Journal of Animal Ecology 2009, 78, 109– 118 doi: 10.1111/j.1365-2656.2008.01484.

Predictors of floater status in a long-lived bird:

Blackwell Publishing Ltd

a cross-sectional and longitudinal test of hypotheses

Fabrizio Sergio*, Julio Blas and Fernando Hiraldo
Department of Applied Biology, Estación Biológica de Doñana, C.S.I.C. Avenida Maria Luisa s/n, Pabellón del Perú, Apdo
1056, E-41013 Sevilla, Spain

Summary
1. Few studies have been capable of monitoring the nonterritorial sector of a population because
of the typically secretive behaviour of floating individuals, despite the existing consensus over the
demographic importance of floating. Furthermore, there is almost no information on floating
behaviour for migratory species.
2. The factors that determine whether an individual will be a floater or a territory owner have been
framed into five, non-mutually exclusive hypotheses: (i) territory holders are morphologically
superior to floaters (resource-holding potential hypothesis); (ii) age confers skills and fighting
motivation which lead to social dominance and territory ownership (age hypothesis); (iii) occu-
pancy time of a site determines asymmetries in its knowledge, familiarity and value for potential
contenders (site-dominance hypothesis); (iv) contenders use an arbitrary rule to settle contests
leading to pre-defined cut-off points for a biologically meaningful trait (e.g. age, body size) separ-
ating floaters from territory holders (arbitrary convention hypothesis); and (v) floaters set up a
‘war of attrition’ at arbitrarily chosen territories (arbitrary attrition hypothesis).
3. We tested these hypotheses using long-term data on a long-lived, migratory raptor, the black kite
Milvus migrans Boddaert.
4. Floating status was best explained by the concerted action of mechanisms consistent with the
age and site-dominance hypotheses.
5. In both cross-sectional and longitudinal analyses, acquisition of a territory was determined by
a complex interaction between age and early arrival from migration, suggesting: (i) a progressive
incorporation of early arriving individuals in the territorial contingent of the population, and (ii)
the existence of an alternative restraint strategy of delayed territoriality mediated by long-term
acquisition of social dominance.
6. Such results suggested that territory acquisition was mediated by the establishment of site
dominance through pre-emption and, secondarily, despotism. In this population, age and arrival
date aligned individuals along a demographic continuum ranging from successful breeders mono-
polizing high-quality resources to floaters with no resources, consistent with the notion of floating as
an extreme form of breeding failure.
Key-words: dominance, floating, nonbreeders, territoriality, territory acquisition

Jonsson & Lundquivst 2003). Floating is usually viewed

Introduction
Many populations of territorial species are composed of both
territory holders and nonterritorial, usually nonbreeding
individuals, commonly defined as ‘floaters’. The existence
of such floating contingent has been demonstrated in taxa
as diverse as fish, mammals and birds (Newton 1992;
Solomon & Jaquot 2002; Voigt & Streich 2002; Brännäs,
*Correspondence author. E-mail: fsergio@ebd.csic.es, fabrizio.
sergio8@tin.it
© 2008 The Authors. Journal compilation © 2008 British Ecological Society
either as a constraint imposed by territory holders on a
‘doomed surplus’ of lower-quality individuals (Newton 1992),
or as a restraint integral to a strategy of delayed breeding
and fitness enhancement (Smith & Arcese 1989; Zach &
Stutchbury 1992). Whatever the underlying mechanism,
there is growing recognition of the importance of floating
behaviour for individual fitness and population regulation
(e.g. Hunt 1998; Kokko & Sutherland 1998; Newton 1998;
Penteriani et al. 2005). However, studying floaters has
invariably proven difficult because of their elusive behaviour
110 F. Sergio, J. Blas & F. Hiraldo

Table 1. Hypotheses explaining the determinants of the status of individuals as floaters or territory holders. Predictions have been adapted to
the model system under analysis (a migratory, long-lived bird)

Mechanism determining
Hypothesis territorial status Predictions Supported†

1. Resource-holding Morphological characteristics 1a Body measures predict territorial status‡

potential
2. Age (social dominance;
undivisive asymmetry)
3. Site dominance
(value asymmetry)
4. Arbitrary convention
(uncorrelated asymmetry)
5. Arbitrary attrition
hypothesis
yield capabilities to
acquire resources
Age confers
dominance status
Site occupancy and
familiarity determine value
asymmetries between
potential contenders
Territorial status determined
by an arbitrary rule (e.g. the
resident always wins)
Floaters set up a war of
attrition at a small number
of arbitrarily chosen
territories
1b Body measures predict the transition
from floater to territorial status§
1c In takeovers, body measures predict
success in territory acquisition
2a Floaters are younger than territory *
holders‡
2b Floaters are younger than ‘first-time *
territory holders’§
2c In takeovers, age predicts success in *
territory acquisition
3a Floaters arrive later from migration *
than territory holders‡
3b Floaters arrive later than ‘first-time *
territory holders’§
3c Longitudinal improvements in arrival *
date enhance floaters’ likelihood of
acquiring a territory¶
3d In takeovers, arrival date predicts
success in territory acquisition
4a In all comparisons, floaters and holders
are separated by a cut-off value (of body
size, age, or arrival date)
4b In takeovers, evicted and winners are
separated by a single cut-off value.
5a Floaters and holders do not differ for any
of the tested traits (body measures, age,
arrival)††
5b Floaters and ‘first-time territory holders’§
do not differ for any of the tested traits††
5c In takeovers, evicted and winners do
not differ for any of the tested traits††

†Symbol legend: *the prediction was supported by the Results of this study; ‡this prediction compares floaters with any type of territory holder;
§this prediction compares floaters with territory holders at their first year of successful territory acquisition (first-time ‘breeders’), that is,
between floaters that succeeded or not in acquiring a territory that year. ¶For individuals repeatedly sampled in consecutive years, longitudinal
improvements in arrival dates are more pronounced for individuals in transition between floater and territory holder status in successive years
than for individuals that remain floaters in consecutive years. ††This hypothesis predicts no clear separation between territory holders and
floaters in terms of quality traits such as age or body size, because floaters can eventually obtain a territory even if they lose all fights with owners
(Stutchbury 1991).

(review in Zach & Stutchbury 1992), to the point that they
have been referred to as a ‘shadow’ population living in a
secretive ‘underworld’ (Smith 1978; Rohner 1997). Further-
more, there is almost no information on floating behaviour
for migratory species (Porter 1988; Stutchbury 1991).
The factors that determine whether an individual will be a
floater or a territory owner have been usually framed into the
following five, non-mutually-exclusive hypotheses (summa-
rized in Table 1): (i) the resource-holding potential (RHP)
hypothesis suggests that, compared to floaters, territory
holders will have characteristics (larger size, strength etc.) that
promote higher capabilities to acquire and defend resources
(Mönkkönen 1990; Lozano 1994; Pryke & Andersson 2003).
This hypothesis predicts that territory holders will be larger
than floaters or morphologically different, and that such
© 2008 The Authors. Journal compilation © 2008 British Ecological Society, Journal of Animal Ecology, 78, 109–118
characteristics will lead to higher success during contests over
territories. (ii) The age (or social dominance) hypothesis
proposes that age (and its associated decline in residual
reproductive value) confers skills and fighting motivation
which lead to social dominance and thus to higher likelihood
of successfully acquiring and defending a territory (Rohwer,
Ewald & Rohwer 1981; Grafen 1987; Shutler & Weatherhead
1991). This hypothesis predicts that younger individuals
will be more likely to be floaters and to loose contests over
territories. Some authors consider this hypothesis as a special
case of RHP (Shutler & Weatherhead 1991). (iii) The site
dominance (or value asymmetry) hypothesis states that the
occupancy time of a site determines asymmetries in its know-
ledge, familiarity and value for potential contenders (Krebs
1982; Beletsky & Orians 1989; Tobias 1997). It predicts that,

all else being equal, territory holders will have spent more
time at the site than floating individuals which did not succeed
in acquiring it. (iv) The arbitrary convention (or uncorrelated
asymmetry) hypothesis postulates that contenders use an
arbitrary rule to settle contests, such ‘the resident always
wins’, or ‘the older always wins (Davies 1978; Rohwer 1982).
It predicts the existence of some pre-defined cut-off point for
a biologically meaningful trait (e.g. age, body size), below
which all individuals will be either all floaters or all territory
holders. (v) The arbitrary attrition hypothesis claims that
floating individuals choose an arbitrary number of territories
where they regularly intrude and challenge the owner in a
‘war of attrition’ won by who is able to persist the longest
(Stutchbury 1991). It predicts no clear separation between
territory holders and floaters in terms of quality traits such as
age or body size, because floaters can eventually obtain a
territory even if they lose all fights with owners (Stutchbury
1991). It is important to note that the above hypotheses are
not mutually exclusive.
Overall, the relative importance of these hypotheses in
predicting territorial status remains unclear, as inconsistent
results have even been obtained for different populations of
the same species (e.g. Eckert & Weatherhead 1987; Pryke &
Andersson 2003). Therefore, there is a great need for further
empirical tests in order to reach a consensus over a potential
general pattern. To date, studies on floaters have mainly
included: (i) model species which are small-sized, short-lived
and year-round residents (reviews in Newton 1992; and
Zach & Stutchbury 1992); (ii) theoretical models examining
the consequences of floating for individual fitness and
population regulation (e.g. Zach & Stutchbury 1992; Kokko
& Sutherland 1998); (iii) estimates of floater abundance
(Rohner 1996; Kenward et al. 2000; Newton & Rothery 2001);
(iv) removal experiments (Newton 1992); and (v) analyses
of territory acquisition where floater characteristics are
indirectly inferred by examining individuals at their first
incorporation into the territorial sector of the population (e.g.
Porter 1988; Arcese 1989). Given the secretiveness of floating,
unavoidably some of these methods yield an incomplete
picture of floater characteristics because of inherent biases
and assumptions. Commonly reported biases include: (i) the
difficulty to detect and monitor all floater age classes in
observational studies (e.g. Smith & Arcese 1989; Stutchbury
1991; Kenward et al. 2000); (ii) the incapability to distinguish
between nonbreeding territorial holders and true floaters
(e.g. Eckert & Weatherhead 1987; Kenward et al. 2000); (iii)
the younger age classes of floaters never reclaim territories and
are thus missed in some removal experiments (e.g. Shutler &
Weatherhead 1994); (iv) studies on the age of first territorial
acquisition unavoidably miss a large portion of floaters which
die without ever being territorial; (v) in removal experiments,
the contests between the removed owner trying to regain its
territory against its artificially induced replacement yield
unclear information because both contenders may believe to
be the owner, an unrealistic condition in a natural setting (e.g.
Shutler & Weatherhead 1992); (vi) floaters occupy areas and
habitats disjunct from those of breeders (e.g. Ferrer & Harte
© 2008 The Authors. Journal compilation © 2008 British Ecological Society, Journal of Animal Ecology, 78, 109– 118
Predictors of floater status 111
1997), which confounds direct comparisons between floater
and holder characteristics (e.g. differences in body condition
caused by habitat rather than territorial status). Therefore, as
noted by other authors (Newton 1992; Zach & Stutchbury
1992; Danchin & Cam 2002), there is a great need for empirical
studies where floaters co-exist with breeders and can be
observed before they become territorial and without removing
territory holders. This would represent a much needed com-
plement to previous experimental or indirect analyses.
Here we use a long-term data set on the population of a
long-lived migratory raptor, the black kite Milvus migrans
Boddaert, in which floaters co-exist with territory holders and
can be easily identified. Previous studies on floaters in raptors
have been extremely few, most of them focused on small-sized
short-lived species, none of them was conducted on migrants,
and they were mostly based on indirect methods (removal
experiments, or estimation of floater numbers based on breeders
demographic rates) (Village 1990; Rohner 1996; Ferrer &
Harte 1997; Hunt 1998; Kenward et al. 2000; Newton & Rothery
2001). None of them was framed as a test of the above hypotheses.
Methods
STUDY AREA
Kites were studied in a 430-km2 plot located in Doñana National
Park (south-western Spain). The landscape was characterized by
seasonally flooded marshland, scrublands, grasslands, and mobile
sand dunes along the sea shore.
STUDY SPECIES
The black kite is a medium-sized, migratory raptor. In our popula-
tion, the age of first territorial establishment ranges between 1 and 7
years and the maximum recorded life span is 23 years (Blas, Sergio
& Hiraldo 2009). On return from migration, individuals settle on
breeding territories, which usually include a nest site and an area of
20–200 m around it, while foraging occurs over wider, undefended
(communal) areas. The population was stable during the study
period at around 500 breeding pairs plus 400–500 nonbreeding indi-
viduals congregated in six communal roosts (Blas 2002). The roosts
are regularly distributed within the matrix of breeding territories
and the floaters use hunting areas that overlap widely with those
of breeders (i.e. floaters and breeders fully co-exist). Breeding
and natal dispersal distances are short (median 302 m and 4800 m
respectively) and extensive surveys suggest the absence of emigration
to other populations (Forero et al. 1999; Forero, Donázar & Hiraldo
2002).
FIELD PROCEDURES
In all analyses, we employ reproductive data collected between 1997
and 2000. However, ringing activities started earlier in the following
staggered manner: since 1964, as many nestlings as possible have
been marked every year with metal rings. Since 1986, the nestlings
were also marked with a white plastic ring with a black, three-
character alphanumeric code, which can be read by telescope
without disturbing the birds. In addition, since 1986, we used can-
non nets and padded leg-hold traps throughout the study area to
112 F. Sergio, J. Blas & F. Hiraldo
capture and mark adults with metal and plastic rings. Many of these
had been metal-banded as nestlings before 1986, which subsequently
allowed us to sample all the age classes in the population. For each
trapped adult, we measured body mass to the nearest 5 g, tarsus
length to the nearest 0·1 mm, and wing length and tail length to the
nearest 1·0 mm. Overall, up to 2000, 2367 nestlings were marked
only with metal bands, and 4257 nestlings plus 1076 adults were
marked with both metal and plastic bands. Individuals were sexed by
molecular analysis of a blood sample (Ellegren 1996) or by multiple
observations of copulation behaviour.
Marked territory holders were searched intensively by visiting
breeding territories every 4–5 days. A marked individual was assigned
to a specific territory when it was observed there in more than
three separate visits. In our population, floaters regularly gather to
sleep at communal roosts throughout the spring–summer (authors’
unpublished telemetry data on > 40 floaters). Roost sites were
checked in the evening from hides every 4–5 days throughout the
spring–summer and we classified as floaters all individuals that
were (i) detected at roosts in ≥ two visits separated by ≥ 15 days
in the period 15 April–15 June, and (ii) never observed holding a
breeding territory in that year. Telemetry data on > 40 confirmed
floaters indicated that such procedure could reliably identify floater
status.
Arrival date from migration was expressed as the date of first
observation of a marked individual, which is known to be a reliable
estimate of true settlement date (details in Sergio et al. 2007a).
The extremely frequent visits to territories allowed detection of
numerous cases in which a marked bird was observed to hold a
territory and to be replaced by another individual at a later date. In
all the cases in which this was directly observed, it was accompanied
by fighting, which sometimes escalated into talon-locking, with
individuals grasping each other’s talons in flight and falling to the
ground, where fighting continued until one of the contestants
escaped. For this reason, we assume that most of the observed
replacements were aggressive evictions. Hereafter, we refer to such
contests as ‘takeovers’, to the expelled bird as the ‘evicted’ and to the
new occupant as the ‘winner’.
HYPOTHESIS TESTING AND STATISTICAL ANALYSES
Because univariate metrics have been criticized as measures of
body size (Freeman & Jackson 1990), we estimated size by means
of the first axis (PC1, hereafter ‘body size’) of a principal components
analysis built using tarsus, wing and tail length. The PC1 explained
62% of the variation in size and had high positive loadings for wing
length (r = 0·87), tarsus length (0·52) and tail length (0·68). We used
body mass corrected for skeletal size as an index of condition. In
particular, as mass varied with year and body size, we standardized
it by using the residuals of such relationship as an index of body condi-
tion (see Sergio et al. 2007b, unpublished data and references therein).
The five testable hypotheses and their main predictions are sum-
marized in Table 1. Throughout, we assume that early arrival from
migration allowed individuals more time to acquire general familiarity
with the area and to establish site dominance through pre-emption
(Sergio et al. 2007a; predictions 3a–d).
To test the effect of age, arrival date and body measures on terri-
torial status (predictions 1a, 2a, 3a, 5a), we built a generalized linear
mixed model (GLMM, Littel et al. 1996) with year (as a random
factor), individual identity (as a random factor), sex, age, arrival
date, body size, body condition and their first-order interactions
as explanatory variables and territorial status (floater vs. territory
holder) as the dependent variable (n = 170 individuals, details in
© 2008 The Authors. Journal compilation © 2008 British Ecological Society, Journal of Animal Ecology, 78, 109–118
Table 3). Because body size and condition were never significant
and were available for a sub-sample of trapped birds, we re-ran the
analysis without fitting body measurements. This allowed to increase
the sample size to 714 individuals.
To further test whether body size, age or arrival date afforded a
higher likelihood of territory acquisition for floaters (prediction 1b,
2b, 3b), we built the same GLMM as above, but restricted the anal-
ysis to territory holders at their first year of successful territory
acquisition (first-time ‘breeders’, n = 111 individuals when including
the effect of body measures, n = 280 individuals without including
the effect of body measures, details in Table 3). While the previous
models compared floaters with all territory owners, this analysis dis-
criminated cross-sectionally between floaters that succeeded or not
in acquiring a territory that year.
To gain a further understanding of the link between arrival date
and territorial status, we conducted a longitudinal analysis of the
improvement in arrival dates in relation to age and territorial status
for individuals sampled in consecutive years (prediction 3c). For
each marked individual whose arrival date was known in consecutive
years, we: (i) calculated whether its arrival improved (became earlier
or not) in consecutive years; and (ii) tested by means of a binomial
test whether there was a preponderance of individuals improving
their arrival for each age class category (e.g. transition from age 1 to
2 years old, 2 to 3 years old, etc) and for each of three territorial
status categories: (i) floater, (ii) territory holder, and (iii) individuals
in transition between the status of floater and territory holder in
consecutive years.
To test whether age, arrival date or body measures predicted
successful territory acquisition during a takeover (predictions 1c, 2c,
3d, 4a, 5c), we built a GLM with takeover success as the dependent
variable and sex, age, arrival date, body size and body condition as
the explanatory variables. To increase sample size, for this analysis
we employed all the takeover episodes recorded between 1992 and
2006. Because winners of takeovers included both individuals
that were previously floaters and individuals that previously held a
territory elsewhere (individuals switching to another territory),
we conducted two separate analyses. In the first, we compared floaters
that succeeded in usurping a territory (i.e. which abandoned floater
status) with floaters that failed to evict the previous territory owner
during a takeover attempt (n = 39). In the second, we compared
floaters that failed to acquire a territory during a takeover attempt
with experienced territory holders that succeeded in switching to
another territory and evicting its previous owner (n = 50).
All models were built through a backward stepwise procedure
where the least significant terms or interactions were sequentially
removed until obtaining a minimal adequate model that only retained
significant effects at the 5% probability level (Crawley 1993). All
tests are two-tailed, statistical significance was set at α < 0·05, and
all means are given ± 1 SE.
Results
PRELIMINARY, DESCRIPTIVE TESTS
Compared to territory holders, floaters of both sexes were
smaller, younger and arrived later from migration (Table 2;
Figs 1 and 2). Body condition did not vary significantly
between floaters and territory holders (Table 2), but its
seasonal variations differed between the two status categories
(Fig. 3): the body condition of territory holders declined
rather constantly through the breeding cycle. In contrast, the
 Predictors of floater status 113

Table 2. Generalized linear mixed model logistic regressions (with binomial errors and a logit link function) testing the effect of sex, age, arrival
date from migration, body size and body condition on the territorial status of black kites in Doñana National Park (Spain). (a) comparison
between floaters and territory holders for 72 males and 98 females for which body measurements were available; (b) comparison between floaters
and territory holders for an extended sample of 361 males and 353 females for which body measurements were not available‡; (c) comparison
between floaters and individuals acquiring a territory for the first time in their life (n = 146 males and 134 females); (d) comparison between
floaters that succeeded in usurping a territory and floaters that failed to evict the previous territory owner during a takeover attempt (n = 17
males and 22 females); (e) comparison between floaters that failed to acquire a territory and experienced territory holders that succeeded in
switching to another territory and evicting its previous owner during a takeover attempt (n = 19 males and 31 females). Random factors (year
and individual identity) are only shown when their effect was significant

Percentage of
Variable Parameter estimate ± SE F P deviance explained

a. Dependent variable: territorial status 66·6

(with body measures, n = 170)†
Age −0·39 ± 0·04 91·0 < 0·001
Arrival date 0·02 ± 0·004 16·5 0·002
Intercept 5·22 ± 0·21 – –
b. Dependent variable: territorial status 53·6
(without body measures, n = 714)‡
Age −19·38 ± 1·86 108·7 < 0·001
Arrival date 2·18 ± 0·15 198·1 < 0·001
Interaction term: age × arrival date −0·23 ± 0·02 123·3 < 0·001
Intercept −176·08 ± 13·96 – –
c. Dependent variable: territorial status 34·6
(first time territory holders, n = 280)§
Age –0·13 ± 0·04 10·03 0·003
Arrival date 0·02 ± 0·001 3·27 0·075
Interaction term: age × arrival date 0·07 ± 0·03 4·75 0·033
Intercept 0·44 ± 0·13 – –
d. Dependent variable: floater success in a
takeover attempt (n = 39)¶
No variable entered the model – – –
e. Dependent variable: territorial status in 42·3
a takeover attempt (n = 50)††
Age 1·07 ± 0·03 36·9 < 0·0001
Intercept −6·21 ± 1·81 – –

†1, territory holder; 2, floater. This analysis employs only individuals for which body measurements were available. ‡1, territory holder;
2, floater. Because the effect of body measures was not significant in Model a, the model was repeated on a sample of individuals for which body
measurements were or not available. This allowed a marked increase in sample size. §1, individual holding a territory for the first time in its life;
2, floater. ¶1, floater which failed to usurp a territory; 2, floater that succeeded to evict a previous owner. ††1, floater which failed to usurp a
territory; 2, experienced territory holders that succeeded in switching to another territory and evicting its previous owner.

Table 3. Mean (± SE) age, arrival date from migration, body size and body condition of floaters and territory holders of both sexes in a black
kite population of Doñana National Park (Spain)

Variable Floater (n) Territory holder (n) t P

Age 2·39 ± 0·06 (400) (167†) 6·76 ± 0·13 (637) (63†) 24·93 < 0·0001
Arrival date 123·9 ± 1·6 (300) (134‡) 86·9 ± 0·6 (697) (146‡) 27·94 < 0·0001
Body size −0·49 ± 0·09 (128) (33§) 0·28 ± 0·06 (228) (141§) 7·49 < 0·0001
Body condition −11·85 ± 8·18 (127) (33 §) 6·72 ± 5·82 (224) (138§) 1·88 0·06

†Individuals of unknown sex included in the sample; ‡individuals of unknown sex or age included in the sample; §individuals of unknown age
included in the sample.

body condition of floaters declined from pre-laying to incu-

HYPOTHESIS 1: RESOURCE-HOLDING POTENTIAL
bation and steeply recovered thereafter. Because of such
different trends, in the following analyses body condition Once accounting for the effects of age and arrival date in a
was recalculated as the residuals of body mass on body size, GLMM, body size and condition did not predict territorial
year and breeding stage. status (in all tests, F ≤ 0·53, P ≥ 0·47; Table 3). Similarly,
© 2008 The Authors. Journal compilation © 2008 British Ecological Society, Journal of Animal Ecology, 78, 109– 118
114 F. Sergio, J. Blas & F. Hiraldo
Fig. 1. Age structure of the floater and territorial portions of the
black kite population of Doñana National Park (Spain).
Fig. 2. Arrival date from migration for floater and territorial black
kites in relation to age (Doñana National Park, Spain), when
including (a) all territory holders and (b) only individuals at their first
territorial establishment. Sample sizes in the seven age categories
were: (a) 1, 21, 78, 75, 81, 59, 287 for breeders and 47, 157, 48, 22, 14,
6, 6 for floaters; (b) 1, 21, 56, 25, 16, 8, 3 for breeders and 47, 157, 48,
22, 14, 6, 6 for floaters.
© 2008 The Authors. Journal compilation © 2008 British Ecological Society, Journal of Animal Ecology, 78, 109–118
Fig. 3. Body condition of floater and territorial black kites along
four successive stages of the breeding season (Doñana National Park,
Spain). Body condition was estimated as the residuals of a regression
of body mass on body size and year (see Methods). For floaters,
breeding stage was assigned on the basis of the yearly mean laying
date of breeding individuals, assuming an incubation period of 30
days and a nestling period of 48 days. For example, if a floater was
trapped and measured 15 days after the yearly mean laying date of the
population, its body condition was assigned to the incubation stage.
Sample sizes in the four breeding stages were: 107, 40, 72, 5 for
breeders and 55, 20, 42, 10 for floaters.
during territory takeovers, body size and condition did
not predict floater success in acquiring a territory (F ≤ 0·36,
P ≥ 0·56; Table 3). There was no support for the RHP hypothesis.
HYPOTHESIS 2: AGE (SOCIAL DOMINANCE)
Age entered all GLMMs discriminating between floaters
and territory holders (Table 3a,b,c; Fig. 1). Territory holders
were consistently older than floaters.
During takeovers, territory holders that succeeded in
switching to a new territory and evicting its previous owner
were older than floaters which failed to usurp a territory
(Table 3e). The age hypothesis was supported.
HYPOTHESIS 3: SITE DOMINANCE (VALUE
ASYMMETRY)
Arrival date from migration and its interaction with age
entered all GLMMs discriminating between floaters and
territory holders (Table 3a,b,c). In such cross-sectional
analyses, arrival date improved constantly with age for
territory holders, while for floaters, arrival date improved
steeply only during the first three years of life (Fig. 2).
In longitudinal analyses, arrival date improved significantly
for floaters between 1 and 2 years old and marginally so
between age 2–3 years (Table 4, Fig. 4). For floaters that
succeeded in acquiring a territory (successful transition to
territoriality), arrival date improved between ages 1–2 and
 Predictors of floater status 115

Table 4. Percentage of cases in which the arrival date from migration improved from 1 year to the next for a given individual. For territory
holders, only data up to 5 years of age were considered, in order to make them comparable to the data on floaters. Improvements above 75% are
highlighted in bold

Transition between floater

Age Floaters (n†) P‡ and territory holder (n†)§ P‡ Territory holders (n†) P‡

1–2 90% (10) 0·021 100% (11) 0·001 −¶ −¶

2–3 79% (14) 0·057 97% (29) < 0·001 50% (4) −¶
3–4 40% (5) −¶ 75% (8) 0·28 69% (32) 0·052
4–5 −¶ −¶ 100% (5) 0·06 44% (32) 0·60
Pooled 76% (29) 0·009 94% (53) < 0·001 56% (68) 0·40

†Number of individuals, each one sampled in two successive years; ‡tested by means of a binomial test; §individuals sampled in the year of their
first territorial establishment and in the previous year (their last year as floaters); ¶not available, or sample size too low for any meaningful
estimate or test.

Furthermore, in at least two cases, an individual known to

hold a territory in previous years was evicted from it and the
process was accompanied by much fighting. Therefore, resi-
dence in previous years did not guarantee successful defence
from eviction by floaters, falsifying the ‘resident always wins’
arbitrary rule.

HYPOTHESIS 5: ARBITRARY ATTRITION

In all comparisons, age and/or arrival date predicted

Fig. 4. Longitudinal improvement in arrival date from migration for
floaters, territory holders and individuals in transition between the
status of floater and territory holder (i.e. between the year before and
after their first acquisition of a territory). The y-axis represents the
percentage of individuals that improved their arrival date in
successive years (i.e. how many birds arrived earlier in the second of
the 2 years in which they were consecutively sampled).
2–3 (Table 4, Fig. 4). Similar rates of improvement were apparent
for ages 3–4 and 4–5, but sample size was too low to reach
significance (Table 4, Fig. 4). For territory holders, there was
only a weak indication of longitudinal improvements in arrival
dates from migration (Table 4, Fig. 4). Therefore, improvements
in arrival dates were minimum for territory holders, inter-
mediate for floaters and maximum for individuals that man-
aged to make the transition between floating and territorialism.
During takeovers, arrival date from migration was never a
predictor of floater success in acquiring a territory. Therefore,
the site-dominance hypothesis was partially supported.
HYPOTHESIS 4: ARBITRARY CONVENTION
(UNCORRELATED ASYMMETRY)
There was no evidence of any cut-point above or below which
all individuals were either floaters or breeders for any of the
examined variables and in any comparison. In the four cases
in which both the evicted and the winner were marked and in
which fighting was directly observed (i.e. cases of sure takeo-
ver attempts), there was no clear separation between contend-
ers in terms of body measures, age or arrival date, all variables
overlapping considerably.
© 2008 The Authors. Journal compilation © 2008 British Ecological Society, Journal of Animal Ecology, 78, 109– 118
territorial status (Table 3), disproving predictions 5a,b,c.
Furthermore, in an ongoing study, telemetry data from
> 40 floaters showed that they occupied very large home
ranges, encompassing 100–250 breeding territories and
overlapping widely with those of all other floaters
(authors’ unpublished data). Therefore, floaters did not
seem to queue for specific territories contained within
restricted home ranges, nor intruded consistently in certain
territories over others. The arbitrary attrition hypothesis
was not supported.
Discussion
Of the five tested hypotheses, there was no support for the
RHP, arbitrary convention and arbitrary attrition hypothesis.
The idea that morphological characteristics (i.e. RHP) can
determine dominance and territorial status is contradictory,
having received support from some studies (Mönkkönen 1990;
Lozano 1994; Pryke & Andersson 2003; Sol et al. 2005) but
not others (Eckert & Weatherhead 1987; Shutler & Weather-
head 1991, 1992; Peer, Robertson & Kempenaers 2000). The
arbitrary convention hypothesis has been disproved in
various empirical investigations (e.g. Beletsky & Orians 1989;
Smith & Arcese 1989), has been criticized on theoretical
grounds by Grafen (1987), and seems to be poorly suited to
avian territorial systems and long-lived species which occupy
territories for long periods (Grafen 1987; Beletsky & Orians
1989). Finally, the arbitrary attrition hypothesis is likely to
apply only to certain species, because other previous studies
have also reported that floaters do not hold restricted home
ranges encompassing only a few territories (e.g. Shutler &
Weatherhead 1992, 1994).
116 F. Sergio, J. Blas & F. Hiraldo
In contrast, there was support for the age hypothesis and,
partly, for the site-dominance hypothesis. Age can lead to a
higher likelihood of territory acquisition in five non-exclusive
ways: (i) age can directly confer or be associated with higher
dominance status, as shown in many studies (e.g. Smith
et al. 1980; Rohwer et al. 1981; Cronin & Field 2007); (ii)
older individuals have lower residual reproductive value and
are thus expected to be ready to fight harder for a territory
(Grafen 1987; Shutler & Weatherhead 1994). This is con-
sistent with the age effect that we observed in takeover fights.
(iii) Conversely, in long-lived species with dangerous weap-
onry, like raptors, younger individuals are more likely to
refrain from physical confrontations given the risk of lethal
injuries and their likelihood of surviving to the next breeding
season. (iv) In long-lived species, the cost of reproduction is
often high (e.g. Tavecchia et al. 2001), selecting for delayed
territoriality as a voluntary restraint in order to enhance
longevity and the quality of the territory eventually acquired
(Zach & Stutchbury 1992; Hunt 1998; Kokko & Sutherland
1998). Consistent with this idea, in our population delayed
breeding led to higher longevity and older birds monopolized
the best territories (Sergio et al. 2007a, unpublished; Blas et al.
2009). Such strategy may contribute to accumulate older
birds in the territorial sector of the population. (v) Finally,
even assuming that vacancies are filled randomly by the first
individuals encountering them, older individuals will have
accumulated more time and thus a higher probability to
chance upon a vacancy than younger ones, contributing to
the older age-structure of territory holders (Shutler & Weath-
erhead 1994). However, this mechanism is unlikely in our
model-system, because it assumes that territory holders do
not suffer takeovers (Eckert & Weatherhead 1987) and it
cannot explain why the winners of takeovers were older than
the evicted individuals.
In agreement with the above ideas, most previous studies
have shown that floaters are usually younger than territory
holders (Porter 1985; Smith & Arcese 1989; Shutler & Weath-
erhead 1991; Sol et al. 2005; Blanco et al. 2007; but see also
Stutchbury 1991). However, in our population age alone
could not be the sole predictor of territorial status because
there was (i) wide overlap in the age structure of floaters and
holders (Fig. 1), and (ii) pronounced variation among indi-
viduals in the age of first territorial establishment (range: 1–7
years old). Therefore, two individuals of the same age could
still belong to different categories of territorial status.
Variation in arrival dates may help to explain such differences
in territorial status within age classes (Fig. 2, see below).
Independent of age, early arriving birds had a higher
likelihood of acquiring a territory (Table 4, Fig. 2). Early
arrival from migration may allow floaters to: (i) find more
numerous vacant territories; (ii) have more time to encounter
vacancies and gain familiarity with the general area; (iii)
pre-empt territories (sensu Rodenhouse, Sherry & Holmes
1997) for long enough to develop subsequent site dominance.
All this is consistent with the site-dominance hypothesis. On
the other hand, in our system, occupying a territory too early
could lead to subsequent eviction by an older individual,
© 2008 The Authors. Journal compilation © 2008 British Ecological Society, Journal of Animal Ecology, 78, 109–118
suggesting that the effect of site dominance can be overcome
by direct aggression coupled with social dominance. This
implies a concerted action of the mechanisms proposed by
the site-dominance and age hypotheses. In agreement with
this, territorial status was consistently predicted by the
interaction between age and arrival date.
In particular, the comparison of Figs 2 and 4 suggested
a complex interplay between age and arrival in determining
territorial status. In the cross-sectional analysis shown in
Fig. 2a, territory holders constantly improved their arrival
with increasing age, particularly so in the youngest age
classes. However, in the longitudinal analysis of Fig. 4,
territory holders did not show such pronounced improve-
ment in arrival dates. By exclusion, this implies a process of
progressive incorporation of early arriving birds into the
territorial pool of the population, confirmed by the longitu-
dinal analyses in which the transition from floater to terri-
torial status coincided with pronounced improvements in
arrival dates (Table 4, Fig. 4). In turn, such progressive
removal of early arriving birds from the floater sector of the
population may explain why floaters did not show any further
cross-sectional improvement in arrival date beyond the age
of 3 years (Fig. 2a). Therefore, only late-arriving birds seemed
to accumulate in the older floater age classes. Such ‘late, old’
floaters could be either low-quality individuals incapable
to advance their arrival, or individuals with an alternative,
delayed breeding strategy. The latter explanation seems more
likely, because such delayed breeders had a better nutritional
status (Blas 2002) and a higher longevity (Blas et al. 2009),
which is usually a reliable predictor of lifetime reproductive
success (Newton 1989).
The above reasoning is consistent with two alternative
‘career-decisions’ co-existing in the population: (i) an invest-
ment in early arrival and early incorporation into the ter-
ritorial sector of the population; and (ii) an investment in
long-term survival and in acquisition of social dominance.
The first strategy entails a route to territoriality mainly medi-
ated by the establishment of site dominance, while the second
may rely more on the dominance status conferred by age.
However, considering that 84% of the individuals become
territorial within 4 years of age, the site-dominance route to
territoriality seems to be the predominant one in the popula-
tion. All the above reinforces the idea of previous analyses,
and extends it to the floater contingent, that access to breed-
ing territories in this population is mediated by pre-emption
and, secondarily, despotism (see Sergio et al. 2007a).
To date, we are aware of only one previous study which
focused on the potential link between arrival date and terri-
torial status (Porter 1985). Its results were remarkably similar
to ours: in Kittiwakes Rissa tridactyla, improvements in
arrival dates were particularly pronounced in the younger age
classes, and, within each age class, floaters arrived later than
established breeders. Longitudinal improvements in arrival
were particularly marked for individuals passing from floater
to territorial status. Although more studies will be needed,
the above observations suggest that the patterns found for our
population may be common to many other migratory species.

Furthermore, the five hypotheses we tested could also be valid
for resident, nonmigratory species. However, site dominance
would be attained through different mechanisms. For exam-
ple, rather than investing in early arrival, floaters of resident
species could invest in frequent visits to breeding areas in
order to improve their site familiarity and their chances of
pre-empting a vacancy. Dispersal movements consistent with
such idea have been observed in some studies (Ferrer & Harte
1997; Kenward et al. 2000).
In conclusion, annual breeding in the study population was
competitively structured by age along a sequence of arrival
dates. Older breeders arrived early and pre-emptively mono-
polized resources. Lower quality individuals (e.g. younger)
arrived progressively later and accessed lower-quality resources
(Sergio et al. 2007a). At the end-tail of such demographic
continuum were the even younger and later arriving floaters,
consistent with the previously proposed notion of floating as
an extreme form of breeding failure (e.g. Danchin & Cam
2002). Aggressive takeovers reinforced the age-structured
determination of territorial status and, possibly, weakened
the advantages provided by early arrival. This may have partly
insured the honesty of arrival date as a cue to individual
quality and site dominance for potential, later-arriving
contenders. Exceptions to this general pattern existed in the
form of alternative strategies, where some individuals were
capable of acquiring territories by arriving very early despite
a young age, or by investing in long-term acquisition of social
dominance, suggesting the importance of individual varia-
tion in life-history decisions through time.
Acknowledgements
We thank F.G. Vilches, R. Baos, S. Cabezas, J.A. Donázar, M.G. Forero,
G. García, L. García, M. Guerrero, and A. Sánchez for help in the field and
L. Marchesi, K. Norris, an anonymous associate editor and two anonymous
referees for comments on a previous draft of the manuscript. Part of this study
was funded by the research projects PB96-0834 of the Dirección General de
Investigación Científica y Tecnológica, JA-58 of the Consejería de Medio
Ambiente de la Junta de Andalucía and by the Excellence Project RNM 1790
of the Junta de Andalucía.
References
Arcese, P. (1989) Territory acquisition and loss in male song sparrows. Animal
Behaviour, 37, 45–55.
Beletsky, L.D. & Orians, G.H. (1989) Territoriality among male red-winged
blackbirds III. Testing hypotheses of territorial dominance. Behavioral
Ecology and Sociobiology, 24, 333–339.
Blanco, G., Lemus, J.A., Frías, O., Grande, J., Arroyo, B., Martínez, F. &
Beniandrés, N. (2007) Contamination traps as trans-frontier management
challenges: new research on the impact of refuse damps on the conservation
of migratory avian scavengers. Environmental Research Trends (ed. A.M. Cato),
pp. 153–204. Nova Science Publishers, Hauppauge, New York.
Blas, J. (2002) Age and reproduction in the black kite (Milvus migrans). PhD
dissertation thesis, University of Madrid, Madrid, Spain.
Blas, J., Sergio, F. & Hiraldo, F. (2008) Age-related improvement in reproduc-
tive performance in a long-lived raptor: a cross-sectional and longitudinal
study, Ecography, in press.
Brännäs, E., Jonsson, S. & Lundqvist, H. (2003) Influence of food abundance
on individual behaviour strategy and growth rate in juvenile brown trout
(Salmo trutta). Canadian Journal of Zoology, 81, 684–691.
Crawley, M.J. (1993) GLIM for Ecologists. Blackwell Scientific Publications, Oxford.
Cronin, A.L. & Field, J. (2007) Social aggression in an age-dependent domin-
ance hierarchy. Behaviour, 144, 753–765.
© 2008 The Authors. Journal compilation © 2008 British Ecological Society, Journal of Animal Ecology, 78, 109– 118
Predictors of floater status 117
Danchin, E. & Cam, E. (2002) Can non-breeding be a cost of breeding dis-
persal? Behavioral Ecology and Sociobiology, 51, 153–163.
Davies, N.B. (1978) Territorial defense in the speckled wood butterfly (Pararge
aegeria): the resident always wins. Animal Behaviour, 26, 138–147.
Eckert, C.G. & Weatherhead, P.J. (1987) Owners, floaters and competitive
asymmetries among territorial red-winged blackbirds. Animal Behaviour,
35, 1317–1323.
Ellegren, H. (1996) First gene on the avian w chromosome (CHD) provides a
tag for universal sexing of non-rattie birds. Proceedings of the Royal Society
B. Sciences, 263, 1635–1641.
Ferrer, M. & Harte, M. (1997) Habitat selection by immature Spanish imperial
eagles during the dispersal period. Journal of Applied Ecology, 34, 1359–1364.
Forero, M.G, Donázar, J.A., Blas, J. & Hiraldo, F. (1999) Causes and con-
sequences of territory changes and breeding dispersal distances in the Black
Kite. Ecology, 80, 1298–1310.
Forero, M.G., Donázar, J.A. & Hiraldo, F. (2002) Causes and fitness con-
sequences of natal dispersal in a population of black kites. Ecology, 83,
858–872.
Freeman, S. & Jackson, W.M. (1990) Univariate metrics are not adequate to
measure avian body size. Auk, 107, 69–74.
Grafen, A. (1987) The logic of divisively asymmetric contests: respect for
ownership and the desperado effect. Animal Behaviour, 35, 462–467.
Hunt, W.G. (1998) Raptor floaters at Moffat’s equilibrium. Oikos, 82, 191–
197.
Kenward, R.E., Walls, S.S., Hodder, K.H., Pahkala, M., Freeman, S.N. &
Simpson, V.R. (2000) The prevalence of non-breeders in raptor populations:
evidence from rings, radio-tags and transect surveys. Oikos, 91, 271–279.
Kokko, H. & Sutherland, W.J. (1998) Optimal floating and queuing strategies:
consequences for density dependence and habitat loss. American Naturalist,
152, 354–366.
Krebs, J.R. (1982) Territorial defence in the great tit (Parus major): do residents
always win? Behavioral Ecology and Sociobiology, 11, 185–194.
Littell, R.C., Milliken, G.A., Stroup, W.W. & Wolfinger, R.D. (1996) SAS
System for Mixed Models. SAS Institute Inc., Cary, North Carolina.
Lozano, G.A. (1994) Size, condition, and territory ownership in male tree
swallows (Tachycineta bicolour). Canadian Journal of Zoology, 72, 330–333.
Mönkkönen, M. (1990) Removal of territory holders causes influx of small-
sized intruders in passerine bird communities in northern Finland. Oikos,
57, 281–288.
Newton, I. (ed.) (1989) Lifetime Reproduction in Birds. Academic Press,
London.
Newton, I. (1992) Experiments on limitation of bird numbers by territorial
behaviour. Biological Reviews, 67, 129–173.
Newton, I. (1998) Population Limitation in Birds. Academic Press, London.
Newton, I. & Rothery, P. (2001) Estimation and limitation of numbers of floaters
in a Eurasian Sparrowhawk population. Ibis, 143, 442–449.
Peer, K., Robertson, R.J. & Kempenaers, B. (2000) Reproductive anatomy and
indices of quality in male tree swallows: the potential reproductive role of
floaters. Auk, 117, 74–81.
Penteriani, V., Otalora, F., Sergio, F. & Ferrer, M. (2005) Environmental
stochasticity in dispersal areas can explain the ‘mysterious’ disappearance
of breeding populations. Proceedings of the Royal Society B: Biological
Sciences, 272, 1265–1269.
Porter, J.M. (1988) Prerequisites for recruitment of Kittiwakes Rissa tridactyla.
Ibis, 130, 204–215.
Pryke, S.R. & Andersson, S. (2003) Carotenoid-based epaulettes reveal male
competitive ability: experiments with resident and floater red-shouldered
widowbirds. Animal Behaviour, 66, 217–224.
Rodenhouse, N.L., Sherry, T.W. & Holmes, R.T. (1997) Site-dependent
regulation of population size: a new synthesis. Ecology, 78, 2025–2042.
Rohner, C. (1996) The numerical response of great horned owls to the snow-
shoe hare cycle: consequences of non-territorial ‘floaters’ for demography.
Journal of Animal Ecology, 65, 359–370.
Rohner, C. (1997) Non-territorial ‘floaters’ in great horned owls: space use
during a cyclic peak of snowshoe hares. Animal Behaviour, 53, 901–912.
Rohwer, S. (1982) The evolution of reliable and unreliable badges of fighting
ability. American Zoologist, 22, 531–546.
Rohwer, S.A., Ewald, P.W. & Rohwer, F.C. (1981) Variation in size, dominance
and appearance within and between sex and age classes in Harris’ sparrows.
Journal of Field Ornithology, 52, 291–303.
Sergio, F., Blas, J., Forero, M.G., Donázar, J.A. & Hiraldo, F. (2007a)
Sequential settlement and site-dependence in a migratory raptor. Behavioral
Ecology, 18, 811–821.
Sergio, F., Blas, J., Forero, M.G., Donázar, J.A. & Hiraldo, F. (2007b)
Size-related advantages for reproduction in a slightly dimorphic raptor:
opposite trends between the sexes. Ethology, 113, 1141–1150.
118 F. Sergio, J. Blas & F. Hiraldo
Shutler, D. & Weatherhead, P.J. (1991) Owner and floater red-winged blackbirds:
determinants of status. Behavioral Ecology and Sociobiology, 28, 235–241.
Shutler, D. & Weatherhead, P.J. (1992) Surplus territory contenders in male
red-winged blackbirds: where are the desperados? Behavioral Ecology and
Sociobiology, 31, 97–106.
Shutler, D. & Weatherhead, P.J. (1994) Movement patterns and territory acqui-
sition by male red-winged blackbirds. Canadian Journal of Zoology, 72, 712–720.
Smith, J.N.M. & Arcese, P. (1989) How fit are floaters? Consequences of
alternative territorial behaviours in a nonmigratory sparrow. American
Naturalist, 133, 830–845.
Smith, J.N.M., Montogomerie, R.D., Taitt, M.J. & Yom-Tov, Y. (1980) A
winter feeding experiment on an island song sparrow population. Oecologia,
47, 164–170.
Smith, S.M. (1978) The ‘underworld’ in a territorial sparrow: adaptive strategy
for floaters. American Naturalist, 112, 571–582.
Sol, D., Elie, M., Marcoux, M., Christovsky, E., Porcher, C. & Lefebvre, L.
(2005) Ecological mechanisms of a resource polymorphism in Zenaida
Doves of Barbados. Ecology, 86, 2397–2407.
© 2008 The Authors. Journal compilation © 2008 British Ecological Society, Journal of Animal Ecology, 78, 109–118
Solomon, N.G. & Jaquot, J.J. (2002) Characteristics of resident and wandering
prairie voles, Microtus ochrogaster. Canadian Journal of Zoology, 80, 951–
955.
Stutchbury, B.J. (1991) Floater behaviour and territory acquisition in male
purple martins. Animal Behaviour, 42, 435–443.
Tavecchia, G., Pradel, R., Boy, V., Johnson, A.R. & Cézilly, F. (2001) Sex- and
age-related variation in survival and cost of first reproduction in Greater
Flamingos. Ecology, 82, 165–174.
Tobias, J. (1997) Asymmetric territorial contests in the European robin: the role
of settlement costs. Animal Behaviour, 54, 9–21.
Village, A. (1990) The Kestrel. Poyser, Calton, UK.
Voigt, C.C. & Streich, W.J. (2002) Queuing for harem access in colonies of the
greater sac-winged bat. Animal Behaviour, 65, 149–156.
Zach, S. & Stutchbury, B.J. (1992) Delayed breeding in avian social systems: the
role of territory quality and ‘floater’ tactics. Behaviour, 123, 194–219.
Received 17 April 2008; accepted 1 September 2008
Handling Editor: Jonathan Wright