Ethology
Size-RelatedAdvantagesfor ReproductioninaSlightly
DimorphicRaptor: OppositeTrendsbetweentheSexes
FabrizioSergio*,JulioBlas*t,ManuelaG.Forero*,Jose´ AntonioDona´zar*FernandoHiraldo*
* DepartmentofAppliedBiology,Estacio´nBiolo´gicade Don˜ana,Seville,Spain
tDepartmentofBiology,UniversityofSaskatchewan,Canada
Correspondence
DrFabrizioSergio,DepartmentofApplied Biology,EstacionBiologicade Don˜ana,Avda. de MariaLuisas⁄n,PabellondelPeru´,41013
Seville,Spain.E-mail:;
Received:May9,2007
Initialacceptance: June13,2007
Finalacceptance: July12,2007 (K.Reinhold)
doi:10.1111/j.1439-0310.2007.01417.x
Abstract
Despite many comparativeanalyses and more than20proposedhypoth- eses, thereisstilllittle consensusover the factors promotingthe evolu- tion of reversedsexual dimorphism(RSD) in raptorial species. Furthermore,intrapopulationstudies, which may elucidatehow RSDis maintained once evolved, have been surprisingly scarce and only focused on ahandfulofspecies with mediumto high dimorphism.We examined thereproductiveadvantages associated with bodysizeand condition,measuredin the pre-layingperiod, in a diurnalraptor with lowsexual dimorphism,the blackkite(Milvusmigrans).Thestudy popu- lation was essentiallymonomorphic in size. For females, therewas an evidenceofreproductivebenefits associated with larger sizeand⁄orwith better body condition.Larger females had also access to higherquality partnersand territories,consistentwith the ‘intrasexualselection’ hypothesis, by which members of the larger sex enjoy size-related advantagesin intrasexualcompetitionover ascarce resource,the smal- ler sex. Opposite trends emergedfor males: smaller, leanermales had higherbreedingoutput,consistentwith the ‘small efficient male’ hypothesis.Overall, the fact that we observedin an essentiallymono- morphicpopulation the same selection pressures previously found in specieswith markeddimorphismsuggests that such reproductiveadvan- tages may becounterbalancedin our study model byopposite selection pressuresduringotherstages ofthe lifecycle.Thiscastssome doubts on the evolutionarysignificance of studies focusing exclusivelyon repro- ductionand callsforthe need ofmore comprehensiveanalyses incorpo- rating trait-mediateddifferentialsinsurvival and recruitment.
Introduction
In reptiles, birds and mammals, males are usually larger than females (Andersson 1994). However, even withinthese classes,therearenumerousexam- ples of reversed sexual dimorphism (RSD, males smaller than females) (Ralls 1976; Patton etal.
1994). Todate, most ofthe studies on the evolution- ary significance of this trait have focused on avian groups, especially birdsofprey. Forthese, more than
20 hypotheseshave been formulatedto explain the
evolutionandmaintenance ofRSD(e.g.reviews in Mueller& Meyer 1985; Mueller1986; Olsen Cockburn 1993; Massemin etal. 2000; Kru¨ger
2005). These have usually been classified into four
main categories. (1) Ecological hypotheses propose that RSDispromoted bysize-mediated niche parti- tioningbetweenthe sexes, which lowers intersexual competition forlimited resources(Newton1979; Mueller1986; Shine 1989). (2) Behaviouralhypoth- eses suggest that larger female sizefavours pair for- mationand maintenanceand allows the female to
force the male to provision food (Smith 1982; Mueller1986). (3) Sex-roledifferentiationhypothe- ses focus on the reproductive advantages of large femalesizeandsmallmalesize.Forexample, larger female size may promote higher fecundity, larger eggs,more effectiveincubation andhigherfasting endurance (big mother hypothesis), while smaller male size may lead to higher agility, lower flight costs and higherhunting efficiency (small efficient malehypothesis;Lundberg1986;Hakkarainen & Korpima¨ki1991; Ydenberg Forbes1991; Massemin etal.2000). (4) Sexual selectionhypothesesinclude intersexualselection,bywhich females prefer smal- lermalesbecause of thereproductiveadvantages associated with their smallsize,intrasexualselection, by which large female size affords advantagesin female competition for high quality males, which are in scarce supply (Safina 1984; Olsen Olsen
1987; Olsen Cockburn 1993; McDonald etal.
2005) and display agility, by which small male size confers benefits associated with mate attraction dur- ing acrobatic courtshipdisplays (Figuerola1999; Szekely etal.2000).
Despitethemany hypotheses formulatedandthe extensiveliterature,thereisstilllittleconsensusover the determinants of RSD in raptorial species. The only generalagreementseems to bethe expectation thatthetraitshould conferreproductive advantages associated with an increase in female size and⁄or a decreaseinmale size(Ydenberg Forbes1991).
Todate, therehave been two main types ofinves- tigationson the subject: (1) comparativestudies that examinephylogeniesto infer the evolutionofRSD (e.g. Olsen Cockburn1993; Kru¨ger2005); and (2) intrapopulationanalyses that explore the reproduc- tive advantagesof body size at the individual level and inform on the factors which may maintainRSD onceevolved (e.g.Newton1989;Hakkarainen & Korpima¨ki 1991; Massemin etal. 2000; McDonald et al. 2005). While there have been numerous comparative studies, intrapopulation analyses have been extremelyfew and have focused on ahandful ofspecies (Newton1989; HakkarainenKorpima¨ki
1991; Masseminetal. 2000; McDonald etal. 2005).
Furthermore,someofthem have demonstrated reproductiveadvantagesassociated with large female size, while others have reported only benefits of small male size, with limited consensusreachedto date. Finally, all of these investigationsfocused on specieswith mediumtohigh dimorphism,underthe assumptionthat selective pressuresto maintainRSD should bemore evidentinhighly dimorphicspecies. Thisraises the questionofwhethersimilar selective
pressures areinplaceornotinspecieswith low dimorphism.Therefore,there isan urgentneed for similar studies in distantly related groups, such as other birds of prey genera, and in species which exhibit low dimorphism. Below, we use the term
‘raptor’ to includeboth diurnaland nocturnal birds
ofprey.
Here,wereport theresultsof anintrapopulation analysis conducted on a raptor with low dimor- phism, the black kite (Milvusmigrans).Our dataset, which includes informationon body measurements and breedingperformance,doesnot allow ustocon- ductanextensive review ofalltheproposed RSD hypotheses.Therefore, wefocusour effortonthe evaluationofsixhypothesesforwhich the available data allow quantitativetests of specific predictions (see Cartry et al.1999 forsimilar reasoning).The hypotheses and associated predictions are listed in Table1 and described below: (1) the reproductive efforthypothesis suggests that larger females produce larger clutches (Reynolds 1972; Mueller Meyer
1985); (2) the starvation hypothesisstates that larger
females have a higher capability of accumulating body reserves which buffer them from periods of foodscarcity,resulting inmore efficient incubation (Lundberg1986). Thisleads tothreetestable predic- tions:(2a)female bodysizeandcondition willbe positively correlated;(2b) larger females willhave a higherhatchingsuccess; and (2c) the reserves accu- mulated by large females will be especially useful during periods of food stress, e.g. for birds laying early in the season when food is scarce and the weatherlessfavourable.Thisgeneratesthe expecta- tion of a negativecorrelationbetween female size and laying date (Masseminetal. 2000). The above two hypothesesreflect the bigmotherhypothesismen- tioned above. (3) The femalesupplementary feeding hypothesisproposes that larger female sizeallows lar- ger prey to be capturedby the female in the late- nestlingperiod, whenfemale raptors often start to huntadditionalprey for the chicks (Reynolds1972; Newton 1979;Massemin etal.2000). (4)Theintra- sexual selectionhypothesis predicts that larger females will gain access to higherqualitymales, leading to higheroverall breedingperformance(Olsen Olsen
1987; Olsen Cockburn1993). (5)Thesmallefficient
malehypothesisstates that smaller male sizewilllead to higherforaging efficiency and thus better breed- ing output (Reynolds1972; Hakkarainen Kor- pima¨ ki 1991; Hakkarainenetal. 1996). (6) Finally, the intersexual selectionhypothesissuggests that females should prefersmallmales, if smallmalesizeconfers reproductivebenefits (Safina 1984). Thishypothesis
Table1:Hypothesesforreversedsexualdimorphisminraptorialspe- ciesandtheirassociatedpredictions(re-adaptedfromMasseminetal.
2000).Onlythehypothesesthatcouldbetestedwiththedataavail- ableinthecurrentstudywereincluded.Forsomehypotheses,only somepredictionscouldbetestedwithourdataset.For example,the intrasexual competitionhypothesispredictsapositivecorrelation betweenfemalesizeandmalequalityandbetweenfemalesizeand breedingsuccess (seeIntroduction).Ourdatasetallowedustotest thesecondpredictionbutnotthefirst.Suchpartialtestshavebeen pointedoutinthetablefootnotes
Hypothesis Prediction
Largesizeoffemales
ReproductiveeffortClutchsizepositivelyrelatedtofemalesize
StarvationBodycondition,layingdateandhatching successpositivelyrelatedtofemalesize
Supplementaryfeeding Numberoffledglingspositivelyrelated tofemalesizea
IntrasexualselectionLargerfemalesgainaccesstohigher
qualitymales.Breedingsuccesspositively correlatedwithfemalesizeb
Smallmalesize
Small efficientmaleBreedingsuccessnegativelyrelated tomalesizeb,c,d
IntersexualselectionSmallermalesgainaccesstohigher qualityfemales.Breedingsuccess negativelyrelatedtomalesizeb,c
aThishypothesisgeneratesfurtherpredictionsthatcouldnotbe testedwiththeavailabledata(e.g.apositivecorrelationbetween femalesizeandhercontribution tochickprovisioningrates,which werenotmeasuredinthisstudy).
bThesehypothesesdonotpredictacorrelationwithanyspecificcom-
ponentofbreedingsuccess.
cThesmallmaleandtheintersexualselectionhypothesesgenerate thesamepredictionsintermsoftherelationshipbetweenmalesize andbreedingperformance.
dThishypothesisgeneratesfurtherpredictionsthatcouldnotbe
testedwiththeavailabledata(e.g.higherforagingsuccessbysmaller males).
thus builds on the previousone and predicts both a negative relationship between male size and both female qualityand breedingperformance(Massemin etal.2000).
StudySpecies
Theblack kite isadiurnal,medium-sized,migratory raptor with low dimorphism(see Results). Roledivi- sion betweenthe sexes follows the usual scheme for raptors: duringbreeding,the male provides most of the prey for the female and offspring, while the female performs mostoftheincubation, broodsthe nestlingsand guards the nest. Inthe late chick-rear- ingperiod, shemay start huntingtocomplementthe food provisionedby the male. The diet composition
in our study area isvery broad and dominated by rabbits, waterbirds and carrion (Veiga Hiraldo
1990; Vin˜uela Veiga1992).
StudyAreaandMethods
We studied black kites between1996 and 1998 in a
430km2 plot located in Don˜ana National Park (south-westernSpain). Thelandscapewascharacter- ized by seasonally flooded marshland, scrublands and grasslands and mobile sand dunesalong the sea shore (seeForero etal.1999, 2002 fordetails).
FieldProcedures
Adult black kites were trappedusing a cannonnet and markedwith a white plastic ring with a black, three-character alphanumeric code, which can be read byspotting-scopeswithoutdisturbingthe birds. Foreach trappedindividual,wemeasuredbodymass to the nearest 5g, tarsus length to the nearest
0.1mm, and wing length and tail length to the
nearest1.0 mm. Thesexofeach trappedbirdwas assessedbymolecular analysisofabloodsample (Ellegren 1996). After trapping,all breedingterrito- rieswere intensivelysearched formarkedadults. Wheneverpossible, nests ofthe markedindividuals were repeatedlyvisitedatleastthree times:(1)in incubationto assess egglaying and to record clutch size;(2) at hatching,to record hatchingsuccess and to backdate the hatching date based on a linear regressionrelatingthe lengthofthe eighthprimary ofthe oldest nestlingto age (Vin˜uela Bustamante
1992; Forero etal. 2002) and (3) when nestlings
were 40–45 d old to record the numberof fledged young.Pairs were classified as laying or non-laying only whentheir nests could bechecked weekly dur- ing the incubationperiod. Thisensuredthat we did not classifyasnon-layingfemales those that actually laid eggs but lost their clutch soon afterwards. In addition,whennests were located high on trees and the time for climbing themwas judged to be excessive, visits were minimizedduringthe incuba- tion and hatchingperiod to minimizedisturbance. For these reasons, sample size varies somewhat betweenanalyses.
StatisticalAnalyses
Analyses were restrictedtobirdscapturedduringthe pre-layingperiod for threereasons: (1) to minimize the statistical noise of seasonal variations in mass, which may bepronounced(e.g. Newtonetal.1983;
Wijnandts 1984; Meijer etal. 1989; Korpima¨ki
1990); (2) because this isconsideredas one of the most crucial periods ofthe breedingcycleforraptors, as females must accumulateenough body reserves forsubsequentsuccessful reproduction(e.g. Newton
1979; Meijer etal. 1989) and (3) because many of
the selective pressureshypothesisedtopromoteRSD are most evidentin the early stages ofthe breeding cycle (e.g. Safina 1984). Both males and females were captured on average 20d before laying (females: 19.8 ±2.1;males: 20.2 ±2.3).
Because univariatemetrics have been criticized as measuresofbody size, we estimatedsize by means of the first axis (PC1) of a principalcomponents analysis (PCA) built using tarsus, wing and tail length(McGillivray 1985;Rising& Somers 1989; FreemanJackson1990; McDonald et al. 2005). Except fordirect comparisonsbetweenthe sexes, we conductedseparatePCAsformales and females soas to accountfor shape differencesbetween the sexes (McGillivray 1985; Tornberg etal. 1999; Massemin etal. 2000; Phillips etal. 2002). For females, the PC1explained62% ofthe variationin sizeand had high positive loadings forwing length(r=0.85), tar- sus length(0.71) and tail length(0.47). For males, the PC1explained70% ofthe variationinsizeand had high positive loadings for wing length (0.68) and tarsus length (0.39) but not for tail length (0.08).
Toestimatebody condition,foreach sex we used the residuals ofaregressionofmass on PC1(hereaf- ter ‘mass residuals’) (Olsen Cockburn1993; Phil- lips et al. 2002), following the recommendations outlined by Green (2001) and Schulte-Hostedde etal.(2005).We added year asacategoricalcovari- atetotheregression because bodymassvariedsig- nificantlyormarginallysignificantly among years (females: F2,42 =4.84, p=0.013; males: F2,34 =2.56, p=0.092).
We examinedthe effects of body measurements onreproductive performance throughageneralized linear mixed model (GLMM) procedure(Littel etal.
1996), which allows the incorporationof indepen- dent variables as randomeffects in the model. All GLMMswere built throughthe Macro GLIMMIXof SAS(Littel etal.1996). Multipleand logisticregres- sion GLMMswere built separatelyforthe two sexes byfitting year (asarandomfactor), date ofcapture, body size and mass residuals as explanatory vari- ables,andbyusing asdependent variables: (1)the probability that eggswould belaid(dichotomous variable: 0=no subsequent laying, 1=eggs were subsequentlylaid); (2) laying date; (3) clutch size;
(4) hatchingsuccess (the probabilitythat ‡1 egg in the clutch hatched)and (5) the numberof young raised to fledging age. Each dependentvariable esti- mated adifferentcomponentofreproductiveperfor- mance, as the individualsdemonstratedtheir capability to reach successive stages ofthe breeding cycle(laying, incubationand chickrearing).The breedingprocess was so decomposedto gain addi- tional insights into some hypothesized advantagesof RSD(e.g. the capability oflarger females to lay lar- ger clutchesor incubateeggs more efficiently).All tests are two-tailed,statistical significance was set at a0.05, and allmeansaregiven ±1SE.
Results
There was neither differencebetweenthe sexes in tarsus, wing or tail length,nor in the PC1estimate ofbodysize(Table 2).Theonly measurestovary significantlybetweenthe sexes were body mass and massresiduals.Bodymasswas15% larger infemales (Table2).
ReproductiveEffort and StarvationHypotheses
None ofthese two hypotheseswas supported:laying date, clutch size and hatching success were not related to female size or female condition (GLMM, in all tests F1.58, p0.05). Furthermore,there was no significant correlationbetweenfemale body conditionand female size(r=0.06, p=0.68).
FemaleSupplementaryFeedingHypothesis
Thenumberofchicksraised tofledging agewaspos- itively related tofemale bodysize(Table3b,Fig.1a). However, given that direct estimates of the provi- sioning contribution of the female were not col- lected, support for this hypothesis could only be partial (Table1).
IntrasexualSelectionHypothesis
Larger females and females with highermass residu- als had a higher probabilityof subsequently laying eggs(Table3a, Fig.2a,b). Larger females also raised more young to fledging age (Table3b, Fig.1a,b). Furthermore,we have previouslyshownthat larger females had a higherprobabilitythan smaller ones of winning physical fights over territories and of gaining access to high qualitypartners(Sergio etal.
2007). These results are consistent with the intra- sexual selectionhypothesis.
Table2:Bodymeasures(xx±1 SE) of 38 maleand51femaleblackkitestrappedduring thepre-layingperiod
BodymeasureMalesFemalestp
Tarsuslength(mm)56.53±0.5456.74±0.43)0.310.58
Winglength(mm)445.1±2.2447.3±2.0)0.720.47
Taillength(mm)263.6±3.0258.0±1.771.700.11
Bodysize(PC1)a)0.07±0.170.05±0.13)0.550.58
Bodymass(g)796.1±15.3914.7±13.4)5.80<0.0001
Cuberootofbodymassb9.25±0.069.70±0.04)5.76<0.0001
Massresidualsc)64.9±14.845.8±12.6)5.67<0.0001
aEstimatedasthefirstaxisofaprincipalcomponentanalysis(PC1)oftarsus,wingandtaillength.
bThecuberootofbodymassisgivenforcomparisonwithlinearmeasurements.
cResidualsoftheregressionofbodymassonPC1(seeStudyAreaandMethods).
Table3:Effectofbodysizeandmassresidu- alsontheprobabilityofsubsequentlylaying eggs,layingdate,clutchsize,hatchingsuc- cessandthe numberof fledgedyoungin
Variable
Females
Parameter
estimate±SEFp
Percentageof explaineddeviance
maleandfemaleblackkites.Forbothsexes,
theeffectofbodymeasuresonlayingdate, clutchsizeandhatchingsuccesswasnever significant.Forclarity of presentation,only modelsinwhichsomeoftheexplanatoryvari- ablesattainedsignificanceareshown
a.Dependentvariable:probabilityoflayingeggsa(n=41)b
Bodysize(PC1)1.27±0.496.69<0.0119.3
Massresiduals / 10.85±4.69 / 5.36 / <0.03Date / 0.04±0.03 / 1.74 / ns
Yearc / 29.71±45.87 / 0.65 / ns
Intercept / )0.24±0.71 / – / –
b.Dependentvariable:numberoffledgedyoung(n=49)d
Bodysize(PC1) / 0.65±0.25 / 6.60 / <0.02 / 24.6Massresiduals / 2.89±1.84 / 2.48 / ns
Date / )0.01±0.02 / 1.55 / ns
Yearc / 0.11±0.28 / 0.38 / ns
Intercept / 0.91±1.43 / – / –
Males
c.Dependentvariable:probabilityoflayingeggsa(n=32)b
Bodysize(PC1) / )1.94±0.89 / 4.79 / <0.05 / 32.6Massresiduals / )0.03±0.02 / 3.54 / <0.08
Date / )0.04±0.06 / 0.71 / ns
Yearc / 6.38±8.39 / 0.76 / ns
Intercept / 7.13±6.03 / – / –
d.Dependentvariable:numberoffledgedyoung(n=34)d
Bodysize(PC1) / )0.19±0.12 / 2.42 / ns / 26.2
Massresiduals / )0.06±0.02 / 5.13 / <0.05
Date / )0.01±0.08 / 0.02 / ns
Yearc / 7.87±12.35 / 0.64 / ns
Intercept / 0.81±0.79 / – / –
SmallEfficientMale Hypothesis
ns,notsignificant.
aDichotomousvariable:1=thepairsubsequentlylaideggs;0=thepairdidnotsubsequently layeggs.
bGLMlogisticregressionwithbinomialerrorsandalogitlinkfunction.
cRandomeffect,testedbymeansofaZ-test.
dGLMmultipleregressionwithPoissonerrorsandalogarithmiclinkfunction.
(Table3, Fig.1b). These results supportedthe small male hypothesis.
For males, body conditionwas positively related to
body size (r=0.36, p=0.03). Smaller males had a higher probabilityof reachingthe incubation stage thanlarger ones (Table3, Fig.2a) and leanermales raised a higher number of young to fledging age
IntersexualCompetitionHypothesis
Despite the negativerelationshipbetween breeding success and male size reported above, previous
(a)
1.5
1.0
0.5
0.0
–0.5
Sex
•Males
°Females*
(a) 1.5
1.0
0.5
0.0
–0.5
*
Subsequently laid eggs
Did not lay eggs
*
–1.0
n =
1928121348
012
–1.0
Females
Sex
Males
(b)
140
90
40
–10
–60
–110
–160
–210
–260
Number of fledged young
Sex
•Males*
°Females
(b)130
100
70
40
10
–20
–50
–80
*
Females
Subsequently laid eggs
Did not lay eggs
Males
Sex
n = 1828
0
121348
12
Fig.2:Bodysize(a) andmassresiduals(b)of femaleblackkites
whichsubsequentlysucceededorfailedtolayeggs,andofmaleblack
Number of fledged young
Fig.1:Bodysize(a)andmassresiduals(b)ofmaleandfemaleblack kiteswhichraisedzero,one,ortwonestlingstofledgingage.Based on49femalesand34malesmeasuredinthepre-layingperiod.*Sig-
nificantrelationship.Errorbarsrepresent1SE.
kiteswhosepartnersubsequentlysucceededorfailedto layeggs. Basedon41 femalesand32 malesmeasuredinthepre-layingperiod. Bodysizewasestimatedasthefirstaxisofaprincipalcomponent analysis (PC1)oftarsus,wingandtaillength.Massresidualsarethe residualsofbodymass onPC1.*Significantrelationship.Errorbars
represent1SE.
analyses showedthat small male sizedid not afford accessto higherqualityfemales (Sergio etal.2007). Hence, the intersexual competition hypothesis was not supported.
Finally,for bothmalesandfemales,addingthe interactionbetweenyear and body measuresdidnot changeany ofthe above models.
Discussion
ReproductiveAdvantagesforFemales
Larger females and⁄or females with higher mass residuals inthe pre-layingperiod had ahigherprob- ability ofsubsequentlylaying eggs. In addition,lar- ger females eventuallyfledged more young.Other
authors have previouslystressed theimportancefor successful reproduction of body reserves accumu- lated byfemale raptors duringthe early stagesofthe breeding cycle (Newton etal. 1983; Village 1983; Hirons etal.1984; Hirons 1985; Meijer etal.1989). Such reserves are thoughttobuffer the female from potentialfood shortagesduringincubation and the earlyperiod ofchickrearing, preventingherfrom exposing the eggsor nestlingsto inclementweather or predation (Newton 1979; Hirons 1985; Meijer etal. 1989; Brodin Jo¨nsson 2003). This may be importantinour population, wherenest predation pressureishigh and wherebrood reductiontriggered byfoodscarcity occurs commonlyinthe early stages of the nestling period (Vin˜uela 2000; Sergio etal.
2005): larger reserves may allow females to devote
to the nestlingsa higherportionofthe food provi- sioned bythe male and tocontinuenest guardingin such period of peak demand (Brodin Jo¨nsson
2003). The fact that larger sizeand better condition
didnot lead tomore efficient incubationand tobet- ter hatchingsuccess, furthersuggests that the bene- fits of large size may be more evidentduringthe nestlingperiod ratherthaninincubation.
Ofthe RSDhypothesestested, therewas no sup- port for the big mother hypothesis (reproductive effort and starvationhypotheses),confirmingthe results ofprevious intrapopulationand comparative studies (e.g. Meijer etal.1989; HakkarainenKor- pima¨ki 1991; Massemin etal. 2000; Kru¨ger 2005) and inagreementwith the recentlyreviewednotion of raptorialspecies as ‘income breeders’ (Meijer & Drent 1999). Instead, results were atleastpartially consistent with thefemale supplementaryfeeding hypothesis. However,in the absence of direct esti- mates ofthe female contributiontochick provision- ing, it is difficult to differentiate such hypothesis fromotherones with similar predictions(Table 1). Therefore, of the hypotheses that could be tested with our dataset,results were most consistentwith the intrasexualselection hypothesis,bywhich the positive effect of body size on reproduction may have been related to the higher competitiveability bylarger females tosecure higherqualitymales and territories(Olsen Olsen 1987; Olsen Cockburn
1993; McDonald etal.2005). Inagreementwith this
idea, inour populationlarger females arrived earlier from the spring migration,occupied the best territo- ries, paired with higher quality males and fledged more young (Sergio etal.2007). Similar results have been reportedfor otherraptors (Korpima¨ ki 1990; McDonald et al.2005). Theintrasexualcompetition hypothesis, which isreceiving increasingsupportby intrapopulationstudies on RSD, has the advantage over others toexplain not only the directionofsex- ual dimorphismbut also its degree in species with both normal andreverseddimorphism(Olsen Cockburn1993; McDonald etal.2005).