Project topics and materials
Get Complete Project Material File(s) Now! »
Community assembly: from genes to communities
Community assembly rules (Diamond 1975) are a set of debatable rules in ecology (Rule 1: Forbidden species combinations, and Rule 2: Reduce niche overlap, Diamond 1975). These rules highlight the role of competition10 in determining the patterns of assemblage composition. Keddy (1992) proposed that the role of community assembly rules is not limited only to the assembly of species within a community, but integrates all factors responsible for the occurrence of a species within the community. He states that: “The process of constructing communities from species pools is in many ways analogous to the processes of evolution through natural selection. Habitats serve as filters for genotypes, with the least suited genotypes being filtered out, and the best suited surviving to reproduce. In the case of assembly rules, habitats are again serving as filters. However, in this case, the filters operate on traits11 and eliminate those sets of traits which are unsuitable to that environment. The species which comprise the community are those which survive the filter”. Thus, species (or individuals) face different filters (stochastic, abiotic and biotic) before being able to participate to the final composition of a community (Figure 5). The stochastic filter acts on the initial species pool to form a regional then a local species pool; here, aleatory processes define species presence. The local species pool is than simultaneously subjected to the abiotic and biotic filters. The abiotic filter (i.e. environmental factors: temperature, soil water content, salinity…) acts on a regional and local scale. Species that tolerate the regional or local abiotic conditions can cross this filter, and ecologically similar species are filtered within the same fundamental niche (Lavorel & Garnier 2002). Thus, the abiotic filter acts in a way to reduce trait variability between species – trait convergence. The biotic filter (i.e. biotic interactions) acts on the individual scale in a way to eliminate similar species, and thus participate to the increase in trait variability between the species of a community – trait divergence (Weiher et al. 1998). This filter incorporates all kinds of interactions between two species (e.g. competition, facilitation, allelopathy12, herbivory, predation, pollination…).
Figure 5. The main filters that structure a plant community. Each species is represented by a different geometrical form and color. Species cross simultaneously (from outside to inside) the stochastic filter, the abiotic filter and the biotic filter. The spatial scale, at which each of the filter operate, decreases from the initial species pool to the final community composition (adapted from Keddy 1992; Zobel 1997; Díaz et al. 1999; Lortie et al. 2004)
However, Pärtel et al. (2011) argued that the term “species pool” should only be used to refer to species that can potentially occupy a particular habitat due to suitable local ecological conditions. They also emphasised that studies comparing species diversity of different ecosystems, regions or taxonomic groups should consider not only the observed local diversity, but also the “dark diversity” (i.e. species that are currently absent from a site but which belong to its species pool). Recently, de Bello et al. (2012) re-evaluated the role of biotic processes in generating trait divergence between the species of a community. They showed that biotic processes such as competition could lead to both trait divergence (through the exclusion of similar species – niche differentiation) and convergence (through exclusion of dissimilar species – weaker competitor exclusion). Thus abiotic and biotic processes can produce similar patterns of traits diversity, and separating them cannot be done by comparing the trait diversity observed within communities to patterns of randomly generated communities based on sampling within a region. Instead, de Bello et al. (2012) proposed a framework, the “functional species pool”, in which they separated abiotic and biotic processes and distinguished opposing biotic effects (convergence and divergences) on community assembly (see Figure 6). A valuable point in their approach is the incorporation of the dark diversity within the species pool.
Figure 6. Given a particular regional flora or fauna, that was formed according to different geographical and historical filters (1), habitat selection will filter out species whose environmental preference falls outside the range of environmental conditions available in a given site, thus creating a convergence in trait values (e.g., compare trait ranges – red dashed arrows – between 1 and 2). This convergence is removed when using the functional species pool approach. The test for assessing the relevance of biotic interactions on community assembly in each single site (3) is performed by comparing the functional diversity13 in the sampled community (FDcomm) with the functional diversity expected within the corresponding functional species pool (FDpool; species that potentially coexist – filled grey circles). The deviance from a 1:1 relationship (i.e., random assembly14) between FDcomm and FDpool corresponds to prevailing biotic assembly processes (FDcomm > FDpool biotic divergence vs. FDcomm < FDpool indicating biotic convergence). Adapted from de Bello et al. 2012.
Overview on biotic interactions and abiotic factors (stress and disturbance)
For long time, ecologists have tried to come up with a generalised principle that holds across the natural world (Graham & Duda 2011). In their work on island biogeography, MacArthur & Wislon (1967) proposed one of the first models, the r/K selection model, seeking to explain and predict species distribution (initially designed to be applicable to all living beings). The r/K model was based on environmental stability to predict species selection. The r-selected species (or r-strategists) do best in unpredictable/disturbed environments; they are characterised by a rapid growth, early maturity, but poor competitive ability. The K-selected species (or K-strategists) do well in more predictable/stable environments; they are characterised by a slow growth rate (long life-span), late maturity, but high competitive ability. Although drawbacks of the r/K theory have been pointed out (Wilbur et al. 1974; Barbault 1987; Kuno 1991), it is still widely used as “enough people have found it a useful framework in which to interpret their observations [and thus,] it must contain an element of truth” (Stearns 1992). In particular, this model is well adapted for understanding tree species functional strategies and forest successions, in particular in benign physical conditions (Michalet et al. 2008). Also, Michalet et al. (2011) used this model to contrast the two different phenotypes of the alpine foundation species Geum rossii in northern Arizona (USA). Later on, more detailed models emerged and included gradients of stress15 and disturbance16.
Early in the seventies, Grime (1973) presented a model showing a unimodal relationship between species density17 and the intensity of stress (and site productivity) or disturbance (e.g. grazing, burning, flooding), commonly known as the “humped-back model” (Figure 7). Later on, this model was considered to be universal by plant ecologists (but see Adler et al. 2011; Fridley et al. 2012). The principle of this model is that in stable conditions (in absence of disturbance) species diversity is low in most productive (resource-rich) environments where stress is low and competition is high due to the abundance of competitive (tall fast-growing) species. In contrast, on the other side of the hump, with the increase in stressful conditions species density decreases due to the decrease in productivity as environmental conditions become too harsh even for stress-tolerant species to persist (left curve in Figure 7). When considering disturbance only in productive communities (without stress), a similar pattern is observed. At low disturbance, species density is low due to competitive exclusion by the abundant competitive species. With the increase of disturbance, ruderal species (i.e. fast-growing short-lived species), which are well adapted to disturbance (Grime 1974), progressively replace competitive species that poorly tolerate disturbance. Species density reaches its highest value at intermediate levels of disturbance, where competitive and ruderal species co-occur. This is similar to the “Intermediate Disturbance Hypothesis” (Connell 1978). At high levels of disturbance, species density decreases as the abiotic constraint become extreme for any species to exist.
The universality of Grime’s “humped-back model” was held true, until Waide et al. (1999) doubted of its scale-dependency, which was later confirmed by Mittelbach et al. (2001) who, however, found it to be the dominant pattern for plants, especially at local to landscape scales. Subsequently, Gillman & Wright (2006) performed a survey on 159 productivity-plant species richness relationships from 131 published studies and concluded that positive relationships were the exclusive form of relationships at continental to global extents, and that unimodal (humped-back) relationships were more likely to occur at small spatial (local) scales. Moreover, Pärtel et al. (2007) argue that the productivity-diversity humped-back relationship is not universal, as it is valid in temperate regions but not in tropical ones where positive relationships are more common. Later on, based on an intercontinental data set (from 48 herbaceous-dominated plant communities on five continents), Adler et al. (2011) challenged the concept of the humped-back model of plant diversity, doubting of its utility by showing no consistent general relationship between productivity and species richness at local, regional or global scales. However, Adler et al.’s work was criticised by Fridley et al. (2012) who showed that the data used was not exactly representative, mostly because it lacked sufficient high-productivity sites and excluded anthropogenic sites for no scientific reasons; by including high-productivity sites (e.g. salt marshes, herbaceous floodplains…), Adler et al.’s data would have revealed a pattern consistent with the humped-back model (Fridley et al. 2012), i.e. a decrease in richness at high productivity levels.
Facilitation: a long-time forgotten interaction due to the predominance of competition
Until the mid-nineties, ecological theories and models have considered only negative interactions (e.g. Grime 1973; Connell 1978; Huston 1979; Tilman 1980, 1982), even though positive interactions (i.e. facilitation) have been reported in experimental studies [Niering et al. 1963 and Turner et al. 1966 in Callaway 2007; Hunter & Aarssen 1988] and ecological theories (Clements 1916). This is because negative interactions (e.g. competition or interference) were thought to be the main biotic filter structuring plant communities (Goldberg & Barton 1992). However, two schools of thoughts regarding the strength of competition along productivity gradients [in plant community ecology] emerged over the time. This divergence in thoughts was known as the “Grime-Tilman debate”. Grime considered that [aboveground] competition decreases from high to low levels of productivity, and interactions vanish under stressful conditions (Grime 1973, 1977). In contrast, Tilman founded his “resource-ratio” theory by arguing that when productivity decreases, competition for limiting resources switches from aboveground to belowground, and thus competition is held constant (Tilman 1980, 1982). Both theories have gained significant attention in the field of plant ecology. Grace (1991) argues that the “Grime-Tilman debate” is due to differences in the definitions of some terms used by each of these authors (e.g. ‘competition’, for Grime, it is the capacity for resource capture and the mechanism by which a plant suppresses the fitness of a neighbour; for Tilman, competitive success is the ability to draw resources to a low level and to tolerate those low levels – to have a low equilibrium resource requirement). Welden & Slauson (1986) tried resolving this ‘debate’ by clarifying the difference between the intensity and the importance of competition (see Box 1 for the definition of ‘competition intensity’ and ‘competition importance’). Competition importance has been proposed to explain Grime’s (1973) theory on competition, whereas intensity explains Tilman’s theory – competition intensity stays constant along the productivity gradient, but switches from aboveground to belowground in unproductive environments – (Welden & Slauson 1986; Grace 1991).
Until recently, negative interaction have been the primary concern of studies in plant community genetics (e.g. Whitlock et al. 2007; Lankau & Strauss 2007; Johnson et al. 2008; Bossdorf et al. 2009; Silvertown et al. 2009; Genung et al. 2011). However, [as already said] facilitation has been found in theoretical and experimental studies (Clements 1916; Hunter & Aarssen 1988). For instance, Clements (1916) argues that plants themselves cause succession to occur by improving site factors (e.g. light capture by leaves, production of detritus, water and nutrient uptake, nitrogen fixation), which allows the establishment of plants of the next succession stage. This means that plants of one stage directly ‘facilitate’ plants of the next succession stage. Though, the little attention given to facilitation and the predominance of competition for a long time in research fields such as ecology is likely because facilitation could go undetected, as it appears weaker than competitive mechanisms (Gross 2008).
Box 1. ‘Competition intensity’ versus ‘Competition importance’ Welden & Slauson (1986) state that: “The intensity of competition is a physiological concept, related to the well-being of individual organisms but only indirectly and conditionally to their fitness, and even more indirectly to the evolution of populations and the structure of communities. The importance of competition is primarily an ecological and evolutionary concept, related directly to the ecology and fitness of individuals but only indirectly to their physiological states. The intensity of competition is not necessarily correlated with the intensities of predation, disturbance, abiotic stress, or other ecological processes. The importance of competition is necessarily relative to the importances of other processes. Intensity refers primarily to the processes of present competition, whereas importance refers primarily to the products of past competition”.
In other words:
Intensity refers to the absolute impact of neighbouring plants on a target plant (negative for competition and positive for facilitation).
Importance is the contribution of biotic interactions relative to other environmental processes such as stress and disturbance (sensu Grime 1973) to the change in the performance of a target plant.
Nevertheless, interest in facilitation in ecological studies has significantly increased over the past two decades, leading Bertness & Callaway (1994) to come up with the Stress Gradient Hypothesis (SGH; Figure 8). The SGH predicts that the outcome of plant-plant interactions extends from competition at intermediate levels of stress and disturbance (under favourable conditions) to facilitation at both extremes of these two gradients (Bertness & Callaway 1994; Brooker & Callaghan 1998). At high levels of stress, facilitation is due to habitat amelioration by stress-tolerant species leading to the extension of the realised niche of [stress-intolerant] competitive species (Bruno et al. 2003). At high levels of disturbance, facilitation is indirect by the means of associational defences (e.g. associational defences against herbivores, Rousset & Lepart 2000; Milchunas & Noy-Meir 2002; Rebollo et al. 2002; Baraza et al. 2006; Smit et al. 2007). Because of the differences in the methods followed (observational versus experimental; Maestre et al. 2005) and the complexity added by factors like the variation in the nature (resource versus non-resource) and the length of the gradients considered (Maestre et al. 2005; Lortie & Callaway 2006; Brooker et al. 2008), the chosen estimator of performance (Maestre et al. 2005; Brooker et al. 2008; Gómez-Aparicio et al. 2008) and the diverse characteristics (chemical and physical, Baraza et al. 2006) and strategies (competitive versus stress-tolerant) of both the nurse18 and the beneficiary19 species involved (Liancourt et al. 2005; Michalet 2007; Maestre et al. 2009), the outcomes of plant-plant interaction studies in the literature were ambiguous and conflicting. Many experimental studies have supported the predictions of the SGH (Callaway et al. 2002; Liancourt et al. 2005; Schiffers & Tielbörger 2006) while others contradicted it and found competition to be important under high stress levels (Maestre & Cortina 2004).
Figure 8. The Stress Gradient Hypothesis (SGH): the shift in biotic interactions along stress and disturbance gradients. Facilitation increases by neighbourhood habitat amelioration with the increase in stressful conditions (in red), and by associational defences with the increase in physical disturbance (in blue; adapted from Bertness & Callaway 1994).
Hacker & Gaines (1997) suggested that at intermediate levels of stress and disturbance, facilitator species that might normally be competitively excluded are released from competition and therefore enhance species diversity from intermediate to very high levels of stress and disturbance. Later on, Michalet et al. (2006) presented a revision of Grime’s humped-back model by integrating facilitation to it (Figure 9; for models integrating facilitation to ecological theory, also see Bruno et al. 2003 and Lortie et al. 2004). According to them, the model remains unchanged in productive environments where diversity is low due to competitive exclusion (part A1 in Figure 9). In conditions of low environmental severity (part A2 in Figure 9), competition is gradually replaced by facilitation, thus increasing diversity by expanding the realized niche of stress-intolerant competitive (stress-intolerant) species. Diversity reaches its maximum at intermediate levels of environmental severity where species of the three strategy types (competitive, ruderal and stress-tolerant species) co-occur. From intermediate to high environmental severity levels (part B1 in Figure 9), facilitation decreases for competitive and then for stress-tolerant species, thus decreasing diversity. At very high environmental severity levels, facilitation ‘collapses’ as environmental conditions become too harsh for the ‘nursing’ plants to facilitate other plants (due to a decrease in the size of the nurse plants; see also Forey et al. 2010; Le Bagousse-Pinguet et al. 2013; Michalet et al. 2014a). Additionally, some authors proposed that at the most severe conditions, all interactions (positive and negative) collapse as both competitive and facilitative species are weakened (Malkinson & Tilbörger 2010), consistent with the results of Maalouf et al. (2012).
The inclusion of positive interactions into ecological theories proved that facilitative interactions also have strong effects on community and ecosystem properties, including structure, productivity and stability (Mulder et al. 2001; Michalet et al. 2006; Callaway 2007; Brooker et al. 2008; Butterfield et al. 2013; Le Bagousse-Pinguet et al. 2014a), specifically in severe environments. Thus, it is not surprising to see the growing number of studies of facilitative interactions among plants in an evolutionary perspective (Valiente-Banuet et al. 2006; Crutsinger et al. 2010, 2013 ; Liancourt & Tielbörger 2011; Michalet et al. 2011; Thorpe et al. 2011; Butterfield et al. 2013; Bailey et al. 2014; Le Bagousse-Pinguet et al. 2014a).
Table of contents :
CHAPTER ONE: LITERATURE OVERVIEW
1- Bridging ‘community ecology’ to ‘evolutionary biology’: the emergence of ‘community genetics’
2- Community assembly: from genes to communities
3- Evolution, natural selection and adaptation
4- Thesis objectives
CHAPTER TWO: STUDY SITES, MODEL SPECIES AND EXPERIMENTAL DESIGNS
1- The study site in the Pyrenees, the model species (Festuca gautieri subsp. scoparia Hackel & Kerner), and the experimental designs
2- The study site in the Mount-Lebanon, the model species (Onobrychis cornuta (L.) Desv.) and the experimental design
CHAPTER THREE: Phenotypic differentiation within a foundation grass species correlates with species richness in a subalpine community
CHAPTER FOUR: Disentangling the heritable and plastic components of the competitive and facilitative effects of two contrasting phenotypes of a foundation species
CHAPTER FIVE: Differential effects of contrasting phenotypes of a foundation legume shrub drive plant-plant interactions in a Mediterranean mountain
CHAPTER SIX: SYNTHESIS
1- Determinism of the observed phenotypic variation: the contribution of genetic variation and phenotypic plasticity
2- Consequences of the phenotypic variation within foundation species on the subordinate species
3- Perspectives for future studies
REFERENCES
