Feeding Patterns and Aggressive Behavior in Juvenile and Adult American ...

The Condor93~916925 0 The CooperOmithologicalSociety1991

FEEDING PATTERNS AND AGGRESSIVE BEHAVIOR IN JUVENILE AND ADULT AMERICAN FLAMINGOS'

KEITH L. BILDSTEIN~ Departmentof Biology, Winthrop College,Rock Hill, SC 29733 and Belle W. BaruchInstitutefor Marine Biologyand CoastalResearch,

Universityof South Carolina, Columbia, SC 29208

PETER C. FREDERICK Departmentof Wildlife and Range Sciences,Universityof Florida, Gainesville,FL 32611

MARILYN G. SPALDING

Departmentof InfectiousDiseases,Collegeof VeterinaryMedicine, Universityof Florida, Gainesville,FL 3261I

Abstract. We studiedthefeedingandaggressivbeehaviorofadultandjuvenileAmerican Flamingos(Phoenicopterusruberruber)at a coastalsalinain Venezuela.Most birdsfed in largeflocksin earlymorning,roostedat mid-day,and resumedfeedingin late afternoonearlyevening.Flamingosrarelyflew,exceptwhendisturbed.Adultsin our studystepped at thesamerateswhilefeedingasdid ChileanFlamingos(P.chilensis)feedingin theChilean and Bolivian Andes.Pairedobservationsof birdswithin mixed-ageflocksrevealedthat adultssteppedmore duringfeedingbouts,but lessbetweenbouts,and spentmore time overallwith their bills in the waterfilter-feedingt,handid juveniles.We calculatethat the food-intakerate of juvenileswas,at most,82% that of adults.Bothadultsandjuveniles walk-fedand stamp-fed.Juveniles,but not adults,weretwiceas likely to be supplanted whenstamp-feedingthan whenwalk-feedingJ. uvenilesweremore ofteninvolved in aggressione, speciallyas recipientst,han wereadults.Aggressivencounterssignificantlyaffectedthe amountof time flamingosspentfilterfeeding.

Keywords: Activitypatterns;fng behaviorA; mericanFlamingo;Phoenicopterursuber ruber;wetlands;Venezuela.

INTRODUCTION

prey. The presumption hasbeen that individuals

In many colonial waterbirdsjuveniles forageless efficiently than do adults (seereview in Bildstein 1983). In severalspeciesthis phenomenon, which is sometimes accompanied by age-relateddifferences in aggressionand habitat use (Goss-Custard et al. 1982a, 1982b), has been linked to deferred maturity (Lack 1968, Burgerand Gochfeld 1981, Greig et al. 1983, Bildstein 1984, MacLean 1986) and higher juvenile mortality (Recher and Recher 1969, Heppleston 1971).

Previous studies of age-related differences in the foragingbehavior of colonial waterbirds have focused on speciesthat stalk and subsequently pursue relatively large and highly mobile prey (Bildstein 1983), and not on speciesthat filterfeed on abundant and relatively immobile small

feeding on limited and relatively large mobile prey require longer periods of time to develop the suite of foraging skills needed to feed at adult ratesthan do speciesthat feedon more abundant, smaller, and lessmobile prey (cf. Amadon 1964, Ashmole 197 1). Recently, however, Giroux and Bedard (1988) have shown that even in herbivorous speciessuch as Snow Geese (Chen caerulescens),juveniles forage less efficiently than adults; a circumstance the researcherslinked to higher juvenile mortality in the species.

American Flamingos (Phoenicopterus ruber ruber) are filter-feeding wading birds that breed in coastal South America, the Caribbean basin, and the GalapagosIslands (Allen 1956, Kear and Duplaix-Hall 1975, Ogilvie and Ogilvie 1986). Researchershave suggestedthat first-year Amer-

ican Flamingos suffer higher mortality than do

I Received14 February 1991. Final acceptance8 older birds (Allen 1956, de Boer 1979). Although

May 1991.

it is unclear why this might be so, de Boer (1979)

ZSendreprint requeststo Departmentof Biology, proposedthat lower juvenile survivorship could

WinthropCollege,RockHill, SC 29733.

result from an inability on the part of juveniles

[9161

AGE DIFFERENCES IN FLAMINGO FEEDING AND AGGRESSION

917

to find sufficient food. Although juvenile flamin- face of the water, a distance of 13 cm. Between

gos begin to filter feed in their second week 11:30 and 11:SOon 27 April, we towed the sled

(Chapman 1905, Studer-Thiersch 1975) the 0.40-0.55 m/set over a distanceof 23.7 m, twice

possibility that first-year birds do not feed as over undisturbed sediment, and twice with one

proficiently as adults has not been examined. person stepping about 5 m ahead of the sled to

Studies of flamingo feeding biology include disturb the flocculentand vegetative layers. Sam-

general and detailed descriptions of the behav- ples were placed in 45% alcohol, stained with

ioral patterns used (Buffon 1781, Allen 1956, Rose Bengal, and examined with a 10 x micro-

Rooth 1965, Ogilvie and Ogilvie 1986) esti- scope. Samples of mud were similarly strained

mates of prey consumption (Rooth 1965) ac- through l,OOO-micron mesh and stored in al-

counts of their mouth parts (Jet&in 1957), and cohol.

descriptions of spatial and temporal patterns of We used binoculars and telescopesto watch

feeding activity (Hurlbert and Keith 1979; Hurl- flamingos from two 3.5 m cinderbrick towers

bert 1982; Britton et al. 1986; Espino-Barros and used by tourists. Almost all of the flamingos

Baldassarre1989a, 1989b). Studiesofindividual whose feeding behavior we studied fed within

variation in feeding behavior, similar to those 300 m of the towers; however, we included fla-

that have been reportedfor other speciesof water mingos feeding and roosting more than a kilo-

birds (e.g.,gulls, Burger and Gochfeld 198 1;Ma- meter from the towers in our activity surveys.

clean 1986; ibises, Bildstein 1983; herons, Rech- Initial observations indicated that flamingo

er and Recher 1969) are lacking.

activity was synchronized within flocks. We re-

Here, we detail and compare the feeding be- corded the numbers of birds in both roosting

havior of adult and juvenile American Flamin- flocks(i.e., birds standingin groups,often on one

gosin coastalVenezuela. Specifically,(1) we doc- leg) and feeding flocks (i.e., groups of walking

ument age-relateddifferencesin the foraging and and stamping birds that engagedin intermittent

aggressivebehavior of flamingos in this popu- filter-feeding) at the site during 16 morning (06:

lation, and (2) discussthe potential consequences 50-08:59), 22 mid-day (09:00-l 3:00), and 8 late

of our findings in terms of the concept of a be- afternoon-early evening (16:00-18:30) scan

havioral bottleneck and its impact on social be- samples(Altman 1974) on 11 different days. We

havior, habitat use,and conservation in this spe- used scan sampling during 8 morning, 6 mid-

cies.

day, and 6 late afternoon-early evening surveys

METHODS

to examine temporal variation in walking, walkfeeding, stamp-feeding, standing, preening, and

We studied flamingos feeding in a coastal salina bathing among birds within feeding flocks. Dur-

(10"58'N, 68"2O'W) at the Cuare National Wild- ing our observations, temperatures ranged from

life Refuge, near Chichiriviche, in Falcon State, 25-3 1?C winds were light, and no precipitation

Venezuela, from 2 1 April-l May 1989. The Chi- fell.

chiriviche salinas,which are usedasfeeding sites Flamingos were counted as walking when they

by flamingos that breed approximately 90 km steppedwith their head up or extended forward

away on Bonaire in the Netherlands Antilles (de without feeding, and as walk-feeding when they

Boer 1979) support at least several thousand walked at a relatively constantpacewith the head

birds each winter and spring (Guzman 1986; M. alternately up or extended forward, and down,

Lentino, pers. comm.; Bildstein, Frederick, and either filter feeding in the water column, or re-

Spalding, pers. observ. in 1988 and 1989). The peatedly probing the water column with "forceps

number of flamingos feeding at the site declines movements" of the bill (cf. Rooth 1965). Birds

during the spring as the dry season progresses were counted as stamp-feeding when they

and water in the salinasevaporates(de Boer 1979; stamped their toes while rotating their body

M. Lentino, pers. comm.).

clockwiseor counterclockwisearound their par-

We towed a 1,OOO-micronmesh, rectangular- tially or fully submergedbill (seealsoAllen 1956,

mouthed tapering plankton net attached to a Rooth 1965). Flocks were defined as groups of

benthic sled (two skis 61 cm apart) to capture flamingos in which the nearest-bird distance av-

potential flamingo prey items in the water col- eraged ~25 m, and where no bird was >50 m

umn. The 61-cm wide net extended from less from the nearest flock member.

than 1 cm above the flocculent layer to the sur- We recorded the ratios of adults and juveniles

918 K. L. BILDSTEIN, P. C. FREDERICK ANDM. G. SPALDING

in seven flocks on five days and compared those at the edgeof the salina, dropped at a rate of 4.5

ratios with the ageratios of the nearestneighbors mm/day. On 27 April, when we towed a plankton

of juveniles within those flocks. Juveniles and sled through a portion of the salina frequently

adults were determined by plumage (Bent 1926, usedby feeding birds, the water was 18 cm deep,

Allen 1956, Rooth 1965). We counted as juve- including a flocculent layer of 5-8 cm.

niles only those individuals with entirely gray- Flamingos rarely flew, even when disturbed by

and-white plumage(i.e., clearlyyoungof the year), people discharging firearms and fishing at the

and as adults individuals with pink or bright red site. About half of the birds took flight briefly

plumage. The few birds that appearedto be molt- and circled above the area on the morning of 30

ing from juvenal to adult plumage were not in- April following the aftershock of a 5.0 Richter-

cluded in our observations.

Scaleearthquake centered 25 km away earlier in

PAIRED FEEDING OBSERVATIONS

the day. Roosting flocks averaged 3 15 -t 247 SD birds

We recorded the behavior of 46 "paired" adult (n = 47). Aside from one singleton, all of the

and juvenile flamingos feeding togetherin mixed birds we observedduring 4-min observationsfed

flocks, as well as the behavior of 70 unpaired in flocks of between 16 and 923 individuals (sz

birds. Paired 4-min observations were recorded = 59 1 birds/flock for juveniles, II = 47; 1z= 419

within 10 min of each other, and were limited birds/flock for adults, IZ = 113; P = 0.0002;

to birds feeding within 20 m of each other. We t-test).

recorded (1) the numbers of stepstaken (stamping and otherwise), (2) whenever the bird sub- POTENTIAL PREY

mergedits bill and wasfeeding,and (3) aggressive Flamingos appeared to filter feed both from the

interactions with other flamingos. Observations, water column and alongthe benthic surface.None

which were made throughout the day, were dic- ofthe four mid-day tows,however, producedany

tated into a cassetterecorder for later transcrip- obvious prey items, despite our "filtering" an

tion. We defined aggressionas threat displaying estimated 7,5 12 liters of salina water. Though

(Allen 1956, Ogilvie and Ogilvie 1986) and the most flamingo prey may have been smaller than

rapid approach of one bird that resulted in dis- our mesh size, we expectedto find at least a few

playing, body contact, or the rapid withdrawal small items caughton the mesh. We believe that

of another individual. Threat displays and ap- flamingos were not filtering substantial amounts

proachwere assumedto be dominant behavioral of prey from the water column. We also strained

patterns;rapid withdrawal wasconsidereda sub- 850 ml of the top 8 cm of sediment collected at

ordinate pattern.

five different siteswithin the salina for potential

OTHER FEEDING OBSERVATIONS

prey items. Although we found large numbers of widgeon grass(Ruppia maritima) seedsin our

We also recorded the stepping rates of feeding sample, all of the 56 seedswe examined closely

birds during 200 20-set observation bouts, and consistedof old hulls from the previous season.

we recorded both the direction and rate of tum- We alsoidentified 2 18 piecesof polychaeteworms

ing (in degrees)of stamp-feedingindividuals dur- (family Capitellidae), most of which appearedto

ing 100 30-set observation bouts. We also mon- represent at least half of a worm (i.e., a total of

itored the behavior of 17 adults for up to 79 min, approximately 109 worms, or O.l4/ml of mud).

to determine average durations of feeding pat- As the water level dropped, birds feeding along

terns and synchrony in feeding behavior among the edgeof the salina plucked dead or moribund

nearest neighbors.

3-10 cm fishesfrom the water surface.

We used Chi-square tests, t-tests and paired t-tests, and analysis of variance to analyze our FEEDING BEHAVIOR

data (Sokal and Rohlf 1969).

Flamingos fed mainly during early morning and

RESULTS

late afternoon-early evening, and roostedduring the middle of the day (Table 1). Feeding activity

The number of flamingos feeding and roosting appeared to be greater late in the afternoon than

at the salina declined from approximately 1,100 earlier in the day (Table 2).

on 2 1 April to lessthan 600 on 1 May. The water Walk-feeding was more common than was

level, which we measured with a marked stake stamp-feeding(Table 2). Individuals walk-fed for

AGE DIFFERENCES IN FLAMINGO FEEDING AND AGGRESSION 919

TABLE 1. Relativeoccurrencoef flamingosin roostingversusfeedingflocksduringmorning,mid-day,and lateafternoon-earlyeveningsurveys.

Time of day'

Morning (n = 16) Mid-day (n = 22) Late afternoon-early

evening(n = 8)

869 79

21

726 28

72

1,000 73

27

' Morning = 06X1-08:59; Mid-day = 09:00-13:oO;Late aftemoonearlyevening= 16:00-18:30.

3). When stamp-feeding, adults steppedat more than three times the pace of walk-feeding birds (50 f 7.6 steps/20-set observation versus 15 + 4.5 steps/20-set, P = 0.0001, n = 200). In 27 of our 4-min observationsindividual flamingosboth walk-fed and stamp-fed. During these observations, feeding bouts were longer, and the number of stepsperbout greater,when the bird wasstampfeeding than when it was walk-feeding (Table 4).

AGE-RELATED DIFFERENCES IN FLOCKING BEHAVIOR

an average 6 12 set (h 1,099 SD, n = 23). Birds stamp-fed for an average of 289 set (2266, n = SO),unless disturbed by another bird, in which case they moved on after having stamp-fed for only 160 set (k95.2, IZ = 15). Walk-feeding flamingosusually moved at least20 m during 4-min observations. Stamping flamingos fed in one location, turning rapidly in place (2 = 493 + 156" SD/30-see observation, n = 50 adults).

Although we found the circular plateaus surrounded by shallow depressionsthat result from stamp-feeding (cf. Gallet 1950) throughout the salina, flamingos appeared to stamp-feed in certain portions of the salina more than in others. Stamping may have been sociallyfacilitated, with its occurrence in one individual increasing the likelihood of its occurrencein neighboring birds. In all four pairs of nearest-neighbor adult flamingos we watched for prolonged periods, both birds frequently switched in unison from walkfeeding to stamp-feeding and back to walk-feeding, and in one instance, two adults even synchronized individual feeding bouts (i.e., submersed and lifted their bills in unison).

Flamingos that both stamp-fed and walk-fed steppedmore, both within and between feeding bouts, than did birds that only walk-fed (Table

Although adults often fed together without juveniles, we never sawjuveniles feeding without adults. Over five days, seven mixed-age feeding flocksranged from 153 to 935 birds, with adults comprising 96-99% of all birds counted (X = 97% adults). Of 111 juveniles feeding within these flocks, 37% had another juvenile as their nearest neighbor: more than 10 times the number expected based on a random distribution (P < 0.001; Chi-square test for goodness-of-fit).

AGE-RELATED DIFFERENCES IN FEEDING BEHAVIOR AND AGGRESSION

Neither the number of individual feeding bouts nor the overall number of steps taken by adult and juvenile flamingos differed during our paired 4-min observations (Table 5). However, adult flamingos steppedmore during individual feeding bouts, and lessbetween bouts, and they spent more time with their bills in the water during feedingbouts, than did juvenile flamingos (Table 5). As a result, adult flamingos spent more time filter feeding overall than did juveniles (Table 5).

Both adult and juvenile flamingos regularly engagedin both walk-feeding and stamp-feeding during 4-min observations (Table 6), and there was no indication of an age-related difference in the frequency of use of these two foraging techniques (x2 = 0.53, P = 0.47).

TABLE 2. Activitiesof flamingosin feedingflocksduringmorning,mid-day,andlateaftemoon+arlyevening surveys.

Time of day'

Walking

Percentengagedin Walk-feeding Stamp-feeding Inactive

Preening

Bathing

Morning Mid-day Late aftemoon-

early evening

10

15

38

9

0

5

9

45

4

0

8

52

12

10

15

2

' Morning = 06:50-08:59, eight surveysof a total of 2,593 birds; Mid-day = 09:00-13:00, six surveysof a total of 1,186 birds;Late aftemoon-

earlyevening= 16:OC-l&50, SIXsurveysof a total of 1,243birds.

920 K. L. BILDSTEIN, P. C. FREDERICK AND M. G. SPALDING

TABLE 3. Behaviorofjuvenile and adult flamingosduring4-min observationswith and without stamp-feeding behavior.

Behavior

Number of feeding bouts

Total number of steps

Number of secondsfeeding

Number of secondsper feeding bout

Number of stepsper feeding bout

Mean number of steps between feeding bouts

Number of stepsduring feeding bouts/number of stepsbetween bouts ' t-tests.

Juveniles

Without

With

stamping

stamping

(n = 24)

(n = 24)

PI

R = 42.7 SD = 14.7

K= 170 SD = 55.4

K= 149 SD = 45.3

x = 3.79 SD = 1.69

x = 2.31 SD = 0.93

K = 2.27 SD = 4.09

55= 2.46 SD = 2.37

.Z = 36.0 SD = 15.8

K = 289 SD = 120

K= 136 SD = 38.7

55= 4.20 SD = 1.76

.z = 5.98 SD = 4.05

K = 2.86 SD = 1.99

K = 2.84 SD = 2.86

0.14 0.0001 0.30 0.42 0.0002 0.54 0.61

Without stamping h = 63)

.z = 40.3 SD = 12.2

K= 153 SD = 55.7

X= 170 SD = 34.7

x = 4.63 SD = 1.86

K = 2.73 SD = 1.21

X= 1.25 SD = 1.19

x = 4.31 SD = 4.27

Adults

With

stamping

cn= 49)

P

x = 37.5 SD = 11.0

x = 326 SD = 146

K= 167 SD = 27.1

x = 4.86 SD = 1.70

.X= 7.47 SD = 4.69

x = 1.96 SD = 1.21

x = 4.66 SD = 3.48

0.20 0.0001 0.66 0.51 0.0001 0.003 0.64

Aggressive interactions occurred in 26% ofour 4-min feeding observations. Juveniles were more likely to be involved in aggressiveencounters(x2 = 10.8, P = 0.00 1, Table 7), and were more likely to be the recipient of aggressionduring theseinteractions (95% of aggressiveencounters versus 32% for adults, x2 = 17.8, P < 0.00 l), than were adults. For juveniles, but not adults, aggression was significantly more frequent during 4-min observations when stamp-feeding occurred than in those observations without stamping (for juveniles, x2 = 4.15, P = 0.042; for adults, x2 = 1.88, P = 0.17, Table 7). Even so, aggressiveinteractions were recorded in 24% of the 4-min adult observations in which stamping occurred, and aggressiveinteractions terminated 23% of the 65 stamp-feedingeventswe recordedamong 17adult flamingos that we watchedfor prolongedperiods. In many of the instances where stamp-feeding birds were harassedby another flamingo, the second bird began stamping at the site vacated by the first. An analysis of variance of the total time

flamingosspentfilter feedingindicatesthat, when ageis held constant, the occurrenceof aggressive interactions (2-way ANOVA, P < 0.01) significantly affects total time spent filtering. For example, feeding adults interrupted by aggressive encounters spent 7% less time filter feeding, and interrupted juveniles 6% less time feeding, than did individuals whose feeding behavior was not interrupted.

Although a number of additional species, including Olivaceous Cormorants (Phahcrucorax turbo), eight speciesof wading birds, at leastthree speciesof shorebirds, Gull-billed Terns (Stenla niloticu), and Black Skimmers (Rynchops niger) fed within severalmetersof flamingos, flamingos did not interact aggressivelywith these birds.

DISCUSSION

ACTIVITY PATTERNS AND FLOCKING BEHAVIOR

The decreasing number of flamingos feeding at the site over the course of our study probably

TABLE 4. Behavior of flamingos during 4-min observationsin which both walk-feeding and stamp-feeding occurred.

Juveniles (n = 11)

Adults (n = 16)

' Pairedt-tests.

Mean time with headdown per bout

Walk feeding

Stamping

P'

x = 2.92 SD = 1.13

K = 3.88 SD = 1.57

_z= 4.13 SD = 1.96

K = 4.89 SD = 1.88

0.01 0.06

Mean numberof stepsper bout

Walk feeding

Stamping

P

x = 2.47 SD = 1.09

x= 3.11 SD = 1.63

K = 7.57 SD = 3.57

X = 9.83 SD = 4.23

0.0003 0.0001

 ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download