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 δ15N 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.
Citation: Dejean A, Orivel J, Rossi V, Roux O, Lauth J, Malé P-JG, et al. (2013) Predation Success By A Plant-Ant Indirectly Favours The Growth And Fitness Of Its Host Myrmecophyte. PLoS ONE 8(3): e59405. https://doi.org/10.1371/journal.pone.0059405
Academic Editor : Nigel E. Raine, Royal Holloway University of London, United Kingdom
Received: June 19, 2012; Accepted: February 16, 2013; Published: March 14, 2013
Copyright: © 2013 Dejean et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was supported by the French Labex CEBA 2012 (The laboratory of Excellence Centre for Studies of Amazonian Biodiversity), Tripartite associations and Plant Nutrition (Tri-Nutri). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Mutualisms, defined as cooperative interactions between species where each partner derives a fitness benefit, are based on “invested benefits” corresponding to an adaptation by each species to obtain benefits from its partner with the return exceeding the costs of the investment –. Such context-dependent outcomes vary according to the interacting partners over space and time, thus influencing the evolutionary fate of a mutualistic relationship and providing information on how and when mutualisms arise, persist, and vanish .
When hosts transmit symbionts by “vertical transmission” to their offspring the mutualisms are evolutionarily stable; yet, most mutualistic interactions are transmitted “horizontally” as the partners disperse separately. In the latter case, the reproduction of each partner might be subject to a trade-off if 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 the symbionts may evolve traits promoting a reduction in cost, engendering the emergence of “cheaters” obtaining benefits at minimal cost . Over time, they end up by completely sterilizing their hosts ,.
The subsequent instability would turn every mutualist into a parasite if not counterbalanced by specific conditions such as partner selection and fidelity, spatial structure (i.e., limited dispersal), and, above all, retaliation against exploiters ,,–. Because mutualisms are stable and ubiquitous in nature, the intrinsic traits of one partner can generate an indirect and positive outcome for the other, resulting in “byproduct benefits”. The resulting absence of cost (or very low cost) permits an equilibrium between the costs and benefits to be easily established between the partners –.
Ant relationships with myrmecophytes, or plants housing a limited number of so-called ‘plant-ants’ in domatia (i.e., hollow branches or thorns and leaf pouches), are interesting models for studying conflicts and breakdown within mutualisms. Also, myrmecophytes usually provide their guest ants with food, particularly extrafloral nectar (EFN) and/or food bodies (FBs), while, in return, plant-ants protect them from herbivores, competitors and pathogens  through their intrinsic predatory, territorial and cleaning behaviours. So, these protections correspond to byproduct benefits for the host myrmecophytes (see ). The same is true when ants provide their host plants with nutrients (e.g. prey remains and faeces) that accumulate in the domatia. These nutrients are absorbed through the rhizomes, roots, protuberances, or the walls of the domatia. This phenomenon, called myrmecotrophy, has been noted for epiphytes and for some phanerophytes adapted to the nutrient-poor, lateritic soils of tropical rainforests ,.
The size of plant-ant colonies can be limited by the availability of space and food. As the host myrmecophyte grows, the guest colonies have more nesting space thanks to a greater number of domatia and so can grow in turn, while the production of EFNs and FBs increases. Yet, certain plant-ant species are somewhat less dependent on their host myrmecophyte to provision them because they tend hemipterans to obtain honeydew and/or are predators . Mutualisms between myrmecophytes and plant-ants are transmitted horizontally, so that the energy invested by the myrmecophytes in producing flowers and fruits is not allocated to producing greater nesting space for the ant colonies. Reciprocally, winged ant sexuals are not involved in protecting the host plant foliage. This gives rise to a conflict of interest between the partners. By destroying flowers, the ants “sterilize” the plant and trigger the reallocation of host-plant resources from reproduction to vegetative growth. The sterilization is partial; otherwise, the ant species are considered parasites of the mutualism ,,–.
We focused this study on Hirtella physophora (Chrysobalanaceae) that houses colonies of Allomerus decemarticulatus (Myrmicinae) in pouches situated at the base of the leaf lamina (Fig. 1) and provides them with extrafloral nectar. Workers build galleries under the stems of their host-plants that serve as traps to capture insects of up to 1800 times the weight of a worker (Fig. 1; ). To build these galleries, the workers first cut plant trichomes along the stems, clearing a path; then, using uncut trichomes as pillars, they build the vault of the galleries by binding together the cut trichomes with the mycelium of a fungus that they manipulate . This third partner, an Ascomycota from the order Chaetothyriales, therefore serves a structural purpose. Allomerus decemarticulatus also supply their host tree with nutrients via the walls of the domatia and the fungus with wastes, whilst the fungus, in turn, also provides the host plant with nutrients ,. Finally, the workers partially castrate their host plant by cutting and chewing both the sterile and fertile parts of the flower buds ,.
(a) Leaves bear leaf pouches (left arrow) at the base of their laminas. Allomerus decemarticulatus workers capture a green locust thanks to their trap: a gallery made using severed host plant trichomes and the mycelium of an Astomycota fungus that the ants manipulate to create a composite material pierced by numerous holes (from under which the workers ambush prey). A wasp is seen robbing a piece of the locust abdomen; the wasp was also captured in turn as was the red Reduviid (right arrow). (b) At the distal position of the branch, flowers are segregated on racemous inflorescences at different stages of maturation from flower buds to fully open flowers. (c) Development of young, green (i.e. unripe) drupes. (d) A dark purple (i.e. ripe) drupe. Scale bars represent 1 cm.
We hypothesized that when particularly successful in catching prey, mutualistic ant colonies increase their own fitness by producing more sexuals while also providing benefits to their host plant. Indeed, the more prey they capture, the more they provide the host plant with nutrients that then has greater growth (more leaves equals more housing available for guest ants) and fitness (more flowers and fruits produced in spite of the partial castration).
Materials And Methods
This study was conducted according to relevant national and international guidelines. Sample collections necessary to scientific research were authorized by the French Office National des Forêts (ONF) provided that their impact upon the environment is considered negligible (details of the permit in ).
Study Site And Model
This study was conducted between November 2007 and July 2008 on the top of a hill (05° 03.697′ N; 52° 58.620′ W – 05° 03.638′ N; 52° 58.612′ W) situated in the pristine forest near the field research station at Petit Saut, Sinnamary, French Guiana. Hirtella physophora, an understory plant found in pristine Amazonian forests and strictly associated with A. decemarticulatus in the study area, has long-lived leaves that bear a pair of leaf domatia at the base of each lamina; the flowers are segregated on racemous inflorescences and produce dark-purple drupes (Fig. 1). These trees have a much longer lifespan (up to ca. 350 years) than their associated ant colonies (ca. 20 years) .
Over-Provisioning The Ants
We investigated the role of the ants in the nitrogen provisioning of their host plant by providing the colonies of experimental H. physophora with surplus prey twice a week for 9 months (colonies normally capture one to two prey items per week ). Captured at a light trap situated in Petit Saut, the prey – large moths and grasshoppers – were cut into pieces and the thoraxes plus legs (ca. 1g) were then provided to the ants by holding them close to the galleries where the workers immediately seized and dismantled them . The 41 control and 31 experimental H. physophora trees selected for the main experiment were of similar height, trunk diameter at their base and number of leaves at the beginning of the experiment; height: 1.4±0.09 m vs. 1.35±0.08 m (t = 0.80; df = 70; p = 0.42), trunk diameter at the base of the trees: 1.58±0.13 cm vs. 1.72±0.15 cm (t = 0.62; p = 0.50), and number of leaves: 22.7±1.3 vs. 23.8±1.9 (t = 0.50; p = 0.61). Because Hirtella trees are patchily distributed (several individuals within a 3–5 m radius), we randomly allocated trees within each patch with respect to the treatment (giving us the final ratio of 41 control and 31 experimental trees). This keeps the majority of either experimental or control trees from being selected in the same zone since differences in the amount of 15N in the soil between zones can occur with repercussions for the δ15N of the plants.
Twenty-four trees (height: 1–1.8 m) that had lost their associated ant colonies before we began the experiment were taken into consideration for certain comparisons. Over the 9-month study, only some queens began to found colonies on 11 trees, but no workers patrolled the foliage. Other trees from this hill were not selected because they were either too small to produce flowers (<0.9 m), too tall (over 1.9 m) to be easily paired, or were in the process of regenerating after being broken by a branch that fell from the canopy.
Each month we counted the number of leaves, flower buds, flowers and fruits on both control and experimental trees. We were therefore able to deduce the number of new leaves produced by each tree as well as its level of flower and fruit production.
Nitrogen exists in two stable (non-radioactive) forms, 14N and 15N, and the isotopic nitrogen composition of animal tissue reflects the isotopic ratio of food eaten with a 15N enrichment of 3–5‰ at each trophic level . Also, the δ15N in plant tissue reflects the δ15N value of the nitrogen source. Therefore, if ants supply their host myrmecophyte with nutrients, we reasoned that plants whose associated ants were experimentally provided with surplus prey must be richer in 15N than those in the control treatment.
At the end of the experiment, a 4-cm2 piece of a “young, well-developed leaf” (see “stage 3” leaf in  and in Fig. 2) was harvested from each host tree, freeze-dried and then ground into a homogeneous powder using a mixer mill. Stable isotope analyses on these plant samples were conducted at the Scottish Crop Research Institute, Invergowrie, Dundee, DD2 5DA, Scotland, UK, using a Thermo-Finnigan Deltaplus Advantage gas isotope-ratio mass spectrometer interfaced with a Costech Analytical ECS4010 elemental analyzer. The natural abundances of 15N were calculated as follows: where X is the element of interest and Rsample and Rstandard the molar ratios (i.e., 15N/14N) of the sample and the standard, respectively .
(a) Juvenile leaves: less than 5 cm long, positioned vertically; the domatia are not fully developed. (b) Expanding leaves: have fully developed domatia, but the blade is still immature. (c) Young leaves: from 15 to 25 cm long; positioned horizontally, mature but still relatively tender, and light green in color. (d) Old leaves: older, non-senescent, mature leaves of up to 30 cm in length, positioned horizontally, stiff, and dark green in color. Scale bars represent 5 cm.
We conducted this study because we needed to be sure that the δ15N of the experimental trees had truly increased compared to the control trees.
Biotic Protection Of Trees
Foliar growth is slow in H. physophora, and leaves can live for several years. While juvenile leaves benefit from the intense biotic protection provided by mutualistic ants, old leaves have their own efficacious mechanical and chemical defences . So, we compared the defoliation rates of the two other kinds of leaves (“stage 2” and “stage 3”, respectively, according to Grangier et al. ; see Fig. 2) that were likely produced during the survey period: “expanding leaves” (“stage 2”) had fully developed domatia and an immature blade 8-to-15-cm-long (to be distinguished from “juvenile leaves” whose domatia were not fully developed), and “young, well-developed leaves” (“stage 3”).
At the end of the survey 28 control trees, 25 experimental trees (among the 41 and 31, respectively, from the beginning) and the 24 trees that had naturally lost their ants before the survey period bore such leaves, enabling us to compare the levels of defoliation using the following scale: (1) intact leaves or less than 5% of the leaf surface destroyed; (2) between 5% and 25% of the leaf surface destroyed; (3) between 25% and 50% of the leaf surface destroyed; (4) between 50% and 75% of the leaf surface destroyed; and (5) more than 75% of the leaf surface destroyed. We assigned the values 0, 1, 2, 3 and 4 for these levels of defoliation and we averaged these numbers (∑defoliation values × number of leaves−1) to obtain a rate of herbivory per tree.
Ant Castes Produced In Control And Experimental Colonies
To preserve some H. physophora plants in the area at the end of the survey, we gathered colonies from only 23 control and 23 experimental trees so that some intact trees and colonies remained to re-colonize the area. Indeed, to gather the ant colonies, we were obliged to cut off most of the host tree branches to collect both the leaf pouches (so the entire leaf) and the galleries where numerous workers were hiding and quickly placed them into plastic bags containing 0.5 L of 96% ethanol. We placed a label identifying the tree inside each bag, closed the bag, and then tagged the outside of the bag with the same code. The bags were then transported to the laboratory to quantify the size of the ant colony on each H. physophora tree.
All of the colonies had a queen. In the laboratory, we counted the number of males, male pupae, winged females, female pupae and larvae. For the production of workers, we could not consider the adult individuals as some of them likely emerged before the experimental period, so that we estimated the number of workers produced only from the number of pupae and larvae. We evaluated the number of worker larvae using the following formula: where Total Number of pupae = male pupae + female pupae + worker pupae.
The Student's t-test was used each time a comparison of two sets of data was necessary. The levels of defoliation between treatments were compared using the Kruskal-Wallis test followed by a Dunn's post-hoc test. The link between the treatment and the numbers of leaves, buds or fruits on each plant was modelled using a generalized linear model (GLM) with a log-Poisson link . The significance of the effect of the treatment was assessed through Likelihood ratio tests. Using the same statistics, the link between thetreatment and the number of buds that developed into flowers was modelled for each plant to obtain the transition rates from bud to flower; the same was done for the number of flowers that developed into fruits (R v. 2.14.2 software).
The numbers of workers and males produced by the experimental and control colonies during the survey period were not significantly different (means ± s.e.; workers: 167.60±13.24 vs. 174.40±19.36, t = 0.289, df = 44, p = 0.77; males: 5.09±1.17 vs. 4.61±1.12, t = 0.294, df = 44, p = 0.77). However, over-provisioned colonies produced significantly more winged females than control colonies (3.17±0.97 vs. 1.00±0.34, t = 2.24, df = 44, p = 0.03).
The very similar rates of defoliation for experimental (median, 25% and 75% percentiles: 1.00, 0.50, 1.33) and control trees (1.00, 0.62, 1.00) were much lower than for trees having lost their guest ant colony (3.33; 3.00; 3.41, Kruskal-Wallis test: H2,77 = 47.22, p<0.0001; Dunn's post-hoc test: rates of defoliation for experimental vs. control trees, not significant; experimental and control trees vs. trees having lost their guest ants, p<0.001). Among the 24 trees that had naturally lost their ants before the survey, six produced inflorescences and a total of 26 flower buds. Defoliating insects destroyed 21 of these buds as well as the five flowers produced.
Both plant growth and reproductive investment were enhanced by over-provisioning the ant colonies as the experimental trees produced significantly more leaves (3.58±0.40 vs. 2.17±0.31, p<0.001), more flower buds (5.81±1.45 vs. 3.27±0.64, p<0.0001), more flowers (3.61±0.83 vs. 1.78±0.39, p<0.0001) and more fruits (1.90±0.50 vs. 0.49±0.19, p<0.0001) than control trees (Likelihood ratio test on the GLM-Poisson model; R statistics). The percentage at which flower buds matured into flowers was not significantly different between experimental and control trees (64.3±5.5 % vs. 49.9±7.3 %, p = 0.48 for the Likelihood ratio test on the GLM-Binomial model). So, it is likely that the workers castrated their host trees regardless of whether the trees belonged to the experimental or the control group. Finally, the percentage of flowers that matured into fruit was significantly higher for experimental than for control trees (49.7±5.4 % vs. 23.3±6.0 %, p = 0.02 for the Likelihood ratio test on the GLM-Binomial model).
At the end of the survey, while the percentage of nitrogen contained in the leaves of both experimental and control trees was not significantly different (mean ± SE: 1.46±0.03 % vs. 1.42±0.02 %; t = 0.95; df = 70; p = 0.35), the experimental trees had significantly higher δ15N values than control trees (2.45±0.18 ‰ vs. 1.59±0.11 ‰; Welch-corrected t = 4.08, df = 51, p<0.001; see Fig. 3).
Here we show that colony fitness was enhanced through over-provisioning with prey as more winged females were produced by experimental colonies. Supplemental protein increased the production of males and gynes, but not of workers, in the Argentine ant Linepithema humile , while the increased production of winged females was due to supplementary carbohydrates rather than proteins in Myrmica brevispinosa .
The experiment did not modify the worker patrolling activity because we noted non-significant differences in the: (1) number of workers produced per colony sheltered, (2) biotic protection (similar defoliation rate; having very low constitutive defences, the leaves at the chosen stages of development depend on ant protection ) and (3) castration rate between control and experimental trees.
On the plant side, experimental individuals had more 15N than did control plants, confirming findings by Leroy et al.  of myrmecotrophy in this mutualism. The novelty of this study is that “surplus” nutrients likely permitted the experimental trees to increase their investment in both growth and reproduction as they produced significantly more leaves, flower buds, flowers and fruits than control trees. Furthermore, the experimental trees had a higher rate of flowers that matured into fruits compared to the control trees, showing that fruiting is likely limited by the quantity of available nitrogen.
Our results support the initial hypothesis that the more insects the ants capture, the more nutrients are available to the host plant for growth. 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 , while the trapped prey attract cleptobionts that are also captured (Fig. 1a; ).
Cases of plant-ants sterilizing their host myrmecophyte have frequently been noted –,. This behaviour corresponds to the “over-exploitation” of the partner, increasing that partner's cost, as opposed to “cheating by defection”, or reducing the partner's benefit ,. In response, myrmecophytes can keep their guest ants from destroying flowers by changing the distribution of the inflorescences and through repellent floral volatiles , or through retaliation where they destroy certain domatia or reduce their size and survival rate ,. For H. physophora, a trade off occurs between the partial castration carried out by the A. decemarticulatus workers , and myrmecotrophy that indirectly favours both plant growth and fitness (this study). Although entirely dependent only on the ants' behaviour, these mechanisms seem to increase the stability of the relationship.
In conclusion, myrmecotrophy, a byproduct benefit for myrmecophytes, may intervene in the stability of myrmecophyte-ant mutualisms because the plant receives a benefit at no cost (or extremely reduced cost) for the ant. Further studies are needed to verify whether myrmecotrophy is more widely found than previously thought, particularly when the plant-ants are predatory and have a tendency to partially castrate their host plant.
Conceived and designed the experiments: AD. Performed the experiments: AD OR. Analyzed the data: AD VR CL. Contributed reagents/materials/analysis tools: AD CL JO JL P-JM RC. Wrote the paper: AD.
- 1. Herre EA, Knowlton N, Mueller UG, Rehner SA (1999) The evolution of mutualisms: exploring the paths between conflict and cooperation. TREE 14: 49–53.
- 2. Trivers RL (1971) Evolution of reciprocal altruism. Quart Rev Biol 46: 35–57.
- 3. Axelrod R, Hamilton WD (1981) The evolution of cooperation. Science 211: 1390–1396.
- 4. Holland JN, Ness JH, Boyle A, Bronstein JL (2005) Mutualisms as consumer-resource interactions. In Barbosa P, ed. Ecology of predator-prey interactions. Oxford: Oxford University Press. Pp 17–33.
- 5. Sachs JL, Simms EL (2006) Pathways to mutualism breakdown. TREE 21: 585–592.
- 6. Leimar O, Hammerstein P (2010) Cooperation for direct fitness benefits. Phil Trans R Soc B 365: 2619–2626.
- 7. Wilkinson DM, Sherratt TN (2001) Horizontally acquired mutualisms, an unsolved problem in ecology. Oikos 92: 377–384.
- 8. O'Keefe KJ, Antonovics J (2002) Playing by different rules: the evolution of virulence in sterilizing pathogens. Am Nat 159: 597–605.
- 9. West SA, Kiers ET, Simms EL, Denison RF (2002) Sanctions and mutualism stability: why do rhizobia fix Nitrogen? Proc R Soc B 269: 685–694.
- 10. Bronstein JL, Wilson WG, Morris WE (2003) Ecological dynamics of mutualist/antagonist communities. Am Nat 162: 24–39.
- 11. Kiers ET, Rousseau RA, West SA, Denison RF (2003) Host sanctions and the legume-rhizobia mutualism. Nature 425: 78–81.
- 12. Sachs JL, Simms EL (2006) Pathways to mutualism breakdown. TREE 21: 585–592.
- 13. Szilágyi A, Scheuring I, Edwards DP, Orivel J, Yu DW (2009) The evolution of intermediate castration virulence and ant coexistence in a spatially structured environment. Ecol Let 12: 1–11.
- 14. Connor RC (1995) The benefits of mutualism – a conceptual framework. Biol Rev Camb Philos Soc 70: 427–457.
- 15. Sachs JL, Mueller UG, Wilcox TP, Bull JJ (2004) The evolution of cooperation. Quart Rev Biol 79: 135–160.
- 16. Leimar O, Connor RC (2003) By-product benefits, reciprocity, and pseudoreciprocity in mutualism. In Hammerstein P, ed. Genetic and cultural evolution of cooperation. Cambridge: The MIT Press. Pp 203–222.
- 17. Rico-Gray V, Oliveira P (2007) The ecology and evolution of ant-plant interactions. Chicago IL: The University of Chicago press. 331p.
- 18. Thornham DG, Smith JM, Grafe TU, Federle W (2012) Setting the trap: cleaning behaviour of Camponotus schmitzi ants increases long-term capture efficiency of their pitcher plant host, Nepenthes bicalcarata. Funct Ecol 26: 11–19.
- 19. Leroy C, Séjalon-Delmas N, Jauneau A, Ruiz-González M-X, Gryta H, et al. (2011) Trophic mediation by a fungus in an ant-plant mutualism. J Ecol 99: 583–590.
- 20. Young TP, Stubblefield CH, Isbell LA (1997) Ants on swollen-thorn acacias: species coexistence in a simple system. Oecologia 109: 98–107.
- 21. Yu DW, Pierce NE (1998) A castration parasite of an ant-plant mutualism. Proc R Soc B 265: 375–382.
- 22. Gaume L, Zacharias M, Borges RM (2005) Ant-plant conflicts and a novel case of castration parasitism in a myrmecophyte. Evol Ecol Res 7: 435–452.
- 23. Frederickson ME (2009) Conflict over reproduction in an ant-plant symbiosis: why Allomerus octoarticulatus ants sterilize Cordia nodosa trees. Am Nat 173: 675–681.
- 24. Malé P-JG, Leroy C, Dejean A, Quilichini A, Orivel J (2012) An ant symbiont directly and indirectly limits its host plant's reproductive success. Evol Ecol 26: 55–63.
- 25. Dejean A, Solano P-JG, Ayroles J, Corbara B, Orivel J (2005) Arboreal ants build traps to capture prey. Nature 434: 973.
- 26. Ruiz-González M-X, Malé P-JG, Leroy C, Dejean A, Gryta H, et al. (2011) Specific, non-nutritional association between an Ascomycete fungus and Allomerus plant-ants. Biol Lett 7: 475–479.
- 27. Orivel J, Lambs L, Malé P-JG, Leroy C, Grangier J, et al. (2011) Dynamics of the association between a long-lived understory myrmecophyte and its specific associated ants. Oecologia 165: 369–376.
- 28. Dejean A, Delabie JHC, Corbara B, Azémar F, Groc S, et al.. (2012) The ecology and feeding habits of the arboreal trap-jawed ant Daceton armigerum. PLoS ONE e37683.
- 29. DeNiro MJ, Epstein S (1981) Influence of diet on the distribution of nitrogen isotopes in animals. Geochem Cosmochem Acta 42: 495–506.
- 30. Grangier J, Dejean A, Malé P-JG, Orivel J (2008) Indirect defense in a highly specific ant-plant mutualism. Naturwissenschaften 95: 909–916.
- 31. Agresti A (2002) Categorical data analysis. 2nd edition. New York: J. Wiley and Sons. 710 p.
- 32. Aron S, Keller L, Passera L (2001) Role of resource availability on sex, caste and reproductive allocation ratios in the Argentine ant Linepithema humile. J Anim Ecol 70: 831–839.
- 33. Bono JM, Herbers JM (2003) Proximate and ultimate control of sex ratios in Myrmica brevispinosa colonies. Proc R Soc B 270: 811–817.
- 34. Dejean A, Carpenter JM, Corbara B, Wright P, Roux O, et al. (2012) The hunter becomes the hunted: when cleptobiotic insects are captured by their target ants. Naturwissenschaften 99: 214–226.
- 35. Douglas AE (2008) Conflict, cheats and the persistence of symbioses. New Phytol 177: 849–858.
- 36. Douglas AE (2010) The symbiotic habit. Princeton and Oxford: Princeton University Press. 214p.
- 37. Raine NE, Willmer P, Stone GN (2002) Spatial structuring and floral avoidance behavior prevent ant-pollinator conflict in a Mexican ant-acacia. Ecology 83: 3086–3096.
- 38. Willmer PG, Nuttman CV, Raine NE, Stone GN, Pattrick JG, et al. (2009) Floral volatiles controlling ant behaviour. Funct Ecol 23: 888–900.
- 39. Izzo TJ, Vasconcelos HL (2002) Cheating the cheater: domatia loss minimizes the effects of ant castration in an Amazonian ant-plant. Oecologia 133: 200–205.
- 40. Edwards DP, Hassall M, Sutherland WJ, Yu DW (2006) Selection for protection in an ant-plant mutualism: host sanctions, host modularity, and the principal-agent game. Proc R Soc B 273: 595–602.