MORPHOLOGICAL AND FORAGING BEHAVIORAL DIFFERENCES BETWEEN SEXES OF THE MAGELLANIC WOODPECKER (CAMPEPHILUS MAGELLANICUS)

529

ORNITOLOGIA NEOTROPICAL 23: 529–544, 2012© The Neotropical Ornithological Society

MORPHOLOGICAL AND FORAGING BEHAVIORAL DIFFERENCES BETWEEN SEXES OF THE MAGELLANIC

WOODPECKER (CAMPEPHILUS MAGELLANICUS)

Laura Chazarreta1, Valeria Ojeda1, & Martjan Lammertink2,3

1Departamentos de Ecología y de Zoología, Universidad Nacional del Comahue-INIBIOMA, Quintral 1250, (8400) Bariloche, Argentina. E-mail: [email protected]

2CICyTTP-CONICET, Matteri y España s/n, (3105) Diamante, Entre Ríos, Argentina.3Cornell Lab of Ornithology, 159 Sapsucker Woods Road, Ithaca, NY 14850, USA.

Resumen. – Diferencias morfológicas y de comportamiento alimentario entre sexos delCarpintero Gigante (Campephilus magellanicus). – La divergencia ecológica producto de diferenciasmorfológicas y comportamentales junto con la dominancia social, promueve la separación de nichosentre sexos en aves. La ausencia de competidores favorecería estas diferencias intersexuales. ElCarpintero Gigante (Campephilus magellanicus) es el único pícido de gran tamaño en los bosques deNothofagus de la Patagonia, sin competidores. Una investigación preliminar documentó diferencias mor-fológicas entre sexos y postuló la existencia de diferencias en el comportamiento alimentario, correla-cionadas con ese dimorfismo. Nuestros objetivos fueron analizar las diferencias morfológicas einvestigar si existen diferencias intersexuales en el comportamiento alimentario. Obtuvimos datos mor-fológicos de colecciones de museos y registros comportamentales de poblaciones de Patagonia Argen-tina. Estimamos amplitud y solapamiento de nichos en variables de alimentación. Los machos fueronmás grandes y tu-vieron picos 12,4% más largos que las hembras. Ambos sexos se alimentaron princi-palmente sobre árboles vivos, pero los machos usaron sustratos más grandes (troncos) que las hem-bras y alturas intermedias (5–10 m), mientras que las hembras usaron sustratos más pequeños (ramas)dentro de la copa, a mayor altura (> 15 m). La tasa de captura de presas fue similar entre sexos (0,28presas/min), pero los machos capturaron presas más grandes (larvas xilófagas) que las hembras (pre-sas superficiales). Registramos dominancia social de los machos. Según el análisis de amplitud denicho, las hembras fueron más generalistas que los machos en el uso de microhábitats y en la posicióndel cuerpo mientras se alimentaban. El dimorfismo sexual en tamaño corporal y largo del pico fue acordea la especialización por sustratos en los sexos, lo cual fue probablemente reforzado por la dominanciade los machos. La falta de competencia interespecífica en este bioma contribuiría a la diferenciaciónobservada entre los sexos.

Abstract. – Ecological differentiation arising from morphological and behavioral differences, togetherwith social dominance, is known to promote niche differentiation between sexes in birds. The absence ofcompeting species would favor intersexual differences. The Magellanic Woodpecker (Campephilusmagellanicus) is the only large woodpecker in the southern beech Nothofagus forests of Patagonia, withno competitors. Sexual divergence in morphology had been documented by preliminary research, anddifferences in foraging behavior were proposed as a correlate. Our aims are to analyze intersexual diver-gence in morphology and to investigate whether the behavioral differences exist. We obtained morpho-logical data from ornithological collections and foraging records from populations from ArgentinePatagonia. We estimated foraging niche breadth and the overlap of foraging variables. Adult males werelarger and had bills 12.4% longer than those of females. Both sexes foraged mostly on living trees, butmales foraged on larger substrates (trunks) at intermediate heights (5–10 m), while females foraged

530

CHAZARRETA ET AL.

higher within the crown (> 15 m) on smaller substrates (branches). The rate of captured prey was similarbetween sexes (0.28 prey/min), but males consumed larger prey (wood-boring larvae) than females(near-surface prey). Social dominance by males was recorded. Based on analysis of niche breadth,females were more generalist than males in microhabitat use and body posture when foraging. Sexualdimorphism in body and bill size seems to be in line with specialization in the use of different substratesbetween the sexes, which is probably reinforced by male dominance. Lack of interspecific competition inthis biome probably contributed to the differentiation observed between sexes. Accepted 23 December2012.

Key words: Magellanic Woodpecker, Campephilus magellanicus, Austral Temperate Forest, insularbiome, niche breadth, sexual dimorphism.

INTRODUCTION

Sexual dimorphism in size and morphologyis widespread among animals (Hedrick &Témeles 1989) and has traditionally beenrelated to sexual selection, especially in birdspecies with a polygamous mating system(Amadon 1959, Selander 1972). Nevertheless,the evolutionary origin of sexual dimorphismhas multiple pathways (Hedrick & Témeles1989), and there are many cases where differ-ent size and morphology between males andfemales can be related to ecological causes(Slatkin 1984, Shine 1989), as in Green Wood-hoopoes (Phoeniculus purpureus, Radford &Plessis 2003), Darwin’s finches (Grant 1986),and Varied Sittella (Daphoenositta chrysoptera)and treecreepers (Climacteridae) (Noske1986).

Segregation between sexes in morphologi-cal traits is thought to alleviate intersexualcompetition (Selander 1966, Ruckstuhl &Clutton-Brock 2005). The absence of com-peting species would favor a larger nichewidth of a species, leading to greater struc-tural dimorphism and the subdivisionbetween sexes of the foraging niche (Selander1966, Shine 1989). Social dominance, wherethe dominant sex (usually male) excludes thesubordinate from favored patches or sub-strates, explains sexual niche segregation inseveral cases (Hogstad 1978, Morse 1980,Peters & Grubb 1983, Matthysen et al. 1991,Osiejuk 1994, Ruckstuhl 2007).

Woodpeckers (Picidae) are often sexuallydimorphic in plumage and morphology, andthus, are an interesting group for studying dif-ferences in foraging behaviors between maleand female (Catry et al. 2005). In particular,sexual differences in bill size have beenrelated to differences in feeding behavior,explored habitats, and consumed items(Selander 1966, Aulén & Lundberg 1991).The southern beech (Nothofagus) temperateforests covering southern Chile and Argen-tina (35°S–56°S), comprise a narrow strip ca.2000 km long and up to 120 km wide thatevolved in isolation from other South Ameri-can forests, constituting a habitat island withhigh levels of endemism (Vuilleumier 1985,Armesto et al. 1996). In the interior of Nothof-agus forests, only two woodpecker speciesoccur, the small Striped Woodpecker (Venilior-nis lignarius, ~ 16 cm, 35–38 g) and the verylarge Magellanic Woodpecker (Campephilusmagellanicus, ~ 40 cm, 276–363 g) (Short1982). In this particular setting, the Magel-lanic Woodpecker has no potential ecologicalcompetitor and therefore would show a broadwoodpecker niche (Short 1970a). Pioneeringresearch conducted by Short (1970a) revealeda noticeable sexual dimorphism in bill size(males larger than females), which he pro-posed would correlate to differences in feed-ing behavior and explored habitat, but thisremained untested. Our aims are to analyzesexual dimorphism in morphology of Magel-lanic Woodpeckers and to explore sex-specific

531

DIFFERENTIATION BETWEEN SEXES OF MAGELLANIC WOODPECKERS

foraging behavior and feeding habitats of thisspecies.

METHODS

Study area. Study sites (n = 8) were locatedwithin the Lanín (39°35’–41°19’S) andNahuel Huapi National Parks (71°17’–71°36’W), north and south of Bariloche city,in northern Argentine Patagonia. The climatein the region is characterized by a relativelylong winter with rain and snow and by drysummers, with mean monthly winter (June–August) temperatures of 2°–3°C, and summer(December–March) temperatures of 12°–14°C (data from Bariloche Airport, 41°09’S,71°10’W).

Subalpine forests between 1100–1700 ma.s.l. are composed of pure stands of thedeciduous lenga Nothofagus pumilio. Our fieldsites were old-growth lenga stands over 1000ha each, which were contiguous with otherNothofagus stands. Lenga forest was selectedbecause of the great extent of this forest typein the study area and its mostly open under-story that allows following woodpeckers formonitoring their behavior.

Species. Magellanic Woodpeckers show anoticeable sexual plumage dimorphism, whichfacilitates field studies involving sex recogni-tion. They reproduce in monogamous pairsand live in stable family groups consisting of2–5 individuals, with offspring delaying dis-persal for up to four years (Chazarreta et al.2011). Family groups move in relatively lim-ited areas of forest (ca. 100 ha) all year round,and members of a group keep in close prox-imity (Ojeda 2004).

Morphology. We measured morphology fromspecimens held in 19 ornithological collec-tions (see Acknowledgments) in order toexplore differences between sexes. Fifty-ninemales and 64 females were measured follow-

ing Baldwin et al. (1931) guidelines. We onlyused adult specimens as determined fromplumage patterns (Ojeda 2004), museumlabels, and size (Short 1970a). We collectedthe following variables from each skin: wingand tail length, culmen cord (bill length), andtarsus, toe (third), and claw (third) length.Some skins were damaged so we were unableto collect all measurements from those speci-mens (n = 7). We took most measurements tothe nearest 0.01 mm when using calipers, andto the nearest 1 mm using a ruler for wingsand tail.

Nearly all specimen measurements (98%)were taken by one of us (LC) and the remain-der was taken by museum curators instructedby LC. Although skin dimensions vary withspecimen age, preparation technique, andstorage conditions (Bortolotti 1984), we con-sidered this variation a random effect in ourdata set.

Foraging behavior. We conducted fieldwork dur-ing three consecutive years from early springto late summer (September–April, 2008–2010). Most behavioral data (85% of observa-tion hours) were obtained from a populationof eight pairs and family groups (Challhuacosite, 15 km southeast of Bariloche) bandedduring a long-term study of the species(Ojeda 2004, Chazarreta et al. 2011). Thus,foraging data were mostly based on color-ringed males and females of known age andreproductive status. Additionally, we collectedsimilar data on adult (according to both plum-age and behavior) unringed birds located atother lenga forest sites. At these sites, when-ever two or more unringed adult birds of thesame sex were encountered, they were studiedonly if they were separated by more than 3km, representing different territorial birds.

We studied foraging behavior by followingfocal individuals under continued sampling ofvariable periods (Martin & Bateson 2008). Weconducted observations during all daytime

532

CHAZARRETA ET AL.

hours (approximately from 08:00–20:00 h,= GMT-0300) and only when birds wereactively foraging. To minimize the effect ofweather, we did not work under conditions ofheavy rain, snow or hard wind.

A territory was searched systematicallyuntil we encountered a woodpecker family.We then followed in turns, one of its mem-bers randomly selected, for as long as itremained in sight, and recorded a continuousaccount of its movements, foraging tech-niques, and other behaviors by dictating theminto a tape recorder, writing field notes, andusing a stopwatch to record time. Distancesbetween members of a pair were taken intoaccount and visually estimated. Behavior dur-ing the first 30 s after detection was excluded(Morrison 1984). In the banded population,the same individual was not observed morethan once a day to increase sample indepen-dence. To obtain detailed observations weused 10 x 40 binoculars. Usually, we couldapproach woodpeckers closely (~ 10 m) with-out apparent disturbance. Most observationswere of adult individuals that occurred infamily groups or pairs, but six observationswere of solitary males. We never met solitaryfemales during the study.

For each adult bird we recorded: sex, for-aging and searching technique, and microhab-itat (portions of the trees) using variablessimilar to those considered in other studies offoraging behavior of woodpeckers (Selander1966, Jackson 1970, Kilham 1972, Askins1983, Pasinelli & Hegelbach 1997, Stenberg &Hogstad 2004, Newell et al. 2009). We definedsearching/foraging techniques as: 1) excavat-ing: subcambial excavation, digging deep (> 2cm) holes; 2) debarking: striking the billagainst the substrate to remove some of theexterior (bark, wood, lichens), or diggingsuperficially; 3) probing: inserting the bill ortongue into cracks or crevices, or into excava-tions created by excavating, or picking insectsoff the bark surface; and 4) scanning: explor-

ing the substrate with lateral head movementsin order to find prey. Surface gleaning (pick-ing prey off the surface) was not consideredas a tactic itself because it was rarelyobserved. Most near-surface prey occurredunder the first layer of bark. Scanning tendsto be a dynamic technique: the birds climb ona stem or branch after and before usinganother foraging technique, usually upward,and alternate soft pecking, listening andmovement hops. During a sequence, we alsorecorded behaviors beyond specific foragingand searching techniques, which were catego-rized as: 5) food handling: when the focal birdhandled prey for itself or its offspring; 6)movement: when changing location withinsubstrate or among substrates; and 7) otherbehaviors, which were recorded as detailed aspossible (preening, resting, vocalizing, etc.).Body position while foraging was alsorecorded, distinguishing between semi-hori-zontal with head up, semi-horizontal withhead down, and vertical with head up. Type ofcaptured prey were assigned to broad catego-ries: wood-boring larvae; near-surface prey(mainly arachnids, adult insects or pupae), andindeterminate, when we were unable to rec-ognize a prey (usually because of their smallsize). We also recorded all agonistic interac-tions involving focal birds: interferences (adominant bird flies towards a subordinate andmakes it fly to another location), and chaseson trunks or in flight.

Foraging substrates. We defined “snag” as astanding dead tree with or without branches,“stump” as a dead tree < 3 m high, and“coarse woody debris” as downed wood. For-aging substrate diameters were estimated rela-tive to the focal bird’s back width (foldedwings), which is around 75 mm in MagellanicWoodpeckers. We estimated foraging heightsin two ways: 1) relative to the tree height, inthree categories (lower, mid, and higher thirdof the tree), and 2) absolute foraging height

533

DIFFERENTIATION BETWEEN SEXES OF MAGELLANIC WOODPECKERS

above ground, in four classes (0–5 m, 5.1–10m, 10.1–15 m, > 15 m). The absolute heightwas estimated using an electronic clinometer.Additional features recorded at foraging loca-tion were: bark/ no bark, branch joint ortrunk fork, and woodpecker-excavated pits(foraging holes).

Data analyses. With the morphological vari-ables recorded for skins, we explored differ-ences between sexes using generalizedlinear models (GLM) with logistic linkfunction and binomial error structure (Craw-ley 2007). We checked for multi-colinearity,transformed the variables as needed, andran models with all variables. We fittedseveral models using different combinationsof the variables as predictors, including theirinteractions. Fitted models were comparedand hierarchically ordered using the AkaikesInformation Criterion (AIC) (Akaike 1973,Richards 2005), choosing the model withthe lowest AIC value. Models were simplifiedby removing non-significant interactionterms first and then, non-significant predic-tors to generate the minimal adequate model(Crawley 2007).

The proportion of time spent by individu-als at each variable state was used in the analy-ses. We arcsine √χ-transformed data to meetnormality for the analyses. To circumvent aunit-sum constraint, variables with two ormore states were reduced by excluding at leastone state (the one that had the lowest valuesin most individuals) (Aebischer et al. 1993).For these analyses we only included statesstrictly referring to foraging techniques (prob-ing, excavating, debarking, and scanning). Weassumed no effect of the year and/or monthon the foraging behavior and pooled data foranalyses. To reduce bias that multiple recordsmay create (Morrison 1984) each individualwas treated as a sample unit when determin-ing proportions and sample sizes for statisticaltests (Airola & Barrett 1985). Therefore, we

added up the total minutes of the observationsessions (for each individual), and then esti-mated a single overall proportion of the vari-ous foraging variables.

To assess whether one of the sexes wasmore efficient than the other in capturingprey, we compared the success rates for eachindividual as the number of captured prey perminute and per number of used substrates,using the unpaired Wilcoxon signed rank testfor independent data.

We analyzed differences in foragingbehavior between sexes by conducting a Prin-cipal Component Analysis (PCA) to collapse aset of variables into a reduced number of newaxes. All components with eigenvalues > 1were retained for subsequent analyses. To bet-ter understand the foraging behavior relation-ships between sexes, we calculated nicheoverlap and breadth values. Foraging nichebreadth was calculated for each individualwith a standardized version of Levins’ index(Hulbert 1978):

For each variable, niche breadth (B) wascalculated separately. The proportion of for-aging time an individual spends foraging atresource state i is pi, and n is the total numberof resource states for the variable under con-sideration. Values range from one (for equaluse of all resource states) to zero, for special-ization on one resource state. To evaluate ifone of the sexes was more stereotyped thanthe other, we performed comparisonsbetween sexes for foraging diversity, using theunpaired Wilcoxon signed rank test for inde-pendent data. The degree of niche overlapwas determined using Schoener’s index(Schoener 1968):

∑ −−= ||211 fim imf ppC

)1(1)/1( 2

−−

= ∑np

B i

534

CHAZARRETA ET AL.

The index was used as an indicator of theamount of overlap between males andfemales in the use of resource states for a par-ticular variable. Values range from zero (nooverlap in resource use) to one (completeoverlap), and pmi and pfi are the mean propor-tion of foraging time that males and females,respectively, spent at resource state i (differ-ence between sexes at each state aresummed).

We performed all statistical analyses usingR statistical software version 2.11.1 (R Devel-opment Core Team, R Foundation for Statis-tical Computing, Vienna, Austria, www.R-project.org). PCA was conducted with thepackage FactoMineR v1.10 (Lê et al. 2008) inR. We tested normality and homocedasticityusing residual plots and used parametric testsunless the data distributions violated theassumptions. All tests were two tailed; valuesof P < 0.05 were considered significant. Wereport median and range, and also means ±SE.

RESULTS

Morphology. Males were larger than females inwing and tail, indicating that males were largeroverall. However, the dimension that best dif-ferentiated the sexes was the bill length (cul-men cord length in Table 1). The modelsincluding bill length showed the highest parsi-mony ranks according to AIC scores. Theselected model included as predictors the cul-men cord, wing, tail, and claw length. Parame-ter estimates for the best model are shown inTable 1.

Foraging behavior. Members of the pair nearlyalways foraged in close proximity (0–50 m)and therefore it was possible to estimate thedistance between the focal bird and its mate.The total time of observations was 25 h for31 males and 22 h for 21 females, spread in164 sequences; median time per sequence was

73.3 min (1.5–94.5 min) for males and 92.0min (1.8–107.5 min) for females. Fifty-eightpercent of males were observed for 5–30 minand the rest (42%) for more than 30 min.Fifty seven of the females were observed for5–30 min and 43% for more than 30 min.

Living trees were used by both sexesmuch more (> 90% of foraging time) thanany other substrate type; hence, comparisonsregarding microhabitat use were made onlyconsidering those substrates. As an exception,successful captured prey comparisonsincluded observations from all substrates.

According to the PCA, the first compo-nent was linked with dimensions (size andheight) of the microhabitat, while the secondcomponent was related to foraging tech-niques and qualitative features at the foraginglocations (Table 2). Males were associatedwith large values of the first and second com-ponents (R2 = 0.4 and 0.2, respectively, bothcomponents P < 0.001, Fig. 1), while femalesdid the opposite (t50 = 5.2, P < 0.001; t50 = -2.9, P = 0.005, PC1 and PC2, respectively).

Males foraged mostly on trunks and atlower heights than females. Females foragedon smaller substrates, which occurred higherin the trees (Fig. 2, Table 2). These correlatedfeatures of the foraging substrates resulted ina vertical separation of the sexes, with femalesforaging in the higher third (crown) morethan in the mid and lower thirds of livingtrees. The highest (terminal) parts of treeswere not used as much as other tree portionsby either of the sexes.

Males were more associated with foraginglocations characterized by the presence ofholes than females. Females foraged in bothmicrohabitats with and without bark, butlocations covered by bark were the most com-mon situation (Table 2). Closeness to limbjoint or trunk fork had no association withthe foraging location of Magellanic Wood-peckers. Because both sexes generally foragedon living trees, the use of dead wood as forag-

535

DIFFERENTIATION BETWEEN SEXES OF MAGELLANIC WOODPECKERS

TABL

E 1

. Mor

phom

etric

mea

sure

men

ts (m

m) o

f adu

lt m

ale an

d fe

male

Mag

ellan

ic W

oodp

ecke

rs. E

stim

ated

coe

ffic

ients,

stan

dard

err

or, a

nd P

valu

e of

the

varia

bles

of

the

best

mod

el ar

e gi

ven.

The

per

cent

age

diffe

renc

e is

expr

esse

d as

the

diffe

renc

e in

mea

n va

lues

bet

wee

n th

e se

xes

in r

elat

ion

to th

at f

orfe

male

s. M

eans

± S

E, s

ampl

e siz

e (n

), an

d sig

nific

ant d

iffer

ence

s (*)

are

show

n.

Male

sFe

male

sE

stim

ated

coe

ffic

ient

% D

iffer

ence

Mea

n ±

SE

n

Rang

eM

ean

± S

E

nRa

nge

Tail

Win

g Cu

lmen

Cor

d Ta

rsus

To

e (th

ird)

Claw

(thi

rd)

176.

80 ±

1.3

020

5.80

± 6

.18

54.8

1 ±

0.4

536

.33

± 0

.41

25.4

7 ±

2.5

620

.25

± 0

.09

59 59 59 59 58 58

144.

50–1

9619

8.50

–229

48.5

0–63

.94

29.1

7–42

.84

21.0

0–36

.15

19.0

0–21

.93

174.

70 ±

1.0

920

2.70

± 4

.79

48.7

5 ±

0.3

934

.26

± 0

.37

24.9

± 2

.68

19.2

9 ±

0.1

0

64 64 63 64 63 63

151–

190

198.

0–22

2.5

43.4

0–56

.00

25.5

3–41

.40

22.7

3–34

.48

16.8

8–21

.18

-0.0

6 ±

0.0

3, P

= 0

.08

0.10

± 0

.06,

P =

0.0

90.

54 ±

0.1

2, P

< 0

.001

*- -

1.59

± 0

.49,

P =

0.0

01*

1.2

2.2

12.4

6.0

2.3

5.0

536

CHAZARRETA ET AL.

ing substrate is explained by females usingdead branches, usually with no bark (Fig. 2).

The use of foraging locations with holesby males relates to their foraging techniques.Males regularly foraged at pits previouslyexcavated, and at crevices or natural holes(Table 2) using excavating and probing tactics.The most common foraging technique for

both sexes was debarking (Fig. 3). Femaleswere more mobile than males during foragingsequences and this was also reflected by theirmore intensive use of the scanning technique(Table 2), alternated with debarking. A higherinvestment of time in movements betweensubstrates by females could be attributed totheir tendency to explore small branches inthe crown. Body positions adopted byfemales were distributed among semi-hori-zontal (both types: with head down and withhead up) and vertical, while males foraged in avertical position most of the time (Fig. 3).

Foraging success rate, measured as: 1) thenumber of prey obtained per minute and 2)per number of used substrates, was almostequal for males and females (median: 0.07,range: 0–0.28 prey/min., Z = 268.0, P = 0.28;median: 0.47 range: 0–5 prey/substrate, Z =304.0, P = 0.7, respectively). In contrast,when comparing the type of prey, males cap-tured more larvae than females (unpaired Wil-coxon signed rank test P = 0.007, Fig. 4).However, we found no differences betweensexes for near-surface or indeterminate prey(unpaired Wilcoxon signed rank test, P > 0.05for both comparisons). For males, 82% per-cent of the successful prey capture events (n= 106) occurred after applying the probingtechnique, and 14% after debarking. Femalescaptured prey by probing in 73% of recordedevents (n = 108), and by debarking in 25% ofthe records. These techniques were usuallyalternated with excavating (especially bymales) or scanning (especially by females).

Niche breadth and overlap. Divergences betweensexes in foraging behaviors were related tolocation within substrate, substrate diameter,and relative height (Table 3), in accordancewith the previous results (Fig. 2). Femalesshowed greater flexibility in body position onthe foraging substrate (P = 0.002), explainedby their use of horizontal and vertical pos-tures.

TABLE 2. Components extracted by PCA from 18foraging variables and constructed for 31 malesand 21 females Magellanic Woodpeckers. Variableswere significant P < 0.001 in their correlations andloadings = |0.45| were used for interpretation(Aspey & Blankenship 1977).

Component

PC 1 PC 2Specific location TrunkSubstrate condition LiveRelative height Mid HigherForaging height 5.1–10 m 10.1–15 m > 15 mSubstrate diameter > 30 cm 10–15 cm < 10 cmFeature of foraging location Bark HoleForaging techniques Excavating Debarking Probing ScanningBody position Semi-horizontal head-up VerticalEigenvalueExplained variance (%)Cumulative %

-0.76

-

-0.700.77

-0.600.360.42

-0.670.690.36

--

0.35---

0.44-0.413.92323

0.34

0.36

-0.370.39

-0.34-

0.44

---

-0.520.75

0.500.320.60-0.42

-0.290.362.81639

537

DIFFERENTIATION BETWEEN SEXES OF MAGELLANIC WOODPECKERS

Social dominance. Sixteen events of displace-ment between adults were recorded (44% ofthe agonistic interactions among family mem-bers), all of which corresponded to males dis-placing their mates from tree trunks.Displacement by females corresponded toadults chasing away their offspring (same oropposite sex, n = 5). Two immature floaterfemales were also recorded displacing unre-lated immature individuals. Adult males werealso seen displacing their offspring: immaturebirds of the same (n = 11) or different (n = 4)sex.

DISCUSSION

Adult male and female Magellanic Woodpeck-ers forage in close proximity and use livinglenga trees. Despite this overlap in foraginghabitat, our data reveal that sexes partitiontheir shared main substrates (live trees) into

smaller niches: differences in resource useoccurred in microhabitat (parts of trees), preytype, and foraging behavior. The differentialutilization of microhabitats and behaviors wasaccompanied by significant morphologicaldifferences, with males being larger andhaving larger bills. Our observations of maledominance support the notion of a possiblecomplementary role of interference as amechanism to maintain the resource special-ization, and the foraging niche partitioningfound between the sexes.

According to how the woodpecker speciesexploit resources, they evolved in differentways. The most common pattern is thatmales are usually larger than females in severalmorphological variables (Short 1982), henceallowing sex-specific differences in resourceutilization, with females normally foragingon smaller substrates (e.g., Picoides spp.,Melanerpes spp., Short 1970b; Crimson-crested

FIG. 1. PCA based on the Magellanic Woodpecker foraging behavior and substrate use variables. PC1 ismainly represented by the specific location within the foraging substrate (trunk vs branch), foraging height(relative and absolute), and substrate diameter (i.e., greater than 30 cm or between 10–15 cm); while PC2 isdepicted primarily by the feature of the foraging location (with or without bark). Empty squared symbolsrepresent barycentres (means) of samples’ placement within the sex categories, with 95% confidence levelswithin a category given by ellipses.

538

CHAZARRETA ET AL.

FIG. 2. Proportion of time invested by males and females of the Magellanic Woodpecker at different for-aging substrates and characteristics of these foraging substrates: A) type of foraging substrates, B) tree sec-tion, C) specific substrate condition, D) characteristic of the foraging location, E) and F) foraging locationheight, and G) width of the specific substrate. Means and SE are shown.

539

DIFFERENTIATION BETWEEN SEXES OF MAGELLANIC WOODPECKERS

Woodpecker, Campephilus melanoleucos, Kilham1972; Great Spotted Woodpecker, Dendrocoposmajor, Osiejuk 1994; White-backed Wood-pecker, D. leucotos, Stenberg & Hogstad2004).

Intersexual differences in morphology aremore pronounced in insular woodpeckersthan in mainland forms of the same or relatedgenera because of a reduced interspecificcompetitive environment on islands (Selander1966, Bennett & Owens 2002). In mostwoodpeckers, sexual dimorphism in billlength does not exceed 10%, but in insularspecies, the bill of the male is, on average,markedly larger than that of the female (e.g.,27% in the Hispaniola Woodpecker, Melaner-pes striatus, Selander 1966). Magellanic Wood-peckers that occur in an insular-like environ-ment show a degree of sexual bill dimorphism(12.4%) that is intermediate between trulyinsular and continental species. This degree ofsexual dimorphism resembles populations ofGila (Melanerpes uropygialis) and Ladder-backed(Picoides scalaris) Woodpeckers occurring at thesouthern end of the peninsula of Baja Califor-nia, which apparently are more dimorphic inbill dimensions than are mainland continental

populations of the same species (Selander1966).

The apparently longstanding existence ofthe Magellanic Woodpecker without competi-tors in an isolated biome (Vuilleumier 1985)may explain its degree of sexual dimorphism.In support of this, Short (1970a) noted thatthe overlap in culmen length between thesexes is smaller in the Magellanic Woodpecker(18%) than in the Ivory-billed Woodpecker(Campephilus principalis) (36%) that coexistedwith potential competitors. Thus, the dimor-phism in bill length between males andfemales of the Magellanic Woodpecker is inaccordance with its possibly broadened“niche” (sensu Short 1970a) in the absence ofcompetitors. In parallel, we found that Magel-lanic Woodpecker males also consumed largerprey from living trunks with thicker bark,compared to small prey obtained on dead orliving branches by females. This is in agree-ment with the specialization hypothesis as amechanism for resource optimization in thisspecies, found both in other woodpeckers(Kilham 1972, Short 1982, Aulén & Lundberg1991) and in other bird species (Payne 1984,Winkler & Leisler 1985, Témeles et al. 2010).

FIG. 3. Proportion of time invested by males and females of the Magellanic Woodpecker in differentforaging behaviors (A), and foraging posture within the substrate (B). Means and SE are shown.

540

CHAZARRETA ET AL.

Although males and females overlappedin their niche breadth, the intensity of use ofdifferent substrate categories varied betweensexes of the Magellanic Woodpecker. Malesand females were separated primarily by for-aging location and secondarily by foragingbehaviors. The most important microhabitatintersexual segregation was the differentiationin trunk (mid and low parts of trees) andcrown (highest tree portion). The scarce useof the terminal parts of crowns possibly actsas an anti-predator strategy from knownpredators, such as Buteo spp. or Chilean Hawk(Accipiter chilensis) (McBride 2000, Chazarretaet al. 2011).

Concerning the niche breadth in behav-iors, males were more stereotyped thanfemales in the body position on substrates,spending much of their time at vertical posi-tions. Flexibility in body posture may repre-sent adaptations to the different subnichesused by female Magellanic Woodpeckers.Although both males and females mostly

used the debarking technique, males exca-vated deep holes more intensively thanfemales. Other studies concerning foragingtechniques of woodpeckers often found thatmales excavate foraging holes significantlymore than females. For example, this wasobserved for Melanerpes spp. (Selander 1966,Wallace 1974), Hairy Woodpecker (Picoides vil-losus, Kilham 1965), Arizona Woodpecker(Picodes arizonae, Ligon 1968), White-backedWoodpecker (Aulén & Lundberg 1991, Sten-berg & Hogstad 2004), and Great SlatyWoodpecker (Mulleripicus pulverulentus, Lam-mertink 2004). Vergara & Schlatter (2004)described the main foraging technique ofMagellanic Woodpeckers in pure and mixedlenga forests further south in Patagonia(54°S) as the excavating of holes into hard-wood followed by debarking, but differencesregarding foraging behavior between sexeswere not assessed. Discrepancies with ourstudy (where debarking was most frequenttechnique) may be due to observations biased

FIG. 4. Proportion of prey type captured by male and female Magellanic Woodpeckers. Means and SE areshown. Male and female sample sizes refer to the amount of individuals observed consuming prey. Preyitem sample size: 47 wood-boring larvae, 21 near-surface, and 38 indeterminate for males; 26 wood-boringlarvae, 32 near-surface, and 50 indeterminate for females. Different lowercase letters indicate statistical sig-nificance (P < 0.05) between sexes.

541

DIFFERENTIATION BETWEEN SEXES OF MAGELLANIC WOODPECKERS

over one of the sexes (males) in the formerstudy and/or due to different classification ofbehaviors.

Agonistic interactions observed forMagellanic Woodpeckers suggest interferencecompetition with dominance by males overfemales. Social dominance where the maleexcludes females from favored areas orresources has been proposed as a factor relat-ing to sexual dimorphism and differentiationin several woodpecker species (Ligon 1968,Hogstad 1978, Peters & Grubb 1983, Matthy-sen et al. 1991, Osiejuk 1994). Although wefound no difference between sexes in preycapture rates, we recorded more deep bigwood-boring larvae obtained through excavat-ing for males and more near-surface arthro-pods for females. This suggests males mayobtain more energy gain per consumed item,and thus an advantage of social dominance.Interestingly, the same pattern was describedfor the size/type of prey carried by each sexto active nests during the rearing period(Ojeda & Chazarreta 2006), when the adapta-tive value of acquiring larger prey becomes abenefit for the offspring survival.

Acting together, morphology and micro-habitat divergence may explain acquisition of

different food resources by each sex of theMagellanic Woodpecker. A causal relationshipbetween these two fields of sexual divergenceis among the most difficult hypotheses to testin biology (cause-effect). Now, with both themorphological and behavioral divergence pat-terns confirmed and quantified, it would beworth designing experimental researchaddressing the most recognized theories ofintraspecific niche differentiation, like past orpresent intraspecific competition, and isolatedevolution (or innate behavioral preference)(Slatkin 1984, Shine 1989).

ACKNOWLEDGMENTS

The following museums allowed access totheir collections: Museo Argentino de Cien-cias Naturales, Buenos Aires; Museo del LagoGutierrez, S. C. de Bariloche; Museo de Cien-cias Naturales de La Plata; Museo de la Pa-tagonia, S. C. de Bariloche; Museo Ornito-lógico de la Patagonia, El Bolsón; InstitutoCRICyT, Mendoza; Instituto Miguel Lillio, S.M. de Tucumán; Museo Nacional de HistoriaNatural de Chile, Santiago; Collections ofCornell University, Ithaca; American Museumof Natural History, New York; Carnegie

TABLE 3. Foraging niche dimensions for male (n = 31) and female (n = 21) Magellanic Woodpeckers.Z-values are from unpaired Wilcoxon signed rank test. Significant differences are given (*P < 0.05).

Foraging category Niche breadth Niche overlap

Male Female Z

Median Range Median RangeSubstrate typeSpecific locationSubstrate conditionSubstrate diameterRelative heightHeight above groundState of foraging location Foraging techniqueBody posture

0.000.270.010.180.320.430.290.530.01

0.0–0.190.0–0.50.0–1.00.0–0.740.0–0.970.0–0.910.0–0.710.19–0.870.0–0.4

0.000.280.400.340.220.320.440.520.13

0.0–0.490.0–0.520.0–1.00.0–0.910.0–0.890.0–0.950.0–0.920.19–0.760.0–0.47

376.5364.5416.5419.5246.5317.0397.5311.5489.0*

0.950.671.00.570.720.810.800.900.89

542

CHAZARRETA ET AL.

Museum of Natural History, Pittsburgh; Nat-ural History Museum of Los Angeles County;Michigan State University Museum, Lansing;Museum of Zoology, University of Michigan,Ann Arbor; Museum of Vertebrate Zoology,Berkeley; National Museum of Natural His-tory, Washington, D.C.; Peabody Museum ofNatural History, New Haven; Western Foun-dation of Vertebrate Zoology, Camarillo; andRoyal Ontario Museum, Toronto.

We are indebted to many people whocontributed to this study. The Diversidad-Neumeyer Refuge group helped with fieldlogistics. IDEA Wild, Birders’ Exchange,Vuilleumier Fund for Research on Neotropi-cal Birds, Manomet-K.S. Anderson Awardand Grant-In-Aid of Research from theNational Academy of Sciences, administeredby Sigma Xi, and The Scientific ResearchSociety (USA) donated equipment fundamen-tal to this research. The morphological mea-surements were completed thanks to thefinancial support from the American Museumof Natural History and the Cornell Lab ofLaboratory. This manuscript was greatlyimproved by comments of the editor andanonymous reviewers. Our study complieswith the current laws of Argentina and wasconducted under permits from the Adminis-tración de Parques Nacionales.

REFERENCES

Aebischer, N. J., P. A. Robertson, & R. E. Ken-ward. 1993. Compositional analysis of habitatuse from animal radio-tracking data. Ecology74: 1313–1325.

Airola, D. A., & R. H. Barrett. 1985. Foraging andhabitat relationships of insect-gleaning birds ina Sierra Nevada mixed-conifer forest. Condor87: 205–216.

Akaike, H. 1973. Information theory as an exten-sion of the maximum likelihood principle. Pp.267–281 in Petrov, B. N., & F. Csaki (eds). 2nd

International Symposium on Information The-ory. Akademiai Kiado, Budapest, Hungary.

Amadon, D. 1959. The significance of sexual dif-ferences in size among birds. Proc. Am. Phil.Soc. 103: 531–536.

Armesto, J. J., P. León-Lobos, M. T. K. Arroyo.1996. Los bosques templados del sur de Chile yArgentina: una isla biogeográfica. Pp. 23–28 inArmesto, J. J., C. Villagrán, & M. T. K. Arroyo(eds). Ecología de los bosques nativos de Chile.Editorial Universitaria, Santiago, Chile.

Askins, R. A. 1983. Foraging ecology of temper-ate–zone and tropical woodpeckers. Ecology64: 945–956.

Aspey, W. P., & J. E. Blankenship. 1977. Spidersand snails statistical tales: application of multi-variate analysis to diverse ethological data. Pp.75–120 in Hazlett, B. A. (ed.). Quantitativemethods in the study of animal behavior. Aca-demic Press, New York, New York, USA.

Aulén, G., & A. Lundberg. 1991. Sexual dimor-phism and patterns of territory use by theWhite–backed Woodpecker Dendrocopus leucotos.Ornis Scand. 22: 60–64.

Baldwin, S. P., H. C. Oberholser, & L. G. Worley.1931. Measurements of birds. Sci. Publ. Cleve-land Mus. Nat. Hist. 2: 1–165.

Bennett, P. M., & I. P. F. Owens. 2002. Evolution-ary ecology of birds. Oxford Univ. Press,Oxford, UK.

Bortolotti, G. R. 1984. Sexual size dimorphism andage–related size variation in Bald Eagles. J.Wildl. Managem. 48: 72–81.

Catry, P., R. A. Phillips, J. P. Croxall, K. E. Ruck-shuhl, & P. Neuhaus. 2005. Sexual segregationin birds: patterns, processes and implicationsfor conservation. Pp. 351–378 in Ruckstuhl, K.E., & P. Neuhaus (eds). Sexual segregation invertebrates: ecology of the two sexes. Cam-bridge Univ. Press, Cambridge, UK.

Chazarreta, L., V. S. Ojeda, & A. Trejo. 2011. Divi-sion of labour in parental care in the MagellanicWoodpecker Campephilus magellanicus. J. Orni-thol. 152: 231–242.

Crawley, M. J. 2007. The R Book. John Wiley andSons, Chichester, UK.

Grant, P. R. 1986. Ecology and evolution of Dar-win’s Finches. Princeton Univ. Press, Princeton,New Jersey, USA.

Hedrick, A. V., & E. J. Témeles. 1989. The evolu-tion of sexual dimorphism in animals: hypothe-

543

DIFFERENTIATION BETWEEN SEXES OF MAGELLANIC WOODPECKERS

ses and tests. Trends Ecol. Evol. 4: 136–138.Hogstad, O. 1978. Sexual dimorphism in relation

to winter foraging behavior and territorialbehaviour of the Three-toed WoodpeckerPicoides tridactylus and three Dendrocopos species.Ibis 120: 198–203.

Hurlbert, S. H. 1978. The measurement of nicheoverlap and some relatives. Ecology 59: 67–77.

Jackson, J. A. 1970. A quantitative study of the for-aging ecology of Downy Woodpeckers. Ecol-ogy 51: 318–323.

Kilham, L. 1965. Differences in feeding behaviorof male and female Hairy Woodpeckers. WilsonBull. 77: 134–145.

Kilham, L. 1972. Habits of the Crimson-crestedWoodpecker in Panama. Wilson Bull. 84: 28–47.

Lammertink, M. 2004. Grouping and cooperativebreeding in the Great Slaty Woodpecker. Con-dor 106: 309–319.

Lê, S., J. Josse, & F. Husson. 2008. FactoMineR: AnR package for multivariate analysis. J. Stat.Softw. 25: 1–18.

Ligon, J. D. 1968. Sexual differences in foragingbehavior in two species of Dendrocopos wood-peckers. Auk 85: 203–215.

Martin, P., & P. Bateson. 2008. Measuring behavior:an introductory guide. Cambridge Univ. Press,Cambridge, UK.

Matthysen, E., T. C. Grubb, & D. Cimprich. 1991.Social control of sex–specific foraging behaviorin Downy Woodpeckers, Picoides pubescens.Anim. Behav. 42: 515–517.

McBride, P. 2000. Magellanic Woodpecker (Campe-philus magellanicus) habitat selection in deciduousNothofagus forests of Tierra del Fuego. M.Sc.thesis, Western Washington Univ., Bellingham,Washington, USA.

Morrison, M. L. 1984. Influence of sample size andsampling design on analysis of avian foragingbehavior. Condor 86: 146–150.

Morse, D. H. 1980. Behavioral mechanisms in ecol-ogy. Harvard Univ. Press, Cambridge, Massa-chusetts, USA.

Newell, P., S. King, & M. Kaller. 2009. Foragingbehavior of Pileated Woodpeckers in partial cutand uncut bottomland hardwood forest. For.Ecol. Managem. 258: 1456–1464.

Noske, R. A. 1986. Intersexual niche segregation

among three bark-foraging birds of eucalyptforests. Aust. J. Ecol. 11: 255–267.

Ojeda, V. S. 2004. Breeding biology and socialbehavior of Magellanic Woodpeckers (Campe-philus magellanicus) in Argentine Patagonia. Eur.J. Wildl. Res. 50: 18–24.

Ojeda, V. S., & L. Chazarreta. 2006. Provisioning ofMagellanic Woodpecker (Campephilus magellani-cus) nestlings with vertebrate prey. Wilson. J.Ornithol. 118: 251–254.

Osiejuk, T. S. 1994. Sexual dimorphism in foragingbehavior of the Great Spotted WoodpeckerDendrocopos major during winters with rich cropsof Scotch pine cones. Ornis Fenn. 71: 144–150.

Pasinelli, G., & J. Hegelbach. 1997. Characteristicsof trees preferred by foraging Middle SpottedWoodpecker Dendrocopos medius in northernSwitzerland. Ardea 85: 203–209.

Payne, R. B. 1984. Sexual selection, lek and arenabehavior, and sexual size dimorphism in birds.Ornithol. Monogr. 33: 1–53.

Peters, W. D., & T. C. Grubb. 1983. An experimen-tal analysis of sex–specific foraging in theDowny Woodpecker, Picoides pubescens. Ecology64: 1437–1443.

Radford, A. N., & M. A. du Plessis. 2003. Billdimorphism and foraging niche partitioning inthe Green Woodhoopoe. J. Anim. Ecol. 72:258–269.

Richards, S.A. 2005. Testing ecological theory usingthe information–theoretic approach: examplesand cautionary results. Ecology 86: 2805–2814.

Ruckstuhl, K. E., & T. H. Clutton-Brock (eds).2005. Sexual segregation and the ecology of thetwo sexes. Cambridge Univ. Press, Cambridge,UK.

Ruckstuhl, K. E. 2007. Sexual segregation in verte-brates: proximate and ultimate causes. Integr.Comp. Biol. 47: 245–257.

Schoener, T. W. 1968. The Anolis lizards of Bimini:resource partitioning in a complex fauna. Ecol-ogy 49: 704–726.

Selander, R. K. 1966. Sexual dimorphism and dif-ferential niche utilization in birds. Condor 68:113–151.

Selander, R. K. 1972. Sexual selection and dimor-phism in birds. Pp. 180–230 in Campbell, B.(ed.). Sexual selection and the descent of man:

544

CHAZARRETA ET AL.

1871–1971. Aldine Publ. Co., Chicago, Illinois,USA.

Shine, R. 1989. Ecological causes for the evolutionof sexual dimorphism: a review of the evi-dence. Q. Rev. Biol. 64: 419–461.

Short, L. L. 1970a. The habits and relationships ofthe Magellanic Woodpecker. Wilson Bull. 82:115–129.

Short, L. L. 1970b. Reversed sexual dimorphism intail length and foraging differences in wood-peckers. Bird-Banding 41: 85–92.

Short, L. L. 1982. Woodpeckers of the world. Dela-ware Museum of Natural History. Foris Publi-cations, Cinnaminson, New Jersey, USA &Dordrecht, The Netherlands.

Slatkin, M. 1984. Ecological causes of sexualdimorphism. Evolution 38: 622–630.

Stenberg, I., & O. Hogstad. 2004. Sexual dimor-phism in relation to winter foraging in theWhite-backed Woodpecker (Dendrocopos leuco-tos). J. Ornithol. 145: 321–326.

Témeles, E. J., J. S. Miller, & J. L. Rifkin. 2010.Evolution of sexual dimorphism in bill size

and shape of hermit hummingbirds (Phaethor-nithinae): a role for ecological causation. Phil.Trans. R. Soc. London B Biol. Sci. 365: 1053–1063.

Vergara, P., & R. P. Schlatter. 2004. MagellanicWoodpecker (Campephilus magellanicus) abun-dance and foraging in Tierra del Fuego, Chile. J.Ornithol. 145: 343–351.

Vuilleumier, F. 1985. Forest birds of Patagonia:ecological geography, speciation, endemism,and faunal history. Pp. 255–304 in Buckley, P.A., M. S. Foster, E. S. Morton, R. S. Ridgely, &F. G. Buckley (eds). Neotropical Ornithology.Ornithol. Monogr. 36. The American Orni-thologists’ Union, Washington, D.C., USA.

Wallace, R. A. 1974. Ecological and social implica-tions of sexual dimorphism in five melanerpinewoodpeckers. Condor 76: 238–248.

Winkler, H., & B. Leisler. 1985. Morphologicalaspects of habitat selection in birds. Pp. 415–434 in Cody, M. L. (ed.). Habitat selection inbirds. Academic Press, San Diego, California,USA.