Several studies have shown that ecological change can drive evolutionary change on contemporary scales, and some studies are now showing the reciprocal pathway. Eco-evolutionary dynamics in the broad sense can thus include the unidirectional effects of ecological changes on evolutionary processes or the unidirectional effects of evolutionary changes on ecological processes. An important goal, however, should be to elucidate bidirectional eco-evolutionary interactions, referred to as ‘eco-evolutionary feedbacks’ (Post & Palkovacs 2009
). Such feedbacks might arise at different scales in space and time and at different levels of biological organization (genes, individuals, populations, communities and ecosystems). Because many biological processes are hierarchical or organized in a network, changing one aspect of the chain can affect many levels (Coulson et al. 2006a
). Many such interactions and feedbacks are possible and we will here provide only a few examples.
At the population level, natural selection and population dynamics are closely linked, simply because both are affected by the birth and death of individuals. Thus, if natural selection acts on a trait (e.g. body size) through survival or reproductive success, it will leave a population dynamical signature (Coulson et al. 2006b
; Pelletier et al. 2007
). This signature is expected to be the strongest when populations are not already density limited (Saccheri & Hanski 2006
; Kinnison & Hairston 2007
). As another example, landscape features can influence the distribution of genetic diversity at the individual and population levels (Jeyasingh & Weider 2007
; Clark et al. 2008
), thus affecting gene frequencies. If the change in gene frequencies translates into phenotypic trait changes that affect demographic rates (such as reproduction, survival or dispersal), then, ultimately, the genetic change will affect population dynamics (Hanski & Saccheri 2006
; Saccheri & Hanski 2006
). Effects at the community level are also possible. When two species, such as a predator and a prey, interact, evolutionary change in one can influence selection and demography of the other—and vice versa (Fussmann et al. 2007
). Finally, changes in the numbers of individuals, the types of resources they use and their efficiency in doing so can influence a suite of ecosystem functions, such as decomposition, nutrient cycling and primary productivity (Thompson 1998
There are many ways to write general expressions for eco-evolutionary dynamics (e.g. Coulson et al. 2006a
; Fussmann et al. 2007
; Bailey et al. 2009
; Post & Palkovacs 2009
). The commonality between all of these expressions is a hierarchical relationship between fitness, phenotypes and genes. As a complement to Dobzhansky's famous quote ‘nothing in biology makes sense except in the light of evolution’ (Dobzhansky 1964
), we therefore suggest that nothing in evolution or ecology makes sense except in the light of the other. The present theme issue is an attempt to underline the value of this view by focusing on several key questions.
(a) How does evolution influence demography?
Understanding the relative contribution of ecological and evolutionary processes to the changes in population dynamics is essential to understand how the dynamics of heritable traits influence population processes. As reviewed previously, several studies have shown that phenotypic and microevolutionary changes can occur on ecological time-scales. These changes presumably influence population dynamics, but the extent of this influence is still generally unknown (Saccheri & Hanski 2006
) because only a few studies have quantified the importance of phenotypic change to population growth and dynamics. The few examples include analyses of Darwin's finches (Geospiza fortis
; Hairston et al. 2005
), Soay sheep (Ovis aries
; Pelletier et al. 2007
) and New Zealand chinook salmon (Oncorhynchus tshawytscha
; Kinnison et al. 2008
). Such work is increasingly needed, however, as it may yield information as to the magnitude and direction of the influence evolution exerts on population persistence and extinction in changing environments (Burger & Lynch 1995
; Gomulkiewicz & Holt 1995
; Orr & Unckless 2008
The first paper in this section of the theme issue (Ezard et al. 2009
) extends the above analyses by quantifying the influence of phenotypic distributions on population dynamics in five natural systems. Specifically, the authors partition the effects of phenotypic and environmental changes on population growth (Hairston et al. 2005
) using five detailed long-term time series of marked ungulate populations. Ezard et al. (2009)
find that juvenile mass (a key heritable life-history trait in ungulates; Albon et al. 1987
; Wilson et al. 2005
) can influence population growth to a similar extent as can environmental variables such as climate and food availability. In short, Ezard et al. (2009)
show that neither phenotypic changes nor ecology can be ignored when considering population growth.
Life histories emerge from transitions between life stages and can therefore influence population dynamics (Saether & Bakker 2000
; Oli & Dobson 2003
). Moreover, life histories can clearly evolve on short time-scales (Stearns 1983
; Reznick et al. 1990
), thus providing a possible conduit for eco-evolutionary dynamics. Several theoretical studies have tried to understand how the variety of life histories observed in nature might have evolved. These models have generally considered the fitness of alternative life histories in stable environments, which may be an unrealistic assumption. In the present theme issue, Tuljapurkar et al. (2009)
explore the fitness landscape of various life histories under varying environmental conditions. Their work suggests that ‘optimal phenotypes’ in a fluctuating environment are likely to differ from those in a constant environment.
Central to all studies of eco-evolutionary dynamics is the concept of ‘fitness’. In a constant environment with no density dependence and no age structure, classic theoretical models hold that evolution maximizes the mean relative fitness (Fisher 1930
; Crow & Kimura 1970
). Natural populations, however, frequently experience variable environments (Engen et al. 2005
) and density dependence (Bonenfant et al. 2009
). Lande et al. (2009)
apply diffusion theory to analyse a system of coupled ecological and evolutionary dynamics to derive an evolutionary maximum principle under these more realistic conditions. They show how ‘the magnitude of environmental stochasticity and the form of density dependence govern the trade-off between r
(b) How do ecological variables influence the potential for evolution?
The recognition that ecological conditions can influence evolution is not new. In fact, genotype-by-environment interactions have long been integrated into evolutionary quantitative genetic theory (Bradshaw 1965
; Falconer & Mackay 1996
). To date, most evidence for such interactions has come from controlled laboratory experiments. For natural populations, some studies have revealed that habitat quality can influence genetic variance (reviewed in Hoffmann & Merilä 1999
; Charmantier & Garant 2005
) and covariance (reviewed in Sgrò & Hoffmann 2004
). However, relatively few attempts have been made to assess the influence of environmental variability on the distribution of genetic diversity at different levels of biological organization. Studies in this section of the present theme issue used a variety of approaches to determine how ecological processes influence evolutionary parameters, as well as their links to short- and long-term population dynamics.
One factor that can influence genetic variation, adaptation and population dynamics is dispersal (reviewed in Garant et al. 2007
). Zheng et al. (2009)
emphasize the need to study these interactions in theoretical models tailored to specific empirical systems. For example, evolutionary models do not usually formalize the actual genetic mechanisms underlying dispersal. Zheng et al. (2009)
do so with an individual-based model of Glanville fritillary butterflies (Melitaea cinxia
) in a fragmented landscape. The authors find a strong coupling between metapopulation dynamics and genetic dynamics at a locus (Pgi) affecting dispersal and other life-history traits. The strength of the coupling depends on spatial scale and is asymmetric; demography has a greater impact on genetics than vice versa. Zheng et al. (2009)
consider the positive role of dispersal on population dynamics, whereas Moore & Hendry (2009)
consider the possible negative role. They specifically test whether gene flow between three-spine stickleback (Gasterosteus aculeatus
L.) populations in different environments (lake versus stream) can cause maladaptation that limits population size and constrains species ranges (suggested by Kirkpatrick & Barton 1997
). Finding only mixed evidence in support of the hypothesis despite a simplified system with high gene flow and trait maladaptation, Moore & Hendry (2009)
discuss how any negative effects of gene flow on population dynamics might be difficult to demonstrate in nature.
Another factor that might influence genetic variation, adaptation and population dynamics is spatial habitat heterogeneity. The integration of geographical information system methods and landscape genetic approaches is well suited to explore this topic, as shown by one contribution to the special issue. Porlier et al. (2009)
explore how the spatial distribution of habitat types influences population structure and individual neutral genetic variability in wild tree swallows (Tachycineta bicolor
). The authors show that individual genetic diversity is greater in poorer habitats, at least partly due to non-random settlement patterns of swallows across the landscape. Dionne et al. (2009)
then provide a direct assessment of how spatial variation in selection influences genetic variability of selected loci. The authors first document differences in parasites and pathogens for Atlantic salmon (Salmo salar
) from distinct genetic groups and ecological regions. They then identify specific major histocompatibility complex alleles that influence resistance to pathogen infection in their wild populations. Finally, they show a decrease in the frequency of the susceptibility allele through time, suggesting that pathogen-driven selection is acting in their systems.
Ultimately, many investigators would like to predict evolutionary patterns among populations as a function of selection and genetic variation within populations. This effort has thus far been hampered by the lack of a precise way to quantify and compare complex multivariate genetic variance–covariance (G
) matrices. To fill this gap, Hine et al. (2009)
present a new method for comparing G
matrices and then ask whether the G
matrix within Drosophila serrata
populations can explain variation in the same traits among those populations.
(c) When will evolution influence community and ecosystem variables?
The idea that contemporary evolution can influence demographic parameters is relatively straightforward as are some links between contemporary evolution and community structure. For instance, the evolution of one species often improves its ability to compete with, or to exploit, another species (Yoshida et al. 2003
; Fussmann et al. 2007
). This adaptation of one species then selects for countermeasures in the other species, initiating a coevolutionary arms race with a variety of possible outcomes (Thompson 2005
). Coevolutionary feedbacks are also particularly obvious for tight mutualistic interactions such as those between ants and aphids, and flowering plants and pollinators. Less straightforward, however, are local links between evolutionary change and higher level ecological variables, such as community composition, food webs, nutrient dynamics and primary productivity. In comparison with population dynamics, these variables are further removed from the genetic variation on which selection acts, and may therefore be less sensitive to evolutionary change. What, then, are the conditions under which these higher level eco-evolutionary dynamics might be important?
One of the earliest formal demonstrations of evolutionary effects, above the population level, was chemostat experiments showing that evolutionary change influenced population cycles of predators (rotifers) and prey (algae) (Yoshida et al. 2003
). Jones et al. (2009)
further explore that system by modelling factors influencing its eco-evolutionary dynamics. The authors show that ecological and genetic ‘details’ matter. For example, changes in environmental variables (chemostat flow rates) or cost/effectiveness trade-offs in prey defence can shift the identity of the surviving prey types. An even more complex range of outcomes follows from a consideration of heritable variation in both multiple predator types and multiple prey types with resulting dynamics ranging from equilibrium to chaos (Jones et al. 2009
Chemostat experiments have proven extremely informative but experiments in nature are also needed. Johnson et al. (2009)
take steps in this direction with a combined theoretical–experimental approach. They first develop a theoretical framework for understanding when natural selection should drive community dynamics. They then demonstrate these conditions in their experimental plant–arthropod system, i.e. natural selection acts on plant biomass, which is heritable and influences arthropod abundance and species richness. These findings are then integrated back into the theoretical framework to predict evolutionary changes in the plants, as well as ecological responses in arthropod communities.
Ecosystem variables should be less affected by genetic variation than are population and community variables. Bailey et al. (2009)
therefore predict that genetic variation will have its greatest effect on population variables, a weaker effect on community variables and the weakest effect on ecosystem variables. Using a meta-analytical approach, the authors confirm that genetic variation in plant populations is most closely associated with plant traits, less associated with community variables and least associated with ecosystem variables. Critically, however, strong effects of genetic variation were sometimes present at all of these levels, highlighting the need for future studies to consider additional factors that might influence the strength of eco-evolutionary dynamics.
Many studies of eco-evolutionary dynamics at the community and ecosystem levels focus on genetic variation in plants as the evolutionary driver. Here, Palkovacs et al. (2009)
instead focus on the potential effects of genetic variation in animals. The authors use mesocosms to study the ecological effects of population-level adaptive differences in guppies (Poecilia reticulata
) and killifish (Rivulus hartii
). They find that guppies adapted to different predation environments (high versus low) have differential effects on the biomass of aquatic invertebrates and algae. Killifish adapted to the presence or absence of guppies were also found to have differential effects on aquatic invertebrates. These results suggest the possible importance of top-down drivers of eco-evolutionary dynamics.
Post & Palkovacs (2009)
provide a capstone by specifically considering the conditions under which eco-evolutionary feedbacks are most likely. They then review five empirical systems in which indirect evidence suggests such feedbacks are present (). Each of these systems shows considerable promise for documenting eco-evolutionary dynamics in nature. It also seems certain that new systems will soon emerge as excellent candidates for such kind of study.