Finding the middle road: Flowers evolve to work with multiple pollinators

“I’ve had a lot of jobs in my life: boxer, mascot, astronaut, baby proofer, imitation Krusty, truck driver, hippie, plow driver, food critic, conceptual artist, grease salesman, carny, mayor, grifter, body guard for the mayor, country western manager, garbage commissioner, mountain climber, farmer, inventor, Smithers, Poochie, celebrity assistant, power plant worker, fortune cookie writer, beer baron, Kwik-E-Mart clerk, homophobe, and missionary, but protecting people, that gives me the best feeling of all.”
—Homer Simpson

In twenty-two seasons of The Simpsons, the eponymous family’s bumbling father Homer has tried his hand at dozens of different jobs, and failed hilariously at most of them. Homer is a one-man illustration of “Jack of all trades, master of none,” the idea that it’s hard to do many different things well. This principle applies more broadly than the curriculum vitae; in biology, it means that living things face trade-offs between different ways of making a living.

A wild radish (Raphanus raphaistrum) flower. Photo by Valter Jacinto.

For instance, a plant whose pollen is carried from flower to flower by just one pollinating animal only needs to match that one pollinator very well. But most plants’ flowers are visited by many different potential pollinators, and matching all of them probably means finding a middle ground among the best ways to match each individual pollinator. A study of one such “generalist” flower, the wild radish, has found exactly this: working with multiple partners takes evolutionary compromise [$a].

Wild radishes are visited by a wide variety of different insects, including honeybees, bumblebees, syrphid flies, and cabbage butterflies, among others. Each of these pollinators comes to a radish flower with a slightly different agenda. Butterflies are there for nectar, but bees like to eat pollen as well—and bumblebees will sometimes bite into the base of a flower and “steal” nectar without ever coming into contact with pollen. Figuring out how natural selection from each of these different pollinators adds up required some clever experimental design.

The study’s authors arrayed potted radish flowers inside a big mesh flight cage, and then introduced either bumblebees, honeybees, cabbage butterflies, or all three pollinators to visit the plants and circulate pollen from flower to flower. They measured the plants’ flowers before putting them in the flight cage, then let the pollinators do their thing. Afterward, the authors collected seeds resulting from the pollinators’ activity, grew them up, and measured the offspring to see whether their traits differed. The procedure was essentially one generation of experimental evolution.

A cabbage white butterfly (Pieris rapae), one of many pollinator species exerting natural selection on wild radishes. Photo by ComputerHotline.

By taking DNA fingerprints of both the parents and the offspring, the authors could also estimate the relationship between each parental plant’s floral measurements and the number of offspring it produced, either from its own seeds or by pollinating another plant.

The results are complex. Depending on the floral measurement under consideration, different pollinators selected in different directions, or the same direction, or not at all. One particularly interesting result, though, was in the effects each pollinator had on the “dimorphism” of the radish flowers’ stamens—the difference between the length of the shortest, and longest, of the male parts of the flower. Flowers only visited by honeybees evolved less dimorphic stamens, while flowers visited by either bumblebees or cabbage butterflies evolved more dimorphic stamens. Flowers in the treatment visited by all three pollinators, however, evolved to find a happy medium, an evolutionary compromise to work with the different partners.

The way these interactions played out in a flight cage probably don’t reflect exactly how they operate in the wild, but this is a pretty cool result all the same. I’ve written in the past about how incorporating multiple interactions can alter the way coevolution works. Gerbils under attack by fleas are less careful about watching for predators; but for the protists living inside pitcher plants, competitors can help distract predators. Here we have an example of multiple similar interactions pulling a generalized plant in different evolutionary directions.


Sahli, H., & Conner, J. (2011). Testing for conflicting and non-additive selection: Floral adaptation to multiple pollinators through male and female fitness. Evolution DOI: 10.1111/j.1558-5646.2011.01229.x

Gardening ants grow their own treetop nests

This post was chosen as an Editor's Selection for ResearchBlogging.orgIf you were to combine ants’ dispersal of seeds and plant protection interactions, and maybe squint a little, you might see something like epiphitic ant gardens. Ant gardens form when tree-nesting ants collect the seeds of some epiphytes—plants evolved to grow in the branches of trees—and the collected seeds sprout. The nests provide congenial conditions for the plants, since gardening ants frequently use dung as a building material. The roots running through the nest help stabilize its structure and suck out moisture to control interior conditions.

Ants cultivate “gardens” of epiphytes like Anthurium gracile to provide nesting space. Photo by gjofili.

This adds up to a mutually beneficial relationship between ant and epiphyte [$a]. A number of tropical epiphytes grow almost exclusively in ant gardens, and the inclusion of plants in the structure of their nests apparently helps gardening ant species to establish nests wherever food is most abundant.

Association with ant gardens has evolved independently in a number of epiphytic species, from arums like Anthurium gracile (pictured to the right) to orchids and philodendrons. When distantly-related species begin to perform the same ecological role, they often evolve convergent traits that facilitate the common role. Almost all ant-dispersed plants attach fatty bodies called elaisomes to their seeds to reward the ants that pick them up. Almost all ant-protected plants grow domatia in which the ants can nest, and nectaries to reward them with sugary sap. But plants that grow in ant gardens don’t seem to have a common trait that prompts ants to collect their seeds. Can it be that every ant-garden plant species has a unique way to be an ant-garden plant?

That’s what studies of ant-garden plants, including a new one just published in PLoS ONE, suggest. Plants associated with ant-gardens don’t have elaisomes on their seeds. Many produce fleshy fruit, but ants will collect their seeds even if no shred of fruit flesh clings to them. In some cases, ants will even collect seeds from the dung of fruit-eating birds and mammals.

This leaves the possibility that ant-garden plants produce some ant-attracting chemical in their seeds. In the new paper, Elsa Youngsteadt and her coauthors set out to identify chemical compounds that might be the common attractant used by nine different ant-garden plants from seven different plant families. Youngsteadt et al. isolated seven different compounds found in the seeds of ant-garden plants but not in closely related species that do not grow in ant gardens. (The absence of the seven compounds from the non-ant-garden relatives is established, rather amusingly, with a blank data table.)

The authors then painted crude extractions of all soluble organic compounds from two ant-garden plants onto seeds from species that gardening ants do not cultivate, and found that the ants were indeed more likely to collect them. (As a control, the ants were also offered seeds coated in the pure solvents used to extract attractive compounds. They didn’t like those.) However, analysis of the extracts failed to find a compound or set of compounds present in all three species.

It’s possible that Youngstead et al. simply failed to isolate the compound or compounds that all three ant-garden plants use to prompt ants to collect their seeds. But it’s not that far-fetched to think that these distantly-related plants might each use different attractive compounds to interact with ants in the same way. Natural selection may often arrive at different solutions when shaping different species for the same ecological role. It might also be that ant-garden relationships were established not by plants evolving a way to prompt ants to pick up their seeds, but by ants evolving to recognize seeds of plants that work well in gardens.


Davidson, D. (1988). Ecological studies of neotropical ant gardens. Ecology, 69 (4), 1138-52 DOI: 10.2307/1941268

Youngsteadt, E., Guerra Bustios, P., & Schal, C. (2010). Divergent chemical cues elicit seed collecting by ants in an obligate multi-species mutualism in lowland Amazonia. PLoS ONE, 5 (12) DOI: 10.1371/journal.pone.0015822

Mutualist matchmaking made simple

This post was chosen as an Editor's Selection for ResearchBlogging.orgBack in September, I wrote about a new economic model of mutualism that proposed mutualists could keep their partner species from cheating—exploiting the benefits of a mutualistic relationship without returning the favor—without explicitly punishing them, so long as failure to play nice led to a reduction in mutualistic benefit [$a]. Now the same research group has published an elaboration of the economic approach to mutualism in the January issue of The American Naturalist, which suggests that mutualists can recruit better partners by manipulating the cost of entering into partnership [$a].

The bobtail squid, whose mutualism with luminescent bacteria is an example for the new model. Photo by megpi.

As a concrete example for their model, the authors refer to the mutualism between bobtail squid and a species of bioluminescent bacteria, which colonize the squid’s light organ and makes it glow. Short of some kind of complicated squid-bacterium signaling system, how does a squid ensure that its light organ is only colonized by bacterial strains that will pay it back and generate light?

They charge a cover.

More specifically, while the squid’s light organ supplies food for colonizing bacteria, it is also full of toxic reactive oxygen compounds. In order to take advantage of the food supply, a bacterium has to clear out these toxins—and conveniently, the bacterial enzyme that generates light consumes oxygen and removes the toxins in the course of the light reaction. So the only bacterial strains that colonize the squid’s light organ are those that can pay the cost of eating up the oxygen-based toxins and still make a “profit” on the food supply provided by the squid, generating light in the process.

The trick in setting up this screening is to find the right balance of cover charge and reward for prospective mutualists. The cover charge paid by high-quality partners has to be high enough that low-quality partners won’t accept it, and the reward offered for paying that high cost must be sufficiently good to make it worthwhile.

The authors suggest that this model should also apply not just to other mutualisms in which a host takes on microbial partners, such as plants’ partnerships with nitrogen-fixing bacteria, or animals’ interactions with the bacteria living in their guts—but also to interactions like obligate pollination mutualism or ants’ protection interactions with some plants.

Yuccas and ant-protected plants have to screen mutualists, too—and may impose their own cover charges. Photos by jby and Alistair Rae.

In the case of obligate pollination mutualism, like the one between yuccas and yucca moths, the cover charge is the effort involved in pollination—to guarantee a supply of yucca seeds for their larvae to eat, yucca moths must deliver plenty of pollen and do relatively little damage to the flower as they lay their eggs in it. There do exist yucca moth species who don’t pollinate, but lay their eggs on yucca flowers after they’ve been pollinated and are starting to develop into fruit. The new model would predict that the lower effort of this strategy is reflected in a lower payoff, maybe a lower rate of survival for the eggs of these “bogus” yucca moths.

In the case of ant-protected plants, the cover charge is the effort involved in defending a host plant from other ant colonies that would like to occupy it. As it happens, parts of an ant-plant that are better protected grow to provide better food and shelter for the ants occupying them, which gives a competitive advantage to a colony of effective defenders trying to fight off a colony of less-effective defenders.

Both of these scenarios, and similar ones in other interactions, suggest ways to test for self-screening mechanisms like the one described in this new model. The model suggests that active screening using signaling between interacting species should be rare in nature, and that a simple cost/benefit structure usually underlies the process of establishing associations between partners. I’ll be very interested to see whether new experimental or observational data further supports the self-screening hypothesis.


Archetti, M., Úbeda, F., Fudenberg, D., Green, J., Pierce, N., & Yu, D. (2011). Let the right one In: A microeconomic approach to partner choice in mutualisms. The American Naturalist, 177 (1), 75-85 DOI: 10.1086/657622

Weyl, E., Frederickson, M., Yu, D., & Pierce, N. (2010). Economic contract theory tests models of mutualism. Proc. Nat. Acad. Sci. USA, 107 (36), 15712-6 DOI: 10.1073/pnas.1005294107

What keeps mutualists honest—cake, or death?

This post was chosen as an Editor's Selection for ResearchBlogging.orgSomewhat like cooperation between members of the same species, mutually beneficial interactions between different species should be prone to fall apart when one species evolves a way to cheat the other. Biologists who study mutualism (myself included) have long believed that the solution to cheating is to punish cheaters—but a new model suggests that the benefits gained from playing nice might be enough to deter cheating [PDF].

I knew I had to write about this one when I saw that the authors use their model to propose a new explanation for the dynamics of my own favorite mutualism, between yuccas and yucca moths. (And, yes, it’s also an excuse to reference Eddie Izzard. I’m only human.)

Cake: definitely preferable to death. Photo by 3liz4.

The new analysis by Weyl et al. applies an economic modeling framework to species interactions in which one species provides some benefit to another, and then itself receives a benefit that at least partially derives from the help initially provided. To take one example the authors cite, many ant species colonize acacia plants, which grow structures in which the ants can nest (or domatia), and often produce nectar or other food rewards for the ants. The ant colony defends the plant from insect herbivores, with the consequence that the plant can devote more energy to growth, including new domatia and new leaves to fuel nectar production via photosynthesis.

In many such interactions, it’s been thought that each species can only keep the other from cheating—taking the benefits of the relationship without returning the favor—by actively punishing such behavior. Weyl et al. argue that instead of punishment, cheaters might be deterred if their refusal to play their role results in reduced payback from the other partner.

In the ant-acacia example, ant-tended plants kill off branches that lose a lot of leaves to herbivores, which can happen if the ants cheat by slacking off on their protection duties. But this isn’t punishment as such, say Weyl et al. Plants that aren’t protected by ants also kill off damaged branches, to conserve resources. Instead, because ant domatia tend to be located on the youngest, most herbivore-vulnerable shoots of ant-tended plants, lazy ants harm themselves by allowing herbivores to trigger a response that the plant would make whether or not it hosted ants.

An ant domatium on a “whistling thorn” acacia tree. Photo by Alistair Rae.

It sounds a bit passive-aggressive on the plant’s part, doesn’t it? But let’s look at the example that caught my attention: yuccas and yucca moths. Yucca moths are the sole pollinators of yuccas, and lay their eggs in pollinated yucca flowers; as a pollinated flower develops into a fruit, the eggs hatch, and the new-born larvae eat some of the seeds inside. Moths have good incentive to cheat on yuccas by laying lots of eggs in a single flower or not providing much pollen, but yuccas abort flowers that receive too many moth eggs, or not enough pollen [PDF]. Those of us who study yuccas have tended to interpret this as punishment, since killing off a pollinated flower also kills off any seeds a yucca might have produced via that flower.

However, as Weyl et al. note, yuccas abort flowers in response to damage to the floral ovules [PDF] (the tissue that will become seeds when pollinated), not to the presence of moth eggs per se. Moths generally damage the ovules a bit when laying eggs inside the flowers; but damage without eggs has the same effect. If floral abortion were punishment, say Weyl et al., it would occur as a result of moth eggs alone, not damage to the ovules in general.

In other words, the mutualists analyzed by this new paper are kept honest not by the threat of punishment (death) but the possibility that cheating will result in reduced rewards (less cake). It’s a clever inversion of perspective, and I’ll be very interested to see whether new empirical studies can back it up.


Marr, D., & Pellmyr, O. (2003). Effect of pollinator-inflicted ovule damage on floral abscission in the yucca-yucca moth mutualism: the role of mechanical and chemical factors. Oecologia, 136 (2), 236-43 DOI: 10.1007/s00442-003-1279-3

Pellmyr, O., & Huth, C. (1994). Evolutionary stability of mutualism between yuccas and yucca moths. Nature, 372 (6503), 257-60 DOI: 10.1038/372257a0

Weyl, E., Frederickson, M., Yu, D., & Pierce, N. (2010). Economic contract theory tests models of mutualism. Proc. Nat. Acad. Sci. USA, 107 (36), 15712-6 DOI: 10.1073/pnas.1005294107

Are mutualists monogamists, while antagonists play the field?

ResearchBlogging.orgTwo of the most diverse groups of living things on Earth are flowering plants and the insects that make their living from flowering plants. Biologists have long thought that the almost incessant, intimate interactions between plants and plant-eating insects might be the evolutionary cause of each group’s spectacular diversity. On a smaller scale, this means that we’re interested in the reasons that specific insects and plants interact in the first place—what evolutionary trails leads one insect species to specialize on a single host while others eat pretty much any plant they land on.

A new study of one group of plant-eating insects suggests that the kind of interaction between insects and their host plants also determines how specific those interactions are. Examining a group of moths that, like the yucca moths I study, pollinate their host plant and then eat some of its seeds, the authors of the new study find that related, non-pollinating moths use more host plant species than the pollinators [$a]. I think it makes a particularly nice companion piece to my post about the evolutionary origins of yucca moths, because it provides an example of one or two other things biologists can deduce from phylogenies—and, as we’ll see, some things they can’t.

Epicephala: like a yucca moth without the snappy name

The moths in question are in the genus Epicephala, and they have an obligate pollination relationship with trees in the genus Glochidion, a diverse group of plant species found in southern Asia. That is, female moths carry pollen between Glochidion flowers in special mouthparts, deliberately apply pollen to the flower, and then lay eggs in the flower so that, when it develops into a fruit, her larvae can eat some of the seeds inside. Epicephala species are highly specialized, with most species only using one species of Glochidion [$a]. That’s a higher degree of specialization than what’s seen in yucca moths, in fact.

Pollinating moths (genus Ephicephala, left) use fewer host plant species than related non-pollinating moths (genus Caloptilia, right). Photos by CharlesLam and Bettaman.

The family of which Epicephala is a member happens to include other moths that interact with Glochidion, but only as herbivores: species in the genera Caloptilia and Diphtheroptila, whose larvae all eat Glochidion leaves. Do these antagonistic moths use more, or fewer, species of the host plant than the mutualistic Epicephala? Kawakita and his coauthors set out to answer that question by reconstructing the phylogenies of Caloptilia and Diphtheroptila.

Finding species in evolutionary trees

Most biologists agree that two groups of organisms are separate species if there is no gene flow between them. A consequence of genetic isolation between species is that, if they’re isolated long enough, they become monophyletic within phylogenies. That is, all the individuals within each species share a common ancestor that is not shared with any other species. You can see this by contrasting two monophyletic species (on the left in the figure below) with two groups that turn out to be paraphyletic—some individuals of the red species are more closely related to individuals of the blue species than to other individuals of their own species.

Monophyletic and paraphyletic groupings. Image by jby.

The reasoning behind this is a bit subtle. Paraphyletic groups might still be separate species—they just haven’t been isolated long enough to become monophyletic. As a good example, I’m a coauthor on a recent study that did this kind of analysis on non-pollinating “bogus” yucca moths that use three different yucca species. In that case, the moths were paraphyletic with respect to which yucca species they used, but more analysis showed that there is currently very little gene flow between moths using different hosts [PDF].

In the case of the Glochidion-using Caloptilia and Diphtheroptila, Kawakita et al. found something more complicated. Each genus broke up into several monophyletic groupings, or clades of genetically similar individuals—but in most cases each clade included moths collected from at least two different Glochidion species. Kawakita et al. note that the clades also correspond to differences in the moths’ wing coloration, larval feeding behavior, and genitalia, and conclude that each clade is a different species. That would mean that the two antagonist genera tend to use multiple host plants.

Interesting question, but is this the way to answer it?

Except I’m not sure I buy this usage of phylogenies to define species. Kawakita et al. have shown that within the clades they call species, the individuals all have very similar genetics, but only for the two commonly-used genetic markers from which the phylogenies are reconstructed. It’s not impossible that within each clade the moths might be adapted to individual host plant species, and reproductively isolated by that adaptation—and this could have happened recently enough that not many genetic differences would have built up in the two markers.

To really answer the question Kawakita et al. have posed would require a study of each clade in the two antagonist genera at a much finer scale. The question of how specialized Caloptilia and Diphtheroptila are hinges on how many species are in each genus, and that’s better addressed by examining population genetics, not ancient relationships among these genera.


Drummond, C., Xue, H., Yoder, J., & Pellmyr, O. (2009). Host-associated divergence and incipient speciation in the yucca moth Prodoxus coloradensis (Lepidoptera: Prodoxidae) on three species of host plants. Heredity, 105 (2), 183-96 DOI: 10.1038/hdy.2009.154

Kawakita, A., & Kato, M. (2006). Assessment of the diversity and species specificity of the mutualistic association between Epicephala moths and Glochidion trees. Molecular Ecology, 15 (12), 3567-81 DOI: 10.1111/j.1365-294X.2006.03037.x

Kawakita, A., Okamoto, T., Goto, R., & Kato, M. (2010). Mutualism favours higher host specificity than does antagonism in plant-herbivore interaction. Proc. Royal Soc. B, 277 (1695), 2765-74 DOI: 10.1098/rspb.2010.0355

Double the mutualists, double the fun?

ResearchBlogging.orgFor all living things, information is critical to survival. Where’s the best food source? Is there a predator nearby? Will this be a good place to build a nest? It probably shouldn’t be surprising, then, that lots of animals do what humans do when faced with a host of hard-to-answer questions—they take their cues from their neighbors.

Red-backed shrikes place their nesting sites near where other shrike species have set up territories. Many bird species recognize each other’s predator alarm calls, and respond appropriately. And a new natural history discovery published in the latest issue of The American Naturalist shows that treehoppers let one species of butterfly know where to find ants that will tend its larvae [$a].

The ant-tended butterfly (Parrhasius polibetes, above) looks for ant-tended treehoppers (Guayaquila xiphias, below) to know where to lay her eggs. Photos from Kaminski et al. (2010), figure 2.

The treehoppers help out the butterfly inadvertently, because both of them are dependent on a common resource: ants. Like many true bugs, treehoppers make their living sucking the sap of a host plant. This gives them a surplus of simple sugars and water, which they excrete as “honeydew” to attract ants for protection. As it happens, the larvae of the butterfly Parrhasius polibetes do the same thing—so the new study’s authors hypothesized that P. polibetes females might prefer to lay their eggs on plants where treehoppers were already present, since those would likely already have ants ready to protect butterfly larvae.

To test this, the authors set up experimental pairs of host-plant branches, one occupied by ant-tended treehoppers, and one not. They excluded ants from accessing the unoccupied branch with Tanglefoot, a water-resistant glue used in insect traps. After 48 hours, they checked the experimental plants for newly-laid butterfly eggs, and found that P. polibetes was both more likely to lay eggs, and laid more eggs at a time, on branches occupied by treehoppers.

To assess the fitness benefit of laying eggs on treehopper-occupied plants, the authors compared the survival of newly hatched P. polibetes larvae artificially introduced onto branches occupied by treehoppers to the survival of larvae introduced to branches unoccupied by treehoppers (and with ants excluded, again, using Tanglefoot). The larvae placed with treehoppers had substantially better odds of survival—about six times better.

These two experiments confound the effect of treehoppers with the effect of ants, however—so the authors performed one additional experiment. In this one, they set up paired branches with and without treehoppers, but allowed ants to reach both the occupied and unoccupied branches—and the general result from the earlier experiment held. Larvae placed near treehoppers were three times more likely to survive for the duration of the experiment even when larvae placed on a branch without treehoppers were able to attract ants on their own.

So it looks like P. polibetes is able to freeload on the treehoppers’ ant-attracting efforts, and benefits from that freeloading. What effect does that freeloading have on the treehoppers, or the ants, or the host plant? It’s hard to say based on the data presented in the current paper, but I’d guess that the treehoppers don’t lose much—in fact, they might gain from having another ant-attracting insect nearby, just as the butterfly larvae do. Similarly, it’s probably helpful for the ants to have more honeydew-producing species in the same location. It’s almost like that commercial for … what was the product?

(I’ll leave it to you, dear reader, to decide which insects correspond to which gendered pair in that video.)

I’d think, though, that this pile-on isn’t so good for the host plant, if plants already hosting treehoppers are more likely to have to deal with butterfly larvae, too. Untangling all the different ways these four species—ants, treehoppers, butterflies, host plants—exert direct and indirect natural selection on each other should keep the authors busy for a long time to come.


Hromada, M., Antczak, M., Valone, T., & Tryjanowski, P. (2008). Settling decisions and heterospecific social information use in shrikes. PLoS ONE, 3 (12) DOI: 10.1371/journal.pone.0003930

Kaminski, L., Freitas, A., & Oliveira, P. (2010). Interaction between mutualisms: Ant‐tended butterflies exploit enemy‐free space provided by ant‐treehopper associations. The American Naturalist DOI: 10.1086/655427

Magrath, R., Pitcher, B., & Gardner, J. (2007). A mutual understanding? Interspecific responses by birds to each other’s aerial alarm calls. Behavioral Ecology, 18 (5), 944-51 DOI: 10.1093/beheco/arm063

Parasites help figs control pollinators

Fresh off the open-access press at PLoS Biology: parasites may help to stabilize the mutualism between figs and fig wasps.

In nature, mutualistic relationships usually conceal a tug-of-war between interacting species. This is especially clear in the case of pollinating seed parasites, like yucca moths (my favorite) and fig wasps. Both these insects pollinate their eponymous host plants, then lay eggs in the fertilized flowers so their larvae can eat some of the seeds produced. Natural selection should push each interactor to overexploit this deal: the pollinator “wants” to lay lots of eggs, but the plant “wants” to get as many seeds as possible. Yuccas keep yucca moths in check by killing off flowers with too many eggs inside [subscription], but there hasn’t been a similar mechanism found in figs.

Until now. In the new paper, Dunn et al. show that the figs might benefit from parasites that attack the pollinating wasps. Fig flowers grow in “synconia,” hollow globes like the one in the photo above, which are lined inside with tiny flowers. Fig wasps climb inside the synconia to pollinate and lay their eggs in the flowers. Another wasp species parasitizes the pollinators by laying its eggs near pollinator eggs, so their larvae can eat the pollinator larvae when they hatch. Turns out, the parasites lay their eggs from the outside of the synconium, and the flowers inside the synconium vary in how close they are to the outer wall. Any pollinator eggs laid too close to the outer wall of the synconium are nailed by parasites – so the pollinators have an incentive to only lay eggs in the innermost flowers. Neato!

Dunn, D.W., S.T. Segar, J. Ridley, R. Chan, R.H. Crozier, D.W. Yu, and J.M. Cook. 2008. A Role for Parasites in Stabilising the Fig-Pollinator Mutualism. PLoS Biol 6(3): e59.