sign in sign up
<<
Home , Ecological fitting, Exaptation

Letters, (2008) 11: 1123–1134 doi: 10.1111/j.1461-0248.2008.01237.x IDEA AND PERSPECTIVE Ecological fitting by phenotypically flexible genotypes: implications for species associations, assembly and

Abstract Salvatore J. Agosta* and Ecological fitting is the process whereby organisms colonize and persist in novel Jeffrey A. Klemens environments, use novel resources or form novel associations with other species as a Department of Biology, result of the suites of traits that they carry at the time they encounter the novel condition. University of Pennsylvania, This paper has four major aims. First, we review the original concept of ecological fitting Philadelphia, PA 19014, USA and relate it to the concept of and current ideas on the positive role of *Correspondence: E-mail: in evolution. Second, we propose phenotypic plasticity, correlated [email protected] trait evolution and phylogenetic conservatism as specific mechanisms behind ecological fitting. Third, we attempt to operationalize the concept of ecological fitting by providing explicit definitions for terms. From these definitions, we propose a simple conceptual model of ecological fitting. Using this model, we demonstrate the differences and similarities between ecological fitting and ecological resource tracking and illustrate the process in the context of species colonizing new areas and forming novel associations with other species. Finally, we discuss how ecological fitting can be both a precursor to evolutionary diversity or maintainer of evolutionary stasis, depending on conditions. We conclude that ecological fitting is an important concept for understanding topics ranging from the assembly of ecological communities and species associations, to biological invasions, to the evolution of biodiversity.

Keywords , , biological invasion, , community ecology, exaptation, space, host shift, operative environment, pre-adaptation, resource tracking.

Ecology Letters (2008) 11: 1123–1134

The concept of ecological fitting developed within the INTRODUCTION historical context of concerns about what Janzen (1980) Janzen (1985) coined the term Ôecological fittingÕ to describe and others (e.g. Holmes & Price 1980; Brooks 1985) the situation in which an organism interacts with its biotic perceived as the overuse of coevolutionary arguments to and abiotic environment in a way that appears to indicate a explain associations among species (Agosta 2006 and shared evolutionary history, when in fact the organismal references therein). One of JanzenÕs main concerns was traits relevant to the interaction evolved elsewhere and in that, when cases of ecological fitting occur, it will be very response to a different set of environmental conditions. difficult to distinguish them from cases of long-term Ecological fitting was presented as a contrasting view to, coexistence because the essential biological result, coexis- and as an appropriate null hypothesis for, the assumption tence and direct or diffuse interaction, is the same. Without that currently observed associations among organisms are an understanding of ecological fitting, biologists, naive to evidence of shared evolutionary history or, more generally, the true histories of organisms present in a community, as a response to explicitly adaptationist arguments to explain would be encouraged to invent spurious adaptive or the presence of a or species in a particular coevolutionary scenarios to describe interactions for which environment. they are not needed.

2008 Blackwell Publishing Ltd/CNRS 1124 S. J. Agosta and J. A. Klemens Idea and Perspective

Ecological fitting was initially recognized by Janzen establish populations under novel conditions beyond those (1985), and subsequently discussed by Brooks & McLennan conditions encountered in their previous evolutionary (2002), as being closely related to the concept of exaptation history, circumvents adaptive processes to produce novel (Williams 1966; Gould & Vrba 1982). The process by which ecological interactions between organisms and the environ- an existing trait is co-opted for a different ment. These novel ecological interactions can then provide (exaptation) is fundamentally the same as that by which an novel selective environments (e.g. novel species associa- existing genotype obtains realized fitness in a novel selective tions) on which can work. environment (ecological fitting). Brooks & McLennan For those familiar with West-EberhardÕs (2003) recent (2002) noted that the frequency with which ecological work, Developmental Plasticity and Evolution, the preceding fitting occurs in nature depends in part on the ability of paragraph should be reminiscent of what some have termed traits to be co-opted for novel functions. Below we will the Ôplasticity theoryÕ of biodiversity (Janz et al. 2006; argue that the process of ecological fitting is essentially the Weingartner et al. 2006; Nylin & Wahlberg 2008). Building process of exaptation taking place on a shorter time scale on a history of ideas on the positive role of phenotypic than that over which it is normally considered. Or, in other plasticity in evolution (e.g. the ÔBaldwin effectÕ: Baldwin words, that ecological fitting is the ecological case of 1896; Robinson & Dukas 1999), West-Eberhard argues (1) exaptation. that genotypes are inherently phenotypically plastic, (2) that Subsequent to Janzen (1985), ecological fitting received this plasticity can allow genotypes (individuals) to obtain little attention in the ecology and realized fitness under novel environmental conditions and, literature even as the essence of the concept continued to be therefore, (3) that Ô…the origin and evolution of adaptive implied by biologists studying species invasions and novelty do not await ; on the contrary, genes are introductions (e.g. Holway et al. 2002; Yeh & Price 2004; followers, not leaders, in evolutionÕ (p. 20). This view has Suarez et al. 2005; Strauss et al. 2006), biogeographers been met with some opposition. de Jong (2005) argues that debating the relative roles of dispersal and colonization vs. West-EberhardÕs (2003) proposed process of genetic assim- vicariance (e.g. Pennington & Dick 2004), and evolutionary ilation– which describes trait evolution resulting from the biologists in discussing adaptation, Ôpre-adaptationÕ and initial exposure of novel phenotypic variants arising from exaptation (see below). The term ecological fitting has developmental plasticity to novel conditions – is (1) not survived, however, and has been used by researchers unique within the Darwinian synthesis if one considers interested in the factors structuring species associations plasticity itself an evolved quantitative genetic trait and (2) (Gill 1987; White & Stiles 1992; Chenuil & McKey 1996; not a major avenue for trait evolution based on current Flowers & Janzen 1997; Yu & Davidson 1997; Brooks & empirical or model support. McLennan 2002; Janzen 2003; Agosta & Janzen 2005; We leave the reader to delve further into these arguments. Agosta 2006; Brooks et al. 2006), whole ecosystems For our purposes, we find the above issues to be largely (Wilkinson 2004) and even the dynamics of emerging immaterial to West-EberhardÕs broader contribution: that infectious diseases (Brooks & Ferrao 2005). Nonetheless, phenotypic plasticity, whether evolved or a developmental there has been little attempt to operationalize the concept of by-product, can allow existing genotypes to obtain fitness ecological fitting and incorporate it into mainstream and therefore persist in novel environments without ecological and evolutionary theory. This is unfortunate awaiting novel , thereby placing existing genotypes because ecological fitting has considerable explanatory into novel selective environments where natural selection power (Brooks & McLennan 2002; Wilkinson 2004; Brooks can potentially act. This is a fundamentally different view of & Ferrao 2005; Agosta 2006; Brooks et al. 2006) and is a phenotypic plasticity compared to its historical role in natural null hypothesis for a range of research programmes evolutionary thought. In this formulation, phenotypic (e.g. any prediction of organismal form or function derived plasticity provides ÔfodderÕ for evolution, rather than being from optimality theory; Agosta 2006). merely the environmental noise that is selected against by However, we argue that ecological fitting has much stabilizing selection or that drags against the efficacy of greater importance than simply acting as a null hypothesis or directional selection (Stearns 1989; Thompson 1991). null explanation in ecology and evolutionary biology. We The connections between recent arguments for a positive argue as did Janzen (1985) that ecological fitting is an role of phenotypic plasticity in evolution (West-Eberhard inevitable and frequent process in nature that results from 1989, 2003; Robinson & Dukas 1999; Gorur 2004; Yeh & the interaction between highly flexible organisms and highly Price 2004; Fordyce 2006; Janz et al. 2006; Weingartner et al. variable biotic and abiotic environments. In what follows, 2006; Nylin & Wahlberg 2008) and JanzenÕs (1985) concept we develop a framework within which to evaluate this of ecological fitting are tangible and ripe for synthesis assertion. We will also show how the process of ecological (Agosta 2006). This paper has four aims. First, we describe fitting, whereby organisms obtain realized fitness and the concept of ecological fitting and outline its basic

2008 Blackwell Publishing Ltd/CNRS Idea and Perspective Ecological fitting 1125 premises. Second, we advance phenotypic plasticity, corre- The third premise is that some species possess Ôrobust lated trait evolution, and phylogenetic conservatism as the genotypesÕ that allow persistence across varied environ- primary mechanisms that allow ecological fitting. Third, we ments without adaptation to local conditions or to the new develop operational definitions for discussing ecological species with which it interacts in the expanded range. Janzen fitting as a process by explicitly defining the terms posited that many widespread species are in an Ôevolutionary ÔenvironmentÕ and Ôspecies traitsÕ and the Ôenvironment-by- quiescentÕ stage of a cyclical pattern that consists of a small species traitsÕ interaction. Fourth, we derive a simple population occupying an initially small area, mass selection graphical model from these definitions to illustrate the on this small population and abrupt expansion of the process of ecological fitting, how it can lead to patterns species range when some particularly Ôrobust genotypeÕ is superficially indistinguishable from evolution, and how it stumbled upon in the ancestral environment. In this stage, can, depending on circumstance, be both a promoter of the species is widespread, evolutionarily static and insensi- evolutionary stasis and a precursor to evolutionary diversity. tive to local selective regimes because it is comprised of many populations under myriad selective pressures that Ô…are fine-scale, heterogeneous and contradictory…Õ (Janzen DEVELOPMENT OF THE CONCEPT OF ECOLOGICAL 1985:308). FITTING Condition (3) is not as obvious as conditions (1) and (2), JanzenÕs (1985) original formulation outlined three basic and requires consideration of two questions. The first is premises of ecological fitting. First, the original, or ancestral, whether it is plausible, based on what we know about the range of many species is smaller than the range they biology and biogeography of species, that species expand currently occupy. Second, communities have porous bor- rapidly into new ranges in the absence of local adaptation. ders such that many of the species that comprise ecological The second, and the subject of the next section of this communities originated elsewhere and immigrated to sites paper, is to ask whether there exist mechanisms that can where they currently exist. Third, many species possess account for the development of JanzenÕs Ôrobust genotypeÕ. Ôrobust genotypesÕ that allow colonization and persistence in One test of the plausibility of condition (3) has already this larger range without evolutionary change prior to range been performed. Human-mediated dispersal of species can expansion. be considered an extreme relaxation of dispersal limitation, results from a sufficient amount of reproduc- and the success of invasive species is evidence that many tive isolation of one or more gene pools (Mayr 1963; taxa easily persist in novel environmental conditions and Futuyma & Mayer 1980), and thus condition (1) is probably with an evolutionarily unfamiliar flora and fauna. The true for many if not most species (e.g. Leigh et al. 2004). It Ôescape hypothesisÕ, whereby invasive species are released will obtain whenever the initial population is a relatively from the constraints imposed by coevolved parasites and small geographical or ecological isolate that evolves in pathogens (Mitchell & Power 2003; Torchin et al. 2003), response to the cast of biotic and abiotic characters in its provides a particularly instructive example. Although the ancestral environment. escape hypothesis is often presented so as to emphasize the Condition (2) is likewise uncontroversial in and of itself, absence of species that have negative effects on the invasive and is equivalent to saying that dispersal and colonization species, this is not a sufficient explanation for biological play important roles in biogeography and community invasions. Implicit in the escape hypothesis is that, following assembly (MacArthur & Wilson 1967; Schluter & Ricklefs release, invaders interact with entirely novel assemblages of 1993; Hurtt & Pacala 1995; Hubbell 2001; Brooks & competitors, pollinators, dispersers, predators and prey, yet McLennan 2002; Pennington & Dick 2004; Urban et al. do not seem to be hindered by a lack of past evolutionary 2008). Together, conditions (1) and (2) imply that the range history with these organisms. Release from parasites and of environmental variables and the number and identity of pathogens alone cannot ensure success in a new habitat. The interacting species in the ancestral environment will be a robust genotype must maintain a broad ability to function subset of or different from the environments experienced when the identities of all other species in the community and species encountered following range expansion. This, in have changed. turn, leads to the conclusion that ecological communities are Perhaps even more germane is MackÕs (2003) discussion comprised of some proportion of relatively recent arrivals of absent life forms and community invasibility. Mack notes with little ÔdeepÕ in situ evolutionary history. These species that due to phylogenetic and biogeographic constraints, carry with them a historical legacy of traits that evolved natural communities do not contain the full range of life elsewhere, and will interact with many other species with forms observed globally. He goes on to describe a number similarly varied histories (Brooks 1985; Janzen 1985, 1986, of cases, particularly invasions of grasslands by woody 2003; Brooks & McLennan 2002; Agosta & Janzen 2005; species, where a life form previously excluded from a habitat Brooks & Ferrao 2005; Agosta 2006; Brooks et al. 2006). by dispersal limitation, in the broad biogeographic sense,

2008 Blackwell Publishing Ltd/CNRS 1126 S. J. Agosta and J. A. Klemens Idea and Perspective becomes invasive or even dominant in a new environment, (2003) and are beyond the scope of this paper. We will simply indicating that there was never an intrinsic biological barrier note that they include the phenomena of hypervariability to the speciesÕ presence in the system. For example, he asks followed by somatic selection, homeostatic , if South American grasslands are inherently inhospitable to hormonal regulation, learning and the tendency of the woody species, or if the absence of pines (Pinus spp.) and modular nature of organisms to produce independence in other woody species adapted to xerophytic conditions form and function of otherwise integrated parts. within the South American biogeographic realm has resulted Regardless of their particular causes, the fact that in grasslands persisting in habitats that would be quickly organisms do possess plastic provides an colonized by woody species in conifer-rich North America. appropriate level of mechanism for understanding ecological In fact, these grasslands are currently undergoing rapid fitting as a process. A central insight from plasticity studies, invasion by introduced pine species (Mack 2003; see also discussed in great detail by West-Eberhard (2003; see also Stohlgren et al. 2008); a forceful demonstration of ecological Robinson & Dukas 1999), is that phenotypic plasticity fitting in practice. allows organisms greater flexibility in the use of novel One consequence of release from dispersal limitation and resources than if mutations and changes in gene frequencies invasion of new habitats is that it adds novel members to were a pre-requisite for their exploitation. Phenotypic ecological communities, which introduces novel genetic plasticity allows organisms to mount a response (i.e. achieve variation, fosters novel ecological interactions and can lead realized fitness) to novel environmental conditions (e.g. Yeh to effects that reverberate throughout the community & Price 2004), and it seems likely that all organisms possess (Urban et al. 2008; Whitham et al. 2008). Cases of the some degree of potential fitness outside the range of incorporation of novel host plants into the diets of conditions under which the species evolved (Fig. 1). We call phytophagous insects provide particularly instructive exam- this region of ecological space Ôsloppy fitness spaceÕ ples of how easily these novel ecological interactions can (Fig. 2a), a key element of the model of ecological fitting form in the first place (Agosta 2006). For instance, Thomas presented in later sections. Interestingly, the existence of et al. (1987) documented the incorporation of a recently Ôsloppy fitness spaceÕ appears to be a prediction of classic acquired host plant into the diet of several populations of quantitative genetic models of optimizing selection on the butterfly Euphydryas editha in California, USA. Females reaction norms (de Jong 2005). normally oviposit on the native plant Collinsia parviflora Although phenotypic plasticity alone would presumably (Scrophulariaceae), but some populations also use Plantago lead to the development of robust genotypes, the correlated lanceolata (Plantaginaceae), which was introduced from evolution of traits (Lande & Arnold 1983) can also be Europe in the last 100 years. Populations of E. editha expected to produce organisms that possess the ability to previously unexposed to P. lanceolata showed evidence that perform in novel environments. For example, Herrera et al. at least some females were Ôpre-adaptatedÕ to oviposit on (2002) showed evidence in the plant Helleborus foetidus for the this novel host. Furthermore, larvae from these populations correlated evolution of pollinator- and -related were able to survive and develop on the novel host. The traits, such that selection by mutualistic pollinators can inference was that, assuming in the study was indirectly affect the plantÕs response to antagonistic herbi- negligible, E. editha populations currently using P. lanceolata vores. Quite simply, if direct selection on trait A causes a initially required no evolutionary change to incorporate the correlated change in trait B, then this could lead to a novel host into their diet. In our terms, E. editha could be phenotype that is somehow Ôpre-adaptatedÕ for some future, said to have possessed a relatively robust genotype that novel environmental condition. allowed use of a novel resource for survival and reproduc- A third mechanism behind the existence of sloppy fitness tion at the moment of contact. space and robust genotypes is simply latent resulting from the conservation of genetic information within a phylogeny. Phylogenetic conservatism, inertia or MECHANISMS constraint of traits related to resource use can, for example, Having established that something akin to JanzenÕs Ôrobust allow parasites to track host resources across taxa and shift genotypesÕ occur in nature, what mechanisms might give rise to novel hosts that possess similar resources (Brooks & to this phenomenon? Phenotypic plasticity, a seemingly McLennan 2002; Murphy & Feeny 2006). The reason, for universal property of organisms (Bradshaw 1965; West- example, an individual parasite may recognize a seemingly Eberhard 1989, 2003; Schlichting & Pigliucci 1998), probably ÔnovelÕ host is because some ancestral species may have long plays a primary role (Fig. 1). The precise mechanisms that ago encountered that host or a sufficiently similar host. give rise to, and in a sense render inevitable, the wide range From that point on, sufficient conditions may be met for of plastic behaviours, morphologies and physiologies exhib- retaining the genetic changes resulting from that past ited by organisms are discussed in detail by West-Eberhard interaction, such that the contemporary possesses a

2008 Blackwell Publishing Ltd/CNRS Idea and Perspective Ecological fitting 1127

Phenotypic plasticity Correlated evolution Phylogenetic conservatism, fitting of traits inertia, and constraint of ecological

Mechanistic basis Potential’ fitness in novel conditions ( sloppy fitness space’) of Raw fitting material ecological

Genotype encounters novel conditions

Realized fitness in novel conditions Process of Population establishes in novel conditions

Existence of species in Existence of novel species Figure 1 Flow-diagram outlining the major novel environments associations components of ecological fitting from its mechanistic basis in factors that give rise to the overall flexibility of phenotypes in Novel selective

dealing with variable environments, to its Products of environments for evolution to potentially act on ecological and potentially evolutionary end- ecological fitting products. See text for details. useful response to that host when it or something similar is from the essential unity of these processes (Brooks & encountered in the future. McLennan 2002). Whether viewed through an ecological or Together, phenotypic plasticity, correlated trait evolution, an evolutionary lens, for ecological fitting or exaptation to and phylogenetic conservatism broadly defined provide the occur, individual organisms must at some point exploit raw material (sloppy fitness space) for ecological fitting to novel environmental conditions with the traits that they occur (Fig. 1). They can all be viewed as contributing to the carry at the moment of contact. overall flexibility of individual organisms in dealing with The remainder of this paper focuses on the process of novel environmental conditions. Furthermore, they can all ecological fitting, its implications for how we view ecological be seen as proximate mechanisms of exaptation. communities and species associations, and its potential Exaptation – the co-option of traits for novel functions evolutionary consequences. We elaborate on the process over evolutionary time scales – is a widely accepted and shown in Fig. 1, which depicts individuals using sloppy historically important concept in evolutionary biology fitness space to obtain realized fitness in a novel condition. (Gould & Vrba 1982). A trait that has been exapted for a This leads to the persistence of species in novel environ- novel function (e.g. Armbruster 1996, 1997) is merely the ments and results in novel species associations. In this way endpoint of a process that must have begun with an initial ecological fitting, as a result of JanzenÕs robust genotypes, case of ecological fitting, as an evolved genotype came into circumvents adaptive evolution to produce ecological contact with a novel environmental condition (Brooks & novelty, which can then, potentially, lead to adaptive McLennan 2002). Compared to some of the most well- evolution. known cases of exaptation (e.g. feathered terrestrial dino- saurs), cases of ecological fitting as we have described it are A CONCEPTUAL MODEL much less likely to leave discrete historical signals in, for example, the fossil record. However, this difference between To develop a conceptual model of ecological fitting, we first ecological and evolutionary time scales should not distract explicitly define the term ÔenvironmentÕ as it is relevant to

2008 Blackwell Publishing Ltd/CNRS 1128 S. J. Agosta and J. A. Klemens Idea and Perspective

(a) Geographic area 1 Geographic area 2 (host species 1) (host species 2)

OV2 OV2 Sloppy fitness space

a a

b b

x y x y OV1 OV1

(b) Geographic area 2 or time 2 Geographic area 1 or time 1 (host species 2) OV2 (host species 1) OV 2 c

a d

b

x y wz OV1 OV1

Figure 2 The left panel in (a) is a simple graphical representation of the interaction between a phenotype and its environment in terms of its potential to achieve realized fitness. In this case, the environment is defined by two operational variables (OVs) and the box defines the combined subset of these variables that comprises the speciesÕ ancestral operative environment (OE). Letters on the axes indicate the range of values over which the OVs occur, which define the OE. The shaded ellipse represents the fitness isocline, i.e. those combinations of OVs where realized fitness can be achieved. That portion of the fitness isocline which lies outside the ancestral OE represents the potential to achieve realized fitness under some combination of OVs never before encountered in the evolutionary history of the species. We refer to this portion of the fitness isocline as Ôsloppy fitness spaceÕ. The entire figure is a graphical representation of (a) ecological resource tracking and (b) ecological fitting. In both cases, the graphs preceding the arrow represent initial conditions. The arrow represents either dispersal to a new area (a and b), rapidly altered environmental conditions without dispersal (b), or the encountering of a novel host species by a parasite (a and b). The graph following the arrow represents the set of conditions following dispersal to a new area (a and b), rapidly altered environmental conditions (b), or the encountering of a novel host species (a and b), respectively. See text for further details. ecological fitting and explicitly define what we mean by Ôoperative environmentÕ (Spomer 1973; Dunham et al. 1989; Ôspecies traitsÕ in terms of the environment. We posit that a Dunham 1993; Dunham & Overall 1994). With its focus on complete description of any organism that bridges evolu- vital rates of the population, the concept of operative tionary and ecological perspectives can be represented as a environment (OE) derives from the concept of the fitness space mapped onto a set of environmental variables. Hutchinsonian niche (Hutchinson 1957; Pulliam 2000). We then illustrate ecological fitting in relation to our Translating the OE to the biogeographic context, we will say definitions of environment, species traits and fitness space. that it is the set of biotic and abiotic variables that will determine whether a species is present or absent, rare or common in any particular area. An explicit definition of environment ÔEnvironmentÕ is a term that has been variously defined by An explicit definition of species traits biologists working at different levels of organization (Peters 1991). For our purposes, we define environment as In the context of the OE, the most appropriate definition of consisting of only those variables that an organism perceives species trait is any intrinsic property of an organism that in an evolutionarily relevant way, which is to say those interacts with the OE to determine growth, reproduction variables that through their effect on the individual affect and survival of the individual (Violle et al. 2007), thereby the vital rates of the population: birth, death and migration. determining the vital rates of the population. By this Collectively, these environmental variables comprise the definition, phenomena such as geographic range, abundance,

2008 Blackwell Publishing Ltd/CNRS Idea and Perspective Ecological fitting 1129 diet breadth, number of interactions with other species, etc., in assembling communities and the associations among are not species traits per se but higher level descriptors that species, we will use this model to explore two non-mutually arise from species trait-by-OE interactions. This definition exclusive cases of ecological fitting: (1) colonization of new serves to integrate evolutionary and ecological perspectives, areas and (2) formation of new species associations. as only properties that affect the vital rates of the population are available to be acted upon by natural selection. For AGRAPHICALMODEL example, the geographic range of a species cannot evolve, but the traits that interact with the OE to determine The left panel of Fig. 2a is a simple way to visualize the geographic range can. interaction between a phenotype or set of phenotypes and the environment in our conceptual framework. The axes represent two OVs and define a vector space of possible Fitness space and fitness isocline environmental conditions, the solid box represents the Natural selection acts on phenotypes, which for our available OE space in the current environment, and the purposes is simply the set of all species traits weighted by shaded ellipse represents the phenotypeÕs fitness isocline their relative abundances within a population. Likewise, OE (k ‡ 1). The shape and breadth of the fitness isocline are variables (OVs) are spatially inter- and autocorrelated and arbitrary. We choose an ellipse for simplicity. More occur with a particular frequency distribution across the importantly, if we consider the OE space defined by the physical landscape of the earth. Therefore, we may consider box to represent the ancestral OE, then the fitness isocline an OE to be the subset (relative to the global maxima and is, at least in part, a direct product of evolution in this OE. minima) of values for all OVs occurring over any arbitrarily The fitness isocline in the left panel of Fig. 2a extends selected area of physical landscape, weighted by the relative beyond the bounds of the OE space in the ancestral abundance of those OVs within that area. Mapping the environment. This overlap indicates that this genotype phenotype onto the OE gives three-dimensional Ôfitness possesses sloppy fitness space in that it has fitness over a spaceÕ, the distribution of fitnesses for a particular pheno- wider range of environmental variation than it encounters in type or set of phenotypes across the OE. Or, working the ancestral environment. backwards, all components of the equation whose output is Figure 2 as a whole is a graphical model of ecological fitness are either the OVs that comprise the OE or the fitting. Using this graphical model, we can visualize what species traits that comprise the phenotype. ecological fitting might look like in our conceptual From this point onwards, we concern ourselves only with framework. We recognize two distinct scenarios that might the potential for a species to be present or absent in some give rise to ecological fitting and consider a colonization and portion of global OE space and not with questions of species association case for each. The first scenario is the relative abundance. Thus we consider only the veil line of simplest case of ecological fitting and is equivalent to fitness space, or the Ôfitness isoclineÕ where the population ecological resource tracking. growth rate, k, is 1 (a concept roughly equivalent to the zero-net-growth isoclines of resource-ratio models: Tilman Scenario 1. Ecological resource tracking 1982; Chase & Leibold 2003). Whether a species is potentially present or absent in any particular area boils The first scenario we consider is the situation where some down to one question: how does its fitness isocline map new area or resource (e.g. host species) presents identical onto the available OE space (e.g. Fig. 2)? OE space to the ancestral OE (Fig. 2a). We therefore We consider a species ancestral OE to be defined by the consider this the simplest case of ecological fitting, which is ranges and frequency distributions of OVs that exist within equivalent to resource tracking in ecological time stricto sensu. its ancestral range. We expect that the fitness isocline of a Some plant community ecologists might also term this species will reflect the OE under which it evolved, but do scenario Ôecological sortingÕ, which is a concept used to not expect it to be perfectly optimized to this OE due to describe how species within plant communities align historical and genetic constraints. Ecological fitting is themselves along edaphic gradients (Ackerly 2003). We therefore a process that takes place when an organism, include this scenario in our graphical model for complete- after some dispersal event or some rapidly altered environ- ness and because, particularly from the perspective of mental condition, comes to rest in some part of physical species associations, it illustrates how resource tracking can space that is outside of its speciesÕ ancestral OE. At this lead to patterns similar to those produced by adaptation and point, it is clear that one of two things will happen: the . organism will either ÔfitÕ into this novel environment, that is, Colonization by ecological resource tracking. This is the case of it will have realized fitness, or it will not. Because we are colonization of new areas by using more of the same OE largely concerned with the role of ecological fitting space. The species evolves in geographic area 1 under the

2008 Blackwell Publishing Ltd/CNRS 1130 S. J. Agosta and J. A. Klemens Idea and Perspective

OE space defined by the box in Fig. 2a, and then expands novel use. Again, the species may appear to be adapted to into geographic area 2, which contains identical OE space. this environment when in fact it is ecologically fit. If, for example, daily variance in temperature and Formation of new species associations by ecological fitting. moisture were the two OVs, areas 1 and 2 would have Referring again to the parasite above, it now forms a novel effectively the same thermal and soil moisture regimes. association with host species 2. However, rather than simply Formation of new species associations by ecological resource tracking more of the same OE space, it is able to form an tracking. Alternatively, we can interpret Fig. 2a in terms of association with a novel host that provides novel OE space the formation of species associations. For example, if the because of the pre-requisite traits (sloppy fitness space) it fitness isocline represents that of some parasite, and we carries at the time. Once again, the process depicted in substitute Ôhost speciesÕ for Ôgeographic areaÕ, then host Fig. 2b can lead to observed species associations that appear species 1 represents the ancestral host and host species 2 adapted but are nonetheless a product of ecological fitting. represents a novel host. From the perspective of a parasite, In summary, the processes of ecological resource tracking a host is a package of resources guarded by a certain suite and ecological fitting, and their differences, can be easily of defences. In this case, the parasite evolves in response visualized using this graphical framework, which is bivariate to a set of host conditions defined by the OE space and therefore greatly oversimplified. It is important to provided by host species 1. Upon encountering the novel realize that (1) real OEs are multivariate and (2) ecological host species 2, this parasite simply utilizes more of the resource tracking and ecological fitting can be non-mutually same OE space packaged in the form of a different exclusive processes. Agosta (2006) offered that many species. This case, although fundamentally identical to instances of host shifts by herbivorous insects can be colonizing new areas with more of the same OE space, is interpreted as examples of ecological fitting, but acknow- not trivial because it leads to patterns that are superficially ledged that these host shifts tend to occur onto taxonom- indistinguishable from those produced by evolution. As ically related plants likely because they tend to possess parasites track resources but biologists tend to track similar oviposition stimulants, nutrients or defences to be species (Brooks & McLennan 2002), what looks like a case overcome (i.e. host shifts tend to involve resource tracking). of increased diet breadth (i.e. the addition of host In fact, these arguments are not contradictory. For example, species 2) is actually a case of the parasite using more of even if a novel host species provides identical nutrients for the same OE space. Although existing fitness space has larval growth and development, it likely differs from the been co-opted to form a novel species association, from original host in many other ways (e.g. plant architecture, the perspective of the parasite, it is simply doing what it microclimate, phenology, abundance, spatial distribution, has always done. suite of heterospecific , suite of natural enemies, etc.). Thus, it seems likely that a host shift, especially one involving dispersal to a new area, often involves ecological Scenario 2. Ecological fitting resource tracking (Fig. 2a) and ecological fitting (Fig. 2b). The second scenario which we consider is the situation Returning to our example of the butterfly E. editha where some new area or host species presents OE space incorporating a novel host into its diet via ecological fitting that is outside the OE space encountered in the ancestral (Thomas et al. 1987), the old question of how successful area or host (Fig. 2b). This scenario is what we define as host shifts occur can be recast in terms of similarities (strict ecological fitting proper. ecological resource tracking) and differences (ecological Colonization by ecological fitting. In this case, organisms can fitting proper) between host-specific OEs. When apparently either physically move from areas 1 to 2 or can remain in the novel OE space has been invaded via ecological fitting, the same area but respond to rapidly altered environmental next questions are by what mechanisms has realized fitness conditions (times 1 to 2). In either case, the organisms been achieved and what, if any, evolutionary responses have encounter some set of novel environmental conditions. That ensued (e.g. adaptation to the novel host). In principal, the is, the available OE space in the new environment is same questions apply when organisms colonize new areas. different than the OE space in the species ancestral environment. Now organisms will persist only if some FURTHER ECOLOGICAL APPLICATIONS portion of the fitness isocline overlaps with some portion of the novel OE space, i.e. only if realized fitness is obtained in We offer that our conception of Dan JanzenÕs original the novel environment. The fitness isocline in Fig. 2b ecological fitting provides one particular framework for permits persistence because some portion of it overlaps with what should be a pluralist approach to studying species some portion of OE space in the novel environment. This distributions and community ecology. At this point, we portion of the fitness isocline is a by-product of history and realize this framework is largely conceptual. It will be very selection in the ancestral OE and it has been co-opted for difficult, for example, to wholly measure and compare real

2008 Blackwell Publishing Ltd/CNRS Idea and Perspective Ecological fitting 1131

OEs, just as it has been to measure a Hutchinsonian or organisms (in addition to explicitly evolutionary processes Grinellian niche (Chase & Leibold 2003). Nonetheless, we such as adaptation). feel that the framework developed here offers some useful Finally, we note the growing fields of community and insights into some of ecologyÕs most pertinent topics, ecosystem , which seek to understand how standing including speciesÕ responses to climate change, biological genetic variation and novel genetic variation introduced via invasions and introductions, and the resulting assembly and dispersal and colonization influence community dynamics dynamics of novel ecological interactions and communities. and ecosystem functioning (Urban et al. 2008; Whitham For instance, we discussed earlier the Ôescape hypothesisÕ to et al. 2008). Ecological fitting not only acts as a mechanism explain successful biological invasions, which posits that for the introduction of novel genetic variation within release from coevolved parasites and pathogens is respon- communities, but also has implications for how other sible for success in the invaded habitat (Mitchell & Power organisms already in the community may respond to this 2003; Torchin et al. 2003). In light of the OE perspective novelty and the community and ecosystem dynamics that developed here, this line of inquiry can produce only an may ensue. incomplete understanding of invasions because it imposes an a priori, top–down approach to defining an organisms EVOLUTIONARY IMPLICATIONS selective environment. In addition, it ignores a role for ecological fitting and the mechanisms by which it occurs. Ecological fitting is not, of course, an endpoint. For Alternatively, the overarching question defined in this example, although many instances of host shifts by native paper in the case of biological invasions is: to what degree insect herbivores onto introduced plants can be interpreted does the former OE differ from the invaded OE, and if they as instances of ecological fitting (Agosta 2006), these host do differ, by what mechanisms is realized fitness achieved in shifts can lead to population divergence and adaptation to the invaded OE? In the aggregate, we want to know the the novel hosts (Strauss et al. 2006). Returning to the degree to which the invasion of new habitats or formation butterfly example given earlier, Thomas et al. (1987) of new species associations involves cases of simple showed that after the initial incorporation of the novel resource tracking, cases of ecological fitting proper (i.e. host plant into their diets, some populations actually the process outlined in Fig. 1) or some likely combination of evolved a preference for this novel host in the last both. To accomplish this task, a more integrative, bottom– 100 years. Therefore, although evolution is not a pre- up approach to defining speciesÕ selective environments requisite for the formation of novel species associations or (OEs) is required. This approach would enable more robust colonizations by introduced species, one outcome of predictions of future speciesÕ responses to environmental ecological fitting is that it can be a precursor to change, which to date rely in large part on correlational evolutionary diversity. As a process whereby organisms habitat modelling (Kearney 2006). Predicting future distri- invade novel environments, use novel resources or form butions and abundance in response to environmental novel associations with other species, ecological fitting can change based solely on the overlap between current place existing genotypes into novel selective environments. distributions and a few environmental variables (1) suffers For example, in Fig. 3, a species invades novel OE space from the same a priori, top–down approach to defining and persists by ecological fitting. However, as opposed to selective environments as imposing the Ôescape hypothesisÕ Fig. 4, the population is relatively isolated (i.e. negligible in the study of invasions, and similarly (2) ignores the gene flow) and responds to directional selection in the new potential for ecological fitting by phenotypically flexible OE.

Novel operative environment Novel operative environment Ancestral operative environment time = 0 time > 0 OV2 OV2 c OV2c

a d d

b

x y w z w z OV 1 OV1 OV1

Figure 3 Graphical representation of ecological fitting as a precursor to evolutionary diversity. From left to right, a species evolves a fitness isocline under a set of conditions in the ancestral OE, then uses sloppy fitness space (see Fig. 2a) to establish a population in some novel OE, and then, because it is sufficiently isolated from other populations, evolves in response to the novel OE (i.e. its fitness isocline changes in response to selection in the novel OE). See text for further details.

2008 Blackwell Publishing Ltd/CNRS 1132 S. J. Agosta and J. A. Klemens Idea and Perspective

OV2 OV2 c c

d d

q r w z OV2 Ancestral operative environment OV1 OV1

a

b

x y OV1 OV2 OV2

e e

f f

s t w z OV1 OV1

Figure 4 Graphical representation of ecological fitting as a promoter of evolutionary stasis. The central graph represents a population experiencing the ancestral OE. The peripheral graphs represent populations experiencing novel OEs. Arrows indicate significant levels of gene flow. Because each population is experiencing a different selective environment but is connected to other populations via gene flow, the species experiences heterogeneous and possibly contradictory selection pressures that promote evolutionary stasis (i.e. there is little to no change in the fitness isoclines). See text for further details.

Alternatively, ecological fitting can promote and maintain We have also posited that phenotypic plasticity is a primary evolutionary stasis if it leads to many populations occupying mechanism behind ecological fitting. Modern ideas regarding different OEs connected by sufficient amounts of gene the role of phenotypic plasticity in producing ecological and flow. In Fig. 4, what was once an initially isolated evolutionary novelty advanced by West-Eberhard (1989, population experiencing a relatively homogeneous OE 2003) and others (Robinson & Dukas 1999; Gorur 2004; Yeh now, via ecological fitting, is many populations experiencing & Price 2004; Fordyce 2006; Janz et al. 2006; Weingartner heterogeneous OEs. In this case, as long as sufficient gene et al. 2006; Nylin & Wahlberg 2008) are integral to the flow is maintained, ecological fitting is a process that concept of ecological fitting. Phenotypic plasticity will be a promotes evolutionary stasis by opposing drift and local major determinant of the shape and breadth of the fitness adaptation. This scenario is one meaning of JanzenÕs (1985) isocline, and moreover, along with correlated trait evolution Ôevolutionary quiescenceÕ and partly what he related to and phylogenetic constraints, will determine the amount and Gould & EldredgeÕs (1977) ideas on stasis in the fossil direction of sloppy fitness space organisms possess. record. Ecological fitting is facilitated by flexible genotypes – it is the outcome of organisms possessing greater flexibility in the environments, resources and other species that they can CONCLUSIONS utilize for survival and reproduction than if mutation and We have argued that ecological fitting plays a fundamental changes in genes frequency were a pre-requisite for escaping role in structuring ecological communities to the extent that the limited context under which species evolve. Because it it is a mechanism for the formation of species associations, can place existing genotypes into novel selective environ- species introductions and invasions, and community assem- ments, ecological fitting by phenotypically flexible organ- bly in general (Janzen 1985; Brooks & McLennan 2002; isms can play an important role in the origin and evolution Wilkinson 2004; Brooks & Ferrao 2005; Agosta 2006; of biological diversity. Brooks et al. 2006). Using a graphical framework, we have illustrated (1) the process of ecological fitting, (2) how the ACKNOWLEDGEMENTS process differs from strict ecological resource tracking, (3) how the process can produce patterns superficially indistin- We especially thank Dan Janzen who first started our guishable from adaptation and (4) how the process can be thinking about ecological fitting. We also thank Dan Brooks both a promoter of evolutionary stasis and a precursor to and Art Dunham for fostering many of the ideas presented evolutionary diversity. in this paper. We also thank Kellie Kuhn and Niklas Janz,

2008 Blackwell Publishing Ltd/CNRS Idea and Perspective Ecological fitting 1133 who independently recommended consultation of Chenuil, A. & McKey, D.B. (1996). Molecular phylogenetic study M.J. West-EberhardÕs recent book on phenotypic plasticity of a myrmecophyte symbiosis: did Leonardoxa ⁄ ant associations and evolution. We acknowledge that none of them may diversify via cospeciation? Mol. Phylogenet. Evol., 6, 270–286. agree with everything we have said. Any faults and omissions Dunham, A.E. (1993). Population responses to global change: physiological structured models, operative environments, and are our own. While writing this paper, SJA was supported by population dynamics. In: Biotic Interactions and Global Change (eds a GIAR grant from Sigma Xi, the Binns-Williams Fund Karieva, P., Kingsolver, J. & Huey, R.). Sinauer Associates, from the University of Pennsylvania and NSF Doctoral Sunderland, pp. 95–119. Dissertation Improvement Grant DEB 0508573. JAK was Dunham, A.E. & Overall, K.L. (1994). Population responses to supported by an NSF Post-doctoral Fellowship in Microbial environmental change: life history variation, individual-based Biology Grant No. 0400833. Significant portions of this models, and the population dynamics of short-lived organisms. Am. Zool. paper were conceived and written while both authors were , 34, 382–396. ´ Dunham, A.E., Grant, B.W. & Overall, K.L. (1989). The interface living and working in the Area de Conservacio´n Guanacaste between biophysical ecology and the population ecology of (ACG), Costa Rica. We acknowledge the entire staff of the terrestrial vertebrate ectotherms. Physiol. Zool., 62, 335–355. ACG for their continued logistic support. Finally, we thank Flowers, R.W. & Janzen, D.H. (1997). Feeding records of Costa David Wilkinson and two anonymous referees for com- Rican leaf beetles (Coleoptera: Chrysomelidae). Fla Entomol., 80, ments and suggestions that significantly improved the final 334–366. version of the manuscript. Fordyce, J.A. (2006). The evolutionary consequences of ecological interactions mediated through phenotypic plasticity. J. Exp. Biol., 209, 2377–2383. REFERENCES Futuyma, D.J. & Mayer, G.C. (1980). Non-allopatric speciation in animals. Syst. Zool., 29, 254–271. Ackerly, D. (2003). Community assembly, niche conservatism, and Gill, F.B. (1987). Ecological fitting – use of floral nectar in adaptive evolution in changing environments. Int. J. Plant Sci., stilesii Daniels by 3 species of hermit hummingbirds. Condor, 89, 164, S165–S184. 779–787. Agosta, S.J. (2006). On ecological fitting, plant-insect associations, Gorur, G. (2004). The importance of phenotypic plasticity in herbivore host shifts, and host plant selection. Oikos, 114, 556– herbivorous insect speciation. In: Insects and Phenotypic Plasticity 565. (eds Whitman, D. & Ananthakrishnan, T.N.). Science Publish- Agosta, S.J. & Janzen, D.H. (2005). Body size distributions of ers, Enfield, pp. 145–171. large Costa Rican dry forest moths and the underlying rela- Gould, S.J. & Eldredge, N. (1977). Punctuated equilibria: the tionship between plant and pollinator morphology. Oikos, 108, tempo and mode of evolution reconsidered. Paleobiology, 3, 115– 183–193. 151. Armbruster, W.S. (1996). Exaptation, adaptation, and homoplasy: Gould, S.J. & Vrba, E.S. (1982). Exaptation—a missing term in the evolution of ecological traits in Dalechampia vines. In: Homoplasy: science of form. Paleobiology, 8, 4–15. The Recurrence of Similarity in Evolution (eds Sanderson, M.J. & Herrera, C.M., Medrano, M., Rey, P.J., Sa´nchez-Lafuente, A.M., Hufford, L.). Academic Press, New York, pp. 227–243. Garcia, M.B., Guitia´n, J. et al. (2002). Interaction of pollinators Armbruster, W.S. (1997). link evolution of plant-her- and herbivores on plant fitness suggests a pathway for correlated bivore and plant-pollinator interactions: a phylogenetic inquiry. evolution of - and antagonism-related traits. Proc. Natl Ecology, 78, 1661–1672. Acad. Sci. U.S.A., 99, 16823–16828. Baldwin, J.M. (1896). A new factor in evolution. Am. Nat., 30, 441– Holmes, J.C. & Price, P.W. (1980). Parasite communities: the roles 451. of phylogeny and ecology. Syst. Zool., 29, 203–213. Bradshaw, A.D. (1965). Evolutionary significance of phenotypic Holway, D.A., Lach, L., Suarez, A.V., Tsutsui, N.D. & Case, T.J. plasticity in plants. Adv. Genet., 13, 115–155. (2002). The causes and consequences of ant invasions. Ann. Rev. Brooks, D.R. (1985). Historical ecology: a new approach to Ecol. Syst., 33, 181–233. studying the evolution of ecological associations. Ann. Mo. Bot. Hubbell, S.P. (2001). The Unified Neutral Theory of Biodiversity and Gard., 72, 660–680. Biogeography. Princeton University Press, Princeton. Brooks, D.R. & Ferrao, A.L. (2005). The historical biogeography of Hurtt, G.C. & Pacala, S.W. (1995). The consequences of recruit- covevolution: emerging infectious diseases are evolutionary ment limitation: reconciling chance, history and competitive accidents waiting to happen. J. Biogeogr., 32, 1291–1299. differences between plants. J. Theor. Ecol., 176, 1–12. Brooks, D.R. & McLennan, D.A. (2002). The Nature of Diversity: An Hutchinson, G.E. (1957). Concluding remarks. Population studies: Evolutionary Voyage of Discovery. University of Chicago Press, animal ecology and demography. Cold Spring Harb. Symp. Quant. Chicago. Biol., 22, 415–427. Brooks, D.R., Leo`n-Re`gagnon, V., McLennan, D.A. & Zelmer, D. Janz, N., Nylin, S. & Wahlberg, N. (2006). Diversity begets (2006). Ecological fitting as a determinant of the community diversity: host expansions and the diversification of plant-feed- structure of platyhelminth parasites of anurans. Ecology, 87, S76– ing insects. BMC Evol. Biol.,6,4. S85. Janzen, D.H. (1980). When is it coevolution? Evolution, 34, 611– Chase, J.M. & Leibold, M.A. (2003). Ecological Niches: Linking 612. Classical and Contemporary Approaches. University of Chicago Press, Janzen, D.H. (1985). On ecological fitting. Oikos, 45, 308–310. Chicago.

2008 Blackwell Publishing Ltd/CNRS 1134 S. J. Agosta and J. A. Klemens Idea and Perspective

Janzen, D.H. (1986). Biogeography of an unexceptional place: what Strauss, S.Y., Lau, J.A. & Carroll, S.P. (2006). Evolutionary determines the saturniid and sphingid moth fauna of Santa Rosa responses of natives to introduced species: what do intro- National Park, Costa Rica, and what does it mean to conser- ductions tell us about natural communities? Ecol. Lett., 9, 357– vation biology? Brenesia,25⁄ 26, 51–87. 374. Janzen, D.H. (2003). How polyphagous are Costa Rican dry forest Suarez, A.V., Holway, D.A. & Ward, P.S. (2005). The role of saturniid caterpillars? In: of Tropical Forests: Spatio-tem- opportunity in the unintentional introduction of nonnative ants. poral Dynamics and Resource Use in the Canopy (eds Basset, Y., Proc. Natl Acad. Sci. U.S.A., 102, 17032–17035. Novotny, V., Miller, S.E. & Kitching, R.L.). Cambridge Thomas, C.D., Ng, D., Singer, M.C., Mallet, J.L.B., Parmesan, C. & University Press, Cambridge, pp. 369–379. Billington, H.L. (1987). Incorporation of a European weed into de Jong, G. (2005). Evolution of phenotypic plasticity: patterns of the diet of a North American herbivore. Evolution, 41, 892–901. plasticity and the emergence of ecotypes. New Phytol., 166, 101– Thompson, J.D. (1991). Phenotypic plasticity as a component of 118. evolutionary change. Trends Ecol. Evol., 6, 246–249. Kearney, M. (2006). Habitat, environment and niche: what are we Tilman, D. (1982). Resource Competition and Community Structure. modelling? Oikos, 115, 186–191. Princeton University Press, Princeton. Lande, R. & Arnold, S.J. (1983). The measurement of selection on Torchin, M.E., Lafferty, K.D., Dobson, A.P., McKenzie, V.J. & correlated characters. Evolution, 37, 1210–1226. Kuris, A.M. (2003). Introduced species and their missing para- Leigh, E.G. Jr, Davidar, P., Dick, C.W., Puyravaud, J., Terborgh, J., sites. Science, 421, 628–630. ter Steege, H. et al. (2004). Why do some tropical forests have so Urban, M.C., Leibold, M.A., Amarasekare, P., De Meester, L., many species of trees? Biotropica, 36, 447–473. Gomolkiewicz, R., Hochberg, M.E. et al. (2008). The evolu- MacArthur, R.H. & Wilson, E.O. (1967). The Theory of Island Bio- tionary ecology of metacommunities. Trends Ecol. Evol., 23, 311– geography. Princeton University Press, Princeton. 317. Mack, R.N. (2003). Phylogenetic constraint, absent life forms, and Violle, C., Navas, M.L., Vile, D., Kazakou, E., Fortunel, C., preadapted alien plants: a prescription for biological invasions. Hammel, I., Garnier, E. (2007). Let the concept of trait be Int. J. Plant Sci., 164, S185–S196. functional! Oikos, 116, 882–892. Mayr, E. (1963). Animal Species and Evolution. Harvard University Weingartner, E., Wahlberg, N. & Nylin, S. (2006). Dynamics of Press, Cambridge. host plant use and species diversity in Polygonia butterflies Mitchell, C.E. & Power, A.G. (2003). Release of invasive plants (Nymphalidae). J. Evol. Biol., 19, 483–491. from fungal and viral pathogens. Science, 421, 625–627. West-Eberhard, M.J. (1989). Phenotypic plasticity and the origins Murphy, S.M. & Feeny, P. (2006). Chemical facilitation of a naturally of diversity. Ann. Rev. Ecol. Syst., 20, 249–278. occurring host shift by Papilio machaon butterflies (Papilionidae). West-Eberhard, M.J. (2003). Developmental Plasticity and Evolution. Ecol. Monogr., 76, 399–414. Oxford University Press, New York. Nylin, S. & Wahlberg, N. (2008). Does plasticity drive speciation? White, D.W. & Stiles, E.W. (1992). Bird dispersal of fruits of Host plant shifts and diversification in nymphaline butterflies species introduced into eastern North America. Can. J. Bot., 70, (Lepidoptera: Nymphalidae) during the tertiary. Biol. J. Linn. Soc., 1689–1696. 94, 115–130. Whitham, T.G., Difazio, S.P., Schweitzer, J.A., Shuster, S.M., Allan, Pennington, R.T. & Dick, C.W. (2004). The role of immigrants in G.J., Bailey, J.K. et al. (2008). Extending to natural the assembly of the South American rainforest tree flora. Philos. communities and ecosystems. Science, 320, 492–495. Trans. R. Soc. Lond., B, Biol. Sci., 359, 1611–1622. Wilkinson, D.M. (2004). The parable of Green Mountain: Ascen- Peters, R.H. (1991). A Critique for Ecology. Cambridge University sion Island, ecosystem construction and ecological fitting. Press, Cambridge. J. Biogeogr., 31, 1–4. Pulliam, H.R. (2000). On the relationship between niche and dis- Williams, G.C. (1966). Adaptation and Natural Selection: A Critique of tribution. Ecol. Lett., 3, 349–361. Some Current Evolutionary Thought. Princeton University Press, Robinson, B.W. & Dukas, R. (1999). The influence of phenotypic Princeton. modifications on evolution: the Baldwin effect and modern Yeh, P.J. & Price, T.D. (2004). Adaptive phenotypic plasticity and perspectives. Oikos, 85, 582–589. the successful colonization of a novel environment. Am. Nat., Schlichting, C.D. & Pigliucci, M. (1998). Phenotypic Evolution: A 164, 531–542. Reaction Norm Perspective. Sinauer Associates, Sunderland. Yu, D.W. & Davidson, D.W. (1997). Experimental studies of Schluter, D. & Ricklefs, R.E. (1993). Species diveristy: an intro- species-specificity in Cecropia-ant relationships. Ecol. Monogr., 67, duction to the problem. In: Species Diversity in Ecological Commu- 273–294. nities: Historical and Geographical Perspectives (eds Ricklefs, R.E. & Schluter, D.). University of Chicago Press, Chicago, pp. 1–10. Spomer, G.P. (1973). The concepts of ÔinteractionÕ and Ôoperational Editor, Tadashi Fukami environmentÕ in environmental analysis. Ecology, 54, 608–615. Manuscript received 14 March 2008 Stearns, S.C. (1989). The evolutionary significance of phenotypic First decision made 17 April 2008 plasticity. Bioscience, 39, 436–445. Second decision made 30 June 2008 Stohlgren, T.J., Barnett, D.T., Jarnevich, C.S., Flather, C. & Kartesz, J. (2008). The myth of plant species saturation. Ecol. Manuscript accepted 3 July 2008 Lett., 11, 313–326.

2008 Blackwell Publishing Ltd/CNRS