The hunter becomes the hunted: When cleptobiotic insects are captured by their target ants

Article (PDF Available)inThe Science of Nature 99(4):265-73 · February 2012with101 Reads
DOI: 10.1007/s00114-012-0895-3 · Source: PubMed
Here we show that trying to rob prey (cleptobiosis) from a highly specialized predatory ant species is risky. To capture prey, Allomerus decemarticulatus workers build gallery-shaped traps on the stems of their associated myrmecophyte, Hirtella physophora. We wondered whether the frequent presence of immobilized prey on the trap attracted flying cleptoparasites. Nine social wasp species nest in the H. physophora foliage; of the six species studied, only Angiopolybia pallens rob prey from Allomerus colonies. For those H. physophora not sheltering wasps, we noted cleptobiosis by stingless bees (Trigona), social wasps (A. pallens and five Agelaia species), assassin bugs (Reduviidae), and flies. A relationship between the size of the robbers and their rate of capture by ambushing Allomerus workers was established for social wasps; small wasps were easily captured, while the largest never were. Reduviids, which are slow to extract their rostrum from prey, were always captured, while Trigona and flies often escaped. The balance sheet for the ants was positive vis-à-vis the reduviids and four out of the six social wasp species. For the latter, wasps began by cutting up parts of the prey's abdomen and were captured (or abandoned the prey) before the entire abdomen was retrieved so that the total weight of the captured wasps exceeded that of the prey abdomens. For A. pallens, we show that the number of individuals captured during attempts at cleptobiosis increases with the size of the Allomerus' prey.
The hunter becomes the hunted: when cleptobiotic insects
are captured by their target ants
Alain Dejean &James M. Carpenter &Bruno Corbara &
Pamela Wright &Olivier Roux &Louis M. LaPierre
Received: 26 October 2011 /Revised: 31 December 2011 /Accepted: 4 February 2012 /Published online: 24 February 2012
#Springer-Verlag 2012
Abstract Here we show that trying to rob prey (cleptobiosis)
from a highly specialized predatory ant species is risky. To
capture prey, Allomerus decemarticulatus workers build
gallery-shaped traps on the stems of their associated myrme-
cophyte, Hirtella physophora. We wondered whether the
frequent presence of immobilized prey on the trap attracted
flying cleptoparasites. Nine social wasp species nest in the H.
physophora foliage; of the six species studied, only
Angiopolybia pallens rob prey from Allomerus colonies. For
those H. physophora not sheltering wasps, we noted clepto-
biosis by stingless bees (Trigona), social wasps (A. pallens
and five Agelaia species), assassin bugs (Reduviidae), and
flies. A relationship between the size of the robbers and their
rate of capture by ambushing Allomerus workers was estab-
lished for social wasps; small wasps were easily captured,
while the largest never were. Reduviids, which are slow to
extract their rostrum from prey, were always captured, while
Trigon a and flies often escaped. The balance sheet for the ants
Communicated by: Sven Thatje
The hunter becomes the hunted
corresponds to the French proverb Tel
est pris qui croyait prendre in the fable Le rat et lHuitre by Jean de la
Fontaine, Livre VIII, 1678. Translation: He is caught who thought to
Electronic supplementary material The online version of this article
(doi:10.1007/s00114-012-0895-3) contains supplementary material,
which is available to authorized users.
A. Dejean (*)
Écologie des Forêts de Guyane (UMR-CNRS 8172), CNRS,
Campus Agronomique, BP 316, 97379 Kourou cedex, France
A. Dejean
Université de Toulouse (UPS), Ecolab,
118 route de Narbonne,
31062 Toulouse, France
J. M. Carpenter
Division of Invertebrate Zoology,
American Museum of Natural History,
Central Park West at 79th Street,
New York, NY 10024, USA
B. Corbara
Laboratoire Microorganismes, Génome et Environnement,
CNRS (UMR-CNRS 6023) Université Blaise Pascal,
Complexe Scientifique des Cézeaux,
63177 Aubière cedex, France
B. Corbara
Clermont UniversitéUniversité Blaise Pascal, LMGE,
BP 10448, 63000 Clermont-Ferrand, France
P. Wright
2833 Nichols Bvd,
Longview, WA 98632, USA
O. Roux
Maladies Infectieuses et Vecteurs, Ecologie, Génétique,
Evolution et Contrôle (UMR-IRD 224), Équipe BEES-IRD, IRD,
BP 171, Bobo-Dioulasso, Burkina Faso
L. M. LaPierre
Department of Biology, Lower Columbia College,
1600 Maple St.,
Longview, WA 98632, USA
Naturwissenschaften (2012) 99:265273
DOI 10.1007/s00114-012-0895-3
was positive vis-à-vis the reduviids and four out of the six
social wasp species. For the latter, wasps began by cutting up
parts of the preys abdomen and were captured (or abandoned
the prey) before the entire abdomen was retrieved so that the
total weight of the captured wasps exceeded that of the prey
abdomens. For A. pallens, we show that the number of indi-
viduals captured during attempts at cleptobiosis increases with
the size of the Allomerusprey.
Keywords Ant predation .Myrmecophyte .Cleptobiosis .
Social wasps .Stingless bees .Flies and Reduviidae
Myrmecophytes live in mutualistic associations with a lim-
ited number of so-called plant-ants that they provide with a
nesting place in pre-existing cavities (domatia) such as leaf
pouches and hollow stems or thorns and usually food in the
form of extra-floral nectar (EFN) and/or food bodies (FBs).
In return, plant-ants protect their host myrmecophytes from
defoliators through their predatory and/or territorial behav-
ior (Heil and McKey 2003; Rico-Gray and Oliveira 2007;
Dejean et al. 2009). Except for myrmecophytic Acacia,
Piper,andMacaranga that produce protein-rich FBs and
whose mutualistic plant-ants do not hunt, in general FBs and
EFN are carbohydrate-rich and comparatively poor in pro-
teins and amino acids (Heil and McKey2003; Rico-Gray
and Oliveira 2007). To obtain the proteins and amino acids
necessary to satisfy their nitrogen requirements, plant-ants
have optimized their ability to capture prey in the restricted
environment represented by the crowns of their host myr-
mecophytes by using elaborate hunting techniques (Dejean
et al. 2008,2009,2010). Allomerus decemarticulatus work-
ers even build a gallery-shaped trap to ambush prey, permit-
ting them to capture insects of up to 1,800 times their weight
(4.5 cm in length and weighing as much as 0.404 g). Yet, it
generally takes the ants a long timesometimes more than
9hto dismantle a prey once it has been captured and
spread-eagle (Dejean et al. 2005).
Consequently, while being spread-eagle on the trap, prey
may attract predatory flying insects so that this situation
likely favors food robbing (cleptobiosis). Yet, due to A.
decemarticulatus workersparticular ability to capture prey,
a cleptobiont risks being preyed upon in turn when alighting
on a trap in an attempt to rob prey. In this situation, the costs
versus the rewards are asymmetrical. For the cleptobionts,
there is a trade-off between a nutritional gain (parts of or
whole prey obtained as immobilized prey is easier to detect)
and the risk of being captured in turn, while from a nutri-
tional standpoint for the ants there is a balance between prey
lost to cleptobionts and cleptobionts captured as prey. Yet, it
is difficult to predict if the balance will be positive either for
certain cleptobionts or for the ants as several factors inter-
vene such as the size of the prey firstly captured by the ants
and the size and ability of the cleptobionts to escape quickly.
Among potential cleptobionts of prey captured by A.
decemarticulatus workers, social wasps are good candidates
because wasps robbing food from plant-ants have already
been reported (LaPierre et al. 2007). Furthermore, some
species of social wasps specifically build their nest in the
foliage of myrmecophytes. For the wasps, it is a means of
being protected from army ants which generally do not
climb up myrmecophytes as they are likely repelled by
compounds spread on the plant by the resident plant-ants
(Herre et al. 1986; Dejean et al. 2001; Corbara et al. 2009).
In such tripartite associations, social wasps are constantly in
the vicinity of ant workers and the food the ants retrieve to
their nest.
In this study, we therefore verified what flying insects can
attempt to rob entire prey or parts of prey from A. decemar-
ticulatus workers with special attention paid to social wasps
that build their nests on H. physophora trees. We also
investigated if, for the A. decemarticulatus colonies, the cost
of having pieces of prey robbed might not be compensated
by the gains made in capturing the robbers.
Materials and methods
Study area
This study was conducted between June 2007 and
November 2010 in the understory of the primary rainforest
surrounding the Petit Saut dam, Sinnamary, French Guiana
(5°0330.0N, 52°5834.6W). The climate is tropical
moist, with 3,500 mm of annual precipitation distributed
over 280 days. There is a major drop in rainfall between
July and November (dry season) and another shorter and
more irregular dry period in March.
The myrmecophyte Hirtella physophora
H. physophora (Chrysobalanaceae) occurs strictly in the
understory of pristine Amazonian forests and mostly in
patches located on the upper slopes of hillsides. These
treelets, which are only exceptionally more than 3 m in
height (but can be up to 6 m in height), have long-lived
leaves that bear extrafloral nectaries and a pair of pouches
(domatia) at the base of each lamina. This species is almost
always associated with A. decemarticulatus colonies
(Myrmicinae) with a single colony per plant; reciprocally,
the latter has never been found in association with another
myrmecophyte species. H. physophora has a much longer
lifespan (up to about 350 years) than its associated ant
colonies (up to about 21 years) (Orivel et al. 2011).
266 Naturwissenschaften (2012) 99:265273
Wasps sheltered by H. physophora treelets
In June and July 2007, we verified if there were social wasp
colonies on H. physophora by examining 545 treelets shelter-
ing an Allomerus colony and 122 others without ants.
In July 2007, we also conducted an experiment consist-
ing of furnishing ca. 2.5-cm-long immature female grass-
hoppers (Tettigoniidae) to the Allomerus colonies on 48 H.
physophora sheltering a wasp colony. In all cases, these
grasshoppers were immediately captured by the Allomerus
workers ambushing in their gallery-shaped traps. We then
noted during 1 h if any of the associated wasps attempted to
steal pieces of prey.
Myrmecophyte visitors
The same experiment as the one described above was con-
ducted between July 2009 and April 2010 on 158 H. phys-
ophora that sheltered Allomerus ants but not wasp colonies.
We noted what social wasp species and other insects attemp-
ted to rob pieces of prey (ca. 2.5-cm-long immature female
tettigoniids) during three series of 1-h-long observations
conducted on sunny days. We also noted if these would-
be-robbers were captured in turn. Using a microbalance, we
weighed 15 individuals from each insect species noted in the
process of robbing prey in order to verify if a relationship
could be established between their size (represented by their
weight) and the rate of their capture by the Allomerus ants.
We also weighed 15 ca. 2.5-cm-long female tettigoniids and
then their abdomens only, resulting in 0.28±0.016 and
0.086±0.005 g, respectively.
As Angiopolybia pallens workers were the wasps most
frequently recorded as robbing pieces of prey captured by
ambushing Allomerus ants, we conducted an experiment on
this species of social wasp. Between July 2009 and
November 2010, in areas where we had already noted for-
aging A. pallens workers during previous experiments, we
furnished calibrated pieces of prey to ambushing Allomerus
ants. These pieces of prey corresponded to portions cut from
the thorax plus the first part of the abdomen of large acridid
grasshoppers. Four size classes were prepared: 1.0 cm in
length and weighing 0.060.08 g, 1.5 cm in length and
weighing 0.130.15 g, 2.0 cm in length and weighing
0.200.22 g, and 2.5 cm in length and weighing 0.27
0.29 g. After 1.5 h, we noted the number of wasp individ-
uals that were captured by the ants while robbing parts of the
pieces of prey. This study was conducted at six sites likely
corresponding to the foraging area of A. pallens colonies
and made up of eight to 14 surveys separated by more than
1 month for a given site. This permitted us to obtain 15 cases
for each prey part class size (60 cases in total). We did not
take into consideration the cases when other insects (e.g.,
Trigona, other wasps or reduviids) robbed the prey parts.
Additional observations were conducted in 2008 and
2009 during photo-documenting sessions using different
prey species. These sessions occurred during a study on
Zelus annulosus (a reduviid Heteroptera frequently present
on H. physophora sheltering Allomerus colonies; Revel et
al. 2010) and during a study where we used pieces of
grasshopper (more attractive to robbing insects than an
entire prey), permitting us to use a net to collect Trigona
bees and social wasps retrieving parts of prey. The latter
were then weighed (ten to 25 pieces of prey per species).
Voucher specimens of wasps were captured, preserved in
70% ethanol, and then deposited in the American Museum
of Natural History, New York, NY, USA. Statistical analyses
were conducted using Past 2.09 software.
Wasps sheltered by Hirtella physophora treelets
Out of 122 H. physophora individuals not occupied by ants,
only one sheltered a social wasp nest (a Mischocyttarus
lecointei lecointei nest; 0.8%), while of the 545 treelets
occupied by an Allomerus colony 87 sheltered the nests of
nine social wasp species (14.5%), with M. lecointei lecointei
being the most frequent (see also Fig. S1in Electronic
supplementary material). The difference between the two
groups of trees is significant (Fisher's exact-text; P< 0.0001;
Table 1).
When we conducted experiments where we provided ca.
2.5-cm-long immature female tettigoniids to ambushing
Allomerus workers whose host trees also sheltered a wasp
nest, only A. pallens workers tried to rob the prey from the
ants. In no case did Mischocyttarus spp. workers try to rob
pieces of prey (Table 2;Polybia spp. and Polistes pacificus
were not noted in the field during this study). During these
observations, we noted that wasps from other species also
tried to rob pieces of prey, explaining why we conducted the
next experiment concerning non-myrmecophyte-nesting
Myrmecophyte visitors
When we furnished ca. 2.5-cm-long Tettigoniidae to ambush-
ing Allomerus workers whose host trees did not shelter a wasp
nest, we noted that the workers of a stingless bee species and
six social wasp species as well as different species of flies and
reduviid heteropterans visited the Hirtella treelets (Table 3).
Whereas stingless bees and social wasps cut off and robbed
pieces of prey, reduviids probed them to feed on the haemo-
lymph and flies licked the fluids that leaked out when the prey
were spread-eagle by the ants or after they had been bitten by
stingless bees or wasps. Stingless bees and social wasps made
Naturwissenschaften (2012) 99:265273 267
several return trips (we were frequently able to follow them
between their nests and the experimental tree; the returnpath
was direct). Stingless bees obviously recruited nestmates. In
social wasps, the same is likely true for a primitive form of
recruitment as the ratio between the number of individuals
robbing prey and the number of visits was greater than 1
(Table 3), and we noted the simultaneous presence of several
individuals from the same species on the same prey (Fig. S2in
Electronic supplementary material).
During this process all of these robbers, at one moment or
another, placed the extremity of a leg on the gallery-shaped
trap so that the Allomerus workers tried to catch them in turn
and were frequently successful (Table 3). The rate of suc-
cessful capture of these robbers was related to: (1) their size
(for social wasps, a relationship between the rate of capture
and size, represented by their weight, was established; the
smallest individuals were more easily captured than the
largest) (Fig. 1), (2) their behavior (stingless bees avoided
landing on the trap), and (3) their ability to fly away very
quickly when Allomerus workers tried to seize the extrem-
ities of their legs (stingless bees and particularly flies). In
this situation, reduviids, which require substantial time to
extract their rostra from the prey, were very vulnerable as all
of them were captured by the ants, including 27 more
individuals noted during additional observations (a total of
33 reduviids out of 33 captured by Allomerus workers).
The additional observations also permitted us to note 352
more cases of wasps robbing pieces of prey from Allomerus
colonies and to confirm the fact that Agelaia myrmecophila
and A. pallens are very easily captured. In contrast, Agelaia
fulvofasciata was captured only once out of 39 cases and
Agelaia testacea was never captured (38 cases). Workers
from these latter two species can be seized by the extremi-
ties of their legs by ambushing Allomerus workers but were
powerful enough to escape immediately or by cutting the
traps into pieces to free their legs before the ants were able
to grasp the extremities of their other legs.
We noted that social wasps involved in robbing pieces of
prey always began by cutting off parts of the prey abdomen (all
cases from Table 3plus those from additional observations
where we used different taxa as prey; Fig. 2; see also Figs.
S3S5in Electronic supplementary material). We therefore
estimated the total weight of wasp individuals likely to be
captured by the Allomerus ants and determined if it exceeded
that of the abdomens of 2.5 cm-long grasshoppers to establish
the potential costsbenefits of robbed prey parts vs. captured
wasps (Table 4). Indeed the Allomerus workers tried to catch
the wasps when they returned to rob new prey parts, seizing the
extremities of their legs so that those wasps that escaped these
attempts at capture stopped returning to the prey before having
retrieved the entire prey abdomen. This was not the case for A.
fulvofasciata and A. testacea that were never captured (but
Table 1 Comparison of the
percentage of social wasp nests
on H. physophora treelets
whether they were occupied or
not by an A. decemarticulatus
Hirtella without ants (N0122) Hirtella occupied by ants (N0545)
Numbers Percentages Numbers Percentages
No wasps 121 99.18 458 85.32
Polistes pacificus 0 0 1 0.18
Mischocyttarus adolphi 0 0 19 3.49
Mischocyttarus l. lecointei 1 0.82 51 9.36
Mischocyttarus oecothrix 0 0 3 0.55
Mischocyttarus punctatus 0 0 4 0.73
Mischocyttarus synoecus 0 0 3 0.55
Angiopolybia pallens 0 0 4 0.73
Polybia affinis 0 0 1 0.18
Polybia occidentalis 0 0 1 0.18
Table 2 Survey of social wasp
species nesting on the foliage
of H. physophora sheltering an
active colony of A. decemarticu-
latus. In each case, the ants were
offered a fresh grasshopper
as prey for each 1-h-long experi-
ment. Only A. pallens robbed
pieces of the prey
No. of trees
with wasp nests
No. of hours
of observation
Robbed prey No. of cases No. of workers
Mischocyttarus adolphi 18 26 No 0 0
Mischocyttarus l. lecointei 19 31 No 0 0
Mischocyttarus oecothrix 310No 00
Mischocyttarus punctatus 417No 00
Mischocyttarus synoecus 28No00
Angiopolybia pallens 4 14 Yes 14/14 29
268 Naturwissenschaften (2012) 99:265273
suffered attempts at capture). So, Allomerus colonies likely
benefit in two ways: from capturing the robbing wasps as the
total weight of the individuals captured can exceed that of a
prey abdomen (Table 4) and because they generally spread-
eagle and dismantle the remainder of the prey of this size which
has not been robbed.
Trigona sp. workers that frequently begin by cutting off parts
of the prey abdomen were more successful as, in this case, the
balance sheet for the Allomerus ants was negative (Table 4).
Because A. pallens was the wasp species most frequently
recorded (see Table 3), we focused on this species by furnishing
pieces of prey of increasing size to Allomerus ants. We noted
that the number of wasp individuals captured while trying to
rob prey parts increased with the size of the pieces of prey. This
relationship is expressed as an exponential curve (Fig. 3).
Wasps sheltered by H. physophora treelets
When sheltering in a myrmecophytes foliage, wasp nests
are protected from rainfall by the host plant leaves, while the
wasps themselves benefit from the protection provided from
Table 3 Insects noted as robbing
pieces of prey (ca. 2.5-cm-long
grasshoppers) captured on
gallery-shaped traps by A. dece-
marticulatus workers. The differ-
ence between the numbers of
cases and the numbers of robbers
involved in the visits is due to the
fact that several individuals from
the same species could be present
at the same time, likely due to
recruitment. This is particularly
true for Trig on a.Thenumberof
cases thus corresponds to visits
by one individual or, for Trigona
or social wasps, one or several
Identity of
the robbers
A B Ratio No. of robbers
% captured
No. of cases No. of robbers
Trigona sp. 73 264 3.61 47 17.8
Angiopolybia pallens 90 157 1.74 59 37.6
Agelaia myrmecophila 23 30 1.30 11 36.7
Agelaia cajennensis 24 37 1.54 11 29.7
Agelaia pallipes pallipes 59 85 1.44 20 23.5
Agelaia fulvofasciata 8 10 1.25 0 0.0
Agelaia testacea 16 16 1.00 0 0.0
Flies 13 13 4 30.8
Reduviidae 6 6 6 100.0
Total 312 618 158
Mean wei
ht ± SD in
y = -1095.3x + 57.64
= 0.9561
Heteroptera (c)
Flies (a, b, d)
Agelaia testacea (e)
Agelaia fulvofasciata (e)
Trigona sp. (a)
Agelaia cajennensis (a, b, d)
Agelaia pallipes pallipes (a, b)
Agelaia myrmecophila (a, b, d)
Angiopolybia pallens (b)
0 0.02 0.04 0.06 0.08 0.1 0.12
Fig. 1 Percentages of Trigona, social wasp workers, reduviids, and flies
captured while robbing pieces of prey from ambushing A. decemarticu-
latus workers and their mean weights. The relationship between these two
values was calculated only for the social wasps (the largest species were
never captured) and constituted a kind of demarcation line permitting us
to note that, comparatively, Tr ig on a are less frequently captured than
wasps. The contrary is true for Reduviids (Heteroptera) that were cap-
tured in all cases. Statistical comparisons of the rate of capture for the
different species: Fishers exact-test and sequential Bonferroni procedure
for multiple comparisons; different letters (between parentheses) indicate
significant differences
Naturwissenschaften (2012) 99:265273 269
army ants by the associated plant-ants (Chadab 1979;
Chadab-Crepet and Rettenmeyer 1982, Herre et al. 1986,
ODonnell and Jeanne 1990; Dejean et al. 2001). Arboreal
ants can divert army ant columns from the base of their host
trees by attacking them (Chadab-Crepet and Rettenmeyer
1982) or by depositing repellent compounds as is the case
for Allomerus octoarticulatus and Pheidole spp. associated
with melastome myrmecophytes (Herre et al. 1986). Wasps
such as A. pallens are protected from the ants sheltering on
the same tree by their nest envelope (Fig. S1in Electronic
supplementary material), while the polistini (Polistes and
Mischocyttarus) rely on compounds deposited on the petiole
(Jeanne 1975,1991;Wenzel1991; London and Jeanne
2000; Smith et al. 2001; see also Fig. S1d in Electronic
supplementary material).
As concerns the behavior of the wasps nesting on
Hirtella treelets towards prey captured on the trap,
Mischocyttarus was never observed robbing (or trying to
rob) prey from Allomerus.Mischocyttarus colonies are very
sparsely populated, which renders cleptoparasitism towards
the highly skilled predatory Allomerus ants very risky not
only for individuals but also in terms of colony survival.
This prudent behavior is probably adaptive and, combined
with the chemical protection of their nest, might help to
explain the successful association of Mischocyttarus with
Allomerus and Hirtella.Onthecontrary,A. pallens rob
pieces of trapped prey from their associated Allomerus col-
onies. They may have learned how to avoid these ants
during repetitive encounters on the host plant (see Jeanne
1991). Moreover, A. pallens colonies, being much more
populated (hundreds of individuals) than Mischocyttarus
colonies, can accept the loss of a few workers without risk
for the colonys survival. Another explanation can be found
in the food selected by these ants as A. pallens frequently
feed on carrion, while Mischocyttarus do not and so were
not attracted by the baits (AD, personal observation).
This illustrates a new kind of cleptobiosis as Allomerus
ants apparently lose out in this association (but see below),
Fig. 2 After a series of return trips, an Agelaia pallipes pallipes wasp
is finishing cutting up an acridid grasshopper captured by ambushing
A. decemarticulatus workers (it has cut off the preys wings to have
better access to the thorax). The acridid was initially caught on the trap,
lowered down on the host plant stem, and then slowly transported by
the ants to the nearest leaf pouch (the green circular plant part in the
lower left corner of the picture). As the ants pull backwards, the prey is
pulled into the trap. The gallery-shaped trap is riddled with holes
beneath which the workers hide, their mandibles wide open, when
ambushing prey. The structure of the trap is a composite material
formed by cut trichomes that the workers assemble into a frame which
they then reinforce with a fungal mycelium that they manipulate. This
system corresponds to a tripartite association as the ants supply the fungus
with pieces of prey, and in turn the fungus provides the host plant
with nutrients (Dejean et al. 2005; Leroy et al. 2011;Ruiz-González
et al. 2011)
Table 4 Parameters for calculating the estimated losses and benefits for the A. decemarticulatus colonies whose workers had caught ca. 2.5-cm-long
grasshoppers which Trig on a sp. or social wasp workers attempted to rob (they always began by cutting off pieces of the prey abdomen)
Identity of
the robbers
Mean weight± SD
of the workers (in
Mean weight of
transported pieces
of prey (in milligrams)
Ratio prey
Probability of being
captured, 1/no. of
trips× 100 (%)
No. of individuals
possibly captured
(ratio C/D)
Net gain
relative to prey
abdomen (mg)
Trigona sp. 16.1±0.3 4.0 21.50 17.8 3.82 24.5
Angiopolybia pallens 15.8± 2.3 3.1 27.74 37.4 10.43 78.8
Agelaia myrmecophila 18.2± 0.9 4.0 21.50 36.6 7.88 57.4
Agelaia cajennensis 29.7± 1.4 6.3 13.65 329.8 3.76 25.7
Agelaia pallipes pallipes 32.9 ±3.7 7.6 11.31 23.5 2.66 1.5
Agelaia fulvofasciata 50.1± 1.4 11.6 7.41 0 (never) None 86.0
Agelaia testacea 123.5± 2.5 28.6 3.00 0 (never) None 86.0
Cratio between the weight of a prey abdomen (0.086 g) and the weight of transported pieces of prey, Dcalculated from Table 2as the ratio between
the number of robbers involved and the number of robbers captured, Ftheoretical weight of wasps likely to be captured (A × E) minus the weight of
a prey abdomen (0.086 g). This is a conservative estimate as the wasps were captured or stopped making return trips before having completely
retrieved the prey abdomens
Because individuals of these species were never captured by ambushing Allomerus ants, 0.086 gcorresponds to the loss of the entire prey
abdomen, sometimes more
270 Naturwissenschaften (2012) 99:265273
whereas, in general, the antwasp association is either
neutral for the ants (neither benefit nor loss) or beneficial
such as when aggressive Polybia rejecta wasps protect their
associated Azteca chartifex ants from ant predators (Herre et
al. 1986).
Insects robbing prey from ambushing A. decemarticulatus
Although plant-ants are well known for their aggressiveness
towards any intruder landing on their host myrmecophyte,
even catching them as prey (Dejean et al. 2005,2009,2010;
Rico-Gray and Oliveira 2007), this study documents the
existence of cleptobiosis involving several insect taxa.
Wasps robbing pieces of prey or food bodies provided by
their host myrmecophytes to plant-ants have previously
been noted (LaPierre et al. 2007). Here we show that this
behavior is frequent in the understory of Neotropical rain-
forests, while, to the best of our knowledge, attempts at
cleptobiosis by reduviids has not yet been reported and
seldom been reported in the case of flies (Wild and Brake
The main result of this study concerns the fact that
cleptobiotic insects can frequently be captured (see
Table 3). An immobilized prey attracts preying insects and
enables the ants to capture these insects in turn. This is
particularly true for reduviids that were always captured.
Flies were sometimes captured and this thus largely com-
pensates for any amount of prey lost to them (four flies were
captured during the study using 2.5-cm-long grasshoppers;
Table 3). On the other hand, for those flies that are not
captured, they are able to lick the haemolymph that leaks
out when the prey is spread-eagle (or if it was already bitten
by cleptobiotic wasps).
Concerning social wasps, we saw that the larger the
individuals, the less likely they are to be captured (Fig. 1).
Notably, the workers of the two larger Agelaia species were
very rarely or never caught. One of the two, A. fulvofasciata,
has already been described as being a particular type of
cleptoparasite, successfully stealing paper from the enve-
lope of the arboricolous nest of the ant Dolichoderus bidens
(Corbara and Dejean 2002). Because workers from the most
common wasp species, A. pallens, were the most frequently
captured (Table 3), Allomerus likely obtain a net gain from
cleptobiosis by wasps. We also noted that the larger the prey,
the more often A. pallens individuals were captured (Fig. 2).
This can be explained by the fact that, when the prey is
large, more ants and more wasps are involved as they recruit
nestmates. Two or three wasps cutting up pieces of prey at
the same time had a tendency to get in each others way so
that one of them placed the extremities of its legs on the trap
and was caught. Because the wasps stopped returning to
gather pieces of a prey captured by the Allomerus workers
after they had successfully escaped an attempt at capture,
they never gathered more than the abdomen of a 2.5-cm-long
grasshopper during our experiment (see also Figs. S4and S5
in Electronic supplementary material). Yet we noted excep-
tions during additional observations (see Fig. S4e in
Electronic supplementary material).
The case of Trigona is different. Although they recruit
nestmates quickly, the number of individuals captured by
the Allomerus ants was relatively low because, in most
cases, when an individual is seized by one or several leg
(s), its nestmates immediately fly away. It is therefore nec-
essary for the Allomerus ants to capture a Trigona individual
as rapidly as possible so as to obtain both the rest of the prey
plus at least one entire Trigona.
The Allomerus ants can obtain a qualitative gain when
they successfully catch a wasp or a Trigona. Indeed, they
profit from the protein-rich alary muscles of either, while
they lose only the preys abdomen (which contains feces) as
wasps always, and Trigona frequently, begin by cutting off
pieces of the preys abdomen.
Because myrmecophytes and plant-ants are not scarce
and constitute a predictable, if not permanent, source of
food, the conditions are right for cleptobiosis. This might
be compared to a more common kind of cleptobiosis which
y = 0.2041e
R² = 0.9514
0.06-0.08g 0.13-0.15g 0.20-0.22g 0.27-0.29g
ht of the pieces of prey
Mean numbers of Angiopolybia pallens
individuals captured
Fig. 3 Relationship between
mean numbers (±) SE of A.
pallens captured while trying to
rob pieces of prey and the size
(represented by the weight) of
the pieces of prey furnished to
ambushing A. decemarticulatus
ants. The data are expressed
as an exponential curve (R
that illustrates the close positive
relationship between these
variables. Statistical
comparisons. One-way
3, 56
03.125; P0
0.033; Tukeyspost-hoc test:
different letters indicate signifi-
cant differences at P<0.05
Naturwissenschaften (2012) 99:265273 271
concerns Argyrodes (Theridiidae) spiders towards larger
spider species (Hénaut et al. 2005 and references cited herein).
In both cases, the trap renders the presence of immobilized
prey more predictable than in any other classical situation of
predation. Indeed cleptoparasitism is more likely to occur in
focal places where prey presence is predictable such as social
nest entrances (this is true of Crematogaster ants robbing food
from Ectatomma tuberculatum; Richard et al. 2004)oron
traps (spider webs and Allomerus galleries). Nevertheless, in
the situation examined here, the robbers were frequently cap-
tured (a quantitative benefit for the ants) so that it is unlikely
that the wasps have specific plants that they trap-linewhile
foraging. Allomerus ants also frequently obtain a qualitative
benefit if the robbers only have time to retrieve the preys
abdomen or just part of it before being captured.
Importance of the recruitment of nestmates in stingless bees
and social wasps
In this study, as typically reported for nectar or carrion
(Roubik 1989; Jarau 2009), Trigona always recruited nest-
mates by laying odor trails to the prey (see also the ratio of
3.61 between the number of individuals robbing prey and
the number of visits; Table 3). We noted up to 12 individuals
robbing parts of the same prey that we had furnished to
ambushing Allomerus workers.
Although they are used during swarm-founding in epi-
ponine wasps, chemical trails to food sources have not yet
been demonstrated (Jeanne et al. 1995; Taylor et al. 2011).
Indeed recruitment to food in social wasps is only known in
a primitive form: in Vespula germanica and Polybia occi-
dentalis, odors from food brought back to the nest stimulate
foraging while permitting new foragers to find the food
source (social facilitation) (Overmyer and Jeanne 1998;
Jandt et al. 2005; Hrncir et al. 2007; Schueller et al. 2010;
Taylor et al. 2011).
We argue that in certain epiponine wasp species a form of
recruitment occurred during cleptobiosis; social facilitation
(when a forager returns to the nest with rich food) could be a
good candidate. First, in Charterginus xanthura, certain individ-
uals prevent Azteca ants from leaving their domatia while nest-
mates freely gather Müllerian bodies from their host
myrmecophytic Cecropia tree (LaPierre et al. 2007). This
well-adapted distribution of tasks is only possible if nestmates
are recruited. Second, in the present study, A. pallens,A. myr-
mecophila,A. cajennensis,A. pallipes pallipes, and, to a lesser
degree, A. fulvofasciata likely recruit nestmates. Indeed the
ratios between the number of individuals robbing prey and the
number of visits was greater than 1 (Table 3;seealsoTable2for
A. pallens), while we also frequently noted two or more
individuals cutting up parts of the same prey without being
aggressive towards each other (see Fig. S2in Electronic
supplementary material)asareTrigona individuals. On the
contrary, P. occidentalis foragers did not treat nestmates
preferentially when gathering food from the same spot
(Raveret-Richter 1990).
In conclusion, we have noted a new kind of cleptobiosis
that concerns stingless bees, social wasps, and flies robbing
pieces of prey from Allomerus plant-ants in the understory
of a Neotropical rainforest. Nevertheless, due to their high
level of specialization in capturing prey, Allomerus workers
can capture cleptobiotic insects in turn so that the balance
sheet between the loss of prey parts and the capture of
cleptobiotic insects can be quantitatively and/or qualitatively
beneficial as they recuperate the protein-rich alary muscles of
these insects.
Acknowledgements We are grateful to Andrea Dejean for proofreading
early versions of the manuscript, to Frédéric Azémar for his technical
help, and to three referees for their very constructive comments. We
would also like to acknowledge the staff of Hydreco field station for
accommodations and technical assistance. Financial support for this
study was provided by the Programme Amazonie II of the French
CNRS (project 2ID).
Ethical standards The experiments comply with the current laws of
the country in which they were performed. The authors declare that
they have no conflicts of interest.
Chadab R (1979) Early warning cues for social wasps attacked by
army ants. Psyche 86:115124
Chadab-Crepet R, Rettenmeyer CW (1982) Comparative behaviour of
social wasps when attacked by army ants or other predators and
parasites. In: Breed MD, Michener CO, Evans HE (eds) The
biology of social insects. Westview, Boulder, pp 270274
Corbara B, Dejean A (2002) Paper stealing on an arboricolous ant nest
by the wasp Agelaia fulvofasciata Degeer (Hymenoptera: Vespidae).
Sociobiology 39:281283
Corbara B, Carpenter JM, Céréghino R, Leponce M, Gibernau M,
Dejean A (2009) Diversity and nest site selection of social wasps
along Guianese forest edges: assessing the influence of arboreal
ants. C R Biologies 332:470479
Dejean A, Orivel J, Corbara B, Olmsted I, Lachaud JP (2001) Nest site
selection by two polistine wasps: the influence of AcaciaPseudo-
myrmex associations against predation by army ants (Hymenoptera).
Sociobiology 37:135146
Dejean A, Solano PJ, Ayroles J, Corbara B, Orivel J (2005) Arboreal
ants build a trap to ambush and capture prey. Nature 434:973
Dejean A, Djiéto-Lordon C, Orivel J (2008) The plant-ant Tetraponera
aethiops (Pseudomyrmecinae) protects its host myrmecophyte
Barteria fistulosa (Passifloraceae) through aggressiveness and
predation. Biol J Linn Soc 93:6369
Dejean A, Grangier J, Leroy C, Orivel J (2009) Predation and aggres-
siveness in host plant protection: a generalization using ants of the
genus Azteca. Naturwissenschaften 96:5763
Dejean A, Leroy C, Corbara B, Roux O, Céréghino R, Orivel J, Boulay
R (2010) Arboreal ants use the Velcro
principleto capture very
large prey. PloS ONE 5:e11331
Heil M, McKey D (2003) Protective antplant interactions as model
systems in ecological and evolutionary research. Ann Rev Ecol
Syst Evol 34:425553
272 Naturwissenschaften (2012) 99:265273
Hénaut Y, Delme J, Legal L, Williams T (2005) Host selection by a
kleptobiotic spider. Naturwissenschaften 92:9599
Herre EA, Windsor DM, Foster RB (1986) Nesting associations of wasps
and ants on lowland Peruvian ant-plants. Psyche 93:321330
Hrncir M, Mateus S, Nascimiento FS (2007) Exploitation of carbohydrate
food sources in Polybia occidentalis: social cues influence foraging
decisions in swarm-founding wasps. Behav Ecol Sociobiol 61:975
Jandt J, Riel L, Crain B, Jeanne RL (2005) Vespula germanica foragers
do not scent-mark food sites. J Insect Behav 18:1931
Jarau S (2009) Chemical communication during food exploitation in
stingless bees. In: Jarau S, Herncir M (eds) Food exploitation in
social insects: ecological, behavioral, and theoretical approaches.
CRC, Boca Raton, pp 223249
Jeanne RL (1975) The adaptiveness of social wasp nest architecture.
Quart Rev Biol 50:267287
Jeanne RL (1991) The swarm-founding Polistinae. In: Ross KG, Matthews
RW (eds) The social biology of wasps. Cornell University Press,
Ithaca, pp 191231
Jeanne RL, Hunt JH, Keeping MG (1995) Foraging in social wasps:
Agelaia lacks recruitment to food (Hymenoptera: Vespidae). J
Kansas Entomol Soc 68:279289
LaPierre LM, Hespenheide H, Dejean A (2007) Wasps robbingfood from
ants: a frequent behavior? Naturwissenschaften 94:9971001
Leroy C, Sejalon-Delmas N, Jauneau A, Ruiz-González M-X, Gryta H,
Jargeat P, Corbara B, Dejean A, Orivel J (2011) Trophic media-
tion by a fungus in an antplant mutualism. J Ecol 99:583590
London KB, Jeanne RL (2000) The interaction between mode of
colony founding and nest architecture to ant defense in polistine
wasps. Ethol Ecol Evol 12:1325
ODonnell S, Jeanne RL (1990) Notes on an army ant (Eciton burchelli)
raid on a social wasp colony (Agelaia yepocapa)inCostaRica.J
Trop Ecol 6:507509
Orivel J, Lambs L, Malé P-JG, Leroy C, Grangier J, Otto T, Quilichini
A, Dejean A (2011) Dynamics of the association between a long-
lived understory myrmecophyte and its specific associated ants.
Oecologia 165:369376
Overmyer SL, Jeanne RL (1998) Recruitment to food by the
German yellowjacket, Vespula germanica. Behav Ecol Sociobiol
Raveret-Richter M (1990) Hunting social wasp interactions: influence
of prey size, arrival order, and wasp species. Ecology 71:1018
Revel M, Dejean A, Céréghino R, Roux O (2010) An assassin among
predators: the relationship between plant-ants, their host myrme-
cophytes and the Reduviidae Zelus annulosus.PloSONE5:
Richard FJ, Dejean A, Lachaud JP (2004) Sugary food robbing in ants:
a case of temporal cleptobiosis. C R Biologies 327:509517
Rico-Gray V, Oliveira P (2007) The ecology and evolution of antplant
interactions. The University of Chicago Press, Chicago
Roubik DW (1989) Ecology and natural history of tropical bees. Cam-
bridge University Press, Cambridge
Ruiz-González MX, Malé P-JG, Leroy C, Dejean A, Gryta H, Jargeat
P, Quilichini A, Orivel J (2011) Specific, non-nutritional associa-
tion between an Ascomycete fungus and Allomerus plant-ants.
Biology Letters 7:475479
Schueller TI, Nordheim EV, Taylor BJ, Jeanne RL (2010) The cues
have it; nest-based, cue-mediated recruitment to carbohydrate
resources in a swarm-founding social wasp. Naturwissenschaften
Smith AR, ODonnell S, Jeanne RL (2001) Correlated evolutionary of
colony defence and social structure: a comparative analysis in
eusocial wasps (Hymenoptera: Vespidae). Evol Ecol Res 3:331344
Taylor BJ, Nordheim EV, Schueller TI, Jeanne RL (2011) Recruitment
in swarm-founding wasps: Polybia occidentalis does not actively
scent-mark carbohydrate food sources. Psyche. doi:10.1155/2011/
Wenzel JW (1991) Evolution in nest architecture. In: Ross KG, Matthews
RW (eds) The social biology of wasps. Cornell University Press,
Ithaca, pp 480519
Wild AL, Brake I (2009) Field observations on Milichia patrizii ant-
mugging flies (Diptera: Milichiidae: Milichiinae) in KwaZulu-Natal,
South Africa. Afr Inv 50:205212
Naturwissenschaften (2012) 99:265273 273
    • "Situations where organisms complete parental investment with some type of protective shield are found in a number of taxa (e.g. Stokes and Boersma 1998; Bourgeois and Vidal 2007; Dejean et al. 2012). By contrast, in many other systems, nestling vulnerability changes slowly and continuously with age and between offspring (Caro 2005 ) making interpretation of nest defence difficult because multiple factors may co-vary during that time. "
    [Show abstract] [Hide abstract] ABSTRACT: Parental protection of offspring is found in numerous animal species. Protection provides offspring with a greater chance of surviving to be able to reproduce, while at the same time, often posing a cost to the parent. Therefore, the net value of defence for the parent can vary depending on the developmental stage of the offspring and their ability to defend themselves. For example, in commonly studied organisms (e.g. birds), defence level increases over time until offspring are able to leave the nest and parental defence wanes. We examine these nest defence decisions in the solitary bee system that poses an interesting variation to traditionally studied organisms. Nest value does not increase in a simple additive manner as offspring are added to the nest. Here, individual investments are allocated to each offspring, then as a final act, the nest entrance is sealed. This sealing action provides increased protection for all developing offspring in the nest, therefore strongly increasing their value. Our observational experiment using Megachile rotundata found that mothers slightly increased nest defence as nest size increased. However, unlike traditionally studied organisms, this increase in defence continued until the nest was completed.
    Full-text · Article · Jul 2015
    • "Because the preys are very attractive to predatory social wasps and stingless bees (Dejean et al. 2012), the faster the siblings arrive, the less chance there is that a successful nymph will have its prey stolen or even be killed in turn. Moreover, arriving siblings can capture these cleptobionts (see Table 2) as do Allomerus ants (Dejean et al. 2012). "
    [Show abstract] [Hide abstract] ABSTRACT: Zelus annulosus is an assassin bug species mostly 28 noted on Hirtella physophora, a myrmecophyte specifically associated with the ant Allomerus decemarticulatus known to build traps on host tree twigs to ambush insect prey. The females lay egg clutches protected by a sticky substance. To avoid being trapped, the first three instars of Z. annulosus nymphs remain grouped in a clutch beneath the leaves where they hatched, yet from time to time they climb onto the upper side to group ambush prey. Long-distance prey detection permits these bugs to capture flying or jumping insects that alight on their leaves. Like for some other Zelus species, the sticky substance of the sundew setae of their forelegs aids in prey capture. Group ambushing permits early instars to capture insects that they then share or not depending on prey size and the hunger of the successful nymphs. Fourth and fifth instars, with greater needs, rather ambush solitarily on different host tree leaves, but attract siblings to share large prey. Communal feeding permits faster prey consumption, enabling small nymphs to return sooner to the shelter of their leaves. By improving the regularity of feeding for each nymph, it likely regulates nymphal development, synchronizing molting and subsequently limiting cannibalism.
    Full-text · Article · Oct 2013
    • "This provides the ants with more housing, in turn providing accommodation for the additional winged females produced and enabling the ants to build longer gallery-traps to capture more prey. In addition to this self-sustaining process, the prey carcasses incorporated into the trap to feed the fungus attract necrophagous insects that are then frequently captured [25], while the trapped prey attract cleptobionts that are also captured (Fig. 1a; [34]). Cases of plant-ants sterilizing their host myrmecophyte have frequently been noted202122232427]. "
    [Show abstract] [Hide abstract] ABSTRACT: Mutualisms, or interactions between species that lead to net fitness benefits for each species involved, are stable and ubiquitous in nature mostly due to "byproduct benefits" stemming from the intrinsic traits of one partner that generate an indirect and positive outcome for the other. Here we verify if myrmecotrophy (where plants obtain nutrients from the refuse of their associated ants) can explain the stability of the tripartite association between the myrmecophyte Hirtella physophora, the ant Allomerus decemarticulatus and an Ascomycota fungus. The plant shelters and provides the ants with extrafloral nectar. The ants protect the plant from herbivores and integrate the fungus into the construction of a trap that they use to capture prey; they also provide the fungus and their host plant with nutrients. During a 9-month field study, we over-provisioned experimental ant colonies with insects, enhancing colony fitness (i.e., more winged females were produced). The rate of partial castration of the host plant, previously demonstrated, was not influenced by the experiment. Experimental plants showed higher δ(15)N values (confirming myrmecotrophy), plus enhanced vegetative growth (e.g., more leaves produced increased the possibility of lodging ants in leaf pouches) and fitness (i.e., more fruits produced and more flowers that matured into fruit). This study highlights the importance of myrmecotrophy on host plant fitness and the stability of ant-myrmecophyte mutualisms.
    Full-text · Article · Mar 2013
Show more