Raptor Nest Decorations Are a Reliable Threat Against Conspecifics

Abstract

Individual quality is often signaled by phenotypic flags, such as bright plumage patches in birds. Extended phenotype signals can similarly show quality, but in these cases the signals are external to the individual, often taking the form of objects scavenged from the environment. Through multiple manipulative experiments, we showed that objects used for nest decoration by a territorial raptor, the black kite (Milvus migrans), act as reliable threats to conspecifics, revealing the viability, territory quality, and conflict dominance of the signaler. Our results suggest that animal-built structures may serve as signaling devices much more frequently than currently recognized.

Full text

DOI: 10.1126/science.1199422
, 327 (2011);331 Science , et al.F. Sergio
Raptor Nest Decorations Are a Reliable Threat Against Conspecifics
This copy is for your personal, non-commercial use only.
clicking here.colleagues, clients, or customers by , you can order high-quality copies for yourIf you wish to distribute this article to others
here.following the guidelines can be obtained byPermission to republish or repurpose articles or portions of articles
): January 20, 2011 www.sciencemag.org (this infomation is current as of
The following resources related to this article are available online at
http://www.sciencemag.org/content/331/6015/327.full.html
version of this article at: including high-resolution figures, can be found in the onlineUpdated information and services,
http://www.sciencemag.org/content/suppl/2011/01/18/331.6015.327.DC1.html
can be found at: Supporting Online Material
http://www.sciencemag.org/content/331/6015/327.full.html#ref-list-1
, 6 of which can be accessed free:cites 23 articlesThis article
http://www.sciencemag.org/cgi/collection/ecology
Ecology subject collections:This article appears in the following
registered trademark of AAAS. is aScience2011 by the American Association for the Advancement of Science; all rights reserved. The title CopyrightAmerican Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005.
(print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by theScience
on January 20, 2011www.sciencemag.orgDownloaded from

15. S. Popy, L. Bordignon, R. Prodon, J. Biogeogr. 37,
57 (2010).
16. J. Lenoir, J. C. Gégout, P. A. Marquet, P. de Ruffray,
H. Brisse, Science 320, 1768 (2008).
17. J. Lenoir et al., Ecography 33, 295 (2010).
18. C. I. Millar, R. D. Westfall, D. L. Delany, J. C. King,
L. J. Graumlich, Arct. Alp. Res. 36, 181 (2004).
19. Materials and methods are available as supporting
material on Science Online.
20. A. E. Wieslander, J. For. 33, 877 (1935).
21. A. E. Wieslander, Madrono 3, 140 (1935).
22. C. Daly et al., Int. J. Climatol. 28, 2031 (2008).
23. A. J. Eckert, M. L. Eckert, Madrono 54, 117 (2007).
24. J. M. Lenihan, R. Drapek, D. Bachelet, R. P. Neilson,
Ecol. Appl. 13, 1667 (2003).
25. K. Zhang et al., J. Hydrol. (Amsterdam) 379, 92 (2009).
26. Supported by NSF (BCS-0819430 to S.Z.D. and
BCS-0819493 to J.A.G.). J.T.A. was supported by NSF’s
Experimental Program to Stimulate Competitive Research
(EPSCoR) (EPS-0814387). A.R.M. was supported by U.S.
Department of Agriculture Forest Service’s(USFS’s) Rocky
Mountain Research Station (JV11221635-201). We thank
C. Daly for providing interpolated climate data; M. Kelly
for Vegetation Type Mapping project data; C. Ramirez
(Region 5 USFS Remote Sensing Lab) for the USFS Forest
Inventory and Analysis data (plot locations were not
provided to the authors); J. Thorne for compiling modern
data sources; and S. Running for helpful comments on
the manuscript.
Supporting Online Material
www.sciencemag.org/cgi/content/full/331/6015/324/DC1
Materials and Methods
Figs. S1 to S4
Tables S1 to S3
References
3 October 2010; accepted 21 December 2010
10.1126/science.1199040
Raptor Nest Decorations Are a
Reliable Threat Against Conspecifics
F. Sergio,
1
*†J. Blas,
1
†G. Blanco,
2
A. Tanferna,
1
L. López,
1
J. A. Lemus,
1
F. Hiraldo
1
Individual quality is often signaled by phenotypic flags, such as bright plumage patches in birds.
Extended phenotype signals can similarly show quality, but in these cases the signals are external
to the individual, often taking the form of objects scavenged from the environment. Through
multiple manipulative experiments, we showed that objects used for nest decoration by a territorial
raptor, the black kite (Milvus migrans), act as reliable threats to conspecifics, revealing the
viability, territory quality, and conflict dominance of the signaler. Our results suggest that animal-
built structures may serve as signaling devices much more frequently than currently recognized.
Animals transfer information through ana-
tomical, physical, and behavioral traits,
such as brightly colored body patches,
extravagant ornaments, or size-revealing calls
(1–3). Some species go to further extents by col-
lecting objects to build external signaling struc-
tures or “extended phenotype signals”(4). Such
external signals may involve higher cognitive
abilities (5) and be subject to cultural transmis-
sion (6). Further, our ability to address questions
about such signals are facilitated by the facts that
(i) they are often flexible and less constrained
than bodily traits (4) and (ii) their constituent
materials may be readily available for experi-
mental manipulation (7,8). Despite the broader
appeal, extended phenotype signals have been
examined in only a handful of passerine bird
and fish species, and these were almost exclu-
sively structures built solely for mate attraction
1
Department of Conservation Biology, Estación Biológica de
Doñana, Consejo Superior de Investigaciones Científicas (CSIC),
C/ Americo Vespucio, 41092 Seville, Spain.
2
Department of Evo-
lutionary Ecology, Museum of Natural Sciences, CSIC, C/ José
Gutiérrez Abascal 2, 28006 Madrid, Spain.
*To whom correspondence should be addressed. E-mail:
fsergio@ebd.csic.es
†These authors contributed equally to this work.
B
A
Fig. 1. Decoration of nests by black kites (Milvus migrans). (A) Highly decorated nest and (B)
selection of materials by kites relative to their availability in the environment. Kites show a preference
for white (c
2
=270.0,P< 0.0001) plastic (c
2
= 57.0, P< 0.0001), whereas all other materials and
colors are used at rates equal to (other materials c
2
=0.1,P= 0.76) or less than (paper c
2
= 68.1, P<
0.0001; cryptic colors c
2
= 77.2, P< 0.0001; other colors c
2
=148.3,P< 0.0001) their availability
in the environment. [Photo credit: F. Sergio]
www.sciencemag.org SCIENCE VOL 331 21 JANUARY 2011 327
REPORTS
on January 20, 2011www.sciencemag.orgDownloaded from

and courtship, such as the bowers of bower-
birds (5–8). However, several hundreds of spe-
cies erect nest structures in manners that could
be compatible with extended phenotype sig-
naling functions. Examples include the addi-
tion of man-made objects, stones, green plants,
snake skins, dung, or carnivore scats by passer-
ines, storks, swans, penguins, corvids, raptors,
and fish and the building of multiple alternative
nests by wrens, weavers, and raptors (4,9–14).
Therefore, although animals are thought to build
structures for communication only rarely (12),
nests could be signaling devices more frequently
than currently recognized.
We combined experimental manipulations and
long-term observations of a long-lived raptor to
show that nest decoration can act as an honest
threat to conspecifics, revealing the viability, ter-
ritory quality, and conflict dominance of the sig-
naler. Our model species, the black kite (Milvus
migrans), is a medium-sized migratory raptor. Pairs
occupy breeding territories that may vary markedly
in resource quality (15). Floating individuals are
prevented from breeding through social dominance
but may attempt violent territory takeovers, which
sometimes succeed (16) (fig. S1). Trespassing can
be frequent, especially in pre-incubation, and is
also aimed at stealing food resources obtained by
the residents. On return from migration after pair
formation, both partners contribute to nest build-
ing. One to 2 weeks before laying, they may both
place various man-made materials in the nest,
typically white plastic objects (Fig. 1A). The
objects seem placed so as to be highly visible and
are collected during the peak period of territory
trespassing and long after pair formation. Thus, we
hypothesized that they serve as an extended phe-
notype signal of threat against intruders (table S1).
We used a well-studied (15,16) (fig. S2)
population of marked, known-age (1 to 25 years
old) kites in Doñana National Park. To investigate
the signaling function of nest decorations (17), we (i)
assessed decoration levels at 127 nests built by
individuals of known age and body size, which
were subsequently intensively monitored to esti-
mate their viability (here defined by subsequent
survival and offspring production); (ii) compared
nest decorations with objects available in the
environment; (iii) offered objects to marked indi-
viduals to examine their propensity to use decora-
tions; (iv) estimated trespass rates and aggression
of intruders for both nonexperimental and exper-
imental pairs (i.e., those whose nest decorations
had been augmented); (v) tested whether nest
decorations predicted the efficiency of resource
monopolization mediated by aggression within a
competitive feeding environment away from the
nest; and (vi) examined the egg predation rates of
experimental dummy kite nests with and without
decorations. Through this series of experiments,
we specifically tested for the presence of an hon-
est extended signal by addressing four hypothe-
ses, as proposed by Searcy and Nowicki (3): (i) is
the signal reliable; (ii) do receivers respond to the
signal in a way that benefits signalers; (iii) is the
signal costly, and is the cost higher for lower-
quality individuals; and (iv) is there deceit (17)
(table S1)?
Nest decorations were observed in 77% of the
nests (n= 127), appeared in the last 20 days
before laying, and declined slightly during the
incubation period and markedly thereafter (fig.
S3). Both sexes were equally likely to carry
decorations (males 15 times, females 12; bino-
mial test, P= 0.70). When compared to avail-
ability in transects (Fig. 1B), kites overselected
plastic materials (c
2
= 57.0, P< 0.0001) and
white color (as assessed by human vision; c
2
=
270.0, P< 0.0001). All other materials and colors
were avoided or used according to availability
(Fig. 1B). When birds were offered a set of ex-
perimental objects in equal availability (17), 29
of 33 pairs that collected objects placed only
white plastic items in their nest, two collected
both white and transparent items, and another
two collected both white and green items. White
was overselected ( c
2
= 100.7, P< 0.0001), and
the other colors were avoided (green and trans-
parent c
2
= 24.5, P< 0.0001).
Is the signal reliable (hypothesis i)? Nest de-
coration increased with territory quality and body
condition and exhibited a quadratic (parabolic)
relationship with age (table S2 and Figs. 2 and 3):
The level of decoration increased up to ages 10 to
12 and declined thereafter with senescence (Figs. 2
and 3). Similarly, 37% of the pairs refrained from
collecting any offered item despite its prolonged
availability. These animals had lower-quality ter-
ritories and differed in age from individuals that
used the offered items (table S3): Individuals that
refrained from signaling were very young or very
old, whereas those that used the decorations to
signal were predominantly in prime age (7 to 12
Fig. 2. Change in mean (T1SE)nest
decoration by black kites (black line)
with age (n= 127). Placement of freely
available items (gray) was highest for
birds in prime age (7 to 12 years old)
andlowestforyoungandoldindivid-
uals (n=49).
Fig. 3. Change in level of nest decoration in black kites. Level of decoration was (A) lowest for young birds, (B) increased up to ages 10 to 12 years old, and
(C) declined thereafter with senescence. [Photo credit: F. Sergio]
21 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org
328
REPORTS
on January 20, 2011www.sciencemag.orgDownloaded from

years old; Fig. 2). The rate and success of attacks on
trespassers increased with decoration levels (table
S4). Similarly, in feeding observations, the percent
of time an individual spent eating and its success
rate in food-related aggressive interactions were
related to its nest decoration level (table S5).
Do receivers respond to the signal and do they
respond in a way that benefits signalers (hypothesis
ii)? In pairs without experimental nest augmen-
tation, territory intrusion rates were lower for highly
decorated nests (table S4), consistent with the
hypothesis that trespassers avoid decorated nests.
However, intruder pressure actually increased when
we experimentally augmented the decorations
(table S6). This suggests that nonterritorial animals
or “floaters”may be stimulated to attack pairs that
suddenly advertise a high-quality territory.
Higher levels of nest decoration predicted
higher subsequent breeding success and survival
(table S7 and Fig. 4), suggesting that birds may
benefit from lower rates of risky fights associated
with higher-quality advertisement.
Is the signal costly, and is the cost higher for
lower quality individuals (hypothesis iii)? Preda-
tion rates of experimental eggs were higher for
deco rated dummy nests (81.3% predated, n=
16) than for nondecorated dummy nests (31.3%,
n=16,c
2
=8.54,P= 0.003), indicating that there
is a cost to advertising quality through nest
decoration. Similarly, trespassing rates increased
for all individuals in the augmentation experi-
ment, but only those in prime age managed to
increase their attack rate in response (table S6 and
Fig. 5). Very young or senescent individuals ap-
peared incapable to counteract the cost of sig-
naling (Fig. 5); thus, it appears this cost is higher
for lower-quality individuals.
Is there deceit (hypothesis iv)? In the equal
availability experiment, lower-quality individuals
almost systematically refrained from signaling
(Fig. 2). One week after the augmentation ex-
periment, 85% of the individuals in prime age
had retained the plastic we placed in their nest,
whereas 87% in the other age classes had
removed it. Age and its quadratic effect were
the only predictors of the likelihood that a pair
would retain or expel the experimental plastic
(table S8). Deceit, if present, seems rare.
Our results show that nest decoration is a
graduated signal that conveys complex informa-
tion on territory quality, individual viability, and
dominance in social interactions. We found no
support for alternative hypotheses (tables S1 and
S9). Reliability of the signal was reinforced be-
cause lower-quality individuals refrained from
dishonest signaling, even though such cheating
would have been possible. Theoretical studies
predict signaling abstention by lower-quality in-
dividuals in the presence of agonistic retaliatory
costs imposed by receivers (18). Such social costs
were evident in this study: Advertisement of high-
quality resources was met with aggressive chal-
lenges. The costs of such challenges could be
disproportionately high for cheating individuals,
because advertising high-quality resources may
attract strong contestants (19) that are able to
displace dishonest signalers through aggres-
sive, and physically damaging, conflicts (16)
(fig. S1). Consistent with such disproportional
costs, only individuals in prime age were able
to respond to the increase in challenges stimu-
lated by experimental nest augmentation. Lower-
quality individuals were not able to match their
defense rates to the cheating decoration levels
that we imposed. Social punishment has been
recurrently shown as a major theoretical deter-
minant of communication honesty, especially for
signals involving minimum production costs
[e.g., (20,21), review in (2)], such as many
extended phenotype signals. This may explain
why social punishment and cheating absten-
tion seem common in such communication sys-
tems [e.g., (7,22–26)].
Extended phenotype signals can be reliable
indicators of both individual and territory qual-
ity. Previous analyses have shown that levels of
external signals may increase from juvenile to
adult age classes [e.g., (23,27)] or be individ-
ually repeatable through time [e.g., (28,29)].
Our findings confirm such earlier studies and
expand them by showing a clear senescent de-
cline in signaling strength. This reinforces the idea
of individuals signaling their physical prowess,
which increases and then declines with age. Fur-
ther, our findings indicate that external signals can
be used in the context of resource defense and do
not solely signal reproductive quality in a sexual
selection context, although reinforcement of pair
bonding may have contributed to our results (10).
In line with these ideas, many characteristics typ-
ical of the objects used to create extended phe-
notype signaling structures make them optimal as
signals of territorial defense: high visibility, ef-
fectiveness in the absence of the signaler, en-
coding of social dominance and motivation, ease
of deconstruction, and reliability enforced by
punishment of cheaters [reviews in (1–4)].
References and Notes
1. J. W. Bradbury, S. L. Vehrencamp, Principles of Animal
Communication (Sinauer, Sunderland, MA, 1998).
2. J. Maynard Smith, D. Harper, Animal Signals
(Oxford Univ. Press, Oxford, 2003).
3. W. A. Searcy, S. Nowicki, The Evolution of Animal
Communication (Princeton Univ. Press, Princeton, NJ, 2005).
4. F. C. Schaedelin, M. Taborsky, Biol. Rev. Camb. Philos. Soc.
84, 293 (2009).
5. J. R. Madden, Proc. Biol. Sci. 268, 833 (2001).
6. J. R. Madden, Anim. Cogn. 11, 1 (2008).
7. J. R. Madden, Proc. Biol. Sci. 269, 1347 (2002).
8. S. W. Coleman, G. L. Patricelli, G. Borgia, Nature 428,
742 (2004).
9. P. H. Wimberger, Auk 101, 615 (1984).
Fig. 4. Nest decoration by
black kites increased with
the viability of the signaler,
as estimated by breeding
performance (number of fledg-
lings produced, black circles,
n= 127) and survival until
spring (white squares, n= 76).
Data are mean T1SE.
Fig. 5. Experimental augmentation
of nest decoration. Mean (T1 SE)
attack rates of black kites on tres-
passers were higher in prime age
(7 to 12 years old; black circles, n=25)
individuals and lower in young or
old (≤6or≥13 years old; white
squares, n=40)animals.
www.sciencemag.org SCIENCE VOL 331 21 JANUARY 2011 329
REPORTS
on January 20, 2011www.sciencemag.orgDownloaded from

10. J. Moreno, M. Soler, A. P. Møller, M. Linden, Anim. Behav.
47, 1297 (1994).
11. M. Hansell, Bird Nests and Construction Behaviour
(Cambridge Univ. Press, Cambridge, 2000).
12. M. Hansell, Animal Architecture (Oxford Univ. Press,
Oxford, 2005).
13. D.J.Levey,R.S.Duncan,C.F.Levins,Nature 431, 39 (2004).
14. J. G. Schuetz, Behav. Ecol. 16, 133 (2005).
15. F. Sergio et al., Oecologia 160, 507 (2009).
16. F. Sergio, J. Blas, F. Hiraldo, J. Anim. Ecol. 78, 109 (2009).
17. Materials and methods are available as supporting
material on Science Online.
18. E. S. Adams, M. Mesterton-Gibbons, J. Theor. Biol. 175,
405 (1995).
19. A. Berglund, A. Bisazza, A. Pilastro, Biol. J. Linn. Soc. 58,
385 (1996).
20. R. A. Johnstone, K. Norris, Behav. Ecol. Sociobiol. 32,
127 (1993).
21. M. Enquist, S. Ghirlanda, P. L. Hurd, Anim. Behav. 56,
749 (1998).
22. J. Diamond, Proc. Natl. Acad. Sci. U.S.A. 83, 3042 (1986).
23. G. Borgia, Am. Nat. 141, 729 (1993).
24. S. Östlund-Nilsson, M. Holmlund, Behav. Ecol. Sociobiol.
53, 214 (2003).
25. J. M. Wojcieszek, J. A. Nicholls, A. W. Goldizen, Behav.
Ecol. 18, 689 (2007).
26. N. R. Doerr, Anim. Behav. 79, 747 (2010).
27. E. C. Collias, N. E. Collias, Auk 81, 42 (1964).
28. B. J. Rushbrook, N. J. Dingemanse, I. Barber, Anim. Behav.
75, 547 (2008).
29. P. T. Walsh, M. Hansell, W. D. Borello, S. D. Healy,
Biol. Lett. 6, 149 (2010).
30. We thank F. J. Chicano, J. Giralt, and F. G. Vilches for
help in the field. S. N. Vignieri and three anonymous
reviewers greatly improved a previous draft of the
manuscript. The study was funded by the research
projects CGL2008-01781, JA-58, RNM 1790, and RNM
03822 granted by the Junta de Andalucía, Ministerio de
Ciencia e Innovación, and Natural Research Limited.
Supporting Online Material
www.sciencemag.org/cgi/content/full/331/6015/327/DC1
Materials and Methods
Figs. S1 to S3
Tables S1 to S9
22 October 2010; accepted 14 December 2010
10.1126/science.1199422
Discovery of a Viral NLR Homolog that
Inhibits the Inflammasome
Sean M. Gregory,
1,2
BeckleyK.Davis,
1
John A. West,
1,2
Debra J. Taxman,
1,2
Shu-ichi Matsuzawa,
3
John C. Reed,
3
Jenny P. Y. Ting,
1,2
Blossom Damania
1,2,4
*
The NLR (nucleotide binding and oligomerization, leucine-rich repeat) family of proteins senses
microbial infections and activates the inflammasome, a multiprotein complex that promotes microbial
clearance. Kaposi’ssarcoma–associated herpesvirus (KSHV) is linked to several human malignancies.
We found that KSHV Orf63 is a viral homolog of human NLRP1. Orf63 blocked NLRP1-dependent
innate immune responses, including caspase-1 activation and processing of interleukins IL-1band
IL-18. KSHV Orf63 interacted with NLRP1, NLRP3, and NOD2. Inhibition of Orf63 expression resulted in
increased expression of IL-1bduring the KSHV life cycle. Furthermore, inhibition of NLRP1 was
necessary for efficient reactivation and generation of progeny virus.Theviralhomologsubvertsthe
function of cellular NLRs, which suggests that modulation of NLR-mediated innate immunity is
important for the lifelong persistence of herpesviruses.
Kaposi’s sarcoma–associated herpesvirus
(KSHV/HHV8) is the etiological agent of
several human cancers including Kaposi’s
sarcoma (KS), primary effusion lymphoma (PEL),
and multicentric Castleman’s disease (MCD)
(1,2). The ability of KSHV to evade host innate
immunity is essential for productive infection,
latency, and lifelong persistence (3).
Several families of pattern recognition recep-
tors (PRRs) have been described, including TLRs
(Toll-like receptors), the NLR (nucleotide bind-
ing and oligomerization, leucine-rich repeat)
family of proteins, RLRs (retinoic acid–inducible
gene I–like receptors), and CLRs (C-type lectin
receptors) (4,5). Upon recognition of pathogen-
associated molecular patterns (PAMPs), PRRs
signal immune cell activation. TLRs play an im-
portant role in the life cycle of KSHV (6,7), but
whether NLRs also do is unknown. More than 20
NLR family members have been identified in hu-
mans, and polymorphisms in several NLRs are
linked to various autoinflammatory diseases (8–12).
Activation of a subset of NLRs by PAMPs causes
the formation of large multimeric complexes termed
inflammasomes, which are composed of oligomers
of a specific NLR, procaspase-1, and an adaptor
protein called ASC (apoptotic-associated speck-
like) (13). Inflammasome formation results in the
proteolytic processing of the proinflammatory
cytokines IL-1band IL-18 by active caspase-1.
Excessive IL-1band IL-18 production in re-
sponse to pathogen infection is associated with
pyroptosis, an inflammatory process involving
caspase-1–mediated cell death.
KSHV Orf63 is an uncharacterized tegument
protein. Basic Local Alignment Search Tool for
proteins (BLASTP) of NLRP1 and KSHV Orf63
revealed that these proteins are homologous (E
value = 0.0002) and that Orf63 shows significant
similarity to the leucine-rich repeat (LRR) domain
of NLRP1 (fig. S1A). A ClustalW2 alignment of
the two proteins also showed homology of the
nucleotide-binding domain (NBD) of NLRP1
(fig. S1A) and full-length NLRP1 (fig. S1B) with
KSHV Orf63. However, KSHV Orf63 does not
contain the effector caspase activation and re-
cruitment domain (CARD) or pyrin domain (PYD)
of NLRP1, which are required for its activation;
this suggests that Orf63 may function as an in-
hibitor of NLRP1.
To investigate whether Orf63 inhibits NLRP1
activity, we primed THP-1 cells, a human mono-
cytic leukemia cell line, stably expressing Orf63
and control cells with lipopolysaccharide (LPS)
to up-regulate IL-1btranscription, followed by
stimulation of NLRP1 with the bacteria cell wall
constituent MDP (muramyldipeptide). We found
that Orf63 expression significantly inhibited MDP-
induced IL-1band IL-18 production relative to con-
trol cells (Fig. 1, A to D, and fig. S2A). Moreover,
no changes in the inflammasome-independent
cytokine TNF-a(tumor necrosis factor–a)were
observed, confirming the specificity of Orf63 in-
hibition of the NLRP1 inflammasome (fig. S2B).
Inflammasome activation leads to produc-
tion of proinflammatory cytokines and eventual
cell death. In cells expressing Orf63, NLRP1-
dependent cell death [as measured by lactate
dehydrogenase (LDH) activity] was significantly
inhibited relative to vector alone (Fig. 1E), demon-
strating that Orf63 protects cells from NLRP1-
dependent cell death.
To further examine the role of Orf63 in block-
ing NLRP1, we transfected 293Tcells with Orf63
and the NLRP1 inflammasome components ASC,
procaspase-1, and pro–IL-1b(14). Reconstitution
of the NLRP1 inflammasome resulted in an in-
crease in IL-1bsecretion, which was inhibited by
expression of Orf63 in a dose-dependent fashion
(Fig. 1, F and G). In contrast, another KSHV viral
protein, RTA (replication and transcription activa-
tor), was unable to inhibit IL-1bsecretion (fig.
S2C). We also investigated the ability of Orf63 to
inhibit caspase-1 enzymatic activity. Transfection
of procaspase-1 and NLRP1 into 293T cells re-
sulted in increased caspase-1 specific activity that
was inhibited by coexpression of Orf63 (Fig. 1H).
Furthermore, NLRP1-induced proteolytic process-
ing of procaspase-1 to activated caspase-1 was
also inhibited (Fig. 1I).
Next, we investigated whether Orf63 could
interact with NLRP1. Orf63 coimmunoprecipitated
with NLRP1, and vice versa, in 293T cells co-
transfected with plasmids to express these proteins
(Fig. 2, A and B). Endogenous NLRP1 was also
showntointeractwithOrf63whenOrf63was
expressed in THP-1 cells (fig. S2D). Thus, Orf63 is
capable of interacting with NLRP1 and/or in a com-
plex with components of the NLRP1 inflamma-
some. To investigate the latter scenario, we tested
1
Lineberger Comprehensive Cancer Center, University of
North Carolina, Chapel Hill, NC 27599, USA.
2
Department
of Microbiology and Immunology, University of North Car-
olina, Chapel Hill, NC 27599, USA.
3
Sanford-Burnham Med-
ical Research Institute, La Jolla, CA 92037, USA.
4
UNC Center
for AIDS Research, University of North Carolina, Chapel Hill,
NC 27599, USA.
*To whom correspondence should be addressed. E-mail:
damania@med.unc.edu
21 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org330
REPORTS
on January 20, 2011www.sciencemag.orgDownloaded from

www.sciencemag.org/cgi/content/full/331/6015/327/DC1
Supporting Online Material for
Raptor Nest Decorations Are a Reliable Threat Against Conspecifics
F. Sergio,* J. Blas, G. Blanco, A. Tanferna, L. López, J. A. Lemus, F. Hiraldo
*To whom correspondence should be addressed. E-mail: fsergio@ebd.csic.es
Published 21 January 2010, Science 331, 327 (2011)
DOI: 10.1126/science.1199422
This PDF file includes:
Materials and Methods
Figs. S1 to S3
Tables S1 to S9
References

1
Supporting Online Material
Material and Methods
Marking and monitoring
We collected data in Doñana National Park (south-western Spain), between 2005-2010, in a 450
km2 plot of seasonally flooded marshland holding ~ 500 breeding pairs (S1). Nestlings have been
ringed since the 1980s with a white plastic ring with a black, three-character alphanumeric code
which can be read by spotting-scopes without disturbing the birds (S2-S4). This allowed us to
work with several individuals of known age and covering the whole lifespan of the population
(1-25 years old, Fig. S2) (S2, S3). Marked territory holders were searched intensively by visiting
territories every 4-5 days throughout the breeding cycle. A marked individual was assigned to a
specific territory when it was observed there in more than three separate visits (S3-S5). The nests
of marked individuals were visited when nestlings were 40-45 days old to record the number of
fledged young and to ring the chicks (S5). Adults of known age were trapped through cannon-
nets to measure their body size (PC1 of tarsus, wing and tail length) and condition (residuals of
mass on PC1: S3-S5); 100 adults were equipped with GPS-satellite tags (PTT-100 manufactured
by Microwave Telemetry). All trapped adults and ringed nestlings were sexed by molecular
methods (S2, S5). Further details of population monitoring can be found in S1-S5.
Whenever possible, we climbed the nest to estimate the number, type and color of the man-
made objects present in the nest and the percentage of the overall nest surface occupied by such
materials (hereafter “nest decoration”, visually estimated to the nearest 5 % by reference to a
graduated score-sheet). Twenty-five nests were visited weekly to estimate the temporal evolution
of decoration-levels through all the stages of the breeding cycle (Fig. S3). Because nest-
decorations peaked in the last three weeks before laying and declined thereafter (Fig. S3), all
subsequent experiments focus on the 20 days before laying.
Parasites of 87 nestlings from 62 broods were surveyed through polymerase chain reaction
(PCR) amplification of blood-samples. The survey method covered 29 types of pathogens (5
protozoan, 7 bacteria, 2 fungi, and 15 viruses) for which PCR-based determination protocols
have been previously described (S6).
To test whether kites were selective in their choice of man-made objects, we estimated the
availability of artificial items of the size-range used by kites by annotating the material and color
of those observed while walking along the major roads present within 2 km of the nests of the
study area. This is where most such objects accumulate and were kites were regularly observed
collecting them.
Behavioral observations
To test the association between nest decorations, territory intrusions and fighting ability, we
conducted standardized 4-hours observations (hereafter “non-experimental observations”) at 20
focal territories occupied by marked individuals. During each session, we recorded: (i) all
trespassing individuals flying within 200 m of the nest; (ii) the number of trespassers physically
attacked by each focal individual; (iii) the time lost in such interactions; (iv) the percentage of
attacks which were successful (i.e. intruder flees from the territory upon attack). All observations
were conducted 10-20 days before laying by the focal pair, the peak period for both intrusions
and nest-decoration additions.
To further verify the association between nest decorations and fighting capability, we
conducted standardized observations of individuals feeding and fighting on baited carrion
(hereafter “feeding-observations”). In each observation-session, we placed 30-35 sheep heads of
equal dimensions in a 10 x10 m quadrat at a fixed feeding site. Once a marked individual was
feeding on a bait-item, an observer placed in a hide 15 m away recorded for 10 minutes: (i) the

2
number of times it was attacked by other individuals trying to dislodge it from the food and their
success; (ii) if dislodged, the number of times it attacked other individuals to dislodge them and
their success; (iii) the cumulative number of seconds spend eating (i.e. the time it managed to
monopolize food in a competitive environment). Nest decoration data were available for 21 of
these marked individuals. We consider this a strong test of the association between decorations
and fighting ability because contenders cannot know the decoration-level of the focal animal
(nests were various km away from the baited carrion) and the focal animal cannot adjust its
aggressiveness level on the basis of its partner-behavior or cheat on its fighting capabilities
through decoration-exaggeration.
Experiment 1: equal availability
To ensure that variations among pairs in decoration-level were not caused by differential
availability, in 2009 we placed three pieces of plastic bags (the material prevalent in nests), each
of 25 cm x 25 cm (i.e. sufficient to cover the nest cup), about 20 cm from each other, 10 m from
the nest and main perch of 49 focal pairs of known age and in full view from them. One piece
was transparent, one green and one white and their collection rate was recorded daily for 10
days. The colors were chosen on the basis of the following hypotheses: (i) over-selection of
transparent items over white ones implies interest in some structural property of the material
rather than its color; (ii) over-selection of green items implies interest in nest camouflage (at
least to human vision) rather than exuberant signalling. Under both the above hypotheses, kites
would rather use man-made materials but pass unnoticed, if given the chance. (iii) Over-selection
of white items implies interest in a shiny, “locally unnatural” color that makes the nest very
obvious from far away, consistent with a signalling function.
Experiment 2: augmentation
To experimentally explore the effect of decorations on intruders, in 2010 we manipulated the
decorations and observed the subsequent behavior of 20 focal pairs with marked individuals.
During each session (hereafter “experimental observation”): (i) an observer in a car placed > 300
m from the focal nest recorded the same behaviors of the non-experimental observations for 2
hours; (ii) a second team in a separate car came from a distant non-visible location, climbed the
nest, placed a 25 x 25 cm piece of white plastic to cover the whole nest-cup and then went away;
(iii) after the birds resumed their normal activities (usually in 10-15 minutes), the observer
resumed data collection for a further two hours. The same procedure was conducted at 20 control
pairs but no plastic was added. The experiment was conducted 15-25 days before laying and all
nests had no decorations when sampled. Pairs were re-checked 7 days later to examine whether
they had subsequently removed or accepted the added plastic. A similar experiment where we
removed the decorations failed because the decorations are sometimes added too late in pre-
incubation and too many such pairs had laid eggs by when we tried to sample them.
Experiment 3: predation costs
To test a potential predation cost of using eye-catching decorations, in 2010 we (i) built 32
dummy nests of the size of a kite nest; (ii) positioned them on trees far away (> 500 m) from any
active kite nest to avoid confounding factors; (iii) added a 25 cm x 25 cm white plastic at a
randomly chosen half of the nests; (iv) placed 2 white chicken eggs in each nest; and (e)
recorded their predation rate every other day for 10 days.
Experiment 4: hatching facilitation
To test whether man-made materials could improve the thermal conditions or sanitation of the
nest resulting in higher incubation efficiency, we removed such materials every 4 days from each

3
of 18 pairs throughout incubation and compared their hatching success with the one of 18 control
pairs climbed every 4 days but without removing anything.
Viability tests
The viability of an individual was defined as its subsequent capability of successful reproduction
and of survival to the next spring. We estimated reproductive success as the number of nestlings
raised to fledging age. Survival of breeding adults to the next year was estimated for a sub-
sample of 76 individuals which were (i) equipped with satellite-tags or (ii) breeding in territories
where between-years breeding-dispersal movements of up to 2 km would have been within areas
of high visibility and easy ring-readings. Given that ring-reading detectability under such
conditions is virtually 1 (S5), we are confident that birds not detected in a year can be reliably
classified as dead (as validated by satellite-tagging data), preventing the need of capture-
recapture analyses incorporating detectability-corrections (details in S5). The relationship
between viability and advertisement strength was tested by fitting nest-decoration as the
explanatory variable and subsequent reproductive success and survival as dependent variables in
generalized linear mixed models (S7, S8).
Hypothesis testing and statistical analyses
A detailed summary of all hypotheses and predictions is given in table S1. Below, we focus on
the main signalling hypothesis. We frame our results as four main questions aimed at
investigating whether a signal can evolve and be honest on average, following Searcy & Nowicki
(S9): (i) is the signal reliable? This question implies a significant association between nest-
decorations and traits which may be of interest to the receiver, such as individual quality,
territory quality or fighting ability. (ii) Do receivers respond to the signal and do they respond in
a way that benefits the signaller? (iii) Is the signal costly and is the cost disproportionately higher
for lower quality individuals? (iv) Is there deceit?
Because nest-decorations peaked in the last three weeks before laying and declined thereafter
(Fig. S3), all statistical analyses refer to data collected in the 20 days before laying. We used χ2
tests to compare the frequency of used and available man-made objects. Generalized linear
mixed models (GLMM) were used to test the effect of multiple explanatory variables on a
dependent variable, while controlling for pseudoreplication by fitting individual identity as a
random factor (S7, S8). In particular, associations between decorations and individual and
territory quality were tested by fitting nest decoration as the dependent variable and sex, territory
quality and age as explanatory variables. Associations between decorations and fighting ability
or intruder pressure were tested for data from non-experimental observations and feeding-
observations by fitting nest decoration, sex and their interaction as explanatory variables and
intrusion rates, time spent feeding, attack rates, time lost in interactions and attack success as
dependent variables. In the augmentation experiment, associations between decorations and
fighting ability or intruder pressure were examined by fitting sex, treatment (augmented vs
control individual), age and their interactions as explanatory variables and intrusion rates, attack
rates and time lost in interactions as dependent variables. Here, the sample unit was the
percentage change in the target variable between the two hours before and after the decoration-
augmentation, which is compared between experimental and control nests.
In all GLMMs, individuals and/or years were fitted as categorical random factors to control
for annual variations and pseudoreplication, where necessary. When no pseudoreplication was in
place, we used generalized linear models (GLM, S10). Some variables were not available for all
individuals (e.g. body size and condition). In such cases, we built the model twice, with and
without the variable of interest in order to reach a compromise between information-
completeness and sample size. Models were built through a backward stepwise procedure: all
explanatory variables were fitted to a (maximal) model, extracted one at a time from such

4
maximal model, and the associated change in model deviance was assessed by a likelihood-ratio
test; at each step, we calculated the AICc of each model and considered as the final competitive
model the one with the lowest AICc containing only significant terms (S7).
Throughout, decoration-levels were estimated as percentage nest cover by man-made objects
because other variables (e.g. number of objects) were strongly correlated with it (all r > 0.85)
and systematically provided a lower fit (larger significance values and AICc) when substituted to
nest cover in all models. Individual quality was expressed as age (i.e. demonstrated capability of
survival), age^2 (to test for a quadratic or parabolic effect of age, consistent with senescence-
effects), body size and condition. Because in this population reproduction, survival and breeding
value peak between 7 and 12 years old and decline thereafter (S2), we tested both the linear and
the quadratic (i.e. parabolic, S11) effect of age in all models. Further, individuals were
sometimes classified as in “prime age” (= 7-12 years old) or not (= 1-6 or 13-25 years old) in
order to simplify some graphical representations and analyses. Territory quality was expressed as
the first axis of a PCA of (i) the percentage availability of wetlands (the optimal foraging habitat)
within 2 km of the nest and (ii) the percentage of years that a territory was occupied (S1, S4, S5,
S12).
All statistics were implemented in R.2.9.2, variables were log or arcsin-square root
transformed as necessary to ensure normality, checking of model assumptions was implemented
by the PLOT function of R (S7), significance levels of multiple tests employing the same dataset
were adjusted through the sequential Bonferroni correction (S13) and all means are given ± 1 SE.

5
1a
1b
Fig. S1. In Doñana National Park, floating black kites are prevented from breeding through social
dominance, but may attempt territory take-overs, which sometimes succeed (S3, S4). Territorial
confrontations can be extremely violent, with birds talon-locking in the air (1a), falling dangerously on
the ground while battling and loosing up to several flight-feathers (S4). Territory trespassing can be
frequent, especially in pre-incubation, and also aimed at stealing food originally caught by the owners
(1b; Photos F. Sergio).

6
Fig. S2. In Doñana, nestlings have been ringed since the 1980s with a white plastic ring with a black,
three-character alphanumeric code, which can be read by telescope (Photo F. Sergio). This allowed us to
work with several individuals of known age and covering the whole lifespan of the population (1-25 years
old).

7
Fig. S3. Variation in mean (± SE) nest-decoration levels at 25 nests sampled longitudinally
during all stages of the breeding cycle. The latter was de-composed in the following progressive
stages:
1. first half of the pre-laying period (from pair-formation up to 20 days before laying eggs);
2. last 20 days of the pre-laying period;
3. incubation;
4. first half of the nestling period (with nestlings up to 20 days old);
5. second half of the nestling period (with nestlings older than 20 days).
Nest-decorations were virtually zero in the first half of the pre-laying period, appeared suddenly
during the last 20 days before laying, declined slightly and non-significantly in incubation, and
then markedly during the first and second half of the nestling period (ANOVA: F4,120 = 41.0, p <
0.0001). In post-hoc tests, nest-decoration was minimal and homogeneous in the first half of pre-
laying and the second half of the nestling period, maximal and homogeneous in the second half
of pre-laying and in incubation, and intermediate in the first half of the nestling period (Duncan’s
post hoc test, p < 0.05).

8
Table S1. Hypotheses explaining the potential function of the placement of man-made objects by black kites in their nests. By the “Signalling or threat
hypothesis”, man-made objects are extended phenotype threats to conspecifics revealing the territory quality, viability and fighting ability of the
signaller. By the “Nest-structure hypothesis”, man-made objects are a nest-material, just like branches, that improves the structural strength of the nest.
By the “thermal conditions hypothesis”, man-made items may improve the thermal conditions of the nest, resulting in improved incubation-efficiency.
By the “Nest sanitation hypothesis”, artificial materials are added to isolate the offspring from prey items in decay and may repel ecto-parasites, as
suggested for the addition of green plants to the nests of some passerine and raptorial birds (e.g. S14-S16).
Hypothesis and rationale Predictions Tests of the predictions Supported?
1. Signalling (or threat)
Man-made objects are extended
phenotype signals of threat against
trespassers
A. Decorations peak when intrusion
rates and territory take-overs are most
frequent
A1. Decorations are added in pre-incubation and decline thereafter Yes
B. Kites are selective with respect to
the decorations employed
B1. Compared to local availability, kites select eye-catching objects visible from
very far.
B2. In the “equal availability experiment” 1, they select white plastic.
Yes
Yes
C. Nest decorations reflect territory
quality, individual quality and fighting
ability
C1. Decoration-levels are related to territory quality, age, body size and
condition.
C2. In the “equal availability experiment” 1, individuals that collect decorations
are of higher quality and hold better territories than individuals that refrain from
doing so.
C3. In “observations of non-experimental pairs” 2, decorations reveal the attack
rate and attack success of residents repelling trespassers.
C4. In “feeding-observations” 3, nest decorations predict the fighting success
and the capability to monopolize food in a competitive environment.
Yes
Yes
Yes
Yes
D. Receivers respond to the nest-
decoration signal in a way that benefits
the signaller
D1. In “non-experimental observations” 2, intrusion rates vary with decoration-
level.
D2. In the “augmentation experiment” 4, intrusion rates vary with decoration-
level.
D3. Breeding success and survival are positively related to nest-decoration level.
Yes
Yes
Yes
E. Decorating the nest bears costs E1. In the “dummy-nest experiment” 5, decorated dummy nests have higher
predation rates than non-decorated ones.
E2. In the “augmentation experiment” 4, addition of decorations causes an
increase in intrusion rates (i.e. higher risks of take-over).
Yes
Yes
F. Costs fall disproportionately higher
on lower quality individuals
F1. In the “augmentation experiment” 4, only higher quality individuals manage
to increase their attack-rate of intruders and match it to their new (manipulated)
Yes

9
decoration-level.
2. Nest structure
Man-made objects are a nest-
material, just like branches, that
improves the structural strength of
the nest
A. Kites are interested in materials, not
colors
A1. Compared to availability, they are not selective according to colors.
A2. In the “equal availability experiment” 1, they collect any of the three offered
items, or avoid white.
No
No
B. Artificial items are added
throughout the breeding cycle and
peak in the nestling period, when
structural wear by the weight and
movements of the nestlings is highest
B1. Nest coverage by man-made items is highest in the nestling period. No
C. Presence of artificial items predicts
the probability that a nest will fall.
C1. Nest coverage by man-made items affects the probability that a nest will fall
during the breeding season.
C2. Nest coverage by man-made items affects the probability that a nest will fall
during the following winter.
No
No
3. Thermal conditions
Man-made items improve the
thermal conditions of the nest,
resulting in improved incubation-
efficiency
A. Artificial items are mostly added in
pre-laying and incubation and peak
during incubation
A1. Nest coverage by man-made items is highest in incubation. No
B. Addition of man-made items
improves incubation efficiency
B1. Nest coverage by man-made items affects hatching success.
B2. In the “hatching facilitation experiment” 6, control pairs have higher
hatching success than pairs that had artificial materials systematically removed.
No
4. Nest sanitation
Artificial materials are added to
isolate the offspring from prey
items in decay and may deter
disease-agents
A. Artificial items are mostly added in
the nestling period, when prey items
accumulate in the nest
A1. Nest coverage by man-made items is highest in the nestling period.
A2. Nest coverage by man-made items increases between the first and second
half of the nestling period.
No
No
B. Man-made materials lower the
nestlings’ infestation rates by
pathogens
B1. Nestlings’ infestations by pathogens are negatively related to the abundance
of man-made materials in the nest
No
1 “Equal availability experiment”: 49 pairs were offered ready availability near the nest of 3 pieces of plastic of white, green and transparent color (details in
Methods).
2 “Observations of non-experimental pairs”: behavioral observations conducted for 4 hours at pairs which were not the subject of any experimental manipulation
(details in Methods).

10
3 “Feeding observations”: 10-min observations of marked individuals feeding and fighting on carrion-baits (details in Methods).
4 “Augmentation experiment”: behavioral observations conducted at 20 experimental (and 20 control) pairs, before and after adding man-made objects to their nest
(details in methods).
5 “Dummy nest predation experiment”: predation-rates were observed at 32 dummy kite nests containing white chicken eggs. Half of the nests contained a 25x25cm
piece of white plastic bag (details in Methods).
6 “Hatching facilitation experiment”: 15 pairs had their artificial materials removed every 5 days throughout incubation and were compared to control pairs (details in
Methods).

11
Table S2. GLMM models testing the relationship between nest-decoration (as measured in the last
20 days before laying) and territory quality, age and body condition.
Results summary: nest-decoration increased with territory quality, body condition and exhibited a
quadratic (i.e. parabolic) relationship with age.
Explanatory variables 1
(sample size):
B ± SE t p
AICc R2
a. Effect on nest-
decoration 2 (n = 127) 3
112.7 15.8
Age 0.06 ± 0.03 2.08 0.043
Age^2 -0.03 ± 0.01 -2.49 0.016
Territory quality 0.34 ± 0.12 2.85 0.006
Constant -0.12 ± 0.16
b. Effect on nest-
decoration 2 (n = 36) 4
32.1 14.5
Body condition 0.15 ± 0.06 2.37 0.024
Constant 0.50 ± 0.61
1 Variables which entered the minimal adequate model after backward stepwise selection (see Methods).
2 Proportional coverage of the nest surface by man-made objects (arcsin-square root transformed).
3 GLMM with normal errors and an identity link function. Explanatory variables fitted to the model: sex,
age, age^2, territory quality, year (random term) and individual identity (random term).
4 GLM with normal errors and an identity link function. Explanatory variables fitted to the model: sex, body
size and body condition.

12
Table S3. GLMM model discriminating between individuals that collected and placed in their nest
artificial objects offered to them in an “equal availability experiment” (value = 2) and individuals
that ignored the experimental items (value = 1). The latter were placed about 10 m and in clear
view of the nest and main perch of each individual and their usage checked daily for 10 days.
Results summary: Individuals refraining from signalling held lower quality territories and were
either much younger or much older than individuals which placed the offered experimental items
in their nest.
Explanatory variables 1
(sample size):
B ± SE t p
AICc R2
Age 0.90 ± 0.36 2.47 0.014 48.3 40.2
Age^2 -0.04 ± 0.01 -2.63 0.009
Territory quality 3.85 ± 1.20 3.21 0.001
Constant -7.96 ± 2.59
1 Variables which entered the minimal adequate model after backward stepwise selection (see Methods).
GLMM with binomial errors and a logit link function. Explanatory variables fitted to the model: sex, age,
age^2, territory quality and year (random term).

13
Table S4. GLMM models testing the relationship between nest-decoration (as measured in the last
20 days before laying) and rates of territory intrusions by floaters, attack-rates of territory holders
on trespassers, the time lost by territory holders in such aggressive interactions and their success
rate in repelling intruders. All behaviors were assessed during 4-hour sessions of standardized
observation of 38 marked individuals of known age and sex.
Results summary: (Model a) receivers responded to the signal and avoided nests with high
decoration-levels; (Models b,c): signallers benefited from the signal by lower rates of attack/hour
on trespassers and less time lost in aggressive interactions; (Model d) the signal honestly portrayed
the fighting success of the territory-owner.
Explanatory variables 1 B ± SE t p
AICc R2
a. Effect on: Trespassing-rate 2 57.4 39.7
Nest-decoration 3 -1.01 ± 0.25 -4.10 < 0.0001
Constant 1.55 ± 1.16
b. Effect on: Attack-rate on
trespassers 2
53.7 27.9
sex -0.33 ± 0.13 -2.60 0.019
Nest-decoration 3 -0.42 ± 0.17 -2.49 0.023
Constant 0.70 ± 0.12
c. Effect on: Time lost in
aggressive interactions 2
28.24 48.2
sex -0.35 ± 0.10 -3.39 0.004
Nest-decoration 3 -0.58 ± 0.13 -4.44 0.003
Constant 0.79 ± 0.09
d. Effect on: Success-rate of
attacks 2
11.5 33.4
Nest-decoration 3 0.58 ± 0.21 2.75 0.019
Constant 0.50 ± 0.10
1 Variables which entered the minimal adequate model after backward stepwise selection (see Methods).
2 GLMM with normal errors and an identity link function. Explanatory variables fitted to the model: sex (1
= male, 2 = female), nest-decoration, sex*nest-decoration and nest (random term). Trespassing-rate =
number of individuals flying within 200m of the nest / hour (log-transformed in the models); Attack-rate
on trespassers = number of attacks on trespassers / hour (square-root transformed); Time lost in
aggressive interactions = proportion of total observation time lost in interactions (arcsin square-root
transformed); Success-rate of attacks = proportion of attacks by the owners which were successful (i.e.
trespassers immediately flee from the territory upon attack; arcsin square-root transformed).
3 Proportional coverage of the nest surface by man-made objects (arcsin square-root transformed).

14
Table S5. GLMM models testing the relationship between nest-decoration (as measured in the last
20 days before laying) and social dominance over feeding resources, expressed as (i) the capability
to monopolize food in an aggressively competitive environment and (ii) fighting success in
defending food from challengers or evicting other kites from their prey. All behaviors were
assessed during 10-min sessions of standardized observation of marked individuals of known age
and sex.
Results summary: Nest-decoration reliably reveals social dominance and fighting ability even in
the absence of potentially confounding factors like the behavior of the partner during contests, or
the nearby presence of the decoration itself.
Explanatory variables 1 (sample
size):
B ± SE t p
AICc R2
a. Effect on: Time spent
feeding 2 (n = 85)
98.4 18.1
Nest-decoration 3 0.49 ± 0.17 2.94 0.009
Constant 0.59 ± 0.10
b. Effect on: Success-rate in
aggressive interactions 2 (n = 85)
126.5 20.6
Nest-decoration 3 0.85 ± 0.24 3.56 0.003
Constant 0.27 ± 0.15
1 Variables which entered the minimal adequate model after backward stepwise selection (see Methods).
2 GLMM with normal errors and an identity link function. Explanatory variables fitted to the model: sex,
nest-decoration, sex*nest-decoration and group size (overall number of kites eating carrion during the
observation session). Time spent feeding = proportion of the observation session spent feeding by the
focal animal (arcsin square-root transformed); Success-rate in aggressive interactions = proportion of
times that a focal animal successfully defended food against a challenger or aggressively dislodged
another individual from food (arcsin square-root transformed). Decomposition of this variable in success
rates in defending food or stealing it gave the same conceptual results (i.e. individuals good at defending
were also good at stealing).
3 Proportional coverage of the nest surface by man-made objects (arcsin square-root transformed).

15
Table S6. GLMM models testing the effect of experimental augmentation of nest-decorations on
rates of territory intrusions by floaters, time lost by territory holders in aggressive interactions and
rates of attack on trespassers by territory holders. The sample unit was the proportional change in
the target variable between the two hours before and after the decoration-augmentation, which was
compared between experimental and control nests.
Results summary: (Model a) receivers responded to the signal and increased their trespassing-
rates when territory-holders were suddenly experimentally forced to advertise high quality
territories; (Model b): the increased intrusion-rates caused territory-holders to loose more time in
aggressive interactions (i.e. a receiver-dependent cost); (Model c) only birds in prime age managed
to raise their attack-rates on trespassers in response to the increased intrusion-pressure.
Explanatory variables 1 B ± SE t p
AICc R2
a. Effect on: Trespassing-rate (n
= 65) 2
32.7 12.9
Treatment 3 0.30 ± 0.10 2.84 0.008
Constant 0.65 ± 0.07
b. Effect on: Time lost in
aggressive interactions (n = 65) 2
47.0 13.4
Treatment 3 0.34 ± 0.10 3.23 0.003
Constant 0.65 ± 0.08
c. Effect on: Attack-rate on
trespassers (n = 65) 2
28.24 25.7
Age -0.12 ± 0.20 -0.63 0.535
Treatment 3 0.17 ± 0.19 0.91 0.369
Age * Treatment 0.63 ± 0.31 2.06 0.047
Constant 1.03 ± 0.14
1 Variables which entered the minimal adequate model after backward stepwise selection (see Methods).
2 GLMM with normal errors and an identity link function. Explanatory variables fitted to the model: sex,
age (1 = individuals ≤ 6 years or ≥ 13 years old; 2 = individuals 7-12 years old), treatment (1 = control
nest; 2 = nest with augmented decorations), age*treatment, and nest (random term). Trespassing-rate =
number of intruders flying within 200m of the nest / hour (log-transformed in the models); Time lost in
aggressive interactions = proportion of total observation time lost in interactions (arcsin square-root
transformed); Attack-rate on trespassers = number of attacks on trespassers / hour (square-root
transformed).
3 Experimental treatment: 1 = control nest; 2 = nest with augmented decorations.

16
Table S7. GLMM models testing the relationship between breeding success, survival and nest-
decoration.
Results summary: breeding success and survival were positively related to nest-decoration levels,
suggesting: (i) that high-quality advertisement could provide subsequent fitness benefits (e.g. by
curtailing aggressive and physically damaging territorial contests); and (ii) that the signal conveys
information on the future viability of the signaler.
Explanatory variables 1 (sample size): B ± SE t p
AICc R2
a. Effect on number of fledged young 2
(n = 127) 3
102.1 19.5
Nest-decoration 3 1.25 ± 0.27 4.56 < 0.0001
Constant -0.97 ± 0.22
b. Effect on survival (n = 76) 4 69.6 9.0
Nest-decoration 3 1.89 ± 0.81 2.33 0.019
Constant 0.58 ± 0.48
1 Variables which entered the minimal adequate model after backward stepwise selection (see Methods).
2 GLMM with poisson errors and a logarithmic link function. Explanatory variables fitted to the model:
nest-decoration and individual identity (random term).
3 Proportional coverage of the nest surface by man-made objects (arcsin-square root transformed).
4 GLMM with binomial errors and a logit link function. Explanatory variables fitted to the model: nest-
decoration and individual identity (random term).

17
Table S8. GLMM model discriminating between individuals that removed the nest-decorations
placed in their nest in an “augmentation experiment” (i.e. who refused to signal; value = 2) and
individuals that retained them (value = 1).
Results summary: Individuals in prime age retained the offered decorations while individuals
who were either very young or very old removed them, consistent with lower-quality birds actively
refraining from cheating.
Explanatory variables 1
(sample size):
B ± SE t p
AICc R2
Age 14.20 ± 6.53 2.17 0.029 15.33 64.6
Age^2 -0.69 ± 0.31 -2.16 0.031
Constant -69.85 ± 32.29
1 Variables which entered the minimal adequate model after backward stepwise selection (see Methods).
GLMM with binomial errors and a logit link function. Explanatory variables fitted to the model: sex, age,
age^2, territory quality and nest identity (random term).

18
Table S9. GLMM models testing the relationship between nest-decoration and structural nest-
strength, hatching success and incidence of pathogens in the nestlings.
Results summary: there was no support for the Nest structure, Thermal conditions or Nest
sanitation hypotheses (described in Table S1).
Explanatory variables 1 (sample size): B ± SE t p
a. Effect on: probability of a nest falling
during breeding (n = 128) 2
Nest-decoration 3 1.60 ± 3.87 0.41 0.68
b. Effect on: probability of a nest falling
during winter (n = 128) 4
Nest-decoration 3 0.75 ± 2.40 0.31 0.76
c. Effect on: hatching success in non-
experimental nests (n = 129)
Nest-decoration 3 0.19 ± 0.15 1.35 0.17
d. Effect on hatching success in hatching-
facilitation experiment (n = 30) 5
Experimental treatment 5 0.10 ± 0.21 0.47 0.64
e. Effect on: number of pathogens found
in blood samples of nestlings (n = 87) 6
Nest-decoration 3 0.23 ± 0.16 1.46 0.15
1 Variables which entered the minimal adequate model after backward stepwise selection (see Methods).
2 Effect on the probability that a nest will fall during the breeding season. GLMM with binomial errors and
a logit link function. Explanatory variables fitted to the model: % nest-decoration, year and nest (random
term).
3 Proportional coverage of the nest surface by man-made objects (arcsin-square root transformed).
4 Effect on the probability that a nest will fall during the following winter. GLMM with binomial errors and
a logit link function. Explanatory variables fitted to the model: % nest-decoration, year and nest (random
term).
5 GLMM with normal errors and an identity link function. Experimental treatment: value 1 = nests that had
their artificial objects removed every 4 days throughout incubation (n = 18, 5 sampled in 2005, 5 in 2006
and 8 in 2007); value 2 = control nests that were climbed every 4 days but no material was removed (n =
18, 5 sampled in 2005, 5 in 2006 and 8 in 2007). Explanatory variables fitted to the model: experimental
treatment and year (random term). Hatching success (mean % of eggs hatched) was 70.1 % and 71.8 %
for experimental and control pairs respectively, which are similar to the overall population-mean.
6 GLMM with poisson errors and a logarithmic link function. Explanatory variables fitted to the model:
nest-decoration, brood order (1 = 1st born nestling in the brood; 2 = 2nd born nestling; 3 = 3rd born
nestling), year (random term) and nest (random term).

19
References
S1. F. Sergio et al., Coping with uncertainty: breeding adjustments to an unpredictable
environment in an opportunistic raptor. Oecologia 161; DOI 10.1007/s00442-010-1795-x
(2010).
S2. J. Blas, F. Sergio, F. Hiraldo, Age-related improvement in reproductive performance in a
long-lived raptor: a cross-sectional and longitudinal study. Ecography 32, 647-657 (2009).
S3. F. Sergio, J. Blas, F. Hiraldo, Predictors of floater status in a long-lived bird: a cross-
sectional and longitudinal test of hypotheses. J. Anim. Ecol. 78, 109-118 (2009).
S4. F. Sergio, J. Blas, M. G. Forero, J. A. Donázar, F. Hiraldo, Sequential settlement and site-
dependence in a migratory raptor. Behav. Ecol. 18, 811-821 (2007).
S5. F. Sergio et al., Short and long-term consequences of individual and territory quality in a
long-lived bird. Oecologia 160, 507-514 (2009).
S6. M. Alcaide, J. A. Lemus, G. Blanco, J. L. Tella, D. Serrano, J. J. Negro, A. Rodríguez, M.
García-Montijano, MHC diversity and differential exposure to pathogens in kestrels (Aves:
Falconidae). Molecular Ecology 19, 691-705 (2010).
S7. M. J. Crawley, The R book (Wiley, Chichester, 2007).
S8. A. F. Zur, E. N. Ieno, N. J. Walker, A. A. Saveliev, G. M. Smith, Mixed effects models and
extensions in ecology with R (Springer, New York, 2009).
S9. W. A. Searcy, S. Nowicki, The evolution of animal communication (Princeton Univ. Press,
Princeton, 2005).
S10. P. Mccullagh, J. A. Nelder, Generalized linear models (Chapman & Hall, London, 1989).
S11. S. P. Otto, T. Day, A biologist's guide to mathematical modeling in ecology and evolution
(Princeton Univ. Press, Princeton, 2007).
S12. F. Sergio, I. Newton, Occupancy as a measure of territory quality. J. Anim. Ecol. 72, 857-865
(2003).
S13. W. R. Rice, Analyzing tables of statistical tests. Evolution, 43, 223-225 (1989).
S14. L. Brower, J. Komdeur, Green nesting material has a function in mate attraction in the
European starling. Anim. Behav. 67, 539-548 (2004).
S15. L. Clark, Starlings as herbalists: countering parasites and pathogens. Parasitology Today 6,
248-360 (1990).
S16. P. H. Wimberger, The use of green plant material in bird nests to avoid ectoparasites. Auk
101, 615-618 (1984)