Niche divergence among closely related species reflects both the selective forces underlying species divergence and the ecological mechanisms structuring current species distributions. Niche partitioning is driven by trade-offs in resource use that prevent the evolution of a single universal niche; such trade-offs are integral components of theory regarding distribution limits (e.g., Kirkpatrick and Barton 1997), ecological speciation (e.g., Rundle and Nosil 2005), and species coexistence (e.g., Tilman 2004). Trade-offs between adaptations to species interactions and other environmental conditions may be particularly important across environmental or resource gradients, where the relative importance of physiological tolerance and competitive ability may vary (Grime 1977; Gaucherand et al. 2006; Liancourt and Tielbörger 2009). Despite long-standing interest in the divergence, distribution, and coexistence of closely related species, a recent review (Sexton et al. 2009) notes an absence of experimental field tests of the relative strengths of species interactions and adaptation to other environmental conditions in setting niche boundaries between recently diverged species. Quantifying the effects of habitat and congeneric interactions in locally sympatric and allopatric habitats allows tests for trade-offs in their relative importance for the divergence and coexistence of closely related species.
Traditionally, species boundaries are generally thought to be maintained by species interactions in productive habitats and physiological limits in stressful habitats (e.g., Connell 1961; Gross and Price 2000). The realized niche is thus conceived as a sub-space of the fundamental niche (Hutchinson 1957); when abiotic mechanisms fail to explain persistent niche boundaries, biological mechanisms such as competition are invoked (e.g., Wethey 1983; Gross and Price 2000; Ettinger et al. 2011). Yet in stressful habitats, facilitative interactions may expand the realized niche to encompass a greater range of environmental conditions (Callaway 1995). Under the Stress Gradient Hypothesis (SGH), species interactions are predicted to switch from primarily facilitative in stressful habitats to primarily competitive in productive habitats (Bertness and Callaway 1994; Maestre et al. 2009). Support for the SGH has been found within communities (e.g., Choler et al. 2001; Liancourt et al. 2005), yet its importance for explaining evolutionary distribution limits (i.e., between closely related species) across environments has to our knowledge not yet been tested.
Niche divergence between closely related taxa across environmental gradients may reflect trade-offs between competitive ability and stress tolerance (Grime 1977; Liancourt et al. 2005; Liancourt and Tielbörger 2009; but see Emery et al. 2001). For example, rapid development in annual plant species may allow stress avoidance, but reduce competitive ability, whereas perennial species exhibit greater vegetative growth that may confer increased competitive ability, but decrease stress tolerance or avoidance (e.g., Tercek and Whitbeck 2004; Cui et al. 2011). Consequently, the strength and sign of species interactions may depend on both the environment in which they are measured and specific functional traits of the interacting taxa (Goldberg 1996).
Even if closely related species exhibit divergence across certain niche axes, other axes may be conserved (Holt 2009). Niche conservatism limits the coexistence of closely related species (MacArthur and Levins 1967). Alternatively, speciation may be associated with niche divergence (e.g., Evans Margaret et al. 2009), potentially reducing competition between close relatives in particular environments. The relative importance of competition with close or distant relatives in determining species distributions remains practically untested (but see Burns and Strauss 2011), and may be habitat-dependent.
When habitat partitioning occurs within the scale of dispersal, species coexistence in locally sympatric habitats may be maintained by either fluctuation-dependent or fluctuation-independent mechanisms (Chesson 2000). Fluctuation-independent mechanisms require that each species exhibits stronger intraspecific competition than interspecific competition independently of environmental variation (e.g., Tilman 1982). Alternatively, fluctuation-dependent mechanisms such as the storage effect rely on covariance between environmental conditions and the strength of competition, allowing intraspecific interactions to be spatially or temporally concentrated relative to interspecific interactions (Chesson and Warner 1981). For example, spatial variation allows positive population growth in favorable habitats to both concentrate intraspecific competition and buffer the effects of unfavorable habitats (Sears and Chesson 2007). In this case, dispersal from competitive refuges into nearby sympatric habitats may maintain species coexistence by preventing competitive exclusion (Amarasekare 2003).
Here, we use two recently diverged plant species that exhibit persistent habitat partitioning to examine the mechanisms and trade-offs underlying niche divergence and local patterns of allopatry and sympatry. Mimulus guttatus thrives in perennial streams, whereas its close relative Mimulus laciniatus occupies nearby fast-drying seeps; either species is absent from the above habitat occupied by its congener (Fig. 1). Both species co-occur in meadow habitats with intermediate water availability. Habitat partitioning in this system occurs within the scale of dispersal, providing an opportunity to examine the mechanisms underlying persistent niche boundaries. We quantify habitat-specific responses and inter- and intraspecific interactions to test whether divergent adaptations to each habitat type, congeneric species interactions, or both are sufficient to explain current patterns of habitat partitioning. We compare the relative importance of habitat type and congeneric interactions in each habitat and for each species to test predictions based on habitat characteristics and life-history trade-offs. Although we focus on habitat adaptation and congeneric interactions in this study, other mechanisms may also contribute to habitat partitioning in this system. For example, interactions with other species in the local community, dispersal among habitats, and/or temporal variation may influence patterns of allopatry and sympatry.