Disentangling the direct and indirect effects of adaptation on ecosystems

Andrew Hendry has been telling me for at least a year now that I will never get a faculty job until I post something on his blog. Matt Walsh’s recent job snag shortly after posting his first post on the blog was enough to convince me that there may actually be something to this. My first post will just be to shamelessly advertise a new paper that Regis Ferriere, Andres Lopez-Sepulcre, Michael Marshall, Joseph Travis, Cathy Pringle and David Reznick and I recently had come out in the American Naturalist titled: “Direct and indirect ecosystem effects of evolutionary adaptation in the Trinidadian guppy” available at:  http://www.jstor.org/stable/10.1086/666611 and also another recently accepted to the American Naturalist that should be out in a few months.

First some background to the system. Those familiar with guppies can skip to the next paragraph and avoid the usual guppy system description. Guppies in Trinidad live along a gradient of fish community types in Trinidad. At one extreme, in so-called “high-predation” communities, they live with many different fish including the killifish, several species of caracins, several species of cichlids and two fish predators: the wolfish and pike cichlid. At the other extreme, in so-called “low-predation” communities, guppies live only with the killifish. These two community types are often separated by barrier waterfalls that restrict fish from moving from the high predation locations below to the low predation communities above. New populations of low predation guppies are believed to evolve when several guppies from the high predation communities make it above the barrier waterfalls into communities that contain only killifish. Decades of research on guppies between these two community types has shown that they exhibit genetically based differences in male coloration, body-shape, behavior and demography. A couple of years ago we asked whether these two phenotypes also impacted the environment in different ways. We conducted mesocosm experiments using guppies from 2 different high and low predation populations. We crossed phenotype treatments with density treatments to yield a relative impact of guppy phenotype compared with guppy density. The results, published in PNAS (http://www.pnas.org/content/early/2010/02/03/0908023107 ), showed that the different phenotypes had different effects on multiple components of the ecosystem and that these effects were often as large as those of guppy density. Several studies like ours have examined the net effects of adaptive evolution on ecosystem properties. However, we do not know if these effects are confined to direct interactions or if they propagate further through indirect ecological pathways. Even less well understood is how the combination of direct and indirect ecological effects of the phenotype promotes or inhibits evolutionary change.

In this most recent paper, we coupled mesocosm experiments with ecosystem modeling to separate not only the direct and total indirect effects of different phenotypes of Trinidadian guppies on the ecosystem, but also the total indirect effect into its constituent parts. We show that although the direct effect of differences in the degree to which different types of guppies eat algae is larger than the total indirect effect, which includes the effect of guppies eating invertebrates that eat algae and the recycling of nutrients by guppies that stimulate the growth of algae, the individual indirect effects are much larger than the direct effect. Some of these individual indirect effects oppose each other and cancel each other out, leading to the small total indirect effect. However, we used the model to show that it is the indirect effects that determine how evolutionary changes in the degree of herbivory alter the amount of algae in the mesocosms and how this can lead to further evolutionary change in guppy herbivory. Along with the relatively recent realization that significant evolutionary change within a species can be rapid enough to happen over observable timescales, the researchers show how these types of dynamics can be used to predict the outcome of feedbacks between ecological and evolutionary processes.

In another recently accepted paper to American Naturalist, we dealt with the other side of eco-evo. This time we were interested in whether eco-evo feedbacks were important in the evolution of the low predation phenotype. We did so by asking whether the simplest type of eco-evo feedback—density-dependent selection—can explain the ability of the low predation phenotype to evolve from a high predation ancestral population. Density and biomass of guppies in low predation communities is higher than in high predation communities, partially due to release from predation in the low predation communities.

We first used a series of density manipulation experiments in 10 natural low predation streams to evaluate whether populations of guppies are regulated via density-dependence. We show that increasing density resulted in changes in vital rates that should decrease the population growth rate; decreasing population density resulted in changes to the vital rates that increased population growth rate. Control populations, where we did not manipulated the density of guppies exhibited population growth rates that did not differ from stasis. Next we conducted a series of mesocosm experiments where we crossed guppy phenotype with population density. At the end of the experiment we calculated rates of phenotypic growth based on measurements of the vital rates at the end of the experiment. The results of this component showed that at low density, high predation guppies had higher rates of growth than their low predation counterparts, but this advantage disappeared at high densities—indicating that low predation guppies are better able to deal with the negative effects of increased densities. In both the field manipulations and the mesocosm experiment, we were able to calculate actual population growth rate or phenotype fitness using integral projection models. Finally, we used the results from both these experiments to parameterize an evolutionary invasion analysis where we formally ask if an incipient low the predation phenotype can invade a population of high predation ancestors. Density here is central because the ability of the low predation phenotype to invade hinges on whether or not they exhibit positive population growth at the population equilibrium value of the resident (high predation) phenotype. The answer appears to be that density dependence allows an incipient population of low predation fish to initially invade, but is not enough to allow them to evolve towards to complete low predation fish.

Stay tuned for more papers coming soon on other ways evolutionary adaptation influences ecosystem processes and what missing pieces are needed to explain how low predation guppies evolve……..