New paper: Conflict and communication in mutualism

*Medicago truncatula*, or barrel clover, a member of the legume family that hosts bacteria in its roots. The bacteria transform nitrogen gas from the atmosphere into fertilizer for their host plant, and the host feeds the bacteria with sugar. Experiments with barrel clover and its mutualists have shown that signals between the plant and the bacteria are important in this interaction, and provide an inspiration for the evolutionary models built by Yoder and Tiffin. (Photo by Jeremy Yoder.) (Flickr: jby)

I’m very excited to see this in virtual print — it’s a new model of coevolution between mutualists that takes into account signals between the partners as well as the benefits they provide each other (or don’t).

Yoder JB and P Tiffin. 2017. Sanctions, partner recognition, and variation in mutualism. American Naturalist doi: 10.1086/693472.

I’ll try to write about this in more depth at some point, but here’s the lay summary at the American Naturalist website:

Mutually beneficial relationships between species, or mutualisms, are ubiquitous in the living world, with examples ranging from flowering plants that rely on animal pollinators to fish that clean the teeth and scales of other fish. Mutualisms are often imperfect — one partner or the other varies in the quality of the help it provides. Evolutionary theory predicts that this should break up the relationship, but most mutualisms hold together in spite of partners that take the benefits of mutualism without properly paying them back.

This paradox may be explained by the fact that there’s more to mutualism than trading goods or services. This is a key result of mathematical evolutionary models published in the American Naturalist by Jeremy Yoder and Peter Tiffin, biologists at the University of British Columbia and the University of Minnesota. Yoder and Tiffin built a mathematical evolutionary model of mutualists that communicate before trading resources, and compared it to simpler models with only resource-trading or only communication. In the model with communication and resource-trading, host could “sanction” by cutting off resources to prevent poor quality partners from taking over, but evolution of the signals sent by partners and the hosts’ response to those signals maintained variation over time. Neither of the simpler models could do this. With only resource-trading, sanctions eliminated all poor-quality partners, and all variation; with only communication, poor-quality partners took over the mutualism.


New paper: Understanding mutualism with population genomics

Comparing metrics of diversity (x axis) and geographic differentiation (y axis) for thousands of genes in the Medicago truncatula genome (gray points) reveals that some symbiosis genes (red points) are genome-wide outliers — but they are not all the same kind of outlier (crosses and triangles). Yoder (2016), Figure 1.

Comparing metrics of diversity (x axis) and geographic differentiation (y axis) for thousands of genes in the Medicago truncatula genome (gray points) reveals that some symbiosis genes (red points, crosses, and triangles) are genome-wide outliers — but they are not all the same kind of outlier. Yoder (2016), Figure 1.

My very latest scientific publication is now online at the American Journal of Botany. It’s sort of an odd paper — something of a review, or an opinion piece, discussing how population genomic data can help us understand why mutualisms stay stable [PDF] in spite of the risk of “cheating” by partners, with a “worked example” with data from the Medicago HapMap Project. Here’s some key bits from the abstract:

Different hypothesized models of mutualism stability predict different forms of coevolutionary selection, and emerging high-throughput sequencing methods allow examination of the selective histories of mutualism genes and, thereby, the form of selection acting on those genes. … As an example of the possibilities offered by genomic data, I analyze genes with roles in the symbiosis of Medicago truncatula and nitrogen-fixing rhizobial bacteria, the first classic mutualism in which extensive genomic resources have been developed for both partners. Medicago truncatula symbiosis genes, as a group, differ from the rest of the genome, but they vary in the form of selection indicated by their diversity and differentiation — some show signs of selection expected from roles in sanctioning noncooperative symbionts, while others show evidence of balancing selection expected from coevolution with symbiont signaling factors.

The paper is my contribution to a Special Section on “The Ecology, Genetics, and Coevolution of Intimate Mutualisms”, which I co-edited with Jim Leebens-Mack. You can view the whole Special Section here, and download my paper here [PDF].


Chapter on coevolution in the Encyclopedia of Evolutionary Biology

Grant (1949).

My visualization of key data from Verne Grant’s 1949 paper showing that floral traits are more likely to be important in the taxonomic descriptions of plant species when those species are pollinated by animals — which suggests that those plant-pollinator interactions play a role in the formation of new species.

I got word this morning that the Encyclopedia of Evolutionary Biology, a huge compendium of current knowledge on evolution, systematics, and ecology, is now online. That’s exciting in and of itself, but it’s particularly so because it means you can finally see my contribution, the introduction to the topic of coevolution. Here’s the opening paragraph, of which I’m rather fond:

No organism is an island. Every living thing contends with predators, parasites, and competitors, and most also receive benefits from mutualists (Table 1). These interactions with other species exert natural selection—and predators, parasites, competitors, and mutualists may also experience selection in return. The mutual evolutionary change that results from this reciprocal selection is ‘coevolution’ (Janzen 1980; Thompson 2005).

The rest of the Encyclopedia includes contributions from a tremendous array of other authors, and I’m grateful to subject editor Andrew Forbes for the invitation to contribute. You can browse the whole thing on the publisher’s website, and download a manuscript-format PDF of the final text of my chapter here.


Coming soon: Crowd-funding a Joshua tree genome

Joshua trees at Tikaboo Valley, Nevada (Flickr: jby)

Joshua trees at Tikaboo Valley, Nevada (Flickr: jby)

I’m very excited to announce a new project, with a new model for doing science: The Joshua Tree Genome Project, in which I’m working with a bunch of smart, accomplished folks to sequence the genome of my favourite spiky desert plant. A sequenced Joshua tree genome will provide the framework to understand how coevolution with highly specialized pollinators has shaped the history of Joshua trees, and to use the landscape genomics skills I’ve developed with the Medicago HapMap Project and AdapTree to understand how the trees cope with extreme desert climates — and how to ensure they have a future in a climate-changed world.

Perhaps most excitingly (terrifyingly?) we’re going to raise some of the funds to do the genome sequencing by crowdfunding, using the platform. So please keep an eye on the project site, follow our Twitter feed, and Like our Facebook page to make sure you don’t miss your chance to help understand Joshua trees’ evolutionary past and ensure their future.


Nothing in Biology Makes Sense: Is a sloth’s best friend its moth-fur?

Three Toed Sloth Is it easier, being green? Photo by Bas Bloemsaat.

This week at Nothing in Biology Makes Sense! I’m discussing a new study that purports to demonstrate that three-toed sloths are in a nutritional mutualism with specialized moths, fueled by algae and poop:

Sloths’ coarse, shaggy fur accumulates its own little microcosm of living passengers. (If you move that slowly in a tropical forest canopy, you’re going to get some hop-ons.) Among these are an assortment of algae, and moths in the genus Cryptoses. It’s been known for a long time that Cryptoses moths lay their eggs in sloth dung, and that their larvae eat it.

To find out why it isn’t completely crazy to think that these poop-eating moths might be helpful to sloths, go read the whole thing.◼


Nothing in Biology Makes Sense: Tropical trees, getting by with a little help from their mutualistic ants

This week at Nothing in Biology Makes Sense! I’m discussing a nifty new study that suggests interacting species can sometimes tolerate stressful environments by helping each other out:

This was the perspective of Peter Kropotkin, a Russian prince and political anarchist who studied the wildlife of Siberia while working as an agent of the Czar’s government. In the harsh conditions of the Siberian winter, Kropotkin reported finding not a bitter struggle over scarce resources, but what he called “Mutual Aid” among species, as well as in the human settlements that managed to eke out a living.

Something like what Kropotkin described is documented in a new paper by Elizabeth Pringle and colleagues. Examining a protection mutualism between ants and the tropical Central American tree Cordia alliodora, Pringle et al. found that drier, more stressful environments supported more investment in the mutualism.

To learn how ants can help a tree deal with drier climates—no, it doesn’t involve little tiny bucket brigades—you’ll have to go read the whole thing.◼


Pitcher plant ants keep their host clean

Via Scott Chamberlain: A species of ants that lives in and around carnivorous pitcher plants isn’t entirely freeloading. They also clean the walls of the plant’s pitfall trap, keeping it nice and slippery for insect-trapping.

Of course, the ants have a vested interest in keeping the trap effective, since they eat some portion of the critters caught by their host. But it seems pretty straightforward to think that this helps the pitcher plant, too. A more definitive test would be to compare the survival and seed production of pitcher plants grown with and without a colony of ants to keep them clean. ◼


Passwords and eviction notices: How do plants keep their bacterial partners honest?

The nodule-y roots of a soybean plant. (Flickr: Pro-Soil AG Solutions)

ResearchBlogging.orgNitrogen is one of the elemental building blocks of life as we know it—it’s a basic component of amino acids, which are in turn the building blocks of proteins, which form the building blocks and moving parts of every living cell. The nitrogen interwoven in our tissues originated as part of the atmosphere we breathe, but the path from atmosphere to living flesh is far less direct than drawing a breath. Atmospheric nitrogen becomes useful to us animals only via an intimate relationship between a plant and bacterial growing in its roots.

The bacteria, called rhizobia, have the rare ability to “fix” free-floating nitrogen into biologically useable form. In return for this nitrogen source, the host plant allows the rhizobia to infect a specialized knob of root tissue, a root nodule, which it supplies with sugar for the benefit of its nitrogen-fixing guests. The plant uses the fixed nitrogen to make proteins for its own use, and anything that eats the plant afterwards benefits.

If all this sounds familiar, it’s because the interaction between plants and rhizobia is the focus of my developing postdoctoral research, and I’ve been writing about it as I’ve done more reading about it. Specifically, I’ve been interested in how plants might be able to make sure their root nodules house helpful bacteria rather than freeloaders, who enjoy the sugar supply inside the nodule without fixing nitrogen in return.

I’ve discussed a couple of different mathematical models that suggest some options. However, models are really just formal ways to follow through the implications of a particular idea, not necessarily descriptions of what actually transpires between a plant and the rhizobia inside its roots. So I thought it might make sense to step back and survey what we presently know about what goes on inside those root nodules.

Continue reading


Choosing your partner is only as helpful as the partners you have to choose from

Picking teammates. Original photo by humbert15.

ResearchBlogging.orgWhen you need partners for some sort of cooperative activity—say, teammates for a game of kickball—you’d probably like to have a choice among several candidates. That lets you weigh considerations about kicking strength and running speed—and who promised to give you his dessert at lunch period—to build a winning team. However, if the other team captain snaps up the good players first, the fact that you have a choice among the others might not make much difference.

Plants and animals looking for mutualists face a similar situation. Being able to choose among possible partners should allow the chooser to work with helpful partners and avoid unhelpful ones, but a new study suggests that in one widespread mutualism the process of choosing between partners can leave the chooser worse off than if it had no choice at all [$a].

Coauthors Erol Akçay and Ellen Simms focus on the effects of partner choice in the mutualism between plants and nitrogen-fixing bacteria—the interaction I’m studying in my current postdoc position, as it happens. All living things need nitrogen, but only some strains of bacteria are able to collect nitrogen from the atmosphere and “fix” it into a form that other organisms can use. Many plants, particularly members of the big and diverse bean family, have evolved to allow nitrogen-fixing bacteria to infect their roots—the plants form a nodule of root tissue around the infection and supply the tissue with sugar for the bacteria to feed on as they fix nitrogen. Eventually the nodule dries up and dies off, and the bacteria are freed into the soil, having multiplied many times over thanks to the food supply from the host plant.

A plant’s root nodules, some cut open to show the interior. Photo by pennstatelive.

To see how this choice might work in practice, Akçay and Simms construct a mathematical model of a plant with two nodules. Each nodule produces some level of nitrogen, and recieves some level of sugar from the plant. The plant negotiates with the two nodules in what’s called a “war of attrition” game: whichever partner wants a better deal cuts off the exchange of services, and holds out until the cost of losing the service it recieves is greater than the benefit it hopes to gain in the war of attrition.

Rather like ant-defended plants, plants that host nitrogen-fixing bacteria don’t seem to screen potential mutualistic bacteria before allowing them to infect their roots. However, after root nodules are established, the success of the mutualism from the perspective of both partners depends on the genetics of each [PDF], and when host plants receive supplemental nitrogen, they put fewer resources into growing nodules [PDF]. Host plants have been observed with different strains of bacteria in different nodules, and some nodules could contain diligent nitrogen fixers while others are full of freeloaders. This may be the point at which the plant has a choice of partners—it can potentially direct sugar to helpful nodules, and cut off unhelpful ones.

Because the plant has two nodules to choose from, it can potentially outlast an uncooperative nodule by relying on the other one. This works if the plant can shunt more resources to the cooperative nodule and recieve more nitrogen from it in return. However, the success of this strategy depends on two traits of the bacteria inhabiting the nodules—how readily they ramp up nitrogen production in response to more sugar, and how stubborn they are in the war of attrition game.

If both nodules are stubborn but responsive to extra sugar, the plant can negotiate with one nodule by giving the other more sugar and receiving extra nitrogen. This lets the plant hold out longer in the war of attrition. On the other hand, nodules that are not responsive to extra sugar but also not very stubborn yield quickly in the war of attrition even though they don’t help much in negotiations. In either of these two cases, the negotiations find an equilibrium in which the plant receives a benefit about intermediate between what it would recieve if both nodules were infected by the same strain of bacteria.

However, if the plant hosts a stubborn-responsive bacterial strain in one nodule and a yielding-unresponsive strain in the other, it finds itself in a trap: the yielding-unresponsive strain is no help in negotiation against the stubborn-responsive strain, and the help provided by the stubborn-responsive strain isn’t an advantage in negotiating with the yielding-unresponsive strain. Over successive negotiations, the stubborn-responsive strain can ratchet up the sugar it extracts from the plant, and the plant ends up worse off than it would be if the two nodules were identical.

Just like humans haggling in a marketplace, the outcome of the interaction depends strongly on whether the other party plays along as expected.

Akçay and Simms find a way out of this trap by adding another wrinkle to the model. Much like the contract-theory models of mutualism I’ve discussed before, they modify the model to allow cooperative nodules to benefit from being cooperative. This makes a good deal of intuitive sense—if a nodule provides a better deal to the plant, the plant can potentially grow more leaves to produce more sugar, which would allow it to offer a better deal to the bacteria it hosts. Akçay and Simms call this “partner fidelity feedback,” and they find that, if it is sufficiently strong, it can allow the plant to out-negotiate a stubborn strain of bacteria.

Although it has a good deal of intuitive appeal, the model presented by Akçay and Simms does a fair bit of speculating in the absence of data. This is also a problem for the contract-theory model, and really all models of this widespread and important interaction. We know a great deal about the chemical details of plants’ interaction with nitrogen fixing bacteria. However, we don’t have a good sense of whether and how plants can redirect resources among nodules to haggle with the bacteria they host, and we don’t know whether and how bacteria could adjust their behavior to haggle with the plant. Akçay and Simms devote a big section of their online appendix [$a] to discussing just this point.

To figure out what’s going on inside those nodules, we need to determine how different models of interaction between plants and their bacterial mutualists may shape patterns in things that are easier to observe—both in the compatibility between plant genotypes and bacterial strains in greenhouse tests, and in the broader population genetics of both partners.


Akçay, E., & Simms, E. (2011). Negotiation, sanctions, and context dependency in the legume-rhizobium mutualism. The American Naturalist, 178 (1), 1-14 DOI: 10.1086/659997

Heath, K. (2010). Intergenomic epistasis and coevolutionary constraint in plants and rhizobia. Evolution DOI: 10.1111/j.1558-5646.2009.00913.x

Heath, K.D., Stock, A.J., & Stinchcombe, J.R. (2010). Mutualism variation in the nodulation response to nitrate Journal of Evolutionary Biology, 23 (11), 2494-2500 DOI: 10.1111/j.1420-9101.2010.02092.x


Freeloading caterpillars get in the way of plant-ant mutualism

Cecropia obtusifolia provides food for ants that come and protect it—unless caterpillars get there first. Photo by wallygroom.

ResearchBlogging.orgImagine you need a team of security guards. To find them, you decide not to place an ad in the local paper or on Craigslist. Instead, you build an apartment complex next to your home, complete with a full-service cafeteria providing free hot meals 24 hours a day. You leave the front doors unlocked, then hope that anyone who shows up to live in the apartments will also keep an eye on your home.

If you took that strategy to protect your assets, you’d have to be crazy. But that’s pretty much what ant-protected plants do all the time. They grow hollow structures called domatia, secrete nectar from special structures, and even produce tasty and nutritious “food bodies.” Then they wait for ants to move into the domatia, eat the nectar and the food bodies, and hopefully chase away anything that might want to do the plant harm. The crazy thing is, it works.

Well, it mostly works.

One gap in the ant-protection mutualism is the period when an ant-protected plant hasn’t grown big enough to support a whole colony of ants. In this early stage, ants won’t colonize the plant, but other insects might be quite happy to take the rewards that are already being offered. That’s exactly what larvae of the butterfly Pseudocabima guianalis do—they make themselves at home on unprotected ant-plants.

The ant-plant Pseudocabima caterpillars target is Cecropia obtusifolia, a shrubby Central American tree that relies on ants in the genus Azteca for protection. Azteca ants make vicious and well-coordinated bodyguards. Here’s video Ed Yong posted last year, showing a bunch of the ants flushing a hapless moth into an ambush.

However, Cecropia saplings can’t produce enough food to support a colony of ants until the plants grow to more than a meter tall. What’s too little for thousands of ants is a feast for a Pseudocabima caterpillar, however. Each caterpillar builds a silk shelter around a region of the plant that grows food bodies, and settles in to eat. As it grow larger, the caterpillar moves into a domatium near its original shelter, covering the entrance hole with silk. Finally the caterpillar pupates inside the domatium, emerging as an adult to lay eggs on another unprotected Cecropia plant.

Eventually the Cecropia saplings grow large enough to attract ants, who run off the caterpillars. However, as the paper I linked to above describes, the caterpillars seem to be able to resist an ant colony’s establishment on the plant—the silk shelters prevent ants from getting to the best sources of food. Cecropia saplings occupied by caterpillars didn’t seem to suffer more herbivore damage than ant-protected plants, but they did grow more slowly over the course of several years’ observations. Caterpillar-infested Cecropia plants were also more vulnerable to infection by a fungus, which the ants removed quite effectively.

Interestingly, though, caterpillar-infested plants also produced less food than those guarded by ants. This is a point of circumstantial evidence for a new model of mutualism I wrote about earlier this year, in which cheating is reduced or prevented when a host like Cecropia better mutualists help create better rewards. An ant-protected plant can divert more resources to feeding its tenants, so their work rewards itself. However, Pseudocabima caterpillars are glad to take the lower level of rewards that Cecropia plants offer up to all comers.

In other words, if you’re going to give out free lunches, you can’t really expect everyone who eats to pay you back.


Roux, O., Céréghino, R., Solano, P.J., & Dejean, A. (2011). Caterpillars and fungal pathogens: Two co-occurring parasites of an ant-plant mutualism. PLoS ONE, 6 : 10.1371/journal.pone.0020538