Short- and long-term consequences of individual and territory quality in a long-lived bird

Oecologia (2009) 160:507–514
DOI 10.1007/s00442-009-1314-0
POPULATION ECOLOGY - ORIGINAL PAPER

Short- and long-term consequences of individual and territory quality in a long-lived bird

Fabrizio Sergio · Julio Blas · Raquel Baos · Manuela G. Forero · José Antonio Donázar · Fernando Hiraldo

Received: 27 December 2007 / Accepted: 17 February 2009 / Published online: 14 March 2009© Springer-Verlag 2009

Abstract Site-quality is a major determinant of Wtnessbut its eVect can be confounded by individual quality, arelationship that has been little studied in large, long-livedvertebrates. The Wtness eVects of quality estimates dependon the assumption of co-variation between individual andterritory quality and can be framed as Wve working hypoth-eses: no eVect on Wtness, exclusive eVect of individual qual-ity, exclusive eVect of site quality, and independent orinteractive eVects of the two. We explored such a frame-work using a medium-sized raptor, the black kite Milvusmigrans, as a model species. Individual and territory qual-ity co-varied, but the strength of the relationship variedacross diVerent estimates of individual quality (age, bodysize, or mass residuals). Short-term production of Xedglingswas related to the independent eVects of both individualand territory quality. However, longer-term production ofrecruits was related solely to territory quality. The disap-pearance of individual quality eVects over the long-termmay be caused by antagonistic selective pressures actingduring diVerent stages of the life cycle. Our results contrib-ute to a growing appreciation of the long-term Wtness-bene-Wts of advantages experienced in early life and highlight theimportance of a long-term perspective in studies assessingthe eVects of individual and territory quality. In our casestudy, prioritizing sites for conservation on the basis of

territory quality may be a feasible pathway to maintain theviability of the population. However, scenarios where sucha method could be ineYcient have been previouslyreported, suggesting caution in its application. More studiesare needed to understand the generality of the eYciency ofpriority-setting approaches based on site quality.

Keywords Early eVects · Habitat quality · Natal eVects · Site quality

Introduction

Site quality is a major determinant of Wtness for territorialspecies (Newton 1989; Komdeur 1992; Laaksonen et al.2004). Both empirical and experimental studies have dem-onstrated the eVect of territory quality on diVerent compo-nents of Wtness, such as mating success, breedingperformance, survival and recruitment (e.g. Alatalo et al.1986; Korpimäki 1988; Newton 1989; Siikamäki 1995;Pärt 2001; Krüger 2005; Hakkarainen et al. 2008). How-ever, in many populations the best individuals monopolisethe best territories and relegate lower quality individualsinto lower quality sites (Fretwell 1972; Rodenhouse et al.1997), generating co-variation between individual and terri-tory quality (e.g. Korpimäki 1990; Newton 1991; Marra2000; Candolin and Voigt 2001; Pärt 2001). Under this sce-nario, individual quality may mediate, oVset, exaggerate orconfound the Wtness eVects of territory quality.

The potential co-variation of individual and site qualitygenerates two controversies. Firstly, the importance of ter-ritory quality is diYcult to ascertain in the many studiesthat did not control for the eVect of individual quality—aproblem recognised by many authors (Carlsson 1998; Ser-gio and Newton 2003; Laaksonen et al. 2004; Carrete et al.

Communicated by Heli Siitari.

F. Sergio (&) · J. Blas · R. Baos · M. G. Forero · J. A. Donázar · F. HiraldoDepartment of Conservation Biology, Estacion Biologica de Doñana, CSIC, Avda. Americo Vespucio, s/n, 41092 Seville, Spaine-mail: [email protected]; [email protected]

123

508 Oecologia (2009) 160:507–514

2006). This has important conservation implications,because threatened species are often preserved by attempt-ing to manage the resources contained in their territories(e.g. Breininger and Carter 2003; Maguire 2006). However,if individual quality rather than site-resources determineWtness, such management actions may be ineYcient (unlessindividual and site-quality are positively related). Secondly,studies that simultaneously weighed the eVects of individ-ual and territory quality have found mixed results (seereferences below), so that the relative contribution of thetwo remains controversial. Furthermore, most studies thatsimultaneously tested the eVect of individual and territoryquality were conducted on small-sized, short-lived species,such as invertebrates, Wsh, amphibians, or passerine birds(e.g. Warner 1987; Roithmar 1994; Candolin and Voigt2001; Maguire 2006 and references therein). Only a hand-ful of studies have focussed on larger, long-lived species(Korpimäki 1988, 1990; Newton 1991; Ferrer and Bisson2003; Carrete et al. 2006).

The above issues may be examined by: (a) testing theassumption that individual and territory quality co-vary,and (b) exploring the relationship between Wtness compo-nents and both individual and territory quality. In particu-lar, quality estimates can aVect Wtness in Wve ways, whichcan be considered as working hypotheses: (b1) territoryquality is the only determinant of Wtness (Alatalo et al.1986; Warner 1987; Pärt 2001; Przybylo et al. 2001); (b2)individual quality is the only determinant of Wtness (Sim-mons 1988; Horn et al. 1993; Carrete et al. 2006); (b3) bothindividual and territory quality independently aVect Wtness(Hill 1988; Bart and Earnst 1999); (b4) individual and terri-tory quality aVect Wtness but their eVect is interactive (Lõh-mus and Vali 2004); and (b5) Wtness is related to neitherindividual nor territory quality (Leonard and Picman 1988;Formica et al. 2004).

Here, we use the above framework to assess the eVectof individual and territory quality on the production ofoVspring and recruits in a medium-sized, long-lived diur-nal raptor, the black kite Milvus migrans. In our studyarea, black kites occupy territories of 50–200 m radiuscentred on a nest-site and forage over nearby undefendedhunting areas (Forero et al. 1999). They actively select ter-ritories located near marshland (Sergio et al. 2005), wherethey feed mostly on wetland birds, freshwater crabs andrabbits Oryctolagus cuniculus, which are also most abun-dant near marshland (Hiraldo et al. 1990; Fernández et al.2003). Breeding success increases with territory-levelavailability of wetlands, but also varies with estimates ofindividual quality, such as age, size and body condition(Blas et al. 2009; Sergio et al. 2005, 2007a, b). Therefore,this population represents a good model for the investiga-tion of the Wtness consequences of individual and territoryquality.

Materials and methods

Study area

We studied black kites between 1996 and 2007 in a430 km2 plot located in Doñana National Park (south-west-ern Spain). The landscape was characterised mainly by sea-sonally Xooded marshland, scrublands, grasslands, andmobile sand dunes along the sea shore (see Forero et al.1999, 2002 for details).

Field procedures

Between 1996 and 1998, adult black kites were trappedusing a cannon net baited with carrion and marked with awhite plastic ring with a black, three-character alphanu-meric code, which can be read by spotting-scopes withoutdisturbing the birds. For each trapped individual we mea-sured body mass to the nearest 5 g, tarsus length to thenearest 0.1 mm, and wing length and tail length to the near-est 1.0 mm. The sex of each trapped bird was assessed bymolecular analysis of a blood sample (Ellegren 1996).Immediately after trapping, all breeding territories wereintensively searched for marked adults. Nests of the markedindividuals were visited when nestlings were 40–45 daysold to record the number of Xedged young and to ring thechicks (further details in Forero et al. 1999, 2002; Sergioet al. 2007a). Territories continued to be searched formarked birds up to 2007, as part of a still ongoing long-term study on this population. This allowed us to identifywhich of the breeders trapped between 1996 and 1998 pro-duced oVspring that subsequently became breeders them-selves (i.e. recruits) between 1997 and 2007. Because theage of Wrst reproduction in this population ranges from 1 to7 years old, and 80% of the recruits begin breeding whenless than 5 years old (Blas 2002), we estimate that all thenestlings included in our sample (i.e. produced up until1998) and which subsequently recruited into the breedingpopulation had done so by 2004 (i.e. 3 years before 2007,the end of the currently available ring-readings).

Individual quality

We used age, body size and an index of body condition asestimates of individual quality. These are the measuresmost commonly employed in studies of this kind (e.g. Alat-alo et al. 1986; Simmons 1988; Horn et al. 1993; Siikamäki1995; Bart and Earnst 1999; Carrete et al. 2006). Becauseunivariate metrics have been criticised as measures of bodysize, we estimated size by means of the Wrst axis (PC1,hereafter “body size”) of a principal components analysis(PCA; Tabachnick and Fidell 1996) built using tarsus, wingand tail length (Rising and Somers 1989; Freeman and

123

Oecologia (2009) 160:507–514 509

Jackson 1990). The PC1 explained 62% of the variation insize and had high positive loadings for wing length(r = 0.87), tarsus length (0.52) and tail length (0.68). Asbody mass varied with year, breeding stage and body size(Sergio et al. 2009), we standardised it by using the residu-als of the regression of body mass on year, breeding stageand body size as an index of body condition (e.g. Schulte-Hostedde et al. 2005). We assume that such a procedureremoved or lowered the bias caused by estimating individ-ual quality based on a single trapping event per individual.

Territory quality

Based on previous work on the species (see below), weestimated territory quality as the Wrst component (PC1,hereafter “territory quality”) of a PCA conducted on twovariables: the percentage of years that a territory was occu-pied (arcsin-transformed; factor loading: r = 0.76) and thepercentage availability of wetland-habitats within 1 km ofthe nest (arcsin-transformed; factor loading: r = 0.73). Theoccupation rate of a territory has been shown recently to bea consistent measure of territory quality in a wide range ofspecies, including black kites (review in Sergio and Newton2003). Territory occupation rates refer to the period 1989–2000 (mean years of occupation § SE = 9.4 § 2.6). This islarger than the average territory tenure of an individual inour population (authors’ unpublished data). Therefore,measured occupation rates should not be biased by long-term settlement on a territory by a single individual (e.g.Krüger 2005). Wetland availability ranged between 0 and89% (mean § SE = 29.1 § 20.8), and was chosen because:(1) black kites have been repeatedly shown to be stronglyassociated with wetlands (Sergio et al. 2003a, b, 2005,2007a); (2) their diet is usually dominated by wetland prey(Hiraldo et al. 1990; Sergio et al. 2003b); and (3) their for-aging eVort and performance peaks over wetlands (Hiraldoet al. 1990; Sergio et al. 2003a; unpublished telemetrydata). Therefore, wetlands can be considered as optimalforaging habitat, and sites with higher wetland availabilityare expected to be of higher quality.

Statistical analyses

Data were available for 131 breeding individuals (62 malesand 69 females). Age was available for a sub-sample of 62individuals (30 males and 32 females) ringed as nestlingsand later captured as breeders. We examined the eVects ofindividual and territory quality through a generalised lin-ear mixed model (GLMM) procedure (Littel et al. 1996),which allows the incorporation of independent variablesas random eVects in the model. All GLMMs were builtthrough the Macro GLIMMIX of SAS (Littel et al. 1996)and through a backward elimination procedure, following

Crawley (1993): all explanatory variables were Wtted to themodel, extracted one at a time from this maximal model,and the associated change in model deviance was assessedby an F-test. We tested the assumption of co-variationbetween individual and territory quality by Wtting year (as arandom factor), sex, estimates of individual quality (age,body size and mass residuals) and their interaction to aGLMM (with normal errors and an identity link function)with territory quality as the dependent variable. Hypothesesb1–b5 were tested by building GLMMs with year (as a ran-dom factor), individual quality of the parents (age, bodysize, and mass residuals) and territory quality as explana-tory variables, and by using as dependent variables: (1) thenumber of young raised to Xedging age (GLMM with Pois-son errors and a logarithmic link function), and (2) theprobability that at least one of the oVspring produced in anesting attempt was later recruited into the breeding popu-lation (GLMM with binomial errors and a logit link func-tion). Age was always calculated as the chronological ageof the parent in the year of birth of an oVspring or recruit.Hereafter, we refer to the eVect of individual and site-qual-ity on the production of Xedglings as a “short-term eVect”and to the eVect of individual and site-quality on the pro-duction of recruits as a “long-term eVect”. There was nopseudoreplication of territories or individuals in any of thedatasets (i.e. each territory or individual appeared once ineach analysis).

Additional explanatory variables were added to theabove models as follows. Firstly, raptorial species showreversed sexual dimorphism, with frequently demonstratedadvantages for reproduction associated with large size andlarge mass residuals for females and small size and smallmass residuals for males (e.g. Massemin et al. 2000;McDonald et al. 2005; Sergio et al. 2007b and referencestherein). Because of such opposing trends between thesexes, we always added the interaction term between sexand body size and between sex and mass residuals to the setof explanatory variables, so as to estimate individual qual-ity in a more biologically meaningful manner. Secondly, totest hypotheses “b3” and “b4” above (independent vs inter-active eVects of individual and territory quality), we addedas explanatory variables: (1) the interaction between terri-tory quality and each of the three estimates of individualquality; and (2) all the three-way interactions between terri-tory quality, sex and estimates of individual quality.Finally, because age was available only for a sub-sample ofindividuals ringed as nestlings, we re-ran all the modelstwice, including and excluding the eVect of age (and itsassociated interactions) from the set of potential explana-tory variables.

We did not employ capture–recapture models to exam-ine the eVects on recruitment for three reasons (see alsoBreininger and Carter 2003; McDonald et al. 2005; Maguire

123

510 Oecologia (2009) 160:507–514

2006): (1) re-sighting probability in this population isnearly one (Forero et al. 1999), and the detectability ofbreeding individuals is extremely high because all territo-ries are repeatedly visited throughout the breeding sea-son—the open landscape allows rings to be read from veryfar away, and only a handful of birds known to be markedeventually escape detection; (2) natal and breeding dis-persal is extremely low in this population and emigration isvirtually absent (Forero et al. 1999, 2002), reducing thelikelihood of some individuals being recruited elsewhere;(3) all oVspring from the last cohort of the trapping periodof study (1998) should have recruited by 2005 (because themaximum observed age at Wrst breeding is 7 years; Blaset al. 2009) and would have had at least two additionalyears of “exposure” to capture. We consider it highlyunlikely that a breeding bird could have escaped detectionfor two consecutive years. All tests are two-tailed, statisti-cal signiWcance was set at � < 0.05, and all means are given§1 SE.

Results

Assumption: do individual and territory quality co-vary?

For both males and females, better territories were occupiedby individuals with larger mass residuals (Table 1: modela). Independently of sex, better territories were alsooccupied by larger-sized individuals, but this relationshipwas only marginally signiWcant (GLMM: parameterestimate § SE = 0.03 § 0.001; t = 1.76, P = 0.08). Therelationship between age and territory quality was positivebut not signiWcant. Therefore, better territories were occu-pied by individuals with larger mass residuals and,possibly, larger body size. The assumption was partiallysupported.

Hypotheses b1–b5: Wtness eVects of individual and territory quality

Short-term oVspring production was related to territoryquality, the interaction between parental sex and body size,and the interaction between parental sex and mass residuals(Table 1: model b; Fig. 1a). As for the Wrst interaction termin the model, oVspring production increased with bodysize in females, while the relationship was more complexfor males (Fig. 1b). For the latter, oVspring productionincreased with body size when considering only successfulindividuals (i.e. those raising at least one chick to Xedg-ling), but the body size of unsuccessful males was largerthan that of males raising only one nestling to Xedging age(Fig. 1b). As for the second interaction term in the model,oVspring production declined with mass residuals for males

but not for females (Fig. 1c). Adding age to the set ofexplanatory variables resulted in the same model as above,but the main eVect of age was also signiWcant (Table 1:model c; Fig. 1d). Therefore, short-term oVspring produc-tion increased with both individual and territory quality andtheir eVect was independent, lending support to hypothesisb3.

There was a positive correlation between the number ofXedglings produced by a nesting attempt in the short termand the number of recruits produced by the same nestingattempt in the long term (rs = 0.36, P < 0.0001). Productionof oVspring that eventually recruited into the breeding pop-ulation was related solely to territory quality (Table 1:model d). Therefore, longer-term eVects lent support tohypothesis b1.

Discussion

In our study population, estimates of individual quality co-varied with territory quality, but the relationship was com-plex. For example, better sites tended to be occupied bymales with larger mass residuals. However, once a territorywas occupied, leaner males tended to produce moreoVspring. Such apparent inconsistency agrees with our pre-vious suggestion that antagonistic selective pressures mayact in diVerent stages of the life cycle (Sergio et al. 2007b).For example, large size or mass residuals were shown to beimportant in competition over territories during the earlystages of the breeding season (Sergio et al. 2007a), but inlater stages the hunting ability of some males may be reX-ected by their capability to reduce their intake rate (and thustheir mass) in favour of provisioning the female andoVspring (Sergio et al. 2007b). Such complexity underlinesthe diYculty of capturing individual quality by means of asingle univariate measure. This diYculty was furtheremphasised by the fact that morphometric measures of indi-vidual quality, such as size and mass residuals, had oppo-site eVects on the productivity of males and females, whichmay be common in raptorial species (Massemin et al. 2000;McDonald et al. 2005; Sergio et al. 2007b and referencestherein). This highlights the importance of an in-depthunderstanding of the general ecology of the study speciesbefore devising biologically meaningful estimates of indi-vidual quality.

Independently of the type of measure of individual qual-ity employed, short-term oVspring production was indepen-dently aVected by both individual and territory quality.Such independent eVects have been previously reported byboth empirical and experimental studies (Hill 1988;Siikamäki 1995; Bart and Earnst 1999; Ferrer and Bisson2003). However, over the longer term, the eVect of parentalquality seemed to vanish, and the production of recruits

123

Oecologia (2009) 160:507–514 511

was related solely to territory quality. The apparent lack oflong-term importance of parental quality may be caused bydiVerent factors. Firstly, the production of Xedglings maybe easily controlled by parents during the nesting periodwhen these provide extensive parental care, such as in mostbirds, but recruit production also depends on the survivalprobabilities of the oVspring during natal dispersal, whichis usually and unavoidably out of parental control. Sec-ondly, such survival probabilities may be more dependenton the oVspring’s own quality and this is unlikely to befully inherited by the parents. For example, mean heritabil-ity of morphological traits such as body size across numer-ous animal taxa is 0.4, which is high but far from perfect(RoV 1997). Thirdly, even if such traits were perfectly heri-table, the characteristics that promote high reproductionmay not be the same that promote high survival. For exam-ple, smaller males may be more productive but survive lesswell (e.g. McDonald et al. 2005). Such antagonistic selec-tive pressures may outweigh the eVect of individual qualityover the long term.

On the other hand, long-lasting eVects of territory qual-ity could be expected for various reasons. Firstly, the bestterritories were typically located at the border of marshland,where food is most abundant and readily accessible, withpotentially positive eVects on the Xedglings’ nutritionalcondition. The latter may be especially important because,shortly after Xedging, the juveniles face a long-distancemigration of more than 2,500 km (unpublished satellite-telemetry data). Secondly, higher food concentrations nearbetter territories may allow the juveniles to better practicetheir hunting skills before the migratory departure. Thirdly,in this population some adults are known to abandon theiroVspring and depart for migration (Bustamante and Hiraldo1990; unpublished satellite-telemetry data). OVspring aban-doned in good nutritional condition and in food-rich territo-ries may fare better than oVspring in depleted condition andin food-poor territories. The above three reasons may pro-vide individuals born in good territories with a “head-start”in early life, with consequent repercussions on subsequentsurvival and long-term performance. Whatever the mechanism

Table 1 Generalised linear mixed model (GLMM) regres-sions testing the intercorrelation between individual and territory quality (model a) and the eVect of individual and territory qual-ity on the production of oVspring (model b, c) and of recruits (model d) in a population of black kites in Doñana National Park (Spain)

Variable Parameter estimate § SE

F P % Deviance explained

Assumption: correlation between individual and territory quality

a Dependent variable: territory quality (n = 131)a

27.2

Mass residuals 0.03 § 0.001 7.11 <0.01

Intercept 0.03 § 0.24 – –

Hypotheses b1–b5: eVects of individual and territory quality

b Dependent variable: oVspring productionb (n = 131)c

33.0

Territory quality 0.40 § 0.10 14.9 <0.001

Body size 0.35 § 0.16 0.65

Mass residuals 0.004 § 0.001 0.22 ns

Sex ¡0.28 § 0.29 0.95 ns

Interaction: sex £ body size ¡0.53 § 0.22 5.67 <0.02

Interaction: sex £ mass residuals ¡0.01 § 0.002 8.88 <0.01

Intercept ¡0.66 § 0.02 – –

c Dependent variable: oVspring productionb (n = 62)c,d

30.4


Age 0.84 § 0.36 5.49 <0.03

Body size 0.48 § 0.31 0.08 ns

Mass residuals ¡0.001 § 0.003 0.19 ns

Sex ¡0.25 § 0.56 0.20 ns

Interaction: sex £ body size ¡1.08 § 0.39 7.57 <0.01

Interaction: sex £ mass residuals ¡0.01 § 0.003 6.11 <0.02

Intercept ¡2.48 § 0.78 – –

d Dependent variable: production of recruits (n = 124)e

27.3


Intercept ¡2.93 § 0.78 – –

a GLMM multiple regression with normal errors and an iden-tity link function (Littel et al. 1996)b Number of young raised to Xedging agec GLMM multiple regression with Poisson errors and a loga-rithmic link function (Littel et al. 1996)d Same as model b, but with the addition of age (and its associ-ated interactions) to the set of potential explanatory variablese GLMM multiple regression with binomial errors and a logit link function (Littel et al. 1996)

123

512 Oecologia (2009) 160:507–514

behind the observed associations, our results contribute to agrowing appreciation of the long-term Wtness-beneWts ofadvantages experienced in early life (Humphries and Bou-tin 2000; Messina and Fox 2001; Krüger 2005; van de Polet al. 2006). They also highlight the importance of a long-term perspective in studies on the Wtness contribution ofindividual and territory quality.

Implications for conservation

When preserving a species or population, various authorshave stressed the importance of planning habitat manage-ment and restoration at the territory level, because this isthe functional demographic unit within the landscape (e.g.Breininger and Carter 2003). Such an approach frequentlyleads to the proposal of identifying high quality territoriesand prioritising their conservation management (e.g. Krü-ger and Lindström 2001; Sergio and Newton 2003). Theassumption behind these approaches is that most of theindividuals contributed to future generations will be pro-duced in a minority of higher quality sites (Newton 1989).The results from our study support this idea. For example,in our population, the hypothetical protection of 40 territo-ries based on site-quality (i.e. the topmost one-third of thesites) would have led to the protection of 70% of the futurerecruits. This compares with protection of 30% of recruitsbased on a random selection of 40 territories.

On the other hand, prioritisation based on site quality, asshown above, may be ineYcient if individual quality is theonly determinant of Wtness, if individual and site quality arenot correlated, and if their Wtness eVects are independent orinteractive. The question is then how general is the patternthat we found here. Many studies have reported a positiveassociation between individual and territory quality (e.g.Korpimäki 1990; Newton 1991; Marra 2000; Candolin andVoigt 2001; Pärt 2001), and a positive Wtness eVect of terri-tory quality (e.g. Alatalo et al. 1986; Pärt 2001; Przybyloet al. 2001; Krüger 2005). However, other studies havereported a nil eVect of territory quality (e.g. Horn et al.1993; Carrete et al. 2006), the absence of co-variationbetween individual and territory quality (e.g. Hill 1988;Matthysen 1990; Ens et al. 1995; Lõhmus and Vali 2004),and their independent or interactive eVects (e.g. Hill 1988;Bart and Earnst 1999; Lõhmus and Vali 2004). Therefore,evidence does exist that, in some cases, such prioritisationapproaches may not be expected to work well. A furtherproblem is that previous prioritisation proposals have beenfrequently based on analyses focussing exclusively onXedgling production (e.g. Krüger and Lindström 2001; Ser-gio and Newton 2003; Carrete et al. 2006). Our resultsshow that such analyses may not always reXect longer-termeVects and thus should be treated with caution.

Overall, detailed long-term data will often be unavail-able in conservation programs targeting threatened species

Fig. 1 OVspring production in relation to a territory quality, b body size, c mass residuals, and d age in male and female black kites in Doñana National Park (Spain). Data for males and females were pooled in the graphs when sex-eVects were non-signiWcant in the models of Table 1

2204168N =

Number of fledged young3210

Ter

riory

qua

lity

(PC

1)

1.5

1.0

.5

0.0

-.51122 236 91931N =


Bod

y si

ze

1.0

.5

0.0

-.5

-1.0

Sex

Males

Females

1122 236 91932N =


Mas

s re

sidu

als

0

120

80

40

-40

-80

-120

-160

Sex

Males

Females

111337N =


Age

(ye

ars)

10

9

8

7

6

5

a b

c d

123

Oecologia (2009) 160:507–514 513

because of time and resource constraints. In these frequentcases, we suggest that priorities be planned with caution ina context-dependent manner. In particular, we suggest: (1)to conduct a pilot study to gain some basic understanding ofthe association between territory quality and some measureof performance, at least over the short-term; (2) to employan adaptive management perspective, so as to be able toalter previous plans as new data become available; and (3)to include in the set of priority sites a minor proportion ofrandomly selected sites, so as to attenuate the damage ofpotential mistakes. Some authors have stressed the impor-tance of protecting a portion of lower quality sites becausein some populations these are “training or queuinggrounds” for young individuals while they wait for betteroptions (e.g. Matthysen 1990; Heg et al. 2000).

Acknowledgements We thank O. Krüger, L. Marchesi, H. Siitari,H. Ylonen and two anonymous referees for comments on a previousdraft of the manuscript and F.G. Vilches, G. García, J.M. Grande,S. Cabezas, A. Sánchez and A.M. Guerrero for help in the Weld. Thestudy complies with the current Spanish laws on animal trapping andhandling. Part of this study was funded by the research projectsPB96-0834 of the Dirección General de Investigación CientíWca yTecnológica, CGL2008-01781/BOS of the Ministerio de Ciencia eInnovación, JA-58 of the Consejería de Medio Ambiente de la Junta deAndalucía and by the Excellence Project RNM 1790 and RNM 03822of the Junta de Andalucía.

References

Alatalo RV, Lundberg A, Glynn C (1986) Female pied Xycatchers chooseterritory quality and not male characteristics. Nature 323:152–153

Bart J, Earnst SL (1999) Relative importance of male and territoryquality in pairing success of male rock ptarmigan (Lagopusmutus). Behav Ecol Sociobiol 45:355–359

Blas J (2002) Edad y reproducción en el Milano Negro (Milvusmigrans). PhD Thesis, University of Madrid

Blas J, Sergio F, Hiraldo F (2009) Age-related improvement in repro-ductive performance in a long-lived raptor: a cross-sectional andlongitudinal study. Ecography (in press)

Breininger DR, Carter GM (2003) Territory quality transitions andsource-sink dynamics in a Florida Scrub-jay population. EcolAppl 13:516–529

Bustamante J, Hiraldo F (1990) Factors inXuencing family rupture andparent-oVspring conXict in the Black Kite Milvus migrans. Ibis132:58–67

Candolin U, Voigt H (2001) Correlation between male size andterritory quality: consequence of male competition or predationsusceptibility? Oikos 95:225–230

Carlsson A (1998) Territory quality and feather growth in the White-backed Woodpecker Dendrocopos leucotos. J Avian Biol29:205–207

Carrete M, Sánchez-Zapata JA, Tella JL, Gil-Sánchez JM, Moleón M(2006) Components of breeding performance in two competingspecies: habitat heterogeneity, individual quality and density-dependence. Oikos 112:680–690

Crawley MJ (1993) GLIM for ecologists. Blackwell Science, OxfordEllegren H (1996) First gene on the avian W chromosome (CHD) pro-

vides a tag for universal sexing of non-ratite birds. Proc R SocLondon Ser B 263:1635–1641

Ens BJ, Weissing FJ, Drent RH (1995) The despotic distribution anddeferred maturity: two sides of the same coin. Am Nat 146:625–650

Fernández N, Delibes M, Palomares M, MladenoV DJ (2003) Identify-ing breeding habitat for the Iberian lynx: inferences from a Wne-scale spatial analysis. Ecol Appl 13:1310–1324

Ferrer M, Bisson I (2003) Age and territory-quality eVects on fecundityin the Spanish Imperial Eagle (Aquila adalberti). Auk 120:180–186

Forero MG, Donázar JA, Blas J, Hiraldo F (1999) Causes and conse-quences of territory change and breeding dispersal distance in theBlack Kite. Ecology 80:1298–1310

Forero MG, Donázar JA, Hiraldo F (2002) Causes and Wtness conse-quences of natal dispersal in a population of Black Kites. Ecology83:858–872

Formica VA, Gonser RA, Ramsay S, Tuttle EM (2004) Spatial dynam-ics of alternative reproductive strategies: the role of neighbors.Ecology 85:1125–1136

Freeman S, Jackson WM (1990) Univariate metrics are not adequate tomeasure avian body size. Auk 107:69–74

Fretwell SD (1972) Populations in a seasonal environment. PrincetonUniversity Press, Princeton

Hakkarainen H, Korpimäki E, Laaksonen T, Nikula A, Suorsa P (2008)Survival of male Tengmalm’s owls increases with cover of oldforest in their territory. Oecologia 155:479–486

Heg D, Ens BJ, van der Jeugd HP, Bruinzeel LW (2000) Local domi-nance and territorial settlement of nonbreeding Oystercatchers.Behaviour 137:473–530

Hill GE (1988) Age, plumage brightness, territory quality and repro-ductive success in the Black-headed Grosbeak. Condor 90:379–388

Hiraldo F, Veiga JP, Mañez M (1990) Growth of nestling black kitesMilvus migrans: eVects of hatching order, weather and season.J Zool (Lond) 222:197–214

Horn AG, Dickinson TE, Falls JB (1993) Male quality and song reper-toires in western meadowlarks (Sturnella neglecta). Can J Zool71:1059–1061

Humphries MM, Boutin S (2000) The determinants of optimal littersize in free-ranging red squirrels. Ecology 81:2867–2877

Komdeur J (1992) Importance of habitat saturation and territoryquality for evolution of cooperative breeding in the Seychelleswarbler. Nature 358:493–495

Korpimäki E (1988) EVects of territory quality on occupancy, breedingperformance and breeding dispersal in Tengmalm’s owl. J AnimEcol 57:97–108

Korpimäki E (1990) Body mass of breeding Tengmalm’s Owls Aego-lius funereus: seasonal, between-year, site and age-related varia-tion. Ornis Scand 21:169–178

Krüger O (2005) Age at Wrst breeding and Wtness in goshawk Accipitergentilis. J Anim Ecol 74:266–273

Krüger O, Lindström J (2001) Habitat heterogeneity aVects populationgrowth in goshawk Accipiter gentilis. J Anim Ecol 70:173–181

Laaksonen T, Hakkarainen H, Korpimäki E (2004) Lifetime reproduc-tion of a forest-dwelling owl increases with age and area of forest.Proc R Soc London Ser B 271:S461–S464

Leonard ML, Picman J (1988) Mate choice by marsh wrens: the inXu-ence of male and territory quality. Anim Behav 36:517–528

Littel RC, Milliken GA, Stroup WW, WolWnger RS (1996) SAS sys-tem for mixed models. SAS Institute, Cary, NC

Lõhmus A, Vali Ü (2004) The eVects of habitat quality and female sizeon the productivity of the lesser spotted eagle Aquila pomarina inthe light of the alternative prey hypothesis. J Avian Biol 35:455–464

Maguire GS (2006) Territory quality, survival and reproductive suc-cess in southern emu-wrens Stipiturus malachurus. J Avian Biol37:579–593

123

514 Oecologia (2009) 160:507–514

Marra PP (2000) The role of behavioral dominance in structuring pat-terns of habitat occupancy in a migrant during the non breedingseason. Behav Ecol 11:299–308

Massemin S, Korpimäki E, Wiehn J (2000) Reversed sexual dimor-phism in raptors: evaluation of the hypotheses in kestrels breedingin a temporally changing environment. Oecologia 124:26–32

Matthysen E (1990) Behavioral and ecological correlates of territoryquality in the Eurasian Nuthatches (Sitta europaea). Auk 107:86–95

McDonald PG, Olsen PD, Cockburn A (2005) Selection on body sizein a raptor with pronounced reversed sexual size dimorphism: arebigger females better? Behav Ecol 16:48–56

Messina FJ, Fox CW (2001) OVspring size and number. In: Fox CW,RoV DA, Fairbairn DJ (eds) Evolutionary ecology: concepts andcase studies. Oxford University Press, Oxford, pp 113–127

Newton I (ed) (1989) Lifetime reproduction in birds. Academic,London

Newton I (1991) Habitat variation and population regulation inSparrowhawks. Ibis 133:76–88

Pärt T (2001) Experimental evidence of environmental eVects on age-speciWc reproductive success: the importance of resource quality.Proc R Soc London Ser B 268:2267–2271

Przybylo R, Wiggins DA, Merilä J (2001) Breeding success in bluetits: good territories or good parents? J Avian Biol 32:214–218

Rising JD, Somers KM (1989) The measurement of overall body sizein birds. Auk 106:666–674

Rodenhouse NL, Sherry TW, Holmes RT (1997) Site-dependent regu-lation of population size: a new synthesis. Ecology 78:2025–2042

RoV DA (1997) Evolutionary quantitative genetics. Chapman andHall, New York

Roithmar ME (1994) Male territoriality and female mate selection inthe dart-poison frog Epipedobates trivittatus (Dendrobatidae,Anura). Copeia 1:107–115

Schulte-Hostedde A, Zinner B, Millar JS, Hickling GJ (2005) Restitu-tion of mass-size residuals: validating body-condition indices.Ecology 86:155–163

Sergio F, Newton I (2003) Occupancy as a measure of territory quality.J Anim Ecol 72:857–865

Sergio F, Marchesi L, Pedrini P (2003a) Adaptive selection of foragingand nesting habitat by black kites (Milvus migrans) and its impli-cations for conservation: a multi-scale approach. Biol Conserv112:351–362

Sergio F, Marchesi L, Pedrini P (2003b) Spatio-temporal shifts ingradients of habitat quality for an opportunistic avian predator.Ecography 26:243–255

Sergio F, Blas J, Forero MG, Fernández N, Donázar JA, Hiraldo F(2005) Preservation of wide-ranging top predators by site protec-tion: black and red kites in Doñana National Park. Biol Conserv125:11–21

Sergio F, Blas J, Forero MG, Donázar JA, Hiraldo F (2007a) Sequen-tial settlement and site-dependence in a migratory raptor. BehavEcol 18:811–821

Sergio F, Blas J, Forero MG, Donázar JA, Hiraldo F (2007b) Size-related advantages for reproduction in a slightly dimorphic raptor:opposite trends between the sexes. Ethology 113:1141–1150

Sergio F, Blas J, Hiraldo F (2009) Predictors of Xoater status in a long-lived bird: a cross-sectional and longitudinal test of hypotheses. JAnim Ecol 78:109–118

Siikamäki P (1995) Habitat quality and reproductive traits in the PiedFlycatcher—an experiment. Ecology 76:308–312

Simmons RE (1988) Food and the deceptive acquisition of mates bypolyginous male harriers. Behav Ecol Sociobiol 23:83–92

Tabachnick BG, Fidell LS (1996) Using multivariate statistics (3rdedn). Harper Collins, New York

Van de Pol M, Bruinzeel LW, Heg D, van der Jeugd HP, Verhulst S(2006) A silver spoon for a golden future: long-term eVects ofnatal origin on Wtness prospects of oystercatchers (Haematopusostralegus). J Anim Ecol 75:616–626

Warner RR (1987) Female choice of sites versus mates in a coral reefWsh, Thalassoma bifasciatum. Anim Behav 35:1470–1478

123

Documents

Short- and long-term consequences of individual and territory quality in a long-lived bird