Retaliation in Response to Castration Promotes a Low Level
of Virulence in an Ant–Plant Mutualism
Pierre-Jean G. Male ´ •Jean-Baptiste Ferdy•Ce ´line Leroy•Olivier Roux•
Je ´re ´mie Lauth•Arthur Avilez•Alain Dejean•Ange ´lique Quilichini•
Je ´ro ˆme Orivel
Received: 7 March 2013/Accepted: 19 June 2013/Published online: 11 July 2013
? Springer Science+Business Media New York 2013
for its own benefit is a major source of instability in hor-
izontally-transmitted mutualisms. This instability can be
counter-balanced by the host’s retaliation against exploit-
ers. Such responses are crucial to the maintenance of the
relationship. We focus on this issue in an obligate ant–plant
mutualism in which the ants are known to partially castrate
their host plant. We studied plant responses to various
levels of castration in terms of (1) global vegetative
investment and (2) investment in myrmecophytic traits.
Castration led to a higher plant growth rate, signalling a
novel case of gigantism induced by parasitic castration. On
the other hand, completely castrated plants produced
smaller nesting and food resources (i.e. leaf pouches and
The diversion of a host’s energy by a symbiont
extra floral nectaries). Since the number of worker larvae is
correlated to the volume of the leaf pouches, such a
decrease in the investment in myrmecophytic traits dem-
onstrates for the first time the existence of inducible
retaliation mechanisms against too virulent castrating ants.
Over time, this mechanism promotes an intermediate level
of castration and enhances the stability of the mutualistic
relationship by providing the ants with more living space
while allowing the plant to reproduce.
Overexploitation ? Mutualism breakdown ? Allomerus
decemarticulatus ? Hirtella physophora
Evolutionary conflict ? Cheater ?
Mutualisms are defined as reciprocally beneficial interac-
tions between organisms, but they can also be viewed as
Electronic supplementary material
article (doi:10.1007/s11692-013-9242-7) contains supplementary
material, which is available to authorized users.
The online version of this
P.-J. G. Male ´ ? J.-B. Ferdy ? A. Avilez
ENFA, UMR5174 EDB (Laboratoire E´volution & Diversite ´
Biologique), Universite ´ de Toulouse, 118 route de Narbonne,
31062 Toulouse, France
P.-J. G. Male ´ ? J.-B. Ferdy ? A. Avilez
UMR5174 EDB, CNRS, Universite ´ Paul Sabatier,
31062 Toulouse, France
P.-J. G. Male ´ (&)
Laboratoire Evolution & Diversite ´ Biologique,
Universite ´ Paul Sabatier, 118 route de Narbonne,
31062 Toulouse Cedex 9, France
UMR 123, AMAP (botAnique et bioinforMatique de
l’Architecture des Plantes), IRD, Boulevard de la Lironde,
TA A51/PS2, 34398 Montpellier Cedex 5, France
UMR 224, Maladies Infectieuses et Vecteurs Ecologie,
Ge ´ne ´tique, Evolution et Contro ˆle, IRD, BP 171,
Bobo Dioulasso 01, Burkina Faso
J. Lauth ? A. Dejean ? A. Quilichini ? J. Orivel
UMR 8172, Ecologie des Fore ˆts de Guyane, CNRS, Campus
Agronomique, BP 316, 97379 Kourou Cedex, France
UMR Ecolab, Universite ´ de Toulouse, 118 route de Narbonne,
31062 Toulouse Cedex 9, France
Evol Biol (2014) 41:22–28
‘reciprocal exploitation’, thus involving both benefits and
costs (Herre et al. 1999). Since natural selection favours
individuals that are able to lower their costs and/or increase
their benefits, conflicts in resource allocation are expected
to cause mutualistic relationships to shift to parasitism
(Bronstein 2001; Edwards et al. 2010). Striking examples
are well-known in the framework of pollination mutual-
isms. Cheating has evolved in both partners and while
some plants produce flowers devoid of nectar, some insects
can rob nectar without pollinating their partner (Gilbert
et al. 1991; Inouye 1983).
The reproduction of each partner engaged in any hori-
zontally-transmitted mutualism is obviously subjected to a
conflict since the resources invested by one partner in its
own reproduction are lost for the other partner. Such a
trade-off is a major source of instability as, over time, the
symbionts may evolve in such a way that they end up
completely sterilizing their hosts. Incomplete sterilization
is thus considered ‘suboptimal virulence’ in the literature
(Jaenike 1996; O’Keefe and Antonovics 2002), which
highlights the existence of counterbalancing mechanisms
such as the local dispersal of both partners, partner choice
and/or retaliation against exploiters (Wilkinson and
Sherratt 2001; Szila ´gyi et al. 2009).
Ant–plant mutualisms provide an interesting framework
for studying such reproductive trade-offs. Myrmecophytic
plants typically provide ants with a nesting space in hollow
structures such as thorns or leaf pouches, and sometimes
food resources, in exchange for protection against predators
and competitors. These relationships are horizontally trans-
evolve to be parasitic on their hosts (Heil and McKey 2003;
Wilkinson and Sherratt 2001). Moreover, because of the
obligatory nature of most of these relationships, the plants
cannot escape such exploitation by ending the interaction
without drastically reducing their fitness. It has been dem-
onstrated in several model systems that the growth of the ant
colony is constrained by the nesting space provided by the
plant (Fonseca 1993; Orivel et al. 2011). The ants can take
advantage of the trade-off between the plant’s reproduction
and growth and induce gigantism in their host plant by cas-
trating it (Frederickson 2009). Such a conflict between the
illustrated by the castration behaviour exhibited by several
Gaumeetal.2005;Male ´ etal.2012).Thisis,however,based
on the major assumption that there is a trade-off between
vegetative growth and reproductionin myrmecophytes. Yet,
given the modular growth and autotrophy of plants, each
module could theoretically be able to pay its own resource
costs, for example, through greater photosynthesis in nearby
leaves and in some of its floral structures (Watson 1984;
Despite the ant-induced limitations to their reproduc-
tion, myrmecophytes manage to reproduce even in the
presence of ants. An intermediate level of castration has
indeed been observed in associations involving Allomerus
plant-ants that probably results from tolerance and/or local
dispersal mechanisms (Edwards and Yu 2008; Izzo and
Vasconcelos 2002; Male ´ et al. 2012; Szila ´gyi et al. 2009).
However, although never studied in these systems, resis-
tance mechanisms are thought to result in the same pattern
of intermediate virulence. Indeed, retaliation mechanisms
have recently been identified as a key factor in preventing
cheating in ant-plant mutualisms, especially in the case of
cheating by defection (Edwards et al. 2006, 2010). But
castration behaviour belongs to another kind of cheating
mechanism; namely, overexploitation. In the case of
defection, cheating results in a reduction in the partner’s
benefits, while overexploitation leads to an increase in its
costs (Douglas 2008, 2010). Moreover, tolerance and
resistance mechanisms lead to very different evolutionary
paths (Best et al. 2009). As a consequence, proving the
presence/absence of resistance mechanisms in ant–plants is
critical to our understanding of the maintenance of such
relationships despite possible exploitation.
In this study, we focus on one of these plant-Allomerus
systems involving Hirtella physophora and A. decemarti-
culatus in which an intermediate level of castration viru-
lence has been demonstrated (Male ´ et al. 2012). We aimed
to quantify (1) the net benefit of castration for the resident
ants in terms of the plant’s vegetative growth and thus
available nesting space, (2) the plant’s response to castra-
tion in terms of its investment in myrmecophytic traits (i.e.
leaf pouches and extra-floral nectaries), and (3) the putative
consequences of any change in the myrmecophytic traits
for the ant colony. We experimentally addressed this issue
by assessing the plant’s investment in myrmecophytic traits
when castration is partial, complete or non-existent and
analysed the implications of the different plant responses
(i.e. gigantism and retaliation) on the continuance of the
Materials and Methods
Study Sites and Model
This study was conducted in French Guiana between
August 2008 and December 2011 on two populations of
H. physophora less than one kilometre apart in the area of
Petit Saut, Sinnamary (05?0303000N; 52?5803400W) and one
population at the Nouragues research station (04?0401800N;
Hirtella physophora is an understory treelet that occurs
strictly in pristine Amazonian rainforest. Plant individuals
Evol Biol (2014) 41:22–28 23
have long-lived leaves that bear a pair of pouches (i.e.
domatia), at the base of each lamina (Leroy et al. 2008).
The domatium results from the curling under of the leaf
margin and the tissues show profound morpho-anatomical
modifications compared to the leaf lamina (Leroy et al.
2010). Circular extra-floral nectaries (EFNs) occur on the
abaxial surface of the lamina and inside the domatia. They
differ morphologically, with the EFNs in the domatia being
three times larger than the ones on the lamina (Leroy et al.
2008). H. physophora is a self-incompatible, entomoga-
mous species and two flowering periods can be distin-
guished annually: from December to February and June to
August (Male ´ 2011).
In the studyarea, plant individuals are almost exclusively
inhabited by A. decemarticulatus with a single colony per
plant (Solano et al. 2003). Moreover, A. decemarticulatus
has never been found in association with another myrm-
ecophyte (Grangier et al. 2009). As in any other protective
mutualism between ants and plants, A. decemarticulatus
workers protect their host plant from defoliators, thus
favouring its vegetative growth(Grangier et al. 2008; Orivel
et al. 2011). However, A. decemarticulatus also destroys
inverylowfruitproduction(Male ´ etal.2012).Thecastration
behaviour of A. decemarticulatus is thus characterizedby an
intermediate level of virulence.
A total of 54 flowering H. physophora inhabited by well-
developed colonies of A. decemarticulatus were randomly
assigned to one of three groups of 18 individuals. The plants
in these groups were similar in terms of size, base diameter,
and number of branches, leaves and inflorescences. Each
group was subjected to one of the three following experi-
an early stage in their development (‘total castration’); (2)
untreated plants underwent the natural, ant-induced partial
castration (‘intermediate castration’); and (3) ants were
excluded from all of the inflorescences by applying rings of
Tanglefoot?on either side of the peduncles on shoots pre-
viously protected by aluminium foil (‘no castration’). At the
beginning of the experiment, the youngest leaf on every
leaves and inflorescences was recorded. Over the course of
the experiment, the plants were subjected to the same
experimental treatment again when necessary (i.e. cutting
off the inflorescences or excluding ants at an early stage of
The experiment ended 17 months later, i.e. at the
beginning of the fourth flowering period. Nine individuals
(i.e. three in the ‘total castration’ group, four in the
‘intermediate castration’ group and two in the ‘no
castration’ group) were not taken into account because of
the death of their associated ant colony or of the plant itself
over the course of the experiment. We recorded the number
of remaining pre-treatment leaves. When possible, the two
youngest pre-treatment leaves and two leaves that had
reached maturity during the experiment were collected to
compare morphological characteristics between them. For
each leaf collected, we estimated its surface area using
ImageJ v1.34 software and quantified the volumes of the
two domatia. Domatia volume was estimated based on the
volume of an ellipsoid for which the three elliptic radii
were equal to the length, width and height of the domatium.
We also counted the number of EFNs and measured the
surface areas of four of them on the lamina and four others
in the domatia.
We then assessed the effect of the experimental treat-
ment on myrmecophytic characteristics by comparing (1)
the domatia volumes, (2) the number of EFNs per leaf, and
(3) the surface areas of the domatia and lamina EFNs based
on the experimental treatment and with pre-treatment
leaves. The plants’ investment in vegetative parts was
assessed by comparing the number of newly-produced
leaves and the abscission rate of old leaves between the
three groups. We also compared the surface areas of the
leaves based on the three experimental treatments and with
the pre-treatment leaves. The effect of the experimental
treatments on plant investment in reproduction was quan-
tified by comparing the final number of inflorescences
between the three experimental groups.
We also sampled the youngest five leaves on one branch
for 20 different plants and the penultimate leaf on up to
eight branches for 20 other H. physophora individuals
inhabited by A. decemarticulatus. The domatia volume and
the number and surface area of the domatia EFNs were
determined as described above. We counted the number of
worker larvae in each domatium and then verified that A.
decemarticulatus larvae were preferentially clustered in the
penultimate pair of domatia. As a consequence, only the
penultimate pair of domatia was used in the subsequent
analyses. Since numerous domatia were devoid of larvae,
we chose to test separately whether the presence/absence of
larvae in the domatia was influenced by domatia volume,
and, when larvae were present, whether their number was
influenced by domatia volume.
All of the statistical analyses were conducted using R
(R Development Core Team 2009). The data were analyzed
using generalized linear mixed-effect models (GLMM),
most of them conducted with the lme function combined
with the Anova function in the nlme and car packages,
respectively (Pinheiro et al. 2009; Fox and Weisberg
24Evol Biol (2014) 41:22–28
2011). Population or plant ID was modeled as a random
block effect in all of the analyses. When appropriate, plant
or leaf ID was nested in population/plant ID. We assessed
model fit through the visual evaluation of residual plots and
by using the Shapiro–Wilk normality test when appropri-
ate. We tested for the autocorrelation and heteroskedas-
ticity of the residuals by conducting the Durbin–Watson
and Breusch–Pagan tests using, respectively, the dwtest
and the bptest functions in the lmtest package (Zeileis and
Hothorn 2002). When necessary, data were Box-Cox-
transformed using the boxcox function in the MASS
package (Venables and Ripley 2002). Because of their
binomial nature, the abscission rate of old leaves and the
probability of larval presence in the domatia were analyzed
using the lmer function in the lme4 package (Bates and
Maechler 2009) combined with the Anova function.
Castrated Plants Produce More Leaves
Non-castrated plants produced fewer leaves than totally or
partially castrated ones (Fig. 1; see also electronic supple-
leaves (GLMM: df = 2; v2= 0.067; p = 0.9672). There
was, however, no significant difference in leaf production
electronic supplementary material). During the flowering
period that immediately followed the end of the experiment,
(GLMM: df = 2; v2= 2.923, p = 0.2319). This result
shows that the investment by plants in reproduction was not
affected by the outcome of the previous flowering periods.
Castrated Plants Produce Smaller Domatia
The surface area of new leaves did not differ significantly
Number of leaves produced
Fig. 1 Barplot of the mean (± SE) number of leaves produced
during the experiment as a function of the treatment (‘No castration’,
plants with inflorescences from which ants were excluded; ‘Interme-
diate castration’, untreated plants experiencing natural, ant-induced
partial castration; ‘Total castration’, plants with all of their inflores-
cences experimentally removed at an early stage). Population ID was
taken into account as a random effect. Letters indicate significant
differences in the number of leaves (GLMM, p values\0.05)
Volume of the domatia (mm3)
Diameter of the extra−floral
nectaries on laminas (mm)
Diameter of the extra−floral
nectaries in domatia (mm)
Fig. 2 Barplot of the mean (± SE) for three morphological traits
[a volume of domatia; b diameter of the extra-floral nectaries located
on the laminas; c diameter of the extra-floral nectaries located inside
the domatia] as a function of the treatment (‘Pre-treatment’, leaves
produced before the beginning of the experiment; ‘No castration’,
leaves produced during the experiment by plants with inflorescences
from which the ants were excluded; ‘Intermediate castration’, leaves
produced during the experiment by untreated plants (i.e. ant-induced,
partial castration); ‘Total castration’, leaves produced during the
experiment by plants with all of their inflorescences experimentally
removed at an early stage). Plant ID was taken into account as a
random effect. Letters indicate significant differences (GLMM,
Evol Biol (2014) 41:22–28 25
leaves (GLMM: df = 3; v2= 3.222, p = 0.3587), but new
leaves on totally castrated plants bore significantly smaller
which experienced partial castration by ants (GLMM: df = 3;
v2= 15.758, p = 0.0013; Fig. 2a; see also electronic sup-
not affect the total number of EFNs per leaf (GLMM: df = 3;
v2= 1.346, p = 0.7184), totally castrated plants produced
leaves with significantly smaller lamina and domatia EFNs
p = 0.0190; domatia EFN: df = 3; v2= 7.903, p = 0.0049;
Fig. 2b, c; see also electronic supplementary material).
df = 3;
Smaller Domatia Shelter Less Ant Larvae
In non-manipulated plants, worker larvae were preferentially
clustered in the domatia of the penultimate leaf (GLMM:
df = 4; v2= 108.75, p\0.0001; Fig. 3; see also electronic
the penultimate pair of domatia increased with domatia vol-
ume but was influenced neither by the number of domatia
Similarly, when present, the quantity of larvae in the penul-
timate pair of domatia was also positively influenced by
that it shelters any larvae, and if it nevertheless does, the less
larvae it shelters.
Allomerus ants are known to destroy the inflorescences of
their host plant (Frederickson 2009; Male ´ et al. 2012;
Edwards and Yu 2008). Our results demonstrate that this
castration behaviour induces an increase in the H. physo-
phora growth rate: castrated plants produce more leaves
than those bearing inflorescences from which ants have
been excluded. This confirms the idea that A. decemarti-
culatus colonies obtain more living space than putative
non-castrating ants thanks to their castrating behaviour. By
castrating their host plant, A. decemarticulatus colonies
thus exploit the trade-off between vegetative growth and
reproduction inH. physophora.
destruction of the plant’s sexual organs can be assimilated
to parasitic castration inducing gigantism (Hall et al. 2007)
which has already been observed in another ant-plant
interaction (Frederickson 2009).
We show that the total castration of H. physophora
induces a decrease in its investment in myrmecophytic
traits. This reduction affects both housing and food
resources with a decrease in the domatia volume and a
reduction in the surface areas of the EFNs, but does not
affect non-myrmecophytic traits such as surface area of
leaves. The loss of space resulting from the decrease in the
volume of the domatia is likely to directly affect the
development and fitness of too virulent colonies, as nesting
space is a major limiting factor to the growth and repro-
duction of the A. decemarticulatus colonies (Orivel et al.
2011). Our results demonstrate that a reduction in the
investment in myrmecophytic traits affects the newly-pro-
duced domatia that are, under natural conditions, the ones
that host most of the ant brood. Moreover, the number of
larvae that are hosted in a domatium depends on the vol-
ume of that domatium. As a consequence, the development
of a putative fully-castrating ant colony is very likely to be
slower than that of a partially-castrating ant colony. This
slowing down in colony growth should directly translate
into a loss of fitness because the production of alates in
A. decemarticulatus has been shown to be positively cor-
related to the number of workers (Orivel et al. 2011).
One still might wonder if the decrease in domatia vol-
ume can be compensated by an increase in the number of
domatia. Because the probability of sheltering larvae, and
not only the number of larvae, is also negatively correlated
to domatia volume, small domatia can be completely lost
for the ant colony. As a consequence, the cost of total
castration could be greater in terms of a loss in fitness for
the ant colony than the net benefit of more leaves, i.e.
nesting units, produced. Moreover, the cost of total cas-
tration also affects the size of the EFNs located inside the
domatia. Since the quantity of nectar produced is propor-
tional to the volume of the nectarial parenchyma
Number of larvae in domatia
One TwoThree FourFive
Fig. 3 Barplot of the mean (± SE) number of larvae sheltered in the
domatia as a function of the leaf location (‘Leaf one’, the youngest
leaf on the branch; the other leaves are numbered subsequently).
Letters indicate significant differences (GLMM, p values\0.05)
26Evol Biol (2014) 41:22–28
(Pacini et al. 2003), the smaller the EFNs, the less nectar
they are believed to produce. Ants inhabiting totally cas-
trated plants should thus experience a potentially harmful
decrease in their carbohydrate intake which might poten-
tially affect the performance of the colony (Dussutour and
Simpson 2012). By analogy with plant-rhizobia interac-
tions, both a reduction in living space and in the amount of
food provided can be considered retaliation by the plant
(West et al. 2002). Such retaliation mechanisms against
castrating parasites are shown here for the first time in the
framework of ant–plant interactions.
The destruction of a fraction of the sexual organs seems
to be the best compromise because it maximizes the cost-
to-benefit ratio for both partners. Retaliation was only
detected when the virulence of the castration was at its
maximum. As a consequence, over the long term, it is
unlikely that ants that totally castrate their host plant gain a
fitness advantage compared to ants with an intermediate
level of virulence. Nevertheless, intermediate castration
virulence is not an optimum situation for H. physophora.
Plants with a lower tolerance to castration should obtain a
higher fitness benefit. But the global fitness of an individual
has to be considered over its lifespan. Indeed, Palmer et al.
(2010) demonstrated that the provisional association with
sterilizing ants enhanced the lifetime fitness of Acacia
drepanolobium. As the energy that a plant invests in
reproduction is lost for vegetative investment, there is a
trade-off between reproduction and survival, leading to a
trade-off between current reproductive output and future
reproduction (Obeso 2002). The ant-induced reduction in
one reproductive event is thus likely to result in an increase
in the life expectancy of the plant, and, consequently, in an
increase in the plant’s subsequent reproductive success.
ment de Petit Saut and the Nouragues scientific station for furnishing
We are grateful to the Laboratoire Environne-
logistical help, to Dr. Jacqui Shykoff and Pr. Doyle McKey for
insightful comments and to Andrea Yockey-Dejean for proofreading
the manuscript. Financial support for this study was provided by a
research program of the French Agence Nationale de la Recherche
(research agreement n?ANR-06-JCJC-0109-01), by the ESF-EURO-
CORES/TECT/BIOCONTRACT program, by the Fondation pour la
Recherche sur la Biodiversite ´ (research agreement n?AAP-IN-2009-
050), by the Programme Convergence 2007–2013 Re ´gion Guyane
from the European Community, and by the Programme Amazonie II
of the French Centre National de la Recherche Scientifique. This
work has benefited from ‘‘Investissement d’Avenir’’ grants managed
by the Agence Nationale de la Recherche (CEBA, ref. ANR-10-
LABX-25-01 and TULIP, ref. ANR -10-LABX-0041).
Bates, D., & Maechler, M. (2009). lme4: Linear mixed-effects models
using S4 classes. (R package version 0.999375-32 ed.).
Best, A., White, A., & Boots, M. (2009). Resistance is futile but
tolerance can explain why parasites do not always castrate their
hosts. Evolution, 64(2), 348–357. doi:10.1111/j.1558-5646.2009.
Bronstein, J. L. (2001). The exploitation of mutualisms. Ecology
Letters, 4(3), 277–287.
Douglas, A. E. (2008). Conflict, cheats and the persistence of
symbioses. New Phytologist, 177(4), 849–858. doi:10.1111/
Douglas, A. E. (2010). The symbiotic habit. Princeton and Oxford:
Princeton University Press.
Dussutour, A., & Simpson, S. J. (2012). Ant workers die young and
colonies collapse when fed a high-protein diet. In Proceedings of
the Royal Society B-Biological Sciences: online. doi:10.1098/
Edwards, D. P., Ansell, F. A., Woodcock, P., Fayle, T. M., Chey, V.
K., & Hamer, K. C. (2010). Can the failure to punish promote
cheating in mutualism? Oikos, 119, 45–52. doi:10.1111/j.1600-
Edwards, D. P., Hassall, M., Sutherland, W. J., & Yu, D. W. (2006).
Royal Society B-Biological Sciences, 273, 595–602. doi:10.1098/
Table 1 Effect of domatia volume, number of domatia EFNs and mean diameter of domatia EFNs on (a) presence/absence of larvae in domatia,
and (b) number of larvae when present in domatia
(a) Presence/absence of larvae in domatia
Domatia volume0.014 0.0071 4.00520.0454*
Number of domatia EFNs-0.1010.1611 0.38800.5334
Mean surface area of domatia EFNs1.556 2.8241 0.3037 0.5816
(b) Number of larvae when present in domatia
Domatia volume0.0160.0051 10.003 0.0016**
Number of domatia EFNs-0.026 0.1321 0.038 0.8460
Mean surface area of domatia EFNs0.6621.9071 0.1210.7285
DF number of degrees of freedom, SE standard error
* Statistical significance at the 5 % level, ** Statistical significance at the 1 % level
Evol Biol (2014) 41:22–28 27
Edwards, D. P., & Yu, D. W. (2008). Tolerating castration by hiding Download full-text
flowers in plain sight. Behavioral Ecology and Sociobiology,
63(1), 95–102. doi:10.1007/s00265-008-0639-8.
Fonseca, C. R. (1993). Nesting space limits colony size of the plant–
ant Pseudomyrmex concolor. Oikos, 67, 473–482.
Fox, J., & Weisberg, S. (2011). An R companion to applied
regression. Thousand Oaks, CA: Sage.
Frederickson, M. E. (2009). Conflict over reproduction in an ant-plant
symbiosis: Why Allomerus octoarticulatus ants sterilize Cordia
nodosa trees? American Naturalist, 173(5), 675–681. doi:10.
Gaume, L., Zacharias, M., & Borges, R. M. (2005). Ant–plant
conflicts and a novel case of castration parasitism in a
myrmecophyte. Evolutionary Ecology Research, 7, 435–452.
Gilbert, F. S., Haines, N., & Dickson, K. (1991). Empty flowers.
Functional Ecology, 5, 29–39.
Grangier, J., Dejean, A., Male ´, P.-J. G., & Orivel, J. (2008). Indirect
defense in a highly specific ant–plant mutualism. Naturwissens-
chaften, 95(10), 909–916. doi:10.1007/s00114-008-0398-4.
Grangier, J., Dejean, A., Male ´, P.-J. G., Solano, P.-J., & Orivel, J.
(2009). Mechanisms driving the specificity of a myrmecophyte–
ant association. Biological Journal of the Linnean Society, 97,
Hall, S. R., Becker, C., & Ca ´ceres, C. E. (2007). Parasitic castration: A
perspective from a model of dynamic energy budgets. Integrative
Heil, M., & McKey, D. (2003). Protective ant–plant interactions as
model systems in ecological and evolutionary research. Annual
Review of Ecology Evolution and Systematics, 34, 425–453.
Herre, E. A., Knowlton, N., Mueller, U. G., & Rehner, S. A. (1999).
The evolution of mutualisms: Exploring the paths between
conflict and cooperation. Trends in Ecology & Evolution, 14(2),
Inouye, D. W. (1983). The ecology of nectar robbing. In B. Bentley &
T. Elias (Eds.), The biology of nectaries. New York, USA:
Columbia University Press.
Izzo, T. J., & Vasconcelos, H. L. (2002). Cheating the cheater:
Domatia loss minimizes the effects of ant castration in an
Amazonian ant-plant. Oecologia, 133, 200–205. doi:10.1007/
Jaenike, J. (1996). Suboptimal virulence of an insect–parasitic
nematode. Evolution, 50(6), 2241–2247.
Leroy, C., Jauneau, A., Quilichini, A., Dejean, A., & Orivel, J. (2008).
Comparison between the anatomical and morphological structure
of leaf blades and foliar domatia in the ant–plant Hirtella
Leroy, C., Jauneau, A., Quilichini, A., Dejean, A., & Orivel, J. (2010).
Comparative structure and ontogeny of the foliar domatia in
three neotropical myrmecophytes. American Journal of Botany,
97, 557–565. doi:10.3732/ajb.0900207.
Male ´, P.-J. G. (2011). Stabilite ´ e ´volutive des mutualismes et
me ´canismes de contro ˆle : le cas d’une relation plante-fourmis.
PhD thesis, Universite ´ de Toulouse, Toulouse, France.
Male ´, P.-J. G., Leroy, C., Dejean, A., Quilichini, A., & Orivel, J.
(2012). An ant symbiont directly and indirectly limits its host
plant’s reproductive success. Evolutionary Ecology, 26(1),
155, 321–348. doi:10.1046/j.1469-8137.2002.00477.x.
O’Keefe, K. J., & Antonovics, J. (2002). Playing by different rules:
The evolution of virulence in sterilizing pathogens. American
Naturalist, 159(6), 597–605. doi:10.1086/339990.
Orivel, J., Lambs, L., Male ´, P.-J. G., Leroy, C., Grangier, J., Otto, T.,
et al. (2011). Dynamics of the association between a long-lived
understory myrmecophyte and its specific associated ants.
Oecologia, 165, 369–376. doi:10.1007/s00442-010-1739-5.
Pacini, E., Nepi, M., & Vesprini, J. L. (2003). Nectar biodiversity: A
short review. Plant Systematics and Evolution, 238, 7–21.
Palmer, T. M., Doak, D. F., Stanton, M. L., Bronstein, J. L., Kiers, E.
T., Young, T. P., et al. (2010). Synergy of multiple partners,
including freeloaders, increases host fitness in a multispecies
mutualism. Proceedings of the National Academy of Sciences,
Pinheiro, J., Bates, D., DebRoy, S., Sarkar, D., & The R Core Team.
(2009). nlme: Linear and Nonlinear Mixed Effects Models. (R
package version 3.1-96 ed.).
R Development Core Team (2009). R: A Language and Environment
for Statistical Computing. In R. F. f. S. Computing (Ed.). Vienna,
Solano, P.-J., Durou, S., Corbara, B., Quilichini, A., Cerdan, P.,
Belin-Dupoux, M., et al. (2003). Myrmecophytes of the under-
story of French Guianian rainforests: Their distribution and their
associated ants. Sociobiology, 41(3), 605–614.
Szila ´gyi, A., Scheuring, I., Edwards, D. P., Orivel, J., & Yu, D. W.
(2009). The evolution of intermediate castration virulence and
ant coexistence in a spatially structured environment. Ecology
Letters, 12(12), 1306–1316.
Venables, W. N., & Ripley, B. D. (2002). Modern applied statistics
with S (4th ed.). New York: Springer.
Watson, M. A. (1984). Developmental constraints: effect on popu-
lation growth and patterns of resource allocation in a clonal
plant. American Naturalist, 123(3), 411–426. doi:10.1086/
West, S. A., Kiers, E. T., Simms, E. L., & Denison, R. F. (2002).
Sanctions and mutualism stability: Why do rhizobia fix nitrogen?
Proceedings of the Royal Society B-Biological Sciences, 269,
Wilkinson, D. M., & Sherratt, T. N. (2001). Horizontally acquired
mutualisms, an unsolved problem in ecology? Oikos, 92(2),
Yu, D. W., & Pierce, N. E. (1998). A castration parasite of an ant-
plant mutualism. Proceedings of the Royal Society B-Biological
Sciences, 265, 375–382. doi:10.1098/rspb.1998.0305.
Zeileis, A., & Hothorn, T. (2002). Diagnostic checking in regression
relationships. R News, 2(3), 7–10.
28Evol Biol (2014) 41:22–28