measured the following to ods of vireos. Individual vireos were followed the nearest0.1 mm using dial calipers:bill length, for variable periods and only sequencesgreater width and depth (at anterior edge of nares); un- than 30 set were used for analysis. Average se- flattened wing cord; tarsus; and tail length. In quenceslasted 77 set for White-eyed Vireos and addition, we obtained an index of wing point- 66 set for Mangrove Vireos. Additional data on edness (Gaston 1974) based on the difference foraging attack methods were gathered on 150 between the length of the longest primary and additional White-eyed and 45 additional Man- 992 RUSSELL GREENBERG ET AL. grove Vireos. We spent a disproportionate censusing during the prior year, this represents amount of time at ecotones and intermediate the peak period of arrival of White-eyed Vireos habitats so that individuals of either speciescould at Sian Ka’an. Trials were conducted by system- be followed. Becauseof this, foraging variables atically selecting points within the most appro- that reflect habitat use (tree height, plant type) priate habitat for each speciesbased on two years can be used to compare the relative habitat use of censuswork. Playback points were established by species,but not to generate estimates of ab- along transects at approximately 200 m. If we solute habitat use. The location variables, re- heard a member of the target speciesbefore we cordedonce per individual, include height of bird, began a trial, the playback point was advanced height of plant, leaf size of plant, and perchbranch an additional 100 m. angle and diameter. Locomotory variables in- A Sony TCM 5000 EV recorder was placed on clude the number and estimated length of hops the ground with the speaker directed upward. (branches changed without the use of wings), Observers (SH and CB) positioned themselves 5 creeps (branch not changed) and flights (branch m on either side of the recorder. Stimuli tapes changed, wings used). For a subset of the se- were constructed to include baseline, playback, quences, we recorded the vertical direction of and residual periods during which the observers flights (up, down, horizontal, approximately 45”). recorded the number of chatterbursts,the num- Foraging maneuvers were classified based on a ber of songs,approaches (bird observed or heard system modified from Remsen and Robinson moving in direction of speaker), approaches (1990). The following attack maneuvers were within 5 m of the speaker, latency to approach recognized: glean, where a perched bird plucks within 5 m of the speaker,and number of flights food item from substrate;reach, where bird cranes over the speaker. upward from a perch to grab prey item; leap, At each playback point, we presented either where a bird attacks upward propelled by legs or chatterburst or song playbacks. One of the three wings (includes upward strike and leap of the types of chatter were presented in a particular Remsen and Robinson system [ 19901); hover, trial: White-eyed Vireo, Mangrove Vireo, or a where bird strikes upward and remains station- non-vireo control consisting of the chatter of the ary with rapid wing beating; lunge, bird rushes Carolina Wren (Thryothorus ludovicianus) or horizontally or downward; and hang, bird grabs White-bellied Wren (Uropsila leucogastra).In the on side or undersurface of substratewhile glean- song trials, each trial presented songsfrom one ing or probing. Other tactics were rare. The tak- of the vireo species.Categories of both song and ing of large prey items or fruit were recorded as chatter were alternated between playback points well. to minimize the indirect arousal of neighboring individuals. Chatter and songtrials were not con- FRUGIVORY ducted in the same areas within 72 hr. Chatter To explore in detail interactions at fruiting trees, and song trials were mixed over an 1 l-day pe- we examined data gatheredduring a study of the riod. use of the fruit of Bursera simaruba by White- From bioacoustic libraries and private collec- eyed Vireos (Greenberg et al., in prep.). We also tions, we obtained recordings of eight different watched two fruiting trees in small patches of songs and two different chatters of Mangrove trees of the grazed acahual, recording all visits, Vireo recorded on the Yucatan Peninsula. We fruit consumed, and aggressiveinteractions. We selectedan equal number of different songsand watched each tree between 05:30-09:30 at least chatter segmentsfrom our recordings of White- five days (one in October and one in March). All eyed Vireos from southwestern Virginia. Re- local White-eyed Vireos and several Mangrove cordings of one chatter segment each were used Vireos were color-banded. In addition, we for the two wren species. watched a fruiting Bursera at the Uxmal ar- All recordingswere band-pass filtered at 500- cheaological site for six hours which also re- 800 kHz, digitized using a 16 kHz sampling rate, ceived visits by both White-eyed and Mangrove resealedso that the peak amplitudes in all stimuli vireos. were matched, reconverted through the D-to-A and input to a cassettedeck for reconstructing PLAY-BACK RESPONSE EXPERIMENTS playback tapes. We conducted playback trials 13-23 October For the song trials, we constructed tape seg- 1989 during the first 3 hr after sunrise. Basedon ments consisting of 2 min of silence, 30 repeti- ECOLOGICAL RELATIONSHIPS IN VIREOS 993 tions of the song in 6-set onset-onset intervals TABLE 1. Measurements of Mangrove Vireo and (3 min), then 2 min more of silence. At the be- White-eyed Vireo (mm, K + SE). ginning/end of each of the three segments, we inserted a 250 ms, 4 kHz tone to signal timing Characteristic MangroveVireo White-eyedVireo during field trials. Bill length 73.1 (1.23) 70.9 (0.06) To prepare the chatter stimulus tapes we ex- Bill width 36.4 (0.08) 36.1 (0.04) tracted a 3-set segment of chatter and used rep- Bill depth 40.0 (0.11) 40.1 (0.04) Tarsus 196.4 (0.21) 188.4 (O.ll)**a etitions of that segment for the entire presenta- Length of tail 436.3 (0.95) 499.2 (1.66)** tion. Each tape began with 2 min of silence, 30 Unflattened wing 537.2 (0.96) 603.4 (0.30)*** repetitions of chatter in 4-set onset-onset inter- Roundnessof wing 7 1.7 (0.48) 109.3 (0.26)*** vals (2 min), followed by a 2-min residual period. Weight (g) 9.2 (0.15) 11.0 (0.19)*** All other details of chatter tape presentation were dStgnificance of differencebased on student’sf testcorrected for mul- tiple comparisons** P < 0.005; *** P < 0.00 I. as described above for the song tape. We used two different 3-see segments from each of the original recordings, giving four chatter playback ments, we found that body masswas 15% smaller tapes for each of the Vireo speciesand four total in the Mangrove Vireo. Based on a 4.5% reduc- control tapes. tion in the cube root of body weight compared to White-eyed Vireo, the Mangrove Vireo has a SEQUENTIAL PLAY-BACK AT ECOTONES disproportionatelylong tarsus,short tail and short We presented 2-min playbacks of songsof each and more rounded wing. speciessequentially within forest and acahual 10 HABITAT DISTRIBUTION m from the same point along the ecotone. The experiment was conducted at a total of eight The two speciesshow a strongly complementary points along the ecotone, located at least 100 m distribution pattern along the successionalgra- apart. The order of habitat was alternated be- dient representedby the censustransects (Fig. 1). tween points, and the order of species presen- Mangrove Vireos are restricted to the pasture, tation was alternated between presentations for farm and scrub habitats with a few individuals a particular habitat. For up to 3 min after song in tintales (seasonallyflooded depressionswithin presentation, we recorded when a bird ap- the forest that support a lower stature vegetation proached within 5 m of the speaker or another type [Olmsted et al. 19831). White-eyed Vireos vireo. We also recorded all interactions between are less specialized, but largely restricted, none- vireos. The purpose of this experiment was two theless,to forest. A more fine-scaled analysis of fold: to determine how far into the lesspreferred the ungrazed acahual, the scrub transect with habitat individuals of each specieswill approach moderate numbers of White-eyed Vireos, also the simulated encroachment of a conspecific,and showsa negative correlation between the number to seehow vireos that are stimulated to approach of sightings of the two specieswithin the 40 x a potential encroaching conspecific, respond to 50 m transect units (r = -0.40, n = 267 sightings the presence of a responding heterospecific on in 20 units). White-eyed Vireo sightings were their territory. positively correlated with the total number of treesper transect unit (r = 0.50,7 1 sightings)and RESULTS plant stems greater than 5 cm (r = 0.45) showed no significant correlation with other vegetation ECOLOGICAL MORPHOLOGY variables (ground cover, mean canopy height). Although morphologically similar, White-eyed Stem density greater than 5 cm and total trees and Mangrove vireos differ significantly in sev- were the only two vegetation variables that were eral mensural characteristics(Table 1) resulting significantly correlated (r = 0.62). Mangrove Vir- in virtually no overall overlap (a DFA based on eos sightings showed a weak, but significantly specimen measurements correctly classified, a negative correlation with total trees/unit (r = posteriori, 98% of the specimens [Wilkinson -0.44) and a positive correlation with percent 19871). The tail, wing chord and wing pointed- ground cover (r = 0.42). ness of the Mangrove Vireo all measure lower than those of the White-eyed Vireo (tail 12%, WINTER SOCIAL ORGANIZATION wing cord 10% and wing pointedness 29%), but Based on the observation of 15 color-marked the tarsii are 4% longer. Based on field measure- birds (seven in acahual and eight in forest) we 994 RUSSELL GREENBERG ET AL.

pasture milpa grazed ungrazed short tall acahual forest

m white-eyed vireo

EB mangrove .rireo FIGURE 1. Mean and standard error of sightingsof White-eyed and Mangrove vireos per 1 km transects(n = 37 counts/transectover two years). found that White-eyed Vireos were consistently pairs throughout the winter and although we had territorial (average territory size was approxi- only a few color-banded, their behavior was con- mately 112ha). All birds were found to be resident sistent with year-long territories. Both species on territory for at least two weeks and seven of commonly sing during the winter months. the eight birds monitored through the winter re- FORAGING BEHAVIOR mained on territory. Trespassingto Burseru trees on neighboring territories was commonly ob- We found no significant difference in the loco- served when conducting focal watches of these motion of foraging vireos (Table 2). The two trees (Greenberg et al., ms.). These individuals species made an average of approximately 13 were almost invariably chased back into their moves per minute ofwhich a majority were hops. territories. Mangrove Vireos occur primarily in This is considerably lower than the 17 moves/

leaf branch other Substrates

m white-eyed vireo

m mangrove vireo FIGURE 2. Proportion of observedinsectivorous foraging attacks directed to different substrates(n = 109 for White-eyed Vireo and 207 for Mangrove Vireo). Within foliage insectivory both species showed a similar preference for leaf under surfaces(75.2 + 2.3% for White-eyed Vireo; 77.8 + 3.4% for Mangrove Vireo). ECOLOGICAL RELATIONSHIPS IN VIREOS 995

TABLE 2. Mean and standard error of major foraging variables recorded during focal watches.

Mangrove Vireo White-eyedVireo

Foraging rate” 1.54 + 0.21 1.01 + 0.14*b Creep rate 1.94 + 0.29 1.34 + 0.18 Hop rate 1.46 + 0.44 7.08 + 0.52 Flight rate 2.92 + 0.20 3.31 + 0.39 Total rate 13.87 + 0.58 12.79 + 0.67 Creep distance’ 13.44 + 2.22 11.02 + 1.80 Hop distance 92.43 + 8.35 81.26 + 12.24 Flight distance 320.71 + 41.09 306.57 + 32.52 Flight orientation (% up, level, down) 19:59:21 (708) 23:61:16 (813) Total distance 426.58 + 43.40 404.86 + 36.19 Feeding height (m) 2.91 + 0.17 (88)’ 5.20 + 0.31 (86)** Relative height (m) 0.65 + 0.02 (86) 0.67 + 0.05 (81) Leaf area (cm2) 25.50 + 3.74 (73) 32.07 + 3.89 (61) Branch diameter 8.7 + 0.31 (178) 9.2 + 0.38 (293) Branch angle 42.3” + 2.7 (173) 37.5 + 3.2 (293) 1All ratesdenote number of maneuvers/minute.Sample sizes are: Mangrove Vireo, n = 143; White-eyedVireo, n = 144. n Significanceof differencebetween species based on student’sf test correctedfor multiple compansons* P < 0.05, ** P < 0.001. LAll distancesdenote cm/minute. Samplesizes are: MangroveVireo, n = 104; White-eyedVireo, n = 102. clAll n values are in parentheses. min found by Williamson (197 1) for White-eyed for White-eyed Vireos of 5.1 m is probably a low Vireos during the breeding season.The estimat- estimate for the population as a whole. We gath- ed distance moved was similar (4.27 m/min vs. ered foraging data in areas where both species 4.06 m/min) for Mangrove and White-eyed vir- could be found (edgesbetween forest and second eos and there was no significant difference be- growth). The high density of White-eyed Vireos tween species in the distance moved by creep, in forest suggeststhat the actual mean may be hop or flight. In terms of distance moved, flights accounted for over 75% of the distance moved. There was a small, but significant difference in the vertical orientation of flights (Table 2). The only rate difference was found in the frequency of prey attacks which was 50% greater in Man- grove Vireo than in White-eyed Vireo (P < 0.05). The value of 1.0 prey attacks per min in the White-eyed Vireo is similar to the 1.2 attack/ min found by Williamson (197 1) on the breeding grounds. The type of foraging attack used and the substrate attacked was not significantly dif- ferent between the two species(Fig. 2 and 3). The similarity index (Schoener 1968) are 0.90 and 0.93 (index range = O-l) respectively. The two largestdifferences in foraging between the two specieswere found in location variables and probably reflect grosshabitat differencesin where individuals were located. White-eyed Vir- glean hover hang leap lunge reach other eosfed higher off the ground. The higher foraging height reflectsthe White-eyed Vireo’s preference Foraging Maneuvers for forest and patchesof trees in second growth. -white-eyed vireo The relative foraging height (bird height/tree BB mangrove vireo height) was nearly identical for the two species FIGURE 3. Proportion of foragingattacks using dif- (0.65 vs. 0.67, Table 2). The mean foragingheight ferent techniques(see Fig. 2 for sample sizes). 996 RUSSELL GREENBERG ET AL.

TABLE 3. Responsesof the White-eyed Vireo and the Mangrove Vireo to a playbackof stimulus species’ song.

Approach Increase/decrease” Stimulus species FlyOVer Target species % Trials Median #/trial % Trials Median latencv Call rate Sonnrate

White-eyed Vireo White-eyed 60 1.5 90 45 80/20 40/30 Mangrove 0 0 0 NAb 30/o 20/o Mangrove Vireo White-eyed 0 10 NA 30/10 10/20 Mangrove :: 1.5 70 40 70/o 70120 dFercent of trials wherethere was an increasedOT decreased rate of vocalizationduring presentation and postpresentation compared to presentation period. h NA = No Approachwithin 5 m. closer for data taken for White-eyed Vireos for- were far less common (5%) and slower (median aging along the transectsin the forest (mean for- = no approach) for heterospecificsongs. aging height = 10.1 + 0.3 mm SE, n = 140). Responsesto homo- versus heterospecificvir- eo calls were similar (Table 4). Flyovers occurred FRUGIVORY in 30% of both the homo- and heterospecificpre- The two speciesdiffered significantly in the per- sentations, close approachesoccurred in 50% of centage of birds observed eating fruit. White- both trial types, and latencies averaged approx- eyed Vireos ate fruit significantly more often than imately 200 set for both. None of these com- Mangrove Vireos (x2, P < 0.001). Twenty-five parisonsinvolved significantdifferences. The only percent of the White-eyed Vireos (n = 247) ver- significant difference was in the proportion of sussix percent of the Mangrove Vireos (n = 158) presentationsthat elicited increasedcalling in the were observedeating fruit. This differenceis based presentation and post-presentation parts of the primarily on the differential use of Burserasima- experiment. Vireos increased their calling sig- ruba, since 96% of all frugivorous observations nificantly (x2 P < 0.01) more in responseto ho- in vireos involved consuming capsules of this mospecificversus heterospecific calls. Vireos were species(Greenberg 1992). more responsive to vireo than to wren chatter. White-eyed Vireos visited the three focal Bur- They approachedvireo chatter twice as frequent- sera trees in scrub habitat on the average of 0.9 ly as wren chatter (50% versus 25%, x2 P < 0.01) vs. 0.3 times per hr (n = 47 hr) for Mangrove and also flew over significantly more often (30% Vireo. The successrate (arils consumed per visit) versus 10%). Vireos did respond to wren chatter was almost 10 times greater as well (1.17 * 0.2 1, by increasing their rate of chatter bursts during vs. 0.11 f 0.07, t = 4.6 df 59, 1. P < 0.001). 70% of the trials. Mangrove Vireos visits were less than half as long as those of White-eyed Vireos (43 ? 3.2 vs. SEQUENTIAL PLAY-BACK AT ECOTONES 102 f 4.7, t = 4.6, P < 0.001). It was during White-eyed Vireos approached within 5 m for observations of Bursera trees that we observed seven out of eight of the forest play-back trials. the only overt aggressionbetween the species; Similarly, Mangrove Vireos responded to six of Mangrove Vireos were chasedout of the trees by eight acahual trials. White-eyed Vireos also ap- the resident White-eyed Vireo during 45% (9/20) proached in seven of eight of the acahual trials, of the visits. whereasMangrove Vireos never entered the for- est to approach the speaker in the forest trials. PLAY-BACK RESPONSE During eight trials, Mangrove Vireos approached Both species responded more strongly to con- to the absolute edge of the forest and sang in specific than heterospecific song. Responses to responseto the conspecificplayback. The habitat conspecificsong were generally the most intense edgewas apparently an absolute territory bound- of all the presentations including frequent fly- ary to the Mangrove but not White-eyed Vireos. overs (median = 1.5/trial, 60% of trials Table 3) During six acahual playback trials, individuals and rapid approach to within 5 m (median la- of both speciesapproached the speakerand were tency = 40 set, 80% trial frequency). Approaches present within 5 m of the speaker(and each oth- ECOLOGICAL RELATIONSHIPS IN VIREOS 997

TABLE 4. Responsesof White-eyedand Mangrovevireos to playbackof stimulusspecies chatter’ call.a

FlyOVer Approach Increase/decrease* Stimulus species Target species % Trials Median #/trial % Trials Median latency Call rate Song rate White-eyedVireo White-eyed 30 0 50 230 50130 30/10 Mangrove 20 0 70 93 lO/lO 60150 Wren 10 - 30 NA 70/o 70/o MangroveVireo White-eyed 30 0 30 NA 40/20 50130 Mangrove 40 0 50 192 so/10 60/20 Wren 10 - 20 NA 70/10 40/10

A See notes for Table 3. er) at the same time. Although vireos remained proach). It is in the weaker responsesto call notes in the vicinity for several minutes actively flying that we failed to find consistent differences in back and forth and apparently searching for in- how vireos responded to homospecific and truders, we recorded no interactions between in- heterospecific calls (although they did respond dividuals. On two occasions, Mangrove and stronger to vireo than to wren calls). The si- White-eyed Vireos perched within 1 m of each multaneous playback experiments demonstrated other on the same branch. that vireos called into conspecificsongs, ignored nearby individuals of the other species.This lack DISCUSSION ofresponseis particularly telling considering that COMPLEMENTARY HABITAT the vireos are approaching the tape recording DISTRIBUTION with aggressivefly-overs and therefore might be Complementary habitat distribution is support- sensitized to attack a wider range of stimuli. ed by the transect census results. Additionally, these results show fine-tuned differences in dis- SIMILARITY OF MORPHOLOGY AND tribution in transitional habitat (shrubby field FORAGING BEHAVIOR with patchesof trees). The transectsprovide rel- White-eyed and Mangrove vireos are morpho- atively little replication for the censusing effort, logically and ecologically similar species with but the degree of complementarity is supported complementary habitat distribution. The longer by the response to play-back experiments. We and more pointed wings of the White-eyed Vireo observed no responsesby White-eyed Vireos to may be related to migratory status (Chapman White-eyed Vireo songsin scrub habitats and no 1940, Leisler and Winkler 1985) or its use of responseby Mangrove Vireos to Mangrove Vir- more open microhabitats for foraging (Poole eo song in the forest habitat. 1938, Morrison 1982). In most aspectsof for- Strongly complementary habitat distribution aging behavior and location, the two specieswere between ecologically similar speciescan result, indistinguishable. This includes locomotory be- in a proximate sense, from aggressiveinterac- havior, attack type and foraging substrate. The tions, including interspecific territoriality. Inter- degreeof similarity, particularly in the type and specific territoriality has also been reported for distance of movement, is surprising given the at least one pair of Vireo species (Rice 1978). differences in branch density and arrangement However, playback experiments did not provide and foliage structure in forest and secondgrowth any evidence for aggressiveinteractions between habitats. This result suggeststhat despite differ- thesetwo species.The strongestresponses in both encesin wing shape,the two speciesshare a par- species were to homospecific song. Vireos re- ticular locomotory stereotypy (Klopfer 1967). sponded to song with rapid approach to within However, reduced wing and tail size could allow 5 m of the speaker and low flyovers. These ap- Mangrove Vireos to perform similar movements proach behaviors probably reflect a high level of to White-eyed Vireos while minimizing feather potential aggression.Responses to call notes were wear. weaker (fewer flyovers, longer latency to ap- Use offruit. The major ecological difference 998 RUSSELL GREENBERG ET AL. between the two speciesis the use of fruit. White- eyed Vireo complex breeds in the Yucatan for- eyed Vireos feed heavily upon the fruit of Bur- ests. Perhaps vireos cannot compete for arthro- sera simaruba throughout the winter. A similar pod resources in mainland forests with their system has been described for the Gray Vireo diversity of foliage gleaning birds. Cox (1985) (Vireo vicinior) in western (Bates 1992). suggestedthat migrants can often occur in trop- In the forestsof Sian Ka’an, Bursera producesa ical areas under situations that resident popu- crop of up to a few thousand capsuleswhich ripen lations cannot, because migrants can use food slowly for a 7-8 month period primarily during sourcesthat are poor sourcesof protein for feed- the stay of White-eyed Vireos (September-early ing young. By this argument, the forests of the April). Individual White-eyed Vireos visit trees Yucatan provide appropriate food that can be with fruit on their territory daily. During these used by vireos for overwintering, but not for visits, they test the indehiscent capsules and breeding. Fruit is protein poor and therefore a swallow the seeds when the capsules fall away poor resourcefor raising young (Morton 1973). from the ripe arils. The importance of the tree Furthermore, the particular fruit involved, Bur- to White-eyed Vireos (Greenberg et al., ms.) is sera, is only available during the winter months. suggestedby the persistenceof vireos visiting the The use of fruit versus insects appears to in- trees,the frequency (25%) of vireos observedvis- volve trade-offs. We observed Mangrove Vireos iting trees during foraging observations, and the attacking arthropod prey at a considerably higher correlation between the occupancy of small for- rate than did White-eyed Vireos. Foliage arthro- est patchesin agricultural areasand the presence pods are more common in scrubby habitats than of a territorial White-eyed Vireo (size of patch in forestunderstory (Greenberg 1992). Small body and canopy height were not correlated). size, short tail, and short rounded wings might We observed Mangrove Vireos feeding on allow Mangrove Vireos to maneuver through Bursera capsules,however such events are rare dense foliage with minimum plumage wear. for two reasons. The abundance of Bursera and Larger body size and pointed wing in the White- per capita fruit crop is much lower in the scrub eyed Vireos may be primarily related to its mi- habitat. We found Bursera fruit to be almost an gratory status, but it could also favor its ability order of magnitude more dense in forest than to dominate a fruit crop by increasing the speed secondgrowth habitats (Greenberg et al., unpubl. and force of an aggressivechase. Thus, White- ms.). The secondprobable causeis the aggression eyed Vireos are able to control a critical non- the White-eyed Vireos’ display towards Man- breeding resource against its resident counter- grove Vireos at fruiting Bursera. Besides being part. chase out of the trees, Mangrove Vireos visited for short periods and were less successfulat lo- ACKNOWLEDGMENTS cating ripe fruit during visits than the local White- We thank the other members of the Sian Ka’an mi- eyed Vireo. The rate of chasing probably under- gratory field team Maro Berlanga,Dana Bradley, Da- estimates the importance of dominance, since vid and Jill Heath, John Sterlingand Rosa Maria Vidal. the short and unsuccessfulvisits also may result We thank Mercedes Foster for help and support from the overall dominance relationship. throughout the project. Jim Lynch, Arturo Lopez Or- The migrant status of the White-eyed Vireo nat and Eugene Morton introduced the senior author to the avifauna of Quintana Roo. The project was fund- raisesthe issue of what limits the distribution of ed by the National EcologyResearch Center ofthe U.S. vireos during the period of the year when only Fish and Wildlife Service. Logistical support was pro- the scrub-breeding Mangrove Vireo is present. vided by Diane Rano and the-Ohio State Cooperative During the temperate zone winter, virtually all Wildlife ResearchCenter. We thank the Secretariatof Urban Development and Ecologyof Mexico for allow- terrestrial habitats on the Yucatan Peninsula ing the use of the facilities at the Sian Ka’an Biosphere support one or the other of the two small vireo Reserve.Jon Barlow, William Hardy and EugeneMor- species. During the summer months, the pre- ton generouslyloaned us recordings of vireo vocali- sumed breeding seasonfor the Mangrove Vireo, zations. only scrub habitats support a member of the White-eyed Vireo group although the much larg- LITERATURE CITED er and distantly related Yellow-green Vireo BARLOW, J. 1980. Patternsof ecologicalinteractions (Johnson et al. 1988) breeds in dry forest. amongmigrant and residentvireos in the winter- It is unclear why no member of the White- ing grounds,p. 79-107. In A. Keast and E. S. ECOLOGICAL RELATIONSHIPS IN VIREOS 999

Morton, [eds.], Migrant birds in the neotropics. MACARTHUR,R. H. 1958. Populationecology of some Smithsonian Institution Press, Washington, DC. warblers of northeasternconiferous forests. Ecol- BATES,J. 1992. Frugivory on Bursera rnicrophylla ogy 39:599-619. (Burseraceae)by wintering Gray Vireos (Vireo vi- MILLER,A. H., H. FRIEDMAN,L. GRISCOM,AND R. T. cinior,Vireonidae) in the coastaldeserts of Sonora. MOORE. 1957. Distributional check-list of the Southwest.Nat. 31:252-258. birds of Mexico, Part II. Pacific Coast Avifauna BLAKE,E. R. 1968. Family Vireonidae, p. 103-138. 33. In R. A. Pavnter. Jr. Ied.1. Checklist of birds of MONROE,B. L. 1968. A distributional survey of the the world, vol 14. Museum of Comparative Zo- birds of . Ornithol. Monogr. 7. ology, Cambridge, MA. MORRISON,M. 1982. The structure of western war- CHAPMAN,F. 1896. Notes on birds observed in the bler assemblages:ecomorphology of the Black- Yucatan. Bull. Am. Mus. Nat. Hist. 8:27 l-290. throated Gray and Hermit Warblers. Auk 99:503- CHAPMAN,F. 1940. A post-glacial history of Zono- 513. trichia capensis.Bull. Am. Mus. Nat. Hist. 77: MORTON,E. S. 1973. On the evolutionary advantages 381-438: and disadvantagesof fruit-eating in tropical birds. CHIPLEY,R. M. 1980. Nonbreeding ecology of the Am. Nat. 107:8-22. Blackburnian Warbler. D. 309-3 17. In A. Keast OLMSTED.I.C.. A. LOPEZORNAT. AND R. DURAN. 1983. and E. S. Morton [eds.j, Migrant birds in the Neo- Vegetation, p. 4 l-64. In Sian Ka’an, estudiospre- tropics. Smithsonian Institution Press, Washing- liminares. Puerto Morelos, Quintana Roo, Mex- ton, DC. ice. SEDUE. Cox, G. 1968. The role of competition in the evo- PARKES,K. C. 1990. A revision of the Mangrove lution of migration. Evolution 22: 180-l 92. Vireo (Vireo pallens) (Aves: Vireonidae). Ann. Cox, G. 1985. The evolution of avian migration svs- Carnegie Mus. 59:49-60. terns between temperate and tropical-regions-in PAYNTER,.!R., R. A. 1955. The ornithogeographyof the New World. Am. Nat. 126:451-474. the Yucatan Peninsula. Peabody Mus. Nat. Hist., GASTON, A. J. 1974. Adaptation of the genus Phyl- Yale Univ. Bull. 9. loscopus.Ibis 116:432-450. POOLE,E. L. 1938. Weights and wing area of North GREENBERG,R. 1986. Competition in migrant birds American birds. Auk-55:5 1 l-5 17: in the nonbreeding season,p. 281-307. In R. J. REMSEN.J. V. AND S. K. ROBINSON.1990. A classi- Johnston, [ed.], Current ornithology, vol. 3. Ple- fication scheme for foraging behavior of birds in num Press,New York. terrestrialhabitats. Stud.-Avian Biol. 13:144-160. GREENBERG,R. 1992. Forest migrants in non-forest RICE, J. 1978. Ecologicalrelationshins of two inter- habitats in Quintana Roo, Mexico, p. 213-287. In specificallyterritohal Vireos. Ecology59:526-538. J. Hagan and D. Johnston[eds.], Ecology and con- RICKLEFS,R. 1992. The megapopulation:a model of servation of Neotropical migrant landbirds. demographiccoupling between migrant and resi- Smithsonian Institution Press, Washington, DC. dent populations,p. 537-548. In J. Hagan and D. HAMILTON.T. H. 1958. Adaotive variation in the Johnston,[eds.], Ecology and conservationof neo- genusVireo.‘ Wilson Bull. ?0:307-346. tropical migrant landbirds. Smithsonian Institu- HAMILTON,T. H. 1962. Speciesrelationships and ad- tion Press, Washington, DC. aptations for sympatry in the avian genus Vireo. RUSSELL,S. M. 1964. A distributional survey of the Condor 64:40-68. birds of British Honduras. Ornithol. Monogr. 1:l- HELLMAYR,C. E. 1935. Catalogue of birds of the 195. Americas and the adjacent islands, part 8. Field SALAMONSON,F. 1957. The evolutionary significance Mus. Nat. Hist. Zool. Ser. 13. of bird migration. Dan. Bol. Medd. 22: l-6 1. JOHNSON,N. K., R. A. ZINK, ANDJ. A. MARTIN. 1988. SCHOENER,T. 1968. The Anohs lizards of Bimini: Genetic evidence or relationships in the avian resourcepartitioning in a complex fauna. Ecology family Vireonidae. Condor 90:428-445. 49~704-126. KLOPFER,P. 1967. Behavioral stereotypy in birds. SMITHE,F. B., AND R. A. PAYNTER,JR. 1963. Birds Wilson Bull. 79:29-300. ofTika1, . Bull. Mus. Comp. Zool. 128. LACK, D. 1944. Ecological aspects of species for- WAIDE, R. 1980. Resourcepartitioning between mi- mation in passerinebirds. Ibis 86:260-283. grant and resident birds: the use of irregular re- LECK,C. 1972. The impact of some North American sources,p. 337-352. In A. Keast and E. S. Morton migrants at fruiting trees in Panama. Auk 89:842- reds.],Migrant birds in the neotronics. Smithson- 850. ian Institution Press, Washington; DC. LEISLER,B.,ANDH. WINKLER. 1985. Ecomornholonv, WILKINSON. 1987. SYSTAT manual. Winston. IL. p. 155-186. In R. F. Johnston [ed.], Current or- WILLIAMSON,P. 1971. Feeding ecology of the’ Red- nithology, vol. 2. Plenum Press, New York. eyed Vireo (Vireo olivaceus)and associated fo- LYNCH, J. F. 1989. Distribution of overwintering liage-gleaningbirds. Ecol. Monogr. 4 1:129-l 52. nearctic migrants in the Yucatan Peninsula: 1. WILLIS, E. 0. 1966. The role of migrant birds at General patterns of occurrence.Condor 9 1:515- swarms of army ants. Living Bird 5: 187-23 1. 545.