The evolution of mutualism, particulary in ant-plant systems

How do individuals from different species end up cooperating with each other? I am interested in the potential routes by which mutualism might evolve and how it is maintained in the face of cheats. I have worked on two ant-plant systems in the process of investigating this: the ants inhabiting epiphytic bird’s nest ferns (Fayle et al. 2012, Fayle et al. 2015; Asplenium phyllitidis and A. nidus) and those inhabiting rattan plants (Edwards et al. 2010; Korthalsia furtadoana).

Bird’s nest ferns are litter-intercepting epiphytes, which are abundant in primary forest, and thus provide patches of cool, damp litter in the canopy, where conditions are usually hot and dry. As a result, individual birds nest ferns often support high abundances of a wide range of arthropods, including, in the largest ferns, colonies of up to 12 different species of ant. Often ants inhabiting epiphytes act mutualistically, defending the plant from herbivores, and sometimes even providing food, in exchange for housing in the form of domatia (ant houses). However, with multiple ant species inhabiting each fern, there is a temptation for individual ant species to be lazy, and let others take up the slack. Indeed, we found that although ferns with more ants suffered less herbivory, no single species of ant actively patrolled the ferns (Fayle et al. 2012). We also looked to see whether there were any potential benefits of helping the ferns for the ant inhabitants. It would be expected that if larger ferns supported larger colonies of particular ant species, then there would be an incentive for those species to help the fern grow larger. However, we found that when ferns grow larger, the extra space is taken by colonies of new ant species, with resident colonies remaining the same size. Therefore the interaction between the ferns and the ants remains at the stage of a two-way by product mutualism: beneficial to both parties, but with no additional investment in the relationship from either side. I have also speculated more broadly on situations in which the evolution of co-operative behaviour may be constrained by asymmetries in power and risk (Fayle 2015). Bird's nest ferns are also found in logged forest and oil palm plantation, where they are still occupied by ants, and maintain their by-product mutualism (Fayle et al. 2015). Despite this, occupancy and worker density are lower in oil palm plantations than in other habitats, probably due to the harsh microclimate, and non-native ant species play an important role in maintaining the mutualism.

However, in many cases, mutualistic interactions in which both partners act altruistically do evolve. Once two species are engaged in a mutualistic interaction there is an incentive for individuals from non-cooperating species to colonise and to cheat, by failing to act altruistically. Often cheats are kept in check by the host by punishing individuals that fail to cooperate. In ant-plant systems this has been observed in the form of failure to grow domatia (ant houses) on shoots which have suffered high levels of herbivory. I worked in collaboration with David Edwards to investigate what happens when a plant is morphologically constrained such that it is unable to punish cheating residents (Edwards et al. 2010). We studied the rattan Korthalsia furtadoana, which forms domatia before new leaves develop, and therefore is unable to tie the growth of domatia to leaf protection. We found that two main species of ant (of the genus Camponotus) inhabited the rattan, one of which protected the plant vigorously, while the other did not. The protecting species only inhabitated approximately half of the rattan plants, a surprisingly low prevalence compared with other ant-plant mutualisms. We also found that plants inhabited by the mutualistic ant species grew faster than those without this species. We suggest that since the rattan is unable to punish the ants that fail to protect it, this allows them to become widespread. There must, however, be some factor limiting the spread of cheats, since the cooperative ant species is still present in the system.


Fayle T.M., Edwards D.P., Foster W.A., Yusah K.M. & Turner E.C. (2015) An ant-plant by-product mutualism is robust to selective logging of rain forest and conversion to oil palm plantation. Oecologia 178: 441–450 [PDF]

Fayle T.M. (2015) Moral hazard in ecology. Frontiers in Ecology and Evolution 3: 3. [PDF]

Fayle T.M., Edwards D.P., Turner, E.C., Dumbrell A.J., Eggleton P. & Foster W.A. (2012). Public goods, public services, and by-product mutualism in an ant-fern symbiosis. Oikos 121: 1279–1286. [PDF]

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. [PDF]