Introduction
A central goal in ecology is to determine what forces structure local communities, and to understand why community membership varies across space. According to theory, the composition of local communities reflects a balance between regional inputs, via speciation and dispersal, and local controls, such as environmental filtering and competition (MacArthur & Wilson 1967; Rosenzweig 1995; Vellend 2016; Leibold & Chase 2018). If local forces control diversity equilibria, communities in similar environments are expected to be similar in composition, regardless of the size of their regional species pools; conversely, if regional forces predominate, regions with more species should also possess greater local diversity (Ricklefs 1987). Debate over whether local or regional forces dominate has reached an uneasy détente, as macroecological patterns suggest local control in some systems, and pervasive regional influences in others (Ricklefs & Schluter 1994; Myers et al. 2013; Cornell & Harrison 2014). What is thus crucial is to identify the reasons for this variation in controls of local diversity, and to establish a framework to explain why one system’s local diversity is set by local processes, while another’s reflects inputs from the regional species pool (Cornell 1999; Harrison & Cornell 2008; Lessard et al. 2012).
We approach this challenge by focusing on a pair of linked mechanisms by which macroevolutionary inputs could ultimately determine community structure. First, we consider the attributes of biogeographic landscapes that foster (or limit) the evolution of diverse regional species pools (McPeek & Brown 2000; Swenson 2011; Zobel et al. 2011; Mittelbach & Schemske 2015; Craven et al. 2019). Second, we consider how these richer regional species pools could then lead to saturated communities (competition-structured communities where richness reaches a stable equilibrium; Rabosky & Hurlbert 2015; Storch & Okie 2019). Crucially, species pools are generated by an interplay of ecological and evolutionary factors at spatial scales intermediate to the local versus regional scales of traditional assembly models (Emerson & Gillespie 2008; Jetz & Fine 2012). The hierarchical organization of regions (e.g., large, evolutionarily-independent areas such as continents or oceanic islands) into environmentally-distinctive ”sub-regions” (e.g., biomes) may shape the evolutionary buildup of species pools, thereby influencing the structure of local communities (Qian & Ricklefs 2000; Zobel et al. 2011; Jetz & Fine 2012; Mittelbach & Schemske 2015).
If competition is relatively weak, increased diversification within a region should increase local community richness by increasing species additions relative to losses (Hubbell 2001; Leibold et al. 2004). Alternatively, if increased diversification produces species that specialize on particular sub-regions (e.g., Mayr 1947; Glor et al. 2003; Rundle & Nosil 2005; Gray et al. 2019), it may spur the emergence of local richness controls. Sub-regional specialists may better monopolize local resources, resulting in competition-structured communities that can resist migrants (MacArthur 1972; Cornell & Harrison 2014). Therefore, under this region/sub-region/local model, a principal determinant of whether local or regional effects on diversity predominate is the availability of evolutionary opportunity – a term we use to describe the conditions necessary for clades to diversify and specialize on unique local ecological conditions (e.g., sufficient time and geographic space for speciation among or within ecologically distinct sub-regions to complete; Losos & Parent 2010; Zobel et al. 2011; Cornell 2013; Algar & Mahler 2015).
If greater evolutionary opportunities indeed facilitate local control, we predict the emergence of several ecological patterns in local-community abundance, species diversity, and organization across space (beta-diversity). First, for communities to exhibit local diversity controls requires that they be saturated at the level of individuals competing for limited resources (Cornell 1999; Gaston 2000). Thus, numbers of individuals in such communities should scale with resource availability, yielding correlations between local abundance and proxies of productivity, such as temperature or elevation (Wright 1983; Gaston 2000). If individuals can utilize local resources efficiently regardless of the diversity of the evolutionary fauna, the scaling of abundance and productivity should be similar across systems that differ in evolutionary diversity. Conversely, greater diversification could fundamentally change how ecological limits manifest by generating specialists on previously unutilized resources and thereby enhancing community-level abundances (Cornell 2013; Storch et al. 2018; Storch & Okie 2019). In that case, regions with greater evolutionary opportunity may exhibit greater local abundances in otherwise-similar environments.
Second, if greater macroevolutionary inputs enhance local control on species diversity then the relationship between environmental resources and local species richness should in theory be similar across regions. However, this relationship will depend on whether clades in different regions have had sufficient opportunity to diversify into the niches available in local environments—in other words, whether they have produced sub-regional environmental specialists that can saturate communities at the species level (Jetz & Fine 2012; Cornell 2013). If a given region has had insufficient time or space for diversification to fill available sub-regional niches, we would expect local richness patterns across regions to diverge (MacArthur 1972; Mittelbach & Schemske 2015).
Finally, if evolutionary opportunity shapes communities by facilitating local control, we expect greater regional diversity to sort primarily among communities (i.e., as beta-diversity) rather than within them (elevated alpha-diversity) (MacArthur 1965; Cornell 2013). Evolutionary opportunity should thus precipitate the emergence of spatial structure within faunas through the evolution of sub-regional environmental specialists. When opportunity is substantial in environmentally similar areas, whole-community properties such as total richness and functional characteristics may converge, even among evolutionarily-independent regions (Orians & Paine 1983; Cornell 1999; Ricklefs 2006). By contrast, faunas lacking in evolutionary opportunity should consist of environmental generalists, and the local communities within these faunas should exhibit comparatively little turnover and potentially low convergence between faunas.
Testing how diversification affects local community structure is difficult, in part because of the challenge locating empirical systems that are ecologically similar, but differ in histories of macroevolutionary diversification (Lessard et al. 2012). Regions differing in evolutionary richness often also differ in climate and habitat, or the deep phylogenetic history of the biota, either of which may confound straightforward comparison (Rosenzweig 1995; Price et al. 2000). Ideally one would seek a system in which replicate regions are similar in environment, and contain faunas that radiated from recent phylogenetic stock, but differ in macroevolutionary diversity.
Here we leverage a natural experiment to examine the effects of macroevolution on the structure of local communities: independent radiations of anole lizards (Anolis ) distributed across similar environmental gradients on the Caribbean island of Jamaica and the northern paleoisland of adjacent Hispaniola (this region corresponds to a biogeographically distinctive unit within the island; Fig. 1). Anoles have diversified to play similar ecological roles on both islands, but Hispaniola has a richer anole fauna than its smaller neighbor Jamaica (Williams 1983; Losos 2009; Mahler et al. 2013). The two islands possess similar climates, vegetation types, and macrohabitat diversity (Ricklefs & Bermingham 2008; Losos 2009). Notably, both feature vast expanses of hot lowland tropical forest that grade into tall inland mountains featuring cool cloud-forest habitats. Their key environmental difference is that Hispaniola’s highlands are much more extensive than Jamaica’s, creating divergent evolutionary prospects in this sub-region type (Fig. 1b). Crucially, while the lowlands of both islands cover more than 3,000km2, the proposed minimum area required forin situ speciation in anoles (Losos & Schluter 2000), only Hispaniola’s highlands clear this threshold, leading to the expectation that Jamaica’s highland communities should be evolutionarily constrained.
We conducted mark-resample surveys of Anolis communities across matched elevational gradients on Jamaica and Hispaniola to test whether greater macroevolutionary diversity, and especially differences in highland evolutionary opportunity, trickle down to structure local communities. We begin by asking (1) whether the abundance ofAnolis within local communities across elevations is similar between islands, as predicted if local temperature is primarily responsible for controlling community size, and macroevolutionary inputs have little effect on abundance. Next, we ask (2) whether differences in sub-regional area trigger divergence in community structure across elevations, such that community convergence exists between islands in the lowlands (where evolutionary opportunity is substantial), but is lacking in the highlands (for which Jamaica is more limited). We then investigate (3) whether greater evolutionary opportunity in the Hispaniolan highlands leads to ecologically distinct sub-faunas via the evolution of local environmental specialists. If so, we predict that Hispaniolan anole communities will exhibit greater beta-diversity than their Jamaican counterparts. Finally, we (4) test the importance of sub-regional faunas in impacting convergence in abundance and richness between islands across the elevational gradient.