The birds and the bees and the microbes

Vannette et al. 2013. “Nectar bacteria, but not yeast, weaken a plant-pollinator mutualism”. Proceedings of the Royal Society B-Biological Sciences.

"When we try to pick out anything by itself, we find it hitched to everything else in the Universe."- John Muir

You can’t help but marvel at the complexity of ecology, at the intricacy and multiplicity of species interactions. But this complexity is also problematic. For many ecologists, it becomes necessary to focus on a single type of interaction, or on interactions limited to only a few species. But real ecological systems are hardly ever limited to a single important process. They might include competition, mutualism, facilitation, predation, environmental constraints and fluctuations, additive and interactive effects, nonlinearities, thresholds and emergent properties. Can knowledge of  omplexity emerge from simplicity? Can simplicity emerge from complexity? These are important and longstanding questions in ecology, the focus of some of our smartest minds. We may not have the answer yet, but if nothing else, it is helpful when experimental work in community ecology attempts to explore multiple interactions.

For example, some of the work from Tadashi Fukami’s lab is focused on how communities of microorganisms (yeast and bacteria species) assemble in Mimulus aurantiacus nectar. In the past, this work has focused particularly on priority effects and resource competition, which plays an important role in structuring these communities. While past work has suggested the importance of pollinators as a dispersal vector for microorganisms, fascinating new work suggests that microbial communities have important effects on pollinators and their mutualistic interactions with the host plant as well.

Vannette et al. (2013) focused on the effects of the two most abundant species in Mimulus nectar, Gluconobacter sp., an acid-producing bacteria, and Metschnikowia reukaufii, a yeast. They then looked at three related questions – how do nectar microbes affect nectar chemistry, how do they affect nectar removal by hummingbird pollinators, and how to they affect pollination success and seed set. Basically, do nectar microbes disrupt important mutualistic interactions between the plant and their pollinators, or are their effects neutral?

This is where the story becomes interesting. Both types of microbes altered nectar chemistry, but in different ways. The bacteria acidified the nectar (to ~ pH 2.0) far more than the yeast species, and tended to also reduce the sugar content of the nectar far more than the yeast. Hypothesizing that these changes could ultimately affect pollinator preference, the authors then filled artificial flowers with nectar containing either the bacteria, yeast, or no microbes. Half of these flowers were bagged to prevent hummingbird access. Compared to the bagged controls, flowers with bacteria-inoculated nectar had less nectar removed than either yeast-inoculated or sterile nectar. It appeared that nectar removal was related to the changes in chemistry driven by bacterial growth in the nectar. Finally,  the authors looked pollination success in relation to microbial inoculation. Flowers inoculated with bacteria did indeed have less pollination success (measured as the number of stigmas closed) and had decreased seed numbers. Microbial communities were not isolated from the ecology of the plant.

Perhaps none of this is that surprising – hummingbirds are intelligent and will preferentially feed, and pollinator choice is important for plant fitness. However, these bacteria and yeast species are specialized for growth in the hypertonic nectar environment and their continued presence in the ecosystem depends on dispersal from one flower to the next before their host flower dies. The transient nature of this nectar habitat suggests that obtaining a dispersal vector should be important. The fact that Gluconobacter alters nectar chemistry in a way which negatively affects their likelihood of movement to other patches suggests an interesting paradox and a complicated relationship between plants, their nectar microbes, and pollinators. Gluconobacter species growing in Mimulus flowers produce acidifying H+ ions and reduce sugar concentrations in nectar – this increases their likelihood of winning competitive interactions with other microbes in the nectar, which should select for the maintenance of acidifying, sugar-reducing characteristics. But on the other hand, these characteristics reduce the likelihood of being transported to new flowers and persisting in the metacommunity. Further, these pollinator-decreasing characteristics may result in selection by the Mimulus plants for nectar compounds that reduce microbial contamination. So understanding competitive interactions in microbial communities, or understanding pollinator-plant interactions, or understanding pollinator-microbial interactions on their own might be inadequate to understand the important ecological and evolutionary processes structuring the entire system.

Given that the question of simplicity vs. complexity is still so difficult and at least for me, uncertain, I would hesitate to draw a general conclusion about whether this is the kind of work all community ecologists should strive for. But it seems that recognizing ecological and evolutionary context is key, whether you work with Arabidopsis, microcosms, or tropical forests.

Seed dispersal: plant height seems to be more important than seed size!

I really like papers that teach me something that I didn’t know. But, I love papers that show me that what I learned is wrong. This is the case of a new paper by Fiona Thomson, Angela Moles, Tony Auld, and Richard Kingsford on seed dispersal that appears in the last issue of the Journal of Ecology. This group from Australia analyzed the effects of seed size and plant height on their dispersal abilities. They reviewed intensively the literature gathering data on 200 species from 148 studies around the world. Surprisingly to me, they found plant height was much better at predicting seed dispersal than seed size. This might not sound so surprising for many people (and after seeing the paper, kind of intuitive), but there was a lot of evidence that seed size was the best predictor of dispersal, with species with smaller seeds dispersing further than species with bigger seeds. For wind dispersed species, their results are more intuitive, but they found this pattern in a number dispersal syndromes analyzed (i.e. unassisted, wind, ballistic, ingestion, and ant dispersal). So, in your next study on seed dispersal consider adding plant height as an explanatory variable.

Thomson, F. J., A. T. Moles, T. D. Auld, and R. T. Kingsford. 2011. Seed dispersal distance is more strongly correlated with plant height than with seed mass. Journal of Ecology 99:1299-1307. DOI 10.1111/j.1365-2745.2011.01867.x

Niche or Neutral? Why size matters.

Metacommunity dynamics (i.e. the effects of dispersal among connected communities) have become an increasingly common lens through which to explain community structure. For example, competition-colonization models explain the coexistence of superior and inferior competitors as the result of a trade-off in colonization and competitive ability. Species are either superior competitors, with high probabilities of establishing in patches, but low ability to move between patches, or superior colonizers, which have tend to lose in competitive interactions but can travel easily between patches. Under this framework, the ability of superior colonizers to reach and maintain populations in patches where their superior competitors are absent allows them to avoid extinction.

One problem with these types of models is that they rarely acknowledge the importance of ecological drift – that is, that chance events also affect species interactions. This despite the fact that we know that in “real life”, chance events likely play a major role in producing assemblages different than those we might predict based on theory. One of the strengths of the Hubbell’s neutral model is that it recognizes and embraces the importance of randomness.

A recent paper by Orrock and Watling (2010) examines how chance events can alter the predictions of the classic competition-colonization model. Orrock and Watling show that the size of communities in a metacommunity (which is assumed to correlate with the strength of ecological drift) determines whether community dynamics are niche-structured or neutral in nature. In large communities, predictions agree closely with those of the classic competition-colonization model, and niche-based interactions (i.e. competitive hierarchies) dominate. It’s in small communities that things get interesting: ecological drift becomes more important, so that differences in competitive ability between species are effectively neutralized. As a result, small communities begin to resemble neutral assemblages in which species abundances don’t relate to differences in competitive ability. An interesting consequence of this outcome is that species who are poor competitors but good colonizers have an additional refuge – simply by escaping to small communities, even if these communities contain superior competitors, they can persist in a metacommunity.

Beyond the theoretical implications of this model, the applied implications are what really matter. Habitat destruction and fragmentation are an growing problem due to human activities. Habitat patches are often smaller, and of lower quality, decreasing the size of the community each patch can support. Even if these patches are still connected and functioning as a metacommunity, species which rely on their strong competitive ability for persistence will lose this advantage as assemblages become increasingly neutral. Under this model, community diversity declines even more as habitat is lost than in the traditional competition-colonization model, and superior competitors face even greater extinction risk than previously predicted.

Since in reality, metacommunities are likely to consist of patches of different sizes, rather than all large or all small patches, the predictions here remain to be extended to more realistic metacommunities. However, Orrock and Watling have produced a useful model for understanding how ecological drift can affect diversity in a metacommunity and alter the expectations of traditional competition-colonization models.


Orrock, J.L. and Watling, J.I. (2010) Local community size mediates ecological drift and competition in metacommunities. Proc. R. Soc. B.

Plant rarity: environmental or dispersal limited?

In order to promote the persistence and possible spread of extremely rare plant species, ecologists need to know why a species is rare in the first place. In 1986, Deborah Rabinowitz identified seven forms of rarity, where rarity could mean several things depending on range size, habitat specificity and population sizes. When considering rarity, it often feels intuitive to look for environmental causes for these different forms of rarity. Habitat alteration is an obvious environmental change that affects abundance and distribution, but are rare species generally limited by habitat or resource availability? The alternative cause of rarity could just be that sufficient habitat exists, but that the rare species is simply unable to find or disperse to other sites. An extreme example of this would be the Devil's Hole pupfish which exists at only a single pool. It can survive elsewhere (such as in artificial tanks) but natural dispersal is impossible as its pool is in a desert.

Photo taken by Kristian Peters and available through GNU free documentation license

In a recent paper by Birgit Seifert and Markus Fischer in Biological Conservation, they examine whether an endangered plant, Armeria maritima subsp. elongata, was limited because of a lack of habitats or if it was dispersal limited. They collected seeds from eight populations and experimentally added these seeds to their original populations and to uninhabited, but apparently appropriate sites. They found that seeds germinated equally well in inhabited and uninhabited sites and seedlings had similar survivorships. They found that variation in germination rates were likely caused by originating population size and that low genetic diversity and inbreeding reduce viability.

These results reinforce two things. First is that conserving species may only require specific activities, such as collect and distributing seeds. Here ideas like assisted migration seem like valuable conservation strategies. Secondly, we really need to be doing these simple experiments to better understand why species are rare. If we fail to understand the causes of rarity, we may be wasting valuable resources when try to protect rare species.

Seifert, B., & Fischer, M. (2010). Experimental establishment of a declining dry-grassland flagship species in relation to seed origin and target environment Biological Conservation DOI: 10.1016/j.biocon.2010.02.028

Adaptation and dispersal = (mal)adapted

Ever since Darwin, we often think of organisms as being in a constant battle against other organisms and local environments. Thus natural selection and the resulting arms race results in organisms highly adapted to local conditions and against local antagonists. At the same time, and especially driven by theoretical advances in the 1990's, researchers began to ask how dispersal -that is, the flow of genetic material from elsewhere, can disrupt local adaptation. On the one hand it may provide genetic variation allowing for novel solutions to new difficulties. On the other hand, dispersal may reduce the prevalence of fitness-increasing genes within local populations.

In a simple but elegant experiment, Jill Anderson and Monica Geber performed a reciprocal transplant experiment, moving Elliott's Blueberry plants between two habitats. One population was from highland, dryer habitats and the other from moist lowlands. They further evaluated performance in greenhouse conditions. Their results, published in Evolution, show that these two populations have not specialized to local conditions. Rather, due to asymmetric gene transfer, lowland individuals actually performed better when planted in highlands than compared to their home habitat. Further, in the greenhouse trials, lowland species did not perform better under higher moisture conditions. While genetic or physiological constraints may also limit adaptation, Anderson and Geber present a fairly convincing case that gene flow is the culprit.

These results reveal that populations may actually be relatively mal-adapted to local conditions, which has numerous consequences. For example, we need to be cognizant of adaptations to particular conditions when selecting populations for use in habitat restoration and when trying to predict response to altered climatic or land-use conditions. Importantly what does this mean for multi-species coexistence? Dispersal seems to limit the ability to adapt, and thus, better use local resources or maximize fitness, making for a better competitor. At the same time, dispersal can offset high death rates, allowing for the persistence of a population that would otherwise go extinct. Understanding how these two consequences of dispersal shape populations and communities is an interesting question, and work like Anderson and Geber's provides a foundation for future studies.

Anderson, J., & Geber, M. (2009). DEMOGRAPHIC SOURCE-SINK DYNAMICS RESTRICT LOCAL ADAPTATION IN ELLIOTT'S BLUEBERRY (

)
Evolution DOI: 10.1111/j.1558-5646.2009.00825.x