Age-related change in breeding performance in early life is associated with an increase in competence in the migratory barn swallow Hirundo rustica

Journal of Animal Ecology

2007

76

, 915–925

© 2007 The Authors.Journal compilation© 2007 British Ecological Society

Blackwell Publishing Ltd

Age-related change in breeding performance in early life is associated with an increase in competence in the migratory barn swallow

Hirundo rustica

JAVIER BALBONTÍN*, IGNACIO G. HERMOSELL*, ALFONSO MARZAL*†, MARIBEL REVIRIEGO*, FLORENTINO DE LOPE* and ANDERS PAPE MØLLER‡

*

Departamento de Biología Animal, Universidad de Extremadura, E-06071 Badajoz, Spain;

†

Department of Animal Ecology, Ecology Building, Lund University, S-223 62 Lund, Sweden; and

‡

Laboratoire de Parasitologie Evolutive, CNRS UMR 7103, Université Pierre et Marie Curie, Bât. A, 7ème étage, 7 quai St Bernard, Case 237, F-75252 Paris Cedex 05, France

Summary

1.

We investigated age-related changes in two reproductive traits (laying date and annualfecundity) in barn swallows

Hirundo rustica

L. using a mixed model approach to di-stinguish among between- and within-individual changes in breeding performancewith age.

2.

We tested predictions of age-related improvements of competence (i.e. constrainthypothesis) and age-related progressive disappearance of poor-quality breeders (i.e. selec-tion hypothesis) to explain age-related increase in breeding performance in early life.

3.

Reproductive success increased in early life, reaching a plateau at middle age (e.g. at3 years of age) and decreasing at older age (

>

4 years). Age-related changes in breedingsuccess were due mainly to an effect of female age.

4.

Age of both female and male affected timing of reproduction. Final linear mixedeffect models (LME) for laying date included main and quadratic terms for female andmale age, suggesting a deterioration in reproductive performance at older age for bothmales and females.

5.

We found evidence supporting the constraints hypothesis that increases in com-petence within individuals, with ageing being the most probable cause of the observedincrease in breeding performance with age in early life. Two mechanisms wereimplicated: (1) advance in male arrival date with age provided middle-aged males withbetter access to mates. Yearling males arrived later to the breeding grounds and thereforehad limited access to high-quality mates. (2) Breeding pairs maintaining bonds for 2consecutive years (experienced pairs) had higher fecundity than newly formedinexperienced breeding pairs.

6.

There was no support for the selection hypothesis because breeding performancewas not correlated with life span.

7.

We found a within-individual deterioration in breeding and migratory performance(arrival date) in the oldest age-classes consistent with senescence in these reproductiveand migratory traits.

Key-words

: breeding experience, constraints and selection hypothesis, fecundity, linearmixed effect models (LME), longevity, senescence, timing of reproduction.


(2007)

76

, 915–925doi: 10.1111/j.1365-2656.2007.01269.x

Correspondence: Javier Balbontín, Departamento de Biología Animal, Universidad de Extremadura, E-06071 Badajoz, Spain. E-mail: [email protected]

916

J. Balbontín

et al.

© 2007 The Authors.Journal compilation© 2007 British Ecological Society,


,

76

, 915–925

Introduction

Empirical studies of free-living vertebrates suggeststrongly that reproductive performance increases withage, reaching a maximum output at middle age, fol-lowed by a clear deterioration at advanced age due tosenescence (Clutton-Brock 1988; Forslund & Pärt 1995).This association between age and fecundity has beenobserved in fish (Hodder 1963), amphibians (Salthe1969), reptiles (Tinkle & Ballinger 1972), mammals(Clutton-Brock 1988) and birds (Newton 1989; Sæther1990; Forslund & Pärt 1995). Three major groups ofnon-exclusive hypotheses based on progressive appear-ance or disappearance of individual phenotypes amongage-classes (delayed breeding and selection hypo-theses; Perrins & Moss 1974; Curio 1983; Newton 1989),age-related improvements of competence (constrainthypotheses; Curio 1983; Nol & Smith 1987; Pärt 2001)and optimization of reproductive effort (restraint hypo-theses; Williams 1966; Pianka & Parker 1975; Clutton-Brock 1988) have been proposed to explain the increasein reproductive output with age in early life. It is impor-tant to realize that changes in phenotypic traits couldresult from within-individual patterns (e.g. ageing) andfrom between-individual changes in the quality ofindividuals entering or leaving the population (e.g.selection). The constraint and restraint hypothesesfocus on the phenotypic changes that take placewithin individuals, and the selection or delayed-breeding hypotheses focus on phenotypic changesamong individuals.

The selection hypotheses assume that there is alwaysa difference in quality among individuals in a popula-tion. The progressive disappearance of individuals withpoor reproductive performance occurs if such individ-uals have lower survival probabilities than individualswith superior reproductive success, and therefore apositive correlation between reproductive performanceand survival or longevity is expected at the populationlevel. Although natural selection is involved in thisprocess, humans can cause mortality on cohorts thatreverse the pattern due to natural selection (Balbontín,Penteriani & Ferrer 2005). Most studies have reportedassociations between age and reproductive traits basedon cross-sectional comparisons. Because individualswithin a population differ in their ability to acquire andutilize resources and in their optimal allocation of suchresources, between-individual variation in quality canmask within-individual ageing patterns in cross-sectional studies (Nol & Smith 1987; Forslund & Pärt1995; Nussey

et al

. 2006; van der Pol & Verhulst 2006).In particular, problems of estimating age-effects forreproductive traits within individuals arise when thereis a difference in quality among individuals causinga difference in longevity. Using longitudinal data onreproductive traits of individuals over their lifetime, theselection hypothesis could be tested readily. Recently,van der Pol & Verhulst (2006) suggested that by usinglinear mixed models appearance or disappearance

effects could be tested easily by adding age at first repro-duction (appearance effects) or longevity (i.e. age at lastreproduction, disappearance effects) as a covariate in amodel in which age is kept as a fixed effect for investi-gating their effects on reproductive traits. Correlationsbetween reproductive traits within individuals arecontrolled by adding individual identity as a randomeffect in these models.

The constraints hypotheses suggest that with increas-ing age, individuals improve in skills that improvereproductive performance. Many different ways ofincreasing competence have been proposed. For instance,a prolonged period of learning is needed to perform allreproduction tasks correctly to raise healthy youngsuccessfully. Therefore, previous breeding experiencehas been suggested to influence breeding performancepositively. Increased experience while ageing could alsobe due to learning to cope with predators or competi-tors, familiarity with the local environment or within-individual improvement in feeding behaviour, all ofwhich could cause an increase in reproductive outputwith age (Burger 1988; Marchetti & Price 1989; Desro-chers 1992). Other extrinsic mechanisms may preventyoung individuals from reproducing as successfully asolder individuals. For instance, improvement in breed-ing success with age could be due to limited access tohigh-quality mates or territories for young individuals.Therefore, limited access to resources by young indi-viduals due to constraints on migratory performanceor dominance status might cause poor reproductiveperformance when young.

The restraint hypotheses suggest that individualsenhance reproductive performance as they grow oldbecause of a decrease in residual reproductive value(Williams 1966; Gadgil & Bossert 1970; Curio 1983;Charlesworth 1994). Life history theory predicts atrade-off between reproduction and survival or betweencurrent and future reproduction (Stearns 1992). Assum-ing such trade-offs, individuals may invest less in repro-duction when young because they trade investment ofresources between current reproduction and survivalto the next breeding season.

Most studies of wild vertebrates failed to describethe effect of age on fecundity in the last part of the lifeof an individual. Studies rarely tested whether declinesin reproductive performance in later life are due tochanges in within- or between-individual performance.Theories of senescence deal with within-individualchanges in reproductive performance with age. However,when using cross-sectional data, changes in perform-ance between individuals might mask patterns of senes-cence in reproductive output. For example, if individualsthat increase reproductive effort in early life pay a costin terms of reduced longevity or accelerated repro-ductive deterioration in late life, then the observedpattern of reproductive senescence would be due to thedisappearance of high-quality breeders rather thana phenotypic change within a given individual.Therefore, separation of within- and between-individual

917

Breeding performance in early and late life and age



,

76

, 915–925

variation in age-related breeding performance consti-tutes a major challenge for the study of senescence atthe population level in the wild (Forslund & Pärt 1995;Nussey

et al

. 2006; van der Pol & Verhulst 2006).Our main aim was to test hypotheses explaining age-

related increase in two reproductive traits (e.g. lay dateand annual fecundity) early in life, using the barn swa-llow (

Hirundo rustica

L.) as a model species. The barnswallow is a small (

c

. 20 g), socially monogamous, se-micolonial passerine feeding on insect prey. It is a long-distance migratory bird covering up to

>

10 000 kmbetween breeding and wintering sites. Barn swallowsbreed on farms, gaining permanent access throughopen doors and windows. Females lay between one andthree clutches containing one to six eggs per breedingseason. Breeding dispersal is very low, because lessthan 0·1% of

>

10 000 banded adults (studied during30 years in populations in Spain, Italy and Denmark,including the population studied here) moved betweenbreeding colonies from one year to the next, and all butone among 450 local recruits was captured in the firstyear of life (Møller, de Lope & Saino 2005). Therefore,age at first reproduction shows little or no variation,with most individuals starting reproduction at the ageof 1 year. On average, only 35% of breeding individualssurvive to the next breeding season (Møller 1994a; Saino

et al

. 1999), with very few reaching very old age (i.e. 1·5%of first-time breeding birds reached the age of 5 years).None the less, longevity or age at last reproduction showsconsiderable variation (range: 1–8 years) compared withage at first reproduction, and therefore only disappear-ance effects are likely to mask within-individual changesin age-related fecundity in this short-lived species.

(

-

)

Selection: progressive disappearance hypothesis

Breeding performance increases with age because poor-quality breeders die young. We predicted annual bree-ding success to be correlated positively with longevity(i.e. age at last reproduction) and negatively with laydate. This association between age at last reproductionand a reproductive trait (i.e. annual fecundity or lay date)must hold when measuring breeding performance con-trolling for age. This hypothesis assumes a difference inquality between individuals with different life spans.Long-lived individuals should be better breeders thanindividuals with short life spans.

Constraint: breeding experience hypothesis

More experienced breeders or breeding pairs repro-duce better than less-experienced breeders or breedingpairs. Breeding experience should increase final repro-ductive output. Because there is no variation in age atfirst breeding, all similarly aged individuals should have

similar breeding experience, and therefore we used pair-bond duration as an index of breeding experience. Wepredicted breeding pairs maintaining bonds to repro-duce better than newly formed breeding pairs. Becauseimprovement in competence occurs within individuals,we controlled for between-individual changes by incor-porating age at last breeding into the analyses whentesting this hypothesis.

Constraint: limited access to high-quality mates or territories hypothesis

Improvement in breeding success with age is due tolimited access to high-quality mates or preferred nest sitesat younger ages. We predicted that migratory perfor-mance (measured as the date that male and female arrivedat the breeding grounds) should decrease with age, atleast in early life. Early-arriving individuals should havebetter access to mates and preferred nest site than late-arriving ones. We assumed that early arrival providesmales and females with a greater advantage in terms ofmating success, laying date and annual reproductiveoutput than late-arriving individuals, as demonstratedpreviously in several studies on this species (Møller1994a,b; Saino

et al

. 1997; Kose & Møller 1999; Kose

et al

. 1999; Møller, de Lope & Saino 2004). Becausechange in improvement in competence occurs withinindividuals, we controlled for between-individualchanges by incorporating age at last breeding intoanalyses when testing this hypothesis.

Finally, we tested whether deterioration in reproduc-tive performance (laying date and annual fecundity)and migratory performance (arrival date) in late life wasdue to trade-offs between reproduction and survival byintroducing curvilinear relationships between age atlast reproduction and breeding performance, as doc-umented previously in wild vertebrates (Reid

et al

. 2003).

Methods

We studied barn swallows at Badajoz, south-west Spain(38

°

50

′

N, 6

°

59

′

W). The study site was mainly openfarmland with pasture, cereals and fruit plantations (deLope 1983). Early during the breeding season we cap-tured adults at dawn by using mist nets across windowsand doors of the breeding sites. We studied barn swallowsin four different colonies: Potosi (POT), Almendral(ALM), Tres Arroyos (TA), and Virgen de Guadalupe(VG) within our study area (mean

±

SD distance betweencolonies: 10781 m

±

8505, range: 1706–20 338 m). Wemade weekly captures of birds during every breedingseason until 98–100% of breeding individuals were cap-tured. Each bird was identified with a numbered metalring and a combination of coloured PVC rings, so wecould recognize each adult visually. We measured rightand left outermost tail feathers with a ruler to the nearest0·5 mm. Tail length was determined as the mean value

918

J. Balbontín

et al.



,

76

, 915–925

of left and right characters. Individuals with brokentails were excluded from the analyses, because the roundedtip of the outermost tail feathers reveals clearly whetherit is broken. Body mass was recorded with a Pesolaspring balance to the nearest 0·5 g. All measures weretaken by the same observer (F. de L.), which eliminatesany noise in the data due to interobserver variability.All birds were provided with an individual combinationof colour markings on their belly feathers using stampink. Individuals were sexed from the presence (females)or absence (males) of a brood patch and from observa-tion of breeding behaviour during the courtship andincubation period.

We tracked all reproductive events and identifiedvisually each individual of pairs using binoculars frominside a hide, to avoid disturbance during reproduc-tion. We checked all nests in our colonies every 2 daysduring the breeding season with the aim of recordedlaying date, clutch size and brood size up to the thirdbrood. Once females started laying we conducted dailyobservation sessions inside a hide, observing eachoccupied nest, with the aim of identifying individualsbased on colour rings and ink marks. We were able toidentify each individual in the pair and their nest for themajority of the occupied nest sites. Because these po-pulations have been subject to several experimental studies(tail-length manipulation), we eliminated from our dataset any individuals involved in these experiments (89manipulated males and 32 manipulated females). How-ever, manipulated birds differed from non-manipulatedbirds for several traits, resulting in a non-random subsetof individuals being involved in experiments. Femalesinvolved in experiments arrived later [mean (SE):–0·02 (0·05)], laid later [standardized lay date: –0·05(0·05)] and had shorter tails [84·6 mm (0·30)] than non-manipulated females [mean (SE), standardized arrivaldate: –0·48 (0·08), standardized laying date: –0·48 (0·06),tail length: 86·9 mm (1·10 mm)], and manipulated malesarrived later [mean (SE): –0·02 (0·05)] and were older[mean (SE): 2·00 (0·06) years] than non-manipulatedmales [standardized arrival date, mean (SE): –0·37(0·06), age: 1·67 (0·09) years] (

t

-tests employed to checkfor differences on tail length, standardized arrival date,standardized lay date, standardized annual breedingsuccess and age). Because tail manipulation was per-formed as birds arrived at our study areas, starting alittle later during the breeding season, it could result inthe sample of manipulated birds being skewed towardslater arrival date, which could explain the observed di-fferences in arrival date, lay date, male age and tail lengthbetween manipulated and non-manipulated individuals.However, the sample of non-manipulated birds (i.e.those used in the present study) encompassed the entirerange of the population for arrival date, tail length, laydate and age. Therefore, excluding manipulated birdswould have reduced the number of later-arriving andbreeding individuals and the number of females withshorter tails from our final samples. In total, we iden-tified individual males and females for pairs for 322

breeding events during 1994–2006. Two differentmeasures of breeding success were employed. Annualfecundity was measured as the total number of chicksraised to fledging age every year (in up to three differentannual broods), and laying date was estimated as thedate of the first egg of the first brood, relative to day 1(1 March). Migratory performance (arrival date) wasmeasured as the first day an individual was captured inour consecutive capture–recapture sessions, relativeto day 1 (1 February). The precision of this estimatehas been checked previously by observations from thebeginning of the breeding season and calculating asecond measure of arrival date as the first day an adultbird was identified from its colour rings and belly inkmarks. The correlation between these two estimateswas very high (Pearson’s correlation coefficient: range:0·98–0·99, calculated in 7 different years for both sexesseparately; Møller

et al

. 2004).Several studies have demonstrated that barn swa-

llows show high site fidelity and rarely leave the colony.Therefore, disappearance of colour-ringed breeders fromthe colony indicated mortality rather than dispersal.Once a 1-year-old individual returned from its firstmigration, it selected a breeding area and remainedloyal to it for the rest of its life. Therefore, we couldassign the age of individuals with accuracy in our studycolonies, assuming unringed birds to be yearlings atfirst capture (Møller 1992; Møller

et al

. 2005).

We used linear models to investigate the effect of age onbreeding success and migratory performance. The keyfeature of longitudinal data is that individuals aremeasured repeatedly through time. Linear mixed effectmodels (LME) are particularly useful when there istemporal pseudo-replication (repeated measurement).We used the LME procedure of S-Plus 2000 (Mathsoft1999) to investigate the effect of age of the female andits mate on breeding performance, timing of breeding(lay date) and migratory performance (arrival date).We controlled for interannual variation on breedingand migratory performance by standardizing the firstegg-lay date, annual fecundity and arrival date by sub-tracting the annual population mean from each observedvalue and dividing by the annual population standarddeviation (Zar 1999). We used a normal error distribu-tion with an identity link function to model standar-dized arrival, first egg-lay date or annual fecundity asresponse variables. Full and reduced models werefitted by using a maximum likelihood (ML) methodand final parameters for final minimal adequatemodel were estimated using the restricted maximumlikelihood (REML) method (McCullough & Nelder1989; Crawley 2002). There was missing informationfor some individuals, which resulted in slightly varyingsample sizes in different analyses.

Because of high adult mortality (

>

65%, Møller 1994a;Saino

et al

. 1999) the duration of pair-bonds was always

919

Breeding performance in early and late life and age



,

76

, 915–925

short, and therefore we could not specify pair identityas a random factor. Accordingly, when modellingreproductive traits, we introduced bird identity (femaleor female’s mate identity) as a random effect, incorpo-rating identity of the female in a first attempt and there-after repeating the analysis using identity of the female’smate, with the aim of checking if conclusions for fixedeffects were similar using these two different approaches.Explanatory variables were female’s and mate’s age,female’s and mate’s age at last reproduction (ALR)(defined as the age a breeding individual was lastrecorded in the breeding colony), colony (factor withfour levels, POT, ALM, TA and VG) and pair-bondduration (i.e. the number of breeding seasons a pairhad bred together). We also fitted full models withfemale’s and mate’s body mass (g) and female’s andmate’s tail length (mm) when modelling standardizedlaying date. Also, female’s and mate’s arrival and layingdate was included to model standardized annual fecun-dity, because they have been shown previously to affectthese reproductive traits (review in Møller 1994a). Whenthe response was standardized arrival date (for eithermales or females), explanatory variables introducedinto the models were age and ALR of either sex. All thesevariables were treated as fixed-effects in all models.

The selective disappearance hypothesis was in-vestigated following van der Pol & Verhulst (2006).ALR was left in all models independent of its effect, insuch a way that we could test specifically for within-individual age effects in the presence of a selectivedisappearance effect (the estimated slope of ALR).Investigation of the dispersion plots suggested incor-poration of female’s and mate’s age as second-orderpolynomials in a maximal model. Curvilinear relation-ships between ALR and breeding performance were also

tested, because a significant effect could suggesttrade-offs between reproduction and survivals (Reid

et al

. 2003; Nussey

et al

. 2006). Therefore, we firstincluded second-order main fixed effects and possibletwo-way interactions in a maximal model, reducing itby eliminating non-significant terms from the fixedstructures (Crawley 2002). The statistical significanceof each covariate and all possible two-way interactionsamong fixed effects were tested in turn, using a back-ward stepwise procedure to select the most parsimoni-ous model. Models with different fixed structures werecompared using

F

-tests, Akaike’s information cri-terion (AIC) and L-ratio tests (Akaike 1973; Pinheiro& Bates 2000; Crawley 2002). The final model wasconsidered to have been reached when all variables(except ALR) had a significant effect at

P

<

0·05.

Results

The age of the two members of a breeding pair wasrelated to the start of breeding. LME showed that themost parsimonious model included both main andquadratic terms for the female and its mate’s age, whileaccounting for the known source of variability due toamong-individual variation (Table 1). This model wassignificant (L-ratio

=

95·91,

P

<

0·0001). Middle-agedfemales (3 years) started incubation earlier than young(1–2 years) and older females (4 years or older). Theage of the female mate’s was related to the start ofreproduction in a similar way. Females that matedwith middle-aged (3 years) males laid eggs earlier thanthose mated with either young (1–2 years) or older mates(4 years or older), independently of their own age(Fig. 1). The interaction between female and male agewas not statistically significant. There was no effectof disappearance of low-quality individuals (esti-mated

β

s

for main and quadratic terms on female’s ALR:

P

>

0·2, Table 2). Female tail length was related to startof reproduction, with long-tailed females starting toreproduce earlier than the average females. The dur-ation of the pair-bond was not related significantly tothe onset of reproduction (

P

>

0·2). Neither maletail length nor male or female body mass were relatedsignificantly to the onset of reproduction. There wasconsiderable between-individual variation in first-egglay date, with female identity accounting for 43% ofthe total variance (calculated in a model withoutfixed effects).

Only age of the female in the breeding pair was relatedsignificantly to annual fecundity. The most parsimoni-ous model retained main and quadratic terms of femaleage, while accounting for the known source of variability

Table 1.

Linear mixed models of standardized first egg-lay date as the responsevariable. Full models built with age and age at last reproduction (ALR) of female, pair-bond duration, colony, age and ALR of female’s mate, female and male body mass andtail length as explanatory variables. Only significant terms are shown, except for femaleALR, which was retained in the model. Sample size is 322 breeding events for known-age breeding pairs

Random effects

SD 95% CI

Female identity 0·566 0·451–0·710Residual 0·581 0·493–0·685

Fixed-effects

Estimate SE d.f.

F P

Intercept (

β

0

) 3·461 0·814 1, 234 0·00 0·96Female age (

β

1w

) –0·663 0·158 1, 67 43·04

<

0·0001***Female age

2

(

β

2w

) 0·087 0·026 1, 67 29·34

<

0·0001***Female ALR (

β

3s

) –0·068 0·161 1, 67 0·00 0·95Female ALR

2

(

β

4s

) 0·011 0·025 1, 67 0·07 0·77Mate’s age (

β

5

) –0·771 0·150 1, 67 17·99

<

0·0001***Mate’s age

2

(

β

6

) 0·113 0·027 1, 67 16·27 0·0001***Female tail length (

β

7

) –0·019 0·009 1, 67 4·34 0·04

920

J. Balbontín

et al.



,

76

, 915–925

due to among-individual variation (Table 2). The finalmodel was significant (L-ratio

=

21·57,

P

=

0·0014).Average annual fecundity for middle-aged females(3 years) was 6·74

±

2·95 fledglings/year, which was largerthan average annual fecundity of young females (1–2 years, mean

±

SD

=

5·44

±

2·66 fledglings/year) or olderfemales (4 years or more, mean

±

SD

=

5·88

±

2·99fledglings/year) (Fig. 2). Neither the effect of male agenor the interaction between female and male age wasretained in the model. There was no significant effect ofdisappearance of low-quality individuals (estimated

β

s

for main and quadratic terms on the female’s ALR:

P

>

0·4, Table 2). The duration of the pair-bond wasrelated significantly to final reproductive output. Thus,the average annual fecundity of pairs that bred togethertwice was significantly larger (mean

±

SD

=

7·32

±

3·03fledglings/year,

n

=

19) than that of pairs that remainedtogether for just one breeding season (mean

±

SD

=

5·41

±

2·74 fledglings/year,

n

=

219, Fig. 3). Male bodymass was also retained in the final LME. Male bodymass was related positively to the number of chicksfledged. Neither female (L-ratio

=

0·81,

P

=

0·36) normale arrival date (L-ratio

=

2·86,

P

= 0·09) or lay date(L-ratio = 2·15, P = 0·14) were retained in the minimaladequate model. However, if we excluded from thefull model the main and quadratic effects of female age,the effect of laying date and arrival date become highlysignificant (estimate ± SD: laying date = –0·184 ± 0·06,L-ratio = 8·29, P = 0·004; female arrival date = –0·167± 0·059, L-ratio = 7·77, P = 0·005; male arrival date = –0·182± 0·05, L-ratio = 9·83, P = 0·001). Between-individual

Fig. 1. (a) Effect of female age on laying date; (b) effects ofmate’s age on laying date having accounted for female ageeffect (i.e. mate’s age effect on residuals from the regression offemale age on laying date).

Fig. 2. Effect of female age on standardized annual fecundity.Breeding females aged 1, 2, 3, 4, 5, 6 and 7 years old. Samplesize is 313 breeding events.

Fig. 3. Box plots for standardized annual fecundity forexperienced (i.e. breeding together for two consecutive years,n = 19 breeding attempts) and inexperienced breeding pairs(i.e. breeding together for just 1 year, n = 219 breeding attempts).

Table 2. Final linear mixed model obtained for standardized annual fecundity. Fullmodel fitted with female age, female age at last reproduction (ALR), pair-bondduration, colony, age and ALR of female’s mate, male and female body mass and taillength as explanatory variables. Only significant terms are shown, except for femaleALR, which was retained in the model. Sample size is 312 breeding events for known-age breeding pairs

Random effects

SD 95% CI

Female identity 0·237 0·036–1·536Residual 0·931 0·805–1·078

Fixed-effects

Estimate SE d.f. F P

Intercept (β0) –0·280 0·847 1, 203 0·20 0·65Female age (β1w) 0·443 0·221 1, 66 3·58 0·06Female age2 (β2w) –0·076 0·038 1, 66 7·35 0·008**Female ALR (β3s) –0·045 0·183 1, 66 0·40 0·52Female ALR2 (β4s) 0·012 0·027 1, 66 0·47 0·49Pair-bond (β5) 0·532 0·239 1, 66 4·88 0·03*Male body mass (β6) 0·097 0·043 1, 66 5·14 0·02*

921Breeding performance in early and late life and age

© 2007 The Authors.Journal compilation© 2007 British Ecological Society, Journal of Animal Ecology, 76, 915–925

variation in annual fecundity (calculated for femaleidentity) accounted for 16% of the total variance (cal-culated in a model without fixed effects).

Linear mixed models showed that age was related to thedate males returned to their breeding grounds. Variationin male arrival date was best explained by a model inclu-ding male age with significant main and quadratic terms,while accounting for the known source of variation dueto among-male identity (Table 3). This model was sig-nificant (L-ratio = 73·76, P < 0·0001). Middle-aged males(3 years) arrived earlier than either young (1–2 years)or older males (4 years or older) (Table 3, Fig. 4a).There was no effect of disappearance of low-qualityindividuals in early life (main effect of ALR, P > 0·2).The quadratic terms of ALR were also not significant(P = 0·21, Table 3). There was considerable between-individual variation in male arrival date, with maleidentity accounting for 33% of total variance (calcu-lated in a model without fixed effects).

Age was related in a quadratic fashion to the dayfemales returned to their breeding grounds. The finalLME model included both linear and quadratic termsfor female age (L-ratio = 74·65, P < 0·001). Female agewas related to their arrival date at the breeding groundsin a similar way to males, with an optimum age of arrivalat 3 years. Younger (1–2 years) and older females (4 yearsor more) arrived later at the breeding grounds (Table 3,Fig. 4b). There was no significant effect of disappear-ance of low-quality individuals on the date of femalearrival [estimate (βs) for main and quadratic termsP > 0·2; Table 3]. There was considerable between-individual variation in female arrival date, with femaleidentity accounting for 40% of total variance (calcu-lated in a model without fixed effects).

Discussion

The main findings of this study were (1) that annualfecundity increased with age in early life, reaching amaximum at mid-life and decreasing in late life; (2) theassociation between breeding performance and agewas due mainly to a change in within-individual ageingrather than to a change in between-individual pattern or

Table 3. Linear mixed models of standardized male and female arrival date as the response variable, and age and age at lastreproduction (ALR) as explanatory variables. Only significant terms are shown except for female ALR. Sample size is 319 and320 arriving males and females, respectively

Standardized male arrival date Random effects

Standardized female arrival date

SD 95% CI SD 95% CI

Male/female identity Residual

0·405 0·223–0·736 0·491 0·306–0·7900·802 0·688–0·935 0·807 0·682–0·954

Fixed-effects

Estimate SE F P Estimate SE F P

Intercept (β0) 1·741 0·245 0·10 0·74 0·750 0·222 0·06 0·80Age (β1w) –1·070 0·200 36·40 <0·0001 –0·754 0·193 13·88 0·0004Age2 (β2w) 0·158 0·035 32·77 <0·0001 0·113 0·033 11·32 0·001ALR (β3s) –0·290 0·198 1·04 0·30 0·128 0·183 1·10 0·29ALR2 (β4s) 0·038 0·030 1·55 0·21 –0·030 0·028 1·14 0·28

Fig. 4. (a) Effect of male age on standardized male arrivaldate; (b) effect of female age on standardized female arrivaldate. Curvilinear relationships (i.e. model prediction) areshown. Sample size was 320 females and 319 males.

922J. Balbontín et al.


selection; (3) the association between age and annualfecundity was due mainly to a female age effect; (4) bothfemale and male age affected the timing of reproduction;(5) two reproductive traits (i.e. annual fecundity andlaying date) and one migratory performance trait (i.e.arrival date) deteriorated in late life; (6) between-individual variation in individual phenotypic qualitydid not affect change in migratory performance withage; and we found that (7) some evidence supporting‘constraints’ hypothesis were found because breedingperformance increases with breeding experience andmigratory performance improve while ageing. Each ofthese main findings are discussed briefly.

Numerous empirical studies have established thatreproductive performance in birds generally improveswith age (Clutton-Brock 1988; Forslund & Pärt 1995;Sanz & Moreno 2000; Green 2001; Pyle, Sydeman &Hester 2001; Laaksonen, Korpimäki & Hakkarainen 2002).However, the reasons for such increases are still poorlyunderstood. Experimental and observational studiesin birds and mammals have found evidence consistentwith the increase in competence and selection hypoth-eses for explaining age-related fecundity in early life(De Steven 1978; Desrochers 1992; Komdeur 1996;Balbontín, Penteriani & Ferrer 2003; Ferrer & Bisson 2003;Penteriani, Balbontín & Ferrer 2003). Reid et al. (2003),in a review of the bird literature, reported evidencesupporting the differential mortality hypothesis (i.e.selection) in seven of 24 (29%) studies. Most empiricalstudies were based on cross-sectional comparisons inwhich, in many cases, these two groups of hypotheseswere confounded. Increases in competence have beenfound to be an important factor in explaining the increasein reproductive success with age. For instance, somestudies have found experienced breeders to performbetter than inexperienced breeders while controllingfor age, providing support for the breeding experiencehypothesis (Ainley, LeResche & Sladen 1983; Pyle et al.1991; Forslund & Larsson 1992; Pärt 1995; Dittmann& Becker 2003). In contrast, other studies did not findthis effect of experience on breeding success (Perdeck &Cavé 1992; Newton, Marquiss & Moss 1981; Boekelheide& Ainley 1989; Raleigh & Rendell 2001).

Here, we found evidence suggesting that increases incompetence within individuals most probably causedthe observed increase in breeding performance withage, at least until reaching middle age. In contrast, wefound no support for the selection hypothesis becausethere was no correlation between life span and breedingperformance. These conclusions were not confoundedby between-individual differences in quality, becausewe controlled for such effects in the analysis concerningwithin-individual changes in breeding success or layingdate with age. Furthermore, we identified two differentmechanisms associated with an increase in competence.First, male and female barn swallows advance theirarrival date at the breeding grounds as they age untilreaching middle age. The fitness benefits of early arrivalhave been documented widely in birds (Forstmeier

2002; Dittmann & Becker 2003), including this species(Møller et al. 2004). Early-arriving individuals have ahigher probability of mating, start to reproduce earlierand have higher fecundity than the average individual(Møller 1994a,b). Arrival date has been shown to becondition-dependent, with males in prime conditionarriving early at the breeding grounds. Therefore, age-ing males and females may increase the skills needed toperform correctly the tasks related to make the long-distance migratory journey safely between winterquarters in Africa and the breeding grounds in Europe.This increase in competence in migratory performancewith age would finally result in both an advance in layingdate and an increase in breeding success because of thepositive relationship between arrival date, laying dateand breeding success (Møller 1994a,b). However, it wasdifficult to disentangle the effects of age from those ofarrival and laying dates on annual fecundity in this cor-relative study. Although the minimal adequate modelretained main and quadratic terms of female age, the maineffect of female age became non-significant (P = 0·10)when including female and male arrival date and layingdate as three new explanatory variables. This was becauseage, arrival dates and laying dates are correlated witheach other, which precludes us from finding an associ-ation between these variables and annual fecundity onthe predicted direction at the same time.

Secondly, we also reported an increase in com-petence due to an increase in breeding experience, asdemonstrated by the larger annual fecundity of breedingpairs that reproduced together for 2 consecutive years,compared with that of breeding pairs that reproducedonly in 1 year and hence were less experienced. Con-versely, Saino et al. (2002), studying a population ofbarn swallows in Italy, did not find a reproductiveadvantage in terms of clutch size, hatching date, fledg-ing success or offspring phenotype between pairs thatre-mated for 2 consecutive years compared with pairsthat divorced from one year to the next. These resultsand those that we have obtained here are not compara-ble, because in the study by Saino et al. (2002), divorcedpairs were considered to be those in which both memberssurvived but did not re-mate, while in the present studyless experienced pairs did not re-mate mainly becauseone of the pair members died between one breedingseason and the next.

A previous study of age-dependent changes in repro-ductive traits and senescence in barn swallows has shownthat reproductive success increased in early life, reach-ing a plateau in middle age and decreasing in old age(Møller & de Lope 1999). However, that study did notexplore the effect of female and male age on breedingsuccess separately and therefore could not establish theindependent effect of female and male age on fecundity.The present study confirmed age-dependent relation-ships with fecundity and showed that it was due speci-fically to an age effect of females. Our study was basedon a large sample size, and therefore the power of thestatistical tests was large. Studies considering the age of



both sexes suggested that performance can vary moreclosely with either sex depending on species-specificreproductive roles of males and females. Examples fora larger male effect were found in passerines (Nol &Smith 1987; McCleery et al. 1996; Green 2001), auklets(Pyle, Sydeman & Hester 2001) and birds of prey (Espieet al. 2000) and a larger female effect in water-birds(Forslund & Larsson 1992) and bee-eaters and passerines(Lessells & Krebs 1989; Desrochers & Magrath 1993;Smith 1993; Komdeur 1996). In barn swallows, parentsfeed the young until a few days after fledging, but femalescontribute more to feeding than their mates (Sainoet al. 2002), providing a possible explanation as to whyfemale age effects were more important than male ageeffects in age-dependent fecundity, at least during theearly part of life.

A novel finding of the present study was that femaleand male age was related to timing of breeding in barnswallows. As far as we know, no association betweenfemale and male age and laying date has been shownpreviously in this or other species. Interestingly, the ageof males was related to timing of breeding in a similarway to females. This means that females could benefitby mating with middle-aged males and males by matingwith middle-aged females, because this will advance theonset of reproduction of the pair. Therefore, it would beadaptive for each individual to mate with middle-agedmates because this is the optimal age for which lyingdate was earliest. Future research could focus on indi-vidual mating preferences with respect to age, with theaim of trying to disentangle age effects from matingpreferences for phenotype traits (for example, taillength). Experimental studies could manipulate thetail length of males and females of different age-classes.For example, tails of individuals of different age couldbe shortened or elongated to an average of middle-agedindividuals, allowing to test for mating preference withrespect to age while controlling for tail length. Matingpreferences for tail length could be tested by manipu-lating tail length within different age-classes, therebycontrolling for age effects.

Future research might also focus upon testing theoptimization of reproductive effort (restraints hypoth-esis), which was not evaluated in the present study. Thishypothesis predicts that the relative amount of resourcesallocated to reproduction should increase with age. Ifsurvival probability decreases only late in life, asexpected in the barn swallow, reproductive effort shouldincrease at old age. Observational studies could evalu-ate investment in reproduction of individuals belongingto different age-classes, with special focus on old ageclasses. Experimental studies, for example, could mani-pulate brood size for breeding individuals of differentages. Old individuals should invest more in reproduc-tion than young individuals, specifically when caringfor an enlarged brood.

Our long-term study allowed us to quantify within-individual deterioration in two reproductive traits (i.e.annual fecundity and laying date) and one migratory

trait (i.e. arrival date) late in life, while controlling forbetween-individual variation. This finding providesevidence of the existence of senescence in fecundity andarrival date, as shown previously (Møller & de Lope 1999),and in laying date. Linear mixed models revealed con-siderable variation among individuals in laying date,annual fecundity and arrival date. Individual differencesin laying date, annual fecundity and female or malearrival date were not associated with age at last repro-duction because neither linear nor quadratic terms weresignificant in LME. Therefore, the quality of femalebreeders with respect to timing of breeding or breedingsuccess or the quality of male and female with respectto arrival date was represented equally among the over-all range of age-classes.

In conclusion, the observed increase in breeding per-formance with age in early life was related to a within-individual increase in breeding experience and migratoryperformance as individuals aged, providing supportfor the ‘constraints hypothesis’. Conversely, we did notfind evidence for the disappearance of poor-qualityindividuals among age-classes and, hence, the ‘selectionhypothesis’ was not supported. Fecundity was affectedspecifically by age of the breeding female, probablybecause females invest more in the late stage of repro-duction, feeding offspring more extensively than theirmates. The age of the two members of a breeding pairaffected the onset of reproduction in a similar manner,which could have consequences for the mating prefer-ences of individuals with respect to age. We confirmedthe existence of senescence in fecundity and arrival dateand found that laying date also deteriorated in late life.

Acknowledgements

We are grateful to all the people that help to obtain datain the field: F. Mateos, C. Navarro, P. Ninni, J. Cuervo,A. Barbosa and S. Merino. The study was supported bythe Spanish Ministry of Education and Science (CGL2006–2913).

References

Ainley, D.G., LeResche, R.E. & Sladen, W.J.L. (1983) Breed-ing Biology of the Adelie Penguin. University of CaliforniaPress, California.

Akaike, H. (1973) 2nd International Symposium on InformationTheory, pp. 267–281. Akademial Kiado, Budapest, Hungary.

Balbontín, J., Penteriani, V. & Ferrer, M. (2003) Variations inthe age of mates as an early warning signal of changes inpopulation trends? The case of Bonelli’s eagle in Andalusia.Biological Conservation, 109, 417–423.

Balbontín, J., Penteriani, V. & Ferrer, M. (2005) Humans actagainst the natural process of breeder selection: a modernsickness for animal populations? Biodiversity and Con-servation, 14, 179–186.

Boekelheide, R.J. & Ainley, D.G. (1989) Age, resource avail-ability, and breeding effort in Brandt’s cormorant. Auk,106, 389–401.

Burger, J. (1988) Effects of age on foraging in birds. Pro-ceedings of the International Ornithological Congress, XIX,1127–1140.

924J. Balbontín et al.


Charlesworth, B. (1994) Evolution in Age-Structures Popula-tions. Cambridge University Press, Cambridge.

Clutton-Brock, T.H. (1988) Reproductive Success. Universityof Chicago Press, Chicago.

Crawley, J.M. (2002) Statistical Computing. An Introductionto Data Analysis Using S-Plus. Wiley, Chichester, UK.

Curio, E. (1983) Why do young birds reproduce less well? Ibis,125, 400–404.

De Steven, D. (1978) The influence of age on the breedingbiology of the tree swallow; Irioprocne bicolor. Ibis, 120,516–523.

Desrochers, A. (1992) Age-related differences in reproductionby European blackbirds: restraint or constraint? Ecology,73, 1128–1131.

Desrochers, A. & Magrath, R.D. (1993) Age-specific fecund-ity in European blackbird Turdus merula: individual andpopulation trends. Auk, 110, 255–263.

Dittmann, T. & Becker, P.H. (2003) Sex, age, experience andcondition as factors affecting arrival date in prospectingcommon terns, Sterna hirundo. Animal Behaviour, 65, 981–986.

Espie, R.H.M., Oliphant, L.W., James, P.C., Warkentin, I.G.& Lieske, D.J. (2000) Age-dependent breeding performancein merlins (Falco columbarius). Ecology, 81, 3404–3415.

Ferrer, M. & Bisson, I. (2003) Age and territory effects onfecundity of the Spanish imperial eagle. Auk, 120, 180–186.

Forslund, P. & Larsson, K. (1992) Age-related reproductivesuccess in the barnacle goose. Journal of Animal Ecology,61, 195–204.

Forslund, P. & Pärt, T. (1995) Age and reproduction in birds:hypotheses and tests. Trends in Ecology and Evolution, 10,374–377.

Forstmeier, W. (2002) Benefits of early arrival at breedinggrounds vary between males. Journal of Animal Ecology,71, 1–9.

Gadgil, M. & Bossert, W. (1970) Life historical consequencesof natural selection. American Naturalist, 104, 1–24.

Green, D.J. (2001) The influence of age on reproductive per-formance in the brown thornbill. Journal of Avian Biology,32, 6–14.

Hodder, V.M. (1963) Fecundity of Grand Bank haddock.Journal of the Fisheries Research Board of Canada, 20, 1465–1487.

Komdeur, J. (1996) Influence of age on reproductive perform-ance in the Seychelles warbler. Behavioural Ecology, 7, 417–425.

Kose, M.R., Mänd, R. & Møller, A.P. (1999) Sexual selectionfor white tail spots in the barn swallows in relation tohabitat choice by feather lice. Animal Behaviour, 58, 1201–1205.

Kose, M.R. & Møller, A.P. (1999) Sexual selection, featherbreakage and parasites: the importance of the white spotsin the tail of barn swallows. Behavioral Ecology and Socio-biology, 45, 430–436.

Laaksonen, T., Korpimäki, E. & Hakkarainen, H. (2002)Interactive effects of parental age and environmentalvariation on the breeding performance of Tengmalm’s owls.Journal of Animal Ecology, 71, 23–31.

Lessells, C.M. & Krebs, J.R. (1989) Age and breeding per-formance in the European bee-eater. Auk, 106, 375–382.

de Lope, F. (1983) La avifauna de las Vegas Bajas del Gua-diana [The avifauna of the Vegas Bajas del Guadiana].Doñana Acta Vertebrata, 10, 91–121.

Marchetti, K. & Price, T. (1989) Difference in the foraging ofjuvenile and adult birds: the importance of developmentalconstraints. Biological Reviews, 64, 51–70.

Mathsoft, I. (1999) S-plus 2000 Guide to Statistics. Data Ana-lysis Products Division, Seattle, USA.

McCleery, R.H., Clobert, J., Julliard, R. & Perrins, C.M.(1996) Nest predation and delay cost of reproduction in thegreat tit. Journal of Animal Ecology, 65, 96–104.

McCullough, P. & Nelder, J.A. (1989) Generalized LinearModels, 2nd edn. Monographs on Statistics and AppliedProbability. Chapman & Hall, London, UK.

Møller, A.P. (1992) Sexual selection in the monogamous barnswallow (Hirundo rustica). Mechanisms of sexual selection.Journal of Evolutionary Biology, 5, 603–624.

Møller, A.P. (1994a) Sexual Selection and the Barn Swallow.Oxford University Press, Oxford, UK.

Møller, A.P. (1994b) Phenotype-dependent arrival time an itsconsequences in a migratory bird. Behavioral Ecology andSociobiology, 35, 115–122.

Møller, A.P. & de Lope, F. (1999) Senescence in a short-livedmigratory bird: age-dependent morphology, migration, repro-duction and parasitism. Journal of Animal Ecology, 68, 163–171.

Møller, A.P., de Lope, F. & Saino, N. (2004) Parasitism, immu-nity, and arrival date in a migratory bird, the barn swallow.Ecology, 85, 206–219.

Møller, A.P., de Lope, F. & Saino, N. (2005) Reproductionand migration in relation to senescence in the barn swallowHirundo rustica: a study of avian ‘centenarians’. Age, 27,307–318.

Newton, I. (1989) Lifetime Reproduction in Birds. AcademicPress, London.

Newton, I., Marquiss, M. & Moss, D. (1981) Age and breedingin sparrowhawks. Journal of Animal Ecology, 50, 839–853.

Nol, E. & Smith, J.N.M. (1987) Effects of age and breedingexperience on seasonal reproductive success in the songsparrow. Journal of Animal Ecology, 56, 301–313.

Nussey, D.H., Kruuk, L.E.B., Donald, A., Fowlie, M. &Clutton-Brock, T.H. (2006) The rate of senescence inmaternal performance increases with early-life fecundity inred deer. Ecology Letters, 9, 1342–1350.

Pärt, T. (1995) Does breeding experience explain increasedreproductive success with age – an experiment. Proceedingsof the Royal Society of London, Series B, 260, 113–117.

Pärt, T. (2001) Experimental evidence of environmental effectson age-specifics reproductive success: the importance ofresources quality. Proceedings of the Royal Society, Series B,268, 2267–2271.

Penteriani, V., Balbontín, J. & Ferrer, M. (2003) Simultaneouseffects of age and territory quality on fecundity in Bonelli’seagle (Hieraaetus fasciatus). Ibis, 145 (online), E77–E82.

Perdeck, A.C. & Cavé, A.J. (1992) Lying date in the coot:effects of age and mate choice. Journal of Animal Ecology,61, 13–20.

Perrins, C.M. & Moss, D. (1974) Survival of young great titsin relation to age of female parent. Ibis, 116, 220–224.

Pianka, E.R. & Parker, W.S. (1975) Age-specific reproductivetactics. American Naturalist, 360, 113–117.

Pinheiro, J.C. & Bates, D.M. (2000) Mixed-Effects Models inS and S-Plus. Springer-Verlag, New York.

van der Pol, M. & Verhulst, S. (2006) Age-dependent traits:a new statistical model to separate within- and between-individual effects. American Naturalist, 167, 766–773.

Pyle, P., Spear, L.B., Sydeman, W.J. & Anley. D.G. (1991) Theeffects of experience and age on the breeding performanceof western gulls. Auk, 108, 25–33.

Pyle, P., Sydeman, W.J. & Hester, M. (2001) Effects of age,breeding experience, mate fidelity on breeding performancein a declining population of Cassin’s auklets. Journal ofAnimal Ecology, 70, 1088–1097.

Raleigh, J.R. & Rendell, W.B. (2001) A long-term study ofreproductive performance in tree swallows: the influence ofage and senescence on output. Journal of Animal Ecology,70, 1014–1031.

Reid, J.M., Bignal, E.M., Bignal, S., McCracken, D.I. &Monaghan, P. (2003) Age-specific reproductive performancein red-billed choughs Phyrrhocorax phyrrhocorax: patternsand processes in a natural population. Journal of AnimalEcology, 72, 765–776.



Sæther, B.-E. (1990) Age-specific variation in the reproductiveperformance of birds. Current Ornithology, 7, 251–283.

Saino, N., Ambrosini, R., Martinelli, R. & Møller, A.P.(2002) Mate fidelity, senescence in breeding performanceand reproductive trade-offs in the barn swallow. Journal ofAnimal Ecology, 71, 309–319.

Saino, N., Bolzern, A.M. & Møller, A.P. (1997) Inmuno-com-petence, ornamentation and viability of male barn swallows(Hirundo rustica). Proceedings of the National Academy ofSciences USA, 94, 549–552.

Saino, N., Calza, S., Ninni, P. & Møller, A.P. (1999) Barnswallows trade survival against offspring condition andinmunocompetence. Journal of Animal Ecology, 68, 999–1009.

Salthe, S.N. (1969) Reproductive modes and the number andsizes of ova in urodeles. American Midland Naturalist, 81,467–490.

Sanz, J.J. & Moreno, J. (2000) Delayed senescence in a south-ern population of the pied flycatcher (Ficedula hypoleuca).Ecoscience, 7, 25–31.

Smith, H.G. (1993) Parental age and reproduction in themarsh tit (Parus palustris). Ibis, 135, 196–201.

Stearns, S.C. (1992) The Evolution of Life Histories. OxfordUniversity Press, Oxford.

Tinkle, D.W. & Ballinger, R.E. (1972) Sceloporus undulatus:a study of the intraspecific comparative demography of alizard. Ecology, 53, 570–584.

Williams, G.C. (1966) Natural selection, the cost of reproduc-tion, and a refinement of Lack’s principle. American Naturalist,100, 687–690.

Zar, J.H. (1999) Biostatistical Analysis. Prentice Hall, NewJersey.

Received 30 January 2007; accepted 30 April 2007

Documents

Age-related change in breeding performance in early life is associated with an increase in competence in the migratory barn swallow Hirundo rustica